@literaturamedica@ CT Thorax

5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page i
Computed Tomography
and Magnetic Resonance
of the Thorax
FOURTH EDITION
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page ii
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page iii
Computed Tomography and

Magnetic Resonance
of the Thorax
FOURTH EDITION
EDITORS
DAVID P. NAIDICH, MD W. RICHARD WEBB, MD

Professor of Radiology and Medicine Professor of Radiology
NYU Medical Center University of California San Francisco
New York, New York Hideyo Minagi Professor of Radiology
San Francisco General Hospital
San Francisco, California
NESTOR L. MÜLLER, MD, PhD IOANNIS VLAHOS, MB, BS, BSC

Professor and Chairman Assistant Professor of Radiology
Department of Radiology NYU Medical Center
University of British Columbia New York, New York
Vancouver, British Columbia, Canada
CONTRIBUTING AUTHOR
GLENN A. KRINSKY, MD MONVADI B. SRICHAI, MD

Valley Hospital Assistant Professor of Radiology
Ridgewood, New Jersey NYU Medical Center
New York, New York
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page iv
Publisher: Lisa McAllister

Managing Editor: Kerry Barrett
Project Manager: Fran Gunning
Manufacturing Coordinator: Kathleen Brown
Marketing Manager: Angela Panetta
Design Coordinator: Terry Mallon
Production Services: Nesbitt Graphics, Inc.
Printer: Edwards Brothers
© 2007 by LIPPINCOTT WILLIAMS & WILKINS, a Wolters Kluwer business

530 Walnut Street
Philadelphia, PA 19106 USA
LWW.com
Third edition, © 1999 Lippincott-Raven Publishers

Second edition, © 1991 Raven Press
First edition, © 1984 Raven Press
All rights reserved. This book is protected by copyright. No part of this book may be reproduced in any form or by any means, includ-
ing photocopying, or utilized by any information storage and retrieval system without written permission from the copyright owner,
except for brief quotations embodied in critical articles and reviews. Materials appearing in this book prepared by individuals as part
of their official duties as U.S. government employees are not covered by the above-mentioned copyright.
Printed in the USA
Library of Congress Cataloging-in-Publication Data
Computed tomography and magnetic resonance of the thorax / editors, David P. Naidich . . . [et al.]; contributing author,
Monvadi B. Srichai. —4th ed.
p. ; cm.
Includes bibliographical references and index.
ISBN-13: 978-0-7817-5765-2 (alk. paper)
ISBN-10: 0-7817-5765-7 (alk. paper)
1. Chest—Tomography. 2. Chest—Magnetic resonance imaging. I. Naidich, David P. II. Srichai, Monvadi B.
[DNLM: 1. Radiography, Thoracic. 2. Magnetic Resonance Imaging. 3. Thorax—pathology. 4. Tomography, X-Ray Computed.
WF 975 C7379 2007]
RC941.N27 2007
617.5’40757—dc22
2006038568
Care has been taken to confirm the accuracy of the information presented and to describe generally accepted practices. However, the
authors, editors, and publisher are not responsible for errors or omissions or for any consequences from application of the information
in this book and make no warranty, expressed or implied, with respect to the currency, completeness, or accuracy of the contents of the
publication. Application of this information in a particular situation remains the professional responsibility of the practitioner.
The authors, editors, and publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in
accordance with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in
government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check
the package insert for each drug for any change in indications and dosage and for added warnings and precautions. This is particularly
important when the recommended agent is a new or infrequently employed drug.
Some drugs and medical devices presented in this publication have Food and Drug Administration (FDA) clearance for limited use in
restricted research settings. It is the responsibility of the health care provider to ascertain the FDA status of each drug or device planned
for use in their clinical practice.
To purchase additional copies of this book, call our customer service department at (800) 638-3030 or fax orders to (301) 223-2320.
International customers should call (301) 223-2300.
Visit Lippincott Williams & Wilkins on the Internet: at LWW.com. Lippincott Williams & Wilkins customer service representatives are
available from 8:30 am to 6 pm, EST.
10 9 8 7 6 5 4 3 2 1
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page v
To my wife Jocelyn, my son Zachary, and to my father Harry

D.P.N.
To the learners and teachers:

Cole, Andy, Sonny, Jessica, Emma, Brett, and Teresa
W.R.W.
To Alison and Phillip Müller

N.L.M
To my wife Jo, my boys Emile and Sebastian, and to my parents

I.V.
To Stacie, Emma, and Reginald

G.A.K.
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page vi
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page vii
Contents
Preface to the First Edition ix

Preface to the Fourth Edition xi
Acknowledgments xiii
1 Heart 1 6 Focal Lung Disease 557

2 Aorta, Arch Vessels, and Great Veins 87 7 Lung Cancer 621
3 Pulmonary Arterial Disease 217 8 Diffuse Lung Disease 671
4 Mediastinum 289 9 Pleura, Chest Wall, and Diaphragm 769
5 Airways 453
Subject Index 885
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page viii
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page ix
Preface to the First Edition
Diagnostic imaging has undergone a profound and aston- 1981, Dr. Zerhouni moved to the East Virginia Medical
ishingly rapid transformation over the last decade, parallel- Center. Despite this separation, the team continued their
ing the rapid evolution of modern computer science. The collaborative endeavors, and in 1982 and 1983 presented
first computed tomography (CT) unit, conceived and instructional courses in chest CT at the annual meeting of
developed by Godfrey Hounsfield, underwent initial clini- the American Roentgen Ray Society. From these presenta-
cal testing at the Atkinson Morley Hospital, Wimbledon, tions, the need for a volume representing the current status
England, in 1971. This early scanner employed two of thoracic CT became apparent.
sodium iodide crystals and produced an image based on an This textbook has been organized primarily around the
80 80 matrix. The scan time of four and one half major anatomic subunits of the thorax. These include: the
minutes effectively limited the machine to examination of mediastinum, the airways, the hila, the pulmonary
intracranial pathology. parenchyma, the pleura, and the chest wall, the peri-
Within three years, Ledley developed a CT unit capable cardium, and the diaphragm. Additional chapters have been
of imaging the body. In 1974 and 1975, whole body scan- added specifically on the role of CT in evaluating lobar
ner prototypes were installed first at the Cleveland Clinic, collapse and the pulmonary nodule, as these represent dis-
then at the Malinckrodt Institute of Radiology and the crete topics best addressed apart. This organizational
Mayo Clinic. Initial reports from these institutions were scheme represents the authors’ view that CT is primarily an
enthusiastic about the role of CT in the evaluation of anatomic imaging modality. Specifically excluded from con-
the pancreas, liver, and retroperitoneum, but pessimistic sideration is the use of CT in evaluating the heart. It is only
about the value of CT in the thorax. appropriate to consider cardiac CT in comparison to other
Further improvements in instrumentation were neces- cardiac imaging modalities, including angiography, echocar-
sary in order to assess the clinical role of thoracic CT. The diography, and nuclear cardiology. It is the authors’ feeling
pace of technological innovation was such that by 1977, that this is outside the intended scope of the present volume
scanners capable of scan times shorter than breath-holding as initially conceived.
had been developed. Additional improvements, including Thoracic CT has become an integral part of the daily
more detectors to increase resolution and finer collimation practice of radiology. With the ever-increasing number of
allowing a reduction in slice thickness to minimize partial diagnostic modalities, the task of deciding which diagnos-
volume averaging, soon became standard. As a result, tic test is the most appropriate for a given clinical problem
initial pessimism about the role of CT in the thorax has become a significant part of medical practice. The
quickly gave way to considerable enthusiasm. This first be- authors believe that an adequate understanding of clinical
came apparent as reports of the value of CT in analyzing issues is necessary for practicing radiologists to best help
mediastinal disease were published. Thereafter, an ever- the referring physician. Consequently, throughout the text,
expanding range of uses for thoracic CT has evolved and a strong emphasis has been placed on discussing CT as it
continues to evolve. relates to clinical issues, especially as compared to other
In December 1977, a CT scanner was installed at the routine imaging modalities. It is to be anticipated that
Johns Hopkins Hospital. At that time, Drs. Naidich further technologic advances in diagnostic imaging will
and Zerhouni were residents under the tutelage of further complicate the role of radiologists. It is hoped that
Dr. Siegelman, under whose auspices the three authors this text will prove valuable in assisting this process.
of this volume enthusiastically began a series of studies
concerning the utility of thoracic CT in a wide range of David P. Naidich
clinical settings. In July 1980, Dr. Naidich joined the staff Elias A. Zerhouni
at New York University Medical Center, and in January Stanley S. Siegelman
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page x
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page xi
Preface to the Fourth Edition
Computed Tomography and Magnetic Resonance of the Thorax thoracic imaging. To meet this challenge, we have elimi-
was first published 21 years ago; it has been 7 years since nated an introductory chapter, instead reviewing technical
the publication of the third edition. In retrospect it is truly aspects of CT, magnetic resonance (MR), and PET scanning
amazing how much has changed in the field of thoracic as individually pertinent to specific chapters. To facilitate
imaging over the course of these years. Initially restricted this approach, the first three chapters have been rearranged
to computed tomography (CT) evaluation of the chest to cover the heart, aorta, and the pulmonary arteries as
without consideration of alternate imaging methodolo- a group to form a single coherent unit. As before, the
gies, the book, in its current edition, attempts to meet the remainder of the text remains organized anatomically,
formidable challenge of keeping pace with the ever more including chapters devoted to the mediastinum, airways,
rapidly expanding technical innovations that are taking lungs, and pleura and chest wall, again reflecting our basic
place in the development of both hardware and software anatomic bias to differential diagnosis. In addition, sepa-
devices. These include, among others, the availability of rate chapters are once more targeted to the evaluation of
Picture Archiving and Communications Systems (PACS), focal lung disease and lung cancer. In distinction, a previ-
as well remarkable advances in methods of image process- ous chapter devoted to pulmonary complications of AIDS
ing, including the growing field of computer-assisted has been deleted, reflecting the considerable advances that
diagnosis (CAD). Needless to say, this expansion in tech- have occurred in the treatment of this disease. We have also
nology has required substantial modifications in the made the judgment, however regretfully, to exclude any
organization of the text. number of newer promising techniques, including, for
Most important has been the inclusion of sections example, the expanding field of functional imaging of the
specifically highlighting cardiovascular imaging. With the lungs, in order to keep this edition as a single volume
introduction and now widespread availability of CT scan- devoted to in-depth analyses of current clinically relevant
ners capable of sophisticated cardiac imaging, the rationale topics rather than superficial overviews.
for excluding this topic can no longer be justified. Similar As previously noted in the Preface to the Third Edition,
considerations apply to the need for evaluating the impact it is the authors’ continuing belief that accurate radiologic
of positron emission tomography (PET) scanning in the interpretation requires in-depth familiarity with the clini-
clinical management of chest disease, again reflecting the cal context in which we practice rather than simple pattern
rapidly growing influence of this technology. recognition. Although this remains a truly daunting task, it
Despite these changes, the overall intent of this text re- is our hope that the current volume will again prove of
mains the same: specifically, the authors’ desire to produce value to all involved with clinical assessment of the entire
a single accessible volume devoted primarily to clinical gamut of diseases of the thorax.
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page xii
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page xiii
Acknowledgments
As is always the case with an undertaking as potentially Cardona and Warren Hendricks at Tisch Hospital; and
overwhelming as the fourth edition of Computed Tomography John Sebellico and Phil Fontana at Bellevue Hospital.
and Magnetic Resonance of the Thorax, the authors are truly We also extend our deeply felt thanks again to Martha
indebted to the many individuals without whose support Helmers and Tony Jalandoni, whose photographic talents
this text would never have seen the light of day. and tireless dedication to perfection in no small way
At New York University we are particularly fortunate to account for the value of this text.
work with a number of dedicated technologists who have Finally, it is our true pleasure to acknowledge the count-
demonstrated past and present willingness to oblige our less contributions, large and small, made by our former and
incessant demands: Tania Yeargin, Nelly Patino, Bernard present residents and fellows whose constant curiosity, skep-
Assadourin, and Emilio Vega at our Faculty Practice; Elba ticism, and questioning have made us all better physicians.
5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page xiv
5636_Naidich_ch01_pp001-086 12/6/06 5:13 PM Page 1
Heart 1
IMAGING TECHNIQUES 2 Tricuspid Valve 58

Computed Tomography 2 Prosthetic Valves 59
Magnetic Resonance Imaging 6
CONGENITAL HEART DISEASE 59
NORMAL ANATOMY 18 Bicuspid Aortic Valve 61
Cardiac Chambers 18 Atrial Septal Defect 62
Cardiac Valves 21 Ventricular Septal Defect 62
Coronary Arteries and Veins 23 Anomalous Pulmonary Venous Connection 62
Pulmonary Arteries and Veins 24 Ebstein Anomaly 64
Pericardium and Myocardium 24 Patent Ductus Arteriosus 64
Coarctation of the Aorta 64
ISCHEMIC HEART DISEASE 24 Tetralogy of Fallot 64
Coronary Artery Imaging 25 Complete Transposition of the Great Arteries 67
Functional Myocardial Imaging 27 Congenitally Corrected Transposition of the
Special Considerations 31 Great Arteries 69
Coronary Artery Anomaly 70
OTHER MYOCARDIAL DISEASE 34
Hypertrophic Cardiomyopathy 34 OTHER APPLICATIONS 74
Dilated Cardiomyopathy 36 Left Atrial Anatomy 74
Restrictive Cardiomyopathy 37 Coronary Venous Anatomy 75
Myocarditis 40
PERICARDIAL DISEASE 40 Cardiac disease is one of the leading causes of morbidity

Pericardial Effusion 41 and mortality in the developed world. Unlike other organs
Acute Pericarditis 42 of the body, the heart is constantly subject to both cardiac
Constrictive Pericarditis 42 and respiratory motion within the thoracic cavity. Addi-
Pericardial Masses 42 tionally, there is a constant flow of blood through the
Congenital Pericardial Lesions 44 cardiac structures that presents unique technical and diag-
nostic challenges for imaging. Depending on the disease
CARDIAC MASSES 45 process, cardiac imaging evaluation has focused on the use
Benign Cardiac Neoplasms 45 of echocardiography, nuclear cardiac imaging techniques,
Malignant Cardiac Neoplasms 50 and invasive coronary angiography. Technologic advances
Pseudotumors 53 in magnetic resonance (MR) imaging and computed
tomography (CT) have led to an explosion of interest in
VALVULAR HEART DISEASE 54 these new approaches to image the heart and vascular
Aortic Valve 55 structures. We can now visualize in great detail cardiac
Mitral Valve 56 structure, motion, and function important for clinical
Pulmonary Valve 57 diagnosis, treatment, and management of cardiac disease.
2 Computed Tomography and Magnetic Resonance of the Thorax
IMAGING TECHNIQUES motion. However, variables including collimation, pitch,

breath-hold period, field of view, reconstruction interval,
rate and volume of intravenous contrast administration,
Computed Tomography reconstruction algorithm, and radiation dose need to be
With the advent of multidetector scanners in 1998, cardiac addressed. Unfortunately, no single set of optimal scan
CT imaging has become one of the fastest growing areas of parameters currently suffices for all cardiac CT applications
major interest. Prior to multidetector scanners, electron beam on all CT systems. Instead, CT techniques often reflect a
computed tomography (EBCT) had been in use for cross- combination of personal or institutional preferences cou-
sectional CT imaging of the heart. With its high temporal res- pled with individual manufacturer capabilities. Table 1-1
olution of 50 to 100 ms, coronary artery imaging, including lists general MDCT guidelines used at New York University
calcium scoring and functional assessment of the heart, has Medical Center for cardiac CT imaging.
been its focus. With the technical evolution of multidetector
CT scanners (MDCT) as well as their widespread availability,
Technical Design
however, much of the focus in recent years has shifted to the
replacement of EBCT with MDCT for cardiac CT imaging. The gantry geometry of all MDCT generations is essentially
Important aspects of cardiac CT imaging include the identical with only minor variations. As the number of slices
following: has steadily increased from 4 to 16 to 64, major changes
have occurred with detector geometry and the necessary
1. High temporal resolution to minimize cardiac motion
post-processing algorithms. Special three-dimensional (3D)
artifacts
backprojection algorithms are needed to correct for artifacts
2. High spatial resolution for visualization of small car-
that arise from the cone beam geometry that results from the
diac structures
oblique x-ray projections of thin collimated slices generated
3. Fast volume coverage during one breath-hold to mini-
from the gantry (1). New MDCT scanners with up to
mize respiratory motion artifacts
64 slices allow for either 64 parallel slices based on widened
4. Electrocardiogram (ECG) synchronization of data
detector arrays or for overlapping projections based on flying
acquisition
focal spots along the z-axis, which allows for an improve-
5. Isotropic voxels to allow for oblique reconstructions
ment of spatial resolution and for reduction of cone beam
without loss of resolution
geometry image artifacts (2). Thin-slice collimations (0.5 to
Optimization of image quality is mainly related to the 0.625 mm) with reconstructed slice thickness of 0.4 to
elimination of artifacts related to cardiac and respiratory 0.7 mm allow for near isotropic voxels, which are important
TABLE 1-1
GENERAL MULTIDETECTOR SCANNER COMPUTED TOMOGRAPHY CARDIAC PROTOCOL
FOR 16- TO 64-SLICE SCANNERS
Patient with HR 65 beats/min Administer beta-blocker (oral or intravenous)
Intravenous contrast
Volume (mL) 70–140 mL (depending on scanner) with 40–50 mL saline flush
Rate (mL/sec) 4–6 mL/sec
Scanning delay Timed for region of interest with test bolus or bolus tracking technique
Collimation 0.6–0.75 mm
Reconstructed slice thickness 0.4–1 mm
Reconstruction interval 0.5 mm
Reconstruction kernel Medium soft tissue
Reconstruction phase Mid-late diastole 40%–70% or between 300 and 550 ms; for evaluation of function, sequential
phases in 5%–10% increments throughout the cardiac cycle
ECG gating Retrospective or prospective
Scan direction Craniocaudad starting at the tracheal bifurcation and extending to the diaphragm (may need to
start above the aortic arch for bypass graft or other vascular assessment)
Sector reconstruction Depends on scanner type; in general: single for HR 70 beats/min, multiple for HR 70 beats/min
Scanning time 10–20 sec
KVp 120
Effective mAs 500–900
Pitch 0.2–0.3
Matrix 512 512
Image display MPR, MIP, 3D volume rendered
HR, heart rate; ECG, electrocardiogram; MPR, multiplanar reconstruction; MIP, maximum intensity projection; 3D, three-dimensional.
Chapter 1: Heart 3
for oblique reconstructions without loss of spatial resolu- Systole Diastole

tion. Gantry speed rotations of 330 to 420 ms per R
360 degrees allow for high temporal resolution in the range R-R interval
of 165 to 190 ms, and new dual-source technology will allow
T
for further improvement in temporal resolution down to
83 ms, which is important for functional assessment of the P
heart as well as for improving image quality for patients with
fast heart rates (HRs). ECG dose modulation schemes can
further reduce the radiation dose received during image Q
acquisition by lowering the tube current during acquisition S
of nondesired phases. Figure 1-1 Schematic diagram of standard electrocardiogram
(ECG) with nomenclature.
Electrocardiogram Referencing
The rapid, constant motion of the heart is a source of signifi-
cant motion artifact on conventional CT imaging. Because ECG-related image reconstruction after scanning (Fig. 1-3).
cardiac motion varies throughout the cardiac cycle, usually In contrast to other CT spiral acquisitions, cardiac algo-
in a predictable manner, synchronization of data acquisition rithms usually require a substantially reduced pitch factor
to the cardiac cycle allows for combination of data acquired (0.2 to 0.3) to allow image reconstruction at any time point
from consecutive gantry rotations in volumetric datasets of the cardiac cycle. A comparison of these two acquisition
with minimal blurring artifact related to cardiac motion. techniques is shown in Table 1-2.
Typically, synchronization is based on the ECG signal One of the major benefits of retrospectively ECG-gated
(Fig. 1-1). The time between two consecutive heartbeats is data acquisitions is the ability to provide image data
defined by the RR interval (the interval between consecutive coverage of the full cardiac cycle. Depending on the patient’s
R waves of the ECG, 1,000 ms for a HR of 60 beats per individual HR, optimal timing of image data reconstruction
minute). Depending on the acquisition mode, the RR inter- can be chosen based on individual experience, multiple pre-
val of the ECG signal is used to either trigger data acquisi- reconstructions, or using semiautomated data evaluation
tion or to retrospectively gate the acquisition. tools. Image data reconstruction can be related to the
With prospective triggering, data acquisition is triggered R-wave trigger in a forward or retrograde manner, and it can
to the ECG R peak after a preselected delay within the be performed as a preset absolute time delay or as a percent-
following cardiac cycle at a stable table position. After table age of the cardiac cycle. In addition, this type of recon-
movement, data are again sampled based on the next avail- struction is less sensitive to aberrancy of the heart rhythm
able R trigger (Fig. 1-2). This type of data acquisition is (e.g., premature ventricular contractions) (3).
similar to the step and shoot mode. With retrospective ECG Dose modulation schemes, in which tube current is low-
gating, a conventional spiral dataset is acquired with simul- ered during acquisition of nondesired phases of the cardiac
taneous tracing and recording of the patient’s ECG. Sub- cycle, can be used with up to 50% reduction in radiation
sequent merging of the image and ECG data allows for dose (Fig. 1-4). However, this type of acquisition requires
TABLE 1-2
BASIC ADVANTAGES AND DISADVANTAGES OF DATA
ACQUISITION SCHEMES
Prospective Triggering Retrospective Gating
Advantages Pulsed radiation Spiral acquisition

Low radiation exposure Volumetric dataset
Allows for dose modulation Full coverage of cardiac cycle
Variable data reconstruction
High reproducibility
Disadvantages Sequential (axial rather than Continuous radiation
volumetric) data acquisition High radiation exposure
Predefined timing necessary
Partial RR interval covered
Irregular heartbeats cause misregistration
Scan
Scan
Scan
z-position
Scan
Delay Table Delay Table Delay Table Delay

Feed Feed Feed
Figure 1-2 Schematic diagram of prospective ECG triggering for CT data acquisition and reconstruction.
z-position
Recon
Recon
Recon
Recon
Recon
Recon
Continuous data acquisition

Figure 1-3 Schematic diagram of retrospective ECG triggering for CT data acquisition and reconstruction.
Chapter 1: Heart 5
Systole Diastole
100% Tube current
20% Tube current
Figure 1-4 Schematic diagram of ECG dose modulation for CT data acquisition and reconstruction.
one to have preset phases of the cardiac cycle desired for single-source scanners, multisector reconstruction algo-
acquisition because the other phases will often be subopti- rithms can further reduce the temporal resolution.
mal for image reconstruction and interpretation. Given the limited temporal resolution offered by current
MDCT technologies, image quality can often be improved
by optimizing the patient’s HR. Reconstruction is usually
performed during diastole, when cardiac motion is mini-
Temporal Resolution
mal. The duration of diastole is approximately two thirds of
Temporal resolution is the amount of time needed to the cardiac cycle; however, the proportion of the cardiac
acquire necessary scan data to reconstruct an image (4). cycle spent in diastole decreases with increasing HR and
Temporal resolution in cardiac MDCT is primarily depend- increases with lowering of the HR (1,6). Beta-blockers have
ent on gantry speed rotation. Half-scan reconstruction been used to both lower the HR and promote regularity of
algorithms result in temporal resolutions corresponding to the heart rhythm or decrease arrhythmias. Depending on the
half of a full 360-degree rotation, or 165 ms with a 330 ms optimal diastolic time interval desired, oral or intravenous
per 360 degree rotation speed. Further improvements in beta-blockers have been shown to be safe when adminis-
temporal resolution can be achieved with multisector tered to patients prior to CT scanning (6–8). In patients
reconstruction algorithms, which are based on merging data unable to tolerate beta-blockers (e.g., severe asthmatics),
of adjacent cardiac cycles (two to four heartbeats) to fill a calcium-channel blockers can provide a similar benefit.
data segment more rapidly. However, in using these algo-
rithms, image quality is often degraded because of blurring
Spatial Resolution
artifacts related to differences in the position of cardiac
structures from cycle to cycle even within the same phase of Spatial resolution is mainly dependent on detector width,
the cycle (4,5). Dual-source CT scanners achieve temporal slice collimations, and data sampling. Coronary arteries
resolution of 83 ms because of simultaneous scanning are small, ranging from 1 to 2 mm in diameter (9). Nitro-
using two sets of sources and detectors, and similar to glycerin (0.4 mg) is often given just prior to scanning to
dilate the coronary arteries (10,11). The spatial resolution In addition to 2D evaluations, volume and shaded surface
of EBCT is 0.7 0.7 3 mm, and that of four detector rendered 3D reconstructions with advanced visualization
MDCT scanners is 0.6 0.6 1.0 mm. In comparison, the algorithms allow for the extraction of specific overlying
spatial resolution of MR coronary angiography is about structures, such as bones, lung tissue and vessels, and soft
1.0 1.0 1.5 mm, depending on the scanning parame- tissue structures, to focus on cardiac anatomy. These
ters. With helical CT volume acquisition and reconstruction volume rendered 3D images (Fig. 1-6) are invaluable for
of overlapping sections, the z-axis resolution is improved, understanding complex cardiac anatomy (e.g., anomalous
leading to near isotropic voxels. The resolution of 16- coronary arteries) and in surgical planning. Shaded surface
detector MDCT is 0.5 0.5 0.6 mm, and that of 64-slice displays are commonly used for visualization of internal
systems is up to 0.4 0.4 0.4 mm (12). These advances structures from a specified point simulating an endo-
have greatly enhanced the visualization of small structures scopic or angioscopic view (Fig. 1-7) and can provide the
such as the coronary arteries, but the resolution remains important information on the internal relationship of
inferior to that of conventional coronary angiography, structures, which may be important for guiding proce-
which has a spatial resolution of 0.2 0.2 mm (13). dures (e.g., pulmonary vein isolation). ECG gating greatly
reduces the amount and degree of stair-step artifact related
to changes in cardiac position between gantry rotations but
Contrast Administration
may still be apparent primarily because of slight changes in
Several considerations are important to be aware of with the position of cardiac structures that occur from beat to
contrast use in cardiac imaging. First, because image qual- beat, especially when there is even minor irregularity in the
ity is greatly affected with irregularity of the heart rhythm, heart rhythm.
nonionic contrast is used to prevent perturbations in the Because CT data are volumetric and are acquired synchro-
HR during image acquisition (14). Second, patients with nized to the cardiac cycle, slices in any arbitrary plane in
cardiac disease may have low cardiac outputs (COs), and, 3D space can be created during multiple phases of the
as such, the timing of contrast arrival into the regions of cardiac cycle. Although currently temporal resolution is still
interest can be variable. Hence, it becomes important to limited (83 ms at best), qualitative evaluation of wall
accurately track contrast arrival times individually for each motion and valve function can be made. Additionally,
patient. Finally, the physiologic impact of receiving con- precise quantitative calculation of ventricular volumes and
trast media must be taken into consideration, as this can ejection fraction can be obtained from the datasets (15–18).
also affect the timing of contrast arrival.
Two types of contrast timing methods are employed in
Magnetic Resonance Imaging
current MDCT scanning systems. The first involves deter-
mining the circulation time by giving a small-volume test The challenge in cardiac MR imaging is to cope with the
bolus (10 to 20 mL) of contrast and observing the time of its cardiac motion and superimposed respiratory motion dur-
arrival in the heart, also known as a timing run. We usually ing the time needed for imaging sequences. The ongoing
subtract 6 seconds from the circulation time to determine development of MR scanners, sequence techniques, and
the time to start image acquisition. The second involves gradient amplifier enhancements has led to increasingly
imaging a specific region of interest (e.g., ascending aorta), widespread cardiovascular MR applications. In addition,
and once the contrast density attenuation reaches a prespec- the use of parallel imaging has led to further im-
ified point, the scanner turns on and starts imaging, also provements in temporal and spatial resolution. With the
known as bolus tracking. Again, exact values are dependent progress in rapid MR techniques, data acquisition is possi-
on scanner type, contrast type, amount and delivery ble in just a fraction of the heart cycle. However, synchro-
method, scan delay time, and individual preference. nization with the heart cycle is still needed to obtain
images without blurring throughout the cardiac cycle for
both structural and functional analysis.
Post Processing
Because of the near isotropic volumetric nature of the
Technical Design
MDCT data, oblique multiplanar and thin maximum
intensity projection reconstructions, including curved Technical improvements in hardware and software design
re-formations, are used in addition to standard axial two- have promoted the growth of cardiovascular MR imaging.
dimensional (2D) evaluation of CT images to better under- Developments of MR systems, coil technology, and
stand cardiac structures. These oblique reconstructions are sequence techniques have produced considerably short-
especially important for visualization of the coronary ened acquisition times and improvements in overall image
vessels throughout their course in both a longitudinal and quality. High field imaging at 3 Tesla and multichannel
axial plane to the vessel of interest for the assessment coils result in improved spatial resolution but may intro-
of atherosclerotic plaque and luminal narrowing (Fig. 1-5). duce other technical issues that need to be addressed.
Chapter 1: Heart 7
A B
Figure 1-5 CT images using multiplanar reconstructions in oblique planes demonstrate severe noncalcified plaque (arrows) in the right
coronary artery on both longitudinal (A) and axial (B) views to the vessel of interest.
Figure 1-6 CT images demonstrating three-dimensional volume rendered reconstructions that show the normal origins of the left coro-
nary system with an intramyocardial bridge (arrowhead) in the mid segment of the left anterior descending (LAD) artery. LCX, left circumflex
artery.
one cardiac cycle to ensure that a complete heart cycle is

covered. With retrospective gating, data acquisition occurs
continuously with recording of the patient’s ECG. This
type of gating is useful for studies of cardiac function, in
which the complete heart cycle is covered during data
acquisition. Image data can then be referenced to the ECG
and reconstructed with a user-definable temporal resolu-
tion. Additionally, arrhythmia rejection can be employed
to reject data acquired outside a user-definable heart cycle
range.
Respiratory Referencing
Cardiac MR sequences are often acquired over multiple
heart cycles and usually require a breath-hold to avoid
blurred images due to respiratory motion. With particu-
larly long sequences and in patients unable to adequately
hold their breaths, multiple averages and navigator tech-
niques that track the movement of the diaphragm have
been used to acquire data during a particular phase of the
respiratory cycle.
Figure 1-7 CT image demonstrates shaded surface rendered
reconstruction of the interior of the left atrium and draining pul-
monary veins. RSPV, right superior pulmonary vein; RIPV, right infe-
rior pulmonary vein; LSPV, left superior pulmonary vein; LIPV, left Cardiac Acquisition Sequences
inferior pulmonary vein; LAA, left atrial appendage.
Tomographic Imaging
General anatomic imaging is performed using either spin-
Acceleration techniques (e.g., parallel imaging) can further echo or gradient-echo techniques for evaluation of cardiac
reduce scan acquisition time but usually at the expense of
spatial resolution. Hence, a combination of factors is often
needed to optimize cardiac imaging protocols. Prospective Gating
R-R interval
Motion Compensation
Electrocardiogram Referencing
Synchronization of image data is usually performed with
ECG referencing to the R wave. However, the magneto-
hydrodynamic effect of the flowing blood can lead to an
increase in the T-wave amplitude of the ECG, which may be
misinterpreted as an R wave. This misinterpretation can Trigger Data Trigger
result in mistriggering, with consequent blurring of the Delay Acquistion Search
image data. The development of vectorcardiography-based
triggering can classify cardiac events by comparing them
with a previously calculated reference vector of the QRS Retrospective Gating
complex of the heart cycle (19). In this way, events with the R-R interval
same magnitude but different phase as the R wave are not
misinterpreted as such.
Similar to ECG referencing with MDCT, detected R
waves act as reference points for data acquisition, which
can be performed as prospective triggering or retrospective
gating (Fig. 1-8). With prospective triggering, the detection
of the R wave triggers the execution of the next part of the ... ...
imaging sequence. The search for the next trigger event is Continuous Data Acquisition
enabled after the completion of the sequence part. Figure 1-8 Schematic diagram of prospective and retrospective
Consequently, a single measurement can cover more than ECG triggering for MR data acquisition.
Chapter 1: Heart 9
structure and morphology. The spin-echo sequences can determines the maximum time savings and is also known
be performed with T1 and T2 weighting, fat saturation, as the turbo factor. Single-shot TSE/FSE sequences use a
inversion recovery (IR), and pre– and post–contrast single excitation followed by multiple RF refocused echoes
administration to provide static high-resolution images to acquire all data in one heart cycle. In addition, the half-
with excellent contrast between the intracavitary blood Fourier technique accelerates the measurement further
flow and myocardium or other structures (20). The gradi- by acquiring and using only a little more than half of the
ent-echo sequences are a quick method for providing raw data.
information about cardiac structure, although tissue con- In a general gradient-echo–based sequence, the 180-
trast for characterization is not as great as with spin-echo degree refocusing pulse is omitted, and spin dephasing
sequences. In general, gradient-echo sequences are widely occurs because of local field inhomogeneities that are not
used at present for assessment of functional characteristics refocused. Hence, flowing blood appears bright because of
of the heart. Methods to shorten acquisition time as well the wash-in effect of fresh blood. The advantage of the
as cardiac gating are used to decrease cardiac motion arti- gradient-echo–based sequences is the potentially short
fact and improve overall image quality. repetition time.
In spin-echo sequences, flowing blood generally The spoiled gradient-echo sequence uses the steady
appears hypointense. A complete signal void usually state of the longitudinal magnetization (small flip angle),
results if all of the spins excited by the 90-degree pulse and the remaining transverse magnetization is destroyed
have flowed out of the slice of interest and hence do not or “spoiled.” This can be accomplished by using a long
experience the 180-degree pulse. This effect can be intensi- repetition time (TR), by adding a phase offset, or by
fied by the use of 180-degree preparation pulses that invert extending the duration of the readout gradient (spoiler or
the blood and tissue signal inside and outside the excita- crusher gradients). Spoiled gradient-echo images may be
tion slice (nonselective inversion) followed by a subse- either T1- or T2*-weighted. The longer the TR and the
quent 180-degree slice-selective inversion pulse, which higher the flip angle, the more T1-weighted the image
results in re-inversion of the signal within the slice of inter- becomes, whereas the longer the echo time (TE), the more
est. Blood flowing into a slice is inverted by the first T2*-weighted the image is (21). This technique is less sen-
180-degree pulse, and acquisition timing is such that the sitive to field inhomogeneities compared with steady-state
blood flows through the slice during the zero-crossing of free precession techniques, with excellent contrast between
its magnetization, resulting in signal from the adjacent blood and tissue in through-plane orientations due to the
tissue and none from the flowing blood. wash-in phenomenon of fresh blood (Fig. 1-9).
Turbo/fast spin-echo (TSE/FSE) sequences shorten the The fully balanced steady-state free precession sequence
acquisition times of general spin-echo sequences consider- uses the steady state of the residual transverse magnetiza-
ably by acquiring multiple-k-space lines using multiple tion (high flip angle) in addition to the steady state of the
phase-encoded echoes acquired following a single excita- longitudinal magnetization. In steady-state free preces-
tion. The multiple 180-degree refocused echoes are also sion, all echo paths that have been dephased for the pur-
referred to as the echo train. The length of the echo train pose of spatial encoding are rephased (21). Theoretically,
A B
Figure 1-9 Tomographic axial MR image using dark blood spin-echo (A) and bright blood gradient-echo (B) imaging.
the achieved contrast between blood and tissue becomes a Tagging

function of the ratio between T1 and T2, rather than a Magnetic tissue tagging is a process in which landmarks on
function of TE and TR. The TR/TE in steady-state free pre- an MR imaging plane are introduced noninvasively prior
cession sequences are much shorter than in the spoiled to cine imaging (Fig. 1-12). Presaturation pulses are used
gradient-echo sequences. At TR<<T2, the image signal is to create linear dark stripes or a grid pattern over the
proportional to the T2/T1 ratio, which is approximately myocardium. As systole proceeds, the pattern distorts in
1:5 in arterial blood and 1:18 in myocardium. Hence, the direction corresponding to myocardial movement.
there is high contrast between blood and myocardium, but This technique can be very useful in the analysis of the
because steady-state free precession sequences are more heart wall motion because it can provide temporal corre-
sensitive to field inhomogeneities, they can exhibit more spondence of material points within the heart wall, which
artifacts (Fig. 1-10). The implementation of steady-state can be assessed reproducibly and quantitatively.
free precession sequences requires high-performance MR
systems capable of delivering fast gradients to ensure short
enough times to preserve transverse magnetization (21). Phase Contrast
Phase-contrast MR methods are based on the sensitivity of
the phase of the MR signal to motion. Using appropriate
Cine Sequences velocity encoding gradients, the motion-dependent phase
Cine sequences used in cardiac MR imaging consist of a effect can be used to measure two datasets with different
collection of images (usually at the same spatial location) velocity-dependent phase at otherwise identical acquisi-
covering one full period of the cardiac cycle. The pulse tion parameters. Subtraction of the two resulting phase
sequence used is commonly a type of gradient-echo images allows for quantitative assessment of the velocities
sequence because these can be acquired with relatively of the underlying motion. In the simplest phase-contrast
short acquisition times. The image acquisition utilizes pulse sequence, bipolar gradients (two gradients with
cardiac gating to yield a multiphase sequence covering the equal magnitude but opposite direction) are used to
cardiac cycle. As previously mentioned, data can be timed encode the velocity of blood flow. Stationary tissue will
to the cardiac cycle in a prospective or retrospective undergo no net change in phase after the two gradients are
manner. Segmented data acquisition is commonly em- applied. Flowing blood will experience a different magni-
ployed to reduce overall imaging time. Segments are then tude of the second gradient compared with the first owing
pieced together from different cardiac cycles to yield an to its different spatial position and a resultant net phase
image loop tracking the motion of tissues within the heart shift. This information can be used directly to determine
and/or vessels over the cardiac cycle (Fig. 1-11). the velocity of the flowing blood.
Figure 1-10 Gradient-echo MR

image of the right ventricular out-
flow tract demonstrating a large
amount of artifact (arrows) in the
steady-state free precession image
(A) related to the presence of tissue
valve prosthesis. On spoiled gradi-
ent-echo imaging (B), there is much
less artifact and the surrounding
structures, including valve leaflets,
A, B can be visualized.
Chapter 1: Heart 11
Figure 1-11 Schematic diagram demonstrating acquisition of gradient-echo cine MR imaging at a basal short axis level of the heart.
Figure 1-12 Schematic diagram demonstrating acquisition of tagged gradient-echo cine MR imaging at a mid short axis level of the
heart. The tagging sequence (dark gray) is applied just prior to cine imaging.
With the phase-contrast method, two sequences are First Pass Perfusion
acquired, one with and one without motion compensa- A saturation recovery technique is used in cardiac MR to
tion. By comparing the phase of signals from each location maximize the T1 contrast between normal and hypoper-
in the two sequences, the exact amount of motion-induced fused myocardium during a dynamic contrast-enhanced
phase change can be determined and mapped, where pixel first pass myocardial perfusion acquisition. This is
color (black vs. white) represents velocity direction and accomplished by using a 90-degree saturation pulse in
pixel brightness is proportional to spatial velocity (Fig. combination with a gradient-echo–based sequence. The
1-13). In the basic sequence, signal only occurs when a preparation pulse saturates the entire tissue in the slice of
voxel of tissue contains a uniform velocity in a specific interest, and the contrast between the different perfused
direction. Acquisition of the sequence in the three orthog- tissues is based on the very short T1 generated by the
onal directions is needed to obtain a complete map of contrast uptake (Fig. 1-14). Fast acquisition sequences
flow regardless of direction. Phase change is measured on allow for the collection of multiple images during a
a circular scale and is designed so that different flow veloc- single cardiac cycle, and the flow of contrast into the
ities are displayed along the 360 degrees of phase shift, imaging planes can be followed during the first pass of
depending on the velocity-encoding factor, or VENC. For the contrast agent.
best results, the VENC is chosen to be slightly more than
the maximum expected flow velocity in a region. However, Contrast-enhanced Magnetic Resonance
if a higher velocity is encountered than that set for the Angiography
VENC, it will be represented as a darker (or brighter, Depiction of blood vessels is based on their T1 values
depending on direction) pixel, otherwise known as alone. The tissue surrounding blood vessels should have a
velocity aliasing (Fig. 1-13). Phase-encoded images are higher T1 value compared with blood itself. Additionally,
generally displayed as corresponding magnitude and T1-shortening agents such as gadolinium-based contrast
phase display, usually in encompassing an entire cardiac can further reduce the T1 value of the blood to below
cycle (cine display). For the phase display, the flow infor- 50 ms. As such, vessels of interest become much brighter
mation on flow direction (bright vs. black) and flow than background fat and visualization of in-plane blood
velocity (signal intensity) can be visually interpreted. vessels is improved. This sequence is commonly performed
B C D
Figure 1-13 Phase-contrast velocity mapping (B–D) at the level of the narrowest point in the descending aorta (dashed line in A, arrows
in B–D). The magnitude images (B) are displayed with phase images demonstrating aliasing (C) when velocity encoding is set at 200 cm/sec
and no aliasing (D) when velocity encoding is set at 300 cm/sec.
Chapter 1: Heart 13
Figure 1-14 Basal (left column), mid (middle column), and apical (right column) images demonstrate pre-contrast (top row), early contrast
(middle row), and late contrast (bottom row) during the first pass of a small bolus of gadolinium contrast agent using a saturation recovery
technique for evaluation of myocardial perfusion. There is delayed uptake of contrast in the inferolateral wall segments (arrows).
as a T1-weighted 3D spoiled gradient-echo pulse sequence contrast agents and newer intravascular agents have also
with zero interpolation in the slice-selected direction and been utilized to improve image contrast. Both 2D and 3D
short TR and TE values. Contrast injection is timed with sequences are available, each with its own pros and cons.
respect to the k-space acquisition such that peak enhance- The most widely available sequence is a 2D segmented
ment coincides with optimal contrast imaging (Fig. 1-15). k-space gradient-echo acquisition in which multiple thick
This can be achieved with educated guess, bolus timing, or slices are acquired through the vessel of interest. Variability
bolus tracking techniques as discussed previously. in breath-holds and thick slices can limit the registration
of images from slice to slice. 3D imaging methods are
commonly performed using bright-blood techniques with
Coronary Magnetic Resonance Angiography free breathing. These sequences can be performed with or
Coronary artery imaging is one of the most technically without gadolinium-based contrast agents (21).
challenging areas of MR imaging. In addition to the spatial Regardless of the sequence used, certain fundamental
requirements necessary to visualize the small coronary components are common to all the sequences. 2D slices or
arteries, respiratory and cardiac motion provide further 3D slabs need to be positioned appropriately to encom-
impediments to optimizing image quality. Numerous MR pass the entire course of the coronary artery to be imaged.
sequences have been used for imaging the coronary Depending on the slab thickness, often at least two slabs
arteries. Bright-blood techniques, including spoiled gradi- are needed to image all three major coronary arteries,
ent-echo, echo planar, and steady-state free precession including an oblique sagittal slab directed along the atrio-
gradient-echo sequences and black-blood techniques, have ventricular (AV) groove and a second oblique sagittal
been proposed. Furthermore, conventional gadolinium acquisition along the anterior border of the heart. Data
A B
Figure 1-15 Contrast-enhanced MR angiography performed as a time-resolved sequence with maximum intensity projection reconstruc-
tions demonstrating primary enhancement of the pulmonary arteries (A) followed by primary enhancement of the aorta (B) in a patient with
absent left pulmonary artery and left lung supplied by collaterals.
acquisition is typically constrained to a window less than hyperintense, scarred myocardium) reaches zero at different
150 to 200 ms per heartbeat during diastole to minimize times. Appropriate selection of inversion time (TI) can
cardiac motion. Respiratory motion artifacts are mini- suppress the signal of relevant tissue and improve contrast
mized by using multiple short breath-hold acquisitions or between different tissue types with different TI relaxation
navigator techniques to track the diaphragm position times. The challenge is in the choice of TI, which depends
during data acquisition (Fig. 1-16). Because the coronary on the amount and time of contrast administration and
arteries lie against the epicardial surface of the heart and continually changes during the examination.
are surrounded by epicardial fat, methods to improve
the image contrast between fat, myocardium, and flow-
ing blood are used, such as fat and myocardial sup- Acceleration Techniques in Magnetic Resonance
pression. Anterior saturation bands and prone imaging
Although the speed of pulse sequences has steadily
can reduce ghosting artifacts related to free-breathing
improved with the performance of gradient hardware, the
techniques (21).
maximum gradient switching rates (slew rate) are still a
limiting factor because further increments will lead to
Inversion Recovery (Delayed Enhancement Imaging) patient nerve stimulation. Because the speed of MR imag-
The IR technique is often used in cardiac MR imaging to ing is generally limited by the number of required phase-
characterize tissue after the administration of contrast encoding steps, several approaches have been introduced
media. Normally perfused myocardium is often sup- to reduce the number of required phase-encoding steps
pressed to better visualize areas with increased contrast and thus further shorten data acquisition time.
uptake that appear hyperintense (Fig. 1-17). In conjunc- With single-shot k-space filling, the goal is to acquire all
tion with TSE imaging, the IR technique can be used to the needed k-space lines during one heart cycle. In TSE/FSE,
suppress fat to identify fatty infiltration seen in arrhythmo- this is performed by a single excitation followed by multi-
genic right ventricular (RV) dysplasia or to visualize edema ple 180-degree refocusing pulses. In segmented k-space fill-
and tissue damage seen in acute myocardial infarctions. ing, only some k-space lines are acquired per heart cycle.
A 180-degree inversion pulse is applied before data Depending on the total number of k-space lines, the
acquisition. As the longitudinal magnetization relaxes, the sequence is repeated until the entire k-space is filled. This
magnetization of different tissue types (fat, normal or technique is used to collect data over a longer period of
Chapter 1: Heart 15
A, B
C, D
Figure 1-16 Navigator techniques are used to track diaphragm position during image acquisition. Navigators (arrows) are positioned
over the diaphragm, and the intersection points are used to track the diaphragm position during image acquisition of a specified slice
through the coronary arteries (arrowheads) (A, B). Image acquisition is set to trigger when the diaphragm is positioned at a certain phase of
respiration, usually end-expiration, indicated by the mode (C, D).
time to improve the spatial and temporal resolution. coil elements aligned along the phase-encoding direction.
Segmented k-space with view sharing is based on the seg- The number of skipped phase-encoding steps is identified
mented technique and involves sharing the outer lines of by the acceleration/PAT factor [e.g., factor of 3 (PAT3)
kspace between adjacent cardiac phases to improve tempo- implies that only every third k-space line is measured]. The
ral resolution. Because the most important information is missing information is replaced using the individual coil
in central k-space, these lines are acquired separately for sensitivity profiles from all involved coil elements.
each image. In nonsegmented k-space filling, one k-space Different reconstruction methods such as sensitivity encod-
line is acquired per cardiac phase, resulting in a long acqui- ing (SENSE) (22) and generalized autocalibrating partial
sition time but higher spatial resolution. parallel acquisition (GRAPPA) (23) can eliminate aliasing
Half- and partial-Fourier techniques decrease scan time artifacts related to the reduction of k-space lines.
by omitting certain lines of k-space, which are recon- Depending on slice orientation and coil arrangement, PAT
structed after measurement using theoretical symmetry of factors of 2 to 3 for 2D measurements and 4 to 6 for 3D
k-space. Shared k-space filling, in which outer k-space lines measurements can usually be obtained. With short acquisi-
between adjacent cardiac phases are “shared,” is used to tion time it is possible to acquire the data in free breathing
shorten the acquisition and improve temporal resolution. and to minimize image blurring during arrhythmias. Data
Parallel acquisition techniques (PAT) involve omitting acquisition usually is set to occur during mid-to-late dias-
certain phase-encoding steps, depending on the number of tole to avoid image degradation from cardiac motion.
A, B C, D
E, F G, H
Figure 1-17 A–H: Inversion recovery technique used in delayed enhancement imaging for visualization of myocardial scarring. Normal
myocardium is suppressed (dark) to improve visualization of areas with increased contrast uptake, which appear hyperintense. Short axis
images from the base (A) to apex (H) of the heart demonstrate areas of myocardial scarring in the inferior, inferolateral, and anterolateral
segments (arrows).
Specific Evaluations 3. Ejection fraction (EF) equals the SV divided by the EDV
multiplied by 100 [EF (SV/EDV) 100] to be
Ventricular Analysis reported as a percentage
Ventricular analysis requires acquisition of a volumetric 4. CO equals SV multiplied by the HR; CO SV HR
cine imaging data set usually acquired in the true short
axis plane from the base of the heart to the apex, Myocardial mass can be calculated by inclusion of the RV
although long axis orientations have also been used (24). and LV epicardial borders during the ED phase. The interven-
Next, representative end-diastole ED and end-systole tricular septum is included with the LV and excluded from
ES cardiac-phase images are chosen as phases with the RV tracing for myocardial mass calculation. The volumes
the largest and smallest ventricular volumes or the phase of all sections are added, and the corresponding EDV is sub-
images obtained immediately before mitral valve closure tracted from the endocardial EDV to yield the myocardial
ED and opening ES. Once the phase images are volume. This result is then multiplied by the specific gravity
chosen, the right and left ventricles are traced along of myocardium (i.e., 1.05 g/mL) to calculate the mass.
the endocardial margin on each section obtained in the Although this measure is useful for the assessment and
selected ED and ES phases from the cardiac apex to follow-up of hypertrophic states, clinicians commonly also
the section just prior to the one that depicts the mitral rely on the one-dimensional ED thickness measurement at
and tricuspid valve annuli. By convention, the tracings specific segments throughout the left ventricle. Normal car-
should exclude trabeculations. Papillary muscles are also diac chamber dimensions using steady-state free precession
commonly excluded but may be included inside the cine MR techniques are shown in Table 1-3 (see ref. 346).
endocardial margin, as long as there is consistency Diastolic function of the ventricles can be measured by
followed for both ED and ES phases (Fig. 1-18). several different methods (25). The use of phase-contrast
Calculations can then be made as follows: velocity mapping at the transtricuspid or transmitral level
for calculation of time-velocity curves is a method similar
1. ED and ES volumes for each section are totaled to yield to that used in Doppler echocardiography. However, the
the RV and left ventricular (LV) end-diastolic volume calculation of RV and LV time-volume curves is generally
(EDV) and end-systolic volume (ESV); ∑EDLV = LVEDV, the most reliable method used in practice (1). Diastolic
∑ESLV LVESV, ∑EDRV RVEDV, ∑ESRV RVESV function can be evaluated by several parameters related to
2. Stroke volume (SV) equals the EDV minus the ESV; LV both early and atrial filling. The early rapid filling phase
SV LVEDV LVESV can be characterized by parameters such as peak filling
Chapter 1: Heart 17
LV Function Quantitative Assessment
End-diastole & End-systole
Bas
e to
ape
x
Figure 1-18 Volumetric acquisition of MR slices for calculation of ventricular volumes.
rate, time-to-peak filling rate, acceleration and deceleration and atrial peak filling rate (E/A ratio) is another parameter
slope of the early filling peak, deceleration time, and dura- that is commonly used. Tagged MR imaging can also be
tion of the rapid filling period (26). The atrial filling peak used to measure global and regional diastolic function by
is often quantified by peak filling rate. The ratio of early calculation of diastolic strain rate (27).
TABLE 1-3
NORMAL RANGE OF LEFT VENTRICULAR AND RIGHT VENTRICULAR CARDIAC DIMENSIONS
Mean SD (n 108) Male (n 63) Female (n 45)
LV ejection fraction (%) 69 6 69 6 (57–81) 69 6 (57–81)

LV mass (g) 112 27 123 21 (81–165) 96 27 (42–150)
LV mass index (g/m2) 59.2 11 62.5 9.0 (45–81) 54.6 12 (31–79)
LV end-diastolic volume (mL) 150 31 160 29 (102–218) 135 26 (83–187)
LV end-diastolic volume index (mL/m2) 80 13 82 13 (56–108) 78 12 (54–102)
LV end-systolic volume (mL) 47 15 50 16 (18–82) 42 12 (18–66)
LV end-systolic volume index (mL/m2) 25 7 25 8 (9–41) 24 6 (12–36)
LV stroke volume (mL) 104 21 112 19 (74–150) 91 17 (57–125)
LV stroke volume index (mL/m2) 55 8 56 8 (40–72) 54 9 (36–72)
RV ejection fraction (%) 61 6 59 6 (47–71) 63 5 (53–73)
RV mass (g) 38 8 41 8 (25–57) 35 7 (21–49)
RV mass index (g/m2) 20.3 3.6 20.6 3.7 (13–28) 20.0 3.5 (13–27)
RV end-diastolic volume (mL) 173 39 190 33 (124–256) 148 35 (78–218)
RV end-diastolic volume index (mL/m2) 91 16 96 15 (66–126) 84 17 (50–118)
RV end-systolic volume (mL) 69 22 78 20 (38–118) 56 18 (20–92)
RV end-systolic volume index (mL/m2) 36 10 39 10 (19–59) 32 10 (12–52)
RV stroke volume (mL) 104 21 113 19 (75–151) 90 19 (52–128)
RV stroke volume index (mL/m2) 55 9 57 8 (41–73) 53 9 (35–71)
LV, left ventricular; RV, right ventricular; SD, standard deviation.

Copyright 2005 from Hudsmith LE, Petersen SE, Francis JM, et al. Normal human left and right ventricular and left atrial dimensions using steady
state free precession magnetic resonance imaging. J Cardiovasc Magn Reson. 2005;7:775–782. Reproduced by permission of Taylor & Francis Group,
L.L.C., http://www.taylorandfrancis.com.
Flow Analysis basis for demonstrating aortic valve and aortic root
Quantitative assessment of blood flow through the cardiac anatomy as well as for obtaining additional images of the
chambers, valves, and blood vessels is important in the ascending aorta. Depending on the structure of interest,
management of cardiac patients. Phase-contrast velocity additional imaging planes directed to best depict the
measurements allow the assessment of the volume flow anatomic relationships may be important for diagnosis
through a cross-sectional area of a vessel according to a (Fig. 1-20). The coronary anatomy is best displayed in con-
simple relationship: Q V A, where Q blood flow, ventional angiographic views, including right anterior
V velocity in cm/sec, and A lumen area in cm2. Phase- oblique, left anterior oblique, and so forth. However, the
contrast velocity measurements have been validated and use of 3D images allows for visualization of the coronary
are currently used to quantify flow and other physiologic arteries in relation to the underlying cardiac chambers
parameters such as SV, valvular stenosis and regurgitation, (Fig. 1-21). Hence, depending on the technique, imaging
and CO, with an overall error rate in flow measurement protocols often need to be customized to accommodate
below 10% (28). It should be noted that peak velocities individual patient variations and distortions that may be
might be underestimated with this method due to factors normal or due to cardiac disease.
such as angle effects in the case of eccentric jets or low
temporal resolution. Hence, imaging planes should be
positioned orthogonal (through-plane) to the direction of Cardiac Chambers
the flow, and at least 16 phases of the cardiac cycle must be The heart is a hollow muscular organ with a conical shape,
sampled with the encoding velocity (VENC) set well above located between the lungs in the middle mediastinum just
the actual peak velocity in the region of interest to avoid to the left of and below the sternum. The heart is enclosed
aliasing (1). in the pericardium. The adult heart measures about 12 cm
in length, 8 to 9 cm in breadth at its broadest section, and
Assessment of Intracardiac Shunts 6 cm in thickness. The heart is subdivided by septa into
Accurate assessment of the ratio of pulmonary (Qp) to sys- right and left halves, and a constriction subdivides each
temic (Qs) flow is important in the management of half of the organ into two cavities, with the upper cavity
patients with intracardiac left to right shunts. Large shunts named the atrium and the lower cavity named the ventri-
with a Qp/Qs ratio exceeding 1.5 are usually closed because cle, resulting in four main chambers: right and left atria
chronic left to right shunting with increased blood flow to and right and left ventricles. The atria are separated from
the lungs may lead to irreversible elevation of pulmonary the ventricles by the AV groove.
resistance, also known as Eisenmenger physiology. The right atrium is located high and anterior in the heart.
The Qp/Qs ratio is usually derived directly from heart It consists of the large, posterior quadrangular cavity situ-
catheterization oximetry, but noninvasive methods for ated between the vena cava and a smaller, anterior portion
calculation of this ratio can be made using Doppler known as the appendage. The main body of the right atrium
echocardiography, radionuclide angiography, and MR. consists of a thin, smooth wall, whereas the wall of the
With phase-contrast velocity mapping MR, flow quantifi- appendage contains raised muscular fibers termed pectinate
cation in both great arteries can be used to quantify Qp/Qs muscles. The separation of the appendage from the main
and has been shown to have good agreement with other cavity of the right atrium is indicated externally by the
methods for calculation of shunt size (29,30). terminal sulcus groove and internally by the vertical muscu-
lar ridge known as the crista terminalis. The superior vena
cava returns blood from the upper half of the body and
opens without a valve into the upper and posterior part of
NORMAL ANATOMY the atrium, directing blood downward and anteriorly
toward the AV valve. The inferior vena cava, larger than the
Given the complexity of the cardiac anatomy and the indi- superior, opens into the lowest part of the atrium near the
vidual variations unique to each patient, cardiac structures atrial septum and directs blood returning from the lower
can have different appearances depending on the imaging half of the body upward and posteriorly toward the atrial
plane. Hence, the most useful imaging planes are those septum. The eustachian valve guards the orifice of the infe-
that are parallel and perpendicular to the cardiac axes, rior vena cava, directing blood through the foramen ovale in
which are based on internal cardiac landmarks. The most the fetus. This valve structure may persist as a small
common standard views of the heart are demonstrated in remnant, cribriform or filamentous in appearance in the
Figure 1-19. The two-chamber and four-chamber long axis adult. The coronary sinus opens into the right atrium
views along with the short axis views are often used for between the orifice of the inferior vena cava and the
visualization of the cardiac chambers and for evaluation of AV opening. It returns blood from the main heart muscles
cardiac anatomy and function. The three-chamber view and is protected by a semicircular valve known as the thebe-
allows for visualization of the aortic root and provides the sian valve (31).
Chapter 1: Heart 19
horizontal
long axis
short axis
vertical
short axis long axis
A 3-chamber view
B, C, D
E, F, G
Figure 1-19 A–G: Standard imaging planes of the heart. RV, right ventricle; LV, left ventricle; RA, right atrium; LA, left atrium; LAA, left
atrial appendage; MV, mitral valve; TV, tricuspid valve; AV, aortic valve; Ao, aorta; CS, coronary sinus; P, pericardium. (Reprinted from Lee VS.
Cardiovascular MRI: physical principles to practical protocols. Philadelphia: Lippincott Williams & Wilkins; 2006, with permission.)
The left atrium is located high and posterior in the stricted at its junction with the main cavity and is longer,
heart. Similar to its right-sided counterpart, the left atrium narrower, and more curved than that of the right side. The
also consists of two main portions, including a large pulmonary veins, four in number, open without valves
cuboidal principal cavity and smaller, anteriorly situated into the upper part of the posterior surface of the left
appendage. Its walls are thicker than the corresponding atrium. The left AV opening is the aperture between the left
right atrium (3 mm). The appendage is somewhat con- atrium and ventricle. The pectinate muscles are fewer and
A, B, C
D, E, F
Figure 1-20 A–F: Additional imaging planes to demonstrate right ventricular, aortic valvular, and pulmonary valvular anatomy. RV, right
ventricle; RVOT, right ventricular outflow tract; RVIT, right ventricular inflow tract; RVA, right ventricular apex; LV, left ventricle; LVOT, left
ventricular outflow tract; LVIT, left ventricular inflow tract; LVA, left ventricular apex; RA, right atrium; LA, left atrium; MV, mitral valve; AV,
aortic valve; PV, pulmonary valve; Ao, aorta; NCC, non-coronary cusp; RCC, right coronary cusp; LCC, left coronary cusp; CS, coronary sinus.
A, B
Figure 1-21 A–D: Volume rendered dis-

play of the coronary artery tree. LM, left main;
LAD, left anterior descending artery; LCX, left
circumflex artery; RCA, right coronary artery;
Dg, diagonal branch; OM, obtuse marginal
C, D branch; CS, coronary sinus.
Chapter 1: Heart 21
smaller than in the right-sided appendage and are con- papillary muscles are large and arise only from the free
fined to the inner surface of the appendage (31). wall surface. The outlet component consists of the portion
The atria are separated by a wall of tissue known as the of the ventricle that gives rise to the aortic valve. Two
interatrial septum. In the right atrium, it forms the poste- leaflets of the aortic valve have muscular attachments to
rior wall in the region superior to the coronary sinus. In the outlet component. The posterior part of the outlet
the left atrium, the septum lies obliquely, running posteri- component is short, with the remaining aortic leaflet sup-
orly from right to left. The fossa ovale is an oval depression ported by the fibrous tissue of the aortic root, which is in
on the septal wall of the atrium, and corresponds to the fibrous continuity with the anterior leaflet of the mitral
location of the foramen ovale in the fetus. The fossa ovale valve (Figs. 1-19 and 1-20). The left AV opening is inferior
is the thinnest section of the interatrial septum, and it is and to the left of the aortic orifice and a little smaller than
situated in the lower part of the septum, above and to the the corresponding aperture on the right side. This aperture
left of the inferior vena cava orifice. The limbus of the fossa is surrounded by a dense fibrous ring, covered by the lin-
ovale is the prominent oval margin of the fossa ovalis. A ing membrane of the heart and guarded by the bicuspid
small slitlike valvular opening is occasionally found, at the AV valve known as the mitral valve. The aortic opening is
upper margin of the fossa (31). circular and located anterior and to the right of the mitral
The right ventricle is pyramidal and extends from the valve. Its orifice is guarded by the aortic semilunar valve.
right atrium to near the apex of the heart. The RV chamber The LV walls are about three times as thick as those of the
can be divided into three components consisting of the right ventricle, and on short axis section its shape presents
inlet, the apical trabecular portion, and the outlet portion an oval or nearly circular outline. Unlike the right ventri-
(Fig. 1-20). The inlet component extends from the AV cle, the LV myocardium has a smooth endocardial surface
junction to the distal attachments of the chordae and normally measures 1 cm thick at ED. The interventric-
tendineae of the tricuspid valve. The apical trabecular com- ular septum separates the left ventricle from the right
ponent extends from the junction of the inlet and outlet ventricle. The greater portion of the septum is thick and
portions to the apex. The trabeculae consist of rounded or muscular, termed the muscular ventricular septum.
irregular muscular columns, which project into the ventric- However, the upper and posterior part of the septum,
ular cavity as either prominent ridges, traversing muscle which separates the LV outlet component from the lower
bundles, or numerous small papillary muscles that arise part of the right atrium and upper part of the right ventri-
from both the septal and anterior free wall and form cle, is thin and fibrous and is termed the membranous
the distal attachments of the chordae tendineae of the tri- ventricular septum (31). For the purposes of standardi-
cuspid valve. A muscular band named the moderator band zation, the LV myocardium is commonly divided into
frequently extends from the base of the anterior papillary 17 segments (Fig. 1-22). This 17-segment model is used
muscle to the ventricular septum, and assists in preventing across all imaging modalities and is useful for describing
overdistension of the ventricle. The outlet component is the locations of myocardial disease and their relationships
the smooth-walled tube of muscle, infundibulum, which to coronary artery territories (32).
supports the leaflets of the pulmonary valve. The right AV
orifice is the large oval aperture of communication
Cardiac Valves
between the right atrium and right ventricle, surrounded
by a fibrous ring and guarded by the tricuspid valve. The The heart has four main cardiac valves that function to
opening of the pulmonary artery is circular, located at keep the blood flow circulating in one direction. The
the summit of the infundibulum and guarded by the pul- valves can be divided by the type of chambers they sepa-
monary semilunar valve. The leaflets of the pulmonary rate and include the AV valves, separating the atria from
valve and the tricuspid valve are widely separated by the the ventricles, and the semilunar valves, separating the
infundibular musculature. The RV free wall is thin, meas- ventricles from the great arteries.
uring 2 to 3 mm in thickness (31). The right-sided AV valve, the tricuspid valve, consists of
The left ventricle is longer and more cone shaped than three triangular cusps. The largest cusp is interposed
the right. Similar to the right ventricle, the left ventricle can between the AV orifice and the infundibulum and is termed
be described in terms of inlet, apical trabecular, and outlet the anterior (or infundibular) cusp. A second, posterior (or
components. The inlet component extends from the left marginal) cusp is in relation to the right margin of the
AV junction to the distal attachments of the chordae ventricle, and the third, medial (or septal) cusp is adjacent
tendineae of the left-sided AV valve. The apical trabecular to the ventricular septum. The cusps are fibrous structures
portion contains fine trabeculations, which are numerous, attached to a fibrous ring surrounding the AV orifice at the
fine muscular projections that give the endocardial surface base to form a continuous annular membrane with their
of the chamber a smooth surface. There are two prominent apices projecting into the ventricular cavity. The atrial
papillary muscles that have attachments to the chordae surfaces are generally smooth, and their ventricular surfaces
tendineae: the anterior lateral and the posterior medial. are often rough and irregular and, together with the apices
Unlike the papillary muscles of the right ventricle, these and margins of the cusps, give attachment to several
anterior
anterolateral
1 anterolateral 17 Apex
2
6
Basal Left Ventricular Segmentation
3
4 5 1
inferoseptal inferolateral
7
inferior
2 6
8 13 12
anterior 14 17 16
anteroseptal
7 anterolateral Horizontal 9 11
3 15 5
8 Long Axis (HLA)
12 (4 Chamber) 10
Mid-Cavity
11 Apex
9
17 4
10
inferoseptal inferolateral
inferior 1. basal anterior 7. mid anterior 13. apical anterior
2. basal anteroseptal 8. mid anteroseptal 14. apical septal
3. basal inferoseptal 9. mid inferoseptal 15. apical inferior
4. basal inferior 10. mid inferior 16. apical lateral
anterior
5. basal inferolateral 11. mid inferolateral 17. apex
13 lateral 6. basal anterolateral 12. mid anterolateral
14
septal Apical
16
15
Vertical
inferior Long Axis (VLA)
(2 Chamber)
Short Axis (SA)
Figure 1-22 Seventeen-segment model of the left ventricular myocardium. (Reprinted from Cerqueira MD, Weissman NJ, Dilsizian V, et
al. Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: a statement for healthcare professionals
from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart Association. Circulation. 2002;105:
539–542, with permission.)
delicate chordae that connect to trabeculae carnae of the in the adult heart. They are attached by their convex mar-
right side. The tricuspid valve is often slightly more apically gins to the wall of the artery at its junction with the ventri-
displaced than the corresponding left-sided AV valve. cle, with their free borders directed upward toward the
The left-sided AV valve, the mitral valve, is a bicuspid lumen of the vessel. The free and attached margins of each
valve that is a little smaller and more basally placed than are strengthened by tendinous fibers, and the former pre-
the corresponding tricuspid valve. It also consists of sents, at its middle, a thickened nodule (corpus Arantii).
fibrous tissue surrounded by a dense fibrous ring, covered From this nodule, tendinous fibers radiate through the
by the lining membrane of the heart. The two triangular segment to its attached margin but are absent from two
cusps are of unequal size and are larger, thicker, and narrow crescentic portions, the lunulae, placed one on ei-
stronger than those of the tricuspid valve. The larger cusp ther side of the nodule immediately adjoining the free
is placed in front and to the right between the AV and margin. Between the semilunar cusps and the wall of the
aortic orifices and is known as the anterior cusp; the pulmonary artery are the three sinuses.
smaller posterior cusp is placed behind and to the left of The aortic semilunar valve directs blood flow from
the opening. Similar to the tricuspid valve, there are chor- the left ventricle into the aorta. The valve consists of
dae tendineae that attach the body and tips of both cusps three cusps, which are often referred to for their relation-
to the anterolateral and posteromedial papillary muscles ship to the orifices of the coronary sinuses; two are ante-
in the left ventricle; however, the chordae tendineae are rior (right coronary and left coronary), and one is poste-
thicker, stronger, and less numerous than on the right. rior (noncoronary). They are similar in structure and
The pulmonary semilunar valve directs blood flow from mode of attachment to the pulmonary semilunar valves
the right ventricle into the main pulmonary artery. The but are larger, thicker, and stronger; the lunulae are more
valve consists of three cusps, two in front and one behind, distinct, and the noduli or corpora Arantii thicker and
again consisting of fibrous tissue. They are officially desig- more prominent. Opposite the valve, the aortic root con-
nated as right, left, and anterior for their positions in the sists of the sinuses of Valsalva and sinotubular junction,
fetal heart but are sometimes termed right anterior, poste- which separates the sinuses from the tubular portion of
rior, and left anterior cusps, respectively, for their positions the aorta (31).
Chapter 1: Heart 23
Coronary Arteries and Veins lateral wall. Some institutions, depending on the location
of the first marginal branch, may refer to this branch as the
Each coronary artery arises from the corresponding aortic high lateral branch of the circumflex, with subsequent
sinus. A segmentation scheme for the coronary arteries is branches named lateral or posterolateral branches, depend-
shown in Figure 1-23 (33). The left main artery arises from ing on their destination. The LCX continues in the AV
the left coronary sinus, is 5 to 10 mm long, and is constant groove for a variable distance. The LCX may be segmented
in diameter. It passes leftward posterior to the pulmonary as proximal (from the origin off the left main to and
trunk and usually bifurcates into the left anterior descend- including the origin of the first major obtuse marginal
ing (LAD) and the left circumflex (LCX) arteries, although branch), and distal or AV LCX (distal to the origin of the
it sometimes trifurcates to include a ramus intermedius. major obtuse marginal branch and running along or close
The LAD courses in the anterior interventricular groove to the posterior left AV groove), which may be very small
and terminates near the apex of the heart, supplying caliber. In patients with a left dominant system, the LCX
numerous septal perforators to the interventricular septum reaches the posterior interventricular groove and gives rise
and a variable number of diagonal branches to the antero- to a posterior descending artery (PDA) branch. The ramus
lateral wall of the left ventricle. The proximal LAD is the intermedius, if present, has a course similar to that of the
portion proximal to and including the origin of the first first diagonal branch of the LAD. For nomenclature
major septal perforator branch. The mid segment of the purposes, the septal, diagonal, and obtuse marginal
LAD corresponds to the segment immediately distal to branches are sequentially numbered as they arise from
the origin of the first major septal perforator branch and each respective vessel.
extending approximately one half the distance to the apex The right coronary artery (RCA) arises from the right
of the heart (coinciding or close to the origin of the second coronary sinus and runs rightward posterior to the pul-
diagonal branch). The distal segment of the LAD runs monary outflow tract and then inferiorly in the right AV
along the interventricular groove from the end of the mid groove toward the posterior interventricular groove, pro-
segment and usually extending beyond the apex. The LCX viding atrial branches (to the right atrium) and marginal
in turn courses along the left AV groove, providing a vari- branches (to the right ventricle), including the acute mar-
able number of obtuse marginal branches to supply the ginal branch, which arises at the junction of the mid and
distal RCA. The RCA is often segmented as proximal (first
one half of the distance to the acute margin of the heart),
mid (latter one half of the distance to the acute margin of
the heart), and distal (portion running along the posterior
Coronary Arteriogram right AV groove, from the acute margin of the heart to the
origin of the PDA). The first branch of the RCA is usually
Main LCA the conus branch, which runs superiorly and supplies the
Aorta 5 RV outflow tract (RVOT); this branch can also have a sepa-
Aorta 6 rate origin directly from the right coronary sinus 50% of
D1
11 7 9 the time (34). The sinus node artery also arises from the
1 CIRC D2 proximal RCA in 60% of individuals (otherwise a left atrial
AV 10 branch of the LCX serves this function) and runs superi-
2 RCA OM
3 13 12 LAD orly and posteriorly. In patients with a right dominant sys-
PL
4 8 tem, the distal RCA divides into the PDA and posterior LV
RPD
PD
14 branches. Dominance refers to which artery gives off a
15
PDA to supply the posterior part of the heart. About 85%
of patients are right dominant, 7% are codominant, and
1. Proximal RCA 6. Proximal LAD 11. Proximal LCX 8% are left dominant. The PDA courses along the inferior
2. Mid RCA 7. Mid LAD 12. First marginal interventricular groove, providing septal perforators to
3. Distal RCA 8. Distal LAD 13. Mid-distal LCX supply the inferior septum (34).
4. Right PDA 9. First Diagonal 14. Posterolateral branch Most of the veins of the heart drain into the coronary
5. Left main 10. Second Diagonal 15. Left PDA
sinus, a wide venous channel about 2.25 cm in length that
Figure 1-23 Coronary artery segmentation. RCA, right coronary runs from left to right in the posterior portion of the AV
artery; RPD, right posterior descending artery; AV, atrioventricular
continuation of the right coronary artery; LCA, left coronary artery; groove. The coronary sinus opens into the right atrium
LCX, left circumflex artery; CIRC, left circumflex artery; OM, obtuse between the right AV orifice and the inferior vena cava ori-
marginal; PL; posterolateral; PD, posterior descending; LAD, left fice. The tributaries are the great, small, and middle cardiac
anterior descending; D1, first diagonal; D2, second diagonal; PDA,
posterior descending artery. (Reprinted from Austen WG, Edwards veins; the posterior vein of the left ventricle; and the
JE, Frye RL, et al. A reporting system on patients evaluated for oblique vein of the left atrium. The great cardiac vein
coronary artery disease. Report of the Ad Hoc Committee for drains most of the area supplied by the left coronary artery
Grading of Coronary Artery Disease, Council on Cardiovascular
Surgery, American Heart Association. Circulation. 1975;51:5–40, system and begins at the apex of the heart, ascends in the
with permission.) anterior interventricular groove with the LAD artery, and
enters the left end of the coronary sinus. The left marginal Pericardium and Myocardium
vein is a tributary of the great cardiac vein and ascends
along the left margin of the heart. The middle and small The pericardium is a double-walled sac surrounding the
cardiac veins drain most of the area supplied by the RCA. heart, composed of a serous membrane, the visceral peri-
The middle cardiac vein begins at the apex, ascending in cardium, which directly overlies the external surface of the
the posterior interventricular groove with the PDA, and heart and epicardial fat, and a fibrocollagenous outer layer,
empties into the right side of the coronary sinus. The small the parietal pericardium, which separates the heart from
cardiac vein runs in the right AV groove and, along with other intrathoracic structures (35,36). The pericardial cav-
the right marginal vein, which ascends along the right ity is the potential space that lies between the parietal and
margin of the heart, drains into the right side of the coro- visceral layers and normally contains about 15 to 50 mL of
nary sinus or directly into the right atrium. The posterior serous fluid, which allows the two membranes to slide
vein of the left ventricle runs along the diaphragmatic sur- smoothly over one another and thus permit free motion of
face of the left ventricle and drains either directly into the the heart within the pericardial sac (37). The parietal peri-
coronary sinus or into the great cardiac vein. The oblique cardium is attached anteriorly via ligaments to the chest
vein of the left atrium is a small vessel that descends wall, posteriorly to the vertebral column, and inferiorly to
obliquely posterior to the left atrium and opens into the the central tendon of the diaphragm. Superiorly, the
left end of the coronary sinus. There are also several smaller fibrous pericardium blends with the adventitia of the great
cardiac veins that drain directly into the right atrium (31). arteries. Posteriorly, it is more loosely attached to the
esophagus and descending aorta. The pericardial space is
limited superiorly by sleevelike attachments of the peri-
Pulmonary Arteries and Veins cardium to the great arteries just distal to their origins. The
The pulmonary artery is a short, wide vessel, about 5 cm in short transverse sinus lies behind the great arteries, ante-
length and 3 cm in diameter, which supplies deoxygenated rior to the atria and superior vena cava. Posteriorly, the
blood from the right side of the heart to the lungs. It com- pericardial space is also fixed at the insertions of the pul-
monly arises from the conus arteriosus of the right ventri- monary veins and venae cavae. This posterior series of
cle and extends obliquely upward and posterior, passing at venous attachments forms a U-shaped arc that is concave
first in front and then to the left of the ascending aorta. inferiorly and creates a blind cul-de-sac behind the left
The right branch of the pulmonary artery is longer and atrium (the oblique pericardial sinus).
larger than the left, runs horizontally to the right, posterior The heart consists of muscle fibers and fibrous rings
to the ascending aorta and superior vena cava and in front that serve for their attachment. It is covered by the visceral
of the right bronchus to the hilum of the right lung, where layer of serous pericardium and lined by the endothelial
it divides into two branches. The larger, lower branch con- membrane. Between these two membranes lies the mus-
tinues to the middle and lower lobes of the lung, and the cular wall, or myocardium. The myocardium is often
smaller, upper branch distributes to the upper lobe. The subdivided into endocardial (adjacent to the endothelial
left branch of the pulmonary artery is shorter and some- membrane), mid wall, and epicardial (adjacent to the
what smaller than the right, passes horizontally anterior to visceral pericardium) layers.
the descending aorta and left bronchus to the hilum of the
left lung, where it divides into two branches to supply each
lobe of the left lung. The ligamentum arteriosum, a rem- ISCHEMIC HEART DISEASE
nant of the ductus arteriosus, connects the pulmonary
artery to the concavity of the aortic arch. The ligament of Atherosclerotic heart disease remains one of the leading
the left vena cava joins the pulmonary artery to the left causes of morbidity and mortality worldwide. The accurate
upper pulmonary vein. detection of coronary artery disease with its functional
The pulmonary veins return oxygenated blood from consequences at an early stage can significantly decrease
the lungs to the left atrium of the heart. In general, there morbidity and mortality associated with the disease
are four, two from each lung. The vein from the middle process by guiding optimal patient management. Just as
lobe of the right lung generally unites with the vein from important is the accurate exclusion of individuals without
the upper lobe prior to entering the left atrium, but occa- significant coronary artery disease because unnecessary
sionally the three veins on the right side may remain invasive testing with their inherent risks can ultimately
separate. Not infrequently, the left-sided veins may open be avoided. Hence, noninvasive imaging techniques have
into the left atrium via a common ostium. At the hilum been developed and used extensively over the years for
of the lung, the superior pulmonary veins lie anterior and these purposes. The ultimate focus of nonimaging in
slightly inferior to the pulmonary artery. The right ischemic heart disease is two-fold: (a) anatomic imaging
pulmonary veins pass behind the right atrium and supe- to directly visualize the coronary arteries and associated
rior vena cava, and the left pulmonary veins are in front atherosclerotic plaques and (b) functional imaging to
of the descending aorta. assess the hemodynamic consequences of obstructive
Chapter 1: Heart 25
coronary artery plaques. Until recently, reliable noninva- 1.5 mm overlap has also been shown to be comparable to
sive anatomic imaging of the entire coronary artery tree EBCT for the quantification of coronary calcification (45).
was not possible because of limits in spatial resolution and Coronary artery calcium is assessed through the measure-
cardiac motion. Currently, noninvasive coronary angiogra- ment of the number of pixels in the CT image with a den-
phy can be performed using MR, MDCT, and EBCT. sity of 130 or more Hounsfield units (HU) to calculate a
Functional imaging has traditionally been performed total calcium score (44), a calcium volumetric score (46),
using stress nuclear [single photon emission computed or an absolute calcium mass (47) (Fig. 1-24). Each
tomography (SPECT) or positron emission tomography method has shown similar reproducibility (45,48,49), and
(PET)] and stress echocardiography techniques, but stress most software processing programs can easily provide all
MR protocols are becoming more available and, hence, three scores simultaneously. A high coronary calcium score
more frequent in clinical practice. is a sensitive but nonspecific marker for obstructive coro-
nary artery disease, and changes in the calcium score have
not been shown to correspond to changes in cardiovascu-
Coronary Artery Imaging lar event risk (43).
Goals in the assessment of coronary artery disease include
identification of flow-limiting coronary stenoses, and calci-
fied plaques, direct visualization of atherothrombotic Coronary Angiography
lesions, measurement of atherosclerotic burden, and char- Noncalcified atherosclerotic plaques in the coronary arter-
acterization of plaque components (38). Noninvasive ies can be visualized using contrast enhanced CT or MR.
imaging of the coronary arteries for anatomic evaluation of With the superior spatial resolution obtained by MDCT
atherosclerotic disease can be accomplished by MR or CT. (0.4 to 1 mm) compared with EBCT (1.5 mm), clinical
evaluation for noncalcified coronary atherosclerosis com-
monly utilizes MDCT technology. The general imaging
Calcium Scoring
protocol has already been described, and aggressive use of
Atherosclerotic coronary calcifications are often found in beta-blockers to decrease the HR to less than 65 beats per
patients with long-standing coronary atherosclerosis. minute has been shown to greatly improve image quality
Elevated levels of coronary calcification have been shown by diminishing image artifact related to cardiac motion
to correspond to increased risk of myocardial events (50–52). The high spatial resolution and soft tissue delin-
(39–42). Current clinical indications for quantification of eation may also provide information about the content
coronary calcium are in patients with atypical chest pain, of atherosclerotic plaques and coronary artery wall (53),
as well as asymptomatic patients with traditional cardio- including distinction between low-density lipid-rich (47
vascular risk factors (43). Noncontrast CT is currently the 9 HU) versus higher density fibrous (104 28 HU)
only method available to accurately quantify the coronary plaques (54,55). Several studies have demonstrated the
calcium plaque burden. EBCT using an ECG-triggered high sensitivity (72% to 99%) and specificity (86% to
acquisition of the entire heart performed with 3-mm con- 97%) and particularly the high negative predictive value
tiguous slices has been considered the “gold standard” for (97% to 100%) of MDCT for the detection of significant
assessment of calcified plaques (44). MDCT with retro- coronary artery disease (52,56–61) (Fig. 1-25). More im-
spective ECG triggering using 3-mm slice thickness with portantly, with an increasing number of detectors and faster
A, B, C
Figure 1-24 A–C: Coronary calcium scoring.
A B
Figure 1-25 A, B: Oblique multiplanar reconstructed CT images demonstrating calcified (white arrow in A), noncalcified (black arrow in
A), and mixed (arrowheads in B) atherosclerotic plaques in the right coronary artery.
gantry speeds leading to an improvement in image resolu- been made since the first clinical results were presented in
tion, there has been a significant increase in the number of 1993, and coronary MR imaging has matured from breath-
evaluable segments than with early-generation scanners. hold 2D gradient-echo imaging to navigator-based meth-
The presence of coarse calcifications with related bloom- ods using targeted 3D gradient-echo, followed by targeted
ing and streak artifact, however, still remains problematic 3D steady-state free precession, to most recently volumetric
in regard to accurate interpretation for degree of luminal “whole heart” approaches mimicking that of cardiac CT.
narrowing (Fig. 1-26). A recent meta-analysis evaluating coronary MR angiogra-
Coronary artery MR angiography has been one of the phy for detection of coronary artery disease demonstrated
most challenging areas of cardiovascular MR because of the moderately high sensitivity and specificity (73% and 86%,
small size and tortuosity of the coronary arteries, their near respectively) for evaluable vessel segments that were com-
constant motion during both respiratory and cardiac cycles, monly limited to the proximal and mid segments (63)
and the adjacent epicardial fat and myocardium (62). (Fig. 1-27). In the only international multicenter trial,
Numerous technical improvements and innovations have coronary MR angiography demonstrated high sensitivity
A B
Figure 1-26 A, B: Oblique multiplanar reconstructed CT images demonstrating calcified atheromatous plaques with associated bloom-
ing and streak artifact, limiting assessment for degree of luminal narrowing.
Chapter 1: Heart 27
A B
Figure 1-27 A, B: Oblique multiplanar reconstructed MR images using a steady-state free precession sequence that demonstrates
atheromatous plaques in the left anterior descending and left circumflex arteries (arrows), including a calcified plaque with artifact obscuring
lumen visualization (arrowhead).
(100%) and negative predictive value (100%) with modest Although coronary artery angiography (invasive or nonin-
specificity (85%) for detection of significant coronary vasive) can depict an anatomic view of the coronary arter-
artery disease in interpretable proximal and middle coro- ies for assessment of coronary artery disease, the effects
nary artery segments (64). Of concern is the fact that 69% that such plaques have on the myocardial territory sup-
to 100% of proximal and mid segments were considered plied by the affected artery is important. Atheromatous
interpretable in these studies, and the inclusion of the plaques that cause obstruction of less than 70% of the
unevaluable segments lowers the overall diagnostic accu- luminal diameter of the vessel rarely cause symptoms of
racy. Additionally, several groups have reported on MR of obstructive coronary artery disease. As the plaques grow in
the coronary vessel wall for plaque characterization. thickness and obstruct more than 70% of the vessel diame-
Although current data demonstrate the potential clinical ter, patients may experience symptoms of ischemic heart
utility and high potential usefulness of coronary vessel wall disease. These symptoms are often initially noted only dur-
imaging for evaluation of coronary artery disease (65–67), ing increased workload of the heart and manifest as exer-
visualization of plaque components such as the fibrous cap tional angina or decreased exercise tolerance. As the degree
is currently not possible because of limited spatial resolu- of luminal narrowing increases, there may be near
tion (68). Hence, based on the data to date, coronary artery complete obstruction of the coronary artery lumen, sever-
MR angiography for the identification of focal coronary ely restricting blood flow to the myocardium. Patients with
artery stenosis has been limited to the evaluation of left this degree of coronary artery luminal narrowing may
main or multivessel disease. suffer from myocardial infarction or have signs and symp-
In patients undergoing potential revascularization, it is toms of chronic coronary ischemia, including symptoms
important to report also on the status of plaques in the of rest angina and congestive heart failure.
ascending aorta and supra-aortic vessels due to the poten- The myocardium can exist in several different states,
tial risk of embolic stroke related to both catheter-based depending on the status of the supplying coronary artery.
and surgical procedures. Additionally, the presence of aor- Limitation of blood flow to the heart causes ischemia of
tic aneurysm or dissection in the proximal aorta should the myocytes. Prolonged lack of blood flow to the myocar-
also be noted for patients undergoing surgical revascular- dium eventually leads to myocyte death and myocardial
ization because bypass grafts will often be anastomosed in infarction. Functional imaging relies on the myocardial
the aortic wall. It is also useful to note anomalies in the assessment for regional perfusion or wall motion abnormal-
internal mammary arteries, which may be used for graft ities at rest or induced during stress. A myocardial territory
conduits, and concomitant aortic valve disease, which can may be described as normal, ischemic, stunned, hibernat-
be repaired at the time of surgery (69). ing, or infarcted (scarred). Ischemia induction is based on
the principle that although resting myocardial blood flow in
territories supplied by a stenotic coronary artery is normal
Functional Myocardial Imaging
with preserved regional function, the increased flow dem-
Functional imaging focuses on the assessment of the and during stress conditions cannot be met, resulting in a
hemodynamic consequences of coronary artery disease. sequence of events referred to as the ischemic cascade (70)
normally enhanced myocardium. After about 1 to 3 min-

utes, microvascular obstruction can be detected. After
Echocardiography/ about 10 minutes, depending on the amount of gadolin-
MRI
Angina ium administered, delayed enhancement in areas of infarc-
Ischemia
Nuclear/ ECG Changes tion or fibrosis can be demonstrated. Unlike standard

MRI SPECT or PET assessment of myocardial perfusion, the
Systolic Dysfunction
high spatial resolution (approximately 2 mm) of MR
Diastolic Dysfunction allows distinction between subendocardial and transmural
Transmural perfusion defects
perfusion defects (Fig. 1-29). Perfusion assessment is
ideally performed under both pharmacologic stress and
Subendocardial perfusion defects
rest conditions. Currently, physiologic stress testing is not
Time possible with MR. Stress MR perfusion has an overall
Figure 1-28 Ischemic cascade. sensitivity of 84% and specificity of 85% for the detection
of coronary artery disease (74) and appears to be more
sensitive for the detection of myocardial ischemia com-
(Fig. 1-28). Initially perfusion abnormalities are induced, pared with nuclear imaging techniques (75).
followed by diastolic and systolic dysfunction. Clinical Pharmacologic stress is often performed using a potent
symptoms of angina are a relatively late phenomenon of the vasodilator, either adenosine or dipyridamole. Adenosine
cascade. Hence, preclinical detection of inducible myocar- has the advantage of being very fast acting, with a half-life
dial ischemia is an important feature of noninvasive imag- of 4 to 10 seconds. Adenosine is typically administered at a
ing in the diagnosis of coronary artery disease. Additionally, rate of 140 mg/kg/min over a 4- to 6-minute infusion,
in the setting of recent ischemia or chronic ischemia, re- depending on the protocol. First pass perfusion imaging
gional myocardial function may become depressed, resulting using a bolus injection of gadolinium-based contrast agent
in stunned (transient dysfunction) or hibernating (pro- at a dose of 0.05 to 0.2 mmol/kg and a rate of 4 to
longed dysfunction) states without overt myocardial infarc- 7 mL/sec is usually performed after 3 minutes of adeno-
tion. Because stunned or hibernating myocardium may sine infusion, at which point maximum vasodilation and
normalize after revascularization of the involved regions, steady state have been achieved. Patients often experience
as opposed to regions with myocardial infarction, the dis- shortness of breath, chest tightness, and occasional palpi-
tinction between these myocardial states is an important tations or headaches, which resolve quickly once the med-
one (71–73). ication is stopped. Because gadolinium perfusion imaging
Currently, functional imaging can be performed using is performed during the adenosine infusion, two sites of
gated SPECT or PET, contrast stress echocardiography, and intravenous access or connector tubing setup need to be
MR, and depending on the imaging protocol utilized, an in place during the study. Dipyridamole has the advantage
assessment of perfusion and/or function can be made of being less expensive and easier to administer than
at rest and during stress conditions. In addition to evalua- adenosine, with fewer side effects, but it has a longer half-
tion for inducible ischemia, distinction between dysfunc- life. Dipyridamole is typically administered at a dose of
tional but viable (stunned or hibernating) myocardium 0.56 mg/kg over a 4-minute infusion. First pass perfusion
and dysfunctional, nonviable (scarred) myocardium can imaging is performed 7 minutes after the start of dipy-
be assessed with all these techniques, depending on the ridamole administration, when peak vasodilation has
imaging protocol. MDCT thus far has not played an occurred. In patients unable to tolerate the side effects of
important clinical role in functional imaging assessment dipyridamole, aminophylline (50 to 100 mg) can be given
of the myocardium, although investigational work is in to reverse the effects. Both stress agents should be avoided
progress. in patients with asthma or high-degree AV blocks. In addi-
tion, caffeine and medications containing aminophylline
and/or theophylline should be withheld for at least
Myocardial Stress Perfusion Assessment
24 hours prior to the exam because of their interaction
Myocardial perfusion assessment with MR relies on the with these agents. Hemodynamic and rhythm monitoring
concept that first passage of gadolinium contrast into the is required during administration of the stress agent.
heart reflects delivery at the myocardial level. Under nor- Because the goal of myocardial perfusion assessment is
mal circumstances, there is uptake of gadolinium contrast to measure the gadolinium concentration in the blood and
in all regions of the heart, which temporarily shortens T1 in the myocardium, single-shot images are acquired during
relaxation time and thereby increases the signal intensity each cardiac cycle to cover as much myocardium as possi-
of the perfused myocardium. In regions with inadequate ble (5 to 8 slices in short axis orientation) while maintain-
blood supply either from significant epicardial coronary ing high spatial and temporal resolution. Challenges with
artery stenosis or microvascular obstruction at the capillary this technique are to reduce the acquisition time per image
level, there will be a delay in contrast delivery, with result- to avoid blurring and artifact related to cardiac motion. The
ant dark or hypoenhanced regions when compared with use of parallel acquisition technology can help reduce
Chapter 1: Heart 29
D
Figure 1-29 Saturation recovery MR images at specific time points (A–C) during the first pass perfusion with gadolinium-containing con-
trast agent with corresponding time-intensity curves (D) of the ventricular cavity (1) and myocardium (2–4) demonstrate significant delay in
perfusion of the inferoseptal wall (2) compared with the corresponding inferolateral (3) anterior (4) walls.
acquisition time. Following image acquisition, in most Myocardial tagging techniques can be used to further
clinical settings, a qualitative visual assessment of perfusion depict regional strain for improved quantification of wall
is made (Fig. 1-29). This provides a quick and easy method motion abnormalities. A tomographic approach allows for
of image interpretation with overall sensitivity of 93% measurement of ventricular volumes without geometric
and specificity of 63%, which is comparable to SPECT. assumptions, resulting in accurate measurements for
Quantitative assessment in the ideal examination would be severely distorted ventricles. Similar to perfusion assess-
to measure the actual myocardial blood flow within each ment, cine MR images are obtained under pharmacologic
segment of the heart. This assessment can be made by ex- stress (primarily dobutamine) and rest conditions. Dobu-
amining the change in signal intensity over time, gradient tamine MR has an overall sensitivity of 83% to 89% and
of the upslope, peak signal intensity, and the time to peak specificity of 81% to 86% for the detection of significant
(Fig. 1-29). The myocardial perfusion reserve index can be coronary artery disease (79–81).
quantified by differences between myocardial perfusion re- Pharmacologic stress is commonly performed using
serve in the stress and rest studies (76–78). dobutamine because it has both chronotropic and inot-
ropic properties and mimics physiologic stress. Dobu-
tamine is commonly infused starting at a rate of 10 mg/
Myocardial Stress Systolic Function Assessment
kg/min and is gradually increased by 10 m/kg in 3-minute
Global and regional systolic LV function can be assessed by increments to a rate of 40 mg/kg/min. Atropine (up to
MR. Gradient-echo techniques demonstrate clear endocar- 2 mg) may be given if target HR is not achieved with dobu-
dial border definition, which is important in the assess- tamine infusion alone and the patient has no contraindi-
ment for regional wall thickening (Fig. 1-11). Myocardial cations to atropine. Evaluation for regional wall motion
tagging techniques can be used to track segmental motion abnormalities is made using cine MR sequences with or
and help in directly distinguishing impaired myocardium without myocardial tissue tagging during each stage of
from myocardium that may move abnormally owing infusion. Myocardial function and contractility should
to proximal tethering to a diseased area (Fig. 1-12). normally increase with each stage of infusion. Myocardial
segments with stress-induced (or worsened) wall motion volumes, aneurysms, intracavitary thrombus, concomitant
abnormalities reflect ischemia of the involved segments. mitral regurgitation), viability assessment by MR has
become a valuable tool for the evaluation of patients with
ischemic cardiomyopathy.
Myocardial Viability Assessment
Several studies have noted that LV dysfunction is not
The incidence of heart failure has been increasing over the necessarily an irreversible process but that improvement in
past several decades, in part due to improvements in car- LV function is possible after revascularization. Over the
diac survival. Coronary artery disease represents the under- years, many studies have focused on identifying dysfunc-
lying cause of heart failure in more than 70% of patients tional but viable myocardium to predict improvement
(82). Current therapeutic options for these patients of function post-revascularization. Although viable myo-
include optimization of medical treatment, heart trans- cardium theoretically includes the entire spectrum ranging
plantation, revascularization (with optional LV restoration from normal myocardium to epicardial regions of viable
and/or mitral valve repair), and cardiac resynchronization myocardium in non-transmural infarction, the term viabil-
therapy (83). Viability assessment is important in guiding ity is only important in regions with chronic contractile
therapeutic options because patients with viable myocar- dysfunction where there may be an improvement in func-
dium may improve in LV function and overall mortality tion with revascularization. This has often been referred to
after revascularization, whereas patients with only scar tis- as hibernating myocardium, a situation in which there is
sue often will not improve (84). In view of the potential chronic hypoperfusion (85) or repetitive stunning (86)
risks inherent in surgical and percutaneous revasculariza- that results in the chronic dysfunction. Several MR tech-
tion techniques, optimal selection of patients most likely niques have been proposed for the assessment of myocar-
to benefit from the procedure is warranted. Traditionally, dial viability. These techniques include resting cine MR for
viability assessment was performed using nuclear imaging evaluation of end-diastolic wall thickness (EDWT), low-
techniques (SPECT and PET) to assess myocardial metabo- dose dobutamine MR for evaluation of contractile reserve,
lism, perfusion, cell membrane and mitochondrial integ- first pass perfusion MR for evaluation of contrast uptake
rity, and echocardiography to assess contractile reserve. kinetics, and delayed contrast enhanced MR for evaluation
Given the high spatial resolution provided by MR, which of scar tissue (Fig. 1-30).
allows distinction between subendocardial and transmural The use of EDWT as a marker of viability is based on
processes, as well as the additional information needed the finding that in the presence of extensive transmural
to optimize revascularization strategies (LV function, infarction, significant wall thinning occurs (87). Recent
A, B
Figure 1-30 Different MR se-

quences used for the evaluation of
myocardial viability, including cine
images with (A) and without (B)
myocardial tagging for evaluation of
wall thickness and segmental func-
tion, first pass perfusion (C), and
delayed enhancement imaging (D).
There is an area of transmural
scarring noted in the anteroseptal
C, D region (arrows).
Chapter 1: Heart 31
studies have shown that segments with EDWT less than and transmurality of the scarred segment. Using a cutoff
5.5 mm virtually never show recovery of function post- value of 50% transmurality, this technique shows high
revascularization. However, segments with EDWT of sensitivity (91% to 99%) and low specificity (37% to 54%)
5.5 mm or more may or may not improve in function for predicting post-revascularization improvement in func-
post-revascularization because they frequently contain an tion (83). The suboptimal specificity is related to the pres-
admixture of subendocardial scar tissue with residual ence of segments with subendocardial infarct and epicar-
viability in the epicardial layers. Overall, however, using dial viability that do not improve function, and more
EDWT to predict functional recovery has a good sensitivity, information regarding the constitution of the epicardial
94% to 100%, but low specificity, 19% to 53% (88–90). regions is needed to predict actual functional improve-
In addition to EDWT, the presence of contractile reserve ment with revascularization (83).
is frequently used to detect viable myocardium. This can
be accomplished by the administration of low-dose dobu-
Special Considerations
tamine 5 to 10 mg/kg/min to assess for improvement of
contraction in dysfunctional myocardium. An increase in
systolic wall thickening greater than 2 mm during dobuta-
Assessment of Stent Patency
mine infusion has been shown to be a reliable marker for With the growth of nonsurgical revascularization tech-
post-revascularization improvement in function (90). niques, the increase in the number of patients with coronary
Overall, dobutamine MR demonstrates high specificity artery disease who receive coronary artery stents has been
with slightly lower sensitivity for assessment of contractile enormous (96,97). Hence, early identification of in-stent
reserve, with a mean sensitivity of 73% (50% to 89%) and stenosis is important in the management of these patients.
a mean specificity of 83% (70% to 95%) (83). MDCT has shown variable success for evaluation of in-stent
Hypoperfused regions on resting first pass perfusion stenosis, with steady improvements noted when using
studies have also been used to assess myocardial viability. newer 64-slice systems offering improvement in spatial
As previously mentioned, areas with inadequate blood resolution and using dedicated reconstruction kernels
supply will often not enhance as normal myocardium. (97–105). However, the variability of stent lumen visibility
Dysfunctional regions that display hypoenhancement on is still high, depending on the stent type, size, orientation,
resting first pass perfusion imaging showed high specificity and surrounding tissue (Fig. 1-31). Sharp reconstruction
(89%) but low sensitivity (19%) for predicting functional kernels in addition to the routine medium kernels may
recovery after revascularization (91–93). provide improvement in visible lumen diameter as well as
Myocardial contrast hyperenhancement of infarct more realistic intraluminal attenuation values (96,97).
regions, defined as increased signal intensity on delayed Because of their metallic composition, currently manufac-
resting T1-weighted MR images acquired more than 5 min- tured stents cause susceptibility artifacts that often obscure
utes after intravenous administration of contrast, was first the stent lumen in MR (106–108). However, MR-compatible
described more than 20 years ago. However, improve- stents based on copper alloy have been developed and are
ments in pulse sequence design have led to a nearly 500% being tested in animal models (109,110).
increase in signal intensity between abnormal hyperen-
hanced regions of myocardium and normal myocardium
(94). These regions of hyperenhancement show excellent
Assessment of Bypass Grafts
anatomic agreement with the histologic extent of necrosis Surgical revascularization of coronary artery disease is
(95). The mechanism underlying the hyperenhancement accomplished by coronary artery bypass grafting, in which
depends on the age of the infarction. In acute infarcts, a graft (arterial or venous) is used to bypass an occluded or
there is loss of sarcolemmal membrane integrity and rup- stenosed coronary artery. In comparison with the native
ture of myocyte membranes, which lead to increased tis- coronary arteries, reversed saphenous vein and internal
sue-level contrast concentration. In chronic infarcts, there mammary artery grafts are easier to visualize owing to the
is trapping of the contrast agent within the interstitial reduced overall motion, larger lumen, and less convoluted
space that is increased between collagen fibers commonly course. Familiarity with the common types of grafts used
seen in scar tissue, compared with densely packed myo- and placement can aid in both planning and interpreta-
cytes in normal myocardium, as well as delayed washout tion of images. A simple classification scheme is shown in
of contrast due to decreased capillary density. The major Table 1-4 and Figure 1-32. Venous conduits are generally
advantage of MR over other imaging techniques is the wider and longer than their arterial counterparts but have
superior spatial resolution that allows detection and differ- reduced long-term patency.
entiation of subendocardial infarction from transmural With the MDCT technique, special considerations must
infarction and normal myocardium. Because delayed be taken into account during data acquisition. Because most
enhancement MR represents anatomic visualization of scar patients have an in situ mammary artery graft, image acqui-
tissue, the accuracy of this technique for predicting func- sition must cover the entire thorax to visualize the proximal
tional recovery depends on the cutoff value used for extent anastomosis and origin of the internal mammary arteries.
A B
C D
Figure 1-31 Oblique multiplanar reconstructed CT images demonstrating patent (A, B) and occluded (C, D) stents.
Given the larger volume of coverage, adjustments in for diagnosing graft occlusion, although somewhat lower
contrast dose and scan delay timing after contrast adminis- sensitivity (75% to 82%) and specificity (88% to 92%) for
tration should be made to ensure adequate contrast detecting significant stenoses (Fig. 1-33) (111,112).
enhancement of the bypass grafts and native coronary artery Given a schematic knowledge of the origin and touch-
vessels during image acquisition. MDCT has been shown to down site of each graft in question, conventional free-
have high sensitivity (97% to 100%) and specificity (98%) breathing ECG gated 2D spin-echo and gradient-echo MR
TABLE 1-4
BYPASS GRAFT TYPESa
Type of Graft Description
Free grafts Venous (saphenous) or arterial (radial, internal mammary) excised grafts in which the origin is connected
to the proximal aorta and the distal end in the coronary artery beyond the blocked segment
In situ grafts Arterial (e.g., left internal mammary artery) graft in its normal origin with the distal end connected to the
coronary artery
Sequential bypass grafts Venous or arterial excised grafts used to supply two adjacent vessels (e.g., two diagonal arteries) with a
proximal anastomosis (side-to-side) in one vessel and a distal anastomosis (end-to-side) in the second
Y or T grafts Two bypass grafts, one of them an in situ or free bypass graft, and the second connected laterally to the
first and with the shape of an inverted Y or T
aSee Figure 1-32.
Chapter 1: Heart 33
A, B
C, D
Figure 1-32 A–D: Volume rendered CT images demonstrating different bypass graft types, including free saphenous vein grafts (black
arrowheads), in situ internal mammary artery grafts (white arrowheads), sequential bypass grafts (white arrows), and “Y” or “T” grafts (black
arrows). OG, occluded graft.
A, B
Figure 1-33 A–D: Curved multi-

planar reconstructed CT images of
bypass grafts demonstrating pate-
ncy of graft anastomosis site with
native vessel (arrows) and atheroma
C, D within the graft (arrowheads).
sequences in the transverse plane have been used to reli- techniques appear to provide excellent results for identify-
ably assess bypass graft patency with a sensitivity of 90% ing focal stenoses (122).
to 100% and specificity of 72% to 100% (113–116).
Patency is generally determined by visualization of a
patent graft lumen in at least two contiguous transverse OTHER MYOCARDIAL DISEASE
levels along its expected course (signal void for spin-echo
techniques and bright signal for gradient-echo techni-
Hypertrophic Cardiomyopathy
ques). If a patent graft lumen is not visualized at any level,
then the graft is considered occluded. Contrast-enhanced Hypertrophic cardiomyopathy refers to the condition in
MR techniques and 3D noncontrast approaches show which there is inappropriate myocardial hypertrophy in the
slight improvements in diagnostic accuracy (117–120), absence of an obvious cause (e.g., aortic stenosis or hyper-
and phase-velocity mapping or flow reserve assessment tension). Histologically, the disease is characterized by
may be superior to graft imaging (121). However, limita- disorganization and malalignment of the myofibrils (123).
tions of MR bypass graft assessment include difficulties The disease is genetically transmitted in about half of the
related to artifact due to implanted metallic objects such as cases. The natural history is variable but is a cause of sud-
hemostatic clips, ostial stainless steel graft markers, sternal den death due to ventricular arrhythmias (124). The mag-
wires, coexistent prosthetic valves with supporting struts or nitude of LV wall thickness has been shown to correlate
rings, and graft stents. Additionally, information on the with the risk for sudden death (125,126).
ability to identify severely diseased but patent grafts is Hypertrophic cardiomyopathy commonly affects the
limited, although targeted 3D coronary MR angiography upper portion of the interventricular septum, and contractile
Chapter 1: Heart 35
function is usually preserved. However, myocardial hypertro- Diagnosis of hypertrophic cardiomyopathy and associ-
phy may be diffuse or preferentially involve the mid portion ated LVOT obstruction can usually be made with echocar-
of the ventricular septum, the apex (Yamaguchi variant), diography. However, reliable quantitative delineation of
the lower portion of the septum, or the anterolateral or even LV wall thickness depends on adequate acoustic windows,
posterior free wall (127). In the typical form of hypertrophic and given the limited views obtained with echocardiogra-
cardiomyopathy, associated dynamic obstruction of the phy, cross-sectional images are often obtained with obliq-
LV outflow tract (LVOT) may be present (Fig. 1-34). uity (129). On the other hand, both CT and MR can pro-
Anterior movement and apposition of the mitral leaflet tips vide high-resolution, true short axis and long axis images
(commonly, the anterior leaflet) to the septum during sys- encompassing all levels and regions of the LV to allow for
tole contribute to the LVOT obstruction. In patients with complete reconstruction of the ventricular chamber and
severe apical hypertrophy, no LVOT obstruction is present, therefore the potential to detect the presence, distribution,
but a classic spade-shaped LV cavity is noted (Fig. 1-35). and severity of segmental wall thickening in any area of
Patients with midventricular hypertrophy may present with the LV wall. It is important that these studies be performed
midventricular obstruction during systole that may progress using cardiac gating to ensure quantitative measurement
to a noncontractile apical aneurysm (128). during ED to obtain a true measurement of resting wall
A,B
C, D
Figure 1-34 CT (A, B) and MR

(C–F) images at end-diastole (ED)
and end-systole (ES) from two differ-
ent patients with hypertrophic car-
diomyopathy, demonstrating the
asymmetric septal hypertrophy and
left ventricular outflow tract obstruc-
tion with increased flow turbulence
(arrowhead) that is commonly seen
in this subtype. In addition, delayed
enhancement inversion recovery
images (E, F) demonstrate hyperen-
hancement (arrows) in the area of
myocardial thickening, suggesting
E, F some degree of myocardial fibrosis.
A,B
Figure 1-35 CT (A, B) and MR (C,

D) images demonstrating a variant
of hypertrophic cardiomyopathy with
C, D apical hypertrophy.
thickness. On CT, simple hypertrophy manifests as an

Dilated Cardiomyopathy
asymmetric wall thickening, usually involving the proxi-
mal septum, and is often hypodense compared with nor- Dilated cardiomyopathy is a syndrome characterized by LV
mal myocardium (130). On MR, focal thickening often or biventricular enlargement and severely depressed sys-
demonstrates abnormal signal intensity on T1-weighted tolic dysfunction. There is usually a combination of hyper-
sequences, low signal intensity on T2-weighted sequences, trophy and atrophy of the myocytes seen morphologically.
and isointensity to muscle on gradient-echo sequences Clinical presentation is variable, but left-sided heart failure
(131). A recent study demonstrated that MR was capable of is the most common symptom. Although the cause is often
identifying regions of LV hypertrophy not readily recog- undefinable (idiopathic), a variety of cytotoxic, metabolic,
nized by echocardiography, improving the overall sensitiv- immunologic, familial, infectious, and valvular mecha-
ity for diagnosis of hypertrophic cardiomyopathy and also nisms can produce the clinical manifestations (138).
for detection of extreme LV hypertrophy (3 cm) (129). Diagnostic studies are focused on evaluating the etiol-
Cine and phase-velocity mapping MR can also be used to ogy of the disease process. Hence, an assessment of size of
assess the degree of LVOT obstruction and diastolic dys- the ventricular cavity, thickness of ventricular walls, and
function. With the use of myocardial tagging techniques, concomitant valvular and pericardial disease is important.
diffuse alterations of the myocardial contraction pattern In addition, the resultant effects of a dilated LV on valvular
can be shown (132–134). Gadolinium contrast adminis- function are important for further management. Although
tration may be useful in differentiating normal from disor- segmental wall motion abnormalities can be present with
ganized myocardial tissue, which appears as areas of dilated cardiomyopathy, diffuse global dysfunction is
abnormally high signal intensity in the LV myocardium, more typically seen.
commonly observed in regions where the right ventricle is MR and CT can depict morphologic and functional
attached, which seem to reflect myocardial ischemia and abnormalities associated with dilated cardiomyopathy. At
fibrosis due to small-vessel disease or myocardial degener- present, MR is probably the preferred technique to quantify
ation and necrosis (135). The presence and extent of ventricular volumes and functional parameters to measure
delayed contrast enhancement has been shown to corre- wall thickness, calculate RV and LV mass, and assess for the
late with increased incidence of ventricular tachycardia presence of RV enlargement. LV mass is significantly greater
and with decreased LV function (136, 137) (Fig. 1-34). in patients with dilated cardiomyopathy, and LV trabeculae
Chapter 1: Heart 37
are larger and more numerous compared with those in scarring. Additionally, patients with hypertrophic cardio-
healthy patients or those with ischemic heart disease myopathy may also demonstrate a restrictive pattern due to
(139,140). Several studies have demonstrated abnormal the increased stiffness of the myocardium.
delayed contrast enhancement in patients with dilated
cardiomyopathy, which may reflect areas of myocardial
Cardiac Sarcoidosis
inflammation, degeneration, necrosis, and fibrosis and pre-
dict the severity of LV dysfunction (141). Unlike the suben- Cardiac involvement is estimated to occur in 20% to 50%
docardial to transmural enhancement patterns ordinarily of patients with sarcoidosis, with sudden death due to ven-
seen in a coronary distribution found in ischemic heart dis- tricular tachyarrhythmias or conduction block accounting
ease, the enhancement patterns in dilated cardiomyopathy for 30% to 65% of deaths (143–145). The early use of cor-
are typically patchy, are mid wall, and do not follow a usual ticosteroid therapy has been shown to improve LV func-
coronary distribution (142). tion and prevent malignant arrhythmias (146). Because
lesions have a patchy distribution, endomyocardial biopsy
to show granulomatous infiltration of the myocardium
Restrictive Cardiomyopathy
may be negative.
Restrictive cardiomyopathy is characterized commonly as MR can provide anatomic and functional information
abnormal diastolic function. Hemodynamically, RV and about the heart that may be useful for the treatment and
LV filling pressures are increased as a consequence of monitoring of patients with sarcoidosis. Sarcoid infiltrates
increased myocardial stiffness. Contractile function may or on MR appear as zones with increased signal intensity that
may not also be impaired. Often restrictive cardiomyopa- are more pronounced on T2-weighted images because of
thy resembles constrictive pericarditis, a condition charac- the associated edema that occurs with inflammation. In the
terized by normal or nearly normal systolic function with nodular type of sarcoidosis, MR may further highlight the
abnormal ventricular filling due to pericardial constraint. central low signal intensity on both T1- and T2-weighted
The distinction between restrictive cardiomyopathy and images (representing hyaline fibrotic tissue) and a periph-
constrictive pericarditis can be difficult but is important eral area with high signal intensity on T2-weighted images
because the latter can often be cured with surgical resec- (representing the edema) (146,147). Furthermore, granulo-
tion of the pericardium, whereas restrictive cardiomyopa- matous lesions have been shown to enhance with gadolin-
thy is managed medically. Clinical imaging with CT and ium contrast administration (Fig. 1-36) (148–151).
MR can often demonstrate pericardial disease involvement
to aid in the distinction.
Cardiac Amyloidosis
A variety of pathologic processes may result in restrictive
cardiomyopathy, although the cause is often unknown. Amyloidosis represents a diverse group of diseases charac-
It can be seen in the setting of myocardial fibrosis, infil- terized by abnormal deposition of amyloid, twisted
tration, myocardial storage disorders, or endomyocardial b-pleated sheet fibrils that may be produced in a variety of
A B
Figure 1-36 A, B: Cardiac sarcoidosis with delayed enhancement inversion recovery images demonstrating mid wall hyperenhancement
(arrows), which is more suggestive of nonischemic fibrosis.
pathologic states. Severity of the disease depends on the anemia. Chelation therapy may delay or prevent the devel-
composition of the amyloid fibril. Accumulation of amy- opment of heart failure.
loid in various organs disrupts tissue architecture and can Initial cardiac findings in iron overload cardiomyopa-
lead to severe organ damage. Amyloidosis commonly thy are a mild increase in LV wall thickness and ED cham-
involves the kidneys, heart, liver, or peripheral nerves. The ber diameter. Diastolic dysfunction usually precedes
primary form of the disease, designated AL amyloidosis, systolic dysfunction with a restrictive filling pattern (157).
consists of light-chain immunoglobulin produced by MR imaging can be used to identify the presence of iron in
monoclonal plasma cells and represents the most com- the heart and potentially to quantify the iron levels based
mon and most severe form of the disease (152). Familial on image signal intensity or calculated T2 relaxation time
amyloidosis is a fairly uncommon autosomal dominant (158). The presence of iron within a tissue disturbs the
disease resulting in the production of unstable serum magnetic field homogeneity and enhances T2 relaxation.
protein transthyretin (ATTR amyloidosis) that leads to MR techniques that are sensitive to tissue T2 changes will
neurologic and/or cardiac dysfunction (33,152). Senile be able to detect the marked decrease in signal intensity on
systemic amyloidosis results from the deposition of wild- T2-weighted images. The degree of signal attenuation is
type transthyretin and is almost exclusively limited to the related to intrinsic tissue iron levels, but not serum ferritin
heart. Reactive systemic amyloidosis is caused by overpro- concentration. Calculation of the T2 relaxation time is the
duction of nonimmunoglobulin protein AA and is rarely best independent parameter to detect the iron content
associated with cardiac involvement (153). Although directly in the heart and can be used for monitoring ther-
effective treatments for certain types of amyloidosis are apy (159,160,160a).
available, options are limited once cardiac involvement
becomes clinically apparent. Cardiac involvement not
Endomyocardial Disease
only signals poor prognosis but also predicts poor toler-
ance to high-dose chemotherapy regimens and stem cell Endomyocardial diseases include Löffler endocarditis and
transplantation (154). endomyocardial fibrosis, and are a common form of sec-
Echocardiography is the most commonly used test to ondary restrictive cardiomyopathy. Löffler endocarditis is
noninvasively diagnose cardiac amyloidosis. Typical find- an aggressive, rapidly progressive disease seen in temperate
ings include small LV size, biventricular and atrial septal climates and related to hypereosinophilia. The underlying
thickening, atrial enlargement, and in late stages, a restric- pathogenesis is not completely understood, but the hyper-
tive LV filling pattern (155). However, early structural and eosinophilia causes a toxic effect on the heart, leading to
morphologic changes are not often seen on echocardiogra- necrosis, endomyocarditis, and formation of intramyocar-
phy, and hence it has limited impact on guiding treatment dial thrombi, and in the later stages, localized or extensive
decisions. replacement fibrosis that produces thickening of the
Contrast-enhanced MR has recently been shown to be of ventricular wall (161,162). Endomyocardial fibrosis most
value in the diagnosis of cardiac amyloidosis and the deter- commonly occurs in equatorial Africa and has no relation
mination of prognosis in these patients, and in particular, to hypereosinophilia. Instead, there is the development of
distinguishing them from patients with LV hypertrophy. intensive endocardial fibrotic wall thickening of the apex
Patients with cardiac amyloidosis are commonly found to and subvalvular regions of one or both ventricles, resulting
have qualitative global and dominant subendocardial distri- in inflow obstruction of blood (163).
bution of gadolinium enhancement of the myocardium on CT and MR can be helpful in diagnosing these entities
T1-weighted IR sequences that matches the transmural and evaluating the effects of medical treatment. Areas of
distribution of amyloid protein (156). Additionally, wash- myocardial fibrosis may appear to be low density on
out of gadolinium from the blood and myocardium is often CT images (164,165). MR can demonstrate high signal
faster, which necessitates earlier imaging and shorter inver- intensity on delayed enhancement IR imaging (162,166).
sion time for visualization of late gadolinium enhancement
in comparison with normal (156).
Arrhythmogenic Right Ventricular
Cardiomyopathy
Iron Overload Cardiomyopathy
Arrhythmogenic RV cardiomyopathy is defined as a pri-
Hemochromatosis can lead to increased iron deposition in mary disorder of the right ventricle characterized by partial
the heart with resultant systolic or diastolic cardiac dys- or total displacement of the myocardium by fibrous or
function that cannot be explained by another process. fatty tissue (167,168). It usually manifests in young people
Primary hemochromatosis is an autosomal recessive dis- as syncope, sudden death from ventricular arrhythmias, or,
ease that usually presents late in life. Secondary hemochro- less commonly, right-sided heart failure. There is a male
matosis is a common cause of mortality in the young adult predominance, and symptoms often occur during exercise.
population related to chronic blood transfusions for Uhl anomaly is an extreme form of arrhythmogenic RV
hereditary anemias such as thalassemia or sickle cell cardiomyopathy known as parchment heart because of the
Chapter 1: Heart 39
paper-thin RV wall characteristic of this disease. Because depiction of small focal aneurysms. The arrhythmic nature
arrhythmias are often of RV origin, arrhythmogenic RV car- of the disease often leads to image degradation. Contrast
diomyopathy must be distinguished from RVOT tachycar- CT findings include (a) abundant epicardial fat, (b) low-
dia, which has a more benign prognosis. attenuation prominent trabeculations, (c) scalloping of
Diagnosis of arrhythmogenic RV cardiomyopathy relies the RV free wall, and (d) intramyocardial fat deposition
on the demonstration of intramyocardial fibrofatty depo- (170). Quantification of RV volumes and visual assessment
sition, wall thinning, and abnormal wall motion. In ar- using cine CT can demonstrate regional dysfunction
rhythmogenic RV cardiomyopathy, the fibrofatty replace- and depressed global RV function (170). MR evaluation
ment almost exclusively involves the RV free wall, with in arrhythmogenic RV cardiomyopathy includes both
septal and LV involvement being rare. Later stages of the morphologic and functional assessment (Fig. 1-37). T1-
disease are associated with focal or diffuse myocardial wall weighted spin-echo sequences can provide some tissue
thinning with bulging of the ventricular wall in association characterization and may show abnormal high signal inten-
with akinetic or dyskinetic wall motion. Because the fibro- sity within the myocardial wall characteristic of intramy-
fatty replacement is often patchy in nature, endomyo- ocardial fatty deposits. Because the RV free wall is thin by
cardial biopsy may miss the disease. Hence, diagnosis is nature, techniques to maximize the spatial resolution, such
based on major and minor criteria that include structural, as a small field of view, can be used. Saturation bands can
histologic, ECG, arrhythmic, and genetic features of the be used to minimize flow signal artifact from the ventricu-
disease process (Table 1-5) (169). lar blood pool. Cine MR sequences used in conjunction
Echocardiography may sometimes demonstrate RV with spin-echo sequences can depict focal or diffuse abnor-
enlargement and dysfunction. However, CT and MR have malities in RV size and regional and global systolic and
the capacity to better assess the RV free wall in multiple diastolic function. Similar morphologic and functional
imaging planes with higher spatial resolution to assess for abnormalities have been reported in patients with RVOT
focal RV wall thinning and dysfunction, including the tachyarrhythmias in the absence of arrhythmogenic RV
TABLE 1-5
DIAGNOSTIC CRITERIA FOR ARRHYTHMOGENIC RIGHT VENTRICULAR CARDIOMYOPATHYa
Major Criteria Minor Criteria
Global and/or regional Severe dilation and reduction of RV Mild global RV dilation and/or
dysfunction and structural ejection fraction with no (or only ejection fraction reduction with normal LV
alterations mild) LV impairment Mild segmental dilation of the RV
Localized RV aneurysms (akinetic or Regional RV hypokinesia
dyskinetic areas with diastolic bulging)
Severe segmental dilation of the RV
Tissue characterization Fibrofatty replacement of myocardium
of walls on endomyocardial biopsy
Repolarization Inverted T waves on right precordial leads
abnormalities (V2 and V3) (age 12 yr; in absence of right
bundle branch block)
Depolarization/conduction Epsilon waves or localized prolongation Late potentials (signal-averaged ECG)
abnormalities (110 ms) of the QRS complex in right Left bundle branch block–type ventricular
precordial leads (V1-V3) tachycardia (sustained and nonsustained)
(ECG, Holter, exercise testing)
Arrhythmias Frequent ventricular extrasystoles
(1,000/24 hr) (Holter)
Family history Familial disease confirmed at necropsy Familial history of premature sudden death
or surgery (35 yr) due to suspected RV dysplasia
Familial history (clinical diagnosis based on
present criteria)
LV, left ventricular; RV, right ventricle/ventricular; ECG, electrocardiogram.

aTo fulfill the appropriate criteria for arrhythmogenic right ventricular cardiomyopathy, patients must meet either two major criteria (one major plus
two minor criteria) or four minor criteria.
Adapted from McKenna WJ, Thiene G, Nava A, et al. Diagnosis of arrhythmogenic right ventricular dysplasia/cardiomyopathy. Task Force of the
Working Group Myocardial and Pericardial Disease of the European Society of Cardiology and of the Scientific Council on Cardiomyopathies of the
International Society and Federation of Cardiology. Br Heart J 1994;71: 215–218, reproduced with permission from the BMJ Publishing Group.
A,B,C
Figure 1-37 CT (A) and T1-weighted turbo spin-echo MR (B, C) images demonstrate the presence of fatty infiltration (arrows) of the right
ventricular free wall and septum with areas of microaneurysms (arrowheads) typical for arrhythmogenic right ventricular cardiomyopathy.
cardiomyopathy, but the abnormalities are usually limited damage that appear as hyperenhanced regions on
to the lower infundibular part of the right ventricle (171). delayed enhancement IR sequences (Fig. 1-38) (172–175).
Although the finding of intramyocardial fat in the RV free Pericardial enhancement may also be seen if it is involved
wall can be demonstrated on MR and CT imaging (Fig. in the disease process. In addition to diagnosis and moni-
1-37), the presence of this finding alone (in the absence of toring of disease activity, MR can be useful in guiding
wall motion abnormalities) has limited sensitivity and endomyocardial biopsy, which is needed to determine
specificity for the diagnosis of arrhythmogenic RV car- etiology (176).
diomyopathy (170). Hence, MR and CT are often used to
provide information related to RV size, global and regional
function, and aneurysm formation and not tissue charac- PERICARDIAL DISEASE
terization as outlined in the diagnostic criteria for arrhyth-
mogenic RV cardiomyopathy (Table 1-5) (169). Diseases of the pericardium encompass a spectrum of disor-
ders, including congenital malformations and infection-
related, infarction-related, metabolic, autoimmune, trau-
Myocarditis
matic, neoplastic, and idiopathic disorders. Depending on
Myocarditis is an inflammatory infiltrate of the my- the disease process, clinical manifestations of pericardial
ocardium that leads to necrosis or degeneration of the adja- involvement can vary. Suspected pericardial disease is usu-
cent myocytes in a distribution not typical of ischemic ally initially evaluated with echocardiography. However,
damage associated with coronary artery disease (172). MR and CT imaging can also provide additional valuable
Clinical presentation is variable with patients, who may information.
present with profound ventricular dysfunction that pro- Both CT and MR have been widely accepted for the
gresses to dilated cardiomyopathy or, more insidiously, evaluation of structural changes in the pericardium. The
with chronic persistent myocarditis with limited ventricular best quality images in the evaluation of pericardial disease
dysfunction despite foci of myocyte necrosis. Myocarditis are obtained with the use of cardiac gating and fast imag-
can be found in a variety of disease processes affecting the ing to minimize motion blurring, although more promi-
myocardium and/or other parts of the heart (pericardium), nent pericardial disease findings may be evident even on
including infection or inflammatory or allergic reactions. conventional studies performed for other indications.
Acute myocarditis is most commonly due to a virus infec- On CT, the pericardium is usually well delineated from
tion (e.g., coxsackievirus). However, myocarditis has also the adjacent low-attenuation fat. Anatomic features of peri-
been seen in association with Lyme disease, Chagas disease, cardial disease, such as pericardial thickening, pericardial
Wegener granulomatosis, and systemic lupus erythemato- calcification, and pericardial effusion or masses, are easily
sus (173). Endomyocardial biopsy is often required for detected and evaluated with CT. Limited tissue characteriza-
classification but, because of the patchy nature of involve- tion of pericardial fluid or masses can make it difficult to
ment, may miss the disease. differentiate thickened pericardium from an exudative peri-
Although its role is still undetermined, MR can be useful cardial effusion with high protein content. Functional and
for the diagnosis of myocarditis. The increase in myocardial hemodynamic information is also limited with CT.
water content due to interstitial edema related to the MR offers several additional imaging options and
inflammation will lead to a hyperintense signal appear- considerations in the evaluation of pericardial disease.
ance of the myocardium on T2-weighted sequences. Anatomic information about the pericardium can be
Furthermore, administration of gadolinium contrast agents obtained from standard static imaging of the heart, and
can be helpful in identifying the regions of myocardial functional information demonstrating the effects of
Chapter 1: Heart 41
A,B,C
Figure 1-38 MR images from a patient with myocarditis, demonstrating diffuse myocardial hyperintensity noted on T2-weighted turbo
spin-echo (A), inversion recovery (B), and postcontrast delayed enhancement inversion recovery (C) images.
pericardial disease on cardiac motion can be obtained with both parietal pericardial stiffness and thickness and
from dynamic cine imaging (177). The pericardium is rate of accumulation, the size as well as location of peri-
typically seen as a dark stripe between the bright layers of cardial effusions is an overall predictor of prognosis.
fat that surround the pericardium and epicardium. Static Although a few approaches are available to quantify the
T1-weighted or T2-weighted imaging may not produce volume of the effusion, in clinical practice, such quantifi-
images with the expected relative signal strength from the cation is generally not necessary, and qualitative assess-
pericardial fluid owing to motion of the fluid within the ment of the effusion size (small, moderate, or large) is
pericardial space, which may limit fluid characterization. sufficient. Clinical presentation, history, and ECG often
Dynamic cine MR imaging demonstrates the motion of can give clues to the presence of significant effusion.
the heart within the pericardial space and can help bring Echocardiography is generally used to evaluate the nature,
out the presence of effusions because of their increased sig- size, and hemodynamic consequences of pericardial
nal intensity. As pericardial fluid may sometimes have effusions, to aid in medical or surgical treatment and
intensity similar to that of the pericardium, depending on follow-up. However, diagnostic pitfalls with echocardiog-
the imaging technique, care must be taken to distinguish raphy include small loculated effusions, hematoma, cysts,
pericardial effusion from pericardial thickening. MR imag- tumors, hernias, lipodystrophia, left pleural effusion,
ing is less sensitive than CT in the detection of pericardial mitral annular calcification, giant left atrium, epicardial
calcification, which often appears as irregularly thickened fat, and inferior left pulmonary vein and LV pseudo-
low-signal areas of the pericardium. aneurysm. In cases in which the diagnosis is not clear, or to
gain further insight into etiology, CT and MR can provide
additional anatomic information as well as some fluid
Pericardial Effusion
characterization and imaging of associated malignant
Pericardial effusion occurs when there is a significant or infectious lesions.
accumulation of fluid (50 mL) within the pericardial The CT and MR appearance of pericardial effusion is
space. Common causes of pericardial effusion include often not very specific for a particular etiology, and further
heart failure, renal insufficiency, infection (bacterial, evaluation is usually necessary to help narrow the range of
viral, or tuberculous), neoplasm (carcinoma of lung or differential possibilities. CT attenuation measurements
breast, or lymphoma), myxedema, and injury (trauma or can provide initial characterization of pericardial fluid.
myocardial infarction). Pericardial effusions can be cir- The appearance of pericardial effusion on MR images is
cumferential, encompassing the entire heart, or localized, affected by fluid content and motion, which tend to mod-
as can sometimes occur when focal scarred areas are pres- ify the effects of the fluid’s relaxation times.
ent, often noted after cardiac surgery. With increasing On CT, simple effusions have attenuation similar to
fluid accumulation, an elevation in intrapericardial pres- that of water. On MR, simple transudative effusions will
sure results, which can lead to cardiac tamponade from tend to look dark on T1-weighted imaging and bright on
compression of the cardiac chambers. Although the vol- T2-weighted imaging because of long T1 and T2 relaxation
ume of fluid needed to cause tamponade varies inversely times. On cine MR imaging, simple pericardial effusions
will generally have an appearance similar to that of the therapy, post–cardiac surgery, and remote nonpenetrating
blood within the ventricular cavity: bright due to through- or penetrating cardiac trauma have been frequently seen.
plane motion for T1-weighted fast gradient-echo imaging Diagnosis requires an astute clinician with a high index of
and bright on steady-state free precession imaging due to suspicion when evaluating a patient with signs of heart fail-
the high T2/T1 ratio of fluid. ure without evidence of valvular or myocardial disease.
In contrast, more proteinaceous fluid (exudates or Accurate diagnosis and differentiation from restrictive
inspissated fluid collections) will have attenuation greater cardiomyopathy is important because curative treatment
than water on CT (178). On MR, these exudative effusions for constrictive pericarditis is possible with surgical peri-
will look bright on T1-weighted imaging and dark on T2- cardiectomy. Initial diagnostic work-up usually includes
weighted imaging because of shorter T1 and T2 relaxation 2D, Doppler, and tissue Doppler echocardiography with
times (179). Fibrinous and hemorrhagic components of analysis of respiratory changes with or without changes of
effusions will also affect the imaging appearance. preload. However, equivocal findings may be present in up
Depending the age of the collection, bloody effusions can to one third of patients with possible pericardial constric-
appear similar to intravascular blood if fresh but often ap- tion, and further testing is usually required.
pear bright on T1-weighted images if older or composed of CT and MR allow for direct visualization of the
mixed serosanguineous collections because of the T1 pericardium for measurement of pericardial thickness
shortening effects of methemoglobin formation. As previ- (Fig. 1-39). However, the finding of a thickened peri-
ously mentioned, loculated fluid will be seen as locally cardium is not necessarily confirmatory of constrictive
confined collections with minimal fluid motion; identifi- pericarditis because up to 20% of patients may present
cation of the presence, size, location, and effects on cardiac with pericardium of normal thickness on current imaging
motion of such loculations can be helpful in deciding on methods (181). CT and MR can also demonstrate character-
therapy. CT and MR characteristics of malignant effusions istic features such as a conical or tubular diastolic ventricu-
may be similar to those of hemorrhagic effusions, depend- lar shape, an associated pericardial effusion, and secondary
ing on blood content, and are often associated with an sequelae, including dilated atria, hepatic veins, and pul-
irregularly thickened pericardium or pericardial nodular- monary veins, which can further support the diagnosis of
ity. In addition, because of the wide field of view, associ- constrictive pericarditis. CT can also demonstrate pericardial
ated abnormalities of the mediastinum and lungs may calcification, and cine MR imaging, including the use of
also be detected during the examination (180). dynamic tissue tagging technique, can characterize global
and regional systolic and diastolic function and assess for
pericardial adhesions and myocardial tethering (Fig. 1-39).
Acute Pericarditis
The use of phase-velocity mapping can also demonstrate
Acute pericarditis may be dry, fibrinous, or effusive, inde- abnormal flow patterns in the superior vena cava in patients
pendent of the underlying etiology. Diagnosis is suspected with constrictive pericarditis (182).
on the basis of clinical symptoms, including a prodrome
of fever, malaise, and myalgia followed often by retroster- Pericardial Masses
nal or left precordial chest pain (often pleuritic), nonpro-
The differential diagnosis of pericardial masses includes
ductive cough, and shortness of breath. Serologic markers
pericardial cyst, hematoma, neoplasm, loculated effusions,
of inflammation can often support the diagnosis. Peri-
and pseudoaneurysms. Although often detected initially
carditis is often accompanied by some degree of myocardi-
with echocardiography, evaluation with CT and MR can be
tis. CT and MR imaging can be used to aid in the diagnosis,
helpful in diagnosis and treatment. CT and MR tissue char-
especially when other tests prove inconclusive. Structural
acteristics, such as signal intensity, degree of contrast en-
changes of the pericardium such as thickening, inflamma-
hancement, and presence or absence of blood flow into
tion, pericardial effusion, and associated myocarditis or
the mass, can help differentiate among pericardial masses.
other concomitant heart disease and mediastinal pathol-
In addition, the size, anatomic extent, associated lesions,
ogy can be demonstrated.
vascularity, and effects on cardiac function of the masses
can be clarified. Furthermore, CT and MR can guide other
Constrictive Pericarditis diagnostic testing, such as biopsy, and facilitate treatment
and follow-up.
Chronic inflammation of the pericardium can lead to con-
strictive pericarditis, which is characterized by impaired
Cysts
ventricular filling. The process may extend to involve the
underlying myocardium, resulting in reduced ventricular Pericardial cysts may be congenital or acquired. Most cysts
function. The etiology of constrictive pericardial disease has are asymptomatic and are often detected incidentally on
often been attributed to antecedent acute pericarditis due chest radiography or echocardiography. However, patients
to infectious, inflammatory, or idiopathic causes. However, may also present with chest discomfort, dyspnea, cough,
other etiologies such as neoplastic diseases, post–radiation or palpitations due to compression of surrounding cardiac
Chapter 1: Heart 43
Figure 1-39 CT (A) and MR (B–D)

images demonstrating thickened
(black arrows) and calcified (white
arrows) pericardium with conical
deformation of the ventricular cham-
bers (C) typically seen in constrictive
pericarditis. Additionally, there may
be a lack of pericardial thickening,
and patients may present with adhe-
sive constrictive pericarditis charac-
terized by myocardial tethering of
the pericardium (arrowheads).
structures. The cysts can also become secondarily infected. hematomas demonstrate homogeneous high signal inten-
Congenital pericardial cysts usually have thin, smooth sity, whereas subacute hematomas that are 1 to 4 weeks
walls without internal septa and are formed when a por- old typically show heterogeneous signal intensity, with
tion of the pericardium is pinched off during early devel- areas of high signal intensity on both T1-weighted and
opment. Inflammatory cysts comprise pseudocysts as well T2-weighted images. On T1-weighted and gradient-echo
as encapsulated and loculated pericardial effusions caused images, chronic organized hematomas often show a dark
by infection, rheumatic pericarditis, trauma, or cardiac sur- peripheral rim and low signal intensity foci within
gery. Echinococcal cysts usually originate from ruptured that may represent calcification, fibrosis, or hemosiderin
hydatid cysts in the liver and lungs. On CT, simple pericar- deposition. High signal intensity areas may be seen on
dial cysts have the same attenuation as water (180). On T1-weighted or T2-weighted images, corresponding to
MR, pericardial cysts with simple fluid collections typically hemorrhagic fluid. Coronary or ventricular pseudoaneury-
have low or intermediate signal intensity on T1-weighted sms or neoplasms may resemble hematomas on static
images and homogeneous high signal intensity on anatomic MR images. However, because hematomas often
T2-weighted images (Fig. 1-40). Cysts containing highly do not demonstrate flow and consequently do not enha-
proteinaceous fluid may have greater attenuation on CT nce, the administration of gadolinium chelate contrast
and high signal intensity on T1-weighted MR images; com- agents allows differentiation of these entities. Additionally,
plicated cysts may be more heterogeneous. Because of their velocity-encoded cine MR imaging may be used to detect
encapsulated nature, they often do not enhance with the internal flow in pseudoaneurysms and thus further differ-
intravenous administration of contrast. Depending on entiate pseudoaneurysms from hematomas.
the etiology, pericardial cysts can be found anywhere in
the mediastinum, although congenital pericardial cysts
Neoplasms
commonly occur along the right cardiophrenic angle.
Pericardial metastases are much more common than pri-
mary pericardial tumors (183). Metastatic spread of tumor
Hematomas
may be via the lymphatic or blood system or through direct
Pericardial hematoma often develops after cardiac invasion from the lung, mediastinum, or breast. Breast and
surgery or myocardial infarction. On MR imaging, acute lung cancers are the most common sources of metastases,
A,B

D) images from two different pa-
tients with congenital pericardial
cyst (*) noted along the border of
the right side of the heart. The cyst
demonstrates hypointensity on T1
(C) and hyperintensity on inversion
recovery (D) images, which is char-
C, D acteristic for serous fluid.
usually by direct invasion, followed by lymphomas and teratoma, and hemangioma. Malignant neoplasms of the
melanomas. CT and MR are useful in the evaluation for pericardium include mesothelioma, lymphoma, sarcoma,
pericardial metastases because of their ability to visualize and liposarcoma. These neoplasms have characteristic
the entire pericardium. Evidence of metastatic involvement findings on CT and MR and have been discussed else-
may be suggested by an irregularly thickened pericardium, where. Primary mesothelioma of the pericardium may be
pericardial mass, or associated pericardial effusion. manifested as pericardial effusion, occasionally accompa-
Visualization of an intact pericardial line may be observed nied by pericardial nodules or plaques. Malignant pleural
if an adjacent tumor extends to the pericardium but not mesothelioma may also be seen to directly invade the peri-
through it. Consequently, local tumor invasion of the pericardium. Lymphoma, sarcoma, and liposarcoma typically
cardium may be recognized by focal obliteration of the appear as large heterogeneous masses, frequently associ-
pericardial line and the presence of associated pericardial ated with serosanguineous pericardial effusion. Given the
effusion. Hemorrhagic pericardial effusions secondary to overlap of characteristics between the different types of
metastases usually have attenuation greater than water on malignant neoplasms, biopsy and histopathologic analysis
CT and high signal intensity on T1- and T2-weighted are often necessary for definitive diagnosis.
images (179). Most malignant neoplasms are characterized
by low but heterogeneous attenuation on CT, low signal
Congenital Pericardial Lesions
intensity on T1-weighted images, and high signal intensity
on T2-weighted images (184). Metastatic melanoma, Congenital defects of the pericardium comprise partial
however, often demonstrates high signal intensity on left, right, diaphragmatic, or total absence of the peri-
T1-weighted images due to paramagnetic metals bound to cardium. Most patients are asymptomatic, although approx-
melanin (185). In addition, the use of contrast agents can imately 30% may have additional congenital abnormali-
aid in diagnosis because sites of malignant disease often ties. Complete absence of the pericardium poses a risk of
show significant contrast enhancement (186). homolateral cardiac displacement, amplified heart mobil-
Primary pericardial neoplasms are rare. Benign ity, and an increased risk of traumatic aortic type A dissec-
neoplasms of the pericardium include lipoma, fibroma, tion. Partial left-sided defects can be complicated by
Chapter 1: Heart 45
cardiac strangulation caused by herniation of the left atrial MR imaging, anatomic and morphologic assessment is per-
appendage, atrium, or left ventricle through the defect. formed using T1-weighted, T2-weighted, fat suppression,
Although the diagnosis can be suggested by chest radiogra- and IR cardiac gated spin-echo sequences performed before,
phy and echocardiography, definitive diagnosis and com- during (first pass perfusion), and after administration of
plete evaluation of the pericardium usually require CT or intravenous gadolinium contrast. Additional morphologic
MR imaging (Fig. 1-41). as well as functional consequences of the mass are obtained
using gradient-echo cine imaging.
CARDIAC MASSES
Benign Cardiac Neoplasms
Cardiac masses may be divided into benign or malignant
Benign cardiac masses include myxoma, papillary fibro-
lesions. Benign cardiac tumors include cardiac neoplasms,
elastoma, lipoma, fibroma, rhabdomyoma, teratoma, and
pseudotumors such as thrombi, and normal structures that
hemangioma. Benign tumors tend to affect the left-sided
resemble masses. Malignant cardiac tumors include both
chambers of the heart. Although these lesions are histo-
primary and secondary malignancies that can affect the
logically benign, they can act malignantly via secondary
heart. Echocardiography is often the initial diagnostic test
effects on the heart and vasculature. Therefore, successful
to evaluate cardiac masses but often provides limited
treatment depends on the early detection and characteriza-
information on tissue characterization and extent of extra-
tion of these masses. Whereas CT provides important
cardiac involvement or presence of metastatic disease.
information on detection of calcification and extent of
Therefore, cross-sectional imaging with CT and MR can
disease, MR offers unsurpassed soft tissue characterization.
provide additional information, including the precise
location of cardiac tumors (i.e., paracardiac, mural, or
intracavitary), the extent of disease, the presence of associ-
Myxoma
ated effusions, and the presence of metastases (187).
Additionally, CT and MR imaging findings that include Myxomas account for nearly half of the primary benign
tissue characterization may suggest the tumor type, which cardiac tumors (188). They occur in all age groups but are
can be important in treatment planning (Table 1-6). particularly found between the third and sixth decades of
Depending on the size and location of the cardiac mass, life and more commonly in women. Myxomas are usually
CT and MR image acquisition is tailored to obtain the nec- found in the left side of the heart but can also been seen in
essary information. Noncontrast CT imaging can provide the right side. They tend to be solitary. Typically, they
limited information regarding the nature of cardiac masses, appear as pedunculated or polypoid masses commonly
but contrast CT imaging with cardiac gating can better assess attached to the interatrial septum adjacent to the edge of
the precise location, size, and extent of cardiac masses. With the fossa ovalis. Depending on characteristic features of
A B
Figure 1-41 A, B: Turbo spin-echo MR images demonstrate a leftward-shifted heart with little evidence of pericardium within the epicar-
dial fat surrounding the heart, suggestive of congenital absence of the pericardium.
TABLE 1-6
CHARACTERISTIC FEATURES OF CARDIAC MASSES ON MAGNETIC RESONANCE IMAGING
AND COMPUTED TOMOGRAPHY
MR Imaging Characteristicsa
Contrast Distribution
Mass T1 T2 Enhancement CT Characteristics and Features
Thrombus Variable, depending Hypointense None Hypodense Atrial appendage,

on age aneurysms
Myxoma Hypointense or Hyperintense Mild to moderate; Heterogeneous Attachment to
isointense heterogeneous hypodense interatrial
septum;
commonly in left
atrium
Papillary Heterogeneous hypo- Hypointense Delayed Hypodense Solitary, mobile,
fibroelastoma to isointense hyperenhancement pedunculated
Lipoma Hyperintense Moderate to None or minimal Hypodense Encapsulated mass
hyperintense with smooth
contour
Fibroma Isointense Hypointense Early rim Hypodense with Large, solitary,
enhancement; areas of calcification intramyocardial
delayed
heterogeneous
hyperenhancement
Rhabdomyoma Isointense Hyperintense Variable Hypodense Multicentric masses,
intramural or
intracavitary
Teratoma Heterogeneous Heterogeneous Heterogeneous Heterogeneous, Multicentric
often with
calcification in
the form of teeth
Hemangioma Isointense Hyperintense Intense Heterogeneous with Commonly in
heterogeneous interdispersed epicardial layer
enhancement calcification of myocardium
(usually peripheral
to central)
Angiosarcoma Irregular nodular Irregular nodular Heterogeneous Hypodense nodular Commonly affect
areas of areas of lesions right atrium;
hyperintensity hyperintensity frequent
extension
Rhabdomyosarcoma Variable (isointense) Variable (isointense) Heterogeneous Hypodense Pericardial
infiltration rare
Lymphoma Hypointense Hyperintense Heterogeneous Hypodense or Multiple
isodense circumscribed
polypoid masses
or ill-defined
infiltrative lesion
a T1 and T2 characteristics are compared with those of myocardium.
the myxoma, patients may present with symptoms of mass with a narrow base of attachment (somewhat
dyspnea or embolic events (189). Patients with myxomas pedunculated); attachment of the mass to the interatrial
of the right side of the heart may present with tricuspid septum virtually cements the diagnosis (192,193). CT often
disease, pulmonary embolism, or pulmonary hyperten- demonstrates a distinct, intracavitary sphere, typically
sion. Patients with myxomas of the left side of the heart with calcification and heterogeneous hypoattenuation con-
may present with mitral valve obstruction, features sug- sistent with its gelatinous nature (192,194,195). On MR,
gestive of endocarditis, or neurologic symptoms from myxomas are generally isointense to myocardium on
embolism (190,191). T1-weighted images and hyperintense on T2-weighted
Various features of myxomas seen on CT and MR can images, and they sometimes show areas of decreased signal
help confirm the diagnosis. Myxoma is commonly seen as a intensity due to calcification or hemosiderin (196–198).
Chapter 1: Heart 47
There is often heterogeneous enhancement with gadolin- However, on CT and MR they often appear as a small
ium contrast agents because of cellular matrix or inflamma- (2 cm in diameter), solitary, mobile, pedunculated hypo-
tion. Functional consequences of large myxomas can be dense (CT) (200) or hypointense (T2-weighted MR)
seen in cine MR or CT in which the tumor mass may homogeneous valvular or endocardial mass that flutters or
prolapse into the ventricles or cause obstruction of the AV prolapses with cardiac motion (Fig. 1-43) (195).
valves (Fig. 1-42). Myxomas may be confused with thrombi,
in large part because of similar intracardiac behavior, and
Lipoma
clinical cardiac history coupled with lack of contrast en-
hancement argues for the presence of thrombus, especially Lipomas typically occur in adults, although they can also
given the low overall incidence of primary cardiac tumors. be seen in children. They generally occur as solitary masses
that can arise from the epicardial surface, spreading into
the pericardial space (201), or from the interatrial septum
Papillary Fibroelastoma
or endocardial surface as a broad base from which they
Papillary fibroelastomas are avascular papillomas lined can grow into any of the cardiac chambers (202). Lipomas
with endothelium. Although rare, they represent the sec- are seen as encapsulated masses demonstrating the signal
ond most common primary benign cardiac neoplasm of fat on CT and MR. There is often lack of contrast
(195). They typically occur in the left heart chambers, at- enhancement, but use of contrast can increase their con-
tached to cardiac valves in more than 90% of cases (199). spicuity. They often appear hypodense on CT (203) and
Most papillary fibroelastomas are found incidentally, but with homogeneous hyperintensity on T1-weighted MR,
affected patients may present with embolic or obstructive with only moderate signal intensity on T2-weighted MR
disease (195). These lesions are commonly diagnosed by (199,204). MR fat suppression sequences can reveal
echocardiography, although CT and MR can be useful decreased signal of lipomas and confirm the diagnosis
when the origin is not typical. Depending on the size, loca- (205). In addition, a smooth contour and capsule can
tion, and mobility, these masses may not be well seen on further distinguish benign lipomas from the irregular,
CT and MR, especially if cardiac gating is not employed. multilobar liposarcomas (195).
A,B
Figure 1-42 CT (A) and MR (B)

images from a patient with a large
myxoma (arrows) attached to the
interatrial septum and T1-weighted
(C) and T2- weighted (D) turbo
spin-echo MR images from a second
patient with a small left atrial
myxoma (arrows). RA, right atrium;
C, D LA, left atrium.
A,B
Figure 1-43 CT images (A, B)

demonstrate a small papillary fibro-
elastoma attached to the noncoro-
nary cusp of the aortic valve (black
arrows). MR images (C, D) demon-
strate a fibroelastoma attached to
C, D the pulmonary valve (white arrows).
Although not a tumor, lipomatous hypertrophy of the and MR imaging with signal intensity similar to subcuta-
interatrial septum can be confused with the truly neoplas- neous fat on T1- and T2-weighted images (Fig. 1-44)
tic lipoma. Lipomatous hypertrophy results from an incr- (205). These lesions tend to be benign but may be associ-
ease in cell number, or hyperplasia, seen in obese, elderly ated with atrial and ventricular arrhythmias due to their
individuals as an unencapsulated fatty infiltration, defined interference with the conduction system (193).
formally as any deposit of fat in the atrial septum at the
level of the fossa ovalis that exceeds 2 cm in transverse
Fibroma
diameter (199). The normal dimension of interatrial fat
anterior or posterior to the fossa ovalis measures less than Fibromas mainly affect young children, in whom they
1 cm (206). In addition to characteristic sparing of the rank as the second most common tumor seen in infancy
fossa ovalis seen on echocardiography, this tissue may (207). They are commonly associated with arrhythmias
appear wedge-shaped or as diffuse septal thickening on CT that may be life threatening (208). However, in about a
A B C
Figure 1-44 MR turbo spin-echo images without (A) and with (B) fat saturation demonstrate a large lipoma (*) involving the right atrium.
CT image (C) demonstrates asymmetric lipomatous hypertrophy with sparing of the fossa ovalis of the interatrial septum (arrow).
Chapter 1: Heart 49
third of cases, patients present asymptomatically into be seen in both RV and LV myocardium and interventricular
adulthood (207,209). Treatment options, especially for the septum. They may be intracavitary or intramural and are
asymptomatic patient, are debated but generally surgical often multiple rather than single. On CT, rhabdomyomas
resection, if possible, is recommended owing to the high appear hypodense to myocardium after contrast administra-
incidence of sudden cardiac death (210). tion (Fig. 1-46). On MR, rhabdomyomas appear isointense
On CT and MR, a fibroma commonly appears as a large to myocardium on T1-weighted sequences and hyperin-
solitary, calcified mass within the ventricular myocardium tense on T2-weighted sequences (215).
(Fig. 1-45). This is in contrast to other intramyocardial
tumors, including rhabdomyomas, which are often multi- Teratoma
centric, and rhabdomyosarcomas, which are frequently
Cardiac teratomas generally occur in infants and children
cystic or necrotic (211). On CT, fibromas are commonly
as a predominantly right-sided pericardial mass (216).
low attenuation because of the dense, fibrous tissue, with
Although considered benign, these tumors are often
bright areas of calcification. On MR, fibromas appear
accompanied by pericardial effusions, which may progress
isointense to muscle on T1-weighted images but dark on
to cardiac tamponade with subsequent respiratory distress
T2-weighted sequences, with variable enhancement pat-
and cyanosis (216,217). However, surgical resection is
terns seen after contrast administration (212,213).
often curative (218). CT and MR imaging of teratomas
demonstrate a heterogenous, multicystic appearance with
Rhabdomyoma calcification, often in the form of teeth, and can confirm
the diagnosis of teratoma over other tumors (193,211).
Cardiac rhabdomyomas account for the majority of cardiac
tumors seen in infants (193,211). These hamartomas are
Hemangioma
frequently associated with tuberous sclerosis of the brain,
sebaceous adenomas, and various hamartomatous lesions Cardiac hemangiomas are rare, benign vascular tumors of
of the kidney and other organs (209). Unlike other cardiac the heart. They usually occur in the epicardial layer but can
tumors, rhabdomyomas frequently regress over time and involve all the cardiac chambers. Symptoms are usually
hence are treated conservatively. However, complications the result of tumor compression of surrounding structures
can result from an obstructive syndrome or from severe or embolization (219). Because these tumors have unpre-
arrhythmias (214). On imaging studies, rhabdomyomas can dictable outcomes and may resolve, halt progression, or
Figure 1-45 MR turbo spin-echo

images with T1- (A) and T2-weight-
ing (B) demonstrate thickening of
the myocardium in the mid anterior
wall with a small hypointense area.
On postcontrast delayed enhance-
ment images (C), there is a large
intramyocardial mass seen in this
region, again with a small hypo-
intense area within the mass. On CT
(D), the bright central region corre-
sponds to significant calcification
within the mass, consistent with a
large fibroma.
Figure 1-46 MR images from a

patient with tuberous sclerosis and
suspected rhabdomyoma demon-
strate a large intramyocardial left ven-
tricular (LV) mass that is isointense to
myocardium on T1-weighted images
(A), hyperintense on T2-weighted
images (B), and patchy enhancement
on early (C) and late (D) T1-weighted
postcontrast images.
proliferate, surgical excision is usually recommended (219, and leukemia (183,223,224). Involvement of cardiac
220). On noncontrast CT, cardiac hemangiomas have a structures takes place primarily through the lymphatic
heterogeneous density with occasional interspersed calci- pathway, although hematogenous, contiguous, or trans-
fication (195). On MR, they appear isointense to myocar- venous extension does occur (225). Cardiac involvement is
dium on T1-weighted images and hyperintense on commonly seen as a late manifestation of primary tumor
T2-weighted images (195). In addition, they enhance extension (226), but characteristic vascular involvement
intensely with contrast, which may be inhomogeneous signs may help facilitate diagnosis. Pulmonary vein exten-
because of interspersed calcification and fibrous septa sion may signal bronchogenic carcinoma, inferior vena
within the mass (Fig. 1-47) (221). caval extension can be seen with renal cell and hepatocel-
lular carcinoma, and superior vena caval extension may in-
dicate supracardiac tumors such as thymic carcinomas
Malignant Cardiac Neoplasms (184). CT and MR imaging of metastases can demonstrate
Malignant cardiac masses include metastatic neoplasms, nodular masses or pericardial thickening, often with con-
primary cardiac sarcoma, and primary cardiac lymphoma. trast enhancement, due to vascularity of the lesions (186,
Although no single finding is specific, malignant cardiac 227). Other common features of malignancy include
tumors commonly affect the right-sided chambers of the right-sided involvement, ventricular infiltration, and hem-
heart, demonstrate inhomogeneity of tumor tissue, appear orrhagic pericardial effusion (198).
infiltrative, and are associated with pericardial or pleural
effusion (222). Additionally, MR signal intensity features
commonly associated with malignant lesions, although Sarcoma
not specific, include hyperintensity on T2-weighted im- Cardiac sarcomas are extremely rare tumors, occurring in
ages, hypointensity on gradient-echo images, and evidence less than 0.2% of the population, but represent the most
of contrast enhancement (222). common primary malignant cardiac tumor seen in
adults (207). These tumors may originate in the epicar-
dium or pericardium, involve the myocardium and cardiac
Metastatic Disease
valves, and cause nonspecific signs and symptoms that
Most metastases to the heart occur from lung or breast make clinical diagnosis difficult. There are several subtypes
cancer but can also be seen with melanoma, lymphoma, of sarcomas, including angiosarcoma, osteosarcoma, fibro-
Chapter 1: Heart 51
Figure 1-47 MR (A–C) and CT (D) images

from a patient with cardiac hemangioma that
appears mildly hyperintense to myocardium
on T1-weighted sequences (A), hyperintense
on T2-weighted sequences (B), and with vivid
enhancement after administration of contrast
on delayed enhancement inversion recovery
sequences (C). On early phase CT images (D),
the hemangioma is slightly heterogeneous in
density.
sarcoma, malignant fibrous histiocytoma, leiomyosarcoma, Rhabdomyosarcomas are striated-muscle tumors that
myxosarcoma, synovial sarcoma, neurofibrosarcoma, lym- represent the most common cardiac malignancy in infants
phosarcoma, reticulum cell sarcoma, and undifferentiated and children (235). These tumors may arise anywhere
sarcoma (Fig. 1-48). within the heart, involving any cardiac chamber or valve
Angiosarcoma is a tumor of endothelial cells and repre- (230), although pericardial infiltration is rare (193). On
sents the most prevalent subtype. Patients commonly pre- CT, these tumors demonstrate low attenuation with a
sent with right-sided heart failure or tamponade (207, smooth or irregular contour that enhances with contrast
228). Angiosarcomas most commonly affect the right (236). Signal intensity characteristics on MR imaging are
atrium, are highly vascular lesions with hemorrhagic, variable, but these tumors often appear isointense to my-
necrotic foci, and frequently invade the pericardium with ocardium (237) although heterogeneous signal intensity
associated hemorrhagic effusion that can lead to cardiac and contrast enhancement have also been reported (238).
tamponade (228). On CT, angiosarcomas appear as hypo- Extracardiac extension into the pulmonary arteries, aorta,
dense irregular or nodular lesions, often arising from the or valvular structures can be demonstrated with either CT
right atrial free wall with heterogeneous enhancement, or MR (230).
or, in the case of pericardial infiltration, may appear Fibrosarcomas consist primarily of malignant fibro-
as pericardial effusion or thickening (229). On T1- and blasts and comprise 5% of all primary cardiac tumors
T2-weighted MR sequences, angiosarcomas demonstrate (239,240). They commonly affect the left atrium, and
irregular nodular areas of increased signal intensity that valvular involvement is found in as many as 50% of the
may be focal or peripheral within areas of intermediate lesions. On CT, fibrosarcomas appear as a low-attenuation,
signal intensity to create the classically described “cauli- often obliterative mass. On T1-weighted MR sequences,
flower” appearance (230,231). In cases with diffuse peri- fibrosarcomas may be heterogeneous (241) or isointense
cardial infiltration, linear contrast enhancement along to myocardium (242). Fibrosarcomas may infiltrate the
vascular lakes gives rise to the “sunray” appearance (232). pericardium by direct invasion (241) or tumor deposition
Areas of central necrosis can also be seen in communica- nodules (243) or may primarily involve the pericardium,
tion with the cardiac chambers (233,234). appearing similar to malignant mesothelioma (207).
A,B

D) images demonstrate a large mass
(*) extending from the superior
vena cava into the right atrium,
which demonstrates heterogeneous
density on CT and hyperintensity on
T2-weighted spin-echo (C) and inver-
sion recovery (D) sequences. The
patient underwent resection of the
mass, which was found to be a
C, D synovial cell sarcoma.
Osteosarcomas in the heart consist of malignant bone- Liposarcoma is a malignant mesenchymal tumor that
producing tumor cells (207). Primary osteosarcomas consists of lipoblasts. These tumors are extremely rare
commonly arise in the left atrium and are usually accom- (207). They tend to arise in the atria but have been
panied by signs and symptoms of congestive heart failure reported to occur in any cardiac chamber, the pericardium,
(244). CT may show dense calcifications within a low- and cardiac valves (253,254). Unlike benign lipomas,
attenuation left-sided mass (245). However, early lesions liposarcomas have little or no macroscopic fat (195). CT
may have minimal calcification and can be mistaken for and MR imaging demonstrates a large, multilobular,
dystrophic calcifications. MR imaging demonstrates a heterogeneous mass with areas of necrosis and hemor-
large, irregular mass with heterogeneous signal intensity rhage (230).
on both T1- and T2-weighted sequences (246). Other dis-
tinguishing features include a broad base of attachment,
Lymphoma
an aggressive growth pattern with extension into the
pulmonary veins or atrial septum, or infiltrative growth Primary cardiac lymphoma includes lymphoma that is
along the epicardium (247–249). mostly confined to the heart or pericardium (207), as
Leiomyosarcoma is a malignant tumor with smooth opposed to the more common cardiac spread of non-
muscle differentiation. It may arise from the subendo- Hodgkin lymphoma (255). These are commonly aggressive
cardium of the cardiac chambers, but it more commonly B-cell lymphomas (193,256). Although prevalence is in-
arises from smooth muscle of the pulmonary veins and creased in immunocompromised patients, these lym-
arteries and then spreads into the heart (207). CT imaging phomas have also been diagnosed in immunocompetent
demonstrates lobulated, irregular, low-attenuation masses patients (256). Cardiac lymphoma may appear as either
(250), and MR signal intensity characteristics are nonspe- multiple circumscribed polypoid masses (257) or an ill-de-
cific but have been reported as intermediate on T1-weighted fined infiltrative lesion (258,259). They are less likely than
images and increased on T2-weighted images (251,252). sarcomas to have necrosis (207). They commonly arise in
Chapter 1: Heart 53
Figure 1-49 MR images from a

patient with a large right atrial mass
that demonstrates isointensity to
myocardium on T1-weighted images
(A), hyperintensity on T2-weighted
spin-echo images (B) and inversion
recovery sequences (C), and mild
uptake of contrast with some het-
erogeneity (D). These features were
suggestive of lymphoma, which was
proven on biopsy, and the patient
underwent chemotherapy and radia-
tion treatment, with resolution of
the mass.
the right side of the heart but may involve any chamber. Pseudotumors
Pericardial effusions are common and may be the only
finding at the time of imaging. On CT, lymphomas appear Because of the complex nature of the cardiac anatomy and
hypodense or isodense relative to myocardium and the unique individual variations seen, normal cardiac
demonstrate heterogeneous contrast enhancement (256, structures can often be mistaken for cardiac tumors. In
259). On MR, lymphomas often appear hypointense to particular, fetal remnants (e.g., eustachian valve, Chiari
myocardium on T1-weighted sequences and hyperintense network) or normal variants (e.g., prominent trabecula-
on T2-weighted sequences (258); however, isointensity on tions, false tendons) may give the impression of a cardiac
both T1- and T2-weighted images has also been reported tumor (Fig. 1-50). Given the 3D nature of cross-sectional
(Fig. 1-49) (257,259). imaging with CT and MR, normal cardiac structures can be
A B C
Figure 1-50 A–C, Some pseudomasses that may be confused with tumors include the crista terminalis (arrowheads) seen in the right
atrium, indentation of the atrioventricular groove epicardial fat (arrows), and the eustachian valve (black arrow).
differentiated from cardiac tumors. In addition, the dis- do not enhance with contrast administration. Although
tinction between cardiac neoplasms versus other cardiac distinction from cardiac tumor is possible, tissue differen-
masses can be made. tiation between thrombi and vegetations is generally not
possible with CT and MR, and accurate diagnosis relies on
incorporating the patient’s accompanying clinical features
Thrombi
and the effect on cardiac structures (e.g., location and/or
Intracavitary cardiac thrombi probably represent the most evidence of destruction).
common masses seen in cardiac imaging. In patients with
atrial arrhythmias or other predisposing factor for atrial
blood flow stasis (e.g., mitral stenosis), they can frequently VALVULAR HEART DISEASE
be seen along the posterolateral wall of the left atrium or
within the left atrial appendage (259a). On CT, thrombi Valvular heart disease is an important part of cardiology.
generally exhibit homogeneous hypodense attenuation Often detected by physical examination, imaging studies
after contrast administration (197). On MR, signal inten- can help clarify the affected valve, define the valvular
sity varies with thrombus age (Fig. 1-51). Acute thrombi anatomy, and assess the degree of valve dysfunction and
appear hyperintense to myocardium on T1-weighted the effect on the cardiac chambers and function. The car-
sequences and hypointense on T2-weighted sequences. diac valves are normally very thin, pliable, mobile struc-
Older thrombi exhibit signal intensity dependent on tures. Hence, imaging techniques to assess valve disease
the amount of deoxyhemoglobin and methemoglobin need to have high spatial and temporal resolution.
remaining within the thrombus. Additionally, the admin- Traditionally, transthoracic echocardiography, with its
istration of contrast can usually be diagnostic because acceptable spatial resolution and high temporal resolution
thrombi do not generally enhance as other cardiac tumors with near real-time imaging, is the imaging modality of
do (197). choice in the assessment of valve disease. Quantification of
valvular stenosis and valve area, valvular regurgitation, and
effective orifice area and an assessment of ventricular func-
Vegetations
tion can be easily performed. Transesophageal echocardio-
Intracardiac masses related to infective endocarditis are graphy can further define valvular anatomy and dysfunc-
usually diagnosed with echocardiography in the appropri- tion in patients with infective endocarditis or in patients
ate clinical setting. However, in cases in which the diagno- with poor acoustic windows. MR and CT imaging have the
sis in uncertain, MR and CT imaging can aid in excluding a potential to provide information on cardiac chamber size,
cardiac tumor. In general, vegetations exhibit MR and CT myocardial mass, pulmonary blood flow, pulmonary
imaging characteristics similar to those seen with thrombi. venous pressures, and calcifications in these patients.
Because vegetations are often nonvascular, they generally Although CT can provide high spatial resolution of valvular
A B
Figure 1-51 T1-weighted MR images before (A) and after (B) administration of gadolinium contrast demonstrate left ventricular
thrombus (arrows).
Chapter 1: Heart 55
anatomy, moderate temporal resolution limits assessment hypertrophy. The most common causes of valvular aortic
of valvular flow and function. MR can directly demonstrate stenosis include congenital (e.g., bicuspid), calcific de-
the intracardiac jets of valvular dysfunction as well as quan- generative, and rheumatic diseases. Subvalvular and su-
tify degree of valvular dysfunction and its sequela. pravalvular lesions are often congenital in nature, but
subvalvular obstruction may also be due to sequelae
related to systolic anterior motion of the mitral leaflets
Aortic Valve
that may occur in hypertrophic cardiomyopathy with
Anatomic and functional assessment of the aortic valve is LVOT obstruction. CT and MR are rarely indicated in the
important for the diagnosis and management of patients evaluation of these patients. However, concomitant aortic
with aortic valve pathology. Echocardiography is the pri- pathology may lead to CT and MR evaluation. Valve mor-
mary method of choice in the assessment of aortic valve phology, leaflet function, and concomitant aortic or
disease, and transesophageal echocardiography or ven- myocardial pathology can be evaluated. Etiology of aortic
triculography can be used for further evaluation if findings stenosis, including the level of obstruction, is often diag-
by the transthoracic approach appear inconclusive (260). nosed on echocardiography, but information from CT or
MR and cine CT can both be used for monitoring adapta- MR can be helpful in inconclusive cases. Severity of the
tional changes in the cardiac chambers and assessment of stenosis can be assessed by planimetry of the orifice area
ventricular function. In addition, measurement of blood (Fig. 1-52) (263,264). Additionally, flow quantification of
flow on MR allows for quantitative evaluation of stenosis stenotic jets, which appear as signal void with variable
and regurgitation. Information on valve morphology, extension into the ascending aorta, can be correlated with
leaflet abnormalities, and valve motion is possible with the severity of the valvular gradient.
MR and CT but remains inferior to that obtained with
echocardiographic techniques (261,262).
Aortic Regurgitation
Aortic regurgitation may be due to disease of the valve
Aortic Stenosis
itself or diseases of the aorta that secondarily affect the
Aortic stenosis can be described as valvular, subvalvular, or valve. Common causes of valvular aortic regurgitation
supravalvular, depending on the exact site of obstruction. include rheumatic heart disease, infective endocarditis,
Regardless of the level of obstruction, these lesions all congenital bicuspid aortic valve, and Marfan syndrome.
increase LV myocardial strain with resultant myocardial Aortic diseases associated with aortic regurgitation include
A,B
Figure 1-52 CT (A, B) and MR

(C, D) images from two different
patients with aortic stenosis. The
short axis CT images demonstrate
calcified aortic valve leaflets and lim-
ited opening of the valve in systole.
The MR short axis image (C) also
demonstrates limited opening of the
leaflets and flow turbulence seen in
C, D the ascending aorta (D, arrow).
A B C
Figure 1-53 CT (A, B) and MR (C) images from a patient with aortic regurgitation. There is incomplete coaptation of the leaflets during
diastole (B) and resultant regurgitant flow within the left ventricular outflow tract (arrow) demonstrated on the MR image.
trauma, aortic dissection, inflammatory (e.g., giant cell mitral valve disease has become increasingly important
aortitis) and connective tissue (e.g., Ehlers-Danlos syn- (261). Echocardiography is the primary diagnostic tool used
drome) diseases, and idiopathic dilation of the aortic for assessment of anatomy and pathology of the mitral
annulus. Although acute aortic regurgitation may occur in valve as well as its functional consequences. However, MR
the setting of infective endocarditis, acute aortic dissection, and CT have the potential to provide additional informa-
and thoracic trauma, most commonly aortic regurgitation tion, especially when echocardiography is limited or
results from a slow process of LV dilation with a prolonged impractical (267–269).
asymptomatic phase. In the setting of acute aortic regurgi-
tation, there has not been enough time for ventricular
Mitral Stenosis
adaptation, resulting in decreased forward CO, elevated
left atrial pressure, pulmonary edema, and shock. MR and Mitral stenosis in the adult population is commonly due
CT examination in the acute setting can demonstrate the to rheumatic heart disease. Other causes of mitral stenosis
underlying etiology for acute LV overload. More com- include calcification, carcinoid syndrome, radiation valvu-
monly, in the chronic setting, concentric and eccentric ven- lar disease, and congenital mitral stenosis. Rheumatic
tricular myocardial hypertrophy with ventricular dilation mitral stenosis is a chronic, slowly progressive fibrotic
compensates for the increased wall stress induced by the process instigated by the initial rheumatic inflammatory
increased volume load. Measurement of LV size and func- reaction. The process may take up to 20 to 40 years before
tion can be used to follow up patients with chronic aortic symptoms develop and then another 10 years for the
regurgitation. Valvular characteristics, such as calcifica- symptoms to become disabling. As the mitral orifice nar-
tions, congenital lesions, and poor coaptation, can be rows, left atrial hypertension develops from the increased
demonstrated on CT and MR, and the effective regurgitant diastolic pressure, which is transmitted to the pulmonary
orifice can be directly measured by planimetry. In addi- bed. The left atrium dilates and hypertrophies in response
tion, MR can often demonstrate the jet of aortic regurgita- to the pressure load. Continued pressure overload leads to
tion typically seen as an early diastolic signal void interstitial and eventually alveolar pulmonary edema.
(Fig. 1-53) (265). Direct and indirect findings of mitral stenosis may be
seen on CT and MR. The initial sign of mitral stenosis is an
increase in left atrial and left atrial appendage area and
Mitral Valve
volume. Etiology may be discerned by the structural alter-
The mitral valve apparatus consists of two leaflets and com- ations associated with mitral stenosis. Rheumatic mitral
missures, the mitral annulus, the chordae tendineae, and stenosis often results in thickened, calcified mitral leaflets
the papillary muscles. Normal mitral valve function is a with eventual fusion that typically begins at the leaflet tips
complex process requiring the interaction of all of these and then progresses to involve the leaflet body. The chor-
components along with adequate left atrial and LV function dae tendineae are often involved and become thickened,
(266). Abnormalities affecting any portion of the mitral fused, and nonpliable (270). Radiation valvular disease is
valve apparatus can affect mitral valve function, and the marked by significant calcifications of the leaflet body.
pattern of pathologic involvement often determines the fea- Elevation of pressures in the pulmonary venous bed may
sibility of mitral valve repair (267). Over the past decade, be transmitted across the capillary bed and result in pul-
major advances in mitral valve repair and replacement pro- monary arterial hypertension. This may be identified as an
cedures, both surgical and percutaneous, have improved increase in caliber of the central pulmonary artery seg-
patient management and outcomes. Hence, knowledge of ments, and increased resistance may be reflected as a slow-
the exact morphologic and functional characterization of ing of the pulmonary blood flow, resulting in increased
Chapter 1: Heart 57
signal within the pulmonary arteries on spin-echo images. LV dilatation. The etiology of mitral regurgitation can also
Chronic elevation in pulmonary vascular resistance may be discerned from examination of the mitral valve appara-
be reflected in RV hypertension and RV hypertrophy visu- tus, including the mitral leaflets and annulus, leaflet
alized as a thickening of the RV free wall and interventricu- motion, and regurgitation jet origin and direction. Both CT
lar septum; further changes result in alteration of the and MR can demonstrate structural alterations associated
curvature of the interventricular septum, first reflected as with mitral regurgitation, including myxomatous, infective,
straightening and subsequently as reversal of the systolic or congenital (e.g., cleft) degeneration; annular dilation; or
bowing of the septum toward the right ventricle (265). papillary muscle rupture. In patients with ischemic mitral
regurgitation, there is relatively normal appearing mitral
valve apparatus in the setting of regional wall motion
Mitral Regurgitation abnormalities and coronary atherosclerosis. Functional
Mitral regurgitation can present acutely or chronically. alterations such as prolapse and flail can sometimes be seen
Acute mitral regurgitation may result from sudden changes with CT (Fig. 1-54), although this may be better demon-
in the chordae tendineae anchoring the valvular leaflets strated with MR because of its higher temporal resolution.
(e.g., chordal rupture), papillary muscle dysfunction or MR examination reveals signal void jet of mitral regurgita-
rupture of the head of the papillary muscle (e.g., myocar- tion as an early systolic fan-shaped signal void extending
dial infarction), or acute damage to the leaflets themselves from the mitral annulus into the left atrium (Fig. 1-54).
(e.g., infection). Chronic mitral regurgitation is often due Flow alterations, including regurgitant jet origin, size, and
to mitral valve prolapse syndrome, acute or chronic rheu- direction, can often be demonstrated on cine MR sequences.
matic heart disease, collagen vascular disease, subacute Quantification of mitral regurgitant orifice area can be
infective endocarditis, or ischemic heart disease, or it may directly planimetered on cine CT studies (269). On MR,
be congenital or drug related (271). It can also be seen in quantification of mitral regurgitation volume can be calcu-
the setting of hypertrophic cardiomyopathy. In the acute lated by measurement of total SV of the left ventricle minus
setting, sudden volume overload on the unprepared left SV in the ascending aorta, as measured by phase-contrast
ventricle may result in low CO and pulmonary congestion velocity mapping or by through-plane phase-contrast veloc-
and, if left untreated, may result in death. In the chronic ity mapping performed at the level of the mitral inflow.
setting, compensatory LV hypertrophy and an increase
in LVEDV lead to increased total SV and restoration of
Pulmonary Valve
forward CO (272).
CT and MR findings during the acute phase of mitral The pulmonary valve is a trileaflet valve that separates the
regurgitation are those of severe pulmonary congestion with right ventricle from the pulmonary vasculature. Dysfunc-
nearly normal heart size. Diffuse bilateral infiltrates of alve- tion of the valve can have adverse effects on the right ven-
olar and interstitial edema predominate. In chronic mitral tricle. Echocardiography has well-known limitations in its
regurgitation, dominant findings are those of left atrial and ability to accurately provide hemodynamic and anatomic
A B
Figure 1-54 Mid systolic phase CT (A) and gradient-echo MR (B) images demonstrating mitral valve prolapse (arrow) with associated
regurgitation (arrowhead) seen on MR images as a flow turbulence.
information about the pulmonary valve, primarily because tion are rare, such as in absent, malformed, or fenestrated
of the position of the pulmonary valve in the chest wall. valve leaflets. However, acquired causes are much more
This may be exacerbated further by distortion of the common. These include pulmonary artery hypertension,
anatomy from prior surgery. Especially in patients in endocarditis, and following repair of pulmonary stenosis
whom anatomic findings are unclear, CT and MR imaging or tetralogy of Fallot. Carcinoid and rheumatic heart dis-
can provide helpful additional anatomic information ease may also be associated with pulmonary regurgitation
because of their high spatial resolution and ability to but more commonly are associated with pulmonary steno-
obtain anatomic images in every plane of choice. In addi- sis. Marfan syndrome may also cause pulmonary regurgita-
tion, the effect on ventricular volume and function can be tion secondary to dilatation of the pulmonary artery.
quantitatively assessed. MR imaging also allows for quan- Doppler echocardiography commonly can demonstrate a
tification of flow across the pulmonary valve. significant jet of pulmonary regurgitation and can be used
to assess for the presence of pulmonary artery hyperten-
sion. MR imaging has become the gold standard for the
Pulmonary Stenosis
periodic evaluation and follow-up of patients with pul-
Pulmonary valve stenosis is commonly congenital in ori- monary regurgitation. In addition to visualization of the
gin and represents approximately 10% of all patients regurgitant flow jet (characteristically seen as a signal void
with congenital heart disease. Other etiologies of pul- in the otherwise bright signal intensity of flowing blood)
monary valve stenosis include rheumatic heart disease, on cine MR images (Fig. 1-55), phase-contrast velocity
carcinoid syndrome, and RV outflow obstruction (pseudo- mapping can accurately quantify systolic and diastolic
pulmonary stenosis) from cardiac tumors or aneurysm of flow through the pulmonary valve for calculation of pul-
the sinus of Valsalva. Echocardiography may be able to monary regurgitant fraction (275,276). In addition, MR
demonstrate the site and anatomic cause of the obstruc- imaging quantitative assessment of RV size and function
tion, and Doppler techniques can be used to grade the does not rely on geometric assumptions and is more
severity of obstruction (273). CT and MR imaging can pro- reproducible compared with echocardiographic tech-
vide high spatial resolution for elucidation of anatomic niques. Furthermore, the wide field of view and unre-
causes of obstruction (e.g., valvular, subvalvular, or stricted imaging planes allow for full assessment of RVOT
supravalvular). Phase-contrast velocity mapping MR imag- aneurysmal or akinetic regions (Fig. 1-55) as well as extra-
ing sequences can provide flow information and estima- cardiac lesions, such as conduits or distal pulmonary artery
tion of the stenotic valve gradient (274). Cine MR imaging stenosis (276,277). CT can also provide quantitative and
sequences can be useful in demonstrating anatomic abnor- anatomic information similar to that obtained with MR
malities associated with flow turbulence. imaging but has not been as fully studied.
Pulmonary Regurgitation Tricuspid Valve

Physiologic (mild) pulmonary regurgitation is commonly The tricuspid valve consists of three leaflets (anterior, pos-
detected by color Doppler echocardiography in otherwise terior, and septal) and a fibrous annulus and is commonly
normal hearts. Congenital causes of pulmonary regurgita- more apically displaced than the corresponding mitral
A B
Figure 1-55 Gradient-echo MR cine images during systole (A) and diastole (B) demonstrate akinetic right ventricular outflow patch repair
(arrows) and turbulent flow forward (arrowhead in A) and backward (arrowhead in B) through the pulmonary valve.
Chapter 1: Heart 59
valve. Unlike the mitral valve, the tricuspid valve leaflets used in the implants, which may limit accurate evaluation
receive chordae from anterior, posterior, and septal papil- for prosthetic valve dysfunction. However, the high spatial
lary muscles. resolution and functional imaging of CT and MR may
provide additional information to that obtained with
echocardiography in patients with prosthetic heart valves
Tricuspid Stenosis
and annuloplasty rings (Fig. 1-56).
Tricuspid stenosis is a very rare condition, most commonly
due to rheumatic heart disease (>90% of cases) or carci-
noid syndrome. Echocardiography can usually diagnose CONGENITAL HEART DISEASE
and identify the underlying etiology as well as estimate the
severity of disease. CT and MR imaging can provide correl- The number of adults with congenital heart disease has
ative or supplemental information but are not commonly been growing over the past 5 decades because of the
used in the evaluation of patients with tricuspid stenosis. advances in cardiac surgery, intensive care, and noninva-
Common abnormalities seen include thickening, retrac- sive diagnosis (282). Approximately 85% of infants with
tion, and immobility of the tricuspid valve leaflets and cardiovascular anomalies can be expected to survive
subvalvular apparatus, with diastolic doming a frequent into adulthood, and with further advances in surgical
finding, as well as turbulent flow patterns on cine MR techniques, this number may continue to grow (283).
imaging, indicative of stenosis (278,279). Although surgical correction of an anomaly may
re-establish a relatively normal pattern of blood flow,
these adult survivors still remain at significant risk for
Tricuspid Regurgitation
developing complications from their operative proce-
Tricuspid regurgitation may be a result of abnormalities of dures or from lingering effects of the original anomaly.
the tricuspid valve apparatus or secondary to RV volume or Hence, these patients need to be followed up closely
pressure overload. Primary causes of tricuspid regurgita- regardless of their stage of treatment. As such, com-
tion include rheumatic heart disease, Ebstein anomaly, prehensive imaging evaluation of the cardiovascular
carcinoid syndrome, myxomatous valve disease, trauma anatomy, flow, and function is important in the manage-
(e.g., catheter or pacemaker related), tumor, infective ment of these patients. Echocardiography and cardiac
endocarditis, or papillary muscle dysfunction (273). CT catheterization are the primary cardiac imaging modali-
and MR imaging can provide high-resolution anatomic ties in the assessment of patients with congenital heart
information when the etiology of valve dysfunction is in disease. However, echocardiography, although noninva-
question. Quantitative assessment of right-sided cardiac sive with no radiation, is limited by a small field of view,
chamber size and function can be measured with CT and acoustic windows, and operator dependence. Conven-
MR imaging for management and follow-up of these tional angiography is limited by the 2D view with over-
patients. Quantitative information on severity of regurgita- lapping of adjacent vascular structures, catheter-related
tion can also be made using phase-contrast velocity complications, and relatively high exposure rates to
mapping MR imaging sequences. ionizing radiation and iodinated contrast material.
With the advancement of techniques in both CT and
MR, cardiac imaging of congenital heart disease has
Prosthetic Valves
become an important adjunct to the evaluation of these
Prosthetic heart valves and annuloplasty rings are implan- patients. EBCT and MDCT, with their fast acquisition times
ted every year around the world and have become ex- and capacity to obtain volumetric data with high spatial
tremely useful devices in cardiac disease. Prosthetic valves resolution, have often been used in the morphologic eval-
can be divided into two main classes: mechanical prosthe- uation of congenital heart disease. MR imaging can also
ses and biologic or tissue valves. Mechanical valves are provide high-quality anatomic imaging, but also impor-
classified into three major groups: caged-ball, tilting-disc, tant information on cardiac function and flow. Both tech-
and bileaflet valves. The most widely employed mechani- niques can provide 3D information and evaluation in
cal valve currently used is the bileaflet valve, although virtually any plane needed for evaluation of cardiac and
older caged-ball valves are still seen in patients at follow- vascular disease.
up. Tissue valves include heterografts, homografts, and Advantages of CT are short examination, fewer require-
autografts. CT and MR imaging used to evaluate prosthetic ments for sedation, simultaneous evaluation of airways
valve dysfunction have not been extensively reviewed in and lung parenchyma, and high spatial resolutions.
the literature. For the most part, all prosthetic heart valves Disadvantages are radiation exposure, use of iodinated
and annuloplasty rings in current use are considered safe contrast, and lack of functional information (Table 1-7).
for imaging at 1.5-Tesla MR systems (280), and most have MR imaging protocols vary depending on the structure
been shown to be safe at 3.0 and 4.7-Tesla imaging (281). of interest and the information desired. An overview of
Artifacts vary depending on the amount and type of metal cardiovascular anatomy is evaluated using both spin-echo
A,B
C, D
Figure 1-56 CT images of a tilting bileaflet disc

(A, B) and tissue prosthetic valve (C, D) and MR
images of a tissue prosthetic (E, F) valve seen in the
open (left) and closed (right) positions. On gradient-
echo MR images, turbulent flow across the pros-
thetic valve can be seen during valve opening and
E, F closing (arrows).
and gradient-echo cine techniques. The vascular system the vascular system, image acquisition is usually started a
can be further evaluated using 3D MR angiography. An few centimeters above the aortic arch and continued to the
assessment of blood flow can be made using phase- diaphragm. Because congenital heart disease often involves
contrast velocity encoded mapping for quantification of abnormalities in both the right and left heart chambers, the
both stenotic and regurgitant jets, calculation of shunt timing of data acquisition is set such that there is optimal
fraction, and calculation of differential blood flow (e.g., contrast enhancement seen in both sides of the heart. In
pulmonary artery system). Cine MR sequences can be used cases with intracardiac shunting, imaging can be performed
for quantification of chamber size and function. early (within 5 seconds of contrast administration) and late
The CT imaging protocol is similar to the general CT (approximately 30 seconds after contrast administration)
cardiac protocol, with some adjustments depending on to determine the degree and direction of shunting (284).
the structure of interest. For patients with abnormalities in Images can be evaluated using multiplanar reconstructions,
Chapter 1: Heart 61
TABLE 1-7
MAGNETIC RESONANCE IMAGING VERSUS COMPUTED TOMOGRAPHY
FOR EVALUATION OF CONGENITAL HEART DISEASE
MRI CT
Examination time Long Short

Often requires sedation in children Yes No
Radiation exposure No Yes
Iodinated contrast material No Yes
Difficulty in patients with irregular heart rhythms Maybe Yes
Difficulty in patients with significant calcification No Yes
Imaging patients with pacemakers/internal defibrillators No Yes
Simultaneous evaluation of airways and lung parenchyma No Yes
Functional information obtained Yes Yes
Flow information obtained Yes No
Spatial resolution 1–2 mm 0.4–1 mm
Temporal resolution 50–75 ms 83–210 ms
maximum intensity projections, and 3D volume rendered abnormalities, including aortic dilation, aneurysms, and
image sets. Data can also be reconstructed in multiple dissection.
phases through the cardiac cycle for calculation of cardiac Although diagnosis of bicuspid aortic valve can often be
volumes and function. established by echocardiography, MR imaging can also
clearly demonstrate the anatomy and functional state of the
aortic valve. Axial and oblique–sagittal views targeting the
Bicuspid Aortic Valve
LVOT and aortic valve in long and short axis views can
Bicuspid aortic valve occurs in 1% to 2% of the general U.S. demonstrate the valve leaflet anatomy and motion during
population (285) and is the most common cardiac anom- the cardiac cycle. Clues suggestive of a bicuspid aortic valve
aly seen in adults (286). Although functionally competent include an eccentric coaptation point of the leaflets and pro-
in children, the bicuspid aortic valve is prone to accelerated lapse of the leaflets during diastole seen on long axis views
degeneration, leading to some degree of aortic stenosis or of the LVOT. Additionally, short axis views of the aortic valve
aortic insufficiency in more than 70% of patients. In addi- often show two unequal-sized leaflets, with the larger aber-
tion, owing to abnormal flow turbulence in the congeni- rant leaflet resulting from fusion of two of the aortic cusps
tally deformed valve, bicuspid aortic valve is associated and a characteristic football-shaped appearance of the aortic
with an increased incidence of infective endocarditis. valve opening during systole (Fig. 1-57).
Approximately 30% to 40% of patients require surgical The presence of bicuspid aortic valve is an independent
replacement of the valve during their lifetime (287). In risk factor for aortic dilation, aneurysm, and dissection.
general, bicuspid aortic valve is highly associated with con- Aortic complications were previously thought to be se-
genital abnormalities of the aorta, including coarctation quelae of valvular dysfunction, but more recent reports sug-
and patent ductus arteriosus, as well as acquired aortic gest that accelerated degeneration of the aortic media with
A,B C
Figure 1-57 Gradient-echo MR images of a bicuspid aortic valve in the open (A) and closed (B) positions demonstrate fusion of the right
and left cusps. CT image of a bicuspid aortic valve in a different patient demonstrates fusion of the right and noncoronary cusps (C).
resultant structural weakness of the aortic wall is the main of atrial septal defects include depiction of the defect on at
cause of aortic disease (288–290). least two adjacent anatomic levels on multisection trans-
verse images or at the same anatomic level on single-
section cine images (293). Ostium secundum defects occur
Atrial Septal Defect
in the middle of the interatrial septum in the region of the
Atrial septal defects represent at least 30% of all congenital thin fossa ovalis. Because the fossa ovalis is normally very
cardiac defects in individuals older than 40 years of age thin tissue, this region may exhibit loss of signal intensity
(291). Atrial septal defect may be described according to on both MR and CT images and can sometimes be mis-
the portion of the interatrial septum involved and may be taken for a defect in the septum secundum. However, the
divided into primum, secundum, sinus venosus, and coro- adjacent tissue around the fossa ovalis normally thins
nary sinus defects. Patent foramen ovale is often consid- gradually toward the site of signal intensity loss, whereas
ered a type of secundum atrial septal defect because of the defects in the septum secundum are usually accompanied
presence of shunting, but no defect is actually present; the by thickening of the adjacent septum (294). Ostium pri-
septa primum and secundum membranes are present and mum defects are seen in the lower portion of the atrial sep-
often overlap but have not completely sealed. Defects tum and are commonly associated with AV septal defects,
involving the ostium secundum represent more than three including abnormalities of the AV valves. Sinus venosus
quarters of adults with atrial septal defect, who are invari- defects occur in the portion of the atrial septum between
ably symptom free. However, by the age of 40 years, at the superior vena cava and the left atrium. Use of phase-
least half will develop symptoms of dyspnea, easy fatiga- contrast and cine MR imaging can demonstrate more accu-
bility, and chest pain caused by atrial arrhythmias, infec- rately the size and shape of atrial septal defects as well as
tive endocarditis, right-sided heart failure, or pulmonary quantitate the direction and degree of shunting (30,295).
hypertension related to chronic right-sided volume over- In addition to demonstrating the anatomic presence of an
load. Defects in the septum primum are frequently associ- atrial septal defect, ECG-gated CT imaging performed early
ated with defects in the AV valves. Associated mitral regur- and late after contrast administration can also further
gitation may be present in up to 15% of patients and is demonstrate the direction and degree of shunting (284).
commonly related to poor leaflet coaptation associated Furthermore, the anatomy and size of the right ventricle
with leaflet prolapse or LV cavity deformity (286). In addi- and pulmonary arteries can be evaluated with CT and MR,
tion, abnormal pulmonary venous drainage may also be which is important in patients with long-standing atrial
seen with atrial septal defect. septal defects.
Right atrial and RV enlargement is a common finding
in adults with atrial septal defect. The left atrium is usually
Ventricular Septal Defect
normal sized in younger individuals. However, in patients
with atrial septal defect who go on to develop secondary Ventricular septal defects account for approximately 20% of
pulmonary hypertension with Eisenmenger physiology, all congenital heart disease (296). However, they commonly
reversal of blood flow through the atrial septal defect will close spontaneously or surgically early in life and therefore
lead to left atrial enlargement due to the increased volume are not often seen in adults. Ventricular septal defects can be
of blood. With the onset of secondary pulmonary hyper- classified according to their location and margin and
tension, there is a reduction in peripheral pulmonary include outlet, membranous, trabecular, and inlet defects
blood flow as the central pulmonary arteries enlarge. (Fig. 1-59) (297). The diagnosis of ventricular septal defect
Additionally, the central pulmonary arteries may develop a is usually established by echocardiography. However, CT
symmetric pattern of calcification as a result of long-stand- and MR imaging can provide high sensitivity for detection
ing pulmonary hypertension. of ventricular septal defects as well as an accurate noninva-
Recently, percutaneous catheter–delivered devices have sive alternative imaging modality (284,298). As with other
become a popular choice for closure of atrial septal defect. shunt lesions, phase-contrast and cine MR imaging can pro-
These devices usually consist of two umbrellalike or clam- vide information on degree, direction, and quantification of
shell segments that clamp the atrial septum between shunting (30). ECG-gated CT imaging performed early and
them to cover the defect. They are most commonly used in late after contrast administration can also further demon-
patients with small- to moderate-sized ostium secundum strate direction and degree of shunting (30).
type of defects in which there is an adequate amount of
surrounding septal tissue to anchor the device in place.
Anomalous Pulmonary Venous Connection
Sinus venosus, septum secundum, and septum primum
defects can often be clearly visualized with horizontal long Anomalous pulmonary venous connection occurs when
axis and short axis imaging planes (Fig. 1-58). MR imaging one or more of the pulmonary veins drain into systemic
has been shown to have greater than 90% sensitivity and veins, the right atrium, or the coronary sinus. Total anom-
specificity for detection and localization of atrial septal alous pulmonary venous connection occurs when all four
defects (292). MR imaging criteria to define the presence pulmonary veins have anomalous drainage, resulting in
Chapter 1: Heart 63
A B
C D
Figure 1-58 CT (A, B) and MR (C, D) images from two different patients with a large secundum atrial septal defect (arrows).
A B
Figure 1-59 A, B: MR images from two different patients with membranous ventricular septal defects (arrows).
A B C
Figure 1-60 3D volume rendered MR angiography (A, B) and gradient-echo MR images (C) demonstrate partial anomalous pulmonary
venous return of the left upper lobe veins into the left brachiocephalic vein (arrows).
mixing of systemic and pulmonary venous blood in the of the main pulmonary artery near the orifice of the left
right atrium, and is estimated to occur in 1 in 15,000 live pulmonary artery (303) (Fig. 1-61). MR and CT angiogra-
births (299). Partial anomalous pulmonary venous con- phy of the great vessels can readily depict the anomalous
nection, in which some but not all of the pulmonary connection between the aorta and pulmonary artery
venous flow returns to the systemic venous system, occurs circuits. Additionally, MR phase-contrast velocity mapping
more frequently and is often incidentally noted on cross- at the level of the shunt can allow quantification of
sectional imaging studies including CT and MR. In the set- the amount of flow through the ductus. Because many
ting of an atrial septal defect, about 10% of patients will patients often undergo closure with embolization coils,
also have a pulmonary venous abnormality. Diagnosis is follow-up imaging with CT angiography is preferred
often suspected by clinical signs and echocardiography but because artifact related to the metal in the embolization
can be confirmed with CT and MR imaging. Unlike coils often obscures visualization on MR.
echocardiography, cross-sectional imaging with CT and
MR allows for comprehensive evaluation of the pulmonary
Coarctation of the Aorta
venous and systemic venous systems to more fully evaluate
pulmonary venous return, including the presence of sub- Significant coarctation of the aorta is a potentially lethal
segmental anomalous pulmonary veins (Fig. 1-60). In condition if left untreated (286). Classically, the site of
addition, phase-velocity flow mapping with MR allows for narrowing occurs just distal to the left subclavian artery,
accurate calculation of the degree of shunting, which can with narrowings proximal to and compromising the left
aid in management decisions (Fig. 1-60) (300). subclavian artery rare (304). An aberrant right subclavian
artery can often arise at or just below the coarctation.
Localized coarctation and tubular hypoplasia of the aorta
Ebstein Anomaly
may also coexist.
Ebstein anomaly is defined as displacement of the attach- MR and CT 3D angiography of the aorta can diagnose
ments of the tricuspid valve leaflet from the AV junction to and distinguish aortic coarctation from other congenital
the RV cavity, with resultant atrialization of portions of the anomalies of the aorta, including pseudocoarctation (305),
RV cavity (301). Both MR and CT imaging can depict the and can identify the precise location of the coarctation and
abnormal tricuspid leaflet attachments, which commonly its relationship with the branch vessels (Fig. 1-62). The
involve only the septal and posterior leaflets (301,302). presence of associated aneurysms and collateral vessels can
Additionally, RV size and function, degree of septal leaflet also be defined. MR phase-contrast velocity mapping can
displacement and tethering, degree of tricuspid regurgita- also provide quantitative information on the severity of the
tion, and associated congenital defects such as atrial septal obstruction and contribution of collateral flow (306,307).
defect can be assessed. Both MR and CT imaging can be used to evaluate the status
of repaired aortic coarctations, although in the setting
of aortic stenting, CT may be more useful in evaluating for
Patent Ductus Arteriosus
in-stent patency (Fig. 1-62) (308).
Patent ductus arteriosus is defined as persistent patency of
the ductus arteriosus beyond the functional closure that
Tetralogy of Fallot
normally occurs following birth. Uncomplicated patent
ductus arteriosus connects the proximal descending aorta Tetralogy of Fallot is characterized by a ventricular septal
below the origin of the left subclavian artery with the roof defect, overriding of the aorta, RVOT obstruction, and RV
Chapter 1: Heart 65
A B C
Figure 1-61 Gradient-echo MR images in the sagittal (A) and coronal (B) planes and volume rendered MR angiogram (C) demonstrate a
patent ductus arteriosus (arrows).
hypertrophy (Fig. 1-63). Other commonly associated vessels. Pulmonary stenosis and regurgitation are common
anomalies include right-sided aortic arch, persistent left- long-term postoperative complications due to the repair of
sided superior vena cava, and anomalies of the coronary the RVOT, which is achieved by pulmonary valvotomy,
arteries. Surgical correction of this complex congenital infundibular incision and resection of infundibular mus-
condition is often performed in infancy and focuses on cle, and/or placement of RVOT and/or pulmonary artery
relief of the RVOT obstruction and closure of the ventricu- patches. These procedures destroy pulmonary valve com-
lar septal defect (309). Prior to surgical correction, pallia- petence, and chronic pulmonary regurgitation with associ-
tive shunt procedures may be performed, commonly the ated RV dilation and RV systolic and diastolic dysfunction
Blalock-Taussig shunt, which involves connection of a sub- are often seen in adulthood (Fig. 1-65) (310–313).
clavian artery to the pulmonary artery (Fig. 1-64), to MR and CT imaging can readily demonstrate the
improve the pulmonary blood flow, thereby decreasing anatomic cardiovascular structural alterations seen in pre-
cyanosis and stimulating growth of often small pulmonary and postoperative tetralogy of Fallot patients. Multislice,
A B C
Figure 1-62 Maximum intensity projection MR image (A) and thin maximum intensity projection CT image (B) from two different patients
with aortic coarctation. Follow-up CT image after aortic stenting of a coarctation (C).
A B C
Figure 1-63 Diagram (A) and gradient-echo MR cine images (B, C) from a patient with uncorrected tetralogy of Fallot, demonstrating
ventricular septal defect (VSD) (black arrow), overriding aorta (Ao), right ventricular hypertrophy, and stenosis of the right ventricular
outflow tract (white arrow). RV, right ventricle; RA, right atrium; PA, pulmonary artery; LV, left ventricle; VS, ventricular septum. (Diagram
courtesy of Dr. Michael Argilla, New York University School of Medicine, New York, New York.)
A B C
Figure 1-64 Gradient-echo (A), thick maximum intensity projection (B), and thin maximum intensity projection (C) MR angiography
images from a patient with a right subclavian–to–pulmonary artery shunt (Blalock-Taussig shunt) (arrows).
Figure 1-65 Right ventricular functional assessment in a patient with tetralogy of Fallot.
Chapter 1: Heart 67
multiphase gradient-echo MR imaging and multiphase CT ing of the circulations though shunting at the atrial, ventric-
imaging can be used to provide quantitative measure- ular and/or great arterial level (patent ductus arteriosus)
ments of RV and LV mass, size, and function (314–317). (Fig. 1-66). Surgical treatment in infants requires redirec-
Selected imaging planes of the RVOT and pulmonary tion of systemic venous flow into the pulmonary artery cir-
valve can further depict structural alterations and assess- cuit and of pulmonary venous flow into the aorta. Early
ment of RVOT and/or pulmonary valve stenosis. MR and surgical therapy consisted of an atrial switch procedure
CT angiography can readily depict great vessel anatomy (Senning or Mustard) in which superior and inferior vena
and evaluate for associated pulmonary artery stenoses. caval blood flow is directed by means of a baffle from the
MR phase-contrast velocity flow measurements can pro- right atrium to the mitral valve and left atrium (320,321).
vide quantification of the degree of pulmonary stenosis This blood is then pumped by the left ventricle to the pul-
and/or associated regurgitation and flow differences be- monary arteries, where it is oxygenated and returned to the
tween the right and left pulmonary vasculature system heart by the pulmonary veins, then around the baffle and
(318,319). through the tricuspid valve, pumped by the right ventricle
to the aorta. In 1975, surgical correction by switching the
aortic and pulmonary artery trunks to redirect ventricular
Complete Transposition of the Great Arteries
outflow allowed establishment of a normal circulatory
Complete transposition of the great arteries results from pathway, with each ventricle now returned to its normal
faulty division of the truncus, resulting in a combination of physiologic function within the circulation (322). Since the
AV concordance and ventriculoarterial discordance (i.e., the development of the arterial switch operation, the atrial
ascending aorta arises from the right ventricle and the pul- switch procedures have largely been abandoned. However,
monary artery arises from the left ventricle) (Fig. 1-66). The given the timeframe in which these procedures were per-
systemic and pulmonary circulations are parallel and inde- formed, a significant number of patients that underwent
pendent closed circuits; hence, surviving infants have mix- the prior atrial switch procedure are now in their mid-adult
A B
C D E
Figure 1-66 Diagram (A) of dextro-looped transposition of the great arteries (dTGA) with associated MR images (B–E) demonstrating
typical findings in a patient undergoing atrial switch operation for dTGA. RV, right ventricle; LV, left ventricle; RA, right atrium; LA, left
atrium; SVB, systemic veins baffle; PVB, pulmonary veins baffle; AV, aortic valve; PV, pulmonary valve; Ao, aorta; PA, pulmonary artery.
(Diagram courtesy of Dr. Michael Argilla, New York University School of Medicine, New York, New York.)
years, and late complications of the operations are still mapping may demonstrate turbulence of flow into sys-
being encountered (323). Because patients that underwent temic and pulmonary venous drainage systems and aid in
the arterial switch procedure are mostly in their twenties or localizing the site of obstruction. Both MR and CT multi-
younger, potential late complications of this procedure phase imaging can be used to quantify ventricular mass,
may not manifest until later in life. volumes, and function.
Common complications of the atrial switch procedure In the arterial switch operation, the great arterial trunks
include cardiac arrhythmias, RV dysfunction, and stenosis are severed just above the sinuses of Valsalva and switched
of the systemic or pulmonary venous inflows. Although so that the aorta arises from the left ventricle and the
different in size, shape, and morphology from the left ven- pulmonary artery from the right. The coronary arteries,
tricle, the systemic right ventricle can tolerate the chronic which arise from the sinuses, also need to be reimplanted
need for an almost fourfold increase in its normal systolic above the left ventricle after the aortic trunk has been
pressure (324) with moderate dilation and hypertrophy of switched. Coronary artery stenoses and occlusions, which
its walls. Only 5% to 10% of patients develop serious RV can lead to LV dysfunction, are complications of this
failure (325,326). Caval or pulmonary venous stenoses are procedure but usually manifest early in the postoperative
direct complications of the surgical procedure and often period. To date, follow-up studies have not demonstra-
involve the reconnection sites. Transposition of the great ted additional complications other than supravalvular
arteries, either untreated or following an atrial switch, is pulmonary stenosis and insufficiency of the neoaortic LV
usually obvious on MR and contrast-enhanced CT valve (327). However, because the follow-up patient
anatomic imaging, which demonstrates the aorta posi- cohort is still young, it is unknown whether other poten-
tioned in front of (anterior to) the pulmonary artery and tial complications related to coronary arteries may become
the ventricles in their normal anatomic positions (Fig. more apparent in later years. CT and MR imaging of the
1-66). Additionally, imaging planes directed toward the coronary arteries can be useful in the noninvasive evalua-
RVOT and LVOT show the relationship of the great arteries tion of the coronary arteries for kinking, stenosis, or
to their respective ventricles. Imaging planes aligned along atherosclerosis (Fig. 1-67). In addition, follow-up of the
the inflow of the superior and inferior venae cavae and pulmonary artery, aorta, and semilunar valves can be
pulmonary veins can sometimes demonstrate stenoses. performed to diagnose and follow potential complications
Gradient-echo multiphase MR with and without velocity related to the operation.
A,B
Figure 1-67 A–D: CT assess-

ment of the coronary arteries after
reimplantation with the arterial
switch operation in a patient with
dextro-looped transposition of the
great arteries. There is no kinking or
stenosis noted at the origin of the
left coronary artery (white arrows) or
C, D right coronary artery (black arrows).
Chapter 1: Heart 69
Congenitally Corrected Transposition of the normal, these patients often do not present until adult-
Great Arteries hood, when complications related to the systemic func-
tioning morphologic right ventricle occur. The coronary
Congenitally corrected transposition of the great arteries is arteries are inverted along with the ventricles (i.e., the mor-
a condition in which both the AV and ventriculoarterial phologic right coronary artery is on the left side, and the
connections are discordant (Fig. 1-68). Unlike complete morphologic left coronary artery is on the right side)
transposition of the great arteries described earlier, in this (328). Nearly all of these patients have associated anom-
condition, the patients have a second anomalous connec- alies, including ventricular septal defect, pulmonary steno-
tion (ventricular inversion) that “corrects” the transposi- sis, or left-sided (tricuspid) AV valve abnormalities, which
tion. Essentially, the left-sided ventricle has the morpho- may be a dysplastic valve or Ebstein anomaly (329).
logic appearance of the right ventricle and the right-sided Associated long-term complications include hypertrophy,
ventricle has the morphologic appearance of the left ven- dilation, and dysfunction of the systemic morphologic
tricle. As such, blood flows from the systemic veins into right ventricle, insufficiency of the left-sided AV valve, and
the right atrium, across the right-sided AV valve (often atrial arrhythmias related to left atrial dilation. Although
bileaflet) into the morphologic left ventricle (position of often diagnosed and monitored by echocardiography, con-
the normal right ventricle) and is pumped into the pul- genitally corrected transposition of the great arteries can be
monary artery, where it is oxygenated and returned via the also be diagnosed on MR and CT by recognition of the
pulmonary veins into the left atrium, crosses the left-sided AV and ventriculoarterial discordance. The aortic root is
AV valve (often trileaflet) into the morphologic right ven- usually anterior, to the left, and superior to the pulmonary
tricle (position of the normal left ventricle) and is pumped artery. In addition, MR and CT can aid in differentia-
into the aorta. Because blood flow is hemodynamically ting the left-sided morphologic right ventricle (muscular
A B
C D
Figure 1-68 Diagram (A) of congenitally corrected transposition of the great arteries with associated MR images (B–D) demonstrating
the typical findings. RV, right ventricle; LV, left ventricle; RA, right atrium; LA, left atrium; Ao, aorta; PA, pulmonary artery. (Diagram courtesy
of Dr. Michael Argilla, New York University School of Medicine, New York, New York.)
infundibulum separating the AV and semilunar valves, the opposite artery (Fig. 1-71). Although extremely rare,
prominent trabeculations, and moderator band) from the anomalous origins may also be from the ventricles, pul-
right-sided morphologic left ventricle (no muscular infun- monary artery (Fig. 1-72), aortic arch, branch vessels, or de-
dibulum, fibrous continuity of the AV and semilunar scending aorta. The functional significance of these origins,
valves, and smooth walls) and the right atrium from the however, often depends on the course taken by the artery.
left atrium as well as identify associated lesions, such as Arteries can course (a) anterior to the RVOT (Fig. 1-73), (b)
pulmonary stenosis (298,330–333). As in other types of posterior to the aorta (Fig. 1-74), (c) between the aorta and
congenital heart disease, MR and CT can also be used to RVOT (Fig. 1-75), (d) between the aorta and main pul-
assess ventricular volume, mass, and function. MR can also monary artery (Fig. 1-76), (e) with intramyocardial bridg-
evaluate valvular regurgitation and quantitate intracardiac ing (Fig. 1-77), or (f) intramurally (within the aortic wall).
shunts (293). Clinically significant anomalies of coronary artery origin
and course generally present before midlife and often in
the context of strenuous physical exercise with resultant
Coronary Artery Anomaly
hypoperfusion and myocardial ischemia, which may lead
Anomalies of the coronary arteries affect about 1% of the to angina, cardiac arrhythmias, or syncope. Abnormalities
population, with 87% of these having anomalies of the with coronary terminations or fistulae often affect the right
origin and distribution and 13% having coronary artery coronary artery. Fistulae may drain anywhere from between
fistulae (334,335). The true incidence of ectopic origin of the vena cava or coronary sinus to the pulmonary artery
the coronary arteries from the aorta in the population is (Fig. 1-78) or left atrium (339). More than 90% of fistulae
unknown, but it is estimated to be between 0.17% and drain into the right side of the heart and therefore cause a
0.6%. Several classifications have been suggested for char- shunt, and chronic large volume shunting can lead to
acterizing congenital coronary artery anomalies. Coronary aneurysmal enlargement of the proximal feeding coronary
artery anomalies are classified as such if they are found in artery as well as the receiving vessel or chamber (Fig. 1-72)
less than 1% of the unselected population and can be (340). Symptoms due to coronary artery fistulae are a result
characterized by origin, course, termination or connection, of right-sided heart dilation or of coronary steal leading
and coronary size (336,337). Approximately 20% of coro- to dyspnea, chest pain, or arrhythmias. Intrinsic coronary
nary anomalies have associated clinical consequences, such artery anatomy, such as ostial stenosis; atresia; or single,
as myocardial ischemia or infarction (338). A coronary ar- absent, or hypoplastic coronary arteries, may have clinical
tery is considered to have an anomalous origin when the implications.
ostium (a) is located above the sinotubular junction (Fig. Cardiac catheterization has traditionally been the pre-
1-69), or arises (b) from the opposite coronary sinus ferred imaging modality for characterization of coronary
(Fig. 1-70), (c) from the non–coronary sinus, or (d) from artery anomalies. However, given the complex 3D nature
A,B
Figure 1-69 A, B: CT volume rendered images demonstrate high take-off (above the sinotubular junction) of the right coronary artery
(RCA). LCA, left coronary artery.
Chapter 1: Heart 71
A,B
Figure 1-70 CT volume rendered (A) and oblique multiplanar reconstruction (B) images demonstrate the right coronary artery arising
from the left aortic cusp (arrows).
A,B
Figure 1-71 A, B: CT volume rendered images demonstrate the left coronary system arising from the right coronary artery (arrow).
Figure 1-73 MR coronary angiogram demonstrating an anom-

alous right coronary artery coursing anterior to the pulmonary
artery (arrows).
Figure 1-72 CT volume rendered image demonstrating origin

of the right coronary artery (RCA) from the pulmonary artery (PA).
LCA, left coronary artery; Ao, aorta.
A,B
Figure 1-74 A, B: CT volume rendered images demonstrating the left circumflex (LCX) artery coursing posterior to the aorta (Ao). LM,
left main; LAD, left anterior descending; RCA, right coronary artery; LA, left atrium; Dg, diagonal artery.
Chapter 1: Heart 73
A,B
C, D, E
Figure 1-75 CT volume rendered (A, B) and oblique multiplanar reconstruction (C–E) images demonstrate the left coronary system aris-
ing from the right coronary artery (RCA) and coursing between the right ventricular outflow tract and aorta (Ao). LM, left main; LAD, left an-
terior descending; LCX, left circumflex; RCA, right coronary artery.
A,B
C, D, E
Figure 1-76 A–E: CT volume rendered and oblique multiplanar reconstruction images demonstrate the left coronary artery coursing
between the main pulmonary artery (PA) and aorta (Ao). LM, left main; LAD, left anterior descending; LCX, left circumflex; RCA, right
coronary artery.
A B C
Figure 1-77 CT volume rendered (A) and oblique multiplanar reconstruction (B, C) images from a patient with a myocardial bridge
(arrows) involving the mid segment of the left anterior descending artery.
of these anomalies, conventional angiography not infrequency ablation of early electrical potentials arising from
quently incompletely delineates the anatomic origin and the pulmonary vein ostia. Imaging evaluation of the left
course of the coronary artery. MR and CT angiography of atrial anatomy and pulmonary venous anatomy can aid
the coronary arteries have therefore become the accepted the electrophysiologist in preprocedural planning and can
standards for complete evaluation of coronary artery be used to follow up for postprocedural complications.
anomalies. With the high spatial resolution, volumetric 3D visualization of the left atrium and pulmonary venous
acquisition, and orientation in any plane, CT and MR can anatomy is often used to guide the electrophysiologist in
accurately depict the origin and course of coronary artery navigating the left atrial cavity during catheter ablation.
anomalies with high sensitivity, although specificity is Knowledge of aberrant pulmonary veins can help further
reduced in patients with high HRs because of the limited guide procedures to ensure isolation of the electrical
temporal resolution even with cardiac gating (340,341). potentials arising from all the pulmonary veins. Diagnosis
for the presence of left atrial thrombus, especially preva-
lent in the left atrial appendage, is important to prevent
OTHER APPLICATIONS iatrogenic systemic embolism (Fig. 1-79).
The location, size, and number of veins need to be
defined. Ostial branches, venous branches noted within
Left Atrial Anatomy
5 mm of the atriopulmonary venous junction, are also
There has been a growth of interventional procedures for important to note. Commonly there are four pulmonary
the treatment of atrial arrhythmias. In particular, atrial fib- veins with separate ostia into the left atrium (Fig. 1-80).
rillation, which affects approximately 2 million people in However, accessory pulmonary veins and common
the United States (342), can now be treated using radiofre- or conjoined veins may also be present. Accessory
A B C
Figure 1-78 A–C: CT volume rendered images demonstrate a coronary fistula extending from the left anterior descending artery to the
right pulmonary artery (arrows), and draining into the right pulmonary artery superiorly (*).
Chapter 1: Heart 75
include stroke, hemopericardium, hemothorax, pulmonary

vein thrombosis, and hemodynamically significant pul-
monary vein stenosis, which can result in pulmonary veno-
occlusive disease, including focal pulmonary edema (344).
Pulmonary vein stenosis may develop up to 8 months after
the procedure. Current ablation techniques, however, have
greatly diminished the complication rate of pulmonary
venous stenosis.
Imaging of the left atrial and pulmonary venous
anatomy can be performed with CT or MR. 3D MR angio-
graphy or CT angiography imaging can be performed as a
gated or nongated study and should encompass the area
from the aortic arch through the apex of the heart during a
single breath-hold (345). Post-processing of the image
datasets, including 2D multiplanar reconstructions of the
Figure 1-79 CT image demonstrating thrombus within the left pulmonary venous ostia and atrial appendage, and 3D vol-
atrial appendage (arrow).
ume rendered and shaded surface displays, demonstrates
the complex anatomy (Figs. 1-80 to 1-82). MR phase-
pulmonary veins are additional veins with independent velocity flow mapping of the pulmonary veins can
ostia into the atrium. They are more frequently seen on the demonstrate elevated flow in patients with significant pul-
right side and are named for the pulmonary lobe or monary vein stenosis. In our institution, volumetric
segment that they drain (e.g., right middle lobe or superior datasets acquired through the left atrium are synchronized
segment of right lower lobe) (Fig. 1-81). In contrast, and fused with electrical mapping systems in the electro-
conjoined veins occur when superior and inferior veins physiology (EP) laboratory (Fig. 1-84).
join proximal to the left atrium. Conjoined veins occur
more frequently on the left side (Fig. 1-82). Anomalous
Coronary Venous Anatomy
pulmonary venous return occurs when part or all of the
pulmonary veins drain into a structure other than the left Several interventional procedures are taking advantage of
atrium (Fig. 1-60). the coronary venous system for optimizing therapy. New-
Postprocedural complications of radiofrequency abla- generation pacemaker devices now employ dual-chamber
tion include endocardial scarring (Fig. 1-83), pulmonary pacing for optimization of ventricular filling in patients
vein dissection, and perforation (343). Because vagal nerve with heart failure. New nonsurgical techniques for repair
fibers lie within the walls of the pulmonary veins, signifi- of mitral regurgitation employ a device aimed at adjusting
cant bradyarrhythmias, including asystole, may occur annular size via the coronary sinus. CT and MR imaging of
(343). Small pleural or pericardial effusions, small atrial the coronary venous anatomy can provide valuable infor-
septal defects, and pulmonary venous stenosis are also mation on the optimal placement of coronary venous
complications that may be seen. Serious complications catheters and wires and may be seen more in the future.
A B C
Figure 1-80 Oblique multiplanar reconstruction (A, B) and posterior view volume rendered (C) CT images demonstrate four normal pul-
monary veins emptying into the left atrium.
A B
Figure 1-81 Oblique multiplanar reconstruction (A) and posterior view volume rendered (B) CT images demonstrate that the right mid-
dle lobe veins empty into the left atrium via a separate ostium (arrows).
A B C
Figure 1-82 Oblique multiplanar reconstruction (A, B) and posterior view volume rendered (C) CT images demonstrate that the left-
sided pulmonary veins empty into the left atrium via a common ostium (arrows).
Chapter 1: Heart 77
A B
Figure 1-83 A, B: CT images demonstrate ostial pulmonary vein stenosis (arrows).
A B
Figure 1-84 Synchronization of electrical mapping system (A) with the CT volume rendered images of the left atrium (B). RSPV, right
superior pulmonary vein; LAA, left atrial appendage; LIPV, left inferior pulmonary vein.
REFERENCES 23. Griswold MA, Jakob PM, Heidemann RM, et al. Generalized au-
tocalibrating partially parallel acquisitions (GRAPPA). Magn
Reson Med. 2002;47:1202–1210.
1. Wintersperger BJ, Nikolaou K. Basics of cardiac MDCT: tech- 24. Clay S, Alfakih K, Radjenovic A, et al. Normal range of human left
niques and contrast application. Eur Radiol. 2005;15 (suppl 2): ventricular volumes and mass using steady state free precession
B2–9. MRI in the radial long axis orientation. Magma. 2006;19:41–45.
2. Flohr T, Stierstorfer K, Raupach R, et al. Performance evaluation 25. Paelinck BP, Lamb HJ, Bax JJ, et al. Assessment of diastolic func-
of a 64-slice CT system with z-flying focal spot. Rofo. 2004;176: tion by cardiovascular magnetic resonance. Am Heart J.
1803–1810. 2002;144:198–205.
3. Ohnesorge B, Becker C, Flohr T, Reiser MF. Multi-slice CT in car- 26. Nishimura RA, Abel MD, Hatle LK, Tajik AJ. Assessment of dias-
diac imaging: technical principles, clinical application and future de- tolic function of the heart: background and current applications
velopments. Berlin, Germany: Springer-Verlag; 2002:3–109. of Doppler echocardiography. Part II. Clinical studies. Mayo Clin
4. Flohr T, Ohnesorge B. Heart rate adaptive optimization of spa- Proc. 1989;64:181–204.
tial and temporal resolution for electrocardiogram-gated multi- 27. Edvardsen T, Rosen BD, Pan L, et al. Regional diastolic dysfunc-
slice spiral CT of the heart. J Comput Assist Tomogr. 2001; tion in individuals with left ventricular hypertrophy measured
25:907–923. by tagged magnetic resonance imaging—the Multi-Ethnic Study
5. Schroeder S, Kopp AF, Baumbach A, et al. Noninvasive detection of Atherosclerosis (MESA). Am Heart J. 2006;151:109–114.
and evaluation of atherosclerotic coronary plaques with multislice 28. Lotz J, Meier C, Leppert A, Galanski M. Cardiovascular flow
computed tomography. J Am Coll Cardiol. 2001;37:1430–1435. measurement with phase-contrast MR imaging: basic facts and
6. Boudoulas H, Rittgers SE, Lewis RP, et al. Changes in diastolic implementation. Radiographics. 2002;22:651–671.
time with various pharmacologic agents: implication for my- 29. Brenner LD, Caputo GR, Mostbeck G, et al. Quantification of
ocardial perfusion. Circulation. 1979;60:164–169. left to right atrial shunts with velocity-encoded cine nuclear
7. Boudoulas H, Lewis RP, Rittgers SE, et al. Increased diastolic magnetic resonance imaging. J Am Coll Cardiol. 1992;20:
time: a possible important factor in the beneficial effect of pro- 1246–1250.
pranolol in patients with coronary artery disease. J Cardiovasc 30. Hundley WG, Li HF, Lange RA, et al. Assessment of left-to-right
Pharmacol. 1979;1:503–513. intracardiac shunting by velocity-encoded, phase-difference
8. Shim SS, Kim Y, Lim SM. Improvement of image quality with magnetic resonance imaging. A comparison with oximetric and
beta-blocker premedication on ECG-gated 16-MDCT coronary indicator dilution techniques. Circulation. 1995;91:2955–2960.
angiography. AJR Am J Roentgenol. 2005;184:649–654. 31. Gray H. Anatomy of the human body. Philadelphia: Lea & Febiger;
9. de Feyter PJ, Nieman K. New coronary imaging techniques: 1918.
what to expect? Heart. 2002;87:195–197. 32. Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardized
10. Kopp AF, Schroeder S, Kuettner A, et al. Non-invasive coronary myocardial segmentation and nomenclature for tomographic
angiography with high resolution multidetector-row computed imaging of the heart: a statement for healthcare professionals
tomography. Results in 102 patients. Eur Heart J. 2002; from the Cardiac Imaging Committee of the Council on
23:1714–1725. Clinical Cardiology of the American Heart Association.
11. Gerber TC, Kuzo RS, Karstaedt N, et al. Current results and new Circulation. 2002;105:539–542.
developments of coronary angiography with use of contrast-en- 33. Austen WG, Edwards JE, Frye RL, et al. A reporting system on pa-
hanced computed tomography of the heart. Mayo Clin Proc. tients evaluated for coronary artery disease. Report of the Ad
2002;77:55–71. Hoc Committee for Grading of Coronary Artery Disease,
12. Nikolaou K, Flohr T, Knez A, et al. Advances in cardiac CT imag- Council on Cardiovascular Surgery, American Heart Associ-
ing: 64-slice scanner. Int J Cardiovasc Imaging. 2004;20:535–540. ation. Circulation. 1975;51:5–40.
13. Pannu HK, Flohr TG, Corl FM, Fishman EK. Current concepts in 34. Miller S. Normal angiographic anatomy and measurements. Cardiac
multi-detector row CT evaluation of the coronary arteries: prin- angiography. Boston: Little, Brown and Company; 1984:51–71.
ciples, techniques, and anatomy. Radiographics. 2003;23 (spec 35. Gardner E, Gray D, O’Rahilly R. Anatomy: A regional study of
no): S111–125. human structure, 4th ed. Philadelphia: WB Saunders; 1975.
14. Hill JA. Nonionic contrast use in cardiac angiography. Invest 36. Goss C. Gray’s anatomy of the human body, 29th ed. Philadelphia:
Radiol. 1993;28 (suppl 5): S48–53; discussion S54. Lea & Febiger; 1973.
15. Gilard M, Pennec PY, Cornily JC, et al. Multi-slice computer to- 37. Edwards E. Applied anatomy of the heart. In: Giuliani E, Fuster
mography of left ventricular function with automated analysis V, eds. Cardiology: fundamentals and practice, 2nd ed. St. Louis:
software in comparison with conventional ventriculography. Mosby-Year Book; 1991:47–51.
Eur J Radiol. 2006; 59:270–275. 38. Fayad ZA, Fuster V. Clinical imaging of the high-risk or vulnera-
16. Dirksen MS, Bax JJ, de Roos A, et al. Usefulness of dynamic mul- ble atherosclerotic plaque. Circ Res. 2001;89:305–316.
tislice computed tomography of left ventricular function in un- 39. O’Malley PG, Greenberg BA, Taylor AJ. Cost-effectiveness of
stable angina pectoris and comparison with echocardiography. using electron beam computed tomography to identify patients
Am J Cardiol. 2002;90:1157–1160. at risk for clinical coronary artery disease. Am Heart J. 2004;148:
17. Mahnken AH, Spuentrup E, Niethammer M, et al. Quantitative 106–113.
and qualitative assessment of left ventricular volume with ECG- 40. Rumberger JA, Behrenbeck T, Breen JF, Sheedy PF 2nd. Coronary
gated multislice spiral CT: value of different image reconstruc- calcification by electron beam computed tomography and ob-
tion algorithms in comparison to MRI. Acta Radiol. 2003; structive coronary artery disease: a model for costs and effective-
44:604–611. ness of diagnosis as compared with conventional cardiac testing
18. Juergens KU, Grude M, Maintz D, et al. Multi-detector row CT of methods. J Am Coll Cardiol. 1999;33:453–462.
left ventricular function with dedicated analysis software versus 41. Raggi P, Callister TQ, Cooil B, et al. Evaluation of chest pain in
MR imaging: initial experience. Radiology. 2004;230:403–410. patients with low to intermediate pretest probability of coro-
19. Fischer SE, Wickline SA, Lorenz CH. Novel real-time R-wave de- nary artery disease by electron beam computed tomography. Am
tection algorithm based on the vectorcardiogram for accurate J Cardiol. 2000;85:283–288.
gated magnetic resonance acquisitions. Magn Reson Med. 42. Clouse ME. How useful is computed tomography for screening
1999;42:361–370. for coronary artery disease? Noninvasive screening for coronary
20. Winterer JT, Lehnhardt S, Schneider B, et al. MRI of heart mor- artery disease with computed tomography is useful. Circulation.
phology. Comparison of nongradient echo sequences with sin- 2006;113:125–146; discussion 125–146.
gle- and multislice acquisition. Invest Radiol. 1999;34:516–522. 43. O’Rourke RA, Brundage BH, Froelicher VF, et al. American
21. Lee VS. Cardiovascular MRI: physical principles to practical protocols. College of Cardiology/American Heart Association Expert
Philadelphia: Lippincott Williams & Wilkins; 2006. Consensus document on electron-beam computed tomography
22. Goldfarb JW. The SENSE ghost: field-of-view restrictions for for the diagnosis and prognosis of coronary artery disease.
SENSE imaging. J Magn Reson Imaging. 2004;20:1046–1051. Circulation. 2000;102:126–140.
Chapter 1: Heart 79
44. Agatston AS, Janowitz WR, Hildner FJ, et al. Quantification of 65. Fayad ZA, Fuster V, Fallon JT, et al. Noninvasive in vivo human
coronary artery calcium using ultrafast computed tomography. J coronary artery lumen and wall imaging using black-blood
Am Coll Cardiol. 1990;15:827–832. magnetic resonance imaging. Circulation. 2000;102:506–510.
45. Becker CR, Kleffel T, Crispin A, et al. Coronary artery calcium 66. Botnar RM, Stuber M, Kissinger KV, et al. Noninvasive coronary
measurement: agreement of multirow detector and electron vessel wall and plaque imaging with magnetic resonance imag-
beam CT. AJR Am J Roentgenol. 2001;176:1295–1298. ing. Circulation. 2000;102:2582–2587.
46. Callister TQ, Cooil B, Raya SP, et al. Coronary artery disease: 67. Kim WY, Stuber M, Bornert P, et al. Three-dimensional black-
improved reproducibility of calcium scoring with an electron- blood cardiac magnetic resonance coronary vessel wall imaging
beam CT volumetric method. Radiology. 1998;208:807–814. detects positive arterial remodeling in patients with nonsignifi-
47. Hong C, Becker CR, Schoepf UJ, et al. Coronary artery calcium: cant coronary artery disease. Circulation. 2002;106:296–299.
absolute quantification in nonenhanced and contrast-enhanced 68. Spuentrup E, Botnar RM. Coronary magnetic resonance imag-
multi-detector row CT studies. Radiology. 2002;223:474–480. ing: visualization of the vessel lumen and the vessel wall and
48. Hong C, Bae KT, Pilgram TK. Coronary artery calcium: accuracy molecular imaging of arteriothrombosis. Eur Radiol. 2006;16:
and reproducibility of measurements with multi-detector row 1–14.
CT—assessment of effects of different thresholds and quantifi- 69. Fernandez GC. Bypass graft imaging and coronary anomalies in
cation methods. Radiology. 2003;227:795–801. MDCT. Eur Radiol. 2005;15 (suppl 2):B59–61.
49. Horiguchi J, Yamamoto H, Akiyama Y, et al. Coronary artery 70. Nesto RW, Kowalchuk GJ. The ischemic cascade: temporal
calcium scoring using 16-MDCT and a retrospective ECG-gating sequence of hemodynamic, electrocardiographic and sympto-
reconstruction algorithm. AJR Am J Roentgenol. 2004;183: matic expressions of ischemia. Am J Cardiol. 1987;59:23C–30C.
103–108. 71. Tillisch J, Brunken R, Marshall R, et al. Reversibility of cardiac
50. Knez A, Becker CR, Leber A, et al. Usefulness of multislice spiral wall-motion abnormalities predicted by positron tomography.
computed tomography angiography for determination of coro- N Engl J Med. 1986;314:884–888.
nary artery stenoses. Am J Cardiol. 2001;88:1191–1194. 72. Meluzin J, Cerny J, Frelich M, et al. Prognostic value of the
51. Achenbach S, Giesler T, Ropers D, et al. Detection of coronary amount of dysfunctional but viable myocardium in revascular-
artery stenoses by contrast-enhanced, retrospectively electrocar- ized patients with coronary artery disease and left ventricular
diographically-gated, multislice spiral computed tomography. dysfunction. Investigators of this Multicenter Study. J Am Coll
Circulation. 2001;103:2535–2538. Cardiol. 1998;32:912–920.
52. Nieman K, Oudkerk M, Rensing BJ, et al. Coronary angiography 73. Bax JJ, Poldermans D, Elhendy A, et al. Improvement of left ven-
with multi-slice computed tomography. Lancet. 2001;357: tricular ejection fraction, heart failure symptoms and prognosis
599–603. after revascularization in patients with chronic coronary artery
53. Becker CR, Ohnesorge BM, Schoepf UJ, Reiser MF. Current de- disease and viable myocardium detected by dobutamine stress
velopment of cardiac imaging with multidetector-row CT. Eur J echocardiography. J Am Coll Cardiol. 1999;34:163–169.
Radiol. 2000;36:97–103. 74. Schuijf JD, Shaw LJ, Wijns W, et al. Cardiac imaging in coronary
54. Becker CR, Nikolaou K, Muders M, et al. Ex vivo coronary ather- artery disease: differing modalities. Heart. 2005;91:1110–1117.
osclerotic plaque characterization with multi-detector-row CT. 75. Schwitter J. MR-Impact: Comparison of MR Myocardial
Eur Radiol. 2003;13:2094–2098. Perfusion Imaging with Single Photon Emission Computed
55. Nikolaou K, Becker CR, Muders M, et al. Multidetector-row Tomography in Known or Suspected Coronary Artery Disease: A
computed tomography and magnetic resonance imaging of ath- Multicentre, Multivendor Dose Finding Study. Conference of
erosclerotic lesions in human ex vivo coronary arteries. Athero- the European Society of Cardiology. Stockhold, Sweden, 2005.
sclerosis. 2004;174:243–252. 76. Cullen JH, Horsfield MA, Reek CR, et al. A myocardial perfusion
56. Pugliese F, Mollet NR, Runza G, et al. Diagnostic accuracy of reserve index in humans using first-pass contrast-enhanced
non-invasive 64-slice CT coronary angiography in patients with magnetic resonance imaging. J Am Coll Cardiol. 1999;33:
stable angina pectoris. Eur Radiol. 2006;16:575–582. 1386–1394.
57. Mollet NR, Cademartiri F, van Mieghem CA, et al. High- 77. Gatehouse PD, Elkington AG, Ablitt NA, et al. Accurate assess-
resolution spiral computed tomography coronary angiography ment of the arterial input function during high-dose myocardial
in patients referred for diagnostic conventional coronary perfusion cardiovascular magnetic resonance. J Magn Reson
angiography. Circulation. 2005;112:2318–2323. Imaging. 2004;20:39–45.
58. Ropers D, Baum U, Pohle K, et al. Detection of coronary artery 78. Al-Saadi N, Nagel E, Gross M, et al. Noninvasive detection of
stenoses with thin-slice multi-detector row spiral computed myocardial ischemia from perfusion reserve based on cardiovas-
tomography and multiplanar reconstruction. Circulation. 2003; cular magnetic resonance. Circulation. 2000;101:1379–1383.
107:664–666. 79. Hundley WG, Hamilton CA, Thomas MS, et al. Utility of fast
59. Schroeder S, Kopp AF, Baumbach A, et al. Noninvasive detection cine magnetic resonance imaging and display for the detection
of coronary lesions by multislice computed tomography: results of myocardial ischemia in patients not well suited for second
of the New Age pilot trial. Catheter Cardiovasc Interv. 2001; harmonic stress echocardiography. Circulation. 1999;100:
53:352–358. 1697–1702.
60. Kuettner A, Trabold T, Schroeder S, et al. Noninvasive detection 80. Nagel E, Lehmkuhl HB, Bocksch W, et al. Noninvasive diagnosis
of coronary lesions using 16-detector multislice spiral com- of ischemia-induced wall motion abnormalities with the use of
puted tomography technology: initial clinical results. J Am Coll high-dose dobutamine stress MRI: comparison with dobuta-
Cardiol. 2004;44:1230–1237. mine stress echocardiography. Circulation. 1999;99:763–770.
61. Leber AW, Knez A, von Ziegler F, et al. Quantification of obstruc- 81. Paetsch I, Jahnke C, Wahl A, et al. Comparison of dobutamine
tive and nonobstructive coronary lesions by 64-slice computed stress magnetic resonance, adenosine stress magnetic resonance,
tomography: a comparative study with quantitative coronary and adenosine stress magnetic resonance perfusion. Circulation.
angiography and intravascular ultrasound. J Am Coll Cardiol. 2004;110:835–842.
2005;46:147–154. 82. American Heart Association. Heart disease and stroke statistics,
62. Appelbaum E, Botnar RM, Yeon SB, Manning WJ. Coronary 2005 update. Dallas, TX: American Heart Association;2005.
magnetic resonance imaging: current state-of-the-art. Coron 83. Kaandorp TA, Lamb HJ, van der Wall EE, et al. Cardiovascular
Artery Dis. 2005;16:345–353. MR to access myocardial viability in chronic ischaemic LV dys-
63. Danias PG, Roussakis A, Ioannidis JP. Diagnostic performance function. Heart. 2005;91:1359–1365.
of coronary magnetic resonance angiography as compared 84. Allman KC, Shaw LJ, Hachamovitch R, Udelson JE. Myocardial
against conventional X-ray angiography: a meta-analysis. J Am viability testing and impact of revascularization on prognosis in
Coll Cardiol. 2004;44:1867–1876. patients with coronary artery disease and left ventricular dys-
64. Kim WY, Danias PG, Stuber M, et al. Coronary magnetic reso- function: a meta-analysis. J Am Coll Cardiol. 2002;39:1151–1158.
nance angiography for the detection of coronary stenoses. N 85. Rahimtoola SH. The hibernating myocardium. Am Heart J.
Engl J Med. 2001;345:1863–1869. 1989;117:211–221.
86. Vanoverschelde JL, Wijns W, Depre C, et al. Mechanisms of 107. Sardanelli F, Zandrino F, Molinari G, et al. MR evaluation of
chronic regional postischemic dysfunction in humans. New coronary stents with navigator echo and breath-hold cine gradi-
insights from the study of noninfarcted collateral-dependent ent-echo techniques. Eur Radiol. 2002;12:193–200.
myocardium. Circulation. 1993;87:1513–1523. 108. Hug J, Nagel E, Bornstedt A, et al. Coronary arterial stents: safety
87. Dubnow MH, Burchell HB, Titus JL. Postinfarction ventricular and artifacts during MR imaging. Radiology. 2000;216:781–787.
aneurysm. A clinicomorphologic and electrocardiographic study 109. Buecker A, Spuentrup E, Ruebben A, Gunther RW. Artifact-free
of 80 cases. Am Heart J. 1965; 70:753–760. in-stent lumen visualization by standard magnetic resonance
88. Perrone-Filardi P, Pace L, Prastaro M, et al. Assessment of my- angiography using a new metallic magnetic resonance imaging
ocardial viability in patients with chronic coronary artery dis- stent. Circulation. 2002;105:1772–1775.
ease. Rest-4-hour-24-hour 201Tl tomography versus dobuta- 110. Spuentrup E, Ruebben A, Mahnken A, et al. Artifact-free coro-
mine echocardiography. Circulation. 1996;94:2712–2719. nary magnetic resonance angiography and coronary vessel wall
89. Baer FM, Voth E, Schneider CA, et al. Comparison of low-dose imaging in the presence of a new, metallic, coronary magnetic
dobutamine-gradient-echo magnetic resonance imaging and resonance imaging stent. Circulation. 2005;111:1019–1026.
positron emission tomography with [18F] fluorodeoxyglucose 111. Anders K, Baum U, Schmid M, et al. Coronary artery bypass graft
in patients with chronic coronary artery disease. A functional (CABG) patency: Assessment with high-resolution submillime-
and morphological approach to the detection of residual my- ter 16-slice multidetector-row computed tomography (MDCT)
ocardial viability. Circulation. 1995;91:1006–1015. versus coronary angiography. Eur J Radiol. 2006;57: 336–344.
90. Baer FM, Theissen P, Schneider CA, et al. Dobutamine magnetic 112. Marano R, Storto ML, Maddestra N, Bonomo L. Non-invasive
resonance imaging predicts contractile recovery of chronically assessment of coronary artery bypass graft with retrospectively
dysfunctional myocardium after successful revascularization. J ECG-gated four-row multi-detector spiral computed tomogra-
Am Coll Cardiol. 1998;31:1040–1048. phy. Eur Radiol. 2004;14:1353–1362.
91. Lauerma K, Virtanen KS, Sipila LM, et al. Multislice MRI in assess- 113. Fitzgibbon GM, Kafka HP, Leach AJ, et al. Coronary bypass graft
ment of myocardial perfusion in patients with single-vessel proxi- fate and patient outcome: angiographic follow-up of 5,065
mal left anterior descending coronary artery disease before and grafts related to survival and reoperation in 1,388 patients dur-
after revascularization. Circulation. 1997;96: 2859–2867. ing 25 years. J Am Coll Cardiol. 1996;28:616–626.
92. Sensky PR, Jivan A, Hudson NM, et al. Coronary artery disease: 114. van Geuns RJ, Wielopolski PA, de Bruin HG, et al. MR coronary
combined stress MR imaging protocol-one-stop evaluation of angiography with breath-hold targeted volumes: preliminary
myocardial perfusion and function. Radiology. 2000;215:608–614. clinical results. Radiology. 2000;217:270–277.
93. Wintersperger BJ, Penzkofer HV, Knez A, et al. Multislice MR 115. White RD, Caputo GR, Mark AS, et al. Coronary artery bypass
perfusion imaging and regional myocardial function analysis: graft patency: noninvasive evaluation with MR imaging.
complimentary findings in chronic myocardial ischemia. Int J Radiology. 1987;164:681–686.
Card Imaging. 1999;15:425–434. 116. White RD, Pflugfelder PW, Lipton MJ, Higgins CB. Coronary ar-
94. Simonetti OP, Kim RJ, Fieno DS, et al. An improved MR imag- tery bypass grafts: evaluation of patency with cine MR imaging.
ing technique for the visualization of myocardial infarction. AJR Am J Roentgenol. 1988;150:1271–1274.
Radiology. 2001;218:215–223. 117. Molinari G, Sardanelli F, Zandrino F, et al. Value of navigator
95. Kim RJ, Fieno DS, Parrish TB, et al. Relationship of MRI delayed echo magnetic resonance angiography in detecting occlusion/
contrast enhancement to irreversible injury, infarct age, and patency of arterial and venous, single and sequential coronary
contractile function. Circulation. 1999;100:1992–2002. bypass grafts. Int J Card Imaging. 2000;16:149–160.
96. Maintz D, Seifarth H, Raupach R, et al. 64-slice multidetector 118. Vrachliotis TG, Bis KG, Aliabadi D, et al. Contrast-enhanced
coronary CT angiography: in vitro evaluation of 68 different breath-hold MR angiography for evaluating patency of coro-
stents. Eur Radiol. 2006;16:818–826. nary artery bypass grafts. AJR Am J Roentgenol. 1997;168:
97. Seifarth H, Raupach R, Schaller S, et al. Assessment of coronary 1073–1080.
artery stents using 16-slice MDCT angiography: evaluation of a 119. Bunce NH, Lorenz CH, John AS, et al. Coronary artery bypass
dedicated reconstruction kernel and a noise reduction filter. Eur graft patency: assessment with true ast imaging with steady-state
Radiol. 2005;15:721–726. precession versus gadolinium-enhanced MR angiography.
98. Maintz D, Grude M, Fallenberg EM, et al. Assessment of coro- Radiology. 2003;227:440–446.
nary arterial stents by multislice-CT angiography. Acta Radiol. 120. Wintersperger BJ, Engelmann MG, von Smekal A, et al. Patency
2003;44:597–603. of coronary bypass grafts: assessment with breath-hold contrast-
99. Kruger S, Mahnken AH, Sinha AM, et al. Multislice spiral com- enhanced MR angiography—value of a non-electrocardiograph-
puted tomography for the detection of coronary stent restenosis ically triggered technique. Radiology. 1998;208:345–351.
and patency. Int J Cardiol. 2003;89:167–172. 121. Langerak SE, Vliegen HW, Jukema JW, et al. Value of magnetic
100. Ligabue G, Rossi R, Ratti C, et al. Noninvasive evaluation of resonance imaging for the noninvasive detection of stenosis in
coronary artery stents patency after PTCA: role of multislice coronary artery bypass grafts and recipient coronary arteries.
computed tomography. Radiol Med (Torino). 2004;108:128–137. Circulation. 2003;107:1502–1508.
101. Mahnken AH, Buecker A, Wildberger JE, et al. Coronary artery 122. Langerak SE, Vliegen HW, de Roos A, et al. Detection of vein
stents in multislice computed tomography: in vitro artifact eval- graft disease using high-resolution magnetic resonance angiog-
uation. Invest Radiol. 2004;39:27–33. raphy. Circulation. 2002;105:328–333.
102. Maintz D, Seifarth H, Flohr T, et al. Improved coronary artery 123. Maron BJ, Bonow RO, Cannon RO 3rd, et al. Hypertrophic car-
stent visualization and in-stent stenosis detection using 16-slice diomyopathy. Interrelations of clinical manifestations, patho-
computed-tomography and dedicated image reconstruction physiology, and therapy (1). N Engl J Med. 1987;316:780–789.
technique. Invest Radiol. 2003;38:790–795. 124. Sardanelli F, Molinari G, Petillo A, et al. MRI in hypertrophic
103. Schuijf JD, Bax JJ, Jukema JW, et al. Feasibility of assessment of cardiomyopathy: a morphofunctional study. J Comput Assist
coronary stent patency using 16-slice computed tomography. Tomogr. 1993;17:862–872.
Am J Cardiol. 2004;94:427–430. 125. Spirito P, Bellone P, Harris KM, et al. Magnitude of left ventricu-
104. Hong C, Chrysant GS, Woodard PK, Bae KT. Coronary artery lar hypertrophy and risk of sudden death in hypertrophic car-
stent patency assessed with in-stent contrast enhancement diomyopathy. N Engl J Med. 2000;342:1778–1785.
measured at multi-detector row CT angiography: initial experi- 126. Elliott PM, Poloniecki J, Dickie S, et al. Sudden death in hyper-
ence. Radiology. 2004;233:286–291. trophic cardiomyopathy: identification of high risk patients. J
105. Gilard M, Cornily JC, Rioufol G, et al. Noninvasive assessment Am Coll Cardiol. 2000;36:2212–2218.
of left main coronary stent patency with 16-slice computed to- 127. Maron BJ, McKenna WJ, Danielson GK, et al. American College
mography. Am J Cardiol. 2005;95:110–112. of Cardiology/European Society of Cardiology clinical expert
106. Maintz D, Botnar RM, Fischbach R, et al. Coronary magnetic consensus document on hypertrophic cardiomyopathy. A report
resonance angiography for assessment of the stent lumen: a of the American College of Cardiology Foundation Task Force
phantom study. J Cardiovasc Magn Reson. 2002;4:359–367. on Clinical Expert Consensus Documents and the European
Chapter 1: Heart 81
Society of Cardiology Committee for Practice Guidelines. J Am gadolinium-DTPA-enhanced magnetic resonance imaging. Am J
Coll Cardiol. 2003;42:1687–1713. Med. 2001;110:520–527.
128. Abelmann WH, Lorell BH. The challenge of cardiomyopathy. J 151. Tadamura E, Yamamuro M, Kubo S, et al. Effectiveness of de-
Am Coll Cardiol. 1989;13:1219–1239. layed enhanced MRI for identification of cardiac sarcoidosis:
129. Rickers C, Wilke NM, Jerosch-Herold M, et al. Utility of cardiac comparison with radionuclide imaging. AJR Am J Roentgenol.
magnetic resonance imaging in the diagnosis of hypertrophic 2005;185:110–115.
cardiomyopathy. Circulation. 2005;112:855–861. 152. Lachmann HJ, Booth DR, Booth SE, et al. Misdiagnosis of hered-
130. Stone DL, Petch MC, Verney GI, Dixon AK. Computed tomogra- itary amyloidosis as AL (primary) amyloidosis. N Engl J Med.
phy in patients with hypertrophic cardiomyopathy. Br Heart J. 2002;346:1786–1791.
1984;52:136–139. 153. Kwong RY, Falk RH. Cardiovascular magnetic resonance in car-
131. Park JH, Kim YM. MR imaging of cardiomyopathy. Magn Reson diac amyloidosis. Circulation 2005;111:122–124.
Imaging Clin North Am. 1996;4:269–286. 154. Dubrey SW, Cha K, Skinner M, et al. Familial and primary (AL)
132. Young AA, Kramer CM, Ferrari VA, et al. Three-dimensional left cardiac amyloidosis: echocardiographically similar diseases with
ventricular deformation in hypertrophic cardiomyopathy. distinctly different clinical outcomes. Heart. 1997; 78: 74–82.
Circulation. 1994;90:854–867. 155. Rahman JE, Helou EF, Gelzer-Bell R, et al. Noninvasive diagno-
133. Kramer CM, Reichek N, Ferrari VA, et al. Regional heterogeneity sis of biopsy-proven cardiac amyloidosis. J Am Coll Cardiol.
of function in hypertrophic cardiomyopathy. Circulation. 1994; 2004;43:410–415.
90:186–194. 156. Maceira AM, Joshi J, Prasad SK, et al. Cardiovascular magnetic res-
134. Dong SJ, MacGregor JH, Crawley AP, et al. Left ventricular wall onance in cardiac amyloidosis. Circulation. 2005;111: 186–193.
thickness and regional systolic function in patients with hyper- 157. Liu CP, Ting CT, Lawrence W, et al. Diminished contractile re-
trophic cardiomyopathy. A three-dimensional tagged magnetic sponse to increased heart rate in intact human left ventricular
resonance imaging study. Circulation. 1994;90:1200–1209. hypertrophy. Systolic versus diastolic determinants. Circulation.
135. Koito H, Suzuki J, Nakamori H, et al. [Clinical significance of 1993;88:1893–1906.
abnormal high signal intensity of left ventricular myocardium 158. Chan PC, Liu P, Cronin C, et al. The use of nuclear magnetic res-
by gadolinium-diethylenetriamine penta-acetic acid enhanced onance imaging in monitoring total body iron in hemodialysis
magnetic resonance imaging in hypertrophic cardiomyopathy]. patients with hemosiderosis treated with erythropoietin and
J Cardiol. 1995;25:163–170. phlebotomy. Am J Kidney Dis. 1992;19:484–489.
136. Teraoka K, Hirano M, Ookubo H, et al. Delayed contrast en- 159. Liu P, Henkelman M, Joshi J, et al. Quantification of cardiac and
hancement of MRI in hypertrophic cardiomyopathy. Magn tissue iron by nuclear magnetic resonance relaxometry in a
Reson Imaging. 2004;22:155–161. novel murine thalassemia-cardiac iron overload model. Can J
137. Matsunaka T, Hamada M, Matsumoto Y, Higaki J. First-pass my- Cardiol. 1996;12:155–164.
ocardial perfusion defect and delayed contrast enhancement in 160. Westwood M, Anderson LJ, Firmin DN, et al. A single breath-
hypertrophic cardiomyopathy assessed with MRI. Magn Reson hold multiecho T2* cardiovascular magnetic resonance tech-
Med Sci. 2003;2:61–69. nique for diagnosis of myocardial iron overload. J Magn Reson
138. Dec GW, Fuster V. Idiopathic dilated cardiomyopathy. N Engl J Imaging. 2003;18:33–39.
Med. 1994;331:1564–1575. 160a.Alexopoulou E, Stripeli F, Baras P, et al. R2 relaxometry with
139. Doherty NE 3rd, Fujita N, Caputo GR, Higgins CB. Measurement MRI for the quantification of tissue iron overload in beta-
of right ventricular mass in normal and dilated cardiomyopathic thalassemic patients. J Magn Reson Imaging. 2006; 23:163–170.
ventricles using cine magnetic resonance imaging. Am J Cardiol. 161. Corssmit EP, Trip MD, Durrer JD. Loffler’s endomyocarditis in
1992;69:1223–1228. the idiopathic hypereosinophilic syndrome. Cardiology. 1999;91:
140. Imai H, Kumai T, Sekiya M, et al. [Left ventricular trabeculae 272–276.
evaluated with MRI in dilated cardiomyopathy and old myocar- 162. Plastiras SC, Economopoulos N, Kelekis NL, Tzelepis GE.
dial infarction]. J Cardiol. 1992;22:83–90. Magnetic resonance imaging of the heart in a patient with hy-
141. Koito H, Suzuki J, Ohkubo N, et al. [Gadolinium-diethylenetri- pereosinophilic syndrome. Am J Med. 2006;119:130–132.
amine pentaacetic acid enhanced magnetic resonance imaging 163. Andy JJ. Aetiology of endomyocardial fibrosis (EMF). West Afr J
of dilated cardiomyopathy: clinical significance of abnormally Med. 2001;20:199–207.
high signal intensity of left ventricular myocardium]. J Cardiol. 164. Mousseaux E, Hernigou A, Azencot M, et al. Endomyocardial fi-
1996;28:41–49. brosis: electron-beam CT features. Radiology. 1996;198: 755–760.
142. Zimmermann O, Grebe O, Merkle N, et al. Myocardial biopsy 165. Sarazin M, Caumes E, Cohen A, Amarenco P. Multiple mi-
findings and gadolinium enhanced cardiovascular magnetic reso- croembolic borderzone brain infarctions and endomyocardial
nance in dilated cardiomyopathy. Eur J Heart Fail. 2006;8:162–166. fibrosis in idiopathic hypereosinophilic syndrome and in
143. Valantine H, McKenna WJ, Nihoyannopoulos P, et al. Schistosoma mansoni infestation. J Neurol Neurosurg Psychiatry.
Sarcoidosis: a pattern of clinical and morphological presenta- 2004; 75:305–307.
tion. Br Heart J. 1987;57:256–263. 166. Cury RC, Abbara S, Sandoval LJ, et al. Images in cardiovascular
144. Silverman KJ, Hutchins GM, Bulkley BH. Cardiac sarcoid: a clin- medicine. Visualization of endomyocardial fibrosis by delayed-
icopathologic study of 84 unselected patients with systemic sar- enhancement magnetic resonance imaging. Circulation.
coidosis. Circulation. 1978;58:1204–1211. 2005;111: e115–117.
145. Abeler V. Sarcoidosis of the cardiac conducting system. Am 167. Blake LM, Scheinman MM, Higgins CB. MR features of arrhyth-
Heart J. 1979;97:701–707. mogenic right ventricular dysplasia. AJR Am J Roentgenol. 1994;
146. Sharma OP. Cardiac and neurologic dysfunction in sarcoidosis. 162:809–812.
Clin Chest Med. 1997;18:813–825. 168. Fontaliran F, Arkwright S, Vilde F, Fontaine G. [Arrhythmogenic
147. Doherty MJ, Kumar SK, Nicholson AA, McGivern DV. Cardiac right ventricular dysplasia and cardiomyopathy. Clinical and
sarcoidosis: the value of magnetic resonance imaging in diagno- anatomic-pathologic aspects, nosologic approach]. Arch Anat
sis and assessment of response to treatment. Respir Med. Cytol Pathol. 1998;46:171–177.
1998;92:697–699. 169. McKenna WJ, Thiene G, Nava A, et al. Diagnosis of arrhythmo-
148. Inoue S, Shimada T, Murakami Y. Clinical significance of gadolin- genic right ventricular dysplasia/cardiomyopathy. Task Force of
ium-DTPA-enhanced MRI for detection of myocardial lesions in a the Working Group Myocardial and Pericardial Disease of the
patient with sarcoidosis. Clin Radiol. 1999;54: 70–72. European Society of Cardiology and of the Scientific Council on
149. Schulz-Menger J, Strohm O, Dietz R, Friedrich MG. Visualization Cardiomyopathies of the International Society and Federation
of cardiac involvement in patients with systemic sarcoidosis of Cardiology. Br Heart J 1994; 71:215–218.
applying contrast-enhanced magnetic resonance imaging. 170. Tandri H, Bomma C, Calkins H, Bluemke DA. Magnetic reso-
Magma. 2000;11:82–83. nance and computed tomography imaging of arrhythmogenic
150. Shimada T, Shimada K, Sakane T, et al. Diagnosis of cardiac sar- right ventricular dysplasia. J Magn Reson Imaging. 2004;19:
coidosis and evaluation of the effects of steroid therapy by 848–858.
171. White RD, Trohman RG, Flamm SD, et al. Right ventricular 196. Watanabe M, Takazawa K, Wada A, et al. Cardiac myxoma with
arrhythmia in the absence of arrhythmogenic dysplasia: MR ima- Gamna-Gandy bodies: case report with MR imaging. J Thorac
ging of myocardial abnormalities. Radiology. 1998;207:743–751. Imaging. 1994;9:185–187.
172. Laissy JP, Hyafil F, Feldman LJ, et al. Differentiating acute 197. Restrepo CS, Largoza A, Lemos DF, et al. CT and MR imaging
myocardial infarction from myocarditis: diagnostic value of findings of benign cardiac tumors. Curr Probl Diagn Radiol.
early- and delayed-perfusion cardiac MR imaging. Radiology. 2005;34:12–21.
2005;237:75–82. 198. Schvartzman PR, White RD. Imaging of cardiac and paracardiac
173. Magnani JW, Dec GW. Myocarditis: current trends in diagnosis masses. J Thorac Imaging. 2000;15:265–273.
and treatment. Circulation. 2006;113:876–890. 199. Araoz PA, Mulvagh SL, Tazelaar HD, et al. CT and MR imaging
174. Laissy JP, Messin B, Varenne O, et al. MRI of acute myocarditis: a of benign primary cardiac neoplasms with echocardiographic
comprehensive approach based on various imaging sequences. correlation. Radiographics. 2000;20:1303–1319.
Chest. 2002;122:1638–1648. 200. Alkadhi H, Leschka S, Hurlimann D, et al. Fibroelastoma of the
175. Friedrich MG, Strohm O, Schulz-Menger J, et al. Contrast aortic valve. Evaluation with echocardiography and 64-slice CT.
media-enhanced magnetic resonance imaging visualizes my- Herz. 2005;30:438.
ocardial changes in the course of viral myocarditis. Circulation. 201. King SJ, Smallhorn JF, Burrows PE. Epicardial lipoma: imaging
1998;97:1802–1809. findings. AJR Am J Roentgenol. 1993;160:261–262.
176. Mahrholdt H, Goedecke C, Wagner A, et al. Cardiovascular mag- 202. Grande AM, Minzioni G, Pederzolli C, et al. Cardiac lipomas.
netic resonance assessment of human myocarditis: a compari- Description of 3 cases. J Cardiovasc Surg (Torino). 1998;39:
son to histology and molecular pathology. Circulation. 813–815.
2004;109:1250–1258. 203. Conces DJ, Jr., Vix VA, Tarver RD. Diagnosis of a myocardial
177. Srichai MB, Axel L. Magnetic resonance imaging in the manage- lipoma by using CT. AJR Am J Roentgenol. 1989;153:725–726.
ment of pericardial disease. Curr Treat Options Cardiovasc Med. 204. Kamiya H, Ohno M, Iwata H, et al. Cardiac lipoma in the inter-
2005;7:449–457. ventricular septum: evaluation by computed tomography and
178. Tomoda H, Hoshiai M, Furuya H, et al. Evaluation of pericardial magnetic resonance imaging. Am Heart J. 1990;119:1215–1217.
effusion with computed tomography. Am Heart J. 1980;99: 205. Hoffmann U, Globits S, Frank H. Cardiac and paracardiac
701–706. masses. Current opinion on diagnostic evaluation by magnetic
179. Mulvagh SL, Rokey R, Vick GW 3rd, Johnston DL. Usefulness of resonance imaging. Eur Heart J. 1998;19:553–563.
nuclear magnetic resonance imaging for evaluation of pericar- 206. Broderick LS, Conces DJ Jr, Tarver RD. CT evaluation of normal
dial effusions, and comparison with two-dimensional echocar- interatrial fat thickness. J Comput Assist Tomogr. 1996;20:
diography. Am J Cardiol. 1989;64:1002–1009. 950–953.
180. Wang ZJ, Reddy GP, Gotway MB, et al. CT and MR imaging of 207. Burke A, Virmani R. Atlas of Tumor Pathology, 3rd Series, Fascicle
pericardial disease. Radiographics. 2003;23 (spec no): S167–180. 16. Washington, DC: Armed Forces Institute of Pathology;1996.
181. Talreja DR, Edwards WD, Danielson GK, et al. Constrictive peri- 208. Cina SJ, Smialek JE, Burke AP, et al. Primary cardiac tumors
carditis in 26 patients with histologically normal pericardial causing sudden death: a review of the literature. Am J Forensic
thickness. Circulation. 2003;108:1852–1857. Med Pathol. 1996;17:271–281.
182. White R, Hardy P, VanDyke C. Diastolic dysfunction: dynamic 209. Abushaban L, Denham B, Duff D. 10 year review of cardiac tu-
MRI velocity-mapping of related flow patterns in the superior mours in childhood. Br Heart J. 1993; 70:166-169.
vena cava. J Magn Reson Imaging. 1993;3 (P):65. 210. Yamaguchi M, Hosokawa Y, Ohashi H, et al. Cardiac fibroma.
183. Abraham KP, Reddy V, Gattuso P. Neoplasms metastatic to the Long-term fate after excision. J Thorac Cardiovasc Surg. 1992;
heart: review of 3314 consecutive autopsies. Am J Cardiovasc 103:140–145.
Pathol. 1990;3:195–198. 211. Beghetti M, Gow RM, Haney I, et al. Pediatric primary benign
184. Restrepo CS, Largoza A, Lemos DF, et al. CT and MR imaging cardiac tumors: a 15-year review. Am Heart J. 1997;134:
findings of malignant cardiac tumors. Curr Probl Diagn Radiol. 1107–1114.
2005;34:1–11. 212. Brown JJ, Barakos JA, Higgins CB. Magnetic resonance imaging
185. Mousseaux E, Meunier P, Azancott S, et al. Cardiac metastatic of cardiac and paracardiac masses. J Thorac Imaging. 1989;
melanoma investigated by magnetic resonance imaging. Magn 4:58–64.
Reson Imaging. 1998;16:91–95. 213. Valente M, Cocco P, Thiene G, et al. Cardiac fibroma and heart
186. Funari M, Fujita N, Peck WW, Higgins CB. Cardiac tumors: as- transplantation. J Thorac Cardiovasc Surg. 1993;106:1208–1212.
sessment with Gd-DTPA enhanced MR imaging. J Comput Assist 214. Muhler EG, Turniski-Harder V, Engelhardt W, von Bernuth G.
Tomogr. 1991;15:953–958. Cardiac involvement in tuberous sclerosis. Br Heart J. 1994;
187. Link KM, Lesko NM. MR evaluation of cardiac/juxtacardiac 72:584–590.
masses. Top Magn Reson Imaging. 1995; 7:232–245. 215. Fujita N, Caputo GR, Higgins CB. Diagnosis and characteriza-
188. Reynen K. Cardiac myxomas. N Engl J Med. 1995;333: tion of intracardiac masses by magnetic resonance imaging. Am
1610–1617. J Card Imaging. 1994;8:69–80.
189. Shimono T, Makino S, Kanamori Y, et al. Left atrial myxomas. 216. Seguin JR, Coulon P, Huret C, et al. Intrapericardial teratoma in
Using gross anatomic tumor types to determine clinical infancy: a rare disease. J Cardiovasc Surg (Torino). 1986;27:
features and coronary angiographic findings. Chest. 1995;107: 509–511.
674–679. 217. Beghetti M, Prieditis M, Rebeyka IM, Mawson J. Images in car-
190. Symbas PN, Hatcher CR Jr, Gravanis MB. Myxoma of the heart: diovascular medicine. Intrapericardial teratoma. Circulation.
clinical and experimental observations. Ann Surg. 1976;183: 1998;97:1523–1524.
470–475. 218. Tollens T, Casselman F, Devlieger H, et al. Fetal cardiac tampon-
191. Seemann MD, Brenner P. Demonstration of spiral CT scans and ade due to an intrapericardial teratoma. Ann Thorac Surg.
reconstructions of a right atrial myxoma with bilateral pul- 1998;66:559–560.
monary tumor emboli and a coincidental benign mediastinal 219. Lo LJ, Nucho RC, Allen JW, et al. Left atrial cardiac hemangioma
thymoma. Eur J Med Res. 1996;1:515–519. associated with shortness of breath and palpitations. Ann Thorac
192. Tsuchiya F, Kohno A, Saitoh R, Shigeta A. CT findings of atrial Surg. 2002; 73:979–981.
myxoma. Radiology. 1984;151:139–143. 220. Kocak H, Ozyazicioglu A, Gundogdu C, Sevimli S. Cardiac
193. Shapiro LM. Cardiac tumours: diagnosis and management. hemangioma complicated with cerebral and coronary emboli-
Heart. 2001;85:218–222. zation. Heart Vessels. 2005;20:296–297.
194. Masui T, Takahashi M, Miura K, et al. Cardiac myxoma: identifi- 221. Oshima H, Hara M, Kono T, et al. Cardiac hemangioma of the
cation of intratumoral hemorrhage and calcification on MR left atrial appendage: CT and MR findings. J Thorac Imaging.
images. AJR Am J Roentgenol. 1995;164:850–852. 2003;18:204–206.
195. Grebenc ML, Rosado de Christenson ML, Burke AP, et al. 222. Hoffmann U, Globits S, Schima W, et al. Usefulness of magnetic
Primary cardiac and pericardial neoplasms: radiologic- resonance imaging of cardiac and paracardiac masses. Am J
pathologic correlation. Radiographics. 2000;20:1073–1103. Cardiol. 2003;92:890–895.
Chapter 1: Heart 83
223. Klatt EC, Heitz DR. Cardiac metastases. Cancer. 1990;65: 252. Durand E, Vanel D, Mousseaux E, et al. A recurrent left atrium
1456–1459. leiomyosarcoma. Eur Radiol. 1998;8:97–99.
224. Lam KY, Dickens P, Chan AC. Tumors of the heart. A 20-year 253. Paraf F, Bruneval P, Balaton A, et al. Primary liposarcoma of the
experience with a review of 12,485 consecutive autopsies. Arch heart. Am J Cardiovasc Pathol. 1990;3:175–180.
Pathol Lab Med. 1993;117:1027–1031. 254. Murtra M, Mestres CA, Igual A, et al. Primary liposarcoma of the
225. Schoen FJ, Berger BM, Guerina NG. Cardiac effects of noncar- right ventricle and pulmonary artery: surgical excision and
diac neoplasms. Cardiol Clin. 1984;2:657–670. replacement of the pulmonic valve by a Bjork-Shiley tilting disc
226. Roberts WC. Primary and secondary neoplasms of the heart. Am valve. Thorac Cardiovasc Surg. 1983;31:172–174.
J Cardiol. 1997;80:671–682. 255. Roberts WC, Glancy DL, DeVita VT Jr. Heart in malignant lym-
227. Barakos JA, Brown JJ, Higgins CB. MR imaging of secondary car- phoma (Hodgkin’s disease, lymphosarcoma, reticulum cell sar-
diac and paracardiac lesions. AJR Am J Roentgenol. 1989;153: coma and mycosis fungoides). A study of 196 autopsy cases. Am
47–50. J Cardiol. 1968;22:85–107.
228. Janigan DT, Husain A, Robinson NA. Cardiac angiosarcomas. A 256. Ceresoli GL, Ferreri AJ, Bucci E, et al. Primary cardiac lym-
review and a case report. Cancer. 1986;57:852–859. phoma in immunocompetent patients: diagnostic and thera-
229. Rettmar K, Stierle U, Sheikhzadeh A, Diederich KW. Primary peutic management. Cancer. 1997;80:1497–1506.
angiosarcoma of the heart. Report of a case and review of the 257. Versluis PJ, Lamers RJ, van Belle AF. Primary malignant lym-
literature. Jpn Heart J. 1993;34:667–683. phoma of the heart: CT and MRI features. Rofo. 1995;162:
230. Araoz PA, Eklund HE, Welch TJ, Breen JF. CT and MR imaging of 533–534.
primary cardiac malignancies. Radiographics. 1999;19:1421–1434. 258. Tada H, Asazuma K, Ohya E, et al. Images in cardiovascular med-
231. Kim EE, Wallace S, Abello R, et al. Malignant cardiac fibrous his- icine. Primary cardiac B-cell lymphoma. Circulation. 1998;97:
tiocytomas and angiosarcomas: MR features. J Comput Assist 220–221.
Tomogr. 1989;13:627–632. 259. Dorsay TA, Ho VB, Rovira MJ, et al. Primary cardiac lymphoma:
232. Yahata S, Endo T, Honma H, et al. Sunray appearance on enhan- CT and MR findings. J Comput Assist Tomogr. 1993;17:978–981.
ced magnetic resonance image of cardiac angiosarcoma with 259a.Freedberg RS, Kronzon I, Rumancik WM, Liebeskind D. The
pericardial obliteration. Am Heart J. 1994;127:468–471. contribution of magnetic resonance imaging to the evaluation of
233. Dawson WB, Mayo JR, Muller NL. Computed tomography intracardiac tumors diagnosed by echocardiography. Circulation.
of cardiac and pericardial tumors. Can Assoc Radiol J. 1990;41: 1988;77:96–103.
270–275. 260. Wiegers S, Herrmann H, Plappert T. Valvular heart disease. In: St
234. Mader MT, Poulton TB, White RD. Malignant tumors of the John Sutton M, Rutherford J, eds. Clinical cardiovascular imaging:
heart and great vessels: MR imaging appearance. Radiographics. a companion to Braunwald’s heart disease. Philadelphia: Elsevier
1997;17:145–153. Saunders; 2004:280–338.
235. Hwa J, Ward C, Nunn G, et al. Primary intraventricular cardiac 261. Shanewise JS, Cheung AT, Aronson S, et al. ASE/SCA guidelines
tumors in children: contemporary diagnostic and management for performing a comprehensive intraoperative multiplane
options. Pediatr Cardiol. 1994;15:233–237. transesophageal echocardiography examination: recommenda-
236. Jack CM, Cleland J, Geddes JS. Left atrial rhabdomyosarcoma tions of the American Society of Echocardiography Council
and the use of digital gated computed tomography in its diag- for Intraoperative Echocardiography and the Society of
nosis. Br Heart J. 1986;55:305–307. Cardiovascular Anesthesiologists Task Force for Certification in
237. Watanabe AT, Teitelbaum GP, Henderson RW, Bradley WG Jr. Perioperative Transesophageal Echocardiography. J Am Soc
Magnetic resonance imaging of cardiac sarcomas. J Thorac Echocardiogr. 1999;12:884–900.
Imaging. 1989;4:90–92. 262. Baumert B, Plass A, Bettex D, et al. Dynamic cine mode imaging
238. Siripornpitak S, Higgins CB. MRI of primary malignant cardio- of the normal aortic valve using 16-channel multidetector row
vascular tumors. J Comput Assist Tomogr. 1997;21:462–466. computed tomography. Invest Radiol. 2005;40:637–647.
239. Gilkeson RC, Chiles C. MR evaluation of cardiac and pericardial 263. Debl K, Djavidani B, Seitz J, et al. Planimetry of aortic valve area
malignancy. Magn Reson Imaging Clin North Am. 2003;11: in aortic stenosis by magnetic resonance imaging. Invest Radiol.
173–186, viii. 2005;40:631–636.
240. McAllister HA Jr, Hall RJ, Cooley DA. Tumors of the heart and 264. John AS, Dill T, Brandt RR, et al. Magnetic resonance to assess the
pericardium. Curr Probl Cardiol. 1999;24:57–116. aortic valve area in aortic stenosis: how does it compare to current
241. Itoh A, Okubo S, Nakanishi N, et al. Recurrent epicardial diagnostic standards? J Am Coll Cardiol. 2003;42:519–526.
fibrosarcoma which arose 12 years after the first resection. Eur 265. Rozenshtein A, Boxt LM. Computed tomography and magnetic
Heart J. 1991;12:270–272. resonance imaging of patients with valvular heart disease. J
242. Kim DG, Lee SY, Chung SK, et al. Brain metastasis from myxofi- Thorac Imaging. 2000;15:252–264.
brosarcoma of the heart. Acta Neurochir (Wien). 1997;139: 266. Perloff JK, Roberts WC. The mitral apparatus. Functional
88–89. anatomy of mitral regurgitation. Circulation. 1972;46:227–239.
243. Shih WJ, McCullough S, Smith M. Diagnostic imagings for pri- 267. Schwitter J. Valvular heart disease: assessment of valve morphol-
mary cardiac fibrosarcoma. Int J Cardiol. 1993;39:157–161. ogy and quantification using MR. Herz. 2000;25:342–355.
244. Burke AP, Virmani R. Osteosarcomas of the heart. Am J Surg 268. Alkadhi H, Bettex D, Wildermuth S, et al. Dynamic cine imag-
Pathol. 1991;15:289–295. ing of the mitral valve with 16-MDCT: a feasibility study. AJR
245. Chaloupka JC, Fishman EK, Siegelman SS. Use of CT in the eval- Am J Roentgenol. 2005;185:636–646.
uation of primary cardiac tumors. Cardiovasc Intervent Radiol. 269. Alkadhi H, Wildermuth S, Bettex DA, et al. Mitral regurgitation:
1986;9:132–135. quantification with 16-detector row CT—initial experience.
246. Lurito KJ, Martin T, Cordes T. Right atrial primary cardiac Radiology. 2006;238:454–463.
osteosarcoma. Pediatr Cardiol. 2002;23:462–465. 270. Roberts WC, Perloff JK. Mitral valvular disease. A clinicopatho-
247. Yashar J, Witoszka M, Savage DD, et al. Primary osteogenic logic survey of the conditions causing the mitral valve to func-
sarcoma of the heart. Ann Thorac Surg. 1979;28:594–600. tion abnormally. Ann Intern Med. 1972; 77:939–975.
248. Schneiderman H, Fordham EW, Goren CC, et al. Primary 271. Connolly HM, Crary JL, McGoon MD, et al. Valvular heart dis-
cardiac osteosarcoma: multidisciplinary aspects applicable to ease associated with fenfluramine-phentermine. N Engl J Med.
extraskeletal osteosarcoma generally. CA Cancer J Clin. 1984; 1997;337:581–588.
34:110–117. 272. Zile MR, Gaasch WH, Carroll JD, Levine HJ. Chronic mitral
249. Reynard JS Jr, Gregoratos G, Gordon MJ, Bloor CM. Primary regurgitation: predictive value of preoperative echocardio-
osteosarcoma of the heart. Am Heart J. 1985;109:598–600. graphic indexes of left ventricular function and wall stress. J Am
250. Takamizawa S, Sugimoto K, Tanaka H, et al. A case of primary Coll Cardiol. 1984;3:235–242.
leiomyosarcoma of the heart. Intern Med. 1992;31:265–268. 273. Cheitlin M, MacGregor J. Acquired tricuspid and pulmonary
251. Lo FL, Chou YH, Tiu CM, et al. Primary cardiac leiomyosar- valve disease. In: Topol E, ed. Textbook of cardiovascular medicine,
coma: imaging with 2-D echocardiography, electron beam CT 2nd ed. Philadelphia: Lippincott-Raven Publishers; 2002:
and 1.5-Tesla MR. Eur J Radiol. 1998;27:72–76. 529–548.
274. Wexler L, Higgins CB. The use of magnetic resonance imaging in 301. Choi YH, Park JH, Choe YH, Yoo SJ. MR imaging of Ebstein’s
adult congenital heart disease. Am J Card Imaging. 1995;9: 15–28. anomaly of the tricuspid valve. AJR Am J Roentgenol. 1994;
275. Rebergen SA, Chin JG, Ottenkamp J, et al. Pulmonary regurgita- 163:539–543.
tion in the late postoperative follow-up of tetralogy of Fallot. 302. Garrett JS, Schiller NB, Botvinick EH, et al. Cine-computed to-
Volumetric quantitation by nuclear magnetic resonance velocity mography of Ebstein anomaly. J Comput Assist Tomogr. 1986;
mapping. Circulation. 1993;88:2257–2266. 10:664–666.
276. Bouzas B, Kilner PJ, Gatzoulis MA. Pulmonary regurgitation: 303. Lee WJ, Chen SJ, Wu MH, Li YW. Regression of ductus arteriosus
not a benign lesion. Eur Heart J. 2005;26:433–439. aneurysm in a neonate demonstrated by three-dimensional
277. Helbing WA, de Roos A. Optimal imaging in assessment of right computed tomography. Int J Cardiol. 1999;68:231–234.
ventricular function in tetralogy of Fallot with pulmonary regur- 304. Becker C, Soppa C, Fink U, et al. Spiral CT angiography and 3D
gitation. Am J Cardiol. 1998;82:1561–1562. reconstruction in patients with aortic coarctation. Eur Radiol.
278. Mirowitz SA, Gutierrez FR. MR and CT diagnosis of carcinoid 1997;7:1473–1477.
heart disease. Chest. 1993;103:630–631. 305. Taneja K, Kawlra S, Sharma S, Rajani M. Pseudocoarctation of
279. Mollet NR, Dymarkowski S, Bogaert J. MRI and CT revealing the aorta: complementary findings on plain film radiography, CT,
carcinoid heart disease. Eur Radiol. 2003;13 (suppl 6): L14–18. DSA, and MRA. Cardiovasc Intervent Radiol. 1998;21:439–441.
280. Shellock FG. Prosthetic heart valves and annuloplasty rings: as- 306. Julsrud PR, Breen JF, Felmlee JP, et al. Coarctation of the aorta:
sessment of magnetic field interactions, heating, and artifacts at collateral flow assessment with phase-contrast MR angiography.
1.5 Tesla. J Cardiovasc Magn Reson. 2001;3:317–324. AJR Am J Roentgenol. 1997;169:1735–1742.
281. Edwards MB, Ordidge RJ, Hand JW, et al. Assessment of mag- 307. Engvall J, Sjoqvist L, Nylander E, et al. Biplane transoesophageal
netic field (4.7 T) induced forces on prosthetic heart valves and echocardiography, transthoracic Doppler, and magnetic reso-
annuloplasty rings. J Magn Reson Imaging. 2005;22:311–317. nance imaging in the assessment of coarctation of the aorta. Eur
282. Warnes CA, Liberthson R, Danielson GK, et al. Task force 1: the Heart J. 1995;16:1399–1409.
changing profile of congenital heart disease in adult life. J Am 308. Piechaud JF. Stent implantation for coarctation in adults. J Interv
Coll Cardiol. 2001;37:1170–1175. Cardiol. 2003;16:413–418.
283. Moller JH. Alexander S. Nadas Lecture. Fifty years of pediatric 309. Sousa Uva M, Chardigny C, Galetti L, et al. Surgery for tetralogy
cardiology and challenges for the future. Circulation. 1994;89: of Fallot at less than six months of age. Is palliation “old-fash-
2479–2483. ioned”? Eur J Cardiothorac Surg. 1995;9:453–459; discussion
284. Funabashi N, Asano M, Sekine T, et al. Direction, location, and 459–460.
size of shunt flow in congenital heart disease evaluated by 310. Bove EL, Byrum CJ, Thomas FD, et al. The influence of pul-
ECG-gated multislice computed tomography. Int J Cardiol. 2005 monary insufficiency on ventricular function following repair of
(online). tetralogy of Fallot. Evaluation using radionuclide ventriculogra-
285. Roberts WC. The congenitally bicuspid aortic valve. A study of phy. J Thorac Cardiovasc Surg. 1983;85:691–696.
85 autopsy cases. Am J Cardiol. 1970;26:72–83. 311. Kavey RE, Thomas FD, Byrum CJ, et al. Ventricular arrhythmias
286. Steiner RM, Reddy GP, Flicker S. Congenital cardiovascular dis- and biventricular dysfunction after repair of tetralogy of Fallot. J
ease in the adult patient: imaging update. J Thorac Imaging. Am Coll Cardiol. 1984;4:126–131.
2002;17:1–17. 312. Waien SA, Liu PP, Ross BL, et al. Serial follow-up of adults with
287. Ward C. Clinical significance of the bicuspid aortic valve. Heart. repaired tetralogy of Fallot. J Am Coll Cardiol. 1992;20:295–300.
2000;83:81–85. 313. d’Udekem d’Acoz Y, Pasquet A, Lebreux L, et al. Does right ven-
288. Braverman AC. Bicuspid aortic valve and associated aortic wall tricular outflow tract damage play a role in the genesis of late
abnormalities. Curr Opin Cardiol. 1996;11:501–503. right ventricular dilatation after tetralogy of Fallot repair? Ann
289. Parai JL, Masters RG, Walley VM, et al. Aortic medial changes Thorac Surg. 2003;7:555–561; discussion 561.
associated with bicuspid aortic valve: myth or reality? Can J 314. Niezen RA, Helbing WA, van der Wall EE, et al. Biventricular sys-
Cardiol. 1999;15:1233–1238. tolic function and mass studied with MR imaging in children
290. Veinot JP. Congenitally bicuspid aortic valve and associated aor- with pulmonary regurgitation after repair for tetralogy of Fallot.
tic medical disease. Ann Thorac Surg. 2001;71:1067–1068. Radiology. 1996;201:135–140.
291. Baron MG. Plain film diagnosis of common cardiac anomalies 315. Helbing WA, Niezen RA, de Roos A. Congenital heart disease.
in the adult. Radiol Clin North Am. 1999;37:401–420, vi. In: Bogaert J, Duerinckx A, Rademakers F, eds. Magnetic reso-
292. Diethelm L, Dery R, Lipton MJ, Higgins CB. Atrial-level shunts: nance of the heart and great vessels. Berlin, Heidelberg: Springer-
sensitivity and specificity of MR in diagnosis. Radiology. Verlag;1999:95–108.
1987;162:181–186. 316. Raman SV, Cook SC, McCarthy B, Ferketich AK. Usefulness of
293. Higgins CB. Radiography of congenital heart disease. In: multidetector row computed tomography to quantify right ven-
Higgins CB, ed. Essentials of cardiac radiology and imaging. tricular size and function in adults with either tetralogy of Fallot
Philadelphia: Lippincott;1992:49–90. or transposition of the great arteries. Am J Cardiol. 2005;95:
294. Wang ZJ, Reddy GP, Gotway MB, et al. Cardiovascular shunts: MR 683–686.
imaging evaluation. Radiographics. 2003;23 (spec no): S181–S194. 317. Eichhorn J, Schoenberg S, Ulmer HE. Cardiac-gated multislice
295. Holmvang G, Palacios IF, Vlahakes GJ, et al. Imaging and sizing computerised tomographic angiography in the preoperative
of atrial septal defects by magnetic resonance. Circulation. evaluation of an infant with tetralogy of Fallot and pulmonary
1995;92:3473–3480. atresia. Cardiol Young. 2003;13:466–468.
296. Therrien J, Webb GD. Congenital heart disease in adults. In: 318. van Straten A, Vliegen HW, Hazekamp MG, et al. Right ventricu-
Braunwald E, ed. Heart disease: a textbook of cardiovascular medi- lar function after pulmonary valve replacement in patients with
cine, 6th ed. Philadelphia: WB Saunders;2001:1592–1621. tetralogy of Fallot. Radiology. 2004;233:824–829.
297. Anderson RH, Lenox CC, Zuberbuhler JR. The morphology of 319. Kang IS, Redington AN, Benson LN, et al. Differential regurgita-
ventricular septal defects. Perspect Pediatr Pathol. 1984;8:235–268. tion in branch pulmonary arteries after repair of tetralogy of
298. Didier D, Higgins CB. Identification and localization of ventric- Fallot: a phase-contrast cine magnetic resonance study.
ular septal defect by gated magnetic resonance imaging. Am J Circulation. 2003;107:2938–2943.
Cardiol. 1986;57:1363–1368. 320. Senning A. Surgical correction of transposition of the great ves-
299. Correa-Villasenor A, Ferencz C, Boughman JA, Neill CA. Total sels. Surgery. 1959;45:966–980.
anomalous pulmonary venous return: familial and environ- 321. Mustard WT. Successful two-stage correction of transposition of
mental factors. The Baltimore-Washington Infant Study Group. the great vessels. Surgery. 1964;55:469–472.
Teratology. 1991;44:415–428. 322. Jatene AD, Fontes VF, Paulista PP, et al. Anatomic correction of
300. Festa P, Ait-Ali L, Cerillo AG, et al. Magnetic resonance imaging transposition of the great vessels. J Thorac Cardiovasc Surg. 1976;
is the diagnostic tool of choice in the preoperative evaluation of 72:364–370.
patients with partial anomalous pulmonary venous return. Int J 323. Baron MG, Book WM. Congenital heart disease in the
Cardiovasc Imaging. 2006 (online). adult:2004. Radiol Clin North Am. 2004;42:675–690, vii.
Chapter 1: Heart 85
324. Tulevski, II, van der Wall EE, Groenink M, et al. Usefulness of 335. Yamanaka O, Hobbs RE. Coronary artery anomalies in 126,595
magnetic resonance imaging dobutamine stress in asympto- patients undergoing coronary arteriography. Cathet Cardiovasc
matic and minimally symptomatic patients with decreased car- Diagn. 1990;21:28–40.
diac reserve from congenital heart disease (complete and cor- 336. Ropers D, Moshage W, Daniel WG, et al. Visualization of
rected transposition of the great arteries and subpulmonic coronary artery anomalies and their anatomic course by
obstruction). Am J Cardiol. 2002;89:1077–1081. contrast-enhanced electron beam tomography and three-
325. Turley K, Verrier ED. Intermediate results from the period of the dimensional reconstruction. Am J Cardiol. 2001;87:193–197.
Congenital Heart Surgeons Transposition Study: 1985 to 1989. 337. Freedom RM, Mawson JB, Yoo SJ, Benson LN. Congenital heart
Congenital Heart Surgeons Society Database. Ann Thorac Surg. disease: textbook of angiocardiography. Armonk, NY: Futura;
1995;60:505–510. 1977.
326. Burkhart HM, Dearani JA, Williams WG, et al. Late results of 338. Goo HW, Park IS, Ko JK, et al. CT of congenital heart disease:
palliative atrial switch for transposition, ventricular septal de- normal anatomy and typical pathologic conditions. Radio-
fect, and pulmonary vascular obstructive disease. Ann Thorac graphics. 2003;23 (spec no): S147–165.
Surg. 2004; 77:464–468; discussion 468–469. 339. Nieman K, Cademartiri F, Lemos PA, et al. Reliable noninvasive
327. Losay J, Touchot A, Serraf A, et al. Late outcome after arterial coronary angiography with fast submillimeter multislice spiral
switch operation for transposition of the great arteries. computed tomography. Circulation. 2002;106:2051–2054.
Circulation. 2001;104: I121–I126. 340. Manghat NE, Morgan-Hughes GJ, Marshall AJ, Roobottom CA.
328. Freedom RM, Dyck JD. Congenitally corrected transposition of Multidetector row computed tomography: imaging congenital
the great arteries. In: Emmanouilides GC, Allen HD, coronary artery anomalies in adults. Heart. 2005;91:1515–1522.
Riemenschneider TA, P. GH, eds. Moss and Adams: Heart disease 341. White CS, Laskey WK, Stafford JL, NessAiver M. Coronary MRA:
in infants, children, and adolescents, 5th ed. Baltimore: Williams & use in assessing anomalies of coronary artery origin. J Comput
Wilkins;1995:1225–1245. Assist Tomogr. 1999;23:203–207.
329. Mullins CE. Ventricular inversion. In: Garson A Jr, Bricker JT, 342. Go AS, Hylek EM, Phillips KA, et al. Prevalence of diagnosed
McNamara DG, eds. The science and practice of pediatric cardiol- atrial fibrillation in adults: national implications for rhythm
ogy. Philadelphia: Lea & Febiger;1990:1233–1245. management and stroke prevention: the AnTicoagulation and
330. Guit GL, Bluemm R, Rohmer J, et al. Levotransposition of the Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA.
aorta: identification of segmental cardiac anatomy using MR 2001;285:2370–2375.
imaging. Radiology. 1986;161:673–679. 343. Lacomis JM, Wigginton W, Fuhrman C, et al. Multi-detector row
331. Park JH, Han MC, Kim CW. MR imaging of congenitally cor- CT of the left atrium and pulmonary veins before radio-
rected transposition of the great vessels in adults. AJR Am J frequency catheter ablation for atrial fibrillation. Radiographics.
Roentgenol. 1989;153:491–494. 2003;23 (spec no): S35–48; discussion S48–50.
332. Takasugi JE, Godwin JD, Chen JT. CT in congenitally-corrected 344. Ghaye B, Szapiro D, Dacher JN, et al. Percutaneous ablation for
transposition of the great vessels. Comput Radiol. 1987;11: atrial fibrillation: the role of cross-sectional imaging. Radio-
215–221. graphics. 2003;23 (spec no): S19–33; discussion S48–50.
333. Schwab JO, Ehlgen A, Sommer T. Congenitally corrected trans- 345. Pilleul F, Merchant N. MRI of the pulmonary veins: comparison
position of the great arteries in a 70-year-old woman diagnosed between 3D MR angiography and T1-weighted spin echo. J
using single-detector helical CT. AJR Am J Roentgenol. 2003; Comput Assist Tomogr. 2000;24:683–687.
181:598. 346. Hudsmith LE, Petersen SE, Francis JM, et al. Normal human left
334. Baltaxe HA, Wixson D. The incidence of congenital anomalies and right ventricular and left atrial dimensions using steady
of the coronary arteries in the adult population. Radiology. 1977; state free precession magnetic resonance imaging. J Cardiovasc
122:47–52. Magn Reson. 2005;7:775–782.
Aorta, Arch Vessels, 2

and Great Veins
IMAGING TECHNIQUES 88 Surgery of the Aortic Arch and Descending

Computed Tomography Angiography 88 Thoracic Aorta 176
Magnetic Resonance 89 Observations of the Stented Aorta 182
Techniques of Reconstruction 96
AORTITIS 185
NORMAL ANATOMY 99 Clinical Features 185
The Aorta and Its Branches 99 Imaging Features 185
Infectious Aortitis or Aneurysm 186
CONGENITAL AORTIC DISEASE 104
Aortic Arch Anomalies 104 GREAT VEINS, INCLUDING THE AZYGOS
Left Aortic Arch Anomalies 105 AND HEMIAZYGOS SYSTEMS 191
Right Aortic Arch Anomalies 106 Acquired Venous Abnormalities 199
Symptomatic Arch Anomalies 110 Magnetic Resonance Venography 204
Interrupted Aortic Arch 112
Coarctation 112
Pulmonary Sequestration 117 Over the past decade, rapid changes in imaging technology
have revolutionized our ability to study the great vessels
ACQUIRED AORTIC DISEASE 119 of the thorax. Parallel developments in multidetector
Thoracic Aortic Aneurysms 119 computed tomography (MDCT), magnetic resonance (MR),
and transesophageal echocardiography (TEE) now present
ACUTE AORTIC SYNDROMES 131 referring clinicians with a wide range of options for evaluat-
Ruptured Aortic Aneurysms (Contained and Overt) 132 ing diseases of the aorta and great vessels.
Aortic Dissection 135 All of these techniques provide excellent diagnostic
Penetrating Atherosclerotic Ulcer 148 accuracy for a wide range of thoracic aortic diseases (1,2).
Primary Aortic Intramural Hematoma 155 Because TEE is widely available, can be performed in the
intensive care unit or the operating room, and is extremely
AORTIC TRAUMA 163 versatile in the appropriate hands, it has emerged as a
Clinical Features 163 valuable tool for evaluating the thoracic aorta in the acute
Imaging Features 165 setting (1,3,4). In addition to being indispensable for
evaluating the aortic valve, the introduction of multiplane
ADVANCES IN SURGICAL AND INTERVENTIONAL probes has extended the utility of TEE in the diagnosis of
TECHNIQUES/THE POSTOPERATIVE AORTA 167 protruding aortic atheromas of the thoracic aorta, a newly
The Postoperative Aorta 168 recognized etiology for stroke and peripheral embolization.
Surgery for Aneurysm or Dissection Involving the Because many elderly patients with acquired thoracic
Ascending Aorta and Aortic Root 173 aortic disease have concomitant coronary artery disease
and 16- or 64-slice MDCT can simultaneously evaluate interval, and largest available pitch. In our experience this
both the thoracic aorta and the coronary arteries in a sin- is most usefully accomplished by obtaining 1- to 2.5-mm
gle breath-hold, it has become the standard examination sections with scans obtained in a cephalocaudad direction
for evaluation of patients with acute aortic disease. It is during a single breath-hold period using a pitch of 1 to 2.
easy to perform, relatively noninvasive, and less operator The introduction of subsecond (0.4 second) scanners
dependent than TEE. MDCT also offers a range of recon- makes acquisition of thinner sections throughout the entire
struction and display options. Most important, unlike thorax more practicable, increasing the scan volume
TEE, MDCT allows simultaneous evaluation of the entire while maintaining the same collimation and pitch. The
thorax, markedly extending the range of potential diag- routine use of ultra thin (1 mm) sections should be
noses, especially in patients with nonspecific symptoms limited to evaluation of the coronary arteries because of
only indirectly suggestive of cardiovascular disease. image processing time, storage requirements, and, most
Although rarely performed in the acute setting, MR and importantly, image review time. It may be anticipated that
MR angiography (MRA) have also emerged as versatile with further technologic improvements, routine studies
tools for the evaluation of both congenital and acquired may result in literally thousands of images; routine electro-
diseases of the thoracic aorta. This is a result of the rapid cardiographic (ECG)–gated aortic studies reconstructed at
technologic advances in both hardware and pulse multiple phases are already 1 to 2,000 images. Clearly,
sequences as well as the use of contrast agents that shorten selection of optimal scan parameters will continue to
the T1 relaxation rate of blood (5). This evolution is now necessitate compromise with the exigencies of image stor-
reflected in many institutions where MRA has replaced age, display, and interpretation. The benefit of the faster
diagnostic thoracic aortography in stable patients with scanners (64 detectors) is that the entire thorax can be
nontraumatic thoracic aortic disease. evaluated in as little as 5 seconds, whereby even the sickest
Given this wide range of potential choices, the purpose patients may be able to suspend respiration.
of this chapter is to focus on a review of disease entities
affecting the aorta and great vessels, with special emphasis
placed on the advantages and limitations of CT and MR in Contrast Administration
clinical practice. In this regard, recent advances in thera-
peutic techniques that have equally revolutionized our Iodine Concentration, Flow Rate, and Volume
approaches to these disorders are also discussed. Optimal contrast administration is critical to the successful
performance of CTA (10–18). Considerations include
volume, route and rate of injection, iodine concentration
IMAGING TECHNIQUES and osmolality, use of a saline flush, timing of scan acqui-
sition, and methods to diminish scan artifacts related to
contrast administration.
Computed Tomography Angiography
Although adequate vascular opacification can be
Optimization of image quality for multidetector CT obtained with as little as 50 mL of 300 mg I/mL contrast
angiography (CTA) requires careful attention to methods of media followed by a saline flush (10), most investigators
data acquisition (6–9). The ability to acquire data volumet- use a considerably higher volume of contrast for CT
rically has resulted in a wide variety of potential scan proto- angiographic applications. Typically, 75 to 100 mL of non-
cols for axial imaging. Variables that need to be selected ionic contrast material (averaging 300 mg I/mL) is injected
include collimation, pitch, breath-hold period, field of through an antecubital vein using an 18- to 20-gauge
view (FOV), reconstruction interval (or index), rate and catheter, at rates varying between 4 and 5 mL/second, with
volume of intravenous contrast administration, reconstruc- images obtained following a test bolus of contrast or an au-
tion algorithm, and radiation dose. tomated triggering approach. A right arm approach is pre-
To date, numerous articles have addressed the issue of ferred to minimize streak artifact across the aortic arch and
optimal data acquisition protocols (6–9). Unfortunately, proximal vessels and to eliminate the possibility of contrast
no single set of optimal scan parameters currently suffices obstruction as the left innominate vein crosses the sternum.
for all intrathoracic CTA applications. CTA techniques Nonionic contrast agents are used to minimize the effects of
often reflect a combination of personal or institutional nausea and vomiting as well as complications arising from
preferences coupled with individual manufacturer capabil- extravasations caused by the rapid infusion of ionic contrast
ities. Nonetheless, general guidelines can be derived from agents. Patients with renal insufficiency can be studied with
review of the literature. a reduced dose (50 mL) of a low-osmolar agent or with
gadolinium diethylenetriaminepenta-acetic acid (DTPA).
The use of injectable saline solution via a power injector
Scan Parameters
as a means to deliver contrast material is widely employed
In practice, it is necessary to compromise between use of in coronary CTA to clear the right ventricle of residual
the thinnest possible collimation, smallest reconstruction contrast, allowing better visualization of the right coronary
Chapter 2: Aorta, Arch Vessels, and Great Veins 89
artery. Before the advent of dual-head power injectors, aortic CT is unclear. Image quality, especially at the aortic
Hopper et al. (18) first loaded 75 mL of 60% nonionic root, is clearly superior to ungated aortic MDCT. Potential
contrast medium into a power injector in a vertical posi- disadvantages of routine use of ECG triggering include
tion, followed by 50 mL of normal saline solution loaded longer examination times (lead set-up), longer breath-
so as to lie on top of the denser nonionic contrast material. holds to cover the same anatomic region (except for the 64-
Meant to decrease the volume and cost of nonionic slice machine, in which it is negligible), two- to threefold
contrast material utilization, while using the saline to clear increase in radiation dose, and the reconstruction of at least
the syringe and intravenous catheter of contrast material a thousand more images. The benefits include simultane-
following the injection, this approach has the added bene- ous evaluation of the coronary arteries in patients with aor-
fit of significantly decreasing the number of streak artifacts tic dissection and aneurysms and the ability to evaluate
from adjacent venous structures. Newer power injectors synthetic valve function in patients who have undergone
contain multiple ports to provide an automated saline prosthetic valve replacement or composite graft placement.
flush following the bolus of iodinated contrast with freely A recent study compared ECG gated with nongated MDCT
selectable flow rates. in the emergency department setting in patients referred for
traumatic thoracic injury or acute aortic dissection (20).
The motion artifacts of the thoracic aorta and the supra-
Scan Delay aortic vessels were significantly reduced in the ECG-gated
In addition to considerations of flow rates and volume, data acquisition compared with the nongated technique,
it is also essential to optimize scan delay to ensure but image quality of the lung parenchyma, the spine, and
adequate opacification of targeted vessels. Although a the ribs was inferior in the ECG-gated data. This did not,
20- to 30-second scan delay most often is sufficient to however, compromise the detection rate of traumatic le-
ensure adequate opacification of the thoracic aorta, large sions and fractures, and overall examination time did not
variations between patients will be encountered. As re- differ between the two groups (20). The authors of this text
ported by Van Hoe et al. (19), in their evaluation of scan advocate ECG gating in patients with known dissections or
delay times in spiral CTA using a test bolus injection, aneurysms of the ascending aorta and in patients who are
the time to peak attenuation within the aorta varied being evaluated for the postoperative ascending aorta. If
between 11 and 32 seconds (mean 20, standard deviation the prevalence of acute aortic disease is less than 1% to 2%
6.1 seconds). This problem is easily resolved by ad- in the patients sent for MDCT of the thoracic aorta, the
ministering 20 mL of contrast media injected at the same additional radiation dose is difficult to justify. However, a
rate as the planned injection, with images obtained at recently published study from Massachusetts General
the same preselected level (usually the middle ascend- Hospital of ungated MDCT showed 18% of cases referred
ing aorta or arch) using 80 to 90 mAs. As discussed in for acute aortic disease had positive findings (21). Under
Chapter 3, a similar approach may be used to optimize these circumstances of relatively high true positive cases,
evaluation of the pulmonary arteries. Alternatively, it is ECG gating should be considered essential. Similarly,
now possible to obtain equivalent data using automated patients referred for the “triple rule out” to exclude
scan acquisitions to monitor vascular opacification with- pulmonary embolic disease, acute aortic disease, and coro-
out the use of a test dose [C.A.R.E bolus (Siemens) or nary artery occlusive disease also require ECG gating.
Smartprep (GE Medical Systems, Milwaukee, WI)].
Although administration of a test dose of contrast does
Magnetic Resonance
have the theoretical deleterious effect of increasing the
background attenuation of abdominal organs, rendering It cannot be overemphasized that optimal use of MR
visualization of small vessels more difficult with CTA, this requires meticulous attention to technique. Considerable
consideration is of much less concern within the thorax. emphasis is placed on the many fine points of scan tech-
An advantage of using a test dose (timing bolus) over an nique that must be mastered to ensure diagnostic accuracy.
automated triggering technique is the ability to test the Once mastered, it should be apparent that MR represents
integrity of the intravenous line without having to give the an extraordinarily powerful tool for the evaluation of
entire bolus of contrast. Furthermore, it is not uncommon cardiovascular disease, especially in the nonacute setting.
for an oncologic patient to have an occlusion of a great The advantages of MR for evaluation of the aorta include
vein, resulting in collateral pathways of contrast to the (a) nonreliance on the use of iodinated contrast agents to
aorta with subsequent suboptimal aortic enhancement image blood vessels, (b) the ability to image in multiple
that may preclude accurate triggering. planes, (c) a lack of ionizing radiation, (d) the ability to
evaluate valvular competency and to quantify flow and
determine pressure gradients across stenotic regions, and
Electrocardiographic Triggering
(e) the long safety record of the use of gadolinium chelates
The use of retrospective ECG triggering or “gating” is piv- as contrast agents (22). Limitations of MR compared with
otal in coronary CTA (8,9), but its routine use for thoracic CT include (a) decreased spatial resolution; (b) restriction
in the numbers of patients eligible to be scanned owing to aortic arch vessel disease. In addition, because acquisi-
the presence of various life-support systems and monitor- tion times are determined by heart rate, patients with
ing devices, such as pacemakers or ventilators; and (c) cost bradycardia have long acquisition times and patients
and availability. Until recently, severe arrhythmias that with abnormal cardiac rhythms often have poor ECG
precluded accurate ECG gating were a significant detriment triggering, resulting in suboptimal or even nondiagnostic
to the widespread use of aortic MR; however, new pulse examinations. Newer black blood techniques using a
sequences can reliably image patients with cardiac arrhyth- double inversion pulse can image the thoracic aorta in a
mias in almost any situation (real time imaging). The relative single breath-hold with excellent image quality (26).
insensitivity to calcification can be viewed as favorable Recently, real-time bright blood cine imaging has been
when evaluating patients with severe calcification that implemented by the major vendors [balanced steady-
would render CTA very difficult to postprocess. However, state free precession (SSFP) imaging, True FISP (fast im-
the inability to detect displaced intimal calcifications in aging with steady-state precession) and FIESTA (fast im-
aortic intramural hematoma (IMH) is a major limitation, aging employing steady-state acquisition)], providing
as is the inability to detect a “porcelain aorta,” one so heav- excellent image quality of the aorta and great vessels
ily calcified that it would preclude safe cross clamping for without the need for ECG triggering or breath holding
either proximal or distal control (Fig. 2-1). (Fig. 2-2) (27).
MR angiographic techniques used in evaluating the
thoracic aorta and arch vessels include these cine tech-
Magnetic Resonance Techniques
niques and gadolinium-enhanced three-dimensional (3D)
Contemporary MR imaging of the thoracic aorta usually MRA. These bright blood techniques can differentiate slow
includes multiplanar ECG-triggered spin-echo (SE) flow from thrombus, demonstrate branch vessel disease,
“black blood” imaging supplemented with a “bright and supply additional physiologic information that com-
blood” MRA technique. Whereas SE MR imaging of the plements SE MR imaging. Perhaps most important, these
thoracic aorta is an effective diagnostic tool in evaluating images resemble conventional angiograms and therefore
a wide range of anatomic and pathologic conditions are easier for referring clinicians to interpret.
(23–25), images may be degraded by pulsatility and flow Before the advent of gadolinium-enhanced MRA, time
artifacts, and this technique cannot accurately evaluate of flight (TOF) techniques were widely used. However,
A B
Figure 2-1 Limitations of aortic MR; reduced sensitivity to the presence of calcium. A: Oblique sagittal maximum intensity projection
image from breath-hold, gadolinium-enhanced three-dimensional MR angiogram (TR/TE/FA 3/1.6/50 degrees) performed with 20 mL of
gadolinium in conjunction with a phased-array coil and a timing examination demonstrates severe atherosclerotic arch vessel disease,
including occlusion of the left common carotid artery and severe stenosis at the origin of the innominate (large arrow) and subclavian (small
arrow) arteries. B: Axial multidetector computed tomography (MDCT) shows a heavily calcified (porcelain) aorta that could not safely be
clamped for direct aortic revascularization requiring a combination of stenting and extra-anatomic bypass.
A B
Figure 2-2 Rapid diagnosis of Stanford type A aortic dissection without electrocardiographic (ECG) gating using a balanced steady-state
free precession (SSFP) technique [True FISP (fast imaging with steady-state precession)] in a single breath-hold. Coronal (A) and sagittal
(B) True FISP MR images show both true (T) and false (F) lumen in this Stanford type A aortic dissection. Note that the true lumen (T) is
contiguous with the aortic outflow tract and the false (F) lumen on the right side of the ascending aorta demonstrates thrombus forma-
tion (arrow).
this approach requires ECG gating and breath holding to aortic enhancement irrespective of the patient’s cardiac out-
maximize flow-related enhancement, minimize satura- put or hemodynamic status.
tion effects, and eliminate respiratory misregistration. Early studies using a slow infusion of 0.2 to 0.3
When performed in the oblique sagittal [left anterior mmol/kg gadolinium during a 2- to 4-minute acquisition
oblique (LAO) equivalent] plane, this technique may be demonstrated high sensitivity and specificity for both con-
limited by low spatial resolution, in-plane saturation ef- genital and acquired thoracic aortic disease (29,31).
fects, and TOF signal loss in areas of turbulent flow or sta- However, images are often degraded by respiratory artifacts
sis. Nevertheless, when combined with SE MR imaging, that result in considerable image blurring, especially of the
very high diagnostic accuracy rates have been achieved for arch vessels (Fig. 2-3) and the aortic root (31). In a small
a wide range of congenital and acquired thoracic aortic percentage of cases, image quality of non–breath-hold
diseases (28). gadolinium-enhanced 3D MRA is so poor that the exami-
Gadolinium-enhanced 3D MRA depends on infusion of nation is entirely nondiagnostic. Significant enhancement
paramagnetic contrast agents, which results in a decreased of the brachiocephalic veins is almost always present and
T1 relaxation of blood (29). Using this approach, anatomic may result in obscuration of the aortic arch vessels on
images of blood vessels are generated and distinguished maximum intensity projection (MIP) images. In addition,
from background tissues based solely on T1 relaxation rates. because of the long acquisition times this technique
This contrast mechanism minimizes saturation effects provides no physiologic information regarding flow
and allows for rapid in-plane imaging of large anatomic dynamics; this precludes the differentiation between true
segments. The intravascular signal enhancement of gadolin- and false lumen in aortic dissection based on preferential
ium chelates allows the use of 3D Fourier-transform enhancement. The advantage of the long acquisition time
imaging with intrinsically high spatial resolution and high is the uniform, excellent aortic enhancement, which is
signal-to-noise ratio (30). This technique results in excellent virtually always achieved.
Optimization and Timing of Gadolinium- optimal contrast between the aortic lumen and back-
Enhanced Three-Dimensional Magnetic ground tissues. However, it may be difficult to predict
Resonance Angiography a given patient’s circulation time. Earls et al. (35), using a
1-mL test bolus of gadolinium administered at 2 mL/
With high-performance gradient systems now increasingly second, demonstrated that the time to peak arterial en-
available, the TR and TE can be shortened such that an entire hancement (circulation time from injection site to peak
3D acquisition can be obtained in a single breath-hold (32). aortic enhancement) can vary from 10 to 60 seconds. By
It is well established that there is only a small temporal win- using a timing examination in conjunction with a MR-
dow between peak brachiocephalic arterial enhancement compatible power injector, all patients had diagnostic
and jugular venous enhancement (33,34). Rapid acquisition quality thoracic aortograms with only 20 mL of gadolin-
times may allow for a “pure” arterial study without con- ium administered. Many centers use a timing examination
founding venous enhancement, and breath holding elimi- because it can be performed with any MR machine (i.e., it
nates most of the artifacts seen with the longer, non–breath- is not a vendor-specific product) and requires no addi-
hold strategies. Optimized thoracic aortic studies should tional capital purchase. More importantly, it can test the
demonstrate excellent arterial enhancement, minimal or no integrity of the intravenous line prior to the administra-
venous enhancement, and sufficient spatial resolution to re- tion of the entire bolus and can provide pertinent physio-
solve vessels as small as the vertebral arteries (Fig. 2-3). In logic information concerning the vessels of interest (36)
addition, optimized timing, with phased array or multichan- and overall cardiac function (37).
nel coils suitable for parallel imaging, allows for a reduction Other, more sophisticated methods of optimizing first
in the dose of contrast used without compromising image pass contrast-enhanced MRA have been recently de-
quality or signal-to-noise (34) (Fig. 2-4). scribed. A novel technique developed at the Mayo Clinic
However, with shorter acquisition times, new con- uses fluoroscopic or real time triggering (38). Following
straints regarding coordination of data acquisition with the bolus injection of contrast, multiple two-dimensional
timing of the contrast injection have become apparent. (2D) gradient-recalled echo (GRE) images are acquired
Peak arterial enhancement should coincide with the with a temporal resolution of approximately 1 frame per
acquisition of low spatial frequency lines of k-space for second. When the leading edge of the contrast bolus is
Figure 2-3 The effects of breath

holding on image quality of
gadolinium-enhanced three-dimen-
sional (3D) MR angiography of the
aortic arch vessels. A: Oblique sagit-
tal maximum intensity projection
(MIP) image from non–breath-hold
gadolinium-enhanced 3D MR an-
giogram (TR/TE/FA 21/6/30 deg-
rees) performed in the body coil,
during slow infusion of 40 mL of
contrast over the first 90 seconds of
a 2-minute acquisition, demonstrates
blurring and ghosting artifacts, resul-
ting in poor image quality and a non-
diagnostic examination. B: Oblique
sagittal MIP image from breath-hold,
gadolinium-enhanced 3D MR angio-
gram (TR/TE/FA 5/2/50 degrees)
performed with 20 mL of gadolinium
in conjunction with a phased-array
coil and a timing examination demo-
nstrates vastly improved image qual-
A, B ity and normal aortic arch vessels.
Figure 2-4 Total body aortography: comparison of

maximum intensity projection images from non–breath-
hold, gadolinium-enhanced three-dimensional (3D) MR
angiography with standard gradient strength (TR/TE/
Flip angle 21/6/30 degrees) obtained during slow
infusion of 30 mL of gadolinium (A) with breath-hold
gadolinium-enhanced 3D MR angiography with high-
performance gradient strength (TR/TE/Flip angle
5/2/30–50 degrees) and bolus administration of 15 mL
of gadolinium with a timing examination (B). The latter
demonstrates less confounding venous enhancement
and no central pulmonary arterial enhancement, both of
A, B which degrade (A).
visually detected, breath-holding instructions are given Technical Aspects of Gadolinium-Enhanced

and the 3D acquisition is performed. Another approach Three-Dimensional Magnetic Resonance
developed by Foo et al. (39) uses an automated sequence Angiography
that detects the arrival of gadolinium with subsequent
triggering of the 3D acquisition. The latter technique uses
an SE sequence with a large voxel placed over the vessel of Coil Selection
interest. The arrival of the gadolinium bolus is detected Gadolinium-enhanced 3D MRA examinations are usually
and triggers a 3D GRE sequence with centric phase-encod- performed on a 1.5 or 3 T system with a high perform-
ing. Therefore, the important low spatial frequency lines ance gradient system that allows a TR of less than 4 ms
of k-space (which determine image contrast) are acquired and a TE of less than 1.5 ms. A 4-, 8-, or 16-channel
at the beginning of the acquisition, coinciding with peak multiarray torso coil is preferable, as it provides an
arterial enhancement. advantage over the body coil by virtue of increased sig-
Alternatively, a time-resolved approach can be used, nal-to-noise and provides for higher spatial resolution or
which eliminates the need for these techniques by acquir- faster imaging using parallel imaging techniques. This
ing an entire 3D dataset with sufficient temporal resolution benefit, however, is proportional to the patient’s body
(every 2 to 5 seconds) (40–42) and excellent spatial resolu- habitus; the increase in signal-to-noise is substantial in
tion. This ensures that at least one frame of the dataset thin patients but may be less in obese patients. There are
will demonstrate a selective arterial study of the vessel of distinct disadvantages in using the multiarray coil that do
interest. Even greater temporal resolution can be achieved not apply to the body coil. The FOV is limited in the
with a time-resolved 2D technique (43) but at the expense craniocaudad dimension, which may result in the need
of decreased spatial resolution. With stronger gradient for a second angiogram to evaluate the infrarenal aorta in
systems, and parallel imaging techniques coupled with patients with aortic dissection and diffuse aneurysmal
novel pulse sequences that allow sharing of certain lines disease. However, recent technology allows for coil cou-
of k-space (TRICKS or TREAT), time-resolved MR imag- pling to image as many anatomic segments as one desires
ing will become the standard way to optimize contrast- during the same breath-hold. For example, a 16-channel
enhanced MRA. neurovascular coil can be linked to a multichannel torso
coil, allowing imaging of the thoracic aorta to the circle

of Willis with a single injection. The bright signal of the
subcutaneous fat may also obscure vessels of interest on
MIP images, especially if a strong rectangular FOV is
used. However, use of subtraction algorithms can mini-
mize this problem if there is good coregistration between
the precontrast and gadolinium-enhanced acquisitions.
In addition, portions of the aorta closest to the coil often
have more signal than segments deeper in the chest, mak-
ing the MIP images difficult to window appropriately
(Fig. 2-5). This can also be minimized with the use of a
postprocessing filter function. Finally, in some patients
(especially children), the weight of the anterior compo-
nent of the coil is uncomfortable and may potentiate
claustrophobia. For these reasons, some centers still use
the body coil when evaluating the thoracoabdominal
aorta (Fig. 2-4). However, the body coil is clearly subopti-
mal when evaluating small vessels, such as the vertebral
arteries, because the small FOV needed to resolve these
vessels (25 cm) results in poor signal-to-noise. This is
compounded by the fact that the very short echo time
intrinsic to these 3D sequences (1 to 2 ms) necessitates
the use of high bandwidths. This in turn also results in Figure 2-5 Oblique sagittal maximum intensity projection
a signal-to-noise penalty. image from breath-hold gadolinium-enhanced three-dimensional
MR angiogram demonstrates two pseudoaneurysms arising from
When studying infants or neonates, the head or knee the ascending aorta. These appear to have higher signal than the
coil can be used, and spine coils may be used for young descending aorta because of their proximity to the anterior
children and adolescents with thin body habitus. Larger elements of the phased-array coil.
children and adolescents can be studied in either the
multiarray coil or the body coil.
axillary vein draining the injection site (Fig. 2-6). For this
reason, when specifically evaluating the subclavian, axillary,
Patient Positioning or brachial arteries, the injection should be administered in
Regardless of the coil selected, the patient’s arms are posi- the side contralateral to the expected pathology so that
tioned at the sides, and the right antecubital fossa is concentrated gadolinium in the draining vein will not result
catheterized so that contrast draining the injection site is in artifactual signal loss from susceptibility artifacts. In addi-
not present in the left brachiocephalic vein (which could tion, the use of a large-volume saline flush (20 mL) will
obscure arch vessels on MIP images). Another pertinent help minimize these artifacts. If bilateral disease is antici-
reason not to inject from the left side is that some patients pated, there are two options: (a) place the intravenous site
with thin body habitus may have physiologic compression in a lower extremity or (b) dilute the gadolinium to a 33%
of the left innominate vein against the sternum, resulting in solution of normal saline, increase the flow rate to 3 to
restriction of flow past the lesion with filling via collateral 4 mL/second, and use the shortest possible echo time the
vessels (44). In a study by Lee et al. (44) 9.1% of 475 scanner will allow (45). One important point in differen-
carotid MRA studies were poorly enhanced because of left tiating real from artifactual stenosis is the absence of
arm injections. Compression of the left brachiocephalic collateral vessels seen in pseudolesions.
vein between the sternum and aorta was confirmed in four
cases by venography, CT, and MRI. Scan Parameters
Although some authors advocate placing the patients’ Given the wide range of variables currently available for per-
arms over their heads (to minimize aliasing artifacts), this forming MRA, we recommend a 3D GRE sequence using the
should not be performed for arch studies, as this position following parameters: [2–4/1–2/30–50 (TR/TE/flip angle)],
may cause compression of the subclavian artery and vein matrix of 256 512 with frequency encoding superiorly to
against the first rib. This compression can be significant inferiorly, a 4/8 to 7/8 rectangular FOV with a maximum
in normal patients (without pathologic thoracic outlet dimension of 25 to 45 cm (dedicated arch vessel studies
syndrome) and can lead to a false-positive diagnosis of require a smaller FOV than thoracic aorta examinations).
occlusive disease in the subclavian arteries as well as suscep- A coronal or oblique sagittal/left anterior oblique equiva-
tibility artifacts from concentrated gadolinium in the lent slab is used (thickness 9 to 12 cm), 60 to 100 partitions,
saline, both infused at 2 to 3 mL/second while 1-cm axial

magnetization-prepared GRE images (TR/TE/FA/inversion
time 11 ms/4.2 ms/15 degrees/300 ms/), one every 2 sec-
onds for 60 seconds, are obtained (46). If a dark blood GRE
pulse sequence is not available, one can use an ordinary GRE
pulse sequence performed in the sagittal plane. Under these
circumstances, because of in-plane saturation effects, a test
bolus of gadolinium will result in transient aortic enhance-
ment significantly greater than the precontrast image.
The time to peak arterial enhancement is determined by
selecting the slice with the greatest qualitative enhancement.
This is less cumbersome than drawing region-of-interest
cursors over the desired vessel.
Using the circulation time, an imaging delay between
the start of the infusion of contrast and the start of the 3D
acquisition is calculated with the following equation (32):
Imaging delay circulation time (infusion time/2)
(imaging time/2)
This delay time is optimized for sequential phase-
Figure 2-6 Artifactual stenosis of the subclavian artery, caused encoding pulse sequences, whereby the important low spatial
by T2 shortening effects from concentrated gadolinium, secondary frequency lines that determine image contrast (center of k-
to transient venous obstruction from abducted arms (arms over
the head position). Coronal maximum intensity projection image space) are sampled in the center of the sequence acquisition.
from breath-hold gadolinium-enhanced three-dimensional MR When using centric or elliptical centric phase-encoding MRA
angiogram demonstrates artifactual stenosis of the left subclavian pulse sequences, these important lines are acquired at the be-
artery (arrow). This results from compression of the axillary vein
draining the injection site because of abduction of the arms. A sec- ginning of the acquisition, and the formula may need to be
ond acquisition, after an additional saline flush, failed to demon- adjusted by adding an extra imaging delay of 2 to 4 seconds.
strate an abnormality. Of note is that some patients, those without Once the imaging delay time is calculated, the identical
thoracic outlet compression, can occlude their subclavian vessels
with arm abduction. (From reference 34, with permission.) 3D gradient-echo sequence is repeated with injection of
20 to 30 mL of gadolinium at 2 to 3 mL/second with an
MR-compatible power injector followed by a 20-mL saline
flush; an inadequate flush may result in artifactual signal
effective interpolated section thickness of 1 to 2 mm, 1 ac- loss from concentrated gadolinium, as previously de-
quisition (acquisition time of 5 to 20 seconds). Saturation scribed (Fig. 2-6). The examination is performed at end-
pulses and gradient moment nulling are not used. Partial inspiration, as most patients find this more comfortable
Fourier and parallel imaging can be used to decrease acqui- than breath holding at end-expiration, although the latter
sition time, as can TRICKS or TREAT. usually provides for better subtraction imaging with less
It is important to position the slab from axial images misregistration. A second acquisition is performed either
at multiple levels to avoid excluding portions of ex- immediately after the first or after a 10- to 15-second
tremely ectatic aortas. An unenhanced acquisition is ini- respite during which patients can catch their breath. The
tially performed to ensure that the vessels of interest are purpose of this second acquisition is to better opacify the
included in the imaging volume, to optimize the rectan- false lumen in patients with communicating dissection
gular FOV that can be used without aliasing artifacts, to and as a “bail-out” if the first acquisition is suboptimal
determine the patient’s ability to suspend respiration, from acquiring data too early.
and for use as a mask for subsequent image subtraction. Following gadolinium-enhanced 3D MRA breath-hold
This subtraction technique is helpful in eliminating the T1-weighted 3D axial gradient-echo images can be
bright, subcutaneous fat, which may degrade the quality acquired through the chest to evaluate for extraluminal
of MIP images (46). disease and as a general anatomic survey (47). These
images are easy to interpret, as they are very similar in
Timing Examination appearance to contrast-enhanced CT and can readily detect
To precisely match peak arterial enhancement to the acquisi- pulmonary nodules as small as 1 cm as well as mediastinal
tion of the central lines of k-space, a timing examination can lymphadenopathy (48). In active vasculitis and arteritis,
be performed to measure circulation time to the region of these delayed 3D axial MR images provide the best con-
interest (antecubital vein to aortic arch). This technique trast to demonstrate mural enhancement, often seen in
employs a 1-mL bolus of Gd-DTPA followed by 20 mL of early disease prior to the development of stenotic lesions.
Black Blood Imaging and to determine if flow in the false lumen is antegrade
Prior to gadolinium-enhanced MRA, axial and either or retrograde.
oblique sagittal or coronal ECG-triggered black blood
T1-weighted SE images were acquired through the chest.
Techniques of Reconstruction
This technique provides excellent anatomic images and is
extremely useful to evaluate for mural pathology, such as CTA (and MRA) can be performed using either multiplanar
IMH and the presence of pleural or pericardial effusions. reformations (MPRs) or 3D renderings (Fig. 2-7) (53–64).
Although it is not always possible to achieve high-quality
black blood luminal images because of poor ECG trigger-
Multiplanar Reformations
ing or flow artifacts, the use of longer echo times (15 ms)
and specialized dephasing gradients may be helpful. Some MPRs are one-voxel-thick 2D “tomographic” sections in-
authors believe that aortic diameter is more reproducibly terpolated along an arbitrary plane or curved surface.
measured using black blood techniques, because of the These have the considerable advantage of computational
excellent contrast between the aortic wall and the adja- efficiency, requiring only a few seconds to reconstruct, as
cent lung, unless the aorta is extremely tortuous. In this well as the advantage of instantaneous window width and
situation, true short axis aortic measurements can be level manipulation. Offsetting any potential decrease in
achieved with reformatted images from the gadolinium- spatial resolution caused by partial volume averaging, it is
enhanced 3D dataset. Although many centers still possible to interactively obtain a series of single oblique
routinely perform ECG-triggered T1-weighted SE MR for reconstructions derived from corresponding transverse im-
black blood imaging, some institutions have replaced this ages, allowing individual vessels to be displayed to best ad-
time consuming technique with dark blood half-Fourier vantage along their nearest long axes. Unfortunately, MPRs
single shot turbo spin-echo (HASTE) imaging (26,49). may be susceptible to stair-step artifacts, limiting their in-
This technique is insensitive to patient motion, can terpretation. This effect may be overcome to some degree
provide excellent image quality even with poor ECG by decreasing the reconstruction interval.
gating, does not require breath holding, and takes less
than 1 minute to acquire 20 images. Dark blood images
Three-Dimensional Rendering
are obtained by using two separate inversion pulses (dou-
ble IR) so the signal from incoming blood is nulled. Under this heading are included a number of different
Because of the limited signal-to-noise inherent in single techniques, all of which have in common the use of an
shot imaging, a phased-array coil must be used. Contrast is entire volume (or preselected subvolume) of scan data to
also different with this technique and reflects the echo generate 2D images that convey 3D spatial information.
time of the pulse sequence used. Three broad categories may be identified: multiplanar
Alternatively, an ECG-triggered double-inversion, volume reconstructions (MPVRs), including MIPs; shaded
single-slice breath-hold technique can be used for higher surface displays (SSDs); and volumetric rendering. Of
spatial resolution techniques, such as detailed evaluation these, the latter two may be adapted to allow both external
of thoracic aortic atheromas (50). By applying an addi- and internal rendering of vessels.
tional inversion pulse, a triple IR pulse sequence can be
obtained with fat suppression. It is important to point out Multiplanar Volume Reconstructions
that this fat suppression technique is based on inversion MPVRs are a variant of MPRs that allow a user to obtain
time and is not specific for fat. For example, a bron- 2D representations from a 3D volume (or slab) at any
chogenic cyst with the same T1 of fat will also lose signal angle using average, ray sum, MIP, or even minimum
on a triple IR pulse sequence. intensity projections (MINIPs). Familiar as a common
ECG-triggered turbo SE T2-weighted or turbo short T1 method for visualizing both MRA and CTA data, MPVRs
inversion recovery (STIR) images have a limited role in the may be obtained in any direction, including the axial
evaluation of aortic disease with the exception of patients plane, thereby simulating true cross-sectional images (57).
suspected of having graft infections or aortitis. In the latter Of these, the most commonly used for imaging vessels are
clinical setting, high signal intensity within the aortic wall MIPs (58). MIPs are derived by projecting onto an imaging
may correlate with disease activity. plane the brightest voxel encountered in a ray through the
Contrast-enhanced MRA provides excellent anatomic scan volume. The result is a 2D image in which blood
images but cannot quantify flow or pressure gradients. vessels are typically highlighted (Fig. 2-7A). Alternatively,
Under these circumstances, cine phase-contrast MR with voxel values along individual rays may be either averaged
velocity encoding can supply additional information, or summed; averaged images mimic the appearance
including measurement of pressure gradients across of routine MPRs, whereas summed vessels appear translu-
stenotic valves and areas of aortic narrowing (coarcta- cent. MIPs encode variations in voxel attenuation or
tion) (51,52). These sequences can also be used to differ- intensity, allowing differentiation between calcium and
entiate the true from false lumen in aortic dissection thrombus; they are not threshold dependent. As a result,
A B
Figure 2-7 Diffuse thoracic aneurysm: reconstruction options. See Color Figure 2-7B. A: Maximum intensity projection image depicts
full extent of the aneurysm’s enhanced lumen but fails to show the low-attenuation thrombus (arrow) as well as osseous structures. B: Volu-
metric rendered image retains the full range of tissue densities represented within the volume scanned, allowing a greater range of three-
dimensional images, resulting in improved anatomic orientation. (Courtesy of Dominik Fleischmann, Stanford University, Stanford, California.)
they also may be of value by depicting differences in flow MIPs are simple and quick to perform, but they retain only
rates, as, for example, occurs in aortic dissection or even approximately 10% of the data from the original image.
within single vessels as a result of antegrade flow. However,
unlike other volume rendering techniques, those derived
using MPVRs have no depth cues. As a consequence, MIPs
Shaded Surface Displays and Volumetric
are best suited for displaying anatomy in which superim-
Rendering
position of structures is minimized. This limitation may be
overcome in part by multiple projections viewed in a cine Included in this category are external and internal image
mode as well as volume editing, restricting MIPs to only renderings. External rendering allows visualization of
a few transaxial images—so-called thin slab MIPs—to structures from an external vantage point using a simu-
remove unwanted overlying structures. Nonetheless, MIP lated external light source. This can be accomplished
images, as all other MPVRs, are best interpreted in con- using either surface thresholding or volume rendering
junction with the original axial source images (58). The techniques. In the former, a single threshold is selected
MIP algorithm is usually most effective when the objects and used to generate surfaces at the boundaries of struc-
of interest are bright and there is little gradation in the tures, effectively discarding all other data within the scan
image. High signal intensity from sources other than ves- volume. The result is often exquisite depictions of the
sels can interfere with visualization; however, this can be relationships between anatomic structures, including vas-
mitigated on MR imaging by using image subtraction. cular branches. Unfortunately, despite the striking visual
appearance of these images, external rendering techniques 3D volume-rendered images in the diagnosis of thoracic
may require time-consuming editing to create appropriate aortic anomalies in 14 pediatric patients and young
subvolumes for analysis using either manual trace or con- adults. All images were reviewed by three radiologists for
nectivity-based editing techniques. In addition, the use of position of the aortic arch, coarctation, vascular compres-
a single threshold makes shaded surface techniques sion of the airway, collateral vessel formation, and aor-
susceptible both to noise and to artifacts related to partial topulmonary shunts (patent ductus arteriosus). Average
volume averaging; this can result in surface discontinu- accuracies were greater than or equal to 96% for diag-
ities or holes, jagged edges, and floating pixels. More im- noses of aortic position and airway narrowing on all
portant, focal areas of high density, in particular, vascular image types. For the diagnosis of coarctation, average was
calcifications and metallic stents, may be completely 73% for axial, 100% for multiplanar, and 100% for 3D
obscured. It cannot be overemphasized that surface ren- volume-rendered images. For the diagnosis of patent duc-
dering techniques reflect voxel boundaries, not true tissue tus arteriosus, average sensitivities were 78% for axial,
interfaces. 94% for multiplanar, and 89% for 3D volume-rendered
In distinction to SSDs, volume rendering uses linear or images. They concluded that axial, multiplanar, and 3D
continuous scaling techniques in which every voxel volume-rendered images serve equally well as methods
within the original dataset is assigned a proportional for assessing the side of the aorta to diagnose anomalies.
value based on the full range of tissue densities repre- For evaluation of coarctation and patent ductus arteriosus,
sented (59). By use of a transfer function, Hounsfield multiplanar and 3D volume-rendered images perform
numbers may then be mapped to brightness, color, slightly better than axial images.
and opacity (60). This approach results in fewer artifacts
and allows a greater range of possible 3D images (includ-
Magnetic Resonance Image Analysis
ing the ability to create transparent tissue planes, allowing
anatomic structures to be viewed at various depths, as well Similar to CT, multiple display options and postprocessing
as stereoscopic imaging). Most important, volumetric techniques can be used to evaluate the gadolinium-
rendering allows depiction of vascular calcifications and enhanced acquisition and, ultimately, to present the imaging
thrombus and provides improved visualization of small information to the referring clinician in a user friendly
vessel anatomy (Fig. 2-7B). Although previously limited fashion. To maximize diagnostic accuracy and minimize
by the need for extensive, time-consuming image process- interpretation error, the source data, MPR, and MIP images
ing, 3D CT now can be performed using real-time interac- should be evaluated on an interactive console. Because the
tive volume rendering (60,61). MIP algorithm includes only the voxels with the greatest
Similar to external rendering, internal renderings can be signal intensity within a given ray tracing, structures of lower
performed using either volumetric rendering or 3D surface signal intensity, such as thrombus or intimal flap, may be
shaded reconstructions to simulate the endoluminal excluded. Therefore, evaluation of the source data with perti-
appearance of vessels—so-called virtual angioscopy (62). nent reformations is mandatory for diagnosis of branch
Internal rendering allows visualization of the internal vessel involvement in aortic dissection and for detection and
structures of organs from a point source at a finite dis- accurate measurement of aneurysmal disease. Furthermore,
tance, simulating human perspective ostensibly as seen MIP images interpreted in isolation may be unreliable in
through an endoscope. Images are viewed through the estimating vessel caliber and can overestimate the degree of
tip of a cone with either a narrow (15 degree) or wide stenosis in a given vessel. In addition, diminutive vessels
(60 degree) FOV (63). A flight path of sequential images may not be readily identified on MIP images but are easily
can also be constructed and navigated either sequentially seen on source data (Fig. 2-8). When overlapping vessels
or in a cine loop, further simulating endoscopy. (usually veins) obscure the artery of interest, a restricted or
Although 3D shaded-surface reconstructions are rela- subvolume MIP can often eliminate the obscuring vessel.
tively fast, they are limited by their greater susceptibility to Multiplanar reformations are invaluable in demonstrating
noise and partial volume averaging. Volumetric rendering abnormalities of small vessels, such as vertebral artery ostial
has the advantage of allowing perivascular tissues to be lesions, and in detecting aortic arch atheromas. Other post-
seen through vessel walls, allowing more precise anatomic processing capabilities include SSD algorithms, volume
localization, although shaded surface reconstructions also rendering, and virtual endoscopic renderings (63,64). How-
allow precise localization of perivascular structures by ever, it is unclear whether these time-consuming techniques
means of simultaneous annotated displays of correspond- provide any incremental diagnostic yield when compared
ing axial, coronal, and sagittal images oriented to conform with evaluation with just the MIP and MPR images (65). For
to any given plane identified from the endoluminal dis- evaluating the renal arteries, Mallouhi et al. (54) demon-
play (62,63). strated that the volume-rendered renal MRA enabled more
With respect to clinical studies comparing MDCT ren- accurate detection and quantification of renal artery stenosis
dering techniques for thoracic vascular imaging, Lee et al. than did MIP, with significantly improved vascular delin-
(56) compared the accuracies of axial, multiplanar, and eation. In a similar study with 28 patients referred for renal
A B
Figure 2-8 Artifactual occlusion on maximum intensity projection (MIP) image—a pitfall in rendering a diagnosis based only on MIP
image without evaluation of source data. A: Coronal MIP image from breath-hold gadolinium-enhanced three-dimensional MR angiogram in
a patient with giant cell arteritis demonstrates apparent chronic, bilateral subclavian artery occlusions with extensive collateral vessels in the
shoulder region. Note the direct origin of the left vertebral artery from the aortic arch. B: Source image from same acquisition shows a
patent but severely stenotic left subclavian artery (arrows). (From reference 34, with permission.)
MRA, volume-rendered and MPR images were both better and sinuses of Valsalva are often seen on contrast-enhanced
than MIP images for evaluating stenosis when compared CT scans and always seen on ECG-gated examinations.
with digital subtraction angiography (55). Above the aortic root, the ascending aorta continues for a
distance of 4 to 5 cm. This segment of the aorta is well
demonstrated on CT.
NORMAL ANATOMY The aortic arch begins at the innominate artery and
consists of two segments. The proximal part of the arch is
Regional anatomy of the mediastinal great vessels and their the longest and gives rise to the innominate, left carotid,
relationships to each other and other mediastinal structures and left subclavian arteries, although variations in the
are discussed in Chapter 4. In this section, the appearances branching pattern of these vessels and their divisions are
of specific vascular structures are reviewed.
The Aorta and Its Branches

The appearance of the aorta is characteristic, although it
can vary somewhat in size and shape in different individu-
als (66). The thoracic aorta is usually considered to consist
of an ascending segment, a transverse segment or arch, and
a descending segment (Fig. 2-9).
The proximal portion of the ascending aorta is termed
the aortic root; it consists of the aortic valve, annulus, and
the sinuses of Valsalva. The right and left coronary arteries Figure 2-9 Anatomy of the thoracic aorta and significant land-
arise from the right and left sinuses of Valsalva; the poste- marks. The ascending aorta extends from the aortic valve to the
origin of the innominate artery. Its proximal portion, in relation to
rior sinus is unassociated with a coronary artery and is the aortic valve and sinuses of Valsalva, is termed the aortic root.
referred to as the noncoronary sinus. The aortic root is The aortic arch begins at the innominate artery and ends at the
surrounded by cardiac structures, with the right atrium to ligamentum arteriosum. Its most distal part, which is often slightly
narrowed, is termed the aortic isthmus. The descending aorta
its right, the right ventricle and pulmonary outflow tract begins at the ligamentum. Its proximal portion may appear slightly
anterior to it, and the left atrium posterior. The aortic valve dilated and has been termed the aortic “spindle.”
common. The distal part of the arch, between the origin The ascending and descending portions of the aorta
of the left subclavian artery and the ligamentum arterio- appear rounded, and in these locations, the diameter of the
sum, is known as the aortic isthmus; it is relatively short, aorta can be accurately determined. As measured in more
measuring 1 to 2 cm in length, and its lumen may be a few than 100 normal subjects (68) the average diameter of the
millimeters narrower in adults than the aorta immediately proximal ascending aorta is 3.6 cm (range 2.4 to 4.7 cm),
distal to the ligamentum. Slight narrowing of the isthmus the ascending aorta just below the arch is 3.51 cm (range
is most common in young children. 2.2 to 4.6 cm), the proximal descending aorta is 2.63 cm
Distal to the ligamentum arteriosum is the descending (range 1.6 to 3.7 cm), the middle descending aorta is
thoracic aorta. The most proximal portion of the descend- 2.48 cm (range 1.6 to 3.7 cm), and the distal descending
ing aorta may appear slightly dilated, a finding that is aorta is 2.42 cm (range 1.4 to 3.3 cm) (Fig. 2-10). Aortic
more common in children than adults. This dilated seg- diameters have also been shown to increase with age, corre-
ment of the proximal descending aorta has been termed late with vertebral size, and be larger in men then women
the “aortic spindle” because of its fusiform shape (Fig. 2-9) (Table 2-1).
(67); a more focal dilatation of the anterior aorta at this It is important to recognize that the aorta can vary con-
level may be caused by a “ductus diverticulum.” The largest siderably in its diameter in different patients, as indicated
branches of the descending aorta are the intercostal and by the ranges of values listed previously. However, in an
bronchial arteries. individual patient, the aorta should taper in a consistent
Usually, narrowing of the aortic lumen in the region of fashion along its length. Any significant deviation from
the aortic isthmus is not visible on transverse images; this should arouse suspicion of an aneurysm.
however, this finding may be seen on oblique sagittal CT In children, the aortic diameters at these same levels
reformations or MR imaging. It is important to keep in have been shown to correlate closely with age and increase
mind that the aortic arch also varies in appearance, in a linear fashion with growth (Fig. 2-11) (69). In every
depending on its curvature and orientation relative to the case, the diameter of the ascending aorta was greater than
scan plane. Scans traversing different parts of the arch that of the descending aorta at the same level.
along its length can result in a false representation of aor- The wall of the aorta measures several millimeters in
tic arch diameter. When scans are centered through the thickness in normal subjects but increases with age. On
lower portion of the arch, it may appear constricted in its unopacified CT scans, the aortic wall usually appears with
mid or posterior portions. Also, although the diameter of the same attenuation as blood in the aortic lumen and
the aortic arch decreases gradually from anterior to poste- cannot be distinguished from it. However, in some anemic
rior, the arch may appear elliptical on scans through the patients, the wall may appear slightly denser than aortic
top of the arch. blood on unopacified scans; focal or concentric increased
density of the aortic wall associated with thickening can
also be seen in patients with IMH (70) and arteritis (71).
Normal Aortic Diameter
In the presence of atherosclerosis and aortic plaques, focal
The thoracic aorta tapers progressively from its origin and areas of wall thickening may appear lower in attenuation
consequently varies in diameter at different levels, with than adjacent aortic wall and aortic blood because of their
the ascending aorta and anterior arch being about 1 cm lipid content (72); these plaques are often irregular in
larger than the posterior arch and descending aorta; in attenuation. On contrast-opacified scans, the aortic wall
unusual cases, the descending aorta appears larger than the may sometimes be seen as lower in attenuation than blood
ascending aorta. in the aortic lumen.
Figure 2-10 Mean values for aortic diameter measured

using CT in 100 normal subjects of varying age. At each level,
the range of values seen in normal persons was approximately
1 cm more or less than the mean. (From Aronberg DJ, Glazer
HS, Madsen K, Sagel SS. Normal thoracic aortic diameters
by computed tomography. J Comput Assist Tomogr. 1984;8:
247–250, with permission.)
TABLE 2-1
AVERAGE AORTA DIAMETER AT DIFFERENT LEVELS, BY AGE AND SEX
Men (age, yr) Women (age, yr)
Level 21–40 41–60 60 21–40 41–60 60
Proximal ascending aorta (level A) (cm) 3.47 3.63 3.91 3.36 3.72 3.50
Distal ascending aorta (level B) (cm) 3.28 3.64 3.80 2.80 3.47 3.68
Proximal descending aorta (level A) (cm) 2.21 2.64 3.14 2.06 2.63 2.88
Middle descending aorta (level B) (cm) 2.25 2.39 2.98 1.91 2.45 2.64
Distal descending aorta (level C) (cm) 2.12 2.43 2.98 1.89 2.43 2.40
Modified from Goss CM. Gray’s anatomy. Philadelphia: Lea & Febiger; 1996:57.
Aortic Branching lobe. At the level of origin of the left subclavian artery, the
posterior portion of the arch should not be confused with
The three great arterial branches of the aorta arise se- a mediastinal mass.
quentially and are often seen at different levels. The Although this pattern of branching is typical, variations
innominate artery arises as the first branch of the arch and are common (73,74). A “normal” branching pattern was
is usually seen more caudally than the other branches. It found in only 74% of 300 arteriograms reviewed by
is located near the midline of the trachea or slightly to the Sutton and Rhys Davies and in 56% to 70% of autopsy
right of midline in most normal patients and is in close cases (Fig. 2-12) (73,74). The most common variation is
proximity to the anterior tracheal wall (Fig. 2-12). The a combined origin of the innominate and left common
innominate artery is usually the largest aortic branch, as it carotid arteries (Figs. 2-12 and 2-13), which is seen in
gives rise to the right subclavian, common carotid, and about 20% of patients at arteriography and in 22% to
vertebral arteries. The left common carotid artery arises 36% of cases at autopsy. In about 4% to 6% of cases, the
next and at a more cephalad level. It lies to the left and left vertebral artery arises as a separate branch of the aorta,
slightly posterolateral to the innominate artery; generally between the left common carotid and subclavian
it has the smallest diameter of the three major arterial branches, instead of arising from the left subclavian artery
branches. The left subclavian artery is the last arch branch itself (Figs. 2-8, 2-12, and 2-14).
and arises from the posterosuperior portion of the aortic Above the level of the innominate artery bifurcation, the
arch; it is usually visible at the most cephalad aspect of right subclavian and right common carotid arteries can be
the arch seen on CT. The left subclavian artery is a rela- identified as separate structures and, although usually lo-
tively posterior structure throughout most of its course, cated somewhat more anteriorly, may appear quite similar
lying to the left of and frequently directly lateral to the to the left subclavian and common carotid arteries in size
trachea. The lateral border of the left subclavian artery and location. The exact position of the bifurcation of the
typically indents the mediastinal surface of the left upper brachiocephalic artery is variable, depending on the length
Figure 2-11 Approximate mean values for

aortic diameter in children aged 5 and 10 years.
Ninety-five percent confidence limits include
values approximately 3 mm above and below the
values shown. (From Fitzgerald AW, Donaldson JS,
Poznanski AK. Pediatric thoracic aorta: normal
measurements determined with CT. Radiology.
1987;165:667–669, with permission.)
Figure 2-12 The most common branching patterns of the arch great vessels, with their approximate frequencies and typical CT appear-
ances. IA, innominate artery; LCCA, left common carotid artery; LSA, left subclavian artery; LVA, left ventricle artery; RVA, right ventricle
artery; RSA, right subclavian artery.
A B
Figure 2-13 Common origin of the left common carotid and innominate artery (bovine aorta). A: Oblique sagittal maximum intensity pro-
jection image from breath-hold gadolinium-enhanced three-dimensional MR angiogram fails to demonstrate the origin of the left common
carotid artery because of extreme tortuosity. B: Reformatted image from same dataset demonstrates that the left common carotid does
arise from a common trunk with the innominate artery and loops inferiorly just after its origin.
A B
Figure 2-14 Direct origin of the vertebral artery from the aortic arch. Oblique sagittal maximum intensity projection image from breath-
hold gadolinium-enhanced three-dimensional MR angiogram (A) and axial MDCT image (B) demonstrate a direct origin of the left vertebral
artery from the aortic arch (arrow). Note that the vertebral artery (arrow) is anterior to the left subclavian and posterior to the left common
carotid artery in B.
and degree of tortuosity of this vessel. In a significant per-

centage of cases, the brachiocephalic artery bifurcates some-
what distally; in these cases the right subclavian artery has
an oblique course, and if thin sections are not obtained near
the thoracic inlet, it may not be seen at all. When the right
common carotid and right subclavian arteries are visible,
they normally are found only in sections through the upper-
most portion of the superior mediastinum. Visualization of
these vessels in a more inferior position, just above the aor-
tic arch, frequently indicates the presence of some type of
vascular mediastinal anomaly, usually involving the aortic
arch. This has been termed the “four vessel sign” by
McLoughlin et al. (75); it is common in double aortic arch.
Although the great vessels should be recognizable by
their characteristic appearances and locations, tortuosity
or ectasia of these vessels may present a confusing picture
and be easily mistaken as pathologic. Use of intravenous
contrast media, with or without multiplanar reformations,
will almost always resolve this problem. Alternatively, Figure 2-15 Redundant common carotid artery presenting as
MRA readily demonstrates tortuosity (Fig. 2-13), ectasia, a palpable neck mass with pseudo-occlusion of the right axillary
artery. Coronal maximum intensity projection image from breath-
and vessel loops that may present clinically as a palpable hold gadolinium-enhanced three-dimensional MR angiogram in a
neck mass (Fig. 2-15). patient referred for a palpable neck mass demonstrates the mass
The subclavian arteries exit and enter the mediastinum as a common carotid artery loop (short arrow). Diffuse signal loss
in the right axillary artery (long arrow) is artifactual and caused by
by crossing over the first ribs, behind the proximal stagnant, concentrated gadolinium in the adjacent vein draining
portions of the clavicles. The subclavian vein lies anterior the injection site. This resulted from failure to administer a flush
and the subclavian artery lies posterior to the anterior after the gadolinium bolus. (From reference 34, with permission.)
A B
Figure 2-16 A: Hypothetical double aortic arch of Edwards. (From Shuford WH, Sybers RG. The aortic arch and its malformations. Springfield,
IL: Charles C Thomas; 1974, with permission.) B: Schematic representation of embryonic double aortic arch. Normally, there is interruption of the
right arch distal to the right subclavian artery. The result is a normal left arch. Right aortic arches result when there is interruption of some por-
tion of the left aortic arch. Five potential sites of interruption have been identified. If there is interruption distal to the left subclavian (at either
1 or 2), the result is a right aortic arch with mirror image branching (types 1 and 2). If there is interruption between the left subclavian and left
common carotid arteries (at 3), the result is a right aortic arch with an aberrant left subclavian (type 3). If the arch is interrupted proximal to the
left common carotid artery (at 4), the result is a right arch with an aberrant innominate artery (type 4). Finally, interruption may occur both distal
to the left subclavian artery and proximal to the left subclavian artery (at 1 and 3); this results in an isolated left subclavian artery, connected to
the left pulmonary artery through the left ductus arteriosus (type 5). DAo, descending aorta; AAo, ascending aorta; LSA, left subclavian artery;
LCC, left common carotid artery; Ld, left ductus arteriosus; Rd, right ductus; RCC, right common carotid artery; RSA, right subclavian artery.
scalenus muscle, which attaches to the superior border of double arch system, described by Edwards (76), is used as
the first rib. The transition from the mediastinal to the ax- a framework. As shown in Figure 2-16, in this system there
illary portions of the subclavian arteries is difficult to visu- is an aortic arch and a potential ductus arteriosus on each
alize because they angle sharply as they cross over the first side; the descending aorta is in the midline posteriorly.
rib. The axillary portions of the subclavian arteries lie pos- Interruption of this arch at different locations can explain
terior to the corresponding veins. the various aortic arch anomalies. These can be divided
into three main groups: left aortic arch anomalies, right
aortic arch anomalies, and double aortic arch anomalies.
CONGENITAL AORTIC DISEASE Normally, there is interruption of the hypothetical right
arch distal to the right subclavian artery. The right com-
mon carotid and subclavian arteries fuse to become the
Aortic Arch Anomalies
right brachiocephalic artery as the proximal portion of the
To recognize anomalies of the aorta and great vessels, it is embryologic right arch becomes incorporated into the left
helpful to have a working knowledge of embryology. These arch. The result is the normal left-sided aortic arch. This is
anomalies are most easily understood if the hypothetical demonstrated in Figure 2-17.
Figure 2-17 Hypothesized embryologic development of the normal left arch. The shaded portion of the right arch, distal to the right
subclavian artery, is interrupted. (From Shuford WH, Sybers RG. The aortic arch and its malformations. Springfield, IL: Charles C Thomas;
1974, with permission.)
of the embryologic right aortic arch between the right

common carotid and the right subclavian arteries. The
right subclavian artery then originates from the posterior
portion of the left-sided arch and crosses the medi-
astinum obliquely from left to right, lying posterior to
the trachea and esophagus (Figs. 2-18 and 2-19).
Figure 2-18 Schematic representation of an aberrant right

subclavian artery.
Left Aortic Arch Anomalies

The most common congenital anomaly of the aorta is an
aberrant right subclavian artery originating from an oth-
erwise normal left-sided arch. This occurs in approx-
imately 0.5% of the normal population (76). According
to Edward’s model, this occurs when there is interruption
A
B C
Figure 2-19 Aberrant right subclavian artery. A–F: Sequential CT images through the mediastinum from the great vessels to the aortic
arch, inferiorly, respectively. An aberrant right subclavian artery is present, arising from the medial-posterior portion of the aortic arch on
the left. This artery passes posterior to the esophagus and trachea (arrow in C) and then proceeds superiorly to eventually lie in a normal
position in the superior mediastinum (arrows in A–C). G: Sagittal MR image in another patient with an aberrant right subclavian artery. Again,
the aberrant subclavian artery (labeled A in G) is posterior to the esophagus and trachea (arrow in G) with compression of these structures.
(continued)
D E
F G
Figure 2-19 (continued)
Dilatation of the artery at its origin (Kommerell divertic- this occurs almost exclusively in patients with aberrant
ulum) is common (Fig. 2-20), occurring in up to 60% of subclavian arteries (78).
cases, and may present as dysphagia. True aneurysms of
an aberrant right subclavian artery (Fig. 2-21) have a mor-
Right Aortic Arch Anomalies
tality as high as 50% and should be surgically repaired
(77). Many arch vessel anatomic variants coexist with Using Edward’s double aortic arch model, five potential
aberrant subclavian arteries; the most frequent is a com- anomalies can occur, although only two are relatively
mon origin of both carotid arteries, which occurs in 48% common. The type of anomaly encountered will depend
of cases (Fig. 2-22) (78). Less common is a direct origin on the exact point at which the left aortic arch is
of the vertebral artery from its ipsilateral carotid artery; interrupted.
A B
Figure 2-20 Enlargement of an aberrant subclavian artery at its origin (diverticulum of Kommerell). Oblique sagittal maximum intensity projec-
tion image (A) and axial reformation (B) from breath-hold gadolinium-enhanced three-dimensional MR angiogram show enlargement of the origin
of an aberrant right subclavian artery. (From reference 34, with permission.)
The most common right aortic arch anomaly is a

right aortic arch with an aberrant left subclavian artery
(Fig. 2-23). This occurs in approximately 1 in 2,500
patients. The sequence of events leading to this malforma-
tion is illustrated in Figure 2-24. In this case, there is inter-
ruption of the left arch between the left common carotid
and left subclavian arteries. This results in an anterior left
common carotid artery as the first branch of the ascending
aorta and a retroesophageal left subclavian artery, which
may also result in dysphagia when aneurysmal. If the right
arch itself becomes aneurysmal, it may also compress the
esophagus (Fig. 2-25). This variant is usually not associ-
ated with congenital heart disease. In both aberrant right
and left subclavian arteries, the portion of the aberrant
artery that is adjacent to the spine may be compressed,
resulting in vessel narrowing. A right aortic arch with
mirror-image branching occurs if the hypothetical left arch
Figure 2-21 Aneurysm, aberrant right subclavian artery. is interrupted distal to the left subclavian artery. This
Contrast-enhanced CT image at the level of the aortic arch anomaly is significant because of the high incidence of
demonstrates an aneurysm of an aberrant right subclavian artery
(arrow) with considerable amount of thrombus. The aneurysm concomitant congenital heart disease, usually tetralogy of
causes considerable mass effect on the adjacent esophagus Fallot. Other rare left aortic arch anomalies may result in
(curved arrow).
Figure 2-22 Aberrant right subclavian

artery associated with common origin of both
common carotid arteries. A, B: Oblique sagit-
tal maximum intensity projection images
from breath-hold, gadolinium-enhanced
three-dimensional MR angiogram demon-
strate plaque within an aberrant right subcla-
vian artery (long arrow in A), which is stenotic
at its origin. There is a severe stenosis of the
ipsilateral vertebral artery (straight arrows)
and the origin of the left subclavian artery.
Both carotid arteries arise from a common
trunk. There is a deep ulcerated plaque on the
inferior aspect of the aortic arch (curved
A, B arrows). (From reference 34, with permission.)
Figure 2-23 Right-sided aortic arch with aberrant left subcla-

vian artery. Coronal (A) and oblique sagittal (B) maximum inten-
sity projection images from breath-hold, gadolinium-enhanced
three-dimensional MR angiogram show a right-sided aortic arch
with a stenosis at the origin (long arrow in A) and midportion
(short arrows) of an aberrant left subclavian artery. The left
vertebral artery is occluded and the right vertebral artery is
prominent with a focal stenosis (curved arrows). The injection of
contrast was administered through the right arm. (From refer-
A, B ence 34, with permission.)
Figure 2-24 Hypothesized embryonic development of a right aortic arch with an aberrant left subclavian artery. RSA, right subclavian
artery; RCCA, right common carotid artery; LCCA, left common carotid artery; LSA, left subclavian artery; LPA, left pulmonary artery; RPA,
right pulmonary artery. (From Shuford WH, Sybers RG. The aortic arch and its malformations. Springfield, IL: Charles C Thomas; 1974, with
permission.)
A B
Figure 2-25 Aneurysmal dilatation of a right-sided aortic arch

with aberrant left subclavian artery: CT and MR demonstration.
A–C: Sequential MDCT images through the mediastinum from the
great vessels to the aortic arch, inferiorly, respectively. There is
aneurysmal dilatation of the ascending, arch, and descending por-
tions of a right aortic arch. The left subclavian artery arises from
the descending aorta (arrows in A and B). D: T1-weighted SE MR
image at the same level as A demonstrates displacement of the
esophagus anteriorly (arrowhead). E: Axial reformation from
breath-hold gadolinium-enhanced three-dimensional MR an-
giogram at the same level as A and D demonstrates a four-vessel
arch configuration. Note that because of the arch aneurysm the
superiormost portion of the aorta is at the thoracic inlet and
C appears similar to a cervical aortic arch (continued).
D E
the syndrome of the interrupted aortic arch (IAA), which is more usually, left sided. The two arches may be of equal
discussed in detail elsewhere in this chapter. size, although the right arch is generally larger (Fig. 2-26).
In the usual case of right aortic arch with mirror-image Less common is the double arch in which atresia of some
branching, interruption of the left arch occurs distal to the portion of the left arch has occurred. There is similarity
ductus arteriosus. The result is that there is no structure between this condition and some of the right aortic arch
posterior to the trachea or esophagus. Rarely, interruption anomalies. Theoretically, the major difference is that with
occurs distal to the left subclavian artery but proximal to double aortic arches, some portion of the left arch persists
the ductus (79). If the ductus on the left side persists the despite atresia, and a vascular ring around the trachea and
result is a true vascular ring. esophagus is present. With right aortic arch anomalies
there is true interruption of the left aortic arch, and a true
vascular ring is rarely formed.
Symptomatic Arch Anomalies
ECG-triggered SE MR (82–84), 2D TOF, cine MR, and
Symptomatic arch anomalies include double aortic arch, phase-contrast MR (85,86) can readily demonstrate symp-
right arch with aberrant left subclavian artery and persistent tomatic arch anomalies (Fig. 2-27) but require multiple
left ligamentum arteriosum, and cervical arch. All of these imaging planes, which are time consuming to perform
abnormalities can result in tracheal and esophageal com- and result in limited postprocessing potential. A single
pression, but the former two often present in infancy with volume gadolinium-enhanced 3D MRA acquisition offers
respiratory or feeding problems. Cervical aortic arches are higher spatial resolution, higher signal-to-noise, and faster
rare anomalies that are usually right sided and are associ- acquisition times and provides for high-quality surgically
ated with unusual great vessel origins (80). The innominate pertinent multiplanar reformations (87). However, with-
artery may fail to form, and there are usually separate out breath holding, respiratory-induced blurring can result
origins of the subclavian and carotid arteries. They are often in significant degradation of image quality with this
circumflex, crossing the chest or lower neck in a near technique.
coronal orientation that may result in compression of the Helical CTA has successfully been used in the evalua-
esophagus (81) and may present as a pulsatile neck mass. tion of pediatric great vessel anomalies (56,88–91).
Double aortic arches have been classified into two Advantages over MR imaging include rapid scanning
types, depending on the patency of the arches (79). The time, which allows for examination of critically ill chil-
most frequent form is a functional double aortic arch. In dren without the constraints of a closed-bore magnet, and
this case, both arches remain patent; the right common better evaluation of the relationships of aberrant vessels
carotid and subclavian arteries arise from the right arch, with the airways and the esophagus. Hopkins et al. (90)
whereas the left common carotid and subclavian arteries evaluated 15 children (aged 1 month to 12 years) with he-
arise from the left arch. Both arches join posteriorly to lical CTA using 3-mm collimation and a bolus of 2 mL/kg
form one descending aorta, which may be midline or, of nonionic contrast injected by hand at 1 mL/second.
A B
C D
Figure 2-26 Double aortic arch. A: Most superior image shows carotid (CA) and subclavian (SA) vessels on both sides of the trachea. B:
Contiguous inferior image shows a larger right aortic arch (r) and a smaller left aortic arch (l) giving rise to the great vessels. C: The right aor-
tic arch (r) crosses to the left of the midline and joins the descending left aortic arch (l). Note that the trachea is not severely compressed at
this level. D: Most inferior image at the level of the aortic pulmonary window. C, superior vena cava; AA, ascending aorta; DA, descending
aorta.
A B
Figure 2-27 Double aortic arch: MR demonstration in two patients. A: T1-weighted SE MR image demonstrates a double aortic arch. In
this case both arches are the same size. B: Cine gradient-echo MR image in a different child demonstrates a double aortic arch, with the
right arch larger than the left.
Scanning began immediately after the infusion of contrast (25%), B (70%), and C (5%) in the Celoria and Patton
finished and was performed with breath holding in two classification (93). A large patent ductus arteriosus invari-
children and during quiet respiration in 13. Using 2D axial ably supplies blood to the descending aorta. IAA is also
images and 3D SSD, they were able to diagnose anomalies associated in the majority of cases with intracardiac defects
correctly in 13 children (87%), including double and and subaortic outflow obstruction. IAA usually presents
right-sided aortic arches, aberrant subclavian arteries, within the first few days of life, as the ductus closes.
innominate artery compression syndrome, and pulmonary Rapidly progressive congestive heart failure ensues. The
artery abnormalities (90). More recently 16-slice MDCT mortality rate at 10 days is 50%, and, by the end of the
was used to diagnose infants with double aortic arches first month, 75%. Initial management includes prost-
with thinner collimation, faster imaging times (mean 4 aglandin therapy to maintain the patency of the PDA.
sec), and reduced contrast dose (91). Echocardiography has traditionally been used to quickly
diagnose IAA, but distinguishing between IAA type A and
severe coarctation is difficult. Hemodynamically stable
Interrupted Aortic Arch
patients, as well as those who have undergone surgical
IAA is an interruption in the continuity of the aortic arch, repair, can be evaluated with cine (94) or gadolinium-
often with a fibrous cord between the two segments. It enhanced 3D MRA (95). Critically ill patients and those
occurs in three places: (a) just distal to the origin of the left who cannot tolerate anesthesia are better studied with
subclavian artery, (b) between the origins of the left com- MDCT (96).
mon carotid and left subclavian arteries (Fig. 2-28), and
(c) between the origins of the innominate and left
common carotid arteries (92). These correspond to types A Coarctation
Clinical Features
Coarctation of the aorta represents a focal narrowing in
the proximal descending thoracic aorta, usually in the
region of the ductus arteriosus. When associated with
tubular hypoplasia of the aortic arch or descending aorta
(Fig. 2-29), it is usually referred to as preductal or infantile.
Patients with aortic coarctation may present with conges-
tive heart failure as infants when a patent ductus arteriosus
closes; alternatively, the condition may be discovered
during workup of a coexistent cardiac anomaly, the most
common being a bicuspid aortic valve, which occurs in
75% to 85% of patients. In 2% of patients coarctation
may occur as an isolated segment of narrowing of the
abdominal aorta.
Treatment
Three types of surgical repair of coarctation remain in
common use: resection of the stenosed segment of aorta
with end-to-end anastomosis, use of a subclavian flap that
is mobilized proximal to the internal mammary artery and
reflected down as a flap to bridge the coarctation, and
patch aortoplasty (using synthetic material for the patch)
(97). The choice of operative technique is influenced by
the morphology of the coarctation, the age of the patient,
Figure 2-28 Interrupted aortic arch with hypoplastic ascend- and the personal preference of the surgeon, although
ing aorta in a 10-day-old infant. Coronal reformation from among surgeons specializing in congenital heart disease,
non–breath-hold gadolinium-enhanced three-dimensional MR end-to-end anastomosis has become the most popular
angiogram, performed in the head coil, demonstrates a hypoplas-
tic, 4-mm ascending aorta (black arrow), which bifurcates into both option, as it probably provides the best anatomic relief of
common carotid arteries (white arrows). The arch was interrupted obstruction, has the lowest risk of recoarctation, and prob-
distal to the common carotid arteries. (From Krinsky G, Weinreb J. ably has the lowest incidence of late aneurysm formation.
Gadolinium-enhanced three-dimensional MR angiography of the
thoracoabdominal aorta. Semin Ultrasound CT MR. 1996:17:280– Dacron patch aortoplasty was a commonly used repair
303, with permission.) technique that resulted in a low rate of recurrence and
A B
Figure 2-29 Aortic coarctation and a hypoplastic descending aorta in an 18-year-old man treated with extra-anatomic bypass from the
ascending to descending aorta. This case represents the importance of subvolume maximum intensity projection (MIP) images in removing
extraneous anatomy that may obscure the vessels of interest. Oblique sagittal MIP image from breath-hold gadolinium-enhanced three-
dimensional (3D) MR angiogram (A) demonstrates an extra-anatomic bypass graft from the ascending aorta to the descending aorta. Arrow
represents a portion of the distal hypoplastic native aorta. Subvolume oblique sagittal MIP image from breath-hold gadolinium-enhanced
3D MR angiogram (B) demonstrates the complete extent of the coarctation and hypoplastic aorta (short arrows) as well as the initial bypass
graft (long arrow) that was unable to adequately decrease left ventricular afterload and hypertension. The extra-anatomic bypass graft from
the ascending aorta to the descending aorta was subsequently placed and resulted in normalization of blood pressure and a decrease in left
ventricular mass. (From reference 29, with permission.)
stenosis, but because of an alarming incidence of late study in children, angioplasty was associated with a
aortic rupture and sudden death resulting from aneurysm higher incidence of aneurysm formation and iliofemoral
formation (98), this technique is now rarely performed. arterial injury than was surgery (101). However, long-
Recoarctation (in reality a mixture of residual and recur- term results from stent placement for coarctation or
rent coarctation) occurred in approximately 10% of cases recoarctation have proven much better than angioplasty
in most published surgical series up until the early 1990s, alone, with a negligible recoarctation rate and no
but in the last decade refinement of the end-to-end aneurysm formation (102).
anastomosis has resulted in a reduction in recoarctation to Adults with coarctation usually present with hyperten-
less than 4% (99). sion and discrepant blood pressure measurements in the
More recently, transluminal balloon angioplasty with arms and legs and less commonly with symptoms of
or without stenting has been used successfully in selected aortic stenosis, aneurysms, and, rarely, aortic dissection.
cases and may be the procedure of choice for recoarcta- Almost all patients with coarctation of the aorta will have
tion without aneurysm formation. This nonsurgical treat- concomitant left ventricular hypertrophy. Coarctation
ment works best when the stenosis is focal and located in adults is often associated with medial degenerative
just distal to the left subclavian artery. In a randomized disease of the aorta; thus, resection with interposition
trial comparing surgery and balloon angioplasty for na- graft is a widely used surgical approach. If primary resec-
tive coarctation, aneurysms visible on angiography were tion is not feasible, extra-anatomic bypass grafting from
present in 20% of cases after angioplasty compared with the ascending to descending aorta can be performed (Figs.
none after surgery (100). In another recent randomized 2-29 and 2-30) under cardiopulmonary bypass (103).
A B
Figure 2-30 Patent extra-anatomic aortic bypass graft in an adult patient with untreated aortic coarctation associated with a native aortic
pseudoaneurysm. The patient was being evaluated for an ascending aortic aneurysm and aortic valve insufficiency. A: Oblique sagittal
source image from non–breath-hold gadolinium-enhanced three-dimensional acquisition clearly demonstrates an extra-anatomic graft
(curved arrow) and pseudoaneurysm (p) of the native aorta. There is also narrowing of the proximal left subclavian artery. Susceptibility
artifact is present from a prosthetic mitral valve (arrow). B: Oblique sagittal maximum intensity projection (MIP) image shows an ascending
aortic aneurysm (arrow) and a normal abdominal aorta. Note the susceptibility artifact is obscured in the MIP image. (From reference 31,
with permission.)
A less invasive alternative is transluminal angioplasty cessfully used in the evaluation of coarctation, to assess the
with stent graft placement, but the long-term results are location, the degree of anatomic narrowing, and the extent
unknown. of collateral circulation. These include multiplanar ECG-
Pseudocoarctation of the aorta results from a congenital triggered TI-weighted SE techniques (108) supplemented
elongation of the aortic arch with resultant redundancy by phase-contrast (109) or cine MRA (110,111). The latter
and kinking (Figs. 2-31 and 2-32) (104). Unlike coarcta- technique, when performed in the oblique sagittal plane,
tion, however, there is no obstruction to blood flow and can readily demonstrate the dephasing that occurs from
therefore little or no demonstrable pressure gradient and turbulent flow across the site of narrowing. Cine phase-
no evidence of collateral circulation (105,106). contrast techniques, including velocity-encoded cine MR,
are used to measure the peak velocity across the site
of coarctation (107). With this information, the pressure
Imaging Features gradient across the coarctation can be estimated. One
study demonstrated that peak coarctation jet velocity
MR Evaluation measured by MR velocity mapping was comparable to re-
Although both MDCT and MR have been successfully used sults obtained with continuous wave Doppler sonography
to diagnose coarctation, in our opinion, MR has proved of (51), and a more recent study showed it was similar but
greater value, especially in the pediatric age group as well added additional morphologic evaluation not assessable
as for routine postsurgical evaluation, as pressure gradients by sonography (112). This technique can also be used
across the stenotic lesions can be determined (107). to determine the extent of collateral circulation by measur-
Various traditional MR imaging techniques, perfor- ing flow immediately distal to the coarctation and at
med in the axial and oblique sagittal plane, have been suc- the level of the diaphragm. In normal aortas, flow will
A B
C D
Figure 2-31 Pseudocoarctation. A: Posteroanterior radiograph shows unusual configuration of the superior mediastinum on the left side
(arrow). Differentiation between an aneurysm and a mediastinal mass is difficult. B, C: Sequential coronal MR images through the root of
the aorta and posterior portion of the aortic arch, respectively. The ascending aorta is elongated as indicated by the considerable distance
between the arch superiorly and the main pulmonary artery inferiorly (arrows in B). There is aneurysmal dilation of the aorta (arrow in C),
below which there is abrupt narrowing of the lumen, suggesting possible coarctation (curved arrow in C). D: Coned-down view from a sub-
traction aortogram shows characteristic kink in the course of the aorta (arrow) without actual narrowing. No pressure gradients could be
measured across this region. (Case courtesy of Patricia Redmond, MD, Staten Island, NewYork.)
decrease by an average of 8% as it reaches the diaphragm, segment of coarctation and the presence of collateral
whereas in patients with hemodynamically significant vessels (114) (Fig. 2-34).
coarctation, the flow will be augmented by an average of MR may be used to predict the response of patients
80% (113). who are candidates for balloon angioplasty and/or stent
Gadolinium-enhanced 3D MRA can readily demon- placement (115,116). The best results with balloon angio-
strate the presence of collateral circulation (Fig. 2-33) as plasty occur when the stenosis is focal and located just distal
well as the degree of associated arch hypoplasia (Fig. 2-29) to the left subclavian artery. When the stenosis involves
and concomitant branch vessel abnormalities. It also the proximal arch or involves a long segment, balloon
provides excellent anatomic images of the extent of coarc- angioplasty is less successful. As documented by Bank et al.
tation without dephasing artifacts from turbulent flow and (116) in their study of 12 children with suspected coarctation
allows reformation of 3D data into multiple surgically evaluated both before and after surgical repair, MR proved
pertinent imaging planes in patients with associated tubu- accurate in identifying those patients who were unlikely to
lar hypoplasia (Fig. 2-29). CT (90) and, more recently, benefit from angioplasty. In addition, MR proved valuable by
ECG-gated MDCT can also demonstrate the anatomic determining the appropriate balloon size for angioplasty.
A, B C
Figure 2-32 Two patients with aortic pseudocoarctation demonstrated with MR angiography (MRA). Coronal (A) and oblique sagittal
(B) maximum intensity projection images from breath-hold gadolinium-enhanced three-dimensional MR angiogram demonstrate pseudo-
coarctation. Note the redundant aorta associated with buckling and inferior displacement of the left subclavian artery and the complete
absence of collateral vessels. In a second patient with pseudocoarctation (C), note the presence of previous supracoronary ascending aortic
graft replacement as areas of graft puckering (arrows). (A, B from reference 29, with permission.)
Figure 2-33 Aortic coarctation—evaluation with gadolinium-

enhanced MR angiography (MRA). Oblique sagittal maximum
intensity projection image from breath-hold gadolinium-enhanced
three-dimensional MRA demonstrates aortic coarctation with
extensive collateral circulation. (From reference 29, with permission.)
A B
C D
Figure 2-34 Aortic coarctation—CTA evaluation. A–D: Sequential contrast-enhanced helical CTA from superior to inferior demonstrates
multiple collateral vessels, including intercostal (short arrows in B) and internal mammary (long arrows in B) arteries. Left ventricular
hypertrophy is present as well. The site and degree of narrowing are not well visualized.
All patients with surgically repaired coarctation Pulmonary Sequestration

should undergo surveillance either with MR (111,
112,117) or with MDCT imaging to detect recurrent Pulmonary sequestrations are defined as areas of non-
coarctation and pseudoaneurysm formation (Fig. 2-35). functioning lung tissue that are not in continuity with the
One study using ECG-triggered 2D TOF MRA with 3D tracheobronchial tree and derive their blood supply
surface renderings demonstrated a sensitivity of 100% from systemic vessels. Traditionally divided into intra- and
for the detection of aneurysms (98). Recurrent coarcta- extralobar forms, they are now considered by most authors
tion after primary repair may result from multiple to be two entirely separate entities. Intralobar sequestrations
causes, including inadequate removal of ductal tissue, are contiguous with normal lung parenchyma, enclosed by
failure to repair associated aortic arch hypoplasia, and normal pleura. In distinction, extralobar pulmonary seques-
suture line tension (118). These patients can be treated trations lie outside the normal pleura, frequently appearing
with transluminal angioplasty and/or stenting with as solid masses of tissue. Intralobar sequestrations are only
excellent results (102). Alternatively, coarctation resec- rarely associated with other malformations and typically are
tion and/or extra-anatomic bypass grafting from the first identified in later life, often as an asymptomatic finding
aortic arch to the descending aorta (Figs. 2-29, 2-30, and on routine chest radiographs. In distinction, extralobar
2-35) can be performed with excellent long-term pulmonary sequestrations are frequently seen in association
results (103). with other congenital abnormalities and are most often
A B
Figure 2-35 Aortic coarctation and postoperative pseudo-

aneurysm formation detected by CT and MR in three different
patients. A: Oblique sagittal maximum intensity projection image
from breath-hold gadolinium-enhanced three-dimensional MR
angiogram demonstrates an aneurysm at the site of previous coarc-
tation repair. Note the enlarged, tortuous internal mammary arter-
ies with areas of focal stenosis. B: Axial contrast-enhanced CT
image through the level of the right pulmonary artery demon-
strates a pseudoaneurysm (long arrow) adjacent to the aortic graft
(small arrows). Curved arrow denotes the native aorta. C: Sagittal
reformatted image from multidetector CT (MDCT) shows a large
pseudoaneurysm with thrombus (T) at the site of previous coarcta-
tion repair. Note segment of extra-anatomic bypass (arrow) from
ascending aorta. D: Volume-rendered image in same patient shows
exclusion of the pseudoaneurysm and extra-anatomic bypass from
innominate artery to distal abdominal aorta with a second graft
revascularizing the distal left subclavian artery. (Images C and D
courtesy of Gary Israel, MD, New Haven, Connecticut.) D
A B
Figure 2-36 A, B: MR demonstration of vascular supply to intralobar sequestration. Coronal source (A) and maximum intensity projection
(B) images from breath-hold gadolinium-enhanced three-dimensional MR angiogram demonstrate a feeding vessel (arrows in A and B) to a
pulmonary sequestration.
identified in either the neonatal period or early childhood. involve one or more aortic segments (aortic root, ascending
In patients with intralobar sequestration, the feeding artery aorta, arch, or descending aorta) and are classified accord-
arises from the thoracic aorta in 73% of cases, the abdominal ingly. Sixty percent of thoracic aortic aneurysms (TAAs)
aorta or celiac axis in 20%, and the intercostal arteries in 4% involve the aortic root and/or ascending aorta, 40% involve
(119). Venous drainage typically is to normal pulmonary the descending aorta, 10% involve the arch, and 10% in-
veins. Although extralobar sequestrations also typically de- volve the thoracoabdominal aorta (with some involving 1
rive blood from the arterial circulation, these frequently are segment). The etiology, natural history, and treatment of
considerably smaller vessels than seen in patients with thoracic aneurysms differ for each of these segments (125).
intralobar sequestrations. Unlike intralobar sequestrations, Although medial and aortic valve disease is a more com-
however, venous drainage is almost always to systemic veins, mon etiology for ascending aortic aneurysm, most descend-
including the azygos and hemiazygos veins, as well as the in- ing TAAs are due to both atherosclerosis and medial disease,
ferior vena cava (IVC) and even the portal vein. similar to abdominal aortic aneurysms (AAAs). Medial de-
Regardless of the type of sequestration, both aberrant generation leads to weakening of the aortic wall, which in
arteries and veins are easily identified using either MDCT turn results in aortic dilatation and aneurysm formation.
or MR (120–122) (Fig. 2-36). A potential advantage When such aneurysms involve the aortic root, the anatomy
of MDCT in these cases is the ability to simultaneously is often referred to as annuloaortic ectasia. Cystic medial
evaluate lung parenchymal abnormalities that fre- degeneration occurs normally to some extent with aging,
quently are present, especially in patients with intralobar but the process is accelerated by hypertension. Patients with
sequestrations (121). bicuspid aortic valves develop ascending aortic aneurysms
unrelated to valvular disease but rather to concomitant
medial degeneration (125). As previously discussed, aortic
ACQUIRED AORTIC DISEASE size is age related, and a 4-cm ascending aorta is clearly
aneurysmal in a 30-year-old but normal in an octogenarian.
The etiology of thoracic aneurysmal disease is broad,
Thoracic Aortic Aneurysms
including degenerative disease, connective tissue disorders,
trauma, aortitis (infective and inflammatory), previous
Clinical Features
surgery, hemodynamic alterations (from aortic stenosis
Aortic aneurysm is defined as an irreversible dilatation of and regurgitation), and congenital abnormalities. The
the aorta with all components of the aortic wall present in histologic changes occurring in the diseased aorta include
the dilated segment (123,124). Thoracic aneurysms may medial degeneration, medial necrosis, atherosclerosis, and
inflammatory infiltration. Pseudoaneurysms, or false aneur- The most serious complication of aneurysmal disease of
ysms, represent saccular dilatations of the aorta that do not the thoracic aorta is death from rupture, the incidence of
contain an intimal layer and are usually associated with which varies from 42% to 70% depending on the series
previous penetrating atheromatous ulcers, aortic surgery, (128). Davies et al. (124) reported on the yearly rupture or
trauma, or infections (Fig. 2-37). dissection rates in 721 patients with TAA. A total of 304 pa-
Most patients with TAAs are asymptomatic at the time of tients were dissection free at presentation; their natural
diagnosis, because the aneurysms are typically discovered history was followed up for rupture, dissection, and death.
incidentally on imaging studies (chest radiograph, CT, or Patients were excluded from analysis once operation oc-
echocardiogram) ordered for other indications (125). curred. Not surprisingly, the authors reported that
Aneurysms of the root or ascending aorta may produce sec- aneurysm size had a profound impact on outcomes; the
ondary aortic regurgitation, so a diastolic murmur may be cumulative risk of rupturing a TAA is related to aneurysm
detected on physical examination or, less often, patients diameter. For example, based on their modeling, a patient
may present with congestive heart failure. Patients with with an aneurysm exceeding 6 cm in diameter can expect a
aneurysms at the level of the sinus of Valsalva that rupture yearly rate of rupture or dissection of at least 6.9% and a
may present with heart failure or a new murmur. When death rate of 11.8%. Although the mean rate of growth of
TAAs are large, patients may suffer a local mass effect, such thoracic aneurysms was only 0.1 cm/year in this series, the
as compression of the trachea or mainstem bronchus (caus- growth rate varied by location. The serial growth rate
ing cough, dyspnea, wheezing, or recurrent pneumonitis), was greater for aneurysms of the descending aorta versus
compression of the esophagus (causing dysphagia), ascending aorta, was greater for dissected aneurysms versus
compression of the recurrent laryngeal nerve (causing nondissected ones, and was greater for those with Marfan
hoarseness), or compression of the great veins (plethora). syndrome versus those without. In another recent series of
Rarely, chest or back pain may occur with nondissecting 133 patients with TAA, risk of rupture at 5 years was 0% for
aneurysms as a result of direct compression of other aneurysms with a diameter less than 4 cm, 16% for those
intrathoracic structures or erosion into adjacent bone. with a diameter of 4 to 5.9 cm, and 31% for aneurysms
The overall incidence of TAAs from a large autopsy series greater than 6 cm in diameter (129). Another study with
over a 27-year period was 489 per 100,000 men and 437 1,600 patients with TAA showed a 31% risk of rupture
per l00,000 women (126). This study also demonstrated an or dissection for ascending aortic aneurysms at a diameter
increasing prevalence of asymptomatic thoracic aneurysms of 6 cm and 43% for descending TAAs at a diameter of
with advancing age. Aneurysmal disease often is multifocal 7 cm (130). This author recommends elective repair of
or diffuse; up to 28% of patients with thoracic aneurysms ascending aneurysms at 5.5 cm and descending aneurysms
have concomitant infrarenal AAAs, making evaluation of at 6.5 cm for patients without any familial disorders such
the entire aorta mandatory in patients presenting with non- as Marfan syndrome. Patients with Marfan syndrome or
traumatic thoracic aneurysms (123,127). familial aneurysms should undergo earlier repair, when
A B
Figure 2-37 Ascending aortic pseudoaneurysm resulting from prior mediastinitis complicating previous aortic surgery. A: Axial
electrocardiogram-triggered T1-weighted SE MR image shows pseudoaneurysm arising from the anterior portion of the ascending aorta
(long arrow) and associated thrombus or granulation tissue (short arrow). B: Oblique sagittal source image from breath-hold gadolinium-
enhanced three-dimensional MR angiogram demonstrates the larger superior pseudoaneurysm (long arrow) and the neck of the inferior,
smaller pseudoaneurysm (short arrow).
Figure 2-38 Ascending aortic aneurysm with involvement of the

sinus segment in a patient with cystic medial necrosis. Oblique
sagittal black blood half-Fourier single shot turbo-SE (HASTE) image
performed with electrocardiogram triggering (TR/TE 800 ms/
42 ms) (A) and oblique sagittal source image from breath-hold
gadolinium-enhanced three-dimensional (3D) MR angiogram
(B) demonstrate the tulip-shaped aortic root characteristic of cystic
medial necrosis. Note the smooth tapering to a relatively normal
aortic arch. C: Oblique sagittal maximum intensity projection image
from breath-hold gadolinium-enhanced 3D MR angiogram better
shows the full extent of the normal great vessels, but overlapping
B chambers preclude evaluation of the aortic root.
the ascending aorta grows to 4 to 5 cm or the descending 0.56 cm/year for aneurysms involving the aortic arch
aorta grows to 5 to 6 cm (130). (133); these expansion rates are faster than those re-
Measuring the size of TAA from axial MDCT and ported for AAAs. As with AAAs, the size at presentation
MR imaging studies is not an exact science, and both can be used to stratify the cohort at risk for rupture; the
interobserver (131) and intraobserver variability exist. cumulative 5-year risk of rupture increases fivefold for
These errors occur most commonly at the aortic root thoracic aneurysms greater than 6 cm in diameter.
and in tortuous arch and descending TAAs, where size The operative technique for aortic aneurysms, as with
may be under- or overestimated. Recently, both auto- dissection of the aorta, depends on the extent of involve-
mated and semiautomated techniques have become ment. Previously, conventional angiography was required
available to determine a true short axis measurement of for preoperative planning, but ECG-gated MDCT or a com-
an aneurysm. Without these techniques a double bination of MR techniques can supply more information,
oblique orthogonal image can usually be obtained including the ability to evaluate the coronary arteries in
accurately at a workstation when using a 3D dataset. patients with aneurysms involving the aortic root (134)
Although MR has been the preferred technique to eval- and the intercostal arteries that supply the spinal cord in
uate the size of the aortic root due to motion-related patients with descending thoracic and thoracoabdominal
artifacts with MDCT, the use of retrospective ECG aneurysms (135–137). These techniques can readily evalu-
gating with 16- and 64-slice CT has rendered this tech- ate aortic size, mural disease, thrombus, and the proximal
nique better than MR. Some authors advocate using rel- and distal extent of aneurysms. The last is extremely
ative aortic size controlled for body surface area important, as this dictates the operative approach and
information (the “aortic size index”). In one recent technique used. For aneurysms involving the ascending
study, increasing aortic size index was a significant pre- aorta, the type of proximal repair is determined by
dictor of increasing rates of rupture as well as the com- whether or not the sinotubular ridge and aortic valve
bined endpoint of rupture, death, or dissection (132). are compromised, and the distal extent may determine
These authors believe that relative aortic size is more the need for hypothermic circulatory arrest. If the distal
important than absolute aortic size in predicting com- ascending aorta is not aneurysmal or diseased and the
plications and allows for the stratification of patients aortic root is relatively normal, then simple aortic cross
into three levels of risk, enabling appropriate surgical clamping with cardiopulmonary bypass without hypo-
decision making. thermic circulatory arrest may be performed. In dis-
Prior to the use of automated or semiautomated volu- tinction, if the aneurysm extends to within 1 cm of the
metric techniques, a longitudinal CT study demonstrated innominate artery, involves the aortic arch, or is associated
growth rates of 0.42 cm/year for thoracic aneurysms and with extensive atheromatous plaque, then hypothermic
A B
Figure 2-39 Ascending aortic aneurysm—value of electrocardiographic (ECG) gating to minimize cardiac motion. See Color Figure 2-39E.
Axial source (A), coronal (B), and sagittal (C) reformatted images from 16-slice CT angiography (CTA) performed without ECG gating
demonstrate fusiform dilatation of the ascending aorta, with relative sparing of the sinus segment (arrow in B) and extension into the trans-
verse arch. Note the extensive artifact from cardiac motion. Axial (D) and oblique sagittal volume-rendered (E) images from another patient
with ascending aortic aneurysm involving the left coronary sinus (arrow in D) performed with ECG gating show the precise relationship of
coronary arteries to aneurysm (arrows in E).
the etiology) will have aneurysmal involvement of the

sinuses of Valsalva with smooth tapering to a normal-sized
C aortic arch. This results in the classic tulip- or pear-shaped
aorta seen in patients with Marfan syndrome, Turner
syndrome, Ehlers-Danlos syndrome (Fig. 2-38), bicuspid
aortic valve, or familial TAA syndrome with cystic medial de-
generation (125). In contrast, atherosclerotic or degenerative
aneurysms usually spare the sinus segment but extend into
the transverse arch (Fig. 2-39). Aneurysms related to aortic
stenosis may have a peculiar appearance with the greatest di-
ameter at the level of the turbulent aortic jet (Fig. 2-40).
The distinction is not always so clear cut, as atheroscle-
rotic aneurysms may also involve the sinus segment
(Fig. 2-41), especially when associated with aortic regurgi-
tation. However, they will almost never demonstrate the
complete effacement of the sinotubular junction seen with
cystic medial necrosis.
Crawford Classification
Descending TAAs originate beyond the left subclavian
artery and may extend into the abdomen. Crawford
devised a classification system that includes both
aneurysms and chronic dissections based on length and
location of disease and relative risk of paraplegia (138).
Type I involves the descending thoracic aorta from the
left subclavian artery to the suprarenal aorta. Type II ex-
tends from the left subclavian artery to below the renal
arteries and may continue distally to the aortic bifurca-
E tion. Type III begins at the mid-to-distal descending tho-
racic aorta and involves most of the abdominal aorta as
far distal as the aortic bifurcation. Type IV (also a
circulatory arrest with antegrade cerebral perfusion is com- suprarenal AAA) extends from the upper abdominal
monly performed. aorta and all or none of the infrarenal aorta. Descending
The morphology of aneurysms of the ascending aorta and TAAs may compress or erode into surrounding structures,
aortic arch is predictable based on the disease process. including the trachea, bronchus, esophagus, vertebral
Typically, patients with cystic medial necrosis (regardless of body, and spinal column.
Figure 2-40 Ascending aortic aneurysm with relative sparing Figure 2-41 Atherosclerotic ascending aortic aneurysm with
of the sinus segment due to congenital aortic stenosis. Oblique some dilatation of the sinus segment. Oblique sagittal maximum
sagittal maximum intensity projection image from breath-hold intensity projection image from breath-hold gadolinium-enhanced
gadolinium-enhanced three-dimensional MR angiogram demon- three-dimensional MR angiogram demonstrates fusiform dilatation
strates dilatation predominantly involving the anterior wall of the of the ascending aorta, with some involvement of the sinus seg-
ascending aorta, with some involvement of the sinus segment, ment, which extends into the transverse arch. This patient had
without extension into the mid transverse arch. The unusual significant aortic regurgitation. Note the concomitant proximal
morphology of the aneurysm is due to the lifelong jet through a descending aortic aneurysm. Involvement of the transverse arch is
stenotic valve in a patient with congenital aortic stenosis. characteristic of atherosclerotic or senescent aortic aneurysms.
Because of the aortic arch involvement, surgical repair would
require hypothermic circulatory arrest.
For thoracoabdominal aneurysms it is essential that technique (see Postoperative Aorta). Because patients
imaging includes a complete evaluation of the renal and who have undergone repair of thoracic aneurysms are at
visceral vessels, as aneurysmal involvement or occlusive risk for a second noncontiguous aneurysm (123), as well
disease may alter surgical management (Fig. 2-42). These as anastomotic pseudoaneurysms, they require follow-up
aneurysms invariably contain thrombus and at times may surveillance imaging.
be difficult to distinguish from chronic, thrombosed aortic
dissections. The presence of a sharp, spiraling interface
with compression of the patent lumen and extension over Imaging Features
a 7-cm segment points to the latter diagnosis (Fig. 2-43).
Displaced intimal calcifications also favor a chronic throm- Computed Tomography Evaluation
bosed aortic dissection. Early CT literature using nonhelical scanning demonstrated
Complete replacement of the thoracic aorta may be high accuracy rates for the diagnosis of thoracic aneurysms
necessary in aneurysms that involve the ascending aorta, and their complications (140–149). Contrast-enhanced
aortic arch, and proximal descending aorta (Figs. 2-7 and MDCT can demonstrate all the gross pathologic features
2-44). This can be performed in a single operation or as of aortic aneurysms, including dilatation, extraluminal
a two-stage procedure (139) using an “elephant trunk” thrombi, displacement or erosion of adjacent structures
Figure 2-42 Thoracoabdominal aneurysms. Oblique

sagittal maximum intensity projection images from
non–breath-hold (A) and breath-hold (B) gadolinium-
enhanced three-dimensional MR angiograms demon-
strate thoracoabdominal aneurysms with sufficient
proximal necks for cross clamping (curved arrows).
Poor visualization of the renal and visceral vessels in
(A) results from a combination of severe renal
insufficiency and the known limitations of evaluation
of branch vessels using a non–breath-hold technique.
C: Oblique sagittal reformatted image from multi-
detector CT (MDCT) in a third patient demonstrates
A B similar findings.
(Fig. 2-45), and perianeurysmal thickening and hemor-

rhage (140–145). CT has also proven valuable in detecting
unusual types of aneurysms, including mycotic (Fig. 2-46)
and ductus aneurysms (Fig. 2-47), as well as complications
arising from aneurysms, including those secondary to
rupture (145–149). CT also plays an important role in dis-
tinguishing TAAs from pulmonary masses adjacent to the
mediastinum.
Thoracic aneurysms secondary to atherosclerosis are
readily demonstrated as fusiform dilatation of a segment
of the aorta, which usually contains mural thrombus, and
may demonstrate calcification of the thrombus or the
aortic wall (Fig. 2-48). The pattern of mural thrombi and
calcifications within aortic aneurysms has been described
(146,147). Calcification within aneurysms is common,
occurring in up to 85% of cases, both mural (Fig. 2-45)
and within the thrombus itself. The latter type of calcifica-
tion has been reported to occur in approximately 25% of
patients with both thoracic and abdominal aortic calcifica-
tions (147). The significance of calcification within
thrombi is that this appearance can be mistaken for the
displaced intimal calcification seen in aortic dissection
C and IMH.
A B
C, D E
FIG. 2-43. Thoracoabdominal aneurysm with thrombus and chronic thrombosed type B dissection (A and B). Oblique sagittal maximum
intensity projection (MIP) image from breath-hold gadolinium-enhanced three-dimensional MR angiogram (A) and axial, breath-hold gradient-
echo image (B) performed immediately after (A) demonstrate crescentic thrombus in the descending thoracic aorta, without mass effect on the
lumen. Findings are most consistent with a thoracoabdominal abdominal aneurysm. C–E: Chronic thrombosed dissection with aneurysmal dila-
tation of false lumen. C: Oblique sagittal source image from non–breath-hold gadolinium-enhanced three-dimensional MR angiogram shows
low signal intensity thrombus (arrow) with sharp luminal interface, typical for thrombosed dissection. Note secondary aneurysm formation in
the abdominal aorta (curved arrow). D: MIP image is deceptively normal appearing. E: Axial reformation from same dataset better demon-
strates the sharp interface (arrow) between the enhanced true lumen and the thrombosed false lumen. (From Krinsky G, Weinreb J. Gadolinium
enhanced three-dimensional MR angiography of the thoracoabdominal aorta. Sem Ultrasound CT MR. 1996;17:280–303, with permission.)
Aortic thrombi may be crescentic (Fig. 2-49) or cir-

cumferential, in which case the residual lumen appears
circular. As a rule, the residual aortic lumen is smooth,
although the aortic lumen may develop a very irregular
contour, especially in large aneurysms, depending on
the extent of thrombus (Fig. 2-48). Unlike IMHs (which
are discussed later), thrombus is often circumferential
and does not displace a calcified intima. When the
thrombus forms a sharp margin with the enhanced
lumen, and no displaced intimal calcifications are pres-
ent, it may be difficult to distinguish a thoracic
aneurysm from a chronic aortic dissection with a throm-
bosed false lumen.
CT can be used to follow up aneurysms, although
this involves repeated use of intravenous contrast and
ionizing radiation not inherent in MR. An increasing
diameter is readily detected with serial scans and helps
prompt a surgical decision. CT may also have value in
assessing patients with iatrogenic pseudoaneurysms,
which may occur following aortic valve replacement,
cardiopulmonary bypass, or coronary artery bypass
grafting.
With the introduction of volumetric CT, interest has
focused on the potential use of retrospective reconstruction
techniques to evaluate aortic aneurysms. Quint et al. (150)
evaluated the diagnostic contribution of helical CT with
multiplanar reformations in the evaluation of 49 patients
Figure 2-44 Oblique sagittal maximum intensity projection with thoracic aortic disease (36 aneurysms, 6 penetrat-
image from breath-hold gadolinium-enhanced three-dimensional
MR angiogram demonstrates aneurysmal dilatation of the entire ing ulcers, 5 dissections, and 2 pseudoaneurysms). Overall
thoracic aorta (mega-aorta). (From reference 29, with permission.) diagnostic accuracy was 92%, both with and without multi-
planar reconstructions. Although MPRs were of value by
displaying pathology in a format more accessible to non-
radiologists, their addition failed to change the diagnosis in
any patient (150). Furthermore, whereas in two patients
A B
Figure 2-45 A, B: Long-standing thoracic aortic aneurysm resulting in bone erosion. Contrast-enhanced CT scans at the same level
imaged with narrow and wide windows, respectively. There is dense calcification of the walls of both the ascending and descending aorta.
An aneurysm is present within the proximal descending aorta, with considerable surrounding thrombus. The adjacent vertebral body has
been significantly eroded (arrow in B).
Figure 2-46 Longstanding mycotic aneurysm of the ascending

aorta. Sagittal reformation from contrast-enhanced helical CT
demonstrates a mycotic aneurysm of the ascending aorta with
thrombus (long arrows) that eroded through the sternum (short
arrow) to present as a chest wall mass.
B
Figure 2-47 A, B: Ductus aneurysm. Axial contrast-enhanced he-
lical CT images with sagittal reformation demonstrate an aneurysm
in the region of the ductus arteriosus.
Figure 2-48 A, B: Atherosclerotic thoracic

aortic aneurysm. Sequential images (superior
to inferior) from contrast-enhanced helical
CT demonstrate a fusiform aneurysm of the
descending thoracic aorta with irregular,
A mostly circumferential thrombus.
B
extension of an aneurysm into the aortic arch necessitating In distinction, Kimura et al. (151) have shown that both
the use of hypothermic circulatory arrest was shown more SSD and especially CT angioscopy are superior to axial
clearly using MPRs, in a third patient multiplanar recon- images for assessing the relationship between distal arch
structions incorrectly predicted this need. aneurysms and the origin of the great vessels. Comparing
A B
Figure 2-49 A, B: Thoracoabdominal aneurysm with crescentic thrombus. Axial and oblique sagittal reformations from contrast-
enhanced helical CT demonstrate crescentic mural thrombus in both the midthoracic aorta (arrow in A) and at the level of the diaphragm
(arrow in B). Note the similar morphology in Figure 2-43.
the accuracy of 2D axial source images with SSDs and Preoperative Evaluation of the Artery of
endoscopic renderings in 12 patients with distal aortic Adamkiewicz at Multidetector Row Helical
arch aneurysms, these authors found that depiction of the Computed Tomography
relationship between the aneurysm and the orifice of the Mapping of the intercostal arteries prior to surgical repair of
left subclavian artery was correct in only 5 (42%) of 12 pa- Crawford type I to III thoracoabdominal aneurysms and
tients, using axial images; in distinction, using 3D display dissection is now possible. The great radicular artery (artery
only, this relationship could be discerned in 9 of the 12 of Adamkiewicz) can now be determined prior to surgical
(75%), whereas use of 3D endoscopic renderings allowed intervention (or endograft placement) to allow selective
accurate assessment in 11 of 12 (92%) patients (151). The reimplantation. Using 4-detector MDCT, Takase et al. (161)
reasons for failure of axial images to accurately identify the evaluated 70 patients with a preoperative diagnosis of thora-
anatomic relationship between aneurysms and the left coabdominal vascular disease for detection of the artery of
subclavian artery included the presence of atheromas ob- Adamkiewicz, using source data and curved planar reforma-
scuring continuity and gradual sloping of the proximal tions. In 63 (90%) of the 70 patients, at least a single artery
portion of the aneurysm obscuring its precise margin. Of of Adamkiewicz was clearly visualized from the interverte-
the three renderings, CT angioscopy most clearly delin- bral foramen to the hairpin-shaped union with the anterior
eated the distance between the arterial orifice and the bor- spinal artery. Two arteries of Adamkiewicz were identified in
der of the aneurysm. 15 (24%) of 63 patients. Fifty-five arteries of Adamkiewicz
An even more compelling use of CT for evaluating aortic (71%) originated from the left side. Seventy-two (92%)
aneurysms has been reported by Rubin et al. (152,153). originated between T8 and L1. Neither the intercostal vein
Using helical CT angiographic data, these authors demon- nor the posterior spinal vein was visualized in 57 of 63 pa-
strated that it is possible to automatically derive accurate tients. Continuity of the entire length, starting from the
quantitation of arterial cross sections, segment lengths, and stem of the intercostal or lumbar artery and proceeding to
absolute curvature by continuously describing variations in the artery of Adamkiewicz and finally to the anterior spinal
cross-sectional dimensions and as a function of position artery, was traceable on cine-mode displays or on curved
along the median vessel path. Previously limited, especially planar reformation images in 20 of 63 patients (161). It is
in patients with markedly tortuous aortas, use of quantita- fully expected that these results will improve with 64-detec-
tive vascular CT should now allow more accurate assess- tor CT.
ment of longitudinal changes in aortic dimensions with
time, as well as more accurate treatment planning, espe-
cially in those cases for which endoluminal stent-grafts are Magnetic Resonance Evaluation
planned. This is in addition to the already well established Aortic aneurysms and pseudoaneurysms are easily identi-
role of CT in pre- and postoperative evaluation of patients fied on ECG-triggered SE or double inversion recovery (IR)
with both surgical and endovascular stent-graft repair of MR sequences because of the signal void usually associated
TAAs (154–158). with flowing blood (25). When present, organized thrombi
In our experience, the extent of aortic aneurysms, and in within the lumen of aneurysms may appear as areas of
particular their relationship to adjacent structures, are intermediate or variegated signal intensity on black blood
better visualized using either MPRs, curved MPRs, or 3D imaging, and unorganized thrombi may appear hyperin-
renderings. In this regard, the use of MIPs to delineate the tense. As there is little to no macrophage activity within
exact location of calcification may be of particular value in thrombus, extracellular methemoglobin is present much
assessing the necks of aneurysms. longer than what is typically seen in the brain. In fact, it is
CT does have some recognized limitations in the evalua- not uncommon to see high signal intensity thrombus on
tion of aneurysmal disease, aside from the risks of iodi- black blood images in the chronic setting, and its presence
nated contrast and radiation. Evaluation of the aortic root should not be a cause for alarm as long as the aneurysm sac
is often difficult secondary to motion artifacts unless ECG is well defined and the patient is not acutely ill. The ability
gating is performed (Fig. 2-39). Many patients with to reliably differentiate thoracic and thoracoabdominal
advanced aneurysmal disease often have concomitant aneurysms from aortic IMH is critical, especially when
renal insufficiency; however, with a 64-detector CT, optimi- clinical symptoms are vague or not readily attainable.
zed timing, and judicious use of a saline flush, adequate Aortic IMH is usually crescent shaped, exerts no mass effect
studies can be performed with as little as 50 mL of contrast. on the lumen, has a relatively homogeneous signal inten-
Gadolinium chelates have been successfully used for CTA sity, and often extends over a long craniocaudad dimen-
as an alternative to iodinated contrast (159,160), but the sion. In contradistinction, thrombus in a thoracic
high dose needed for a diagnostic quality scan (0.4 mg/kg) aneurysm often has a layered, variegated appearance, may
is much more expensive than nonionic iodinated contrast have mass effect on the lumen, and is usually seen over a
and has recently been linked to a rare kidney disease. short segment of aorta (Fig. 2-50).
A B
Figure 2-50 Differentiation of thoracic aneurysm with thrombus from aortic intramural hematoma. Axial double inversion recovery (IR)
images in a patient with aortic intramural hematoma (A) and thoracic aneurysm with thrombus (B). The aortic intramural hematoma is cres-
cent shaped, exerts no mass effect on the lumen, has a relatively homogeneous signal intensity, and often extends over a long craniocaudad
dimension; note involvement of both ascending (arrow) and descending aorta. In contradistinction, thrombus in this thoracic aneurysm has
a layered, variegated appearance, has mass effect on the lumen, and is usually seen over a short segment of aorta (not shown).
Thrombus is almost always markedly hypointense on performing unenhanced CT prior to thoracic surgery in
gadolinium-enhanced 3D MRA as well as on fat-suppressed patients whose entire preoperative evaluation is with MR.
volumetric interpolated breath-hold examination (VIBE) or
liver acquisition with volume acceleration (LAVA) sequences,
as these pulse sequences are optimized for background ACUTE AORTIC SYNDROMES
suppression, the echo times are very short (1 to 2 ms), and
the intravascular gadolinium changes the dynamic range of Acute aortic syndrome (AAS) is characterized clinically by
contrast (Fig. 2-43). aortic pain in a patient with a coexisting history of hyper-
As documented by Glazer et al. (162), measurements of tension (165). It is the current clinical term that includes
the maximum diameter of aneurysms with MR show near aortic dissection, IMH, penetrating atherosclerotic ulcer,
perfect correlation with the measurements obtained with CT and both contained and overtly ruptured aneurysms or
scans. In addition, MR provides accurate information pseudoaneurysms. Aortic pain may be distinct from angina
concerning the relationships of aneurysms to branch vessels in that it is characterized as intense, tearing, migratory chest
and coronary arteries as well as surrounding mediastinal pain that may radiate to the jaw or back (166). Anterior
structures. Although the distal coronary arteries cannot be re- chest, neck, throat, and even jaw pain is related to involve-
liably evaluated with MR imaging, the proximal vessels can ment of the ascending aorta, whereas back and abdominal
usually be seen in patients referred for ascending aortic pain more often indicates that the affected segment is the
aneurysms (163,164). Because of the relatively large FOV (50 descending aorta. The chest pain and clinical presentation
cm) available in MR, the entire aorta can be studied without of patients with penetrating aortic ulcer and aortic IMH are
having to reposition the patient. These complete aortograms similar to that of classic aortic dissection and ruptured
are extremely helpful when evaluating patients with thora- thoracic aneurysms; however, the latter patients are usually
coabdominal aneurysms (Figs. 2-42 to 2-44) or multiple, hypotensive from volume loss or shock. Severe, acute chest
noncontiguous aneurysms (Fig. 2-41). As already noted, an pain consistent with acute aortic syndrome may also occur
important limitation of MR is its inability to detect calcifica- in patients with pronounced aortic root dilatation but
tions; this can be especially problematic when dealing with a without histopathologic evidence of aortic lesions (165).
totally calcified porcelain aorta that may not tolerate cross Aortic root stretching and distension are likely mechanisms
clamping (Fig. 2-1). Because of this, the authors advocate for the pain experienced by these patients (167). Aortic
pain may be confounded with that of ischemic syndromes. rupture, many unnecessary off hour laparotomies were per-
Moderate to severe hypertension is a universal risk factor formed. The speed of MDCT and near universal availability,
for the development of AAS. Inheritable disorders of elastic coupled with the fact that many of these lesions can be
tissues also predispose to the development of classic aortic treated with endografts, have resulted in the newer para-
dissection, but hypertension is the most frequently associ- digm that every patient is imaged prior to therapy unless he
ated condition. Hypertension is also the most common co- or she is in shock. This marked increase in the utilization of
morbid disease associated with penetrating aortic ulcers MDCT (and TEE) has resulted in a better understanding of
and intramural aortic hematomas. Although laboratory the wide spectrum of aortic disease ranging from stable
tests (creatine kinase and troponin), ECG changes, and aneurysms to both contained and frank aortic rupture.
chest radiograph may help to differentiate them, MDCT
has become the procedure of choice to diagnose and distin-
guish acute aortic syndromes. Because of the widespread Imaging Findings
use of nonsurgical techniques (stent-grafts) to treat acute
aortic disease, the MDCT study should be performed with Computed Tomography Evaluation
thin collimation (1 mm or less) and overlapping recon- Noncontrast images should always be obtained first in all pa-
structions with the reconstructed data sent to a commer- tients with acute aortic syndromes. As renal failure is one of
cially available workstation. If needed, the volumetric the leading causes of death after surgical repair of aneurysms
dataset can be used to size an endograft, demonstrate and dissections, intravenous contrast should be used judi-
preferential cannulation sites, and demonstrate both entry ciously in patients who may already be “prerenal” or in hy-
and re-entry tears. potension-induced acute renal failure. This is especially im-
portant in situations in which TEE may provide enough
information to adequately characterize a lesion prior to surgi-
Ruptured Aortic Aneurysms cal intervention, such as in ascending aortic IMHs. Because
(Contained and Overt) subtle calcification of thrombus may be present within stable
aneurysms, it may be difficult or impossible to differentiate
Physiology focal contrast extravasations into a degenerated thrombus (a
sign of contained rupture) from calcified thrombus without
The wall stress related to blood pressure in the nonaneurys-
performing an initial unenhanced CT (Fig. 2-51).
mal aorta is relatively low and uniformly distributed,
whereas within the aortic aneurysm, regions of high- and
low-stress distribution are present (168). Increased tension
stress results in progressive vessel dilatation and weakening
of the aortic media. According to Laplace’s law, wall tension
is proportional to the vessel radius for a given blood pre-
ssure. When an artery wall develops a weak spot and expands
as a result, it might seem that the expansion would provide
some relief, but in fact the opposite is true. The expansion
subjects the weakened wall to even more tension. Unfortun-
ately, in an expanding aneurysm, forming a near spherical
shape cannot give sufficient tension relief. Aortic aneurysm
rupture is believed to occur when the mechanical stress on
the wall exceeds the strength of the wall tissue (168).
Contained Aortic Rupture

The syndrome of contained aortic rupture or leak is a clini-
cal and radiologic entity describing a patient who presents
with chest or back pain with a history of aneurysm or
chronic dissection and the absence of frank aortic rupture
on imaging studies. The pain is identical to that of acute
aortic syndrome, but the patients are usually not hyperten-
sive and are clinically stable. Prior to the widespread use of Figure 2-51 Value of unenhanced CT imaging in patients with
MDCT, the surgical dictum was that all patients with known acute aortic syndromes. Contrast-enhanced MDCT demonstrates
aneurysms and acute chest or back pain should be taken im- a crescent of high attenuation material (arrow) measuring 170
Hounsfield units within thrombus in an aortic aneurysm. Without
mediately to the operating room and explored. Because the unenhanced CT it is difficult to distinguish contrast intravasation
majority of these patients did not have overt or frank aortic into degenerated thrombus from benign calcifications.
A B
Figure 2-52 Contained aortic rupture of a chronic type B dissec-

tion. A: Routine surveillance enhanced multidetector CT (MDCT)
shows a chronic dissection in the descending thoracic aorta with
the smaller, higher attenuation true lumen (arrow) and both
contrast and thrombus (T) within the false (F) lumen. The patient
presented with back pain 3 weeks after A, and axial enhanced (B)
and unenhanced (C) MDCT now show a new crescent of blood
(arrow in B) in the outer third of the aortic media, confirmed at
surgery. The presence of an intramural hematoma in a patient with
classic dissection confirms that the location of the former lesion
C occurs closer to the adventitia than classic dissection.
Both unenhanced and contrast-enhanced MDCT may

demonstrate findings specific for contained or overt aor-
tic rupture. A high attenuation crescent of blood within a
mural thrombus has been called the crescent sign in
AAAs (169–171) and represents acute hemorrhage into a
pre-existing mural thrombus or aneurysm wall (171)
(Fig. 2-52); this sign may imply acute or impending con-
tained rupture. Mehard et al. (169) reviewed unen-
hanced CT scans of 149 consecutive patients operated on
for AAA. The presence of a peripheral high-attenuating
crescent on CT scans was correlated with surgical find-
ings of aneurysm complication. Aneurysm diameter was
correlated with presence or absence of pain at the time
of CT, high-attenuating crescent, and aneurysm compli-
cation. Sensitivity of the high-attenuating crescent sign
as an indication of complicated aneurysm was 77%;
specificity, 93%; and positive predictive value, 53%. The
sign showed a statistically significant correlation with
large aneurysm size and presence of pain at the time of
Figure 2-53 Contained aortic rupture of a thoracic aneurysm
CT. The authors concluded that in patients without CT with the “crescent sign.” Axial unenhanced multidetector CT
evidence of frank aneurysm leak, the high-attenuating (MDCT) image demonstrates a large thoracic aneurysm with
crescent sign should be regarded as a sign of impending crescent of active bleeding into the thrombus or aortic wall
(arrow). Patient refused surgery and died the next day.
AAA rupture, particularly in patients with pain. The
results can be extrapolated to include TAAs as well
(Fig. 2-53). frank aortic aneurysm ruptures may also demonstrate the
Another CT finding of a contained rupture, or a deficient “missing calcium sign,” a focal gap of otherwise circumfer-
posterior aortic wall, is the draped aorta sign (172). This ential wall calcification (170) (Fig. 2-55).
occurs when the posterior aspect of the aneurysm is in close Mediastinal, pericardial, and pleural hematomas may
apposition to the spine and drapes around a vertebral body also be present and should strongly suggest the presence
(Fig. 2-54). Of the 10 patients with this sign, 7 had con- of overt rupture (Fig. 2-56). Contrast-enhanced CT may
tained rupture of their aneurysm (172). Contained or demonstrate deformities of the enhanced lumen, with or
without active extravasation of contrast. It is likely that
many cases of previously reported ruptured thoracic
aneurysms actually represented penetrating ulcers that had
ruptured. Often the history of a TAA is the only clearcut
way to differentiate these two surgical emergencies.
A B
Figure 2-54 Two patients with contained aortic rupture showing the draped aorta sign. A, B: Note the loss of the normal concavity to the
posterior wall of the aorta as it is draped over the thoracic vertebral bodies in both patients on axial multidetector CT (MDCT) images (arrows).
Compare these images with the normal posterior aortic wall in Figure 2-51. Note also the high-attenuation crescent sign in A (large arrow) and
how much better it is seen on unenhanced than enhanced images (Fig. 2-52). (A courtesy of Elliot Fishman, Baltimore, Maryland.)
A B
Figure 2-55 Ruptured abdominal aortic aneurysm demonstrating the “missing calcium sign,” a focal gap of otherwise circumferential wall
calcification. A: Axial enhanced multidetector CT (MDCT) image performed 2 months prior to admission demonstrates a calcified abdominal
aortic aneurysm. Patient was without symptoms at this time. B: Axial enhanced MDCT image performed at the time of severe back pain
shows a segment of “missing calcium” from the 9 to 10 o’clock position (arrow). Note the hazy appearance of the sac at the site of rupture,
which was confirmed at surgery.
Magnetic Resonance Evaluation sity on black blood images and should not be interpreted
Caution should be used when evaluating the signal inten- as evidence of contained rupture. However, this high signal
sity of thrombus at MR imaging to diagnose rupture. intensity unorganized thrombus is usually in a lamellated
Because methemoglobin may remain within an unorgan- configuration. Clefts of high signal intensity thrombus
ized thrombus for much longer than has been reported in within outer layers of low signal intensity degenerated
the brain, a portion of thrombus may be high signal inten- thrombus may represent contained rupture. The draped
aorta sign is helpful in MR to evaluate the integrity of the
posterior wall as it is in CT (Fig. 2-57). Periaortic
hematoma and active extravasation of gadolinium are
direct signs of rupture (Fig. 2-57).
Aortic Dissection
Clinical Features
Epidemiology
The incidence of aortic dissection is probably understated
owing to the declining autopsy rate; however, published
studies suggest an incidence of 3 cases per 100,000 person-
years (167). According to the International Registry of
Acute Aortic Dissection (IRAD), an ongoing multicenter
prospective study performed on 464 patients since 1996,
two thirds of patients with acute dissection were male
Figure 2-56 Ruptured thoracic aortic aneurysm: multidetector
CT (MDCT) evaluation. Axial contrast-enhanced helical CT image (mean age 63 years), and women were significantly older
through the lower mediastinum demonstrates a ruptured thoracic than men, with a mean age of 67 years (173).
aortic aneurysm. Note the enlarged deformed lumen, as well as Predisposing factors to the development of aortic dis-
the subtle hemorrhage within the wall of thrombus (crescent sign)
(arrow). This would be more obvious on a non–contrast-enhanced section include hereditary and congenital disorders
examination. Mediastinal and left pleural hematomas are present. (syndromes such as Marfan, Turner, and Ehlers-Danlos),
Figure 2-57 Signs of both contained and overt aortic aneurysm

rupture on MR imaging. A: Oblique sagittal maximum intensity
projection image from breath-hold gadolinium-enhanced three-
dimensional MR angiogram shows an infrarenal abdominal aortic
aneurysm. B: Delayed post-gadolinium fat-suppressed 3D MR
image (VIBE) shows the draped aorta sign extending over the left
psoas muscle (open arrow), loss of integrity of the anterior wall of
the sac (thin arrow), and active extravasation of gadolinium into
A the left anterior pararenal space (thick arrow).
aortic valve defects, coarctation, arteritis, and trauma. and potentially re-enter the true aortic lumen at a point
Iatrogenic causes of aortic dissection include manipulation distant from the initial intimal tear. The inner two thirds of
of the aorta during surgery (aortic cross clamping or the media remain contiguous with the aortic intima, creat-
cannulation) or during percutaneous procedures, including the intimomedial flap visualized on imaging modalities
ing catheterization and placement of intra-aortic balloon (157). It is unclear whether the intimal tear is the inciting
pumps. Hypertension was the most common predisposing factor that results in aortic dissection or whether the tear is a
risk factor in the IRAD study, seen in 72% of patients secondary phenomenon from hydraulic stress on an aorta
(173), whereas atherosclerosis was present in 31% of weakened from a spontaneous medial hematoma (174).
patients. Analysis of patients younger than 40 years These stresses or mechanical forces that contribute to aortic
revealed a lower incidence of hypertension (34%), essen- dissection include flexion forces of the vessel at fixed sites,
tially no systemic atherosclerosis, and a higher incidence the radial impact of the pressure pulse, and the shear stress
of Marfan syndrome, bicuspid aortic valve, and/or prior of the blood. During the cardiac cycle, the heart and aorta
aortic surgery (173). produce rhythmic movements, allowing all but fixed
segments to move (175). These fixed points of the aorta are
exposed to the most significant flexion forces. Classic aortic
Pathophysiology
dissections produce an intimomedial tear at the areas of
The pathology of aortic dissection represents a separation of greatest hydraulic stress: the right lateral wall of the ascend-
the aortic media (between the middle and outer third) by a ing aorta (Fig. 2-2 and 2-58) or the descending aorta in
stream of blood that usually is associated with a primary proximity to the ligamentum arteriosum. Decreased flow in
intimomedial tear and underlying medial wall degeneration the vasa vasorum, occurring in arterial hypertension, may
or cystic medial necrosis. This stream creates a false lumen increase the stiffness of the outer ischemic media of the
that may remain contained or dissect longitudinally (paral- aorta to produce interlaminar shear stresses contributing to
lel to the long axis) in an antegrade or retrograde manner the development of aortic dissection (175).
Classification dissections involving the ascending aorta as Stanford A

and those that began distal to the left subclavian artery as
Dissection of the aorta has been classified in various type B. Dissections that involve the ascending aorta,
ways. Acute dissection refers to dissections less than 2 irrespective of the site of entry tear, are classified as
weeks old; chronic, older than 2 weeks. Early studies on Stanford type A and require surgical repair. Although type
mortality involving aortic dissection found that 75% of B dissections are stated to occur only distal to the left
deaths from dissection occur in this 2-week period, mak- subclavian artery (178), this definition has evolved into
ing the distinction prognostically significant (176). In any dissection that does not involve the ascending aorta
addition, dissections are classified based on the extent (including primary arch dissections without retrograde
of involvement of the thoracic aorta according to the extension into the ascending aorta). Acute proximal dis-
DeBakey or Stanford classification (177,178). Debakey sections involving the ascending aorta (Stanford A,
et al. (177) classified aortic dissections into three types Debakey I and II) account for 75% of all cases of aortic
related to anatomic and pathologic features (158). dissection (Fig. 2-58). These dissections are repaired im-
Types I and II involve the ascending aorta, with type I dis- mediately to prevent potentially fatal complications,
sections extending beyond the aortic arch and type II which include rupture into the pericardial or pleural
dissections terminating proximal to the arch vessels. space or extension into the coronary arteries or aortic
Type III dissections involved only the descending root; the latter two complications can result in myocardial
thoracic aorta distal to the left subclavian artery, with IIIA infarction and aortic insufficiency, respectively. Because
confined to the descending aorta (rare) and IIIB extend- some of these patients are clinically unstable, the majority
ing into the abdomen. Dailey et al. (178) refined the will undergo operative repair based on the results of TEE
classification into more clinically relevant terms as alone, as this modality can usually provide all of the
A B
Figure 2-58 Acute Stanford type A dissection. A: Electrocardiogram (ECG)–triggered

turbo spin-echo T1-weighted image demonstrates an intimal flap within the ascending
aorta. It is unclear if the right-sided lumen (arrow) is patent or thrombosed. B: ECG-
triggered black blood (double inversion) half-Fourier single-shot turbo spin-echo image
demonstrates flow within the right-sided lumen. Note the marked decreased magnetic
susceptibility effects from the sternotomy wires in B due to the long chain of 180-degree
minimizing spin dephasing. C: Coronal reformation from gadolinium-enhanced three-
dimensional MR angiography dataset demonstrates patency of the right lumen with
the classic location of the false lumen on the right in ascending aortic dissections. A single
coronary artery bypass graft arises from both the false lumen on the right and the true
lumen on the left (which enhances to a greater degree). Differential signal intensity in
the lumens results from the rapid acquisition time (20 seconds) of this first pass
C technique.
A, B C
Figure 2-59 Acute Stanford type B dissection with a normal size aorta. A: Oblique sagittal electrocardiogram-triggered T1-weighted SE
MR image demonstrates abnormal signal (curved arrow) just distal to the left subclavian artery. B: Oblique sagittal source image from
non–breath-hold gadolinium-enhanced three-dimensional MR angiogram shows an intimal flap starting just distal to the left subclavian
artery and spiraling through the abdomen. C: The origin of the dissection is poorly visualized on this maximum intensity projection, and only
the inferior extent of the intimal flap is definitively identified in the right iliac artery. The signal intensity within each lumen is equal, as the
long acquisition time (2 minutes) results in a recirculation technique whereby the gadolinium concentration in each lumen equilibrates. (From
reference 31, with permission.)
surgically pertinent information at the patient’s bedside in aneurysm exceeds 6 cm in diameter or is growing at a rate
intensive care settings, or MDCT and intraoperative TEE. exceeding 1 cm in diameter per year (Fig. 2-60).
Chronic type A dissections are often associated with cys-
tic medial necrosis (Fig. 2-38) and are repaired as soon as
Clinical Presentation and Imaging Utilization—
medically possible. Acute dissections that begin distal to the
Results From the International Registry of Acute
left subclavian artery (Stanford B, Debakey III) (Fig. 2-59)
are initially treated medically primarily by controlling
Aortic Dissection Study
hypertension, unless complications such as rupture, end- The sudden onset of severe, sharp chest pain was the single
organ ischemia, refractory pain, and/or progression of most common presenting symptom in the IRAD study and
dissection develop. A small percentage of patients with was seen more commonly in type A than type B dissec-
medically treated type B dissections will require early tions (79% vs. 63%), whereas both back and abdominal
surgery. However, in certain centers interventional tech- pain were significantly more common in type B dissection.
niques, including percutaneous intimal flap fenestration Interestingly, the 4.6% of patients who presented with pre-
and/or branch vessel stent placement, have supplanted dominantly abdominal pain had a higher mortality than
surgery in patients with acute malperfusion syndrome those with typical symptoms and a trend toward a delay in
(179,180). More recently, stent-graft placement across diagnosis (185). Patients who presented with syncope
the primary tear may also relieve the malperfusion syn- (13%) were more likely to die in the hospital (34% vs.
drome while potentially also treating the underlying dissec- 23%) and were more likely to have cardiac tamponade,
tion by redirecting blood flow to the true lumen (181–184). neurologic deficits, and type A dissections (186). Pulse
As patency of the false lumen is the most important risk deficits were more common in patients with acute type A
factor for distal secondary aneurysm formation, aortic dissection (19% to 30%) than in those with type B dissec-
stent-grafts may be curative if they result in thrombosis of tions (9% to 21%), and patients with pulse deficits had a
the false lumen. Surgery is indicated in chronic distal dissec- higher mortality rate than those without (173). Overall,
tions in symptomatic patients, or when an associated classic physical findings, such as aortic regurgitation and
A B
Figure 2-60 Chronic Stanford type B dissection with aneurysmal

dilatation of the false lumen and thrombus formation. A: Oblique
sagittal electrocardiogram-triggered turbo SE MR image shows
enlargement of the descending thoracic aorta with differential
luminal signal intensities. B: Oblique sagittal source image from
non–breath-hold gadolinium-enhanced three-dimensional MR angi-
ogram distinguishes enhancing slow flow from unenhanced throm-
bus (T) and better delineates the intimal flap. C: Oblique sagittal
maximum intensity projection (MIP) image fails to demonstrate
thrombus and the majority of the intimal flap. Overlapping high sig-
nal intensity jugular venous blood (curved arrows) obscures por-
tions of the left vertebral artery and right common carotid artery.
This overlap can be eliminated with a more restricted subvolume
MIP or by changing the angle of projection. Evaluation of source
data is not degraded by overlapping vessels. (From reference 31,
C with permission.)
pulse deficit, were noted in only 31.6% and 15.1%, re- Unenhanced CT provides important information that
spectively. Poor prognostic signs for patients with acute may be obscured with intravenous contrast. Ruptured
type B dissection include absence of chest pain on presen- aortic dissection can usually be inferred only from the pres-
tation, branch vessel involvement, and hypotension. The ence of blood in the mediastinum or pleural or pericardial
electrocardiogram was normal in 31% of patients and spaces in a patient with aortic enlargement but cannot be
showed nonspecific ST and T wave changes in 42%, is- reliably differentiated from the other radiologic entities
chemic changes in 15%, and evidence of acute myocardial that may cause acute aortic syndromes (Fig. 2-61).
infarction in 5% (173). Displaced intimal calcification from the aortic wall is usu-
Of the 427 patients who underwent chest radiography, ally diagnostic of aortic dissection (Fig. 2-61), and, rarely,
12.4% had normal studies; a widened mediastinum was an intimomedial flap itself can be identified with a
the most common finding, seen in 63% of patients with nonopacified aortic lumen in patients with severe anemia
type A and 56% of patients with type B dissections. (187). Calcification of mural thrombus within an
Abnormal aortic contour was seen in 47% of patients with aneurysm and chronic thrombosed aortic dissections may
type A and 53% of patients with type B dissections. be confused with acute aortic dissection, and the use of a
Displaced aortic calcifications were seen in 11% of type thin section contrast-enhanced MDCT is almost always
A dissections and in 18% of type B dissections. Although warranted (146,147). In the updated IRAD study, periaortic
most patients had multiple imaging studies, the initial hematomas were seen in 23% of patients with aortic dissec-
clinical study consisted of CT in 61%, echocardiography in tion at presentation, and their clinical outcomes were sig-
33%, aortography in 4%, and MRI in only 2%. nificantly worse than in patients without periaortic
hematomas, including significantly greater mortality
(33% vs. 20.3%) (188).
Imaging Features Following contrast enhancement, diagnosis of aortic dis-
section rests on identification of two lumens separated by
Computed Tomography Evaluation an intimomedial flap (Fig. 2-62) or a line or strandlike
MDCT is a reliable means for diagnosing or excluding structure that projects into the opacified lumen (Fig. 2-61).
aortic dissection. The diagnosis rests on findings observed There are many configurations of the intimomedial flap,
both on unenhanced and contrast-enhanced scans. some of which can be used as prognostic factors for the
A B
Figure 2-61 Aortic dissection diagnosed on unenhanced CT. A: Axial unenhanced multidetector CT (MDCT) image shows a calcified
intimal flap in the descending thoracic aorta (arrow) and an intramural hematoma of the ascending aorta (arrow). B: Contrast-enhanced
study confirms the chronic dissection in the descending thoracic aorta with the larger false lumen (F) seen on the left side. However, intimal
flap is now seen in the ascending aorta (arrow), consistent with a Stanford type A dissection. Although the descending thoracic dissection is
clearly chronic, the ascending dissection is acute (high-attenuation hematoma) and may have evolved from a type A intramural hematoma.
Note that the angle between the false lumen and its outer wall is acute (arrow), consistent with the beak sign that is only seen in the false
lumen.
A B
C D
Figure 2-62 Acute Stanford type A dissection with resultant ischemic brain injury—multidetector CT evaluation. A–D: Sequential con-
trast-enhanced helical CT scans from the mediastinum to the thoracic inlet demonstrate thin intimal flap in the ascending aorta, arch, and
descending aorta. Intimal flap is present in all three proximal arch vessels (arrows in C), and thrombus is present within both common carotid
arteries. Thrombus in the right common carotid artery (long arrow in D) is crescentic and in the left common carotid artery results in a sharp
interface with the enhanced lumen (short arrow in D). Note also the presence of an endotracheal tube, as the patient was obtunded from
an acute stroke as a result of the dissection. Also note that the anterior true lumen of the descending thoracic aorta (arrow in A) is virtually
collapsed with the “C-shape” appearance.
development of branch vessel ischemia and malperfusion as static when the dissection intersects and narrows the
syndrome (179,189). It is important to remember that aor- vessel origin or as dynamic when the dissection spares
tic dissection and its intimomedial flap is a fluid process the vessel origin but the dissection flap appears to com-
that changes over time with the cardiac cycle and may press the aortic true lumen at or above the origin or covers
appear different in configuration at different anatomic the origin of the aortic branch vessel. Although both
levels. Although this has been well documented with mechanisms may be simultaneously present, Williams
intravascular ultrasound (US), TEE, and MR imaging, non- et al. (179) have demonstrated that the shape of the true
helical and single-detector CT provide for relatively slow lumen may be a marker for the dynamic and potentially
temporal resolution as the thorax is scanned, precluding more severe form of branch vessel ischemia. When the true
a dynamic study of the intimomedial flap. With 16- and lumen is C shaped or concave toward the false lumen (Figs.
64-detector CT the fluid nature of the intimomedial flap 2-62A and 2-63B), there is a higher incidence of a pressure
can be appreciated, especially when evaluating the images deficit in the true lumen and its tributaries, a situation that
in a cine mode. may result in end-organ ischemia (179). Conversely, a flap
Mechanisms responsible for branch vessel compromise that is oriented convex toward the false lumen is usually
with ischemia during aortic dissection are now classified associated with absence of a pressure deficit in the true
lumen. This “malperfusion syndrome” is clinically defined susception (Fig. 2-64). This unusual type of acute aortic
as a complication in an organ system associated with dissection occurs circumferentially with intussusception of
ischemia and resulting in organ dysfunction and systemic the intimomedial flap distally (190). This entity is com-
metabolic abnormalities. This syndrome resulting in monly associated with neurologic deficits and has a higher
branch vessel ischemia is further exacerbated by a large mortality than classic aortic dissection (Fig. 2-65). The char-
entry tear and a small exit tear, resulting in additional true acteristic appearance is a relatively small, circumferential
lumen compression, and is often alleviated with direct sur- intimomedial flap at the aortic root, absence of a flap in the
gical repair (which redirects flow into the true lumen) with ascending aorta, and windsock linear or curvilinear filling
type A dissections and stent-graft closure of the entry tear in defects in the aortic arch (191).
patients with type B dissection. Vernhet et al. (189) recently A more stable configuration, usually seen in chronic
evaluated abdominal CT angiographic findings as a predic- aortic dissections, is absence of flap curvature (straight
tor of postoperative death in patients with acute aortic flap). This appearance of the flap is probably caused by
dissection. Among the 48 patients, postoperative death oc- cellular and mechanical changes in the aorta wall (192).
curred in 4 of 5 patients with clinical mesenteric ischemia, As the flap heals, fibrosis and neointima formation lead
8 of 18 patients with clinical acute renal failure, and 6 of 6 to thickening and rigidity of the flap and sometimes calci-
patients with clinical lower limb ischemia. Although the fication as well (Fig. 2-61).
number of postoperative deaths did not correlate with the
number of dissected peripheral branch vessels, CT demon- Differentiation of True Versus False Lumen
stration of a C-shaped compressed aortic true lumen was Prior to the advent of endograft or stent placement within
63.6% sensitive and 78.4% specific for postoperative death, the aorta or branch vessels, differentiation of true versus
supporting the theory that dynamic compression is poten- false lumen was irrelevant, as it did not change surgical
tially more lethal than static compression (Fig. 2-63) (189). outcome or technique. In acute dissection, resection of the
Additionally, patients with visible evidence of decreased intima in the diseased aorta redirected blood flow to the
organ perfusion on CTA also had a greater risk of death fol- true lumen and usually cured the malperfusion syndrome.
lowing surgery. This clinical study supports both the ex vivo However, operative mortality for acute type B dissections is
and in vivo models described by Williams et al. (179). as high as 25% to 50% and surgery should be performed
Another potential high-risk intimomedial flap appear- only for rupture, severe dilatation, or rapid growth. For
ance is circumferential 360-degree separation of the these patients, and patients with type A dissections with
true lumen, as this may result in intimointimal intus- malperfusion syndrome who are too sick for surgical
A B
Figure 2-63 Dynamic true lumen collapse associated with decreased end-organ perfusion—axial multidetector CT (MDCT) images. A: At
the level of the supraceliac aorta, complete “sandwich” collapse of the true lumen is demonstrated. B: At the level of the kidneys the true
lumen is slit-like and concave toward the ostia of the right renal artery. Note the decreased perfusion of the right kidney due to dynamic
true lumen collapse.
repair, it is important to determine the entry tear, both true

and false lumens, and luminal origins of branch vessels
before, to plan optimal deployment and avoid end-organ
ischemia (192). In most cases, the true lumen may be
identified on CT by its continuity with an undissected por-
tion of aorta (Figs. 2-2 and 2-58). Sometimes, however,
this rule is difficult to apply, particularly in cases with
involvement of the aortic root and especially in those with
circumferential dissection of the root (192). Aortic
cobwebs (ribbons of media that are incompletely sheared
off by the dissection) (193) are seen only in the false
lumen but are so uncommonly seen that the sign is
of limited clinical value as an insensitive indicator of the
false lumen. In a CT study by LePage et al. (192), aortic
cobwebs were present in only 9% of cases of acute dissec-
tions and 17% of chronic dissections. In the same article,
the authors describe the beak sign, the presence of an acute
angle between the dissection flap and the outer wall;
the space formed by the acute angle could be filled with
high-attenuation material (contrast-enhanced blood) or
Figure 2-64 Circumferential 360-degree intimomedial tear with low-attenuation material (hematoma). This sign was seen
“floating” true lumen. Axial multidetector CT (MDCT) image only in the false lumen (100% specificity), and it was
shows a complete 360-degree intimomedial tear with the central present in all cases, both acute and chronic (192). The
true lumen enhancing to a greater degree than the surrounding
false lumen. This flap configuration places the patient at risk for
beak sign (Fig. 2-61B) is the cross-sectional imaging
potentially lethal intimointimal intussusception.
A B
Figure 2-65 Early intimointimal intussusception associated with type A aortic dissection and neurologic deficits. A: Oblique sagittal max-
imum intensity projection image from gadolinium-enhanced three-dimensional angiogram shows narrowing of the innominate artery and
right carotid artery occlusion. B: Electrocardiogram-triggered black blood (double inversion) half-Fourier single-shot turbo spin-echo image
demonstrates early intimointimal intussusception (arrow).
manifestation of the wedge of hematoma that cleaves a Recently, Kapoor et al. (194) described the “intimome-
space for the propagating false lumen and that is present dial rupture sign” as a specific finding to differentiate the
microscopically in all dissections. Outer wall calcification true from false lumen as well as to detect the primary entry
always indicates the true lumen on scans of acute dissec- tear. This sign is a result of direct visualization of an entry
tions. In chronic dissections, however, this finding is tear as it “erupts” into the false lumen with the free edges
somewhat unreliable. The true lumen is almost always of the dissection flap uniformly pointing toward the false
smaller than the false lumen, is usually enhanced to a lumen (Fig. 2-66). This sign was seen in 8% (5 of 59) of
greater degree (especially with faster scan times), and patients in this series (194).
rarely contains thrombus. False lumen thrombus is signifi- There is a strong emphasis in the radiologic literature
cantly more frequent in chronic dissections than acute dis- on the detection of intimal flap extension into the great
sections. Thrombus formation in acute dissection is due vessels (2). Although the presence of arch vessel exten-
partly to the thrombogenic exposed media, which lines sion may be pertinent to the management of the small
the false lumen, and stasis in low-flow segments of the number of patients with type A dissection who present
false lumen. In cases of chronic dissection, the false lumen with acute neurologic deficits (Figs. 2-62 and 2-65),
remains prone to thrombus formation because of stasis in patients who are otherwise neurologically intact, this
related to aneurysmal enlargement and because of athero- information will usually not alter surgical management.
matous changes in the neointima, which can outstrip Because concomitant replacement of the aortic arch in
atheromatous degeneration in the native intima (192). patients with acute type A dissection is associated with
From a geographical perspective, the false lumen is usually higher mortality rates than simple ascending aortic graft
seen in the right side of the ascending aorta (Figs. 2-2 and replacement, it is usually reserved for patients whose pri-
2-58) and the left side of the descending aorta (Figs. 2-61 mary intimal tear is in the transverse arch or those with
and 2-63). In the aortic arch, the true lumen may be arch extension, resulting in weakening of the outer wall,
surrounded by the false lumen on both sides. fragmentation of the inner layer, aneurysmal dilatation,
A B
Figure 2-66 Ruptured Stanford type A dissection with direct visualization of the entry tear. A: Axial multidetector CT (MDCT) image
shows an ascending aortic dissection with primary entry tear (solid arrow) seen with the larger true lumen and the free edges of the dissec-
tion flap pointing toward the false lumen. This has been called the “intimomedial rupture” sign (194). Note the presence of mediastinal
hematoma (open arrow). As the ascending aorta and pulmonary artery share a common adventitia, hematoma that enters the adventitial
layer of the ascending aorta may extend along the right pulmonary artery with subsequent compression (158). B: Sagittal reformatted image
shows both the true (T) and false (F) lumen as well as relative sparing of the aortic root and extension of aneurysmal disease into the arch
(arrow), likely requiring hypothermia and circulatory arrest for repair.
and/or transmural rupture (128,195–197). However, Magnetic Resonance Imaging Evaluation

a more recent article (198) has advocated routine arch Optimal evaluation of the aorta for suspected dissection
resection (and subsequent more extensive resection of with MR mandates imaging from the aortic root to the
intima) for patients with ascending aortic dissections to iliac vessels to define the presence and location of intimal
decrease the need for a second operation due to a patent flap, to assess overall aortic diameter and branch vessel
false lumen with secondary aneurysmal dilatation, although patency, and to evaluate for the presence of intramural,
this remains controversial. mediastinal, pericardial, and pleural hematomas. Evalua-
tion of aortic valvular competence is readily achieved
with cine GRE MR; however, virtually all patients who
Results of Imaging Studies of Aortic Dissection present to tertiary care centers with suspected acute dissec-
Numerous studies have documented a role for conventional tion and valvular incompetence will undergo TEE for this
CT and MDCT in assessing aortic dissection (2,21,199– purpose.
202). In 1993, Nienaber et al. (199) reported a sensitivity of Differentiation of true from false lumen can be accom-
93.8% and a specificity of 87.1% using nonvolumetric plished with many techniques, as previously described.
technique. Sommer et al. (2) compared the efficacy of Morphologically, the true lumen is generally smaller, is oval
single-slice spiral CT with TEE and MR imaging in 49 symp- in configuration, hugs the inner curvature of the aorta,
tomatic patients with clinically suspected aortic dissection; and rarely harbors thrombus, whereas the false lumen is
these authors found the sensitivity of all three to be 100%, usually larger, is crescent shaped, extends along the outer
with corresponding specificities of 100% and 94% for spiral curvature of the aorta, and often contains thrombus (29).
CT compared with both TEE and MRI, respectively (2). The true lumen usually demonstrates faster flow (or greater
As all of these modalities are excellent for the diagnosis of enhancement after the administration of contrast)—this
noncoronary acute aortic syndromes, for ethical reasons it is can be demonstrated with cine GRE, phase-contrast, or
unlikely that a similar comparative study among MDCT, gadolinium-enhanced 3D MRA (Fig. 2-58) (29). However,
TEE, and MRI will be performed. More recently, Yoshida only phase-contrast (203) or time-resolved MRA (with a
et al. (202) demonstrated 100% accuracy with single-slice temporal resolution of under 3 seconds) can demonstrate
helical CT in 57 patients with surgically confirmed type the presence of retrograde flow in the false lumen, if pres-
A aortic dissection (n 45) or IMH involving the ascending ent. It is critical to remember that the images obtained
aorta (n 12). The sensitivity, specificity, and accuracy, from balanced SSFP techniques do not represent flow
respectively, were 82%, 100%, and 84% for detecting an per se, but rather the inherent T2 and T1 of blood, and
entry tear; 95%, 100%, and 98% for arch branch vessel interpretations of these images in acute dissection require
involvement; and 83%, 100%, and 91% for pericardial caution.
effusion. In 2006, Hayter et al. (21) demonstrated 100% Gadolinium-enhanced 3D MRA has been shown to be
sensitivity and specificity for the diagnosis of aortic dissec- both sensitive and specific for the diagnosis of aortic
tion with MDCT. dissection but is insensitive to IMH and extraluminal
The most common region of the aorta to result in a pathology (5,31). Therefore, black blood techniques
false-positive diagnosis at CT is at the aortic root, usually should always be employed in the setting of acute dissec-
due to motion artifact. When evaluating the aortic root for tion to evaluate the aortic wall and pleural and pericardial
suspected dissection with MDCT, motion of the aortic wall spaces for evidence of hemorrhage. The use of breath
between end diastole and end systole may simulate an holding vastly improves the image quality of gadolinium-
aortic dissection. This curvilinear motion artifact is com- enhanced 3D MRA, especially when evaluating the aortic
monly seen when using a 360-degree linear interpolation root for dissection; it is now possible to demonstrate from
algorithm. Clearly there is a learning curve whereby a more which lumen the native coronary arteries, or bypass grafts,
experienced reader can dismiss pulsation artifacts or valve originate (Fig. 2-58). In addition, as with CT, small intimal
leaflets at the root as a normal variant. Examination of the flaps may be missed with gadolinium-enhanced 3D MR
adjacent pulmonary artery for pulsation on the same image imaging in the ascending aorta when the vessel is enlarged
can be helpful. These artifacts occur less commonly with and turbulent flow is present. ECG-gated cine MR images
64-slice CT scanners but may still be present to a lesser de- are often helpful under these circumstances (Fig. 2-67).
gree. The use of prospective or retrospective ECG triggering The interpreting radiologist should be aware of certain
will eliminate these artifacts but results in a much higher artifacts with gadolinium-enhanced MRA that can result in
radiation dose and is impractical to perform on every a false-positive diagnosis of aortic dissection. Differen-
patient in a busy emergency department. In the authors’ tiation of these artifacts from true dissection can be accom-
clinical practice, the test of choice to further evaluate the plished by analysis of the geometry of the suspected
aortic root in cases of indeterminate nongated MDCT intimal flap: In aortic dissections the intimal flap virtually
examinations is to perform MR imaging in stable patients always maintains a spiral course, whereas ghosting and
and TEE or ECG-gated MDCT in acutely ill patients, pro- ringing artifacts will not change configuration over their
vided there are no contraindications to either study. craniocaudad course (Fig. 2-68). On SE MR images, the
A B
Figure 2-67 Chronic aortic dissection near the aortic

root seen only on cine MR images in the setting of cystic
medial necrosis and aneurysmal dilatation. A: Oblique
sagittal black blood half-Fourier single shot turbo-SE
(HASTE) image performed with electrocardiogram trigger-
ing (TR/TE 800 ms/42 ms) and oblique sagittal maximum
intensity projection image from breath-hold gadolinium-
enhanced three-dimensional MR angiogram (B) demon-
strate the tulip-shaped aortic root characteristic of cystic
medial necrosis. Note the smooth tapering to a relatively
normal aortic arch. C: Coronal cine gradient-echo MR
shows circumferential intimomedial tear just above the
level of the aortic valve (thick arrows). The patient had no
symptoms, and this was an incidental finding at surgery, also
not diagnosed prospectively with transesophageal echocar-
diography. Note the origin of the left main coronary artery
(thin arrow) is well visualized and arises from the aneurysm,
C necessitating reimplantation during surgical repair.
Figure 2-68 Pseudodissection in a patient with aortic coarctation from

central line artifact with breath-hold gadolinium-enhanced MR angiography
(MRA). A: Oblique sagittal maximum intensity projection image from breath-
hold gadolinium-enhanced three-dimensional MRA in a patient with aortic
coarctation demonstrates a linear structure in the center of the ascending
and descending aorta. This does not spiral, unlike an intimal flap from an
aortic dissection. This artifact likely results from acquiring the central lines of
k-space during a period of rapidly rising arterial concentrations of gadolin-
ium. B: Axial cine GRE image through the ascending and descending aorta,
A performed immediately after A, is normal.
superior pericardial recess should not be confused with a patients who underwent operative repair for type A dissec-
proximal aortic dissection (Fig. 2-69). tion was 26% and for type B dissection was 31.4%. The
majority (80%) of patients with type B dissection were
treated medically, with a mortality rate of 10.7%; surgery
Natural History and Treatment from the
was performed in the remaining 20% of cases. When re-
International Registry of Acute Aortic ported, the most common causes of death in patients with
Dissection type A dissection were aortic rupture (Fig. 2-70) or cardiac
Left untreated, almost all patients with aortic dissection tamponade (41.6%) and visceral ischemia (13.9%). Aortic
will die within 3 months, especially with involvement of rupture (38.5%) and visceral ischemia (15.4%) were the
the ascending aorta. Despite improved diagnostic and most common causes of death in patients with type B dis-
therapeutic techniques, overall in-hospital mortality in the section (173). In a more recent update, the three greatest
IRAD study was 27.4% (173). The mortality rates among risk factors for in-hospital death in patients with aortic
A B
Figure 2-69 Pseudodissection––the superior pericardial recess. A, B: Axial and oblique sagittal electrocardiogram-triggered SE MR
images demonstrate a prominent superior pericardial recess (curved arrow in A and straight arrow in B). This normal configuration should
not be confused with an intimal flap.
A B
Figure 2-70 Ruptured type A dissection with extensive mediastinal hematoma. A, B: Two axial multidetector CT (MDCT) images show an
intimal flap within a dilated ascending aorta. The true lumen on the left is smaller, higher in attenuation, and contiguous with the aortic out-
flow tract. Note that the hematoma dissects posteriorly to the bronchus intermedius (arrow in A).
type B aortic dissection were shock, absence of chest or

back pain at presentation, and branch vessel involve-
ment/malperfusion syndrome (204). In another study,
death was also more common in patients older than
70 years (205). In patients who underwent surgery for type
A dissection, risk factors associated with higher mortality
included history of aortic valve replacement, migratory
chest pain, hypotension as a presenting symptom, shock
or tamponade, and preoperative limb ischemia (206).
Penetrating Atherosclerotic Ulcer
Clinical Features
Penetrating atherosclerotic ulcer [also referred to as ulcer-
like projection (ULP)] is a distinct radiologic/pathologic
entity in which ulceration penetrates the internal elastic
lamina into the aortic media and is invariably associated
with secondary IMH, pseudoaneurysm, and, rarely, aortic
rupture (207–217). It is readily differentiated from primary
aortic IMH, which has an intact intima, by the ulcer extend-
ing beyond the aortic lumen. Extension of the ulcer beyond
the expected margin of the aortic wall and the presence of
a localized hematoma cap over the ulcer help to differenti-
ate this entity from the common atheromatous ulcer (158).
Although the clinical symptoms of chest or back pain
may be identical to those seen in classic aortic dissection, Figure 2-71 Incidental, asymptomatic aortic ulcers. Oblique
most penetrating ulcers are diagnosed in asymptomatic sagittal maximum intensity projection image from breath-hold
gadolinium-enhanced three-dimensional MR angiogram demon-
patients who undergo aortic imaging for other reasons (Fig. strates a saccular arch aneurysm (curved arrow) and concomitant
2-71). Unlike classic aortic dissection or IMH, penetrating penetrating atherosclerotic ulcers (arrows).
ulcers usually occur in elderly patients with severe systemic (older than typical dissection patients), and comorbidi-
atherosclerotic disease and typically occur in the descending ties included hypertension in 97 (92%), tobacco use in
thoracic aorta, although they often occur in the aortic arch 81 (77%), and coronary artery disease in 48 (46%). Of
(Fig. 2-72) and, rarely, in the ascending aorta (Fig. 2-73). nonoperated patients with follow-up studies, the mean
In one large series of 105 patients with penetrating thickness of the IMH decreased at 1 month in 89%
atherosclerotic ulcers the descending aorta was involved and completely resolved at 1 year in 85%. There were
in 94%, with the remaining ulcers occurring in the 3 deaths (4%) within 30 days among 76 patients treated
ascending aorta, arch, or a combination of vascular terri- medically and 6 deaths (21%) among 29 patients treated
tories (212). However, as many patients with ascending surgically. Failure of medical therapy defined as surgery
aortic and arch ulcers will die of aortic rupture prior to or death was predicted by rupture at presentation (odds
hospitalization, involvement of the proximal aorta is ratio 20.6) and era of treatment (before 1990) but
clearly underrepresented. Eighty-five patients had a con- not by aortic diameter, ulcer size, or extent of hema-
comitant IMH and 20 did not. The mean age was 72 years toma (212).
A B
C D
Figure 2-72 Penetrating atherosclerotic ulcer of the undersurface of the aortic arch with extensive intramural hematoma and aortic
rupture. A–D: Sequential contrast-enhanced helical CT images from superior to inferior demonstrate a penetrating atherosclerotic ulcer
(arrow in A) associated with extensive aortic intramural hematoma (small arrows in B and C), mediastinal hematoma, and a hemorrhagic
pericardial effusion (curved arrows in B and C). Aortogram (D) confirms the ulcer (arrow) arising from the undersurface of the aortic arch.
A B
Figure 2-73 Aortic rupture from penetrating ulcer of the ascending aorta. A: Axial multidetector CT (MDCT) image at the level of the
right pulmonary artery shows extensive mediastinal hematoma with mass effect on the right pulmonary artery and a small ascending aortic
hematoma (arrow). B: A more inferior axial MDCT image shows a small ulcer outside the lumen of the posterior wall of the aorta (arrow).
Note the heavily calcified left anterior descending coronary artery; atherosclerotic disease of the aorta and coronary arteries is commonly
seen in patients with aortic ulcers.
Imaging Features the former technique better demonstrates the concomi-

tant IMH and allows evaluation of pleural or pericardial
Computed Tomography Evaluation surfaces, the latter technique can better depict the ulcer
A penetrating ulcer is usually depicted on contrast- crater. Breath-hold 2D or 3D images performed after
enhanced CT as a focal collection of contrast material gadolinium-enhanced aortography can also demonstrate
that projects beyond the confines of the expected aortic enhancement within the wall of a pseudoaneurysm or
lumen (Figs. 2-71 to 2-73) (207). These lesions can be contained rupture (Fig. 2-76). It should be emphasized
isolated or multifocal (Fig. 2-71). An IMH (discussed in that MR remains inferior to CT in detecting displaced inti-
detail later in the chapter) is invariably present (179) mal calcifications because of decreased spatial resolution
and either may be localized to a short segment or, rarely, and relative insensitivity to small amounts of calcium,
may be quite extensive (Figs. 2-72 and 2-73). Displaced which clearly define the aortic hematoma as subintimal
intimal calcifications are often present, which confirm or intramural in nature (Fig. 2-77).
that the hematoma is subintimal or within the media
(Fig. 2-74). An aneurysm or pseudoaneurysm may also Natural History and Therapy
be present (Fig. 2-75).
CT findings that should raise suspicion of aortic rupture The natural history of penetrating atherosclerotic ulcers is
include the presence of high-attenuation mediastinal (Figs. difficult to predict but may result in progressive aortic
2-72 and 2-73) or pleural/pericardial fluid collections. A enlargement, including increasing ulcer size (212). For
retrocrural hematoma may be present as well. Rarely, active these reasons, CT or MR imaging surveillance is neces-
extravasation of contrast material into the mediastinum or sary to depict these abnormalities and to stratify patients
pleural space may be visualized. Under these circumstances who may be at risk for aortic rupture from enlarging
it is often difficult, if not impossible, to differentiate a pseudoaneurysms. In general, the associated limited IMH
ruptured thoracic aneurysm from a penetrating ulcer; will resorb within 1 year (212) (Fig. 2-78), and the ulcer
however, both entities would require immediate surgical will either remain stable or enlarge. The resorption of the
management. hematoma may weaken the aortic wall, and fusiform
(Fig. 2-79) or saccular aneurysms may develop in as many
Magnetic Resonance Evaluation as 48% of patients (212). Quint et al. (217) followed
Optimal MR evaluation of penetrating ulcers requires a 33 aortic ulcers, of which 21 were stable over a period of
combination of black blood imaging and MRA. Although up to 6 years (mean, 18 months). In 10 ulcers, disease
A B
Figure 2-74 Penetrating atherosclerotic ulcer of the distal tho-

racic aorta with displaced intimal calcifications. Unenhanced (A) and
enhanced (B) multidetector CT (MDCT) examinations demonstrate
a small crescentic intramural hematoma (arrow in A) in the descend-
ing thoracic aorta. Note the hematoma is better seen without
contrast, and the displaced intimal calcification is difficult to visual-
ize on enhanced CT (arrow in B). C: The offending atherosclerotic
ulcer is seen more inferiorly, projects outside of the expected aortic
lumen, and demonstrates displaced intimal calcification (arrow) and
C a concomitant intramural hematoma.
A B
Figure 2-75 Acute and chronic aortic arch pseudoaneurysm resulting from penetrating atherosclerotic ulcers. A: Enhanced multidetector
CT (MDCT) image shows large penetrating arch ulcer with acute intramural hematoma (thick arrow). Note the presence of thrombus within
the superior vena cava (thin arrow). B: Contrast-enhanced CT image at the level of the aortic arch demonstrates a pseudoaneurysm with
thrombus resulting from a previous penetrating ulcer (arrow). Note that the right upper lobe bronchus is displaced posteriorly (curved
arrow). This is characteristic of cicatrization atelectasis within the right upper lobe secondary to previous granulomatous infection with
resultant scarring.
151
Figure 2-76 Contained aortic rupture from penetrating atherosclerotic ulcer

demonstrated with non–electrocardiogram-gated (rapid) MR imaging. A: Oblique
sagittal maximum intensity projection image from breath-hold gadolinium-enhanced
three-dimensional MR angiogram demonstrates multiple penetrating ulcers and a
giant ulcer (arrow) resulting in contained aortic rupture in this patient with chest pain.
B: The thrombus and enhancing adventitial tissue (curved arrows) are better demon-
strated on this magnetization prepared, rapid gradient-echo sequence performed
A after the aortogram. (From reference 29, with permission.)
B
Figure 2-77 Subtle intramural hematoma seen on CT but not
seen on MR. A: Axial multidetector CT (MDCT) image examina-
tion demonstrates a small crescentic intramural hematoma with
displaced intimal calcifications in the descending thoracic aorta.
B: Axial black blood half-Fourier single shot turbo-SE (HASTE)
image performed with electrocardiogram triggering shows non-
specific thickening of the descending thoracic aorta that may be
due to atherosclerosis, arteritis, or intramural hematoma. The
decreased spatial resolution and relative insensitivity to calcium
A preclude evaluation of the subtle intramural hematoma.
progression was manifested as one of the following three distal embolization, or rapidly enlarging aneurysm or
pathways: (a) increase in aortic diameter only, without pseudoaneurysm. Patients with penetrating ulcers of the
change in the ulcer (one lesion; 15% increase in aortic ascending aorta associated with IMH should probably un-
diameter); (b) incorporation of the ulcer into the aortic dergo surgery as soon as possible if no contraindications
wall contour, with increase in aortic diameter (two lesions; are present. It should be noted that repair of a penetrating
16% and 45% increase in aortic diameter); or (c) increase ulcer requires more extensive surgery than a Stanford B
in the size of the lesion as well as increase in the aortic dissection, and often long segments of an ulcerated, ather-
diameter (seven lesions; mean increase in lesion size, 39%, osclerotic thoracic aorta need to be excised.
mean increase in aortic diameter, 10%). Demers et al. (218) recently reported the midterm
Although an early report advocated surgical repair of follow-up of aortic stent-graft placement in 26 patients
all symptomatic patients (207), a more conservative with penetrating atherosclerotic ulcer. Fourteen patients
approach has emerged (212), as this cohort of patients (54%) were not candidates for open surgical repair and
often has significant comorbid atherosclerotic disease 23% presented with frank aortic rupture. Three patients
involving the heart, brain, and peripheral vasculature. The died within 30 days and two had an early type I endoleak.
current consensus is that all patients receive aggressive Actuarial survival estimates at 1, 3, and 5 years were
antihypertensive therapy, with stent-graft placement or 85%, 76%, and 70%, respectively, with previous cere-
surgery restricted to patients with hemodynamic instabil- brovascular accident an independent risk factor for
ity, refractory pain, frank or contained aortic rupture, death. Treatment failures were due to progressive aortic
A B
Figure 2-78 Regression of localized intramural hematoma over a 1-year period. A: Axial multidetector CT (MDCT) image at the time of
acute aortic syndrome shows a small intramural hematoma in the descending thoracic aorta with displaced intimal calcifications and a small
sympathetic left pleural effusion. B: Axial MDCT performed 6 weeks later shows resolution of the pleural effusion and decreased size of
intramural hematoma. C: Axial MDCT performed 1 year later demonstrates near complete resolution of intramural hematoma. Without prior
imaging studies for comparison, the minimal residual hematoma could be mistaken for intraluminal thrombus. D: Coronal reformatted image
demonstrates the offending ulcer that was unchanged in size since the acute event (not shown) (continued).
C Figure 2-78 (continued)
A B
Figure 2-79 Intramural hematoma of the aortic arch due to penetrating ulcer with progression to fusiform aneurysm. A: Axial
unenhanced multidetector CT (MDCT) image demonstrates an acute intramural hematoma on both sides of the aortic arch, with displaced
intimal calcifications. B: Enhanced MDCT image shows the etiology as a flask-shaped penetrating atherosclerotic ulcer (arrow). C: Enhanced
MDCT image obtained 6 months later shows resorption of hematoma with new fusiform aneurysm of the aortic arch. Note that the ulcer is
unchanged in size and the pleural effusion has resolved.
There remains considerable clinical and imaging over-

lap among the entities referred to as IMH, penetrating
atherosclerotic ulcer, and limited aortic dissection with
thrombosed false lumen. Nevertheless, CT, MR imaging,
and TEE have greater than 96% sensitivity in the diagno-
sis of IMH (225).
Clinical Features and Natural History

Of 1,010 patients in the IRAD study, 6% had IMH (94%
had classic aortic dissection), and these patients tended to
be older and more likely to have distal (type B) involve-
ment than patients with classic dissection (226). Patients
with IMH described more severe initial pain than did
those with classic dissection but were less likely to have
ischemic leg pain, pulse deficits, or aortic valve insuffi-
ciency (226). Overall mortality of IMH was similar to that
of classic dissection (20.7% vs. 23.9%), as was mortality
in patients with IMH of the descending aorta (8.3% vs.
13.1%) and the ascending aorta (39.1% vs. 29.9%) com-
pared with classic aortic dissection. Among the 51 patients
whose initial diagnostic study showed IMH, 8 (16%)
C progressed to classic aortic dissection on a serial imaging
Figure 2-79 (continued) study.
Evangelista et al. (227) followed 50 patients with IMH
over an approximately 4-year period with serial imaging.
In the first 6 months, complete IMH regression was
enlargement over time. Similar results were obtained with observed in 14 and progression to aortic dissection in
stent-grafts by Eggebrecht et al. (219); overall survival rates 18 patients; in 14 of these, the dissection was localized and
were 100% at 30 days, 100% at 1 year, 82.5% at 2 years, 12 later developed pseudoaneurysms. At the end of follow-
and 61.9% at 5 years. up, the IMH had regressed completely without dilatation
in 17 patients (34%), progressed to classical dissection in
6 (12%), and evolved to fusiform aneurysm in 11 (22%),
Primary Aortic Intramural Hematoma
saccular aneurysm in 4 (8%), and pseudoaneurysm in 12
(24%). Evolution to dissection was related to echolucency
Pathogenesis
on TEE and to longitudinal extension of IMH. The authors
IMH, or aortic dissection without intimal flap, is a patho- concluded that maximal aortic diameter at presentation
logic entity with risk factors and clinical findings similar to was an important predictor of adverse clinical events,
those of conventional aortic dissection (70,220–223) and including death and pericardial or pleural effusions, and
is seen in 5% to 20% of all patients with acute aortic syn- mediastinal hematoma was seen more frequently in IMH
dromes. Whereas classic dissection usually results from a than in classical dissection and was associated with a
primary intimomedial tear, with subsequent separation of worse prognosis.
aortic wall components, IMHs represent localized hemor- Using CT with close interval follow-up, Sueyoshi
rhage confined to the aortic media (125,175,220,224). et al. (228) evaluated 52 patients with IMH (of which
The etiology of IMH is not well defined; theories include 16 patients had type A and 36 had type B) for the pres-
intimal fracture of an atherosclerotic plaque with propaga- ence of a ULP, diameter and progression of the projec-
tion of blood into the media (penetrating ulcer), sponta- tion, and overall aortic diameter. In 17 (33%) of the
neous rupture of the vasa vasorum with subsequent aortic 52 patients, 17 ULPs were newly identified during
wall weakening (70,220–223), and complete thrombosis the follow-up period. In these 17 patients, 12 (70%) of
of the false lumen in classic aortic dissection, in which an 17 projections progressed to complications, including
intimal tear is not identified by imaging modalities. enlargement (n 10) or progression to overt aortic dis-
Wilson and Hutchins (224) reviewed the autopsy results section (n 2). A significant predictor of progression of
from 204 patients with aortic dissection, and 13% had no a ULP was based on location in the ascending aorta, as
identifiable intimal tear; this negates the theory that these patients with type A IMH had a significantly higher fre-
represent classic dissections in which the intimal flap was quency of new development of ULP than did patients
“missed” by imaging studies. with type B IMH.
The outcome of patients with IMH without imag-

ing documentation of a ULP is different from the outcome
in those with interval development of these lesions (229).
Overall, current predictive factors of disease progression
for patients with primary IMH (without an associated
ulcer or intimal erosion) include involvement of the
ascending aorta, maximum aortic diameter of 50 mm or
greater on initial CT scan, persistent pain, progressive aor-
tic wall thickness, and enlarging aortic diameter. Predictors
of disease progression in patients with IMH and an
associated ULP include interval increase size of associated
pleural effusion, recurrent pain, ulcer located in the
ascending aorta or arch, initial maximum ulcer diameter
of 20 mm or more, and initial maximum ulcer depth
of 10 mm or greater (229).
Using the data from multiple published studies, it
appears that patients with involvement of the ascending
aorta, unless they have significant comorbid disease
or hematoma thickness less than 1 cm, should undergo
early surgical intervention to prevent potentially fatal
complications, which include aortic rupture, pericardial
tamponade, and conversion to classic dissection (230).
Stable patients with type A IMH and hematoma thick-
Figure 2-80 Primary aortic intramural hematoma (not secondary
ness less than 1 cm who are moderate to poor surgical to ulcer) involving the entire aorta. Unenhanced multidetector CT
candidates may be followed up with serial CT or MR (MDCT) image shows crescent of blood in the wall of the ascending
examinations during the acute hospitalization for early (A) (top arrow) and descending (bottom arrow) (D) aorta. Note the
presence of a hemopericardium (H). Note the absence of mass
regression (231). Patients with type B IMH have better effect on the aortic lumen, which explains why the malperfusion
long-term prognosis than patients with classic aortic syndrome almost never occurs in patients with primary intramural
dissection, with older age and appearance of a ULP dur- hematoma. Ascending aortic involvement (type A), especially when
the ascending aorta is greater than 5 cm and the hematoma is
ing serial imaging predictive for disease progression thicker than 1 cm, is an absolute indication for surgery.
(232). However, because the natural history of these
lesions is still in evolution, all patients should undergo
routine serial imaging at least every 3 months during the
first year to evaluate for spontaneous progression to clas-
sic dissection, early saccular aneurysm, or pseudo-
aneurysm formation. Long-term follow-up is required for enhance, that it appears hypodense when compared with
the detection of fusiform aneurysms, which may develop the enhanced lumen, and that a recognizable intimal flap is
in the cohort of patients without ULPs. absent (Fig. 2-81). Absence of enhancement is helpful
when differentiating IMH from active arteritis, as the latter
will demonstrate mural enhancement and IMH will not.
Contrast is clearly necessary to detect a penetrating athero-
Imaging Features sclerotic ulcer as the etiology of an IMH and to detect ULPs
that may be seen on the initial or follow-up examinations.
Computed Tomography Evaluation However, the IMH secondary to a penetrating atheroscle-
The CT appearance of acute IMH is characteristic. Non- rotic ulcer is almost always localized, as atherosclerotic
contrast examination demonstrates a crescent-shaped area disease will limit propagation, whereas those due to IMH
along the wall of the aorta that has higher attenuation than are diffuse.
blood, without mass effect on the patent lumen (Fig. 2-80) Once a ULP is discovered in an asymptomatic patient,
(70,223,228). Intimal calcifications may be displaced by serial imaging is recommended every 3 months to evaluate
the hematoma but are much less common in the ascending for progression to classic dissection, enlarging pseudo-
than descending aorta. Pericardial (Fig. 2-80) and pleural aneurysm, or fusiform aneurysm. IMH may be differenti-
effusions are common and seen more often in type A IMH ated from a chronic thrombosed classic aortic dissection in
than classic dissection, possibly related to the fact that the that the latter may extend into the iliac arteries, whereas
cleavage plane in IMH is closer to the adventitia than in IMH will not and there are no “straight lines” in IMH,
classic dissection (233) (Fig. 2-52). Contrast-enhanced CT which explains the relative low frequency of end-organ
examinations demonstrate that the hematoma will not malperfusion.
A B
C D
Figure 2-81 Primary aortic intramural hematoma involving the entire aorta in three patients. A: Unenhanced multidetector CT (MDCT)
image shows crescent of blood in the wall of the ascending and descending aorta consistent with type A intramural hematoma. Both pleural
and pericardial effusions are present. B: On the enhanced MDCT image, the hematoma now appears hypodense relative to the enhanced
lumen. Similar findings are seen in a second patient (C) with the presence of a small ulcerlike projection (ULP) (arrow). D: Axial MDCT in
a third patient with ruptured ascending aortic intramural hematoma demonstrates a small ULP (arrow). The interval development of a ULP in
a patient with a primary intramural hematoma prognosticates a higher incidence of rupture, conversion to classic dissection, or progression
to pseudoaneurysm. It is thought that this ULP may represent the earliest sign of an entry tear.
With IMH, unlike classic dissection, the presence of a to confuse fluid in the superior pericardial recess with type
hemopericardium is not tantamount to rupture but may A IMH (Fig. 2-83).
represent blood weeping across the wall of a diseased aorta.
The subadventitial location of IMH may explain this pheno- Magnetic Resonance Evaluation
menon. Classic signs of rupture include active extravasation The MR appearance of IMH depends both on the age of
of contrast, mediastinal hematoma, and, usually, hemor- the hematoma and on the pulse sequence used. The
rhagic pleural effusions (Fig. 2-82). Care must be taken not morphology consists of focal crescentic wall thickening
A B
Figure 2-82 Ruptured primary intramural hematoma of the

aortic arch and descending thoracic aorta. A–C: Three enhanced
MDCT images demonstrate an intramural hematoma of the aortic
arch and descending thoracic aorta, with extensive mediastinal
hematoma (surrounding the trachea in A) with associated high-
attenuation pleural and pericardial effusion (curved arrow). Note
the presence of multiple ULPs (arrows). These ULPs are well seen
(arrows) on the aortogram (D) and have been shown to represent
small entry tears associated with intercostal artery origins. This
image has also been dubbed the “Chinese ring-sword sign” (234). D
without mass effect on the aortic lumen or evidence of MRA (Fig. 2-85). In acute and subacute IMH, there is
intimal flap (Figs. 2-84 to 2-86) (220). good correlation between the findings of MDCT and
The signal characteristics reflect the age of the those of MR imaging (Fig. 2-87).
hematoma; acute IMH (symptoms 7 days) is usually Both fat-suppressed 3D T1-weighted and magnetiza-
isointense to muscle on ECG-gated T1-weighted spin- tion-prepared, gradient-echo sequences performed after
echo or double IR images (Fig. 2-84), whereas subacute gadolinium-enhanced MRA will usually demonstrate IMH
IMH (symptoms 7 days) is usually hyperintense from as a crescentic area of relatively low signal intensity when
methemoglobin (Fig. 2-86). These black blood images are compared with adjacent gadolinium-enriched blood.
essential for the diagnosis because, as previously men- However, on unenhanced VIBE and LAVA MR imaging, the
tioned, IMH may be occult on gadolinium-enhanced signal may be hyperintense (Fig. 2-88).
A B
Figure 2-83 Aortic intramural hematoma: pitfalls in multidetector CT (MDCT) diagnosis. A: Enhanced MDCT shows motion artifact
typically seen in the ascending aorta and fluid in the anterior superior pericardial recess. Region-of-interest interrogation demonstrated
simple fluid. B: In type A intramural hematoma on enhanced MDCT, note involvement of greater than 50% of aortic circumference, mild
mass effect on the superior vena cava, and involvement of both ascending and descending aorta (arrows). Region-of-interest measurements
demonstrated blood attenuation.
A B
Figure 2-84 Acute intramural hematoma involving the ascending aorta (type A)—MR demonstration. A, B: Axial electrocardiogram-
triggered T1-weighted SE MR images demonstrate acute aortic intramural hematoma involving both the ascending and descending aorta
(arrows in A and curved arrows in B). Note that the hematoma is isointense to skeletal muscle in both patients.
A B
C D
Figure 2-85 Subacute intramural hematoma—MR demonstration. A and B: Axial and oblique sagittal electrocardiogram black blood
half-Fourier single shot turbo-SE (HASTE) images demonstrate subacute aortic intramural hematoma involving both the ascending and
descending aorta (arrows in A and B). Note the high signal intensity of the subacute hematoma, the crescentic shape, and the absence of mass
effect on the aortic lumen. Oblique sagittal source (C) and maximum intensity projection (D) images from breath-hold gadolinium-enhanced
three-dimensional MR angiogram fail to demonstrate the hematoma. This case demonstrates the importance of obtaining black blood images
in the setting of an acute chest syndrome. Note the presence of an aberrant right subclavian artery (long arrow in D) and the common trunk
of both common carotid arteries (short arrow in D).
A B
Figure 2-86 Evolution of signal intensity in type B intramural hematoma demonstrated on double inversion recovery MR imaging.
A: Axial electrocardiogram-gated black blood half-Fourier single shot turbo-SE (HASTE) image performed at the time of acute aortic syn-
drome demonstrates acute intramural hematoma of the descending thoracic aorta as isointense to skeletal muscle (arrow). B: Repeat study
in 1 week shows evolution to high signal intensity methemoglobin (arrow). This crescent of high signal intensity may be present for up to
3 months following the acute event.
A B
Figure 2-87 Descending propagation of type A intramural hematoma on CT and MR imaging. A: Enhanced multidetector CT (MDCT)
image obtained 1 week after the onset of chest pain shows a crescent of blood in the ascending and descending aorta consistent with type
A intramural hematoma (arrows). The ascending component measures 20 HU, consistent with the subacute phase, and the descending
hematoma measures 60 HU. B: Similar findings are seen on axial electrocardiogram-gated black blood half-Fourier single shot turbo-SE
(HASTE) images; the subacute ascending component (top arrow) is hyperintense, and the more acute descending component is isointense
to muscle (bottom arrow).
Figure 2-88 Subacute type B intramural hematoma

(IMH): evaluation with fat-suppressed three-dimensional
(3D) breath-hold pulse sequences. A: Axial unenhanced vol-
umetric interpolated breath-hold examination (VIBE) se-
quence shows high signal crescent of blood consistent with
descending aortic intramural hematoma. B: Note the “flip
flop” on imaging performed 5 minutes after A in which the
IMH now appears hypointense compared with the enhanced
lumen. C: Dynamic arterial phase gadolinium-enhanced 3D
MR imaging performed with VIBE sequence shows long seg-
ment type B intramural hematoma (I) isointense to skeletal
muscle. There is adventitial enhancement (upper arrow) that
should not be confused with atelectatic lung. Note the pres-
ence of an ulcerlike projection (lower arrow) and aortic root
morphology typical of cystic medial disease. C
Figure 2-89 Ulcerlike projection (ULP), with rapid pro-

gression to pseudoaneurysm formation. A: Oblique sagittal
maximum intensity projection image from breath-hold
gadolinium-enhanced three-dimensional MR angiogram
(MRA) demonstrates a small ULP (arrow). This occurred
3 months after an intramural hematoma (not shown).
B: Coronal MIP image from breath-hold gadolinium-
enhanced three-dimensional MRA, obtained 18 months after
A, demonstrates rapid progression to large saccular
A, B pseudoaneurysm (arrow). This was surgically confirmed.
As with CT, a ULP may be readily apparent (Fig. 2-89) injuries occur in the region of the aortic isthmus, the por-
and serial imaging is recommended every 3 months to tion of the descending thoracic aorta between the origin of
evaluate for progression to classic dissection, enlarging the left subclavian artery and the site of attachment of the
pseudoaneurysm, or fusiform aneurysm. ligamentum arteriosum (Fig. 2-9) (239). Horizontal decel-
eration results in shearing forces at the aortic isthmus,
the junction between the relatively mobile aortic arch and
AORTIC TRAUMA the fixed descending aorta (237). Bending stress may also
occur as the aorta is flexed over the left pulmonary artery
and left mainstem bronchus (240,241). Some suggest that
Clinical Features
aortic isthmus injuries result from an osseous pinch; the
The majority of aortic trauma results from blunt trauma aorta is squeezed between the anterior bony thorax and
caused by high-speed motor vehicle accidents (235), plane the spine from anterior compression.
crashes, and falls from great heights. Injury to the thoracic The ascending aorta is the second most common site of
aorta or arch vessels is present in approximately 15% of injury and is involved in 5% to 20% of cases (238,239).
road-traffic deaths and as such represents a major cause These injuries are invariably fatal because of associated
of out-of-hospital mortality (236–238). Rupture of the cardiac tamponade, aortic valve rupture, and coronary
thoracic aorta is immediately lethal in 75% to 90% of artery compromise. Injuries to the ascending aorta may
cases (236,238). Of the patients with thoracic aortic injury result from sudden vertical deceleration, which displaces
who reach the hospital alive, 60% to 70% will survive the heart inferiorly and into the left pleural cavity. This
if appropriately treated. Interestingly, those who survive torsion results in traction on the ascending aorta just
until hospitalization do not usually die from their aortic above the aortic valve. Another hypothesis is the water-
injury but from concomitant trauma to other vascular or hammer effect, caused by a rapid increase in intra-aortic
organ injuries. This observation has led to delayed repair pressure, which may lead to rupture into the pericardium
or stent-graft placement after other life-threatening injuries (241). Rarely, aortic injury may occur at the level of the
are repaired. diaphragmatic hiatus.
The mechanism of thoracic aortic injury results from These injuries may result in a spectrum of pathology from
deceleration or traction. Ninety percent of thoracic aortic benign intimal tears with medial hematomas to complete
Figure 2-90 Traumatic descending aortic rupture with active

extravasation of contrast and intra-aortic air from aortobronchial fis-
tula. A: Axial image from contrast-enhanced multidetector CT
(MDCT) demonstrates active extravasation of iodinated contrast
into the mediastinum from traumatic aortic injury (arrow). Note
displacement of the nasogastric tube to the right of the spine from
the mediastinal hematoma. B: Axial image at a more inferior level
demonstrates intra-aortic air (arrow) from aortobronchial fistula (not
shown). C: Sagittal reformatted image shows isthmus transection
(arrow) proximal to the long segment aortic injury and extension of
B periaortic hematoma into the abdominal aorta.
aortic transection and/or arch vessel avulsion (238). Dissec- tion, and ratio of mediastinal width to chest width (245).
tion, hemorrhage, and vascular thrombosis may subse- Although a normal chest radiograph can never exclude an
quently occur. Traumatic pseudoaneurysms develop when aortic injury, its negative predictive value is 96% for supine
the intima and media are disrupted and the adventitia films and 98% for erect films (246).
bulges outward but remains intact (238). Thrombi usually
form within these false aneurysms and can serve as a nidus
Computed Tomography Evaluation
for embolization. These pseudoaneurysms tend to expand
over time and may cause mass effect on nearby structures,
and Findings
fistulize to adjacent organs (Fig. 2-90), and may sponta- Contrast-enhanced MDCT has become an important test in
neously rupture (242). the algorithm of the diagnosis of traumatic aortic rupture
because a completely normal examination in a stable
patient virtually excludes a significant injury (247–250).
This in turn has dramatically reduced the number of
Imaging Features
aortograms performed. Mirvis et al. have demonstrated that
reliance on findings at admission CT rather than chest radi-
Radiographic Evaluation
ography resulted in a $365,000 savings owing to the elimi-
In most patients, the initial radiographic examination will nation of unnecessary aortograms.
be a chest radiograph, usually a supine portable film. The CT signs of aortic injury are best classified as direct
most important findings are a loss of the normal aortic or ancillary. Direct findings include active extravasation of
contour with widening of the mediastinum, tracheal devi- contrast material (Fig. 2-90), pseudoaneurysms, intimal
ation, deviation of a nasogastric tube to the right of the T4 flaps (Fig. 2-91), abrupt caliber change, pseudocoarctation,
spinous process, and depression of the left mainstem and focal occlusion of segments of the aorta or arch vessels.
bronchus (241). If the patient is hemodynamically unsta- Indirect signs are those that are supportive of the diagnosis,
ble with an abnormal chest radiograph, either MDCT or but not specific, and include mediastinal and or retrocrural
surgical exploration may be performed. This will vary by hematomas, which are more likely to be caused by a bony
the clinical scenario, the suspicion of other life-threatening fracture or a venous injury than an aortic injury. However, if
injuries, and the treating institution’s trauma protocol. If the hematoma is periaortic and obliterates the fat planes
the patient is stable, an erect posteroanterior chest radi- surrounding the aorta, the specificity is markedly increased
ograph should be performed. to 77% (251). Wong et al. (252) recently described the
The presence of mediastinal abnormalities has a diag- clinical significance of periaortic hematoma near the level
nostic sensitivity of 90% to 95% for a thoracic arterial of the diaphragm using MDCT in patients referred for
injury but a low specificity (5% to 10%) and positive pre- traumatic aortic injury. Among the 97 patients with medi-
dictive value (10% to 15%) (243,244). The specificity can astinal hematoma, 14 had both periaortic hematoma near
be markedly improved when using a formula including the level of the diaphragm and aortic injury, 6 had aortic
three critical signs—loss of aortic contour, tracheal devia- injuries without periaortic hematoma, 5 had periaortic
A B
Figure 2-91 A, B: Traumatic aortic injury at the level of the isthmus. Axial image from contrast-enhanced helical CT demonstrates subtle
intimal flap (black arrows) and extensive mediastinal hematoma (white arrows). Image is grainy owing to the use of detail (bone) algorithm.
hematoma near the level of the diaphragm without Scaglioni et al. (256) used helical CT to evaluate 1,419
aortic injury, and 72 had no evidence of periaortic hema- consecutive patients with suspected thoracic aortic injury.
toma near the diaphragm and no aortic injury. Sensitivity Direct CT findings considered diagnostic of aortic injury
for periaortic hematoma near the level of the diaphragm as included intimal flap, pseudoaneurysm, contour irregular-
a sign of aortic injury was 70%; specificity, 94%; positive ity, lumen abnormality, and extravasation of contrast
predictive value, 74%; and negative predictive value, 92%. material. On the basis of these direct findings, no further
Similarly, Curry et al. (253) demonstrated that the presence diagnostic investigations were performed in 23 patients.
of periaortic hematoma on abdominal MDCT (Fig. 2-90) Isolated mediastinal hematoma on CT scans was considered
was also predictive of thoracic aortic injury. an indirect sign of injury; in these cases, thoracic aortogra-
Mirvis et al. (250) used contrast-enhanced CT to evaluate phy was performed even if CT indicated a normal aorta.
677 patients with positive or equivocal findings at chest Among the 23 patients with direct CT signs, acute tho-
radiography. For aortic injury and mediastinal hemorrhage, racic aortic injuries were confirmed at thoracotomy in 21.
respectively, specificity for traumatic aortic injury was 99% Two false-positive cases were observed. The 54 remaining
and 87% and sensitivity was 90% and 100%. Therefore, the patients had isolated mediastinal hematoma without aortic
CT finding of mediastinal hemorrhage alone was sensitive injuries at CT and corresponding negative angiograms. The
for traumatic aortic injury, but the findings of direct aortic authors concluded that careful evaluation of potential arti-
injury were more specific. Gavant et al. (254) prospectively facts from pulsatility is necessary to avoid a false-positive
evaluated 1,518 patients with blunt chest trauma; 21 of these result and that patients with direct findings do not require
patients had aortic injuries. Helical CT was more sensitive angiography. The use of ECG-gated 16- and 64-slice MDCT
than aortography (100% vs. 94.4%) but less specific should eliminate false-positive results and provide detailed
(81.7% vs. 96.3%) in detection of aortic injuries in patients evaluation of the brachiocephalic vessels, thus overcoming
who underwent both examinations (254). Potential causes a severe limitation of single detector CT.
of false-positive or indeterminate CT examinations were Iatrogenic causes of aortic injury include aortic cross-
motion artifacts, prominent mediastinal vessels, and pro- clamp injuries (Fig. 2-92), pseudoaneurysms from aortic
truding aortic atheromata. The same authors used CTA to cannulation sites, and catheter-induced dissection. In
evaluate traumatic aortic disease in 3,229 patients and the chronic phase, traumatic isthmic tears may result in
demonstrated that MPR and MIP reconstructions offered no enlarging pseudoaneurysms, some of which may calcify.
additional diagnostic information when compared with the
axial source images (255). However, the use of overlapping
Magnetic Resonance Evaluation and Findings
reconstructed axial images was helpful in depicting subtle
intimal injuries confined to the distal aortic arch above the Traditionally, MR imaging had little role in the evaluation
level of the pulmonary artery (255). of patients with acute aortic injury and was reserved for
A B
Figure 2-92 Iatrogenic aortic cross-clamp injury during coronary artery bypass surgery. A, B: Sequential MDCT images show contrast
outside of the ascending aorta (arrows in A) and traumatic intramural hematoma (arrow in B) at the site of aortic cross clamping.
patients in the chronic phase. The inability to detect mortality during the mean follow-up of 46 months (260).
displaced intimal calcifications for subtle traumatic IMH is Chronic descending aortic injury with pseudoaneurysm is
problematic, and the calcified nature of some chronic best treated with stent grafting unless symptoms result
pseudoaneurysms may not be readily seen with MR imag- from mass effect; under these circumstances, surgery may
ing (Fig. 2-93). Nevertheless, a recent case report describes be indicated.
the utility of MR for acute transection (257).
ADVANCES IN SURGICAL AND

Treatment
INTERVENTIONAL TECHNIQUES/
Treatment for traumatic aortic injury has evolved rapidly THE POSTOPERATIVE AORTA
over the past 5 years, particularly for aortic isthmic injuries
(258–261). Historically, all patients were surgically explored Advances in the imaging of aortic disease have been paral-
and the aortic injuries repaired urgently, usually with an leled by advances in surgical treatment of many of these
interposition graft placed during aortic cross clamping or diseases (262,263). In addition to diagnosing aortic dis-
under a brief period of circulatory arrest. More recently, it has ease, it is incumbent on the radiologist to provide pertinent
been shown that aortic stent-graft placement has become the information to the referring clinician that may influence
treatment of choice, with a much lower morbidity and operative approach and technique. To that end, the radiolo-
mortality (258–260). Furthermore, it has been shown that gist must be familiar with various operative techniques and
patients with comorbid disease that would prove lethal if understand the pathologic and anatomic conditions that
not otherwise treated immediately can be observed during dictate the operative approach. In addition, it is helpful to
hospitalization with repair or graft placement in a semielec- be aware of common postoperative complications and
tive fashion (260). to know the follow-up imaging schedule most appropriate
A recent study of 64 patients with traumatic aortic to each postoperative situation. This will help the referring
injury, of which 35 patients were treated surgically and clinician in monitoring the patient for disease evolution or
29 patients were treated with stent grafting, demonstrated progression.
surgical mortality and paraplegia rates of 21% and 7%, The ability to induce deep hypothermia with circula-
respectively. No death or paraplegia was observed in the tory arrest has led to profound reductions in morbidity
subset of patients with delayed surgical repair. With stent and mortality in surgical repair of aneurysms and dissec-
grafting, complete exclusion of the pseudoaneurysmal sac tion of the thoracic aorta, especially those that involve
was observed in all patients, with no major morbidity or the aortic arch (264,265), aortic root (263), and heavily
A B
Figure 2-93 MR evaluation of chronic traumatic injury at aortic isthmus. A: Oblique sagittal maximum intensity projection image from
breath-hold gadolinium-enhanced three-dimensional MR angiogram demonstrates large ulcer (arrow) in the isthmic region. This finding is
nonspecific and may represent penetrating atherosclerotic ulcer except for the fact that the remaining visualized aorta is free of plaque.
B: Axial multidetector CT (MDCT) image shows the chronic nature of this injury as a calcified isthmic pseudoaneurysm (arrow).
calcified aortas that cannot be clamped safely. With this matically decreased the incidence of spinal cord ischemia
technique, core body temperature is cooled to below are based on the paradigm of maximizing collateral blood
18°C, the patient is placed in Trendelenburg position, flow in an attempt to improve spinal cord perfusion
and cardiopulmonary bypass is halted or slowed. Aortic during aortic occlusion. These include the use of cere-
cross clamping, which may result in embolization of ath- brospinal fluid drainage (272–275), monitoring of spinal
erosclerotic debris, tearing of a fragile aorta (especially in somatosensory evoked potentials (276) or transcranial
patients with cystic medial necrosis), and aortic dissec- electrical stimulation (277), distal perfusion (278–280),
tion, is avoided. It also provides for a relatively bloodless administration of low-dose naloxone (279), direct
surgical field, minimizes the amount of surgical dissec- epidural cooling (281), and mild hypothermia (without
tion needed, and provides for cardiac, visceral, and neuro- circulatory arrest) (282). Most institutions favor some
logic protection. form of intercostal reimplantation when used with the
Subsequently, retrograde cerebral perfusion via a supe- aforementioned adjuncts, as they allow for a longer clamp
rior vena cava (SVC) cannula was introduced as an adjunct time without inducing critical cord ischemia; however,
for cerebral protection during hypothermic arrest. It was for acute dissection this may not be practical owing to
believed that this technique would improve outcomes by the fragile tissue (283). Using various combinations of
providing nutrients and oxygen to the brain and flushing these techniques has resulted in a marked decrease in
out particulate matter from the cerebral and carotid arter- postoperative paraplegia to about 5% to 10%, a consider-
ies that would otherwise embolize. However, systematic able improvement when compared with 16% in the
studies of retrograde cerebral perfusion have shown no largest surgical series (278). Importantly, a recent study of
improvement in outcomes (266,267). More recently, the surgical repair of thoracoabdominal aortic aneurysms
technique of selective antegrade cerebral perfusion was found that surgical mortality was substantially higher
introduced. With this method, perfusion cannulas are in low-volume hospitals and for low-volume surgeons
inserted directly into the cerebral vessels to perfuse the compared with high-volume hospitals and high-volume
brain during all but brief periods of surgery. In fact, surgeons (284).
the use of the multilimbed, prosthetic aortic graft has made
such selective antegrade cerebral perfusion easier and
The Postoperative Aorta
more effective. The data available thus far suggest that selec-
tive antegrade cerebral perfusion significantly reduces the Precise knowledge of the surgical technique performed,
incidence of temporary neurologic dysfunction, although and its anatomic consequences, is crucial to an accurate
its impact on stroke risk remains unclear (267). postoperative imaging evaluation. This will enable the
In descending aortic repairs, including aortic dissection radiologist to differentiate normal postoperative findings
and thoracoabdominal aneurysms, aortic cross clamping from those indicative of postoperative complications, pos-
has profound effects on perfusion of the spinal cord, sibly requiring intervention. Two techniques are widely
depending on hemodynamic alterations and on the nature used for repair of aortic dissection or aneurysm: placement
of vascular supply from the descending aorta to the cord. of a simple interposition graft (Fig. 2-94) (also known
During cross clamping, dysfunction of autoregulatory as the open technique) or use of an inclusion graft (wrap)
circuits results in decreased perfusion; this has been postu- technique (Fig. 2-95). With an interposition graft, the
lated as an etiology for the significant rate of postoperative diseased segment of aorta is excised and a graft is sewn end
paraplegia. From a practical standpoint, the risk of para- to end between the two segments using full-thickness bites
plegia increases with the amount of aorta replaced, cross- (Fig. 2-94). Critical vessels are usually directly reimplanted
clamp time, and operation for rupture or acute dissection into the graft. Advocates of this technique argue that
(125,268,269). sewing the graft to all three layers of the aortic wall
Multiple strategies have been advocated to decrease the provides greater durability, thus minimizing the risk of
incidence of spinal cord ischemia resulting from arch or anastomotic pseudoaneurysms (128).
descending aortic cross clamping. Some advocate deep The inclusion graft technique involves aortotomy, graft
hypothermia with circulatory arrest, thus completely insertion, and subsequent enclosure of the graft within
avoiding cross clamping entirely (270), although this the remnant of diseased aorta; this results in a poten-
method is not widely used because of the threat of coagu- tial space between the graft and the aortic wall (Fig. 2-95).
lopathy and perioperative hemorrhage with systemic hep- The perigraft space, defined as the potential space
arinization (271). Routine reimplantation of the largest between the graft and the native aortic wrap, may con-
backbleeding intercostal arteries from T8 to L1, as the sole tain thrombus, flowing blood, or both (Fig. 2-96) (285).
method to prevent spinal cord ischemia, has fallen out Pseudoaneurysm formation, or persistent flow within the
of favor at many institutions because it increases clamp perigraft space, usually results from partial dehiscence of
times and may be technically difficult in an acutely dis- either the proximal (Fig. 2-97) or distal suture line (or the
sected or heavily calcified aorta. Adjuncts that have dra- coronary anastomosis if reimplantation was performed)
A B
Figure 2-94 Interposition (open) graft technique for ascending aortic aneurysm. A: Diagram of an atherosclerotic ascending aortic
aneurysm that spares the sinus segment but involves the transverse arch. This may require hypothermic circulatory arrest to avoid cross
clamping the aortic arch and to operate in a bloodless field. B: Completed repair after aortic transection and supracoronary proximal anas-
tomosis with a beveled distal anastomosis to the transverse arch.
A B
Figure 2-95 Continuous suture graft inclusion technique. A–C: With use of hypothermic circulatory arrest, the aneurysm is opened and
the distal anastomosis is performed first (B and C) and followed by the proximal anastomosis, both with continuous suture technique. In this
illustration the hemiarch technique is shown, with a tongue of graft material extending along the inferior surface of the aortic arch, beyond
the origins of the great vessels (straight arrow in B). The superior edge of the graft material does not extend beyond the origin of the
brachiocephalic artery (curved arrow in B). D: The native aortic sac is trimmed and wrapped around the prosthesis, and its cut edges are
sutured together. (From Rofsky NM, Weinreb JC, Grossi EA, et al. Aortic aneurysm and dissection: normal MR imaging and CT findings after
surgical repair with the continuous-suture graft-inclusion technique. Radiology. 1993;186:195–201, with permission.) (continued)
C D
A B
Figure 2-96 Periprosthetic flow and thrombus: CT and MR demonstration. A: Diagram demonstrates the potential space between
the native aorta and the aortic wrap, termed the perigraft space. Perigraft flow is present anteriorly, and perigraft thrombus is present
posteriorly. B: Axial image from non–contrast-enhanced CT demonstrates enlargement of the ascending aorta, with a thin, oval rim of high-
attenuation material (arrows) depicting the graft identified in a portion of the lumen. C: Contrast-enhanced CT at the same level demon-
strates flow within the graft (open arrow) and adjacent to the graft (black arrow). The perigraft flow as well as the adjacent thrombus
(white arrow) is contained within the adventitial wrap. D: Axial electrocardiogram-triggered T1-weighted SE MR image at the same level
demonstrates the signal void characteristic of flow within the graft (arrowhead) and in the perigraft space (arrow). Intermediate signal inten-
sity thrombus in the perigraft space is seen as well (open arrow). If this patient did not have a surgical history, one could mistake these
findings for a chronic aortic dissection. Note that the presence of perigraft flow and/or thrombus is not an indication for reoperation unless
there is symptomatic mass effect or rapid growth. (From Rofsky NM, Weinreb JC, Grossi EA, et al. Aortic aneurysm and dissection: normal
MR imaging and CT findings after surgical repair with the continuous-suture graft-inclusion technique. Radiology. 1993;186:195–201, with
permission.)
C D
A B
Figure 2-97 Proximal suture line dehiscence with massive perigraft hematoma after aortic root replacement. A: Oblique sagittal electro-
cardiogram (ECG)–gated spin-echo MR image shows massive perigraft hematoma (H) presenting as a pulsatile chest mass (arrow). The graft
(G) is seen as a signal void, and the proximal anastomosis has dehisced (arrow). B: Axial ECG-gated spin-echo MR image shows graft (G), mas-
sive perigraft hematoma (H), and mass effect on the superior vena cava (SVC) (arrow) and right pulmonary artery (arrow). Note the presence
of left lower lobe atelectasis (curved arrow). C: Oblique sagittal source image from breath-hold gadolinium-enhanced three-dimensional MR
angiogram shows graft dehiscence (small arrow) and aortic coarctation (large arrow). Note how poorly the hematoma (H) is seen on the angio-
graphic source image (continued).
B
Figure 2-98 Persistent perigraft flow after descending aortic
dissection repair. A: Unenhanced CT shows enlargement of the de-
scending thoracic aorta with mass effect on elliptical shaped graft
C (G). B: Enhanced axial multidetector CT (MDCT) image shows the
compressed graft (G) as the highest attenuation structure as well
Figure 2-97 (continued) as persistent flow (arrow) and thrombus (T) within the perigraft
space. Surgical repair demonstrated suture line dehiscence.
but can also result from leakage created from needle holes volves anastomosing the native aortic wrap to the right
with an otherwise intact suture line. This technique can be atrial appendage; the goal of this shunt is to decompress
used at the aortic root, arch, and descending thoracic aorta a potentially tense perigraft space, especially with exces-
(Fig. 2-98). Advocates of this technique believe that wrap- sive intraoperative bleeding, which could compromise the
ping the graft provides for an extra measure of hemostasis coronary artery ostia. Patients who undergo the Cabrol
if bleeding is severe and can prevent fatal exsanguination shunt may have a small amount of perigraft flow; how-
from suture dehiscence. ever, this technique may prevent progression to large
In patients who undergo the inclusion graft technique, pseudoaneurysm formation (285). This iatrogenic left to
a Cabrol procedure (286) may be performed. This in- right shunt may never close, and close interval follow-up
is needed to detect complications. Perigraft thrombus, to avoid anticoagulation in this relatively young patient
which should slowly absorb, is a normal postoperative population, David et al. (289) have described a valve-
finding. preserving aortic root reconstruction technique. This tech-
Niederhauser et al. (287) compared 193 consecutive nique involves excising the sinuses of Valsalva while
patients who underwent surgery for ascending aortic sparing the aortic leaflets, sewing a Dacron graft to the base
dissection with either the open (interposition) technique of the aortic annulus, and reimplanting the aortic valve
(n 93) or the inclusion technique (n 100). The overall leaflets within the graft to restore their normal anatomic
early mortality was 24%. Following graft inclusion it was configuration. When successful, this procedure avoids the
31% compared with 16% in the open technique group. need for prosthetic valve replacement and the associated
Survival at 8 years was significantly increased in the open long-term risks and reduces the risk for repeated valve
technique group. Pseudoaneurysm formation occurred in surgery. Despite the fact that the valve cusps often appear
3% of patients and only after graft inclusion. Freedom functionally normal, there is evidence that they have
from reoperation was 80% at 8 years and did not differ abnormalities similar to those in the aortic wall tissue,
between groups. Graft inclusion was an independent signi- potentially limiting their durability. Some younger patients
ficant predictor of early and late mortality. The authors have a dilated aortic root, but because of intrinsic valve
conclude that the open resection technique is the method dysfunction, the aortic valve cannot be spared. For those
of choice, showing superior early and late results and wishing to avoid the prosthetic valve required with the
avoiding pseudoaneurysm formation. composite aortic graft, one option has been a pulmonary
Quint et al. (288) have reviewed the CT findings in autograft, also known as the Ross procedure. In this proce-
114 patients who underwent interposition graft placement dure, the patient’s aortic valve and root are excised. The
of the ascending aorta (n 93), aortic arch (n 11), patient’s own pulmonary root is then excised and trans-
and descending thoracic aorta (n 25). Low-attenuation planted into the aortic position. The pulmonary root is, in
material was seen adjacent to the ascending graft in 55% turn, replaced with a cryopreserved pulmonary (or aortic)
to 82% of patients and adjacent to the descending graft homograft. Although conceptually attractive, the procedure
in 60% to 79% of patients, showing diminishing frequency has not been free of complications. One of its major limita-
and thickness over time. CT scans in 30 of 53 patients tions is the development of late autograft root dilatation,
showed residual low-attenuation material adjacent to the which is particularly problematic among those who had
graft more than 1 year after surgery. For all types of aortic aortic root dilatation preoperatively (125).
grafts, felt rings were often used to buttress an anasto- The operation of choice in patients with aortic root
mosis, and felt pledgets were used to reinforce the graft disease in which the valve cannot be preserved is place-
or the native aortic wall at sites of intraoperative can- ment of a composite graft (Dacron-impregnated graft
nula placement (Fig. 2-99). These felt rings and pledgets with mechanical valve). The original operation intro-
are generally visible on CT images because of their high duced by Bentall and Debono used an inclusion graft
attenuation, and for aortic root repair, rings are seen at the technique with the coronary arteries directly anasto-
site of coronary artery reimplantation (288). Unenhanced mosed to the composite valve conduit. Pseudoaneurysms
MDCT images are critical to obtain, as a pledget could occurred at the coronary artery anastomosis owing to
be misinterpreted as a pseudoaneurysm on a contrast-only tension on the suture line. For this reason, Cabrol et al.
study. (286) devised a composite graft with a second coro-
nary artery tube graft, which is anastomosed directly
to the proximal right and left main coronary arteries.
Surgery for Aneurysm or Dissection Involving
Recognition of this type of graft is important because if
the Ascending Aorta and Aortic Root
imaged in an axial plane it may resemble a focal dissec-
Aneurysms of the aortic root involving the sinuses of tion at the aortic root (Fig. 2-100).
Valsalva, the sinotubular junction, and the proximal ascend- Further refinement of the composite graft technique
ing aorta are a distinct entity from atherosclerotic ascending includes an interposition (open) technique with reim-
aortic aneurysms. Classically, aortic root aneurysms occur plantation of the coronary arteries using a generous
in patients with connective tissue disorders. However, the button of original aorta sewn into the graft (Fig. 2-99).
majority of patients lack phenotypic expression of a This composite valve conduit reconstruction results in a
connective tissue disorder. Cystic medial degeneration is the more durable result compared with the valve-preserving
pathologic feature for either etiology (263). procedure. Lewis et al. (290) reported that only 9 of
In patients with aortic dissection or aneurysm that 241 patients required reoperation for problems related to
involves the aortic root (cystic medial disease), graft place- the initial composite graft reconstruction procedure.
ment using either the open or inclusion technique may Svensson et al. (291) had no reoperation in 67 patients
be supplemented by reimplantation of the coronary arter- at intermediate-term follow-up. Other predictors of the
ies and resuspension of the aortic valve. In an attempt need for reoperation in this series were Marfan syndrome
A B
Figure 2-99 Normal appearance after interposition graft repair

for ascending aortic aneurysm with aortic root replacement and
coronary artery reimplantation using the button technique. See
Color Figure 2-99C. A: Axial enhanced multidetector CT (MDCT)
scan shows circumferential felt reinforcing ring (arrow) at the distal
anastomosis of an ascending aortic interposition graft, 6 months
after surgery. B: Axial enhanced MDCT scan at the level of the right
coronary artery shows felt ring (arrows) surrounding the reim-
planted vessel. C: Coronal volume-rendered image shows felt ring
at distal anastomosis (top arrow) and reimplanted right coronary
C artery button (bottom arrow).
Figure 2-100 A: Diagram of a Cabrol composite graft for aortic

root replacement. It consists of a graft with a prosthetic aortic
valve and an end-to-side coronary tube graft. The remaining native
aorta is then wrapped around the graft, and a fistula between the
perigraft space and right atrial appendage may be created (not
shown). B: Axial reformation from breath-hold gadolinium-
enhanced three-dimensional MR angiogram demonstrates the left
coronary tube from a Cabrol graft that mimics a focal aortic dissec-
tion. C: Oblique axial reformation from same acquisition demon-
strates the true configuration of the end-to-side coronary tube
A graft with patent limbs. (From reference 29, with permission.)
B C
and the need for concomitant procedures at the initial support an aggressive policy of composite root replace-
operation. ment in acute type A dissection.
Gott et al. (292) have shown that elective aortic root At the time of this publication, composite graft replace-
reconstruction carries a significantly reduced morbidity ment of the ascending aorta was generally recommended
and mortality compared with urgent or emergent inter- for patients with cystic medial disease and an aortic diam-
ventions. In a multi-institutional study of 10 centers, eter of 4.5 to 5 cm. However, because of the excellent
675 patients with Marfan syndrome underwent primarily long-term survival and minimal mortality from this series,
composite graft aortic root reconstruction. Elective surgery combined with the unpredictable nature of medial dis-
had an early mortality of 1.5%, urgent cases had a 2.6% ease, some centers advocate ascending aorta replacement
mortality, and the mortality was 11.7% among those at a size as small as 4 cm.
undergoing emergent procedures for dissection. In the ascending aorta, especially after insertion of the
More recently, Halstead et al. (293) reviewed their original Bentall composite graft, false aneurysm formation
experience with 162 patients who underwent emergent at the site of coronary anastomosis may lead to cardiac
surgery of the ascending aorta due to aortic dissection. ischemia, infarction, and sudden death. Pseudoaneurysms
Eighty-nine patients underwent standard supracoronary resulting from dehiscence of the proximal or distal suture
interposition graft placement and 73 underwent compos- line are often seen during routine postoperative surveil-
ite graft root replacement with coronary artery reimplan- lance, especially in patients with cystic medial necrosis
tation. Although hospital mortality rates were similar (Fig. 2-101). Of patients who undergo reoperative repair of
in both groups (12%), actuarial survival estimates at 1, 5, a graft pseudoaneurysm, morbidity and mortality have
and 10 years were significantly greater for patients who been reported to be as high as 75% and 25%, respectively,
underwent composite root replacement. These results and higher in the case of emergent repair (294,295). Thus,
symptomatic from mass effect or hemodynamic com-

promise, or significant growth is documented on serial
imaging (285). Perigraft thrombus is extremely common
and a normal postoperative finding (Fig. 2-102).
Surgery of the Aortic Arch and Descending

Thoracic Aorta
Surgical replacement for aortic arch aneurysms and dissec-
tion is particularly challenging and carries a significant
risk of neurologic damage from embolization of athero-
sclerotic debris or from global ischemic injury during cir-
culatory arrest (125). To replace the dilated arch with a
prosthetic tube graft, the brachiocephalic vessels must be
removed from the arch before its resection and then reim-
planted into the tube graft arch after its interposition.
Traditionally, this involved removing and then reimplant-
ing the brachiocephalic vessels en bloc during hypother-
mic circulatory arrest. However, many surgeons have now
adopted a newer surgical technique that uses a multil-
imbed prosthetic arch graft, to which each arch vessel is in
turn anastomosed individually, which reduces the dura-
Figure 2-101 Axial enhanced multidetector CT (MDCT) image at tion of hypothermic circulatory arrest and the frequency
the level of the aortic root in a patient who has undergone compos- of embolic events (297) (Fig. 2-103). The trifurcated graft
ite interposition graft placement of the ascending aorta shows two technique results in low rates of perioperative mortality,
small pseudoaneurysms at the aortic root (arrows). These should not
be mistaken for enlarged coronary artery buttons or a Cabrol graft.
temporary neurologic dysfunction, and stroke. It may re-
duce cerebral embolization, as it requires no instrumenta-
tion of the aortic arch to establish selective cerebral perfu-
it is essential to maintain close follow-up after ascending sion, and although it mandates hypothermic circulatory
aortic repair, especially in patients prone to develop post- arrest to place the graft, this interval is reliably brief
operative aneurysmal complications such as those with enough to fall within accepted safe limits. This strategy
Marfan syndrome, so that repair of developing pseudoa- leaves no residual arch tissue behind (as does the island
neurysms can be performed in a nonemergent setting or peninsula graft) that may become aneurysmal with
(296). With inclusion graft techniques, the presence of time (297).
a postoperative pseudoaneurysm (perigraft flow) does Some patients requiring surgical repair have aneurysms
not always mandate surgical repair unless the patient is that involve multiple aortic segments (the ascending, arch,
A B
Figure 2-102 Perigraft thrombus after inclusion graft technique repair of aortic dissection: CT and MR findings. A: Axial contrast-
enhanced CT demonstrates perigraft thrombus (long arrow) and persistent intimal flap distal to the graft site in the descending aorta (small
arrows). B: Axial electrocardiogram-triggered T1-weighted SE MR image at the same level demonstrates intermediate signal intensity
thrombus in the perigraft space (arrow) and intimal flap in the descending aorta (small arrows).
A B
Figure 2-103 Normal postoperative appearance on MR after

interposition graft repair for aortic root, ascending aortic, and arch
aneurysm with reimplantation of the arch vessels into a single side
arm graft, end-to-end anastomosis of arch vessels, and placement
of “elephant trunk” in descending thoracic aorta. A: Oblique
sagittal maximum intensity projection (MIP) image from preopera-
tive breath-hold gadolinium-enhanced three-dimensional (3D) MR
angiogram (MRA) shows aneurysm of ascending aorta (including
root) and aortic arch. The descending thoracic aorta has been pre-
viously repaired (arrow). Note the common origin of the innomi-
nate and left common carotid artery. B: Schematic of repair shows
reimplantation of coronary arteries with a button technique,
sidearm arch graft arising from ascending aorta with three sepa-
rate anastomoses and placement of an “elephant trunk” (arrow).
C: Sagittal MIP image from postoperative breath-hold gadolinium-
enhanced 3D MRA shows graft placement from aortic root to
proximal descending thoracic aorta. Note the normal flare of the
reimplanted right coronary artery button (arrow). The arch vessels
have a new configuration, similar to schematic drawing in B.
D: Coronal volume rendered image from postoperative breath-
hold gadolinium-enhanced 3D MRA shows single side-arm graft
reimplanted into the right side of the midascending aorta (arrow)
with direct end-to-end anastomosis of the three arch vessels.
E: MIP image from postoperative breath-hold gadolinium-
enhanced 3D MRA shows similar findings. F: Source image from
postoperative breath-hold gadolinium-enhanced 3D MRA shows
an elephant trunk graft (arrow) left in situ for replacement of
the descending thoracic aorta at a later date. The second proce-
dure can be a conventional repair or placement of a stent graft.
The elephant trunk graft may be confused with aortic dissection;
however, unlike dissection, it is a very short segment and ends
abruptly in the proximal descending thoracic aorta. (B from
Spielvogel D, Mathur MN, Griepp RB. Aneurysms of the aortic
arch. In: Cohn LH, Edmunds LH Jr, eds. Cardiac surgery in the
C adult. New York: McGraw-Hill; 2003:1149–1168, with permission.)
(continued)
D, E F
and/or descending thoracic aorta). Trying to repair both aneurysm is repaired via a standard left thoracotomy.
the ascending and descending segments in one operation However, the surgeon cross clamps the distal end of the
is technically difficult, even for the most experienced sur- existing elephant trunk (rather than clamping the dilated
geons, especially when operating for dissection. Therefore, native aorta), and the proximal end of the new descending
many centers will repair these lesions by performing aortic graft is now anastomosed to the free distal end of the
staged procedures, with repair of one aortic segment elephant trunk. More recently, a hybrid elephant trunk
(ascending or descending) first, to be followed at a later technique has been used in which the descending repair is
date by a second surgery. The order of surgical repair is performed via a stent-graft instead of a traditional left tho-
dictated by the individual’s specific aortic anatomy—for racotomy (299) (Fig. 2-105).
example, depending on which aortic segment is at greatest In patients who have undergone repair of aortic dissec-
risk of rupture and where the surgeon can expect to place tion, CT or MR may demonstrate persistent intimal flap
the aortic cross clamp. distal to the graft site in up to 75% to 100% of cases and in
However, in some cases, the lack of an adequate neck virtually all cases of repaired chronic dissection (Figs. 2-102
(a relatively nondilated segment of the proximal descend- and 2-106) (300–302). Despite tacking down the intima
ing aorta) makes it difficult to perform such a staged proximally, distal communications between the true and
procedure. In such cases, an alternative staged surgical false lumen are common and result in persistence of the
approach known as the “elephant trunk” technique is used false channel. The single greatest risk factor for reoperation
(Fig. 2-104) (298). The initial surgery is, for the most part, of distal aortic dissection is a patent false lumen under
similar to the standard ascending aneurysm and total arch systemic pressure (Fig. 2-107). Immer et al. (303) recently
repair (125). However, the distal end of the aortic arch graft reported that a large false lumen (with the area of the true
is several centimeters longer than the native arch, so that lumen 30% of the false lumen) 6 months after surgery is
after the anastomosis is made to the proximal descending the strongest predictor for secondary dilatation distal to the
aorta (just distal to the left subclavian artery), the excess primary repair. In chronic dissections, the distal anastomo-
length of the graft is left protruding distally within the sis is usually made to both lumens, thus intentionally
lumen of the descending TAA (125). In the second stage of maintaining false lumen patency because it may provide
this procedure, the descending or thoracoabdominal aortic critical blood supply to the kidneys or viscera.
A B
Figure 2-104 Staged repair of the entire thoracic aorta using an

elephant trunk technique. A: Schematic drawing shows previous
supracoronary repair of ascending aorta and arch with reimplanta-
tion of the arch vessels as a single island or patch (compare with
Fig. 2-103). A distal segment of free floating graft is present in an
aneurysmal descending thoracic aorta; this is described as an ele-
phant trunk and allows a second operation without cross clamping
the native aorta. B: Axial multidetector CT (MDCT) image shows typ-
ical appearance of free floating elephant trunk seen in cross section
(arrow). This may be confused with aortic dissection with a 360-de-
gree circumferential tear; however, unlike dissection, it is a very short
segment and ends abruptly in the proximal descending thoracic
aorta. Note the normal caliber of the repaired ascending aorta. C:
Axial MDCT image in another patient shows an elephant trunk within
the false lumen of a chronic aortic dissection (arrow). Note the
hematoma and/or organizing thrombus around the recently repaired
ascending aorta. (A from Riley P, Rooney S, Bonser R, Guest P.
Imaging the post-operative thoracic aorta: normal anatomy and pit-
C falls. Br J Radiol. 2001;74:1150–1158, Figure 10, with permission.)
A
B
Figure 2-105 Complete replacement of the thoracic aorta using

a hybrid elephant trunk technique. See Color Figure 2-105B,C. A,
B: Maximum intensity projection and volume-rendered sagittal mul-
tidetector CT (MDCT) images show complete ascending aorta and
arch replacement with residual thoracic aneurysm and elephant
trunk graft (arrow) floating within the aneurysm sac. See Figure 2-7
for preoperative images. C: An aortic stent-graft was used to com-
plete the elephant trunk repair, avoiding a left thoracotomy and
surgical resection. (Courtesy of Dominik Fleischmann, Stanford
C University, Stanford, California.)
180
A B
Figure 2-106 Oblique sagittal source image (A) and maximum intensity projection image (B) from breath-hold gadolinium-enhanced
three-dimensional MR angiogram demonstrate the typical postoperative appearance of a repaired Stanford A dissection with a supracoro-
nary interposition graft. The residual, distal intimal flap is a common postoperative finding and is best visualized on the source image. (From
reference 29, with permission.)
All patients with native or postoperative communicat- with 2D gadolinium-enhanced magnetization-prepared

ing dissections are at risk for secondary aneurysm forma- GRE MRA found the latter technique more sensitive in
tion from an enlarging false lumen. In a longitudinal MR distinguishing perigraft flow from thrombus (306). The
study in patients with repaired type A dissection, 25% use of breath-hold, gadolinium-enhanced 3D MRA
developed aneurysms distal to the graft site (Fig. 2-107) can better differentiate perigraft flow from thrombus
(304). Patients with Marfan disease are at an even higher (Fig. 2-109) for three separate but distinct reasons. First,
risk for enlarging aneurysms from progressive dilatation of these sequences often have a very short echo time (1 to
the false lumen distal to the repair site (40%) (305) or 2 ms), which minimizes signal loss from turbulent flow.
the development of a separate noncontiguous aneurysm Second, their intrinsic high spatial resolution with small
(296). These patients also have a high incidence of new voxels minimizes artifactual signal loss from intravoxel
dissection, distal to the graft site, in segments of the aorta phase dispersion. Third, the magnitude of T1 shortening
that had previously appeared normal. In addition, failure of gadolinium may also partially compensate for the
to treat the aortic root in a patient with cystic medial dephasing phenomenon.
necrosis and a supracoronary aortic dissection will invari- On unenhanced CT, prosthetic graft often appears as a
ably lead to continued expansion of the residual ascending ring of high attenuation contiguous with the aortic wall
aorta (Fig. 2-108). (Fig. 2-96 and 2-98). Postoperative complications are read-
ily diagnosed with similar criteria as previously discussed
with MR. A significant advantage of CT over MR is in the
Imaging Features
evaluation of patients who have had endovascular stents
In patients who have undergone inclusion graft techniques, and stent-grafts placed, as these may cause significant sus-
enlargement of the perigraft space is readily visualized with ceptibility artifacts with resultant artifactual signal loss. The
contrast-enhanced CT, TEE, and MRI, and perigraft flow is relationship of a stent to an aortic branch vessel can be
reliably distinguished from thrombus with either contrast- demonstrated with source and MPR images (Fig. 2-110).
enhanced CT or TEE. MPRs are invaluable for assessing the interior of the metal-
With MR imaging, distinction between perigraft flow lic prosthesis and determining the presence of intimal
and thrombus may be difficult, as perigraft flow is often hyperplasia or stent deformation Complications such as
turbulent. This may result in dephasing with artifactual stent migration, perigraft leaks, and progressive dilatation
signal loss on GRE imaging. A study comparing cine GRE of an aneurysm sac can readily be detected.
A B
Figure 2-107 Patency of the false lumen—comparison of good and poor prognostic signs. A: Oblique sagittal maximum intensity
projection image from breath-hold gadolinium-enhanced three-dimensional MR angiogram demonstrates a normal postoperative appear-
ance of a repaired Stanford B dissection. Note that the false lumen is almost completely thrombosed (arrow) and terminates in the upper
abdomen, protecting this patient from subsequent downstream dilatation. B: Severe aneurysmal dilatation of the false lumen has developed
in a patient who underwent previous repair of a Stanford type A dissection. Surgical repair is indicated for chronic aortic dissection when the
overall lumen size is 6 cm. Note the presence of higher signal intensity within the smaller, anterior true lumen (arrow).
Observations of the Stented Aorta with endovascular technologies, in the range of 0% to 5%

(308). Type I endoleak incidence rates are related to mor-
As thoracic aortic stent-graft placement for repair of
phologic case complexity; primary frequency rates of 0% to
aneurysm, dissection, ulcer, and traumatic injury evolves
20% are reported in the literature, with 0% to 5% secondary
into a more common treatment, complications similar to
incidence (308). Endoleaks are classified as follows:
those seen in the abdominal aorta will arise. Endoleaks,
paraplegia, stroke, and retrograde dissections are the main Type I, attachment sites. These are usually seen during
specific complications of thoracic endograft placement deployment of the stent-graft and treated with cuff
(307,308). Imaging of complications includes endoleaks expanders or an additional covered stent. During
and aortic neck remodeling over time. A proximal and distal surveillance CT, these are best visualized during the
landing zone of 1.5 cm of normal aorta would be ideal. angiographic phase as a blush of contrast seen at
Creating an appropriate proximal neck—if necessary by the proximal or distal landing zone. The aneurysm
supra-aortic branch remodeling—and deliberate left subcla- sac is under systemic pressure with a type I endoleak
vian artery overstenting are key mechanisms in avoiding and these must be treated to prevent aneurysm
proximal endoleaks. Paraplegia rates are reportedly low rupture.
A B
Figure 2-108 Progressive dilatation of the aortic root in a patient with cystic medial disease who underwent supracoronary graft place-
ment for acute aortic dissection. A: Balanced steady-state free precession (SSFP) [True FISP (fast imaging with steady-state precession)] MR
image shows dilatation of the aortic root to 6.3 cm, a size that requires replacement. Note the correct way to measure the dilated root.
B: Sagittal maximum intensity projection image from breath-hold gadolinium-enhanced three-dimensional MR angiogram also shows root
dilatation and the presence of a kinked short segment supracoronary repair (arrows). Although the original operation should have consisted
of a composite graft with excision of the diseased aortic root, the operation would have been technically more challenging.
A B
Figure 2-109 Postoperative pseudoaneurysm after repair of ascending aortic aneurysm. A: Oblique sagittal segmented k-space cine
gradient-echo MR image demonstrates a large area of signal void from turbulent flow posterior to the ascending aorta. B: Oblique sagittal
source image from breath-hold gadolinium-enhanced three-dimensional MR angiogram demonstrates a pseudoaneurysm at the proximal
suture line with concentric thrombus (T). This case demonstrates the importance of using gadolinium in conjunction with an ultrashort echo
time (1 to 2 ms) to minimize signal loss from intravoxel dephasing resulting from turbulent flow. (From reference 29, with permission.)
A, B C
Figure 2-110 Metal stent in branch vessel resulting in false-positive vessel occlusion. A, B: Oblique sagittal maximum intensity projection
images from breath-hold gadolinium-enhanced three-dimensional MR angiogram demonstrate occlusive disease of all three great vessels.
Note the decreased signal of the left subclavian and vertebral artery that fill retrograde. C: Source image from same dataset shows metal
stent in the common carotid artery, resulting in artifactual signal loss (arrow). Newer stents result in considerably less artifact; however, mul-
tidetector CT (MDCT) still remains the procedure of choice to image stents for patency and in situ thrombosis.
A B
Figure 2-111 Type II endoleak best seen on delayed multidetector CT (MDCT) in a patient after thoracic aortic stent-graft placement.
A: Arterial phase MDCT of a thoracic aneurysm treated with a stent graft shows a small type II endoleak (open arrow). B: The endoleak is
better seen on the 1-minute delayed MDCT image (open arrow).
Type II, branch vessel endoleak. Type II endoleaks result (310). Early diagnosis may be difficult because of the non-
from a backbleeding intercostal or bronchial artery. specific symptoms but is important because the time at
These are best seen by performing a delayed (approx- which treatment is initiated can affect prognosis (310).
imately 90 seconds) scan following the initial CT During this phase mural thickening of the affected vessels
angiogram (Fig. 2-111). If the aneurysm sac size is likely results from inflammation of the tunica media and
decreasing or stable over time, these are not usually tunica adventitia.
treated. If sac size progresses, the backbleeding vessel Late-phase Takayasu disease may be further classified
may be embolized with caution (unless it is thought as classic pulseless disease (type 1), a mixed type (type 2),
to supply the spinal cord), as the aneurysm sac is an atypical coarctation type (type 3), and a dilated type
likely under systemic pressure with a subsequent risk (type 4) (311). Most patients present with a form of late-
of rupture. phase disease.
Type III, device degeneration leading to leak through graft
material or junction dehiscence. These are extremely
Imaging Features
rare and may necessitate placement of a second
stent-graft for support or even graft explantation
and resection if interventional therapy fails.
Type IV, graft porosity. Porosity of graft material may Helical CT has been successfully used in the evaluation
result in an impressive leak at the time of graft of arteritis (Figs. 2-112 and 2-113) (71,311). A small study
placement. The early AneuRx grafts all demon- of 12 patients (eight patients with active disease and four
strated this leak at angiography and CT scanning with inactive disease) examined with CTA in the evalua-
within 24 hours of insertion. The WALLGRAFT tion of Takayasu arteritis demonstrated high attenuation of
(Boston Scientific, Natick, MA) may also show this the aortic wall on precontrast images in 10 patients and
type of leak. One must be aware of this property of mural calcifications in 9 patients (71). Of these 9 patients,
the grafts being used, as this leak is self-limited over calcifications were intimal in location in 3 patients and
a few days. mural in 6 (71). In 6 of 8 patients with active disease
Type V, endotension. The hallmark of endotension is the mural enhancement was demonstrated during the arterial
progressive enlargement of aneurysm sac size with- phase of contrast injection (71). No patients with inactive
out an endoleak identified on CTA or angiography. disease had arterial phase mural enhancement; thus, this
Open conversion may be required for these cases, as finding is specific but not sensitive for the detection of
the sac remains pressurized and no cause for the leak disease activity. Delayed CT images (performed 7 minutes
is identified. It appears that increased pressure after arterial phase imaging) demonstrated mural
within the AAA sac (endotension) can be transmit- enhancement in 8 patients, 7 of whom had clinically
ted through clot. active disease (71). Mural thickness can persist, even with
inactive disease, as healed lesions often contain a prolifera-
tion of fibrous and connective tissue that may undergo
dystrophic calcification.
AORTITIS
Magnetic Resonance Evaluation
MR features of the systemic phase of aortitis are best visual-
Clinical Features
ized with T1-weighted black blood images supplemented
Takayasu disease is the most common cause of aortitis, by breath-hold STIR or T2-weighted images (Fig. 2-114).
usually afflicting young women of Asian descent. It is The presence of increased signal intensity within the aortic
an inflammatory vasculitis of unknown etiology and is wall on the latter two pulse sequences, as well as enhance-
categorized in one of two ways, based on clinical course ment after the administration of gadolinium chelates,
or anatomic location. Sites of disease may include the suggests active disease (Fig. 2-112). A small study of four
aortic arch vessels, the thoracoabdominal aorta and patients with periaortitis showed that the size of the
branch vessels, and the pulmonary arteries. The disease periaortic rind was a better predictor of disease activity than
involves the vessel walls and results in luminal abnor- quantitative enhancement values (312). More recently,
malities, including stenosis, occlusion, dilatation, and fluorodeoxyglucose positron emission tomography (PET)
aneurysm, all of which may coexist within the same has been advocated as a more sensitive technique for
patient (309). assessing disease activity (313).
The acute or systemic/prepulseless phase usually pre- Yamada et al. (314) studied 77 patients with Takayasu
sents as an acute systemic illness with manifestations arteritis using T1-weighted SE and cine MR imaging.
including fever, night sweats, arthralgia, myalgia, anemia, Aortic lesions included stenosis, dilatation, aneurysm,
and elevation of the erythrocyte sedimentation rate (ESR) wall thickening, and mural thrombi. The most frequent
A, B C
Figure 2-112 Active mural disease in a patient with Takayasu arteritis A: Oblique sagittal maximum intensity projection image from
multidetector CT (MDCT) shows questionable narrowing in the proximal abdominal aorta (arrow). B: Axial MDCT image from the same
dataset shows periaortic thickening. C: Axial fat-suppressed three-dimensional gradient-echo MR performed 5 minutes after the administra-
tion of gadolinium shows intense periaortic enhancement consistent with active mural disease.
site of stenosis was in the descending aorta, and dilata- disease, which would require surgical or radiologic inter-
tion was seen predominantly in the ascending aorta. In vention. Angioplasty and/or stenting may be useful in
addition, aortic valve regurgitation was seen in 45% treating focal disease, with surgery reserved for long seg-
of patients and pulmonary artery involvement was seen ment lesions and aneurysms. Because patients with arteri-
in 70% of patients. However, no attempt was made to tis have a significantly higher incidence of developing
correlate imaging findings with disease activity in this aneurysms and aortic dissections (316) than the general
study (314). population, they should undergo some form of follow-
Treatment of acute episodes of arteritis includes up imaging. Patients with giant cell arteritis (temporal
steroids and cytotoxic agents, but a significant percentage arteritis) have similar imaging findings but usually have
of patients will not respond. Both CT (315) and MR (312) disease in the ascending aorta and arch vessels (Figs. 2-8
studies have demonstrated a decrease in wall thickening and 2-118).
following steroid therapy. With CTA and MRA techniques
currently available, both mural disease (wall thickening
Infectious Aortitis or Aneurysm
or enhancement) and luminal abnormalities, including
stenoses and aneurysms, can be evaluated with a single Infection of the aorta usually results from septic emboliza-
examination (Figs. 2-115 to 2-117). Angiography should tion to the vasa vasorum, hematogenous seeding of an
be reserved for patients with symptomatic branch vessel existing aneurysm, or extension from a contiguous site of
B
Figure 2-113 Axial multidetector CT (MDCT) shows mural Figure 2-114 MR evaluation of aortitis—active/systemic phase.
thickening of the arch vessels. Arrow denotes the slightly dilated
A: Axial electrocardiogram (ECG)–triggered T1-weighted SE MR
innominate artery.
image demonstrates circumferential thickening, without narrow-
ing, of the descending thoracic aorta. B: Axial ECG–triggered T2-
weighted turbo SE MR imaging demonstrates high signal intensity
of periaortic rind. After a 3-month course of prednisone, the pa-
tient’s symptoms resolved and the MR appearance of the de-
scending aorta returned to normal.
A
Figure 2-115 CT and MR demonstration of aneurysmal disease
coexisting with stenoses in Takayasu arteritis. A: Axial contrast-
enhanced helical CT at the level of the inferior aortic arch demon-
strates subtle wall thickening of the ascending aorta (arrow).
B: Maximum intensity projection image from breath-hold
gadolinium-enhanced three-dimensional MR angiogram demon-
strates the aneurysm of the proximal descending aorta (curved
arrow) and the concomitant long tapered stenosis of the descend-
ing thoracic aorta (arrows). (Case courtesy of Deborah Reede,
Long Island College Hospital, Brooklyn, New York.) B
A B
Figure 2-116 MR angiography of branch vessel disease in Takayasu arteritis. A: Oblique sagittal maximum intensity projection image
from breath-hold gadolinium-enhanced three-dimensional (3D) MR angiogram demonstrates subtle, smooth narrowing of the left common
carotid artery (arrows). B: Axial reformation from delayed gadolinium-enhanced 3D MR dataset demonstrates marked thickening of the wall
of the left common carotid artery (long arrow), resulting in narrowing of the lumen compared with the right common carotid artery (curved
arrow). Short arrow denotes normal left subclavian artery.
Figure 2-117 Sagittal (A) and coronal (B) reformatted images

from multidetector CT (MDCT) dataset show aneurysmal dilatation
of the innominate and subclavian arteries in a patient with
Takayasu arteritis. Note that the left subclavian artery (arrows in A
A and B) is as large as the aorta.
A B
Figure 2-118 Giant cell arteritis with involvement of the

ascending aorta and coexistent active mural disease and
aneurysm. A: Electrocardiogram-gated black blood double inver-
sion recovery MR image shows circumferential thickening of the
wall of the ascending aorta. Note that unlike aortic intramural
hematoma, involvement is 360 degrees and symmetric (arrow).
B: Axial contrast-enhanced fat-suppressed three-dimensional (3D)
gradient-echo MR performed 5 minutes after the administration of
gadolinium shows intense periaortic enhancement (arrow) consis-
tent with active mural disease. C: Source image from breath-hold
gadolinium-enhanced 3D MR angiogram demonstrates an ascend-
ing aortic aneurysm that spares the sinus segment as well as the
transverse arch.
C
infection (317,318). Infected (mycotic) aortic aneurysms been surgically manipulated is a specific but insensitive
are estimated to represent about 0.7% to 2.6% of all aortic finding seen in only a small number of patients. More
aneurysms. Infected aneurysms are prone to rupture, with commonly, fluid and/or an enhancing soft tissue mass is
reports of 53% to 75% aneurysm rupture at surgery. seen to surround the vessel (Fig. 2-119).
Patients with infected AAAs often present with fever, Macedo et al. (317) recently reviewed the imaging
abdominal pain, palpable abdominal mass, and leukocy- findings in 29 patients with 31 infected aneurysms in the
tosis, with or without positive blood cultures. Patients following anatomic locations: ascending aorta (6%),
with infected TAAs may present with hemoptysis. As a rule, descending thoracic aorta (23%), thoracoabdominal aorta
infection is much more common in aortic grafts than in (19%), paravisceral aorta (6%), juxtarenal aorta (10%),
native aortic aneurysms. infrarenal aorta (32%), and renal artery (3%). CT revealed
In the preantibiotic area, infectious aortitis was 25 saccular (93%) and two fusiform (7%) aneurysms with
largely a complication of infective endocarditis and was a mean diameter at initial discovery of 5.4 cm (range, 1 to
usually caused by group A streptococci, Streptococcus pneu- 11 cm). Para-aortic soft tissue mass, stranding, and/or fluid
moniae, or Haemophilus influenzae. Now a diverse array of was present in 13 (48%) of 27 aneurysms, and early
bacteria and fungi has been associated, most commonly periaortic edema with rapid aneurysm progression and
Salmonella species, which account for nearly one third of development was present in three patients (10%) with
the abdominal aortic infections, and Staphylococcus aureus. sequential studies. Other findings included adjacent verte-
Mycobacterium infection secondary to bacille Calmette- bral body destruction with psoas muscle abscess (n 1,
Guérin (BCG) instillations of the urinary bladder for carci- 4%), renal infarct (n 1, 4%), absence of calcification in
noma may also result in infectious aortitis. the aortic wall (n 2, 7%), and periaortic gas (n 2,
Contrast-enhanced CT is the most useful imaging 7%). They concluded that saccular aneurysms with
modality for the evaluation of aortic infections and lobulated contour or those with rapid expansion or devel-
complications, with gallium- or indium-labeled white cell opment and adjacent mass, stranding, and/or fluid in an
studies reserved for equivocal cases (317). The presence of unusual location were highly suspicious for an infected
air within a graft, native vessel, or aneurysm that has not aneurysm.
A B
Figure 2-119 Multidetector CT (MDCT) demonstration of infectious aortitis. A, B: Coronal and axial MDCT images demonstrate a soft
tissue mass surrounding the abdominal aorta (arrows in A and B) and lifting it off the spine. The findings are not diagnostic for infection, as
air is not present. The differential diagnosis includes lymphoma and retroperitoneal fibrosis. Needle aspiration and subsequent tissue from
surgical debridement grew Staphylococcus aureus.
Medical treatment alone carries a high mortality, mediastinum and terminates inferiorly in the right atrium.
whereas the mortality with surgery combined with Although its position and appearance are characteristic, it
antimicrobial treatment is lower. If fine-needle aspira- is variable in size, and because of its relatively thin wall, it
tion cannot be performed safely, empiric antibiotics can show transient variations in diameter with changes in
effective against S. aureus and gram-negative rods, such fluid status or patient position. Following contrast infu-
as Salmonella, should be initiated in cases identified be- sion into a single arm vein, streaming or layering of
fore microbiologic diagnosis. Surgical debridement and contrast in the SVC is commonly seen and should not be
revascularization with either extra-anatomic bypass or a misinterpreted as thrombus. Inflow of unopacified blood
cryopreserved aortic homograft should be performed can also be seen at the level of the azygos vein. Because of
early because delay may lead to fatal rupture. The intent its thin wall and intimate relationship to mediastinal and
of surgery is to (a) control hemorrhage, if the vessel has right hilar lymph nodes, obstruction is common in the
ruptured; (b) confirm the diagnosis; (c) control sepsis; presence of lymph node enlargement.
and (d) reconstruct the arterial vasculature. The patient Cross-sectional anatomy of thoracic veins is complex
should remain on parenteral or oral antibiotics for at but has been described in detail (319–323) (Figs. 2-120
least 6 weeks, perhaps longer, to ensure full eradication and 2-121). A knowledge of thoracic venous anatomy is
of the pathogen and prevent recurrent infection (318). important to understand alterations in venous flow occur-
The most dreaded complications of aortic infections ring in the presence of SVC obstruction (323). Important
include rupture, overwhelming sepsis, and aortoenteric veins include the subclavian veins, brachiocephalic veins,
fistulas. the azygos and hemiazygos veins, accessory hemiazygos
vein, and the superior intercostal veins.
The brachiocephalic veins are formed by the junction of
the subclavian and internal jugular veins in the superior
GREAT VEINS, INCLUDING THE AZYGOS mediastinum (Figs. 2-120 to 2-124). The subclavian veins
AND HEMIAZYGOS SYSTEMS enter the mediastinum by crossing over the first ribs, ante-
rior to the anterior scalenus muscles, which attach to the
The SVC is most important among thoracic veins, serving superior borders of the first ribs, and posterior to the prox-
as the final tributary for drainage of systemic venous flow imal portions of the clavicles. The internal jugular veins
from many intrathoracic veins (Fig. 2-120). It is formed by can be traced into the neck from the point where they join
the junction of the brachiocephalic veins in the upper the brachiocephalic veins.
Figure 2-120 Schematic drawing of the

azygos, hemiazygos, and superior intercostal
venous systems. Inset is a schematic cross
section at the level of the aortic arch.
A B
C D
E F
Figure 2-121 Normal venous anatomy on CT. A–C: Above the level of the aortic arch, the innominate artery (IA), the left common carotid
artery (LCCA), the left subclavian artery (LSA), the right brachiocephalic vein (RBV), and the left brachiocephalic vein (LBV) are the most con-
spicuous normal structures. At or near the thoracic inlet, the brachiocephalic veins are the most anterior and lateral vascular branches visi-
ble, lying immediately behind the clavicle heads and articulations of the first ribs. Although they vary in size, their positions are relatively
constant. Below the thoracic inlet, the left brachiocephalic vein crosses the mediastinum from left to right, anterior to the arterial branches
of the aorta. The two brachiocephalic veins have very different configurations. The right brachiocephalic vein has a nearly vertical course
throughout its length; the left brachiocephalic vein is longer, and courses horizontally as it crosses the mediastinum. In C the left superior in-
tercostal vein (LSIV) is visible posterior to the left brachiocephalic vein and beneath the mediastinal pleura. D–F: At the level of the aortic
arch (AA), the brachiocephalic veins are again visible, and can be seen to join to form the superior vena cava (SVC). Small veins visible at
these levels include the right internal mammary vein (IntMV), left superior intercostal vein (LSIV), and right superior intercostal vein (RSIV).
Internal mammary veins are often visible, and in some patients can be traced from their point of origin from the brachiocephalic veins, or on
the right, the superior vena cava. The left superior intercostal vein forms a venous arch, sometimes referred to as the arch of the hemiazygos
vein, then courses anteriorly around the aortic arch, near the level of T3–T4, to join the left brachiocephalic vein superiorly. The right supe-
rior intercostal vein drains the right second to fourth intercostal veins, and then courses inferiorly in a paraspinal location to its junction with
the azygos vein. It appears to represent the continuation of the azygos vein above the azygos arch, and lies anterior and to the right of the
spine. G–I: The superior vena cava (SVC), the arch of the azygos vein (AzA), and the azygos vein (AzV) are visible. The azygos vein lies on the
right, anterior and slightly lateral to the vertebral column; it is almost always visible on CT. The azygos vein drains the lower intercostal veins.
At a lower level, the hemiazygos vein (HAzV) is visible to the left of the spine, posterior to the aorta. It is smaller than the azygos vein,
and less commonly seen. The small vein branch connecting the hemiazygos and azygos veins is visible (arrow in I) between the aorta and the
vertebral column.
G H
I
Figure 2-122 Schematic drawing of the aortic arch and great

vessels. Characteristic levels have been labeled A through E.
A B
C D
Figure 2-123 Cross-sectional CT anatomy, great vessels. A–D: Sequential contrast-enhanced CT images through the great vessels, from
below-upward, corresponding to levels B, C, D, and E in Figure 2-122, respectively. Note that the left brachiocephalic vein (LBV) has a much
more horizontal course than the right brachiocephalic vein (RBV) as it crosses the mediastinum. The left brachiocephalic vein is an important
landmark, dividing the mediastinum into pre- and retrovascular compartments. Tr, trachea; LSA, left subclavian artery; LCCA, left common
carotid artery; BA, brachiocephalic artery; RSA, right subclavian artery; RCCA, right common carotid artery; E, esophagus.
A B
Figure 2-124 Obstruction, brachiocephalic vein. A, B: Sequential images through the great vessels and the aortic arch, respectively, fol-
lowing a bolus of intravenous contrast media administered through a left-sided antecubital vein. The left brachiocephalic vein is markedly at-
tenuated, the result of a previous indwelling venous catheter (black arrows in A and B). Note that there is bright opacification of both the
azygos (curved arrows in A and B) and hemiazygos systems (curved arrow in A), including the left superior intercostal vein beginning at the
left brachiocephalic vein (straight white arrow). The azygos and hemiazygos venous systems serve as an important collateral pathway when
there is obstruction of either the brachiocephalic veins or the vena cava.
The azygos and hemiazygos veins represent the thoracic

continuations of the right and left ascending lumbar veins,
respectively, and serve to drain many of the intercostal and
lumbar veins and systemic veins of the mediastinum. The
azygos and hemiazygos venous systems and their branches
serve as important collateral pathways when there is
obstruction or interruption of blood flow through either
of the brachiocephalic vessels or the superior or inferior
vena cava (323) (Fig. 2-124).
The azygos vein lies on the right, anterior and slightly
lateral to the vertebral column; it is almost always visible
on CT. The azygos vein drains the lower intercostal veins
(Figs. 2-121 and 2-125). On the left side, the hemiazygos
vein lies posterior to the descending aorta and adjacent to
the spine, draining the left lower intercostal veins. It is
smaller than the azygos vein, and less commonly seen. The
hemiazygos vein terminates in the azygos vein through
communicating branches that cross the midline, posterior
to the descending aorta, in the vicinity of the T8 vertebral
body. These are sometimes seen on CT (324).
Figure 2-125 Enlargement of a section through the lower
Superiorly, the azygos vein arches over the medial thorax shows a normal-sized intercostal vein draining into the
aspect of the right upper lobe bronchus and then courses azygos vein (arrow).
anteriorly to terminate in the posterior portion of the
SVC (Figs. 2-121G–I and 2-123). Just below its arch, the
azygos vein is joined by the right superior intercostal vein
(Figs. 2-121D–F and 2-126). This vein drains the right sec- inferior vena cava, or azygos continuation of an anomalous
ond to fourth intercostal veins and then courses inferiorly IVC (Fig. 2-127). In some normal subjects, the azygos arch
in a paraspinal location to its junction with the azygos. It appears dilated or contacts the posterior wall of the right
is occasionally seen on CT in normal individuals. upper lobe or right main bronchus; in either instance the
The arch of the azygos is variable in size, but can meas- azygos arch may be mistaken for a mass lesion (325,326).
ure up to 15 mm in diameter in normal subjects in the In most subjects the azygos vein remains unopacified fol-
supine position. Dilatation of the arch may be caused lowing intravenous contrast administration, but occasional
by central venous hypertension, as may occur with right reflux into the azygos arch and azygos vein occurs in nor-
ventricular heart failure, obstruction of the superior or mal subjects (Fig. 2-128).
A B
Figure 2-126 A: CT section through the aortic arch (Ao), corresponding to level A in Figure 2-122. At this level, the anterior mediastinum
has a triangular configuration, with the apex pointing anteriorly (arrows). SVC, superior vena cava; Tr, trachea; E, esophagus. B: CT section at
the level of the aortic arch shows a prominent right internal mammary vein draining into the superior vena cava (open arrow). This appear-
ance should not be mistaken for residual thymic tissue. The curved arrow points to the right superior intercostal vein, which characteristically
lies in a paravertebral location.
A, B
Figure 2-127 Azygos continuation. A–D:

Coned-down views of the mediastinum from
the level of the azygos arch superiorly to the
lower thorax. The dilated hemiazygos vein is
easily identified, especially at the level of the
azygos arch (arrow in A). Note that the azy-
gos vein is as large as the descending aorta
C, D (arrow in D).
On the left side, the accessory hemiazygos vein contin- joined by the left superior intercostal vein (which drains
ues above the point of termination of the hemiazygos vein, the second to fourth intercostal veins) in approximately
ascending posterior to the descending aorta (Fig. 2-124). At 75% of patients (321). The left superior intercostal vein
the level of the aortic arch, the accessory hemiazygos is forms a venous arch, sometimes referred to as the arch of
the hemiazygos vein, that courses anteriorly around the
aortic arch, near the level of T3 to T4, to join the left
brachiocephalic vein superiorly (Fig. 2-129). The left supe-
rior intercostal vein is occasionally seen in normal subjects
(Fig. 2-121); it may opacify in some normal individuals
following contrast injection into the left arm. On routine
radiographs, the left superior intercostal vein is most often
seen end-on, adjacent to the aortic knob. Because of its
appearance, it has been termed the “aortic nipple,” and can
be seen in up to 10% of normal patients (321). Prominence
of the aortic nipple has been reported secondary to congen-
ital absence of the azygos vein (327) or, more importantly,
in the presence of SVC obstruction (327,328) (Figs. 2-125
and 2-130).
Small intrathoracic veins can sometimes be seen, even
when unopacified. Intercostal veins are commonly visible
in the paravertebral regions, in association with the azygos
or hemiazygos veins. Internal mammary veins are also
often visible, and in some patients can be traced from the
point where they communicate with the brachiocephalic
veins, or on the right, the SVC. Paraesophageal veins, veins
Figure 2-128 Reflux of contrast agent into the azygos arch
(arrows) following intravenous contrast infusion. This is seen in in the paravertebral venous plexus, and chest wall veins
normal individuals. occasionally opacify after contrast infusion, even in the
A B
C D
Figure 2-129 Left superior intercostal vein, hemiazygos continuation due to innominate vein stenosis, resulting in potential pitfall in
diagnosing aortic dissection and aortic intramural hematoma. See Color Figure 2-129A,B. A: Axial volume-rendered multidetector CT
(MDCT) image from a left arm injection shows high attenuation in the azygos arch (thin arrow) and in a dilated left superior intercostal vein
(thick arrow). B: Sagittal reformatted volume-rendered MDCT image shows dilated left superior intercostal vein (thick arrow) refluxing into
the hemiazygos vein (thin arrow) due to distal innominate venous stenosis (not shown). C: Axial source image from fat-suppressed three-di-
mensional gadolinium-enhanced MR image in the same patient shows “pseudodissection” from the parallel course of the left superior inter-
costal vein and the aortic arch (arrow). D: Axial MDCT image in another patient shows an unopacified left superior intercostal vein from a
right-sided injection with apparent displaced intimal calcification (arrow) mimicking aortic intramural hematoma.
absence of venous obstruction, but are not easily seen on ways. Four major collateral pathways between the superior
unopacified scans. and inferior vena cava have been identified and described
Intrathoracic veins, other than the SVC and brachio- (323), including the (a) azygos and hemiazygos veins,
cephalic veins, commonly enlarge in the presence of (b) the vertebral venous plexus, (c) the internal mammary
venous obstruction, serving as important collateral path- veins, and (d) the lateral thoracic veins in the thoracic and
A, B C
Figure 2-130 Dilated left superior intercostal vein forming a hemiazygos arch in a patient with a left superior vena cava. A: The most
cephalad scan shows the left brachiocephalic vein (LBV) and left superior vena cava (LSVC). These opacified after contrast infusion into the
right arm. There is no evidence of a right superior vena cava. B: At the level of the proximal descending aorta (Desc Ao), the left superior
intercostal vein (LSIV) is easily seen passing lateral to the aorta. The azygos arch is absent. C: At the level of the left pulmonary artery (LPA),
the posterior portion of the left superior intercostal vein (LSIV) is visible posterior to the descending aorta. In this patient, the left superior
intercostal vein is serving the same function as the azygos arch in a normal patient.
abdominal walls. Opacification of these veins following of normal individuals, some opacification of collateral
peripheral contrast infusion suggests venous obstruction veins can be expected on CT following peripheral contrast
but can also be seen in normal persons. In a CT study by injection, but in recent studies, opacification of one or two
Kim et al. (329) of patients with intrathoracic venous veins has been reported in as many as 34% of normal sub-
obstruction, the presence of opacified venous collaterals jects. Three (6%) of 50 normal subjects studied by Kim
was 96% sensitive and 92% specific for the diagnosis of et al. (329) showed opacification of parascapular chest
symptomatic SVC syndrome; the most common collateral wall veins or veins in the vertebral venous plexus following
veins to opacify were parascapular (71%), vertebral (58%), contrast infusion. In general, opacification of collateral
lateral thoracic (50%), internal mammary (38%), right veins should be considered normal unless associated with
superior intercostal (38%), azygos and hemiazygos (29%), significant dilatation or other evidence of obstruction of
anterior cervical (25%), and left superior intercostal (21%) the SVC or brachiocephalic veins.
(329). Periesophageal and intercostal vein opacification Dilatation of the azygos and hemiazygos veins is classi-
can also be seen but is less common (323). In 5% to 10% cally associated with anomalies of the IVC (Fig. 2-131).
A B
Figure 2-131 Images from a 32-year-old man with inferior vena cava thrombosis (not shown). A, B: Sequential contrast-enhanced helical
CT scans demonstrate a massive azygos arch and the presence of a giant varix of the left internal mammary vein (arrows).
When there is developmental failure of the hepatic or dinal veins and the left horn of the sinus venosus, structures
infrahepatic (prerenal) segment of the IVC, blood returns that ordinarily regress. A right SVC may or may not be
to the heart via the cranial portion of the supracardinal present. The clinical significance of this anomaly is minimal
veins (i.e., the azygos and hemiazygos veins). This anom- unless there is an associated atrial septal defect, with a
aly has been reported in up to 2% of cases in patients with resultant left to right shunt. Partial anomalous pulmonary
congenital heart disease undergoing cardiac catheteriza- venous connection (PAPVC) from the left upper lobe
tion. The association between azygos continuation and the superior pulmonary vein may appear similar to a left-sided
asplenia and polysplenia syndromes has been well docu- SVC, as they both connect to the left innominate vein (330).
mented. Azygos continuation may also be present in However, left upper lobe PAPVR does not drain into the
otherwise asymptomatic patients, and in this setting may coronary sinus (Fig. 2-134), and no aberrant vessel is seen
be misinterpreted as some other form of pathology, specif- anterior to the left mainstem bronchus. When necessary,
ically, a right paratracheal mass, a posterior mediastinal 3D reformatted images may prove useful in differentiating
mass (if the dilated azygos or hemiazygos vein is identified these two anomalies. This anomaly represents an extracar-
along the paravertebral pleural reflections), or a retrocrural diac left to right shunt and may be associated with a sinus
mass or adenopathy. venosus atrial septal defect, although most adult patients
The appearance of azygos continuation is easily defined are asymptomatic (330). Upper lobe PAPVR is less common
on CT by the following constellation of findings: enlarge- on the right side in adults and may drain into the SVC,
ment of the arch of the azygos, enlargement of the azygous vein, right atrium, coronary sinus, or IVC.
paraspinal portions of the azygos and hemiazygos veins Haramati et al. (330) reviewed the imaging findings
(especially if confluent at higher sections with the azygos of partial anomalous pulmonary venous return in
arch), and enlargement of the retrocrural portions of these 29 patients. Seventy-nine percent (23 of 29 patients) had
same veins in the absence of a definable IVC. It should be an anomalous left upper lobe vein connecting to a persist-
noted that once an abnormality of the IVC is diagnosed, ent left vertical vein, only 5% (1 of 23 patients) of whom
careful examination of the abdomen should be performed had a left upper lobe vein in the normal location.
to further define and clarify its specific nature. Seventeen percent (5 of 29 patients) had an anomalous
In addition to congenital (and acquired) abnormalities right upper lobe vein draining into the SVC, 60% (3 of
of the azygos-hemiazygos venous system, the most com- 5 patients) of whom also had a right upper lobe
mon, clinically significant venous anomaly is persistence of pulmonary vein in the normal location. One patient (3%)
the left SVC. The left SVC forms from a confluence of the had an anomalous right lower lobe vein draining into
left subclavian and left jugular veins and courses inferiorly the suprahepatic IVC. CT findings included cardiomegaly
in a position analogous to the normal SVC on the right side in 48% (14 of 29 patients), right atrial enlargement in
(Figs. 2-132 and 2-133). Inferiorly, the left SVC lies anterior 31% (9 of 29 patients), right ventricular enlargement in
to the left hilum and always drains into a markedly dilated 31% (9 of 29 patients), and pulmonary artery enlarge-
coronary sinus. The anatomic course of the left SVC reflects ment in 14% (4 of 29 patients). Pulmonary or cardiovas-
embryologic retention of the left anterior and common car- cular symptoms were present in 69% (20 of 29 patients),
55% (11 of 20 patients) of whom had specific alternative
diagnoses (excluding congestive heart failure and pul-
monary hypertension) to explain the symptoms. Only
1 patient (3%) was diagnosed with a secundum atrial sep-
tal defect.
Acquired Venous Abnormalities

Venous obstruction, caused by extrinsic processes, is easily
detected with both CT and MR imaging (331). Most often
there is direct visualization of the cause of obstruction,
although chronic obstructions may result in a shrunken,
cordlike vein. SVC obstruction most often results from
bronchogenic carcinoma (Fig. 2-135) (332). Less frequently,
obstruction results from either primary mediastinal
tumors, including lymphoma, thymoma, or seminoma, or
from metastatic disease, especially breast cancer. Rarely,
venous obstruction results from primary vascular neo-
Figure 2-132 Schematic drawing of a persistent left-sided plasms, in particular, angiosarcomas, or by direct
superior vena cava, posterior view. The left-sided vena cava passes
in front of the left hilum to drain inferiorly into an enlarged coro- extension of tumor thrombus in the subclavian or innomi-
nary sinus. LA, left atrium. nate veins (Fig. 2-136).
A B
C D
E F
Figure 2-133 Left-sided superior vena cava draining into the coronary sinus. A–F: Enlargements of sequential CT images from the great
vessels superiorly to the aortic arch inferiorly, respectively. In addition to a left-sided vena cava (arrows in A to D), there is persistence of a
right-sided vena cava (S). The vessel forms from a confluence of the left subclavian and jugular veins (not shown). Inferiorly, the left SVC lies
anterior to the left hilum and always drains into a dilated coronary sinus (labeled C in F). The clinical significance of this anomaly is minimal
unless there is an associated atrial septal defect.
200
A B
C D
Figure 2-134 Left upper lobe partial anomalous pulmonary

venous return (PAPVR). A–D: Sequential contrast-enhanced CT
scans demonstrate an anomalous left upper lobe pulmonary vein
(arrows) draining into the left innominate vein. E: Multiplanar recon-
struction shows the anomalous pulmonary vein and its course
(arrow). Unlike a left superior vena cava, left upper lobe PAPVR does
not drain into the coronary sinus and no aberrant vessel is seen
E anterior to the left mainstem bronchus.
A B
Figure 2-135 Superior vena caval thrombosis from metastatic

lung cancer. A–C: Sequential contrast-enhanced CT sections
through the left brachiocephalic vein, the superior vena cava, and
the right atrium, respectively. A well-defined filling defect is appar-
ent (arrows in A–C) extending from the left brachiocephalic vein to
the right atrium. There is considerable nodularity within the medi-
astinal fat, especially at the level of the great vessels. This finding
was initially thought to represent dilated collateral mediastinal
veins. Subsequent mediastinoscopy disclosed metastatic lymph
C nodes consistent with a primary lung cancer.
Indirect signs of SVC obstruction, such as collateral neoplastic in origin. Caution should be exercised so that
circulation in the veins of the chest wall or azygos vein, venous thrombosis is not diagnosed on scans obtained
also are useful. However, it should be emphasized that during or immediately after a bolus injection of contrast
venous enlargement alone should not be relied on as a agent, because flow phenomena may mimic filling defects
sign of obstruction because it is dependent on the respira- in mediastinal veins. This may be especially problematic
tory cycle or a Valsalva maneuver and sometimes repre- in the SVC adjacent to the azygos arch, as densely en-
sents a normal variant. hanced SVC blood draining the injection mixes with
Intrinsic obstruction, caused by thrombosis, is becom- unenhanced azygos blood returning from the lower body.
ing more common owing to the increasing use of central A peculiar abdominal manifestation of SVC occlusion
venous catheters, larger bore catheters for hyperalimenta- is the presence of a wedge-shaped region of increased den-
tion, and the increasing number of procedures involving sity within the anterior aspect of the left lobe of the liver
passage of tubes into the upper body veins. These include (333–335) (Fig. 2-137). This likely occurs because of the
Swan-Ganz catheterization, pacemaker placement, dialysis development of systemic-portal collaterals probably from
catheters, and, more recently, peripherally inserted central the internal thoracic vein and the paraumbilical vein
lines. (336). Other findings on abdominal CT examinations
The classic CT signs of venous thrombosis are when a central chest vein is obstructed include the pres-
(a) enlargement of the vein, (b) a relatively lucent center, ence of dense enhancement in the azygos, hemiazygos,
and (c) enhancing vein wall on contrast-enhanced studies intercostal, epigastric, subcutaneous, and intramuscular
(Fig. 2-136B). If the thrombus itself enhances it may be veins (335–337).
A B
Figure 2-136 Comparison of bland with tumor thrombus. A: There is neoplastic infiltration of the left subclavian vein from direct exten-
sion of an axillary mass (large arrow). Note the tumor thrombus within the left subclavian vein (small arrow) markedly expands the vessel and
enhances. B: Bland thrombosis of the left innominate vein (arrow) does not expand the vessel and has central low attenuation. Note the
absence of collateral vessels due to right-sided injection.
A, B
C, D
Figure 2-137 Superior vena caval ligation. A–D: Sequential contrast-enhanced scans through the mediastinum from the great vessels
to the subcarinal space, respectively. The superior vena cava (arrow in C) had been surgically ligated just below the aortic arch (arrow in
D). There is marked enlargement of the azygos vein (curved arrows in C and D) as well as enlargement of numerous small mediastinal
veins, especially in the prevascular space anteriorly. E, F: Sequential images through the liver in the same patient as seen in A–D are
shown. There is marked contrast enhancement of the azygos and hemiazygos veins. In addition, there is dense opacification of the
inferior vena cava (open arrows), which is presumably filling through retrograde flow through intercostal and phrenic veins (arrows in
E and F). Finally, there is an area of density in the anterior portion of the liver (curved arrow in F), probably the result of focal systemic-
portal collaterals (continued).
E F
Involvement of the IVC from either thrombus or neo- image quality and to minimize in-plane saturation effects,
plasm is evaluated in a similar fashion (337). An the slice orientation should be perpendicular to the vessel of
anatomic variant, which can be mistaken for an intra- interest. The subclavian vein runs in a horizontal direction
caval mass, is the presence of prominent pericaval fat and is best imaged with a sagittal oriented scan to maximize
(338,339) (Fig. 2-138). This is seen more frequently in inflow of unsaturated blood. Similarly, the jugular veins and
patients with cirrhosis and is best evaluated using coro- SVC run in a near vertical course and are optimally imaged
nal reformations. with axial slices. The brachiocephalic veins run in an
oblique course and are more difficult to image for these
anatomic reasons.
Magnetic Resonance Venography
A slice thickness of 2 to 3 mm should be adequate, and
a body phased-array coil can be used. Breath holding is
Technical Considerations
useful for evaluation of the proximal thoracic veins, as it
Recent improvements in MR technology have led to the eliminates respiratory artifacts that degrade the MIP images.
need to seriously reconsider the role of MR not only for Because the venous structures that form the SVC are ori-
routine aortic imaging but also for routine imaging of ented in different planes and may have varying velocities
intrathoracic venous structures. 2D TOF venography may be and flow patterns, TOF imaging may be suboptimal in
used for the evaluation of the thoracic and upper extremity evaluating the entire thoracic venous system. Flow artifacts
veins that run in a vertical orientation to optimize the at branch points and in-plane saturation effects may result
inflow effect (Fig. 2-139). Arterial signal is readily elimi- in suboptimal image quality with artifactual signal loss
nated with a traveling presaturation pulse. To optimize that can mimic thrombus. Finally, 2D TOF is unable to
A B
Figure 2-138 Pericaval fat mimicking intracaval mass. A, B: Sequential contrast-enhanced helical CT scans at the level of the dome of the
liver demonstrate pericaval fat (arrows) that appears to be within the inferior vena cava. This is quite common and is best demonstrated as
pericaval with coronal reformations.
Figure 2-139 Time of flight MR venography in the

chest. Axial section through the great vessels using
gradient-recalled echoes (TR 25 ms, TE 13 ms, flip
angle 20 degrees). In these images, flowing blood
results in considerable signal within vessels, not only in
the mediastinum (straight arrows) but in the chest wall
as well (curved arrow).
visualize collapsed veins with little or no flow and cannot 3D venograms may contain both arterial and venous
reliably differentiate acute from chronic thrombus. signal. A selective venogram can be generated using a
As an alternative approach, excellent opacification of subtraction technique, in which an arterial phase image is
the venous system is attainable (without bothersome flow subtracted from a mixed venous-arterial image. These 3D
artifacts or saturation effects) by using intravenously datasets provide for high-quality multiplanar reformations
administered gadolinium chelates in conjunction with 3D that can demonstrate long segment thrombosis with a sin-
gradient-echo pulse sequences (340–342). As previously gle image. An additional benefit of gadolinium-enhanced
described with arterial imaging, this technique eliminates techniques is that in the setting of acute or subacute deep
the need for flow-related enhancement, as image quality venous thrombosis periadvential venous enhancement
simply depends on the shortened T1 relaxation time of may be present (Fig. 2-141). This can further increase diag-
gadolinium-enhanced venous blood. (Fig. 2-140). In con- nostic confidence in difficult cases.
trast to first pass arterial imaging, longer scan delay time is When performing dynamic gadolinium-enhanced
needed to evaluate the great veins. Because this is a recircu- 3D-GRE studies, several acquisitions may be obtained
lation technique (peripheral vein–aorta–central vein), a sequentially. The initial acquisition is timed with the
dose of 0.2 mmol/kg of gadolinium is required. These gadolinium’s first pass through the arterial system but
prior to substantial venous enhancement. A second acqui-
sition is acquired immediately following the arterial
phase; this phase usually has both arterial and venous
enhancement (Fig. 2-142). The subtraction technique
requires special software, which performs a pixel-by-pixel
subtraction of the arterial-phase study from a mixed
venous-arterial phase study. The arterial signal is nullified,
whereas the subtracted dataset contains only venous
signal (340).
Recently, a new MR venography (MRV) technique, simi-
lar to conventional venography, has been advocated. This
requires the injection of dilute gadolinium (3 mL of
gadolinium in 57 mL of saline) into the extremity of
interest, while performing multiple, dynamic 3D gradient-
echo sequences (343). Limitations of this technique include
potentially difficult intravenous access in the extremity
ipsilateral to the expected abnormality and the fact that only
a single unilateral drainage system is typically assessed.
Finally, nonvisualization of the deep venous system may
occur without inherent pathology in these vessels.
Figure 2-140 Occlusion of all the great central veins due to

chronic indwelling catheters. Coronal maximum intensity projec- Clinical Applications
tion image from gadolinium-enhanced 3D MR venogram shows
absence of all the deep veins of the chest and superior vena cava As in the lower extremities, conventional contrast venogra-
occlusion. Note the extensive collateral vessels. phy is limited in the chest and upper extremities. Patients
A B
Figure 2-141 Suspected right upper extremity deep venous

thrombosis and indeterminate CT in an athletic male. A: Axial
multidetector CT (MDCT) performed via a left arm injection
demonstrates enlargement and haziness of the right axillary
vessels without a definite clot (arrow). B: Axial TurboSTIR MR
image shows intense axillary edema (arrow). C: Coronal reformat
from gadolinium-enhanced three-dimensional MR venogram
shows extensive low signal intensity thrombus surrounded by
intense periadventitial enhancement consistent with acute deep
venous thrombosis. This case shows the importance of injecting on
the side ipsilateral to the swollen extremity on MDCT. Patient
C subsequently underwent thoracic outlet decompression.
may have limited or no venous access, veins may be ob- Hartnell et al. (344) evaluated 84 patients using breath-
scured by adjacent masses, and high flow in the SVC may hold TOF MR to identify sites for central venous access or
make adequate opacification difficult. Sonography is to diagnose and stage central venous occlusion. Although
limited by poor acoustical access from the chest wall and MRV predicted a patent site for central venous access in
air-filled lungs. Because of these limitations, MRV is a use- 28 of 28 patients, contrast venography did not show all
ful modality that eliminates the need for other imaging patent veins. The authors concluded that both patent and
studies (344). occluded chest veins are reliably defined with MRV, includ-
Finn et al. (345) evaluated 30 patients with suspected ing potential sites for central line placement, in a way that
thoracic venous occlusion and found MR imaging provided is not possible with other techniques and proposed that
more comprehensive information on central venous MRV may be the new “gold standard” for defining systemic
anatomy and blood flow than catheter venography. MRV venous anatomy in the chest (344). Shinde et al. (341) con-
also provided useful clinical information in patients with- firmed these findings in a smaller group of patients using
out venous access. MRV is often useful when attempting to contrast-enhanced 3D venography. Pedrosa et al. (346) re-
assess complex venous anatomy in postoperative patients cently evaluated the use of balanced SSFP (True FISP) MR
or in patients with congenital anomalies. MRV can also imaging for deep venous thrombosis and demonstrated a
assist in determining suitable sites for venous access. lower sensitivity and specificity in the diagnosis of venous
Figure 2-142 Acute right-sided upper extremity deep venous

thrombosis confirmed on MR imaging. A: Axial enhanced multide-
tector CT (MDCT) with injection ipsilateral to the swollen arm
shows numerous collateral vessels and a filling defect in the right
axillary vein (arrow) consistent with acute thrombus. B: Coronal
source image from venous phase of gadolinium-enhanced three-
dimensional MR venogram shows acute right axillary and subcla-
vian vein deep venous thrombosis (arrow) as low signal with intense
periadventitial enhancement. C: Coronal maximum intensity projec-
tion image from the same dataset fails to demonstrate low signal
intensity thrombus but demonstrates the most proximal extent of
C the clot (arrow). Note the normal arterial vessels.
thrombosis than gadolinium-enhanced 3D MRV. The notion that tumor thrombus will enhance after the admin-
authors concluded that these balanced SSFP sequences istration of gadolinium owing to tumor angiogenesis,
should not be used for definitive diagnosis of venous whereas bland thrombus will not. This is best demonstrated
thrombosis. using breath-hold gadolinium-enhanced 3D MR imaging
Neoplastic involvement of the IVC from intra-abdominal acquired in the coronal plane. When the precontrast image
malignancies usually occurs from hepatocellular carci- is subtracted from a delayed postcontrast acquisition,
noma, renal cell carcinoma (Fig. 2-143), adrenal cortical enhancement of tumor thrombus is more easily visualized
carcinoma (Fig. 2-144), retroperitoneal neoplasms, and (Fig. 2-144).
endometrial carcinoma. Less commonly, primary sarcomas Traumatic injury to the systemic thoracic veins can
may arise from the wall of the IVC. Using 2D TOF MRV, be evaluated with both MDCT and MR imaging (348)
demonstration of thrombus within the IVC is readily (Fig. 2-145). As with arterial injuries, active extravasation
achieved, but tumor thrombus may not be distinguished and collections of contrast outside of the expected lumen
from bland thrombus (347). The ability to differentiate are direct signs of injury and mediastinal hematomas are
bland thrombus from tumor thrombus is predicated on the indirect signs.
Figure 2-145 Venous injury from chest tube placement dem-

onstrated with multidetector CT (MDCT). Axial MDCT image shows
Figure 2-143 Extensive enhancing tumor thrombus in the infe- a complex right hemothorax with well-circumscribed collection of
rior vena cava from renal cell carcinoma: MR demonstration. contrast (arrow) in the region of the azygous vein resulting from
Coronal source image from venous phase of gadolinium-enhanced right chest tube placement.
three-dimensional MR venogram demonstrates enhancing tumor
thrombus (arrows) expanding the inferior vena cava.
REFERENCES
1. Nienaber CA, Eagle KA. Aortic dissection: new frontiers in diag-

nosis and management: Part I: from etiology to diagnostic
strategies. Circulation. 2003;108:628–635.
2. Sommer T, Fehske W, Holzknecht N, et al. Aortic dissection:
a comparative study of diagnosis with spiral CT, multiplanar
transesophageal echocardiography, and MR imaging. Radiology.
1996;199:347–352.
3. Adachi H, Omoto R, Kyo S, et al. Emergency surgical interven-
tion of the acute aortic dissection with the rapid diagnosis by
transesophageal echocardiography. Circulation. 1991;84:
14–19.
4. Evangelista A, Avegliano G, Elorz C. Transesophageal echocar-
diography in the diagnosis of acute aortic syndrome. J Card
Surg. 2002;17:95–106.
5. Prince MR, Narasimham DL, Jacoby WT, et al. Three-
dimensional gadolinium-enhanced MR angiography of the
thoracic aorta. Am J Roentgenol. 1996;166:1387–1397.
6. Fishman EK, Lawler LP. CT angiography: principles, techniques
and study optimization using 16-slice multidetector CT with
isotropic datasets and 3D volume visualization. Crit Rev Comput
Tomogr. 2004;45:355–388.
Figure 2-144 Adrenal cortical carcinoma invading the inferior
7. Napoli A, Fleischmann D, Chan FP et al. Computed tomogra-
vena cava and right atrium—morphologic findings and discrimina-
phy angiography: state-of-the-art imaging using multidetector-
tion between tumor and bland thrombus. Discrimination is based
row technology. J Comput Assist Tomogr. 2004;28:S32–45.
on enhancement with gadolinium-subtracted coronal image
8. Flohr TG, Schaller S, Stierstorfer K, et al. Multi-detector row CT
(venous phase-precontrast acquisition), which demonstrates a
systems and image-reconstruction techniques. Radiology.
large thrombus in the inferior vena cava (arrows 2 and 3) and right
2005;235:756–773.
atrium (arrow 1). Note the differentiation between the enhancing
9. Fleischmann D, Hallett RL, Rubin GD. CT angiography of
tumor thrombus superiorly (arrow 2) and the nonenhancing bland
peripheral arterial disease. J Vasc Interv Radiol. 2006;17(1):3–26.
thrombus just inferior to it (arrow 3). The neoplasm is shown
10. Utsunomiya D, Awai K, Tamura Y, et al. 16-MDCT aortography
by arrow 4 and the left renal vein by arrow 5.
with a low-dose contrast material protocol. AJR Am J Roentgenol.
2006;186:374–378.
11. Haage P, Schmitz-Rode T, Hubner D, et al. Reduction of contrast 32. Prince MR, Narasimham DL, Londy FJ, et al. Breath-hold
material dose and artifacts by a saline flush using a double gadolinium-enhanced MR angiography of the abdominal aorta
power injector in helical CT of the thorax. AJR Am J Roentgenol. and its major branches. Radiology. 1995;197:785–792.
2000;174:1049–1053. 33. Levy R, Prince M. Arterial-phase three-dimensional contrast-
12. Fleischman D, Rubin GD. Quantification of intravenously enhanced MR angiography of the carotid arteries. AJR Am
administered contrast medium transit through the peripheral J Roentgenol. 1996;167:211–215.
arteries: implications for CT angiography. Radiology. 2005;236: 34. Krinsky G, Rofsky N. MR angiography of the aortic arch vessels
1076–1082. and upper extremities. Magn Reson Imaging Clin North Am.
13. Bae KT, Tran HQ, Heiken JP. Uniform vascular contrast 1998;6:269–292.
enhancement and reduced contrast medium volume achieved 35. Earls J, Rofsky N, DeCorato D, et al. Breath-hold single-dose
by using exponentially decelerated contrast material injection gadolinium-enhanced MR aortography: usefulness of a timing
method. Radiology. 2004;231:732–736. examination and MR power injector. Radiology. 1996;201: 705–710.
14. Brink JA. Contrast optimization and scan timing for single and 36. Wu C, Zhang J, Ladner CJ, et al. Subclavian steal syndrome:
multidetector-row computed tomography. J Comput Assist Tomogr. diagnosis with perfusion metrics from contrast-enhanced MR
2003;27:S3–8. angiographic bolus-timing examination—initial experience.
15. Brink JA. Use of high concentration contrast media (HCCM): Radiology. 2005;235:927–933.
principles and rationale—body CT. Eur J Radiol. 2003;45: 53–58. 37. Francois CJ, Shors SM, Bonow RO, Finn JP. Analysis of
16. Ho LM, Nelson RC, Thomas J, et al. Abdominal aortic aneurysms cardiopulmonary transit times at contrast material-enhanced
at multi-detector row helical CT: optimization with interactive MR imaging in patients with heart disease. Radiology. 2003;
determination of scanning delay and contrast medium dose. 227:447–452.
Radiology. 2004;232:854–859. 38. Wilman A, Riederer S, King B, et al. Fluoroscopically triggered
17. Awai K, Hiraishi K, Hori S. Effect of contrast material injection contrast-enhanced three-dimensional MR angiography with
duration and rate on aortic peak time and peak enhancement at elliptical centric view order: application to the renal arteries.
dynamic CT involving injection protocol with dose tailored to Radiology. 1997;205:137–146.
patient weight. Radiology. 2004;230:142–150. 39. Prince MR, Chenevert TL, Foo TK, et al. Contrast-enhanced
18. Hopper KD, Mosher TJ, Kasales CJ, et al. Thoracic spiral CT: abdominal MR angiography: Optimization of imaging delay
delivery of contrast material pushed with injectable saline solu- time by automating the detection of contrast material arrival in
tion in a power injector. Radiology. 1997;205:269–271. the aorta. Radiology. 1997;203:109–114.
19. Vanhoe L, Marchal G, Baert AL, et al. Determination of scan 40. Korosec F, Frayne R, Grist T, Mistretta C. Time-resolved contrast-
delay time in spiral CT-angiography: utility of a test bolus injec- enhanced 3D MR angiography. Magn Reson Med. 1996;36:
tion. J Comput Assist Tomogr. 1995;19:216–220. 345–351.
20. Schertler T, Glucker T, Wildermuth S, et al. Comparison of retro- 41. Fink C, Ley S, Kroeker R, et al. Time-resolved contrast-enhanced
spectively ECG-gated and nongated MDCT of the chest in an three-dimensional magnetic resonance angiography of the
emergency setting regarding workflow, image quality, and diag- chest: combination of parallel imaging with view sharing
nostic certainty. Emerg Radiol. 2005;12:19–29. (TREAT) Invest Radiol. 2005;40:40–48.
21. Hayter RG, Rhea JT, Small A, et al. Suspected aortic dissection 42. Wu Y, Goodrich KC, Buswell HR, et al. High-resolution time-
and other aortic disorders: multi-detector row CT in 373 cases resolved contrast-enhanced 3D MRA by combining SENSE with
in the emergency setting. Radiology. 2006;238:841–852. keyhole and SLAM strategies. Magn Reson Imaging. 2004;22:
22. Zhang HL, Ersoy H, Prince MR. Effects of gadopentetate dimeg- 1161–1168.
lumine and gadodiamide on serum calcium, magnesium, and 43. Hennig J, Scheffler K, Laubenberger J, Strecker R. Time-resolved
creatinine measurements. J Magn Reson Imaging. 2006;23: projection angiography after bolus injection of contrast agent.
383–387. Magn Reson Med. 1997;37:341–345.
23. Amparo EG, Higgins CB, Hricak H, et al. Aortic dissection: 44. Lee YJ, Chung TS, Joo JY, et al. Suboptimal contrast-enhanced
magnetic resonance imaging. Radiology. 1985;155:399–406. carotid MR angiography from the left brachiocephalic venous
24. Akins EW, Carmichael MJ, Hill JA, et al. Preoperative evaluation stasis. J Magn Reson Imaging. 1999;10:503–509.
of the thoracic aorta using MRI and angiography. Ann Thorac 45. Neimatallah MA, Chenevert TL, Carlos RC, et al. Subclavian MR
Surg. 1987;44:499–507. arteriography: reduction of susceptibility artifact with short
25. Kersting-Sommerhoff BA, Higgins CB, White RD, et al. Aortic echo time and dilute gadopentetate dimeglumine. Radiology.
dissection: sensitivity and specificity of MR imaging. Radiology. 2000;217:581–586.
1988;166:651–655. 46. Lee VS, Flyer MA, Weinreb JC, et al. Image subtraction in
26. Stemmerman DH, Krinsky GA, Lee VS, et al. Thoracic aorta: rapid gadolinium-enhanced MR imaging. AJR Am J Roentgenol. 1996;
black-blood MR imaging with half-Fourier rapid acquisition with 167:1427–1432.
relaxation enhancement with or without electrocardiographic 47. Bader TR, Semelka RC, Pedro MS, et al. Magnetic resonance
triggering. Radiology. 1999;213:185–191. imaging of pulmonary parenchymal disease using a modified
27. Pereles FS, McCarthy RM, Baskaran V, et al. Thoracic aortic breath-hold 3D gradient-echo technique: initial observations.
dissection and aneurysm: evaluation with nonenhanced true J Magn Reson Imaging. 2002;15:31–38
FISP MR angiography in less than 4 minutes. Radiology. 2002; 48. Both M, Schultze J, Reuter M, et al. Fast T1- and T2-weighted
223:270–274. pulmonary MR-imaging in patients with bronchial carcinoma.
28. Hartnell GG, Finn JP, Zenni M, et al. MR imaging of the thoracic Eur J Radiol. 2005;53:478–488.
aorta: comparison of spin-echo, angiographic and breath-hold 49. Simonetti OP, Finn JP, White RD, et al. “Black blood”
techniques. Radiology. 1994;191:697–704. T2-weighted inversion-recovery MR imaging of the heart.
29. Krinsky G, Reuss P. MR angiography of the thoracic aorta. Magn Radiology. 1996;199:49–57.
Reson Imaging Clin North Am. 1998;6:293–320. 50. Mani V, Itskovich VV, Szimtenings M, et al. Rapid extended cov-
30. Nael K, Laub G, Finn JP. Three-dimensional contrast-enhanced erage simultaneous multisection black-blood vessel wall MR
MR angiography of the thoraco-abdominal vessels. Magn Reson imaging. Radiology. 2004;232:281–288.
Imaging Clin North Am. 2005;13:359–380. 51. Mohiaddin RH, Kilner PJ, Rees S, et al. Magnetic resonance
31. Krinsky GA, Rofsky NM, DeCorato DR, et al. Thoracic aorta: volume flow and jet velocity mapping in aortic coarctation.
comparison of gadolinium-enhanced three-dimensional MR J Am Coll Cardiol. 1993;22:1515–1521.
angiography with conventional MR imaging. Radiology. 1997; 52. Riehle TJ, Oshinski JN, Brummer ME, et al. Velocity-encoded
202;183–193. magnetic resonance image assessment of regional aortic
flow in coarctation patients. Ann Thorac Surg. 2006;81: 77. Akers D, Fowl R, Plettner J, et al. Complications of anomalous
1002–1007. origin of the right subclavian artery: case report and review of
53. Ueno J, Murase T, Yoneda K, et al. Three-dimensional imaging the literature. Ann Vasc Surg. 1991;5:385–388.
of thoracic diseases with multi-detector row CT. J Med Invest. 78. Kieffer E, Bahnini A, Koskas F. Aberrant subclavian artery:
2004;51:163–170. Surgical treatment in thirty-three adult patients. J Vasc Surg.
54. Mallouhi A, Schocke M, Judmaier W, et al. 3D MR angiography 1994;19:100–111.
of renal arteries: comparison of volume rendering and maxi- 79. Shuford WH, Sybers RG, Weens HS. The angiographic features of
mum intensity projection algorithms. Radiology. 2002;223: double aortic arches. AJR Am J Roentgenol. 1972;116:125–140.
509–516. 80. Kennard DR, Spigos D, Tan WS. Cervical aortic arch: CT correla-
55. Baskaran V, Pereles FS, Nemcak AA Jr, et al. Gadolinium- tion with conventional radiologic studies. AJR Am J Roentgenol.
enhanced 3D MR angiography of renal artery stenosis: a pilot 1983;141:295–297.
comparison of maximum intensity projection, multiplanar 81. Flamm SD, VanDyke CW, White RD, et al. MR imaging of the tho-
reformatting, and 3D volume-rendering postprocessing algo- racic aorta. Magn Reson Imaging Clin North Am. 1996;4: 217–235.
rithms. Acad Radiol. 2002;9:50–59. 82. Bissett GSI, Strife JL, Kirks DR, Bailey WW. Vascular rings: MR
56. Lee EY, Siegel MJ, Hildebolt CF, et al. MDCT evaluation of imaging. AJR Am J Roentgenol. 1987;149:251–256.
thoracic aortic anomalies in pediatric patients and young 83. Gomes AS, Lois JF, George B, et al. Congenital abnormalities of
adults: comparison of axial, multiplanar, and 3D images. AJR the aortic arch: MR imaging. Radiology. 1987;165:691–695.
Am J Roentgenol. 2004;182:777–784. 84. Kersting-Sommerhoff BA, Sechten UP, Fisher MR, Higgins CB.
57. Napel S, Marks MP, Rubin GD, et al. CT angiography with spiral MR imaging of congenital anomalies of the aortic arch. AJR Am
CT and maximum intensity projection. Radiology. 1992;185: J Roentgenol. 1987;149:9–13.
607–610. 85. Azarow KS, Pearl RH, Hoffmann MA, et al. Vascular rings: does
58. Prokop M, Shin HO, Schanz A, Schaeferprokop CM. Use of magnetic resonance imaging replace angiography? Ann Thorac
maximum intensity projections in CT angiography: a basic Surg. 1992;53:882–886.
review. Radiographics. 1997;17:433–451. 86. Jaffe RB. Magnetic resonance imaging of vascular rings. Semin
59. Fishman EK, Magid D, Ney DR, et al. Three-dimensional imag- Ultrasound CT MR. 1990;11:206.
ing. Radiology. 1991;181:321–337. 87. Lam WW, Chan JHM, Hui Y, Chan FL. Non-breath-hold
60. Smith PA, Heath DG, Fishan EK. Virtual angioscopy using spiral gadolinium-enhanced MR angiography of the thoracoabdomi-
CT and real-time interactive column-rendering techniques. nal aorta; experience in 18 children. AJR Am J Roentgenol.
J Comput Assist Tomogr. 1998;22:212–214. 1998;170:478–480.
61. Johnson PT, Fishman EK, Duckwall JR, et al. Interactive three- 88. Katz M, Konen E, Rozenman J, et al. Spiral CT and 3D image
dimensional volume rendering of spiral CT data: current reconstruction of vascular rings and associated tracheo-
applications in the thorax. Radiographics. 1998;18:165–187. bronchial anomalies. J Comput Assist Tomogr. 1995;19:564–568.
62. Rubin GD, Beaulieu CF, Argiro V, et al. Perspective volume 89. Haramati LB, Glickstein JS, Issenberg HJ, et al. MR imaging and
rendering of CT and MR images: applications for endoscopic CT of vascular anomalies and connections in patients with
imaging. Radiology. 1996;199:321–330. congenital heart disease: significance in surgical planning.
63. Davis CP, Ladd ME, Romanowski BJ, et al. Human aorta: Radiographics. 2002;22:337–347.
preliminary results with virtual endoscopy based on three- 90. Hopkins KL, Patrick LE, Simoneaux SF, et al. Pediatric great
dimensional MR imaging data sets. Radiology. 1996;199:37–40. vessel anomalies: initial clinical experience with spiral CT
64. Davis CP, Hany TF, Wildermuth S, et al. Postprocessing angiography. Radiology. 1996;200:811–815.
techniques for gadolinium-enhanced three-dimensional MR 91. Cerillo AG, Amoretti F, Moschetti R, et al. Sixteen-row multislice
angiography. Radiographics. 1997;17:1061–1067. computed tomography in infants with double aortic arch. Int
65. Hany TF, Schmidt M, Davis CP, et al. Diagnostic impact of four J Cardiol. 2005;99:191–194.
postprocessing techniques in evaluating contrast-enhanced 92. Bailey WW. Interrupted aortic arch. Adv Cardiac Surg. 1994;
three-dimensional MR angiography. AJR Am J Roentgenol. 5:97–114.
1998;170:907–912. 93. Celoria G, Patton R. Congenital absence of the aortic arch. Am
66. White RD, Dooms GC, Higgins CB. Advances in imaging Heart J. 1959;58:407–412.
thoracic aortic disease. Invest Radiol. 1986;21:761–778. 94. Hernandez RJ, Aisen AM, Foo TKF, Beekman RH. Thoracic
67. Goss CM. Gray’s anatomy. Philadelphia: Lea & Febiger; 1996: 579. cardiovascular anomalies in children: evaluation with a fast gra-
68. Aronberg DJ, Glazer HS, Madsen K, Sagel SS. Normal thoracic dient-recalled-echo sequence with cardiac-triggered segmented
aortic diameters by computed tomography. J Comput Assist acquisition. Radiology. 1993;188:775–780.
Tomogr. 1984;8:247–250. 95. Roche KJ, Krinsky G, Lee VS, et al. Interrupted aortic arch: diag-
69. Fitzgerald AW, Donaldson JS, Poznanski AK. Pediatric thoracic nosis with gadolinium-enhanced 3D MRA. J Comput Assist
aorta: normal measurements determined with CT. Radiology. Tomogr. 1999;23:197–202.
1987;165:667–669. 96. Lee HY, Lee W, Chung JW, et al. Interrupted aortic arch with
70. Yamada T, Tada S, Harada J. Aortic dissection without intimal aberrant subclavian artery: a rare form of arch anomaly demon-
rupture: diagnosis with MR imaging and CT. Radiology. 1988; strated with multidetector CT and 3D reconstruction. Pediatr
168:347–352. Radiol. 2006;36:272–273.
71. Park JH, Chung JW, Im JG, et al. Takayasu arteritis: Evaluation of 97. Gibbs JL. Treatment options for coarctation of the aorta. Heart.
mural changes in the aorta and pulmonary artery with CT 2000;84:11–13.
angiography. Radiology. 1995;196:89–93. 98. Parks WJ, Ngo TD, Plauth WH Jr, et al. Incidence of aneurysm
72. Tenenbaum A, Garniek A, Shemesh J, et al. Dual-helical CT formation after Dacron patch aortoplasty repair for coarctation
for detecting aortic atheromas as a source of stroke: comparison of the aorta: long-term results and assessment utilizing mag-
with transesophageal echocardiography. Radiology. 1998;208: netic resonance angiography with three-dimensional surface
153–158. rendering. J Am Coll Cardiol. 1995;26:266–271.
73. Bosniak MA. Analysis of some anatomic-Roentgenologic aspects 99. Backer CL, Mavroudis C, Zias EA, et al. Repair of coarctation
of the brachiocephalic vessels. AJR Am J Roentgenol. 1964;91: with resection and extended end-to-end anastomosis. Ann
1222–1231. Thorac Surg. 1998;66:1365–1370.
74. Sutton D, Rhys Davies E. Arch aortography and cerebrovascular 100. Shaddy RE, Boucek MM, Sturtevant JE, et al. Comparison of
insufficiency. Clin Radiol. 1966;17:330–345. angioplasty and surgery for unoperated coarctation of the aorta.
75. McLoughlin MJ, Weisbrod G, Wise DJ, Yeung HPH. Computed Circulation. 1993;87:793–799.
tomography in congenital anomalies of the aortic arch and 101. Cowley CG, Orsmond GS, Feola P, et al. Long-term, random-
great vessels. Radiology. 1981;138:399–403. ized comparison of balloon angioplasty and surgery for native
76. Edwards J. Anomalies of derivatives of the aortic arch system. coarctation of the aorta in childhood. Circulation. 2005;111:
Med Clin North Am. 1948;32:925–949. 3453–3456.
102. Chessa M, Carozza M, Butera G, et al. Results and mid-long-term 126. Svensjo S, Bengtsson H, Bergqvist D. Thoracic and thoracoab-
follow-up of stent implantation for native and recurrent coarcta- dominal aortic aneurysm and dissection: an investigation based
tion of the aorta Eur Heart J. 2005;26:2728–2832. on autopsy. Br J Surg. 1996;83:68–71.
103. Almeida de Oliveira S, Lisboa D, Dallan La, et al. Extraanatomic 127. Bickerstaff LK, Pairolero PC, Hollier LH, et al. Thoracic
aortic bypass for repair of aortic arch coarctation via ster- aortic aneurysms. A population-based study. Surgery. 1982;92:
notomy: midterm clinical and magnetic resonance imaging 1103–1108.
results. Ann Thorac Surg. 2003;76:1962–1966. 128. Kouchoukos NT, Dougenis D. Medical progress: surgery of the
104. Soler R, Pombo F, Bargiela A, et al. MRI of pseudocoarctation of thoracic aorta. N Engl J Med. 1997;336:1876–1888.
the aorta: morphological and cine-MRI findings. Comput Med 129. Clouse WD, Hallett JW, Schaff HV, et al. Improved prognosis of
Imaging Graph. 1995;19:431–434. thoracic aortic aneurysms: a population-based study. JAMA.
105. Gay WA, Young WG. Pseudocoarctation of the aorta, a reap- 1998;280:1926–1929.
praisal. J Thorac Cardiovasc Surg. 1969;58:739–745. 130. Elefteriades JA. Natural history of thoracic aortic aneurysms:
106. Choi BW, Choe KO, Kim YJ. Magnetic resonance angiography of indications for surgery, and surgical versus nonsurgical risks.
pseudocoarctation. Heart. 2004:90:1213. Ann Thorac Surg. 2002;74:S1877–1880.
107. Nielson JC, Powell AJ, Gauvreau K, et al. Magnetic resonance 131. Cayne NS, Veith FJ Lipsitz EC, et al. Variability of maximal aor-
imaging predictors of coarctation severity. Circulation. 2005;111: tic aneurysm diameter measurements on CT scan: significance
622–628. and methods to minimize. J Vasc Surg. 2004;39:811–815.
108. Von Schulthess GK, Higashino SM, Higgins CB, et al. Coarctation 132. Davies RR, Gallo A, Coady MA, et al. Novel measurement of rel-
of the aorta: MR imaging. Radiology. 1986;158:469–474. ative aortic size predicts rupture of thoracic aortic aneurysms.
109. Riehle TJ, Oshinski JN, Brummer ME, et al. Velocity- Ann Thorac Surg. 2006;81:169–177.
encoded magnetic resonance image assessment of regional 133. Hirose Y, Hamada S, Takamiya M, et al. Aortic aneurysms:
aortic flow in coarctation patients. Ann Thorac Surg. 2006; growth rates measured with CT. Radiology. 1992;185:
81:1002–1007. 249–252.
110. Simpson IA, Chung KJ, Glass RF, et al. Cine magnetic resonance 134. Gertz SD, Cherukuri P, Bodmann BG, et al. Usefulness of multi-
imaging for evaluation of anatomy and flow relations in infants detector computed tomography for noninvasive evaluation of
and children with coarctation of the aorta. Circulation. 1988;78: coronary arteries in asymptomatic patients. Am J Cardiol. 2006;
142–148. 97:287–293
111. Rees S, Somerville J, Ward C, et al. Coarctation of the aorta: MR 135. Kawaharada N, Morishita K, Hyodah H, et al. Magnetic reso-
imaging in late postoperative assessment. Radiology. 1989;173: nance angiographic localization of the artery of Adamkiewicz
499–502. for spinal cord blood supply. Ann Thorac Surg. 2004;78:
112. Didier D, Saint-Martin C, Lapierre C, et al. Coarctation of the 846–851.
aorta: pre and postoperative evaluation with MRI and MR 136. Hyodah H, Kawaharada H, Akiba M, et al. Usefulness of
angiography; correlation with echocardiography and surgery. Int preoperative detection of artery of Adamkiewicz with dynamic
J Cardiovasc Imaging. 2005;3:1–19. contrast-enhanced MR angiography. Radiology 2005;236:
113. Steffens JC, Courne MW, Sakuma H, et al. Quantification of 1004–1009
collateral blood flow in coarctation of the aorta by velocity- 137. Takase K, Sawamura Y, Igarashi, et al. Demonstration of the
encoded cine magnetic resonance imaging. Circulation. artery of Adamkiewicz at multi-detector row helical CT.
1994;90:937–943. Radiology. 2002;223:39–45.
114. Lembcke A, Dohmen P, Rodenwaldt J, et al. Images in cardio- 138. Crawford ES, Crawford JL, Safi HJ, et al. Thoracoabdominal
vascular medicine. Recoarctation of the aorta associated with aortic aneurysms: preoperative and intraoperative factors deter-
ascending aortic aneurysm demonstrated by ECG gated multi- mining immediate and long-term results of operations in 605
slice CT. Circulation. 2003;107:e80–81. patients. J Vasc Surg. 1986;3:389–404.
115. Soulen RI, Kan J, Mitchell S, White RI. Evaluation of balloon 139. Safi HJ, Miller CC 3rd, Estrera AL, et al. Staged repair of exten-
angioplasty of coarctation restenosis by MR imaging. Am sive aortic aneurysms: long-term experience with the elephant
J Cardiol. 1987;60:343–345. trunk technique. Ann Surg. 2004;240:677–684.
116. Bank ER, Aisen AM, Rocchini AP, Hernandez RJ. Coarctation 140. Godwin JD, Herfkens RL, Skioldebrand CG, et al. Evaluation of
of the aorta in children undergoing angioplasty: pretreatment dissections and aneurysms of the thoracic aorta by conventional
and posttreatment MR imaging. Radiology. 1987;162: and dynamic scanning. Radiology. 1980;136:125–163.
235–240. 141. White RD, Lipton MJ, Higgins CB. Noninvasive evaluation of
117. Hager A, Kaemmerer H, Leppert A, et al. Follow-up of adults thoracic aortic disease by contrast-enhanced computed tomog-
with coarctation of the aorta: comparison of helical CT and raphy. Am J Cardiol. 1986;57:282–290.
MRI, and impact on assessing diameter changes. Chest. 2004; 142. Gonda RL, Gutierrez OH, Azodo MV. Mycotic aneurysms of the
126:1169–1176. aorta: radiologic features. Radiology. 1988;168:343–346.
118. Rajasinghe HA, Reddy VM, Van Son JA, et al. Coarctation repair 143. Kurich VA, Vogelzang RL, Hartz RS, et al. Ruptured thoracic
using end-to-side anastomosis of descending aorta to proximal aneurysm: unusual manifestation and early diagnosis using CT.
aortic arch. Ann Thorac Surg. 1996;61:840–844. Radiology. 1986;160:87–89.
119. Stocker JT. Sequestrations of the lung. Semin Diagn Pathol. 144. Landtman M, Kivisari L, Bondestom S, et al. Diagnostic value
1986;3:106–121. of ultrasound, computed tomography, and angiography in
120. Cho SY, Kim HC, Bae SH, et al. Demonstration of blood supply ruptured aortic aneurysms. Eur J Radiol. 1984;4:248–253.
to pulmonary sequestration by MR and CT angiography. 145. Taneja K, Gulati M, Jain M, et al. Ductus arteriosus aneurysm
J Comput Assist Tomogr. 1996;20:993–995. in the adult: role of computed tomography in diagnosis. Clin
121. Ahmed M, Jacobi V, Vogl TJ. Multislice CT and CT angiography Radiol. 1997;52:231–234.
for non-invasive evaluation of bronchopulmonary sequestra- 146. Torres WE, Maurer DE, Steinberg HV, et al. CT of aortic
tion. Eur Radiol. 2004;14:2141–2143. aneurysms: the distinction between mural and thrombus calcifi-
122. Deguchi E, Furukawa T, Ono S, et al. Intralobar pulmonary cation. AJR Am J Roentgenol. 1988;150:1317–1319.
sequestration diagnosed by MR angiography. Pediatr Surg Int. 147. Heiberg E, Wolverson MK, Sundaram M, Shields JB. CT charac-
2005;21:576–577. teristics of atherosclerotic aneurysm versus aortic dissection.
123. Pressler V, McNamara JJ. Aneurysm of the thoracic aorta: review J Comput Assist Tomogr. 1985;9:78–83.
of 260 cases. J Thorac Cardiovasc Surg. 1985;89:50–54. 148. Moore EH, Farmer DW, Geller SC, et al. Computed tomography
124. Davies RR, Goldstein LJ, Coady MA, et al. Yearly rupture or in the diagnosis of iatrogenic false aneurysms of the ascending
dissection rates for thoracic aortic aneurysms; simple prediction aorta. AJR Am J Roentgenol. 1984;142:1117–1118.
based on size. Ann Thorac Surg. 2002;73:17–28. 149. Thorsen MK, Goodman LR, Sagel SS, et al. Ascending aorta
125. Isselbacher EM. Thoracic and abdominal aortic aneurysms. complications after cardiac surgery. J Comput Assist Tomogr.
Circulation. 2005;111:816–828. 1986;10:219–225.
150. Quint LE, Francis IR, Williams DM, et al. Evaluation of thoracic 175. Macura KJ, Corl FM, Fishman EK, Bluemke DA. Pathogenesis in
aortic disease with the use of helical CT and multiplanar recon- acute aortic syndromes: aortic dissection, intramural hema-
structions: comparison with surgical findings. Radiology. 1996; toma, and penetrating atherosclerotic aortic ulcer. AJR Am
201:37–41. J Roentgenol. 2003;181:309–316.
151. Kimura F, Shen Y, Date S, et al. Thoracic aortic aneurysm and 176. Hirst AE Jr, Johns VR Jr, Klime SW Sr. Dissecting aneurysm of
aortic dissection: new endoscopic mode for three-dimensional the aorta: a review of 505 cases. Medicine (Baltimore). 1958;
CT display of aorta. Radiology. 1996;198:573–578. 37:217–279.
152. Rubin GD, Paik DS, Johnston PC, Napel S. Measurement of 177. Debakey ME, Henley WS, Cooley DA, et al. Surgical manage-
the aorta and its branches with helical CT. Radiology. 1998; ment of dissecting aneurysms of the aorta. J Thorac Cardiovasc
206:823–829. Surg. 1965;49:130–149.
153. Rubin GD. CT angiography of the thoracic aorta. Semin 178. Dailey PO, Trueblood HW, Stiinson EB, et al. Management
Roentgenol. 2003;38:115–1134. of acute aortic dissections. Ann Thorac Surg. 1970;10:
154. Roos JE, Hellinger JC, Hallet R, et al. Detection of endograft 237–247.
fractures with multidetector row computed tomography. J Vasc 179. Williams DM, Lee DY, Hamilton BH, et al. The dissected aorta:
Surg. 2005;42:1002–1006. 3. Anatomy and radiologic diagnosis of branch-vessel compro-
155. Hassoun HT, Mitchell RS, Makaroun MS, et al. Aortic neck mor- mise. Radiology. 1997;203:37–44.
phology after endovascular repair of descending thoracic aortic 180. Slonim SM, Nyman UR, Semba CP, et al. Aortic dissection:
aneurysms. J Vasc Surg. 2006;43:26–31. percutaneous management of ischemic complications with
156. Kusagawa H, Shimono T, Ishida, et al. Changes in false lumen endovascular stents and balloon fenestration. J Vasc Surg.
after transluminal stent-graft placement in aortic dissections: six 1996;23:241–253.
years experience. Circulation. 2005;111:2951–2957. 181. Nienabar CA, Fattori R, Lund G, et al. Nonsurgical reconstruc-
157. Bhalla S, West OC. Spiral CT angiography of the aorta. Semin tion of thoracic aortic dissection by stent-graft placement.
Ultrasound CT MRI. 2005;26:281–304. N Engl J Med. 1999;340:1539–1545.
158. Chiles C, Carr JJ. Vascular disease of the thorax: evaluation 182. Czermak BV, Waldenberger P, Fraedrich G, et al. Treatment of
with multidetector CT. Radiol Clin North Am. 2005;43: Stanford type B aortic dissection with stent-grafts: preliminary
543–569. results. Radiology. 2000;217:544–550.
159. Bae KT, McDermott R, Gierada DS, et al. Gadolinium-enhanced 183. Eggebrecht H, Nienaber CA, Neuhauser M, et al. Endovascular
computed tomography angiography in multi-detector row stent-graft placement in aortic dissection? a meta-analysis. Eur
computed tomography: initial observations. Acad Radiol. 2004; Heart J. 2006;27:489–498.
11:61–68. 184. Dake MD, Wang DS. Will stent-graft repair emerge as treatment
160. Rosioreanu A, Alberico RA, Litwin A, et al. Gadolinium- of choice for acute type B dissection? Semin Vasc Surg. 2006;
enhanced computed tomographic angiography: current status. 19:40–47.
Curr Probl Diagn Radiol. 2005;34:207–219. 185. Upchurch GR, Nienaber C, Fattori, et al. Acute aortic dissection
161. Takase K, Sawamura Y, Igarashi K, et al. Demonstration of the presenting with primarily abdominal pain: a rare manifestation
artery of Adamkiewicz at multi-detector row helical CT. of a deadly disease. Ann Vasc Surg. 2005;19:367–373.
Radiology. 2002;223:39–45. 186. Nallamouthu BK, Mehta RH, Saint S, et al. Syncope in acute
162. Glazer HS, Gutierrez FR, Levitt RG, et al. The thoracic aorta aortic dissection: diagnostic, prognostic, and clinical implica-
studied by MR imaging. Radiology. 1985;157:149–155. tions. Am J Med. 2002;113:468–471.
163. Holland AE, Barentz JO, Skotnicki S, et al. Preoperative MRA 187. Demos TC, Posniak HC, Churchill RJ. Detection of the intimal
assessment of the coronary arteries in an ascending aortic flap of aortic dissection on unenhanced CT images. AJR Am J
aneurysm. J Magn Reson Imaging. 2000;11:324–326. Roentgenol. 1986;146:601–603.
164. Schuijf JD, Bax JJ, Shaw LJ, et al. Meta-analysis of comparative 188. Mukherjee D, Evangelista A, Nienaber CA, et al. Implications of
diagnostic performance of magnetic resonance imaging and periaortic hematoma in patients with acute aortic dissection
multislice computed tomography for noninvasive coronary (from the International Registry of Acute Aortic Dissection).
angiography. Am Heart J. 2006;151:404–411. Am J Cardiol. 2005;96:1734–1738.
165. Vilacosta I, Roman JA. Acute aortic syndrome. Heart. 2001; 189. Vernhet H, Serfaty JM, Serhal M, et al. Abdominal CT angiogra-
85:365–368. phy before surgery as a predictor of postoperative death in acute
166. Wooley CF, Sparks EH, Boudalas H. Aortic pain. Prog Cardiovasc aortic dissection. AJR Am J Roentgenol. 2004;182: 875–879.
Dis. 1998;40:563–589. 190. Goldberg SP, Sanders C, Nanda NC, Holman WL. Aortic dissec-
167. Tsai TT, Nienaber CA, Eagle KA. Acute aortic syndromes. tion with intimal intussusception: diagnosis and management.
Circulation. 2005;112:3802–3813. J Cardiovasc Surg. 2000;41:613–615.
168. Macura KJ, Corl FM, Fishman EK, Bluemke DA. Pathogenesis in 191. Zhan, MF, Zhou QZ, Xiao HM, Xi G. Acute aortic dissection
acute aortic syndromes: aortic aneurysm leak and rupture and with intimal intussusception: MRI appearances. AJR Am J
traumatic aortic transection. AJR Am J Roentgenl. 2003;181: Roentgenol. 2006;186:841–843.
303–307. 192. Le Page MA, Quint LE, Sonnad SS, et al. Aortic dissection: CT
169. Mehard WB, Heiken JP Sicard GA. High-attenuating crescent in features that distinguish true lumen from false lumen. AJR Am J
abdominal aortic aneurysm wall at CT: a sign of acute or Roentgenol. 2001;177:207–211.
impending rupture. Radiology. 1994; 192:359–362. 193. Williams DM, Joshi A, Dake MD, et al. Aortic cobwebs: an
170. Siegel CL, Cohan RH, Korobkin M, et al. Abdominal aortic anatomic marker identifying the false lumen in aortic dissection—
aneurysm morphology: CT features in patients with ruptured imaging and pathologic correlation. Radiology. 1994;190:
and non-ruptured aneurysms. AJR Am J Roentgenol. 1994;163: 167–174.
1123–1129. 194. Kapoor V, Ferris JV, Fuhrman CR. Intimomedial rupture: A new
171. Arita T, Matsunaga N, Takano K, et al. Abdominal aortic CT finding to distinguish true from false lumen in aortic dissec-
aneurysm: rupture associated with the high-attenuating crescent tion. AJR Am J Roentgenol. 2004;183:109–112.
sign. Radiology. 1997;204:765–768. 195. Kouchoukos NT, Marshall WG Jr. Eleven-year experience with
172. Halliday KE, al-Kutoubi A. Draped aorta: CT sign of contained composite graft replacement of the ascending aorta and aortic
leak of aortic aneurysms. Radiology. 1996;199:41–43. valve. J Thorac Cardiovasc Surg. 1986;92:691–705.
173. Hagan PG, Nienaber CA, Isselbacher EM, et al. The 196. Bachet JE, Termignon JL, Dreyfus G, et al. Aortic dissection—
International Registry of Acute Aortic Dissection (IRAD): new prevalence, cause, and results of late reoperations. J Thorac
insights into an old disease. JAMA. 2000;283:897–903. Cardiovasc Surg. 1994;108:199–206.
174. Ergin MA, Griepp RB. Dissections of the aorta. In: Baue E, ed. 197. Crawford ES, Kirklin JW, Naftel DC, et al. Surgery for acute dis-
Glenn’s thoracic and cardiovascular surgery. New York: Appleton & section of ascending aorta: should the arch be included?
Lange; 1996:2273–2298. J Thorac Cardiovasc Surg. 1992;104:46–59.
198. Hirotani T, Nakamichi T, Munakata M, Takeuchi S. Routine 222. Maraj R, Rerkpattanapipat P, Jacobs LE, et al. Meta-analysis of
extended graft replacement for an acute type A aortic dissection 143 reported cases of aortic intramural hematoma. Am J Cardiol.
and the patency of the residual false channel. Ann Thorac Surg. 2000;86:664–668.
2003;76:1957–1961. 223. Sueyoshi E, Matsuoka Y, Sakamoto I, et al. Fate of intramural
199. Nienaber CA, von Kodolitsch Y, Nicholas V, et al. The diagnosis hematoma of the aorta: CT evaluation. J Comput Assist Tomogr.
of thoracic aortic dissection by noninvasive imaging procedures. 1997;21:931–938.
N Engl J Med. 1993;328:1–9. 224. Wilson SK, Hutchins GM. Aortic dissecting aneurysms: causative
200. Thorsen MK, San Dretto MA, Lawson TL. Dissecting aortic factors in 204 subjects. Arch Pathol Lab Med. 1982;106:175–180.
aneurysms: accuracy of computed tomographic diagnosis. 225. Patrick TO, Roman WD. Acute aortic dissection and its variants:
Radiology. 1983;148:773–777. toward a common diagnostic and therapeutic approach. Cir-
201. Stanford W, Rooholamini SA, Galvin JR. Ultrafast computed culation. 1995;92:1376–1378.
tomography in the diagnosis of aortic aneurysms and dissec- 226. Evangelista A, Mukherjee D, Mehta RH, et al. Acute intramural
tions. J Thorac Imaging. 1990;5:32–39. hematoma of the aorta: a mystery in evolution. Circulation.
202. Yoshida S, Akiba H, Tamakawa M, et al. Thoracic involvement 2005 1;111:1063–1070.
of type A aortic dissection and intramural hematoma: diagnos- 227. Evangelista A, Dominguez R, Sebastia C, et al. Long-term follow-
tic accuracy—comparison of emergency helical CT and surgical up of aortic intramural hematoma: predictors of outcome.
findings. Radiology. 2003;228:430–435. Circulation. 2003;108:583–589.
203. Inoue T, Watanabe S, Masuda Y, et al. Evaluation of blood flow 228. Sueyoshi E, Matsuoka Y, Imada T, et al. New development of
patterns of true and false lumens in dissecting aneurysms using an ulcerlike projection in aortic intramural hematoma: CT
MR phase-contrast techniques. Clin Imag. 1996;20:262–268. evaluation. Radiology. 2002;224:536–541.
204. Suzuki T, Mehta RH, Ince H, et al. Clinical profiles and out- 229. Dake MD. Aortic intramural haematoma: current therapeutic
comes of acute type B aortic dissection in the current era: les- strategy. Heart. 2004;90:375–378.
sons from the International Registry of Aortic Dissection 230. von Kodolitsch Y, Csosz SK, Koschyk DH, et al. Intramural
(IRAD). Circulation. 2003;108(suppl 1):II 312–317. hematoma of the aorta: predictors of progression to dissection
205. Mehta RH, Bossone E, Evangelista A, et al. Acute type B aortic dis- and rupture. Circulation. 2003;107:1158–1163.
section in elderly patients: clinical features, outcomes, and simple 231. Kaji S, Akasaka T, Horibata Y, et al. Long-term prognosis of
risk stratification rule. Ann Thorac Surg. 2004;77:1622–1628. patients with type A aortic intramural hematoma. Circulation.
206. Trimarchi S, Nienaber CA, Rampoldi V, et al. Contemporary 2002;106(12 suppl 1):I248–52.
results of surgery in acute type A aortic dissection: the 232. Kaji S, Akasaka T, Katayama M, et al. Long-term prognosis of
International Registry of Acute Aortic Dissection experience. patients with type B aortic intramural hematoma. Circulation.
J Thorac Cardiovasc Surg. 2005;129:112–122. 2003;108(suppl 1):II307–311.
207. Harris JA, Bis KG, Glover JL, et al. Penetrating atherosclerotic 233. Uchida K, Imoto K, Takahashi M, et al. Pathologic characteris-
ulcers of the aorta. J Vasc Surg. 1994;19:90–99. tics and surgical indications of superacute type A intramural
208. Kazerooni EA, Bree RI, Williams DM. Penetrating athersclerotic ul- hematomas. Ann Thorac Surg. 2005;79:1518–1521.
cers of the descending thoracic aorta: evaluation with CT and 234. Wu MT, Wu TH, Lee D. Images in cardiovascular medicine.
distinction from aortic dissection. Radiology. 1992;183: 759–765. Multislice computed tomography of aortic intramural
209. Stanson AW, Kazmier FJ, Hollier LH, et al. Penetrating athero- hematoma with progressive intercostal artery tears: the Chinese
sclerotic ulcers of the thoracic aorta: natural history and clinico- ring-sword sign. Circulation. 2005;111(5):e92–93.
pathologic correlations. Ann Vasc Surg. 1986;1:15–23. 235. Greendyke RM. Traumatic rupture of the aorta: special reference
210. Welch TJ, Stanson AW, Sheedy PF II, et al. Radiologic evaluation to automobile accidents. JAMA. 1966;195:527–530.
of penetrating aortic atherosclerotic ulcer. Radiographics. 1990; 236. Williams JS, Graff JA, Uku JM, Steinig JP. Aortic injury in vehicu-
10:675–685. lar trauma. Ann Thorac Surg. 1994;57:726–730.
211. Yucel EK, Steinberg FL, Efflin TK, et al. Penetrating aortic ulcers: 237. Pretre R, Chilcott M. Blunt trauma to the heart and great vessels.
diagnosis with MR imaging. Radiology. 1990;177:779–781. N Engl J Med. 1997;336:626–632.
212. Cho KR, Stanson AW, Potter DD, et al. Penetrating atherosclerotic 238. Parmley LF, Mattingly TW, Manion WC, Jahnke EJJ. Non-
ulcer of the descending thoracic aorta and arch. J Thorac Cardiovasc penetrating traumatic injury of the aorta. Circulation. 1958;17:
Surg. 2004;127:1393–1399; discussion 1399–1401. 1086–1101.
213. Castaner E, Andreu M, Gallardo X, et al. CT in nontraumatic 239. Groskin SA. Selected topics in chest trauma. Semin Ultrasound
acute thoracic aortic disease: typical and atypical features and CT MRI. 1996;17:119–141.
complications. Radiographics. 2003;23:S93–110. 240. Sevitt S. The mechanisms of traumatic rupture of the thoracic
214. Absi TS, Sundt TM 3rd, Camillo C, et al. Penetrating atheroscle- aorta. BR J Surg. 1977;64:166–173.
rotic ulcers of the descending thoracic aorta may be managed 241. Creasy JD, Chiles C, Routh WD, Dyer RB. Overview of traumatic
expectantly. Vascular. 2004;12:307–311. injury of the thoracic aorta. Radiographics. 1997;17:27–45.
215. Batt M, Haudebourg P, Planchard PF, et al. Penetrating athero- 242. Pretre R, LaHarpe R, Cheretakis K, et al. Blunt injury to the
sclerotic ulcers of the infrarenal aorta: life-threatening lesions. ascending aorta: three patterns of presentation. Surgery. 1996;
Eur J Vasc Endovasc Surg. 2005;29:35–42. 119:603–610.
216. Takahashi K, Stanford W. Multidetector CT of the thoracic aorta. 243. Woodring JH. The normal mediastinum in blunt traumatic rup-
Int J Cardiovasc Imaging. 2005;21:141–153. ture of the thoracic aorta and brachiocephalic arteries. J Emerg
217. Quint LE, Williams DM, Francis IR, et al. Ulcerlike lesions of Med. 1990;8:467–476.
the aorta: imaging features and natural history. Radiology. 2001; 244. Kram HB, Wohlmuth DA, Appel PL, Shoemaker WC. Clinical
218:719–723. and radiographic indications for aortography in blunt chest
218. Demers P, Miller DC, Mitchell RS, et al. Stent-graft repair of trauma. J Vasc Surg. 1987;6:168–176.
penetrating atherosclerotic ulcers in the descending thoracic 245. Huang P, Fong C, Rademaker A. Prediction of traumatic aortic
aorta: mid-term results. Ann Thorac Surg. 2004;77:81–86. rupture from plain chest film findings using stepwise logistic
219. Eggebrecht H, Herold U, Schmerermund A, et al. Endovascular regression. Ann Emerg Med. 1987;16:1330–1333.
stent-graft treatment of penetrating aortic ulcer: results over a 246. Mirvis SE, Bidwell JK, Buddemeyer EU, et al. Value of chest radi-
median follow-up of 27 months. Am Heart J. 2006;151: 530–536. ography in excluding traumatic aortic rupture. Radiology. 1987;
220. Murray JG, Manisali M, Flamm SD, et al. Intramural hematoma 163:487–493.
of the thoracic aorta: MR image findings and their prognostic 247. Mirvis SE, Bidwell JK, Buddemeyer EU, et al. Imaging diagnosis
implications. Radiology. 1997;204:349–355. of traumatic aortic rupture: a review and experience at a major
221. Nienaber CA, von Kodolitsch Y, Peterson B, et al. Intramural trauma center. Invest Radiol 1987;22:187–196.
hemorrhage of the thoracic aorta: diagnostic and therapeutic 248. Raptopoulos V. Traumatic aortic tear: screening with chest CT.
implications. Circulation. 1995;92:1465–1472. Radiology. 1992;182:667–673.
249. Fisher RG, Chasen MH, Lamki N. Diagnosis of injuries of the occlusion of the thoracic aorta by draining cerebrospinal fluid.
aorta and brachiocephalic arteries caused by blunt chest trauma: J Cardiovasc Surg. 1960;1:188–197.
CT vs aortography. AJR Am J Roentgenol. 1994;162: 1047–1052. 273. Blaisdell F, Cooley D. The mechanism of paraplegia after tem-
250. Mirvis SE, Shanmuganathan K, Miller BH, et al. Traumatic aor- porary thoracic aortic occlusion and its relationship to spinal
tic injury: diagnosis with contrast-enhanced thoracic CT—five- fluid pressure. Surgery. 1962;51:351–355.
year experience at a major trauma center. Radiology. 1996;200: 274. Safi HJ, Bartoli S, Hess KR, et al. Neurologic deficit in patients at
413–422. high risk with thoracoabdominal aortic aneurysms: the role of
251. Wong YC, Wang LJ, Lim KE. Periaortic hematoma on helical CT cerebrospinal fluid drainage and distal aortic perfusion. J Vasc
of the chest: A criterion for predicting blunt traumatic aortic Surg. 1994;20:434–443.
rupture. AJR Am J Roentgenol. 1998;170:1523–1525. 275. Cina CS, Abouzahr L, Arena GO, et al. Cerebrospinal fluid
252. Wong H, Gotway MB, Sasson AD, Jeffrey RB. Periaortic drainage to prevent paraplegia during thoracic and thoracoab-
hematoma at diaphragmatic crura at helical CT: sign of blunt dominal aortic aneurysm surgery: a systematic review and meta-
aortic injury in patients with mediastinal hematoma. Radiology. analysis. J Vasc Surg. 2004;40:36–44.
2004;231:185–189. 276. Kotoh K, Fukahara K, Yamashita A, et al. Ischemic changes in
253. Curry JD, Recine CA, Snavely E, et al. Periaortic hematoma on evoked spinal cord potentials during profound hypothermic
abdominal computed tomographic scanning as an indicator circulatory arrest in thoracic aortic surgery. Surg Today. 2005;
of thoracic aortic rupture in blunt trauma. J Trauma. 2002;52: 35:271–274.
699–702. 277. Weigang E, Hartert M, von Samson P, et al. Improved spinal
254. Gavant ML, Menke PG, Fabian T, et al. Blunt traumatic aortic cord perfusion during thoracoabdominal aortic repair. Thorac
rupture: detection with helical CT of the chest. Radiology. 1995; Cardiovasc Surg. 2005;53:69–73.
197:125–133. 278. Svensson LG, Crawford ES, Hess KR, et al. Experience with 1509
255. Gavant ML, Flick P, Menke P, Gold RE. CT aortography of patients undergoing thoracoabdominal aortic operations. J Vasc
thoracic aortic rupture. AJR Am J Roentgenol. 1996;166:955–961. Surg. 1993;17:357–370.
256. Scaglioni M, Pinto A, Pinto F, et al. Role of contrast-enhanced 279. Verdant A, Cossette R, Page A, et al. Aneurysms of the descend-
CT in the evaluation of acute thoracic aortic injuries after blunt ing thoracic aorta: three hundred and sixty-six consecu-
chest trauma. Eur Radiol. 2001;11:2444–2448. tive cases without paraplegia. J Am Coll Cardiol. 1995;23:
257. Sayeed R, Drain AJ, Wells FC. Magnetic resonance angiogram 1085–1090.
3-D reconstruction in acute aortic transection. Emerg Med J. 280. Estera AL, Miller CC 3rd, Chen EP, et al. Descending thoracic
2006;23(2):e14. aortic aneurysm repair: 12-year experience using distal aortic
258. Neuhauser B, Czermak B, Jaschke W, et al. Stent-graft repair for perfusion and cerebrospinal fluid drainage. Ann Thorac Surg.
acute traumatic aortic rupture. Am Surg. 2004;70:1039–1044. 2005;80:1290–1296; discussion 1296.
259. Amabile P, Collart F, Gariboldi V, et al. Surgical versus endovas- 281. Davison JK, Cambria RP, Vierra DJ, et al. Epidural cooling for
cular treatment of traumatic thoracic aortic rupture. J Vasc Surg. regional spinal cord hypothermia during thoracoabdominal
2004;40:873–879. aneurysmal repair. J Vasc Surg. 1994;20:304–310.
260. Rousseau H, Dambrin C, Marcheix B, et al. Acute traumatic 282. DellaCorte A, Scardone M, Romano G, et al. Aortic arch surgery:
aortic rupture: a comparison of surgical and stent-graft repair. thoracoabdominal perfusion during antegrade cerebral perfu-
J Thorac Cardiovasc Surg. 2005;129:1050–1055. sion may reduce postoperative morbidity. Ann Thorac Surg.
261. Nzewi O, Slight RD, Zamvar V. Management of blunt thoracic 2006;81:1358–1364.
aortic injury. Eur J Vasc Endovasc Surg. 2006;31:18–27. 283. Safi HJ, Miller III CC, Carr C. Importance of intercostal artery
262. Nienaber CA, Eagle KA. Aortic dissection: new frontiers in diag- reattachment during thoracoabdominal aortic aneurysm repair.
nosis and management. Part II: therapeutic management and J Vasc Surg. 1998;27:58–68.
follow-up. Circulation. 2003;108:772–778. 284. Cowan JA Jr, Dimick JB, Henke PK, et al. Surgical treatment of
263. Zehr KJ, Orszulak TA, Mullany CJ. Surgery for aneurysms of intact thoracoabdominal aortic aneurysms in the United States:
the aortic root: a 30-year experience. Circulation. 2004;110: hospital and surgeon volume-related outcomes. J Vasc Surg.
1364–1371. 2003;37:1169–1174.
264. Svensson LG, Crawford ES, Hess KR, et al. Deep hypothermia 285. Rofsky NM, Weinreb JC, Grossi EA, et al. Aortic aneurysm and
with circulatory arrest: determinations of stroke and early mor- dissection: normal MR imaging and CT findings after surgical
tality in 656 patients. J Thorac Cardiovasc Surg. 1992;106:19–31. repair with the continuous-suture graft-inclusion technique.
265. Ergin MA, Galla JD, Lansmann SL, et al. Hypothermic circula- Radiology. 1993;186:195–201.
tory arrest in operations on the thoracic aorta: determinations 286. Cabrol C, Pavie A, Mesnildrey P, et al. Long-term results with
of operative mortality and neurologic outcome. J Thorac total replacement of the ascending aorta and reimplantation
Cardiovasc Surg. 1994;107:788–799. of the coronary arteries. J Thorac Cardiovasc Surg. 1986;91:
266. Moon MR, Sundt TM. Influence of retrograde cerebral perfusion 17–25.
during aortic arch procedures. Ann Thorac Surg. 2002;74: 287. Niederhauser U, Kunzli A, Genoni M, et al. Surgery for acute
426–431. type A aortic dissection: comparison of techniques. Eur
267. Hagl C, Ergin MA, Galla JD, et al. Neurologic outcome after J Cardiothorac Surg. 2000;18:307–312.
ascending aorta-aortic arch operations: effect of brain protec- 288. Quint LE, Francis IR, Williams DM, et al. Synthetic interposition
tion technique in high-risk patients. J Thorac Cardiovasc Surg. grafts of the thoracic aorta: postoperative appearance on serial
2001;121:1107–1121. CT studies. Radiology. 1999;1:317–324.
268. Crawford E, Crawford J, Safi H, et al. Thoracoabdominal aortic 289. David TE, Feindel CM. An aortic valve-sparing operation for
aneurysms: preoperative and intraoperative factors determining patients with aortic incompetence and aneurysm of the ascend-
immediate and long-term results of operations in 605 patients. ing aorta. J Thorac Cardiovasc Surg. 1992;103:617–622.
J Vasc Surg. 1986;3:389–404. 290. Lewis CT, Cooley DA, Murphy MC, et al. Surgical repair of aortic
269. Cosselli JS, Lemaire SA, de Figueiredo LP, Kirby RP. Paraplegia root aneurysms in 280 patients. Ann Thorac Surg. 1992;53: 38–45.
after thoracoabdominal aortic aneurysm repair: is dissection a 291. Svensson LG, Longoria J, Kimmel WA, Nadolny E. Management
risk factor? Ann Thorac Surg. 1997;63:28–36. of aortic valve disease during aortic surgery. Ann Thorac Surg.
270. Acher CW, Wynn MM, Hoch JR, Kranner PW. Cardiac function 2000;69:778–783; discussion 783–784.
is a risk factor for paralysis in thoracoabdominal aortic replace- 292. Gott VL, Greene PS, Aleho DE, et al. Replacement of the aortic
ment. J Vasc Surg. 1998;27:821–830. root in patients with Marfan’s syndrome. N Engl J Med.
271. Grabenwoger M, Ehrlich M, Simon P, et al. Thoracoabdominal 1999;340:1307–1313.
aneurysm repair: spinal cord protection using profound hypo- 293. Halstead JC, Spielvogel D, Meier DM, et al. Composite aortic
thermia and circulatory arrest. J Cardiac Surg. 1994;9:679–684. root replacement in acute type A dissection: time to rethink the
272. Miyamoto K, Ueno A, Wada T, Kimoto S. A new and simple indications? Eur J Cardiothorac Surg. 2005;27:626–632; discus-
method of preventing spinal cord damage following temporary sion 632–633.
294. Allen RC, Schneider J, Longenecker L, et al. Paraanostomotic 317. Macedo TA, Stanson AW, Oderich GS, et al. Infected aortic
aneurysms of the abdominal aorta. J Vasc Surg. 1993;18: aneurysms: imaging findings. Radiology. 2004; 231:250–257.
424–432. 318. Foote EA, Postier RG, Greenfield RA, Bronze MS. Infectious
295. Kouchoukos NT, Robicsek F, Griepp RB. Prospective study of the aortitis. Curr Treat Options Cardiovasc Med. 2005;7(2):89–97.
natural history of thoracic aortic aneurysms—discussion. Ann 319. Smathers RL, Buschi AJ, Pope TL, et al. The azygos arch: normal
Thorac Surg. 1997;63:1544–1545. and pathologic CT appearance. AJR Am J Roentgenol. 1982;139:
296. Kawamoto S, Bluemke DA, Traill TA, Zerhouni EA. Thora- 477–483.
coabdominal aorta in Marfan syndrome: MR imaging findings 320. Friedmand AC, Chambers E, Sprayregen S. The normal and
of progression of vasculopathy after surgical repair. Radiology. abnormal left superior intercostal vein. AJR Am J Roentgenol.
1997;203:727–732. 1978;131:599–602.
297. Spielvogel D, Halstead JC, Meier M, et al. Aortic arch replace- 321. Ball JB, Proto AV. The variable appearance of the left superior
ment using a trifurcated graft: simple, versatile, and safe. Ann intercostal vein. Radiology. 1982;144:445–452.
Thorac Surg. 2005;8:90–95. 322. Lane EJ, Heitzman ER, Dinn WM. The radiology of the superior
298. Safi HJ, Miller CC 3rd, Estrera AL, et al. Staged repair of exten- intercostal veins. Radiology. 1976;120:263–267.
sive aortic aneurysms: morbidity and mortality in the elephant 323. Engel IA, Auh YH, Rubenstein WA, et al. CT diagnosis of
trunk technique. Circulation. 2001;104:2938–2942. mediastinal and thoracic inlet venous obstruction. AJR Am J
299. Greenberg RK, Haddad F, Svensson L, et al. Hybrid approaches Roentgenol. 1983;141:521–526.
to thoracic aortic aneurysms: the role of endovascular elephant 324. Takasugi JE, Godwin JD. CT appearance of the retroaortic anas-
trunk completion. Circulation. 2005;112:2619–2626. tomoses of the azygos system. AJR Am J Roentgenol. 1990;154:
300. Yamaguchi T, Naito H, Ohta M, et al. False lumen in type III 41–44.
aortic dissections: progress CT study. Radiology. 1985;156: 325. Landay MJ. Azygos vein abutting the posterior wall of the right
757–760. main and upper lobe bronchi: a normal CT variant. AJR Am
301. Yamaguchi T, Guthaner DF, Wexler L. Natural history of the false J Roentgenol. 1983;140:461–462.
channel of type A aortic dissection after surgical repair. CT 326. Rockoff SD, Druy EM. Tortuous azygos arch simulating a
study. Radiology. 1989;170:743–747. pulmonary lesion. AJR Am J Roentgenol. 1982;138:577–579.
302. Mathieu D, Keta K, Loisance D. Post-operative CT follow-up of 327. Hatfield MK, Vyborny CJ, MacMahon H, Chessare JW. Case
aortic dissection. J Comput Assist Tomogr. 1986;10:216–218. report. Congenital absence of the azygos vein: a cause for “aortic
303. Immer FF, Krahenbuhl E, Hagen U, et al. Large area of the false nipple” enlargement. AJR Am J Roentgenol. 1987;149:273–274.
lumen favors secondary dilatation of the aorta after acute type 328. Carter MM, Tarr RW, Mazer MJ, Carroll FE. The “aortic nipple”
A aortic dissection. Circulation. 2005;112(9 suppl):I249–252. as a sign of impending superior vena caval syndrome. Chest.
304. Gaubert JY, Moulin G, Mesena T, et al. Type A dissection of the 1985;87:775–777.
thoracic aorta: use of MR imaging for long term follow up. 329. Kim H-J, Kim HS, Chung SH. CT diagnosis of superior vena cava
Radiology. 1995;196:363–369. syndrome: importance of collateral vessels. AJR Am J Roentgenol.
305. Bachet JE, Termignon JL, Dreyfus G, et al. Aortic dissection. 1993;161:539–542.
Prevalence, cause, and results of late reoperations. J Thorac 330. Haramati LB, Moche, IE, Rivera, VT, et al. Computed tomogra-
Cardiovasc Surg. 1994;108:199–205; discussion 205–206. phy of partial anomalous pulmonary venous connection in
306. Loubeyre P, Delignette A, Bonefoy L, et al. Magnetic resonance adults. JCAT. 2003; 27:743–749,
imaging evaluation of the ascending aorta after graft-inclusion 331. Tello R, Scholz E, Finn JP, Costello P. Subclavian vein thrombosis
surgery: comparison between an ultrafast contrast-enhanced detected with spiral CT and three-dimensional reconstruction.
MR sequence and conventional cine-MRI. J Magn Reson Imaging. AJR Am J Roentgenol. 1993;160:33–34.
1996;6:478–483. 332. Eren S, Karaman A, Okur A. The superior vena cava syndrome
307. Mitchell RS, Ishimaru S, Criate FJ, et al. Third International caused by malignant disease. Imaging with multidetector row
Summit on Thoracic Aortic Endografting: lessons from long- CT. Eur J Radiol. 2006
term results of thoracic stent-graft repairs. J Endovasc Ther. 333. Ishikawa T, Clark RA, Tokuda M, Ashida H. Focal contrast
2005;12(1):89–97. enhancement on hepatic CT in superior vena caval and brachio-
308. Sunder-Plassmann L, Orend KH. Stent-grafting of the cephalic vein obstruction. AJR Am J Roentgenol. 1983;140:
thoracic aorta—complications. J Cardiovasc Surg (Torino). 2005; 337–338.
46(2):121–130. 334. Maldjian PD, Obolevich AT, Cho KC. Focal enhancement of the
309. Sharma S, Sharma S, Taneja K, et al. Morphologic mural liver on CT: a sign of SVC obstruction. J Comput Assist Tomogr.
changes in the aorta revealed by CT in patients with nonspecific 1995;19:316–318.
aortoarteritis (Takayasu’s arteritis). AJR Am J Roentgenol. 335. Bashist B, Parisi A, Frager DH, Suster B. Abdominal CT findings
1996;167:1321–1325. when the superior vena cava, brachiocephalic vein, or sub-
310. Tanikawa K, Eguchi K, Kitamura Y, et al. Magnetic resonance clavian vein is obstructed. AJR Am J Roentgenol. 1996;167:
imaging detection of aortic and pulmonary artery wall thickening 1457–1463.
in the acute stage of Takayasu arteritis: improvement of clinical 336. Muramatsu T, Miyamae T, Mashimo M, et al. Hot spots on liver
and radiologic findings after steroid therapy. Arthritis Rheum. scans associated with superior or inferior vena caval obstruc-
1992;35:476–480. tion. Clin Nucl Med. 1994;19:622–629.
311. Gotway MB, Araoz PA, Macedo TA, et al. Imaging findings 337. Sonin AH, Mazer MJ, Powers TA. Obstruction of the inferior vena
in Takayasu’s arteritis. AJR Am J Roentgenol. 2005;184:1945– cava: a multiple-modality demonstration of causes, manifesta-
1950. tions, and collateral pathways. Radiographics. 1992; 12:309.
312. Mitnick H, Jacobowitz G, Krinsky G, et al. Periaortitis: gadolin- 338. Miyake H, Suzuki K, Ueda S, et al. Localized fat collection adja-
ium-enhanced magnetic resonance imaging and response to cent to the inferior vena cava: a normal variant on CT. AJR Am J
therapy in four patients. Ann Vasc Surg. 2004;18:100–107. Roentgenol. 1992;158:423–425.
313. Kobayashi Y, Ishii K, Oda K, et al. Aortic wall inflammation due 339. Han BK, Im JG, Jung JW, et al. Pericaval fat collection that mim-
to Takayasu arteritis imaged with 18F-FDG PET coregistered ics thrombosis of the inferior vena cava: Demonstration with
with enhanced CT. J Nucl Med. 2005;46:917–922. use of multi-directional reformation CT. Radiology. 1997;203:
314. Yamada I, Numano F, Suzuki S. Takayasu arteritis: evaluation 105–108.
with MR imaging. Radiology. 1993;188:89–94. 340. Lebowitz JA, Rofsky NM, Krinsky GA, Weinreb JC. Gadolinium-
315. Hayashi K, Fukushima T, Matsunaga N, Hombo Z. Takayasu’s enhanced body MR venography with subtraction technique. AJR
arteritis: decrease in aortic wall thickening following steroid Am J Roentgenol. 1997;169:755–758.
therapy, documented by CT. Br J Radiol. 1986;59:281–283. 341. Shinde TS, Lee VS, Rofsky NM, et al. Three-dimensional
316. Evans JM, Ofallon WM, Hunder GG. Increased incidence of aor- gadolinium-enhanced MR venographic evaluation of patency of
tic aneurysm and dissection in giant cell (temporal) arteritis—a central veins in the thorax: initial experience. Radiology. 1999
population-based study. Ann Intern Med. 1995;122:502–507. 213:555–560.
342. Kroencke TJ, Taupitz M, Arnold R, et al. Three-dimensional 345. Finn J, Zisk J, Edelman R, et al. Central venous occlusion: MR
gadolinium-enhanced magnetic resonance venography in angiography. Radiology. 1993;187:245–251.
suspected thrombo-occlusive disease of the central great veins. 346. Pedrosa I, Morrin M, Oleaga L, et al. Is true FISP imaging
Chest. 2001;120:1570–1576. reliable in the evaluation of venous thrombosis? AJR Am J
343. Li WL, David V, Kaplan R, Edelman RR. Three-dimensional low Roentgenol. 2005;185:1632–1640.
dose gadolinium-enhanced peripheral MR venography. J Magn 347. Roubidoux MA, Dunnick NR, Sostman HD, Leder RA. Renal
Res Imag. 1998;8:630–633. carcinoma: detection of venous extension with gradient-echo
344. Hartnell GG, Hughes LA, Finn JP, Longmaid HE. Magnetic MR imaging. Radiology. 1992;182:269–272.
resonance angiography of the central chest veins—a new gold 348. Wall MJ Jr, Mattox KL, Debakey ME. Injuries to the azygous
standard? Chest. 1995;107:1053–1057. venous system. J Trauma. 2006;60:357–362.
5636_Naidich_ch03_pp217-288 12/7/06 11:34 AM Page 217
Pulmonary Arterial 3
Disease
NORMAL ANATOMY 217 artery diameters in both anatomically normal individuals

and patients with pulmonary hypertension and showed
ACUTE PULMONARY EMBOLISM 219 that in normal persons the main pulmonary artery aver-
Clinical Evaluation 219 aged 24.2 2.2 mm in diameter at a level near its bifurca-
Imaging Evaluation 220 tion. Based on these data, the authors concluded that
Imaging Appearances of Acute Pulmonary 28.6 mm [mean 2 standard deviations (SD)] should be
Embolism 237 considered the upper limit of normal for main pulmonary
artery diameter. These measurements are best made at a
CHRONIC THROMBOEMBOLIC PULMONARY right angle to the long axis of the main pulmonary artery,
HYPERTENSION 258 lateral to the ascending aorta and at the level of its bifurca-
Imaging of Chronic Thromboembolic Hypertension 259 tion (Fig. 3-1). The finding of a dilated main pulmonary
Preoperative Evaluation and Postoperative Changes 265 artery should raise the suspicion of pulmonary hyperten-
sion with a sensitivity of 87% and a specificity of 89% (2).
MISCELLANEOUS ABNORMALITIES 268 The right and left pulmonary arteries should be of approxi-
Congenital Anomalies 268 mately equal size, although the left pulmonary artery
Pulmonary Artery Aneurysms 271 appears slightly larger in most persons. In the same study
Neoplastic Involvement of the Pulmonary Arteries 274 by Kuriyama et al. (1) mentioned previously, the proximal
right pulmonary artery measured 18.7 2.8 mm in diame-
ADDITIONAL CAUSES OF PULMONARY ARTERY ter in normal individuals, and the left pulmonary artery
OBSTRUCTION 277 measured 21.0 3.5 mm. Similar findings have been
reported by Ackman-Haimovici et al. (3): In their study of
patients undergoing organ transplantation, the left pul-
NORMAL ANATOMY monary artery averaged 21 5 mm in those with normal
pulmonary artery pressure.
The main pulmonary artery arises at the base of the right The right main pulmonary artery usually divides into an
ventricle and extends superiorly for a distance of approxi- ascending trunk (truncus anterior) and a descending trunk
mately 5 cm before dividing into the right and left (interlobar branch) posterior to the superior vena cava
pulmonary arteries. The main, right, and left pulmonary (SVC) and anterior to the right main bronchus (Fig. 3-1).
arteries are intrapericardial, with the right pulmonary The truncus anterior originates from within the peri-
artery having a longer course and dividing into two lobar cardium and supplies the right upper lobe, primarily the
branches at the root of the lung (Fig. 3-1). apical and anterior segments. Although the right interlobar
On computed tomography (CT), the main pulmonary branch largely supplies the middle and right lower lobes, in
artery is recognizable as the most anterior vascular structure about 90% of persons, a branch of the interlobar artery
arising from the heart and is characteristically retrosternal supplies the posterior segment of the right upper lobe (4).
at its point of origin. Ranges in the size of the pulmonary The interlobar branch of the right pulmonary artery largely
arteries in both normal individuals and in patients with lies within the major fissure, thus accounting for its name.
pulmonary hypertension have been long established. The left pulmonary artery, after passing over the
Kuriyama et al. (1), for example, measured pulmonary left main bronchus, usually continues as the descending
A B
Figure 3-1 Normal pulmonary artery anatomy: visualization techniques.

A: Routine contrast-enhanced axial image through the main right and left pul-
monary arteries. B: Coronal multiplanar reconstruction (MPR). C: Slightly
oblique maximum intensity projection (MIP) image. D: Parasagittal MIP.
E: Volume-rendered coronal image. Although standard axial images are gener-
ally sufficient to make most diagnoses, in select cases the ability to reconstruct
data using a variety of methods may be preferential, as illustrated throughout
E this chapter.
Chapter 3: Pulmonary Arterial Disease 219
or interlobar left pulmonary artery, which gives rise to hood of diagnosing PE has been significantly improved by
segmental branches of the left upper and lower lobes the widespread availability of CT pulmonary angiography
(Fig. 3-1). On occasion, the left pulmonary artery gives rise (CTPA), initially with single-detector and now, more com-
to a short ascending branch that divides into segmental monly, multidetector CT (MDCT), it remains probable that
branches supplying the upper lobe. a large number of cases still go undetected prior to death
Pulmonary arteries divide by dichotomous branching: (9,10). This is of particular importance, as, compared with
The two branches may be of approximately equal size, or the overall case fatality rate for pulmonary embolism, the
one branch may be significantly larger than the other. The mortality of untreated pulmonary embolic disease approxi-
pulmonary artery has approximately 17 divisions from its mates 30%, similar to the mortality in those patients with
bifurcation, with peripheral branches having a diameter of massive pulmonary embolism (8–14).
0.1 to 1.5 mm (5). Pulmonary arteries are divided into two
broad categories: those with diameters greater than 0.5 mm
Clinical Evaluation
(typically including arteries at the subsegmental level),
referred to as “elastic” arteries serving as a reservoir for Only approximately one half of patients with throm-
blood ejected by the right ventricle, and those smaller than boembolism present with one of the cardinal signs of
0.5 mm in diameter, referred to as “muscular” arteries, PE—namely chest pain, dyspnea, or hemoptysis (11).
which accompany peripheral airways to the level of the Additionally, although laboratory evidence of hypoxia, a
terminal bronchioles. widened alveolar–arterial gradient, hypocapnia, or elec-
The branching patterns of lobar, segmental, and subseg- trocardiographic (ECG) abnormalities are insufficiently
mental pulmonary artery branches show considerable vari- specific to establish the diagnosis, neither do normal val-
ation. Although it is typical for the lobar, segmental, and ues preclude the diagnosis (15–18). Although the Wells
subsegmental bronchi to be paired with a pulmonary criteria (Table 3-1) (19,20) combining clinical and labo-
artery branch, the origins of these branches are variable, ratory tests into low, medium, or high clinical probability
and supernumerary or accessory artery branches supplying for establishing PE have been shown to correlate with ven-
a lobe or segment are often present. Variation is more tilation/perfusion (V/Q) scans, these remain suboptimal
common in the upper lobes than in the lower lobes. To for diagnosis in individual cases. In view of the high mor-
identify a specific lobar or segmental artery branch, it is tality of both late or missed diagnosis, it is imperative that
usually necessary to identify its associated bronchus. a combination of high clinical suspicion, particularly in
When less than 0.15 mm in size, pulmonary arteries patients predisposed to thromboembolism, and timely
are properly labeled arterioles (6). In fact, pulmonary arte- use of appropriate testing remain necessary to exclude the
rioles as small as 300 mm in diameter can be seen on diagnosis (21).
high-resolution CT (7); these vessels roughly correspond
to acinar arteries, being the 16th generation of branches at
D-Dimer Evaluation
the level of the terminal and most proximal respiratory
bronchioles. However, these vessels are visible only as soft The D-dimer test evaluates the presence of fibrinolytic
tissue attenuation structures on lung window scans, and degradation products produced at the site of venous throm-
their opacification is not visible after contrast administra- bosis (22). The traditional enzyme-linked immunosorbent
tion. An increase or a decrease in the size of small periph-
eral pulmonary arteries is sometimes seen in patients with
pulmonary vascular disease associated with increased or
decreased blood flow.
TABLE 3-1
CLINICAL PREDICTION RULES: WELLS CRITERIA
ACUTE PULMONARY EMBOLISM Criterion Points
The diagnosis of few entities has been as markedly affected Prior pulmonary embolus (PE) 1.5
or deep venous thrombosis (DVT)
by advances in imaging technology as pulmonary throm-
boembolic disease. Over the past decade CT has evolved as Tachycardia 1.5
the imaging modality of choice in most institutions for the Recent surgery 1.5
initial evaluation of patients with suspected pulmonary Clinical signs of DVT 3
emboli (PE). An estimated 500,000 cases of acute pul-
Alternative diagnoses to PE less likely 3
monary embolism occur in the United States annually,
associated with an overall mortality of 5% to 10%. Autopsy Hemoptysis 1
data suggest that although the antemortem detection rate of Cancer 1
PE has been steadily increasing in every decade since the
Clinical probability is assessed as low, 0–1 (5%–13% probability of PE);
1950s, as late as the 1990s the rate of antemortem detection intermediate, 2–6 (38%–40% probability of PE); or high, 7 (67%–91%
was still less than 50% (8). Nonetheless, although the likeli- probability of PE).
assay (ELISA) and new quantitative tests, including both of negative CTPAs performed, with estimates ranging
rapid ELISA and turbidimetric latex examinations, have between 36% and 60% (33,34). For example, in a study of
proved more accurate and sensitive for the detection of 247 patients with D-dimer values less than 1,000 ng/mL,
thromboembolic disease than prior qualitative and semi- Abcarian et al. (33) demonstrated that all had negative
quantitative tests such as earlier rapid ELISA, latex, and CTPAs and normal 3-month follow-up. Similarly, Dunn
whole blood agglutination techniques and should prefer- et al. (35) reported a negative predictive value of 99.6% for
entially be obtained when available. negative D-dimer evaluation alone in more than 1,000
Although thresholds for quantitative D-dimer assays vary emergency department patients presenting with a suspi-
among published reports, as documented in a meta-analysis cion of PE.
of 41 studies involving more than 6,000 patients, a thresh- More recently, Kelly and Hunt (36) reviewed several
old of greater than 500 ng/mL has gained widespread prospective clinical outcome studies in a predominantly
acceptance with a demonstrated sensitivity of 90% to 95% outpatient population and demonstrated that a combina-
(22a). Unfortunately, low specificities, averaging between tion of low-intermediate clinical suspicion and a negative
40% and 45% in the same analysis, reflect the fact that D-dimer test safely obviated further testing in 47% to 51%
D-dimer levels may be elevated in numerous other diseases of patients, with rates of subsequently documented throm-
(Table 3-2). Although a positive D-dimer test in the absence boembolic disease at 3 months occurring in only 0.2% to
of clinical suspicion should not be a cause for further inves- 0.4% of cases.
tigation to exclude pulmonary embolism, not surprisingly Based on these data, in our judgment there is now suf-
in clinical practice routine performance of D-dimer testing ficiently strong evidence to support the utility of routine
has been documented to actually increase the number of D-dimer testing to reduce the number of unnecessary
studies performed to exclude PE (23–25). CTPA studies with their associated radiation exposure
In our experience, the most appropriate clinical use of (37). However, appropriate imaging studies to exclude
D-dimer testing is to exclude pulmonary thromboembolic pulmonary embolism should still be obtained when there
disease given the high negative predictive values reported, is high clinical suspicion of disease, particularly in elderly
in most studies ranging between 92% and 100% (26,27). or inpatient populations.
However, it is important to recognize that there are occa-
sional causes of a false-negative D-dimer test (Table 3-2). It
Imaging Evaluation
is also worth emphasizing that the utility of D-dimer test-
ing has largely been established in outpatients and that the
Choice of Modality
use of this test to exclude thromboembolic disease in inpa-
tients is not as well validated (28,29). Chest radiographs may demonstrate several classic signs
Use of D-dimer testing is especially problematic during such as regional oligemia (Westermark sign), peripheral
pregnancy as there is normally a progressive increase in basilar pleural-based wedge-shaped areas of opacity
D-dimer levels. In fact, in a series of normal pregnant (Hampton hump), or enlargement of the central pul-
patients, 75% still had normal D-dimer levels by 14 weeks monary arteries (Fleischner sign). However, these features
and 50% had normal levels at 19 weeks (30–32). D-dimer are neither sensitive nor specific for the diagnosis of PE
evaluation, therefore, may still prove useful in excluding (38,39). The use of chest radiographs, however, remains
embolism during the early stages of pregnancy. important in the determination of other alternative diag-
Several retrospective studies have suggested that nega- noses such as pneumothorax, lobar collapse, congestive
tive D-dimer tests may significantly reduce the number cardiac failure, or rib fractures that may cause similar
symptoms on presentation.
Currently, CTPA, especially with the widespread avail-
ability of MDCT scanners, has evolved in most institutions
TABLE 3-2 into the definitive diagnostic modality for the evaluation of
both acute and chronic PE. MDCT is noninvasive, easily
THROMBOEMBOLISM EXCLUSION BY D-DIMER:
scheduled, rapidly performed, and highly accurate. Its major
FALSE- AND TRUE-POSITIVES
advantage compared with V/Q imaging, the most often used
False-Positives False-Negatives alternative imaging modality, is that unlike V/Q scans that
provide only indirect evidence of pulmonary embolism,
Trauma/surgery Too early (first day)
MDCT allows direct visualization of intra-arterial clot.
Malignancy Too late: 48 hr inpatient Compared with pulmonary angiography, MDCT is less inva-
Pregnancy Elderly sive and provides a global assessment of the entire thorax
Cardiac/renal/hepatic failure Anticoagulant therapy (40–42). In a study of 510 patients reported by van Strijen
et al. (41), although CT demonstrated PE in 24% of
Inflammation with elevated CRP Small thrombus/embolus size
patients, it also proved of value by establishing alternative
CRP, C-reactive protein. diagnoses in an additional 26% (Fig. 3-2).
Several comparative studies have demonstrated the diag-

nostic superiority of CT over V/Q imaging (47–50). CT
directly identifies pulmonary arterial filling defects, whereas
V/Q imaging provides only indirect evidence of clot without
precise anatomic correlation. In one study of 68 patients, of
185 perfusion defects identified on perfusion imaging, only
16% were found to positively correlate with angiographic
findings (51). Although only 10 of 35 segmental or larger
mismatches proved to be due to angiographically identified
emboli, 43 segmental mismatches proved to have no angio-
graphic correlate at all.
The advantage of CT versus V/Q imaging was established
early, even in the era of single-detector CT (SDCT). In one
representative comparison, SDCT demonstrated improved
sensitivity, specificity, negative predictive value (NPV) and
positive predictive value (PPV) (with all test measurements
ranging between 94% and 97%) when compared with
V/Q imaging (with corresponding measurements ranging
between 74% and 82%). Additionally, the interobserver
Figure 3-2 Tumor infiltration of the pulmonary arteries. CT
pulmonary angiogram (CTPA) in a patient presenting with acute agreement of CTPA (k 0.72) was significantly higher than
dyspnea and suspected pulmonary embolism demonstrates the that of V/Q imaging (k 0.22) (48). In a similar study van
presence of extensive tumor infiltration into the left hilum and
Rossum et al. (47) evaluated 123 patients undergoing both
mediastinum, encasing and severely narrowing the left main
pulmonary artery due to documented small cell carcinoma of the V/Q and CTPA imaging, separately evaluating these tests in
lung. Note the presence of bilateral pleural effusions, as well. association with both clinical and chest radiographic find-
ings. Compared with V/Q imaging, CT resulted in fewer in-
conclusive studies (8% vs. 28%), higher sensitivity (75% to
49%), higher specificity (90% to 74%), and a near doubling
of established alternative diagnoses (93% to 51%). As
Computed Tomography of Pulmonary
important, of 35 indeterminate V/Q studies, CT definitively
Emboli
identified 5 normal patients and 13 with PE, and provided
alternative diagnoses in an additional 11 patients.
Computed Tomography Pulmonary Angiography
Versus Ventilation/Perfusion Imaging and
CT Versus Pulmonary Angiography. Conventional pul-
Angiography
monary angiography has long been considered the “gold
CT Versus V/Q Imaging. V/Q perfusion imaging is con- standard” for the evaluation of PE (52,53). However, in the
sidered a sensitive but nonspecific examination for the era of easily acquired MDCT exams, this is no longer consid-
evaluation of PE. In the original Prospective Investigation ered axiomatic. In comparison to SDCT in an animal model
on Pulmonary Embolism Diagnosis (PIOPED) I study of artificially injected subsegmental PE, no significant differ-
(43), an abnormal study was associated with a 98% sensi- ence in sensitivity was identified (both 87%) between CT
tivity for the detection of pulmonary embolism; however, and angiography (54). As noted previously, in comparison
as only 14% had a normal or near normal study, specificity angiography CT has the advantage of being noninvasive
proved exceedingly low (10%). Indeed, in this study the and, more importantly, in a large percentage of cases, pro-
identification either of a high probability (87% incidence viding alternative diagnoses. Despite these advantages, it is
of PE) or normal/near normal studies (4% incidence of worth emphasizing that, although invasive, the significant
PE) occurred in only 27% of cases. Only 13% of patients complication rate of pulmonary angiography is often over-
with PE had a high probability study. In a total of 73% of estimated by clinicians, in most studies on the order of only
cases, studies were interpreted as either of low or interme- 1% to 2% (44,55). As a consequence, pulmonary angiogra-
diate probability (14% and 30% PE risk, respectively) and phy is underutilized in most institutions. In fact, loss of
therefore deemed clinically inconclusive, requiring further operator experience will likely contribute to the already well
diagnostic evaluation (44). Although the introduction of established poor interobserver agreement for angiography.
revised PIOPED criteria (45) has reduced the number of Although agreement among observers is acceptable for
intermediate examinations by approximately 15% (46), emboli within large central vessels (81% to 90%), there is
the number of residual nondiagnostic studies coupled poor two-reader agreement at the subsegmental level (45%
with only a 70% to 75% interobserver agreement for to 66%), further diminishing to 16% when three readers’
intermediate probability studies remains a significant limi- interpretations are compared (43,56,57). Not surprisingly,
tation of this technique (43). the traditional role of pulmonary angiography as a “gold
standard” is evolving, now typically acquired only in select that pulmonary angiography still be performed in equivo-
cases in which all other imaging studies remain equivocal, cal or indeterminate cases.
usually in the setting of a high clinical probability of pul-
monary embolism. In this context, selective angiography
Diagnostic Accuracy of Multidetector Computed
in particular should be considered to clarify indeterminate
Tomography Pulmonary Angiography
CT findings involving single vessels, with the advantage of
Compared with SDCT, MDCT provides markedly
lessening interobserver variability while likely reducing the
improved temporal and spatial resolution, resulting in
complication rate of the procedure.
more consistent demonstration of subsegmental vascular
anatomy. This is due both to the ability to scan a larger
Diagnostic Accuracy of Single-Detector Computed volume of the lungs with thinner collimation and to im-
Tomography Pulmonary Angiography proved resolution of individual sections. As documented
Several meta-analyses have been conducted on the per- in two separate studies, Remy-Jardin et al. demonstrated
formance of SDCT for the detection of PE. Many of these that even with SDCT, a reduction of collimation from 3 to
are limited by their incorporation of both early and newer 2 mm increased the percentage of subsegmental arteries
single-detector hardware and the utilization of differing adequately visualized from 37% to 43% to 61% to 65%,
imaging protocols and gold standards for establishing the respectively (64,65) (Fig. 3-3). Raptopoulos and Boiselle
diagnosis of pulmonary embolism. Nevertheless, these (66) demonstrated that the increased speed of MDCT
studies demonstrate that, even allowing for these differ- using 2.5-mm sections compared with SDCT using 3-mm
ences, the sensitivity and specificity for detecting main, sections resulted in improved visualization of subsegmen-
lobar, or segmental PE range between 80% and 100% tal arteries. In a comparison of SD CTPA using 3-mm
(58–60). In distinction, the sensitivity of SD CTPA for the sections and MD CTPA using both 2.5-mm and 1.25-mm
detection of subsegmental PE, although variable, is consis- sections, Patel et al. (61) demonstrated significant
tently low (25% to 75%) (59,61). In a meta-analysis improvement in reader sensitivity for the evaluation of
of published data from 1986 to 1999, Rathbun et al. (60), subsegmental PE (between 37% and 39%, 53% and 56%,
incorporating both subsegmental data as well as data from and 71% and 76%, respectively). The same authors also
earlier studies performed with suboptimal 5-mm sections, documented improved interobserver agreement using
documented sensitivities and specificities ranging between thinner reconstructions (3 mm vs. 1.25 mm) at both
53% and 100% and 81% and 100%, respectively (60). the segmental (k 0.79–0.80 vs. 0.47–0.75) and subseg-
Based on these data, both the American College of Chest mental (k 0.71–0.76 vs. 0.28–0.54) levels. In a similar
Physicians and the American Thoracic Society concluded study comparing different reconstruction thicknesses
that SD CTPA was an insufficient clinical test to exclude in multidetector PE, Schoepf et al. (67) demonstrated
pulmonary embolism, particularly when there was high a 40% increase in sensitivity for subsegmental PE detec-
clinical suspicion (62,63), leading to the recommendation tion as the slice thickness was reduced from 3 mm to
A, B C
Figure 3-3 Pulmonary emboli (PE) diagnosis: role of thin collimation. Two different patients with suspected PE in vessels oriented in the
axial plane reconstructed at 5 mm (A and D), 3 mm (B and E), and 0.8 mm thickness (C and F), respectively. Note that the use of thinner sec-
tions not only improves the detection of small PE (E and F) but also reduces the incidence of false-positive studies by eliminating partial vol-
ume averaging (A–C).
D, E F
1 mm with concomitant improvement of interobserver Isolated Subsegmental Pulmonary Emboli

agreement. The detection of small subsegmental PE will largely
To date, numerous studies have corroborated signifi- be immaterial in a patient treated for concomitantly
cantly improved detection of subsegmental emboli (75% detected central PE. Therefore, suboptimal detection of
to 95%) using MDCT, with overall sensitivities and speci- subsegmental PE by either SDCT or MDCT will be of
ficities ranging between 90% and 100%, respectively potential clinical significance only when embolic disease
(50,68,69). It is worth emphasizing, however, that most of is limited to vessels of this caliber. In fact, the incidence of
these involve only relatively small numbers of patients. Nor isolated subsegmental PE is unknown, variously reported
do they incorporate advanced image processing techniques, as occurring in between 6% and 36% of cases (43,70,71)
such as multiplanar reconstruction (MPR) or maximum (Fig. 3-5). It has been suggested that some small emboli
intensity projection (MIP) images, now easily reconstructed may be normally filtered physiologically by the lung
from isotropic volumetric datasets obtained with state- capillary bed to prevent systemic embolization (72)
of-the-art CT scanners (Fig. 3-4). (Fig. 3-6).
A B
Figure 3-4 Pulmonary emboli (PE) diagnosis: role of maximum intensity projection images (MIPs). A: Coronal MIP reconstruction from a
contrast-enhanced PE study performed on a 4-detector CT scanner demonstrates the appearance of normal vascular anatomy. B: Coronal
MIP reconstruction in a different patient than in A from a contrast-enhanced PE study on a 16-detector CT scanner demonstrates improved
MIP reconstruction quality with improved visualization of peripheral subsegmental pulmonary arteries (compare with A). Note that in this
case there is an embolus in the right main pulmonary artery associated with an infarct in the right lower lobe, the CT equivalent of
a Hampton hump.
documented, only 1 of which proved fatal (74). In another

study designed to address these issues using CT, 25 of 61
patients with isolated subsegmental PE received no further
treatment, with no long-term adverse effect. At present, the
decision to treat small subsegmental PE varies among clini-
cians, depending on the individual patient’s cardiovascular
reserve and associated comorbid conditions. Although still
controversial, the finding of an isolated subsegmental
pulmonary embolus on CT at the least should prompt
a thorough investigation to identify potential sources of
thrombus that, left untreated, might result in subsequent
emboli (68,69,73).
Negative Computed Tomography Pulmonary

Angiography Studies
Figure 3-5 Pulmonary emboli diagnosis: use of thin collimation. As important as determining the sensitivity of CT is evalu-
Two-millimeter axial reconstruction shows to good advantage a ating the clinical importance of a negative CTPA study.
small isolated subsegmental pulmonary embolus in the posterior
segment of the right lower lobe (arrow). Note that in this case Numerous studies to date have concluded that adequately
thrombus straddles an arterial bifurcation. Points at which arteries performed CTPA is sufficient for the exclusion of PE, with
divide are especially worth close scrutiny, as they are frequent negative predictive values of greater than 98%, comparable
sites in which emboli lodge.
to or better than a negative pulmonary angiogram or nor-
mal V/Q studies, respectively (75–78). This is so, despite
What has yet to be established is whether there is clinical the fact that adequate visualization of all subsegmental
benefit to treating patients in whom isolated small emboli pulmonary arteries is rarely accomplished even using state-
are identified, specifically to lessen the otherwise increased of-the-art multidetector scanners. It is worth emphasizing,
risk for hemodynamic compromise or progression to however, that these studies typically involved only a rela-
chronic thromboembolic disease (73). In this regard, it tively small number of cases, limiting interpretation of
should be noted that in the original PIOPED study, 20 these data.
patients with isolated subsegmental emboli diagnosed by That identification of isolated subsegmental emboli
the expert review panel had initially been viewed as with CT may not be of great clinical importance is further
normal: Of these patients, only 2 had subsequent PE supported by studies comparing the clinical outcome of
A B
Figure 3-6 Pulmonary emboli diagnosis: subsegmental pulmonary emboli. A: Enlargement of an axial section through the right lower
lobe shows a small isolated subsegmental embolus (arrow) associated with peripheral atelectasis. In this case, anticoagulation was withheld.
B: Enlargement of an axial image at the same level as shown in A, obtained 48 hours later, shows spontaneous resolution of the suspected
embolus. Apparent spontaneous resolution of this embolus may reflect normal filtration of small emboli by the lung.
patients evaluated with single versus multidetector CT scan- evaluation, CT sensitivity improved to 90%, with a speci-
ners. In one comparative study of SD CTPA versus MD ficity of 95%, a PPV of 85%, and a NPV of 97%, respec-
CTPA, Remy-Jardin et al. (79) evaluated two consecutive tively. Of 105 patients with a positive CTV, only 3%
groups of patients with suspected PE. Using a single- involved the inferior vena cava (IVC) or pelvic veins.
detector scanner, subsegmental vessels could be adequately Overall, CT had both positive and negative predictive
visualized in only 13% of 125 patients, compared with values of 96% when paired with concordantly high or low
57% of subsegmental vessels identified in a separate but clinical probability, decreasing to 92% with an intermedi-
clinically similar cohort of 134 patients evaluated by ate clinical assessment of PE.
MDCT. In both series, patients with a negative CTPA exami- Based on these data, the following conclusions were
nation were not treated but closely monitored with follow- reached: (a) that the PPV and NPV of CT are high when
up clinical evaluations in 3 months to exclude potentially there is clinical concordance regarding the probability of
recurrent PE. Surprisingly, the NPV for both SDCT and PE but that the false-negative rate of 17% for MD CTA
MDCT approached 100%, suggesting that although MDCT alone indicates the need for additional testing, especially
improves visualization of subsegmental arteries, allowing when there is a high clinical suspicion of PE, and (b) that
more isolated subsegmental PE to be identified, these are the addition of CTV is warranted, given a statistically sig-
unlikely to be clinically significant, especially in patients in nificant improvement in sensitivity when compared with
whom cardiopulmonary reserve is well maintained. MD CTA alone. Surprisingly, these conclusions are not dis-
similar to those obtained from reviews of studies using
single-detector scanners in recommending that, in cases
Prospective Investigation on Pulmonary with discordant CT and clinical findings, additional imag-
Embolism Diagnosis Study ing studies remain necessary.
To address issues relating to the relatively small sample A number of limitations should be noted regarding
size of prior CT studies, the Prospective Investigation on PIOPED II. In addition to the need to rely on composite
Pulmonary Embolism Diagnosis II (PIOPED II) study reference standards given ethical concerns regarding
evaluated the value of MDCT in diagnosing pulmonary requiring DSA in all cases, most studies were performed on
thromboembolic disease in a total of 825 patients with four-detector CT scanners without requiring the use of
suspected PE, using a variety of 4-, 8-, and 16-detector CT bolus tracking. Similarly, there was no requirement for
scanners supplemented by CT venography (CTV) (80). The routinely obtaining D-dimers. Perhaps most important, no
accuracy of CT was established using composite standards data were provided regarding the utility of CT for estab-
for both diagnosing and excluding pulmonary embolism. lishing alternative diagnoses. In this regard, most studies
A positive diagnosis required one of the following: a high were performed on outpatients, presumably with less ex-
probability V/Q lung scan in a patient with no prior his- tensive pulmonary and parenchymal diseases that might
tory of PE, a positive digital subtraction angiogram (DSA), otherwise have limited the utility of V/Q imaging.
or a positive venous ultrasound without a history of prior Although it is likely that these data and recommendations
deep venous thrombosis (DVT) at that site and a nondiag- will set the standard for the use of MDCT for diagnosing
nostic V/Q scan. In distinction, exclusion of PE required PE for the foreseeable future, it is also likely that these con-
one of the following: a negative DSA, a normal V/Q scan, clusions will prove controversial, particularly the signifi-
or a low or very low probability V/Q scan accompanied by cance of a well-performed negative CT study and the need
a clinical score using the Wells criteria 2 (Table 3-1) and to perform concomitant CTV (81).
a negative venous ultrasound.
In 51 (6%) of 825 cases, CT studies were interpreted by
a panel of two certified expert readers as inconclusive, Computed Tomography Technique
usually due to a combination of suboptimal contrast
It cannot be overstated that accurate evaluation of patients
enhancement and motion artifacts. Of the remainder, CT
with suspected pulmonary embolism requires meticulous
sensitivity was 83% (95%, CI 76 to 92), and specificity
scan technique. In addition to choice of scanner, factors
was 96% (95%, CI 93 to 97). The likelihood ratio for a
that affect visualization of the pulmonary arterial tree
positive study was 19.6 (95%, CI 13.3 to 29.0) and the
include, in particular, choice of scan protocol, method and
likelihood ratio for a negative study was 0.18 (95%, CI
rate of contrast administration, and image reconstruction
0.13 to 0.24). CT had a positive predictive value (PPV) of
algorithm (82,83). Proposed scan protocols for different
86% (95%, CI 79 to 90) and a negative predictive value
types of scanners are summarized in Table 3-3.
(NPV) of 95% (95%, CI 92 to 96). When assessed by the
size of involved pulmonary arteries, the PPV was 97% for
emboli in main and lobar arteries, 68% for segmental Scan Protocols
arteries, and 25% for subsegmental arteries, although in Key to optimizing scan quality is the acquisition of
the latter group only eight subsegmental emboli alto- the thinnest possible sections allowable during the short
gether were diagnosed. When CTV was added to the period of peak opacification of the pulmonary arterial
TABLE 3-3
COMPUTED TOMOGRAPHY PULMONARY EMBOLISM PROTOCOLS
4-Slice Adaptive 4-Slice Matrix 16-Slice
Array (Volume Array 16-Slice Sensation
Single Detector Zoom, Philips) (Lightspeed) Sensation Straton Tube 64-Slice
Detector 3.0 mm 4 1.0 4 2.5 16 0.75 16 0.75 64 0.6

configuration
Rotation time (s) 1.0 0.5 0.75 0.4 0.375 0.33
kVp/mAs 140/200 140/120 140/120 120/200 120/200 120/190
PACS recons 3.0 q 2.0 mm 2.5 q 2.5 mm 2.0 q 2.0 mm 2.0 q 2.0 mm 2.0 q 2.0 mm 2.0 q 2.0 mm
med med med med med med
5 q 5 mm 5 q 5 mm 5 q 5 mm 4 q 4 mm 4 q 4 mm 4 q 4 mm
lung lung lung lung lung lung
Recon algorithm Low frequency: Mediastinum. High frequency: Lung
FOV Smallest possible (maximum spatial resolution)
mm/rotation 4.5–6.0 4–6 15 14 14 19–25
IV contrast 150 mL 150 mL 150 mL 120 mL 120 mL 120 mL
volume/rate 3–5 mL/s 3–5 mL/s 3–5 mL/s 3–5 mL/s 3–5 mL/s 3–5 mL/s
Scan delay Timing run Bolus tracking Bolus tracking Bolus tracking Bolus tracking Bolus tracking
20 mL trigger 130 HU trigger 130 HU trigger 130 HU trigger 130 HU trigger 150 HU
Coverage Top of arch to Entire thorax Top of arch Entire thorax Entire thorax Entire thorax
right diaphragm, through bases,
then apices, then apices
then bases
PACS, picture archiving and communication system; recon, reconstruction; med, mediastinum; FOV, field of view; IV, intravenous; HU, Hounsfield units.
vasculature (64). With single-detector scanners it is neces- utility of this approach—in particular, the need to obtain
sary to reduce z-axis coverage to the central vasculature to images through the pelvis with resultant added radiation
permit thin collimation to be utilized during a reasonable exposure—remains uncertain (88). The finding of DVT in
breath-hold. A second acquisition, therefore, is typically a patient with PE is unlikely to alter management unless
necessary to complete the examination with additional there is an isolated subsegmental pulmonary embolus for
images obtained through the apices and lung bases. In this which anticoagulation might not otherwise be initiated. In
setting, it has been suggested that images be acquired in a distinction, in patients evaluated for potential pulmonary
caudocranial direction to minimize respiratory motion arti- embolism with negative CTPA examinations, the addi-
facts, which are generally greatest through the lower lobes. tional finding of asymptomatic lower extremity clot results
In distinction, with newer generation multidetector scan- in higher CT sensitivity, usually leading to alteration of
ners neither partitioning of data acquisition nor scanning in patient management, including, in select cases, insertion
a caudocranial direction is required, as the entire chest can of an IVC filter (89).
be evaluated using thin collimation during a single breath-
hold acquisition. Although the use of ECG gating is tech- Contrast Media Administration
nically feasible, this has not been shown to confer any Choice of iodinated contrast media concentrations in the
diagnostic advantage (84). range of 300 to 370 mg I/mL has little effect on diagnostic
CTPA can be combined with CT leg venography, as rec- image quality. It should be noted that in patients in whom
ommended by PIOPED II, especially in patients in whom iodinated contrast media is contraindicated, one alternative
ultrasound is impractical (85,86). Unfortunately, at present is administration of gadolinium contrast media at a dose of
there is no consensus regarding the optimal method for 0.4 mmol/kg (90). There is, however, some ongoing debate
performing CTV. One approach that has been advocated about the nephrotoxicity of such high gadolinium dosages
involves obtaining contiguous 5-mm sections, starting 2 to in patients with impaired renal function (90a).
3 minutes following contrast administration, beginning Both the volume and rate of administration of intra-
at the superior aspect of the calf and proceeding in a venous iodinated contrast media is increased when com-
caudocranial direction through the proximal IVC (87). The pared to routine thoracic CT examinations. Administration
by pressure injector at 3.5 to 5.0 mL/second necessitates on the patient’s age or heart rate is unreliable and likely to
the use of an 18 G or greater peripheral venous catheter yield a high number of suboptimal examinations (95,96).
and precludes the use of both indwelling peripherally A test injection of 20 mL of intravenous contrast may be
inserted central catheter (PICC)/Mediport catheters or used with dynamic acquisition of a single slice through
wider bore central venous access catheters. The former may the central pulmonary arteries every 3 to 5 seconds for
be damaged by rapid contrast administration, whereas the 20 to 25 seconds. An appropriate triggering delay may then
latter may induce an intracardiac catheter “whip” or fluid be determined by identifying the time point corresponding
“jet” effect, potentially damaging the cardiac endothelium to the slice with the densest contrast opacification and
or valves. Typically, 150 mL of intravenous contrast media adding 3 to 5 seconds to permit opacification of the more
is administered for studies performed with single-detector peripheral vasculature (82). This technique, although a sig-
or earlier four-slice scanners; however, this volume may be nificant improvement over a best guess approach, is still
reduced to 120 mL or less for newer generation MDCT limited principally by the fact that the 20-mL bolus may
scanners owing to their faster data acquisition. It is impor- not accurately predict the manner in which the larger bolus
tant to note that by suspending the injection as soon as the will behave and, in addition, that the available volume of
scanner has completed scanning the central vasculature, the intravenous contrast remaining for the examination is con-
volume of injected contrast may be significantly reduced. If sequently reduced.
a dual injector is available, a saline chaser may be utilized Bolus tracking, available on newer scanners, eliminates
to significantly reduce the volume of contrast required to the need for a test dose with its associated limitations and,
perform the examination (91–94). This is because intra- when available, should be used for all studies. During the
venous contrast administered later in the injection does not main contrast injection, a dynamic enhancement curve
contribute to vascular arterial enhancement but is required obtained from a preselected region of interest typically
merely to maintain both the forward motion of the previ- positioned over the main pulmonary artery is generated,
ously administered contrast and the compact nature of the and at a predetermined threshold the scan acquisition is
bolus, maximizing the enhancement peak. Clearance of commenced. Although there is currently no consensus
dense contrast media from the central veins may also regarding optimal choice of a triggering threshold, as this
reduce artifacts (91). will vary according to the speed of the scanner and the time
required for the gantry to move into position, in our experi-
Contrast Media Administration: Scan Delay ence a minimum of 130 to 150 Hounsfield units (HU)
Timing of the scan acquisition is the single most critical should be selected to ensure optimal contrast enhancement
factor in optimizing CTPA. Estimating a fixed delay based (Fig. 3-7).
A, B C
Figure 3-7 Pulmonary emboli diagnosis: bolus tracking. A–D: Coned-down views of axial images sequentially obtained at the same exact
level following a bolus of intravenous contrast media, with a region-of-interest cursor placed over the main pulmonary artery in each image.
With bolus tracking, automatic densitometry measurements are made every second, allowing precise triggering of scan acquisition to match
a preselected density threshold. E: A real-time graph demonstrating the level of contrast enhancement within the main pulmonary artery as
a function of time. In this case, a threshold of 130 HU was utilized to initiate the CT pulmonary angiography study. Note that in this case
monitoring images demonstrate an incidentally identified right main pulmonary artery embolus (continued).
D E
Dose Modulation of images either on stand-alone workstation or on a Picture-

In larger patients the use of dose reduction techniques Archiving and Communications System (PACS). The use of
during scanning may materially reduce the quality of the cine-viewing, in particular, is of inestimable value, in one
studies and hence should not be considered standard. In study increasing the number of patients identified as having
an attempt to reduce the radiation dose, many scanners PE by 25% (98). The ability to scroll through contiguous or
now employ a proportional dynamic reduction of tube overlapping images not only improves the tracking of
current output when the tube is directed along the pa- individual vessels and detection of intraluminal filling
tient’s thinner anteroposterior (AP) and posteroanterior defects but also permits the dynamic alteration of window
(PA) axis. Unfortunately, when the tube is operating at its levels and widths. Dense contrast may obscure even large
limits, as occurs in larger patients or typically in patients underlying PE, especially when standard mediastinal win-
undergoing CTPA, the tube current may be insufficient in dow settings are used. It cannot be overemphasized that in
the PA or AP axis, with the subsequent reduction in dose the interpretation of CTPA examinations, the pulmonary
resulting in noisy images. With the advent of advanced arteries should not appear “bone white,” as this will lead to
dose-limiting techniques, permitting additional real-time emboli being overlooked. The use of a higher centering
modulation of the mAs in the z-axis as well as in the point and wider window width (e.g., 200/1000) may assist
xy-plane during scan acquisition, tailoring the CT dose in visualizing emboli despite dense contrast enhancement.
index to acceptable image quality is essential for optimiz- For smaller vessels, or those with less dense contrast
ing scan technique. enhancement, these settings may require dynamic, interac-
tive alteration (Fig. 3-8). Brink et al. (99) have suggested
Image Reconstruction and Interpretation that, optimally, windowing should be centered at the
Overlapping 1- to 3-mm images should always be recon- mean density of the right or left pulmonary artery, with
structed using a “mediastinal” or “soft” algorithm to reduce a window width then set at twice the SD. In altering window
the edge-enhancing artifacts that may mimic emboli when width, readers should be aware that visualization of image
“sharp” or “bone” algorithms are utilized. As previously noise, the appearance of laminar contrast flow, and motion-
noted, the use of thinner collimation significantly improves related artifacts may increase, requiring mental adjustment
visualization of peripheral subsegmental arteries. Even with of the thresholds for determining when indeterminate low-
four-detector scanners, a reduction of collimation from 3 density foci truly represent emboli (82).
mm to 1.25 mm improves visualization of opacified sub- The American College of Radiology recommends that
segmental vessels from 37% to 39% to 71% to 76%, respec- official CTPA interpretation reports should include a
tively (61), as well as improving visualization of the next description of the quality of pulmonary arterial opacifica-
two orders of magnitude of vessel size from 47% to 74% tion to guide the level of interpretative confidence. In cases
and 16% to 35%, respectively (97). in which contrast opacification is suboptimal, it is still
It cannot be overemphasized that optimal evaluation of of value to provide a partial interpretation limited to those
CTPA examinations requires interactive dynamic evaluation arteries sufficiently well visualized to allow diagnosis. In
A B
C D
Figure 3-8 Pulmonary emboli diagnosis: effects of image windowing. A, B: Identical contrast-enhanced axial sections through the right
main pulmonary artery imaged with window widths of 400 HU (A) and 1,000 HU (B), respectively, with corresponding window widths of 40 HU
(A) and 150 HU (B). Note that in A, the pulmonary arteries appear “bone” white. It is only when appropriate window levels and widths are
applied that an otherwise unsuspected embolus becomes apparent (B). C, D: Identical coned-down sections through the inferior portion of
the right hilum, in a different patient than shown in A and B, again show to good advantage the use of appropriate windows and levels for
identifying emboli. In this case, note the presence of a filing defect in the anterobasilar segmental artery in D, otherwise obscured in C.
these situations, we believe it is more useful to describe at (100–103). A sliding stack of axial or coronal MIPs may
which specific level of anatomic vascular subdivision it is be especially useful as a screening tool to quickly review
possible to exclude embolic disease (main, lobar, segmen- large datasets resulting from use of thin collimation (104)
tal, subsegmental), rather than to state the absence of (Fig. 3-9). Although paddlewheel reconstructions also
“central” pulmonary embolism or of PE in vessels up to have been advocated by some authors as an additional
the nth order. This is because there is no universally mechanism to enhance visualization of vessels along their
accepted definition of “central” PE and, as important, con- entire length, this technique has not gained widespread
siderable variation in radiologists’ and clinicians’ assess- acceptance, likely because of its nonintuitive format
ment of the order of pulmonary arterial branching with (105–107) (Fig. 3-10). Despite the advantages that may
obvious potential for confusion or misinterpretation. result from advanced image processing techniques, it
should be emphasized that these do not replace the need
Advanced Image Processing to review axial source images (Fig. 3-11).
The use of MPRs may be of assistance by improving diag- Computer-aided diagnosis (CAD) represents another
nostic certainty in cases in which routine axial images may set of techniques currently being evaluated as a method for
be indeterminate, as well simplifying visualization of enhancing the detection of PE. Potentially useful com-
potentially complex three-dimensional (3D) structures puter-aided tools include, among others, interactive 3D
A B
Figure 3-9 Pulmonary emboli (PE) diagnosis: use of maximum intensity projections (MIPs). A: Axial 10-mm maximum MIP image demon-
strates multiple left-sided PE. B: A 7-mm off-axis coronal MIP in the same patient as shown in A, oriented along the length of the left main
pulmonary artery again demonstrates extensive central PE. In select cases, a stack of overlapping sliding MIP images may be used as an initial
means for screening for PE.
A B
C
Figure 3-10 Pulmonary emboli (PE) diagnosis: paddlewheel reconstructions. A–C: Three 1-mm multiplanar reconstruction images
selected from a set of images obtained using a 360-degree paddlewheel reconstruction algorithm. This approach, enabling images to be
viewed as a scrolled image dataset, purposefully elongates central vessels, making PE more apparent.
A B
C
Figure 3-11 Pulmonary emboli diagnosis: use of multiplanar reconstructions (MPRs). A: A 2-mm axial image demonstrates an easily over-
looked, completely occluded subsegmental pulmonary artery (arrow). B, C: Off-axis axial and coronal 5-mm maximum intensity projection
reconstructions, respectively, demonstrate to better advantage the presence of an obstructing subsegmental pulmonary embolus. In select
cases, use of MPRs may enhance diagnostic certainty by identifying otherwise overlooked emboli in cases with a high clinical index of
suspicion or by more definitively establishing emboli in patients with equivocal findings on routine axial images.
surface renderings of the pulmonary arteries with CAD systems for detecting emboli will not emerge until
automatic surface shading of vessels suspected of contain- techniques that allow reliable automated segmentation of
ing emboli and interactive vessel aligned techniques that the pulmonary arteries from the pulmonary veins have
permit real-time MPR images to be generated along the been developed (Figs. 3-12 and 3-13).
long-axis of vessels in which a filling defect is suspected.
Early generation automated second-reader detection sys-
tems are also under evaluation; these automatically iden- Pulmonary Emboli in Pregnancy
tify candidate emboli as low-density intra-arterial foci for Women are at increased risk of thromboembolism
observer review. Although these techniques hold consider- throughout the entire duration of their pregnancy as well
able promise for improving the detection especially of eas- as during the puerperium (108). Although there is no
ily overlooked subsegmental emboli, their general accep- absolute contraindication for performing CT during
tance remains to be validated. To date, the sensitivity and pregnancy, it is clearly advantageous to restrict unnecessary
specificity of currently available CAD algorithms have not radiation exposure, particularly during the first trimester,
been adequately tested. It is likely that the true potential of when fetuses are most susceptible to defective organogene-
Figure 3-12 Computer-assisted diagnosis (CAD). See Color

Figure 3-12. Image depicting an experimental method for provid-
ing automated three- dimensional surface shaded renderings of
the pulmonary vasculature. Using this technique, intraluminal areas
of low density suspicious for pulmonary emboli are automatically
highlighted for inspection. The output is designed to be freely
manipulated in three dimensions on a standalone workstation.
Renderings may be grayscale or color-coded while clicking on can-
didate emboli results in visualization of three orthogonal planes
through the vessel of interest. Although the use of various CAD
methods to diagnose pulmonary emboli is of interest, validation of
these techniques remains to be established (see Fig. 3-13 as well).
(Case courtesy of Carol Novak, PhD, Siemens Corporate Research,
Inc., Princeton, New Jersey.)
sis or death. In fact, the risks of complications occurring

secondary to fetal radiation exposure—including child-
hood malignancy and growth or mental retardation—
typically require doses at least 100 times higher than
those encountered during routine CTPA (0.1 to 0.5 mGy).
At present, termination of pregnancy should not be
considered to be justified below 100 mGy (109,110). Figure 3-13 Computer-assisted diagnosis (CAD). Axial (top)
and coronal images (bottom), as viewed using an experimental
Fetal radiation doses comparable to or less than those CAD device different from that illustrated in Figure 3-12. Using
resulting from a V/Q scan can be achieved with the addi- this approach, axial volumetric datasets are first reviewed by
tional benefit of a reduction in the number of otherwise readers, with subsequent CAD-identified regions of interest auto-
matically placed around suspected pulmonary emboli for reader’s
inconclusive studies. A survey of the clinical practice of evaluation. Simultaneous automatic generation of three orthogo-
members of the Society of Thoracic Radiology confirms nal multiplanar reconstructions further simplifies interpretation.
that 75% of respondents perform CTPA for suspected PE (Case courtesy of Marcos Salganikoff, PhD, Siemens Medical
Solutions, Malvern, Pennsylvania.)
during pregnancy, and of these just over half prefer CTPA
as the initial method of investigation. Eighty percent of
practitioners acquired a PA chest radiograph prior to the
examination, a practice we concur with, considering
of Pulmonary Emboli
the extremely low radiation dose resulting from a single
frontal chest radiograph. Also in keeping with our practice, The utility of magnetic resonance imaging (MRI) in the
approximately half of the respondents modified their evaluation of acute pulmonary embolism is in constant evo-
studies to reduce radiation exposure by reducing z-axis lution due to continual advances in scanning hardware and
coverage while increasing slice width and decreasing mAs imaging sequences. Although MRI for PE detection remains
(Table 3-4) (111). a predominantly research application performed almost
TABLE 3-4
PULMONARY EMBOLISM PROTOCOLS IN PREGNANCY
4-Slice Adaptive 4-Slice Matrix 16-Slice
Array (Volume Array 16-Slice Sensation
Single Detector Zoom, Philips) (Lightspeed) Sensation Straton Tube 64-Slice
Detector 3.0 mm 4 1.0 4 2.5 16 1.5 16 1.5 24 1.2

configuration
Rotation time (s) 1.0 0.5 0.75 0.4 0.375 0.33
kVp/mAs 120/80 120/80 120/80 120/80 120/80 120/80
PACS recons 3.0 q 2.0 mm 2.5 q 2.5 mm 2.0 q 2.0 mm 2.0 q 2.0 mm 2.0 q 2.0 mm 2.0 q 2.0 mm
med med med med med med
5 q 5 mm lung 5 q 5 mm lung 5 q 5 mm lung 4 q 4 mm lung 4 q 4 mm lung 4 q 4 mm lung
Recon algorithm Low frequency: Mediastinum. High frequency: Lung
FOV Smallest possible (maximum spatial resolution)
mm/rotation 4.5–6.0 4–6 15 14 14 29–37
IV contrast 150 mL 150 mL 150 mL 120 mL 120 mL 120 mL
volume/rate 3–5 mL/s 3–5 mL/s 3–5 mL/s 3–5 mL/s 3–5 mL/s 3–5 mL/s
Scan delay Timing run Bolus tracking Bolus tracking Bolus tracking Bolus tracking Bolus tracking
20 mL trigger 130 HU trigger 130 HU trigger 130 HU trigger 130 HU trigger 150 HU
Coverage Restricted: 1 cm above the aortic arch to just inferior to the xiphoid process; no leg venography
PACS, picture archiving and communication system; recon, reconstruction; med, mediastinum; FOV, field of view; IV, intravenous; HU, Hounsfield units.
exclusively in academic institutions, recent improvements increasing the radiofrequency (RF) pulse flip angle for
in image quality suggest that MRI will likely find increasing blood exiting the imaging volume slab, counteracting the
clinical utilization outside the small group of patients for saturation effects of flowing blood. In volunteers, this free
whom it is currently advantageous—patients in whom the breathing technique allows generation of MIP images to
administration of iodinated contrast media or ionizing radi- sixth-order vessels but has proved time consuming and has
ation is contraindicated or undesirable. In particular, as never been clinically validated (121,122).
breath-hold sequences become shorter, overall imaging Image quality and diagnostic accuracy have further
times are reduced, and medical monitoring devices become improved with the development of gadolinium-enhanced
more MR compliant, the MR suite has become a progres- 2D, time-of-flight MRA (123,124) and early breath-hold
sively less inhospitable environment for acutely dyspneic sequences (125–127) but still remained a predominantly
patients. Varying MR techniques offer the ability not only to research application until the introduction of current high-
identify PE but also to evaluate pulmonary parenchymal performance gradient systems (20 mT/m). These systems
perfusion and ventilation, moving toward a global physio- permit the use of short time to repetition (TR) 3D gradi-
logic assessment of the dyspneic patient. ent-echo sequences that allow the acquisition of high spa-
tial resolution volumetric datasets in a single breath-hold.
The reduction of respiratory motion permits visualization
Magnetic Resonance Imaging Techniques of peripheral vessels that would otherwise be obscured by
and Comparative Performance blurring (128–134). Signal averaging from multiple excita-
Initial attempts to employ MR for detecting PE employed tions of the 3D slab also minimizes cardiac motion, and
cine gradient-echo (112–115), two-dimensional (2D) short echo times (1 to 3 ms) result in reduction of signal
(114,116–118) and 3D time-of-flight magnetic resonance loss from flow-related dephasing and susceptibility arti-
angiography (MRA) (119,120). These provided only lim- facts at air-tissue interfaces (129). Additionally, the utiliza-
ited spatial resolution because of respiratory and cardiac tion of specialized surface phase array body coils improves
motion and hence were predominantly limited to the eval- signal-to-noise ratio compared with routine body coils,
uation of the central pulmonary arteries. Modifications of also contributing to improved visualization of the periph-
time-of-flight techniques such as variable-angle uniform eral pulmonary arteries (135,136) (Table 3-5).
signal excitation (VUSE) have improved visualization Monophasic 3D gradient-echo protocols require a 20- to
of the pulmonary arterial vasculature by progressively 30-second breath-hold, achieving higher resolution by
TABLE 3-5
MAGNETIC RESONANCE PULMONARY EMBOLISM PROTOCOL
Preparation ECG leads, right-sided IV, supplemental oxygen if poor breath-hold is suspected
Coils Four- to eight-channel phased-array coil over chest
Sequence Plane Comment
T1-weighted Axial Gated

double IR
Balanced SSFP Axial
3D Gradient-echo Coronal One measure (TR/TE 3.2/1.4 ms, flip 25 degrees, 12 s, FOV 340 mm, 512 384
pixels, 72 partitions, voxel size 0.7 1.2 1.5 mm). Use parallel imaging and
TRICKS/TREAT if available.
Timing run Axial Through main pulmonary artery. 1 mL of Gd at 3 mL/s followed by 20–40 mL
saline at 3 mL/s. Time to pulmonary artery.
3D Gradient-echo Coronal Three measures, no gap. 20–40 mL of Gd at 3 mL/s followed by 20 mL saline at
3 mL/s. Standard timing formula.
VIBE/LAVA Axial One measure
Post-processing MIP subtraction
ECG, electrocardiogram; IV, intravenous; IR, inversion recovery; SSFP, steady-state free precession; TR, time to repetition; TE, time to echo; FOV, field
of view; TRICKS, time-resolved imaging of contrast kinetics; TREAT, time-resolved echo-shared angiographic technique; Gd, gadolinium; VIBE, volu-
metric interpolated breath-hold examination; LAVA, liver acquisition with acceleration; MIP, maximum intensity projection.
utilizing more partitions during this period. The perform- demonstrated comparable sensitivities (71% vs. 73%)
ance of these techniques has been evaluated by several when interpreted by experts (139). In a more recent study
authors. Hatabu et al. (135) demonstrated that vessels (140), the low sensitivity of the utilized CT technique
as small as the sixth to seventh order could be reliably (70% to 72%) rather than the superior sensitivity of MRA
depicted. Meaney et al. (133), in a study comparing classic (79% to 81%) proved the likely explanation for the appar-
angiography with a single-acquisition coronal 3D gradient- ent superiority of MR in detecting central emboli. It is
echo sequence [27-second acquisition, TR/time to echo TE likely that in comparing optimal techniques for both
6.5/1.8 ms; flip angle 45 degrees, slab 8.4 to 11.2 cm, 32 modalities, the sensitivities for central PE would be com-
partitions, 30- to 36-cm field of view (FOV), 40 to 60 mL parable (141,142), whereas MR performance at the sub-
Gd at 2 mL/second followed by a saline chaser] identified 5 segmental level would at best be comparable to results
of 6 lobar PE and 16 of 17 segmental emboli with sensitivi-
ties of 75% to 100% and specificities of 95% to 100%,
respectively. Similarly, in a larger study of 118 patients
undergoing both DSA and MRA [1 breath-hold acquisi-
tion/lung, 15 to 17 seconds, 3D fast low-angle shot
(FLASH), TR/TE 3.6/1.6 ms; flip angle 25 degrees, slab 12.5
cm, 44 partitions, 20- to 32-cm FOV, 44 mL Gd at 2 mL/sec-
ond], readers demonstrated high interobserver agreement
(k 0.75) with sensitivities for central/lobar and segmental
pulmonary arteries of 100% and 84%, respectively,
although sensitivities were low for subsegmental (72%) and
isolated subsegmental (40%) PE (137). Gupta et al. (138)
reported similar sensitivities (85%) and specificities (96%),
detecting 11 of 13 emboli that were predominantly segmen-
tal or larger while missing two isolated subsegmental
emboli. Loubeyre et al. (123) identified all the central but
none of the peripheral PE identified angiographically using Figure 3-14 Acute pulmonary embolism: magnetic resonance
imaging (MRI). Five-millimeter axial T1-weighted, fat-saturated
subsecond MRI gradient sequences (Figs. 3-14 and 3-15). spoiled gradient-echo sequence, obtained following intravenous
To date, there are only a few published studies compar- gadolinium administration, demonstrates bilateral pulmonary
ing the relative performance of optimum MR technique emboli (arrows). Although visualization of the central pulmonary
arteries is adequate using this early technique, note that there are
with state-of-the-art CT for detecting PE. In one early extensive motion artifacts, rendering visualization of segmental
study, MR and CTPA performed with 5-mm collimation and subsegmental pulmonary arteries suboptimal or impossible.
A, B C
Figure 3-15 Acute pulmonary embolism: magnetic resonance angiography (MRA). A–C: Coronal 3D gradient-echo, electrocardiographic-
gated, gadolinium-enhanced MRA (TR/TE 3.2/1.4 ms, 20-second breath-hold) demonstrates improved visualization of the central pulmonary
arteries on both source coronal images (arrows in A and B) as well as oblique multiplanar reconstruction projections (arrow in C). Although
this approach provides improved visualization of peripheral arteries, resolution is still suboptimal compared with multidetector CT
pulmonary angiography.
obtained using SDCT scanners (123,137,138). Animal venous phases (145,146). Although these sequences have
model studies of subsegmental PE have recently provided sufficient temporal resolution to identify pulmonary arte-
equivalent (143) or higher (144) mean reader sensitivities rial perfusion, they offer limited spatial resolution as well
(82% to 92%) for the combined interpretation of MRA as poor depiction of the peripheral pulmonary arteries
and MR perfusion defects when compared with either (147–149). The use of partial Fourier imaging, in combi-
expiratory MDCT (144) or single-detector spiral CT (143). nation with 3D time-resolved sequences that share
However, in one of these studies (144), the MRI technique the central lines of k-space data, can offset this inherent
utilized an impractical 43-second breath-hold sequence. decrease in spatial resolution (150). With recent further
advances in gradient strength (40 mT/m) and a slew rate
of 200 mT/m/ms, Goyen et al. (145) achieved anatomic
Fast Techniques demonstration of emboli and normal anatomy in eight
The use of longer breath-hold sequences requires greater patients at the subsegmental level using only a 4-second
accuracy in the timing of contrast administration because sequence (TR/TE 1.64/0.6 ms). In a study of 39 patients,
of rapid vascular transit through the lung parenchyma. Kluge et al. (151) demonstrated that real-time MR evalua-
Moreover, because saturation pulses are ineffective at tion with true fast imaging with steady-state precession
nulling gadolinium-enhanced vessels, it remains impossi- (True FISP) resulted in an increased yield of diagnostic
ble to achieve pulmonary arterial enhancement without studies for evaluating the central pulmonary vasculature.
overlapping venous enhancement. Despite the use of sup- Parallel processing techniques such as SENSE (sensitivity
plemental oxygen via nasal cannulas, not all patients, par- encoding for fast MRI) and SMASH (simultaneous acquisi-
ticularly those who are dyspneic or in pain, are able to tions of spatial harmonics) offer an alternative mechanism
comply with the required breath-hold period, a significant to reduce breath-hold duration (152–154). These tech-
factor countering the widespread use of MRI for PE. niques utilize the spatial information inherent in the geom-
As an alternative to monophasic sequences, 2D or 3D etry of the surface coils to reduce the number of phase
rapid (5 to 10 second) time-resolved sequences may be encoding steps and hence acquisition time. This advantage
performed to differentiate between pulmonary arterial and may be utilized either to reduce the time of the examination
with preserved spatial resolution or to increase spatial reso- appropriate utilization of advanced image processing tech-
lution while maintaining a constant breath-hold duration niques. Most examinations utilize gadolinium at a dose of
(155). Both techniques, either by reduction of motion or 0.2 mmol/kg administered at a rate of 2 mL/second with a
improvement of spatial resolution, result in significant im- saline flush to maintain a compact contrast profile. The use
provement of visualization of the central and peripheral ves- of significantly higher doses likely yields little additional
sels (156). Although promising, validation in larger studies benefit (157). If a monophasic examination is performed,
is still required to assess the sensitivity for diagnosing sub- a timing run is usually required; however, with ultrafast
segmental emboli (Fig. 3-16). or time-resolved sequences the examination is started
5 seconds after contrast administration (158). The use of
Scan Performance and Interpretation agents restricted to the blood pool (e.g., Gadomer) offer
Successful performance of MRA for detecting PE depends the intriguing possibility of evaluating the pulmonary
not only on successful optimization of scan parameters vasculature without timing restrictions (159), including
but also on optimization of contrast delivery as well as evaluation of peripheral venous thromboembolism (160).
A B
C
Figure 3-16 Acute pulmonary embolism: magnetic resonance angiography (MRA). A–C: Coronal images obtained in a normal subject
using parallel processing fast imaging technique (TR/TE 3.2/1.4 ms, 20-second breath-hold) maintaining the same duration of breath holding
and sequence length as the case illustrated in Figure 3-15, demonstrating improved spatial resolution and visualization of peripheral
pulmonary arteries with progressively thicker maximum intensity projections (MIPs) (A through C, respectively). Thicker MIPs approximate
angiographic appearances but unfortunately may conceal thrombus. Superimposition of venous anatomy remains an additional problem to
date, limiting the routine use of this technique.
The disadvantage of this approach is the inevitable overlap aerosolized gadolinium (177,178) or hyperpolarized
of venous and arterial anatomy. noble gases such as helium (3He) (179). The combina-
As with CTPA, pulmonary arterial MRA requires dynamic tion of pulmonary arterial MRA to directly identify intra-
workstation evaluation. Although MIPs may render aes- luminal emboli with MR V/Q in a single comprehensive
thetically pleasing images that mimic classic angiography, examination without ionizing radiation holds great pro-
the use of these techniques may conceal a low signal inten- mise for the future but as yet has only been evaluated
sity embolus (Fig. 3-16). The best depiction of low signal in animal models (180). It remains to be seen what the
intensity filling defects is achieved by review of the axial clinical impact of these integrated MR data applications
source images or of thin section MPRs (138,155,158,161). will be for patient management.
The clinical utility of surface-rendered images and virtual
endoluminal projections has yet to be demonstrated
(141,162). Imaging Appearances of Acute Pulmonary
Embolism
Emerging and Future Developments Computed Tomography Findings

Evolving applications of MRI include direct thrombus CT findings in acute pulmonary embolism have been
imaging by utilizing the high T1 signal intensity of methe- extensively reviewed (64,68,82,83,181–183). Interpretation
moglobin present in subacute thrombi. A T1-weighted requires knowledge of the pulmonary arterial vasculature
magnetization-prepared 3D gradient-echo sequence may be and its relation to adjacent bronchial anatomy. With the
used following a water-only RF excitation pulse with an in- exception of the posterior subsegmental artery to the left
version time selected to nullify blood (155). This technique upper lobe and the proximal lingular artery, the pul-
has been used successfully in lower limb venography monary arteries run adjacent to accompanying bronchi.
(163–166) and in early studies to evaluate PE (164,167). Characteristically, the arteries course medial to bronchi in
Arterial spin labeling is a noncontrast technique using the upper lobes and lateral to them in the lingula, middle,
tagging pulses that employ plasma as an endogenous tracer and lower lobes (64).
by subtracting the signal from flowing blood from that of
stationary background tissue. One such sequence named
FAIRER (flow-sensitive alternating inversion recovery with Vascular Abnormalities
an extra radiofrequency pulse) has been successfully used in On well-opacified images acute PE appear as central filling
animal models of PE (168) and has now been evaluated in defects typically forming acute angles with adjacent vessel
early human studies (169). walls. Saddle emboli bridging the main pulmonary arteries
Although scintigraphic evaluation of perfusion is well often appear slightly ill defined, possibly owing to mild
established and even CT pulmonary parenchymal perfu- cardiac motion. Filling defects may be serpiginous, creat-
sion has been described using a dual acquisition during ing a tram-track appearance within vessels. Larger vessels
a single breath-hold in animal models (170–172), both may appear dilated either due to the presence of acute
techniques, particularly CT perfusion, are associated with thrombus itself or in smaller vessels due to reactive vasodi-
additional radiation exposure. MR perfusion imaging, latation (Figs. 3-17 and 3-18). Although vessels that
therefore, would be clearly advantageous. MR perfusion is appear cut off or completely nonopacified may result
usually performed using first-pass gadolinium-enhanced from acute emboli, these signs are less specific, as similar
MRA as part of a normal MRA study but can also be findings may be seen in chronic thromboembolic disease
performed with the use of blood pool agents with a long or as the result of nonopacification due to suboptimal
circulation time or with noncontrast techniques such as technique (64,82,83) (Fig. 3-19).
STAR-HASTE (arterial spin tagging followed by half-Fourier
single-shot fast spin echo) (173). In several of the previ- Pleuroparenchymal Findings
ously described studies demonstrating higher sensitivity for Even in the presence of extensive acute emboli the lungs
the detection of emboli, direct visualization of intra-arterial frequently appear entirely normal. The most common
filling defects and identification of perfusion defects in the pulmonary parenchymal findings are segmental and/or
absence of visualized clot were used as diagnostic criteria, subsegmental atelectasis (42). When extensive, these find-
although the contribution of each finding was not specifi- ings may conceal underlying emboli (Fig. 3-20). The pres-
cally analyzed (143,144). Quantitative pulmonary MR ence of focal ill-defined ground-glass attenuation or
perfusion correlates well with scintigraphic (174) as well as parenchymal consolidation, often peripheral and triangu-
porcine microsphere perfusion experiments (175) and has lar in distribution, although a more specific sign, remains
been suggested to have better interobserver agreement than an insensitive sign of acute embolic disease (Fig. 3-21).
scintigraphic perfusion imaging (147,176). Although pulmonary infarction is generally reported to
MR perfusion imaging has been used in animal mod- occur in only 10% of cases of documented PE (10), Coche
els in conjunction with MR ventilation imaging with et al. (40) have demonstrated such triangular opacities to
A B
C D
Figure 3-17 Acute pulmonary embolism. A–D: Sequential axial images following a bolus of intravenous contrast media in several different pa-
tients with documented pulmonary emboli show characteristic findings, including central filling defects forming acute margins to the pulmonary
artery wall (A) and serpiginous filling defects having a “tram-track” appearance (B and C). Vessels may also appear dilated, especially when large
because of the mass effect of a large acute thrombus (C, arrow) or because of reactive vasodilatation around smaller vessels (D, arrow).
A B
Figure 3-18 Acute pulmonary embolism: saddle emboli. A, B: Axial and coronal images, respectively, in a patient with a saddle embolus
bridging the main pulmonary arteries. These sometimes appear slightly ill defined, likely because of pulsation motion artifacts.
A B, C
Figure 3-19 Acute pulmonary embolism: nonspecific findings. A: Axial contrast-enhanced image through the mid thorax shows nonfilling
of the left interlobar pulmonary artery (arrow). This appearance is nonspecific and could be due either to acute or to chronic thromboembolic
disease as well as due to technical reasons, including poor contrast delivery or contrast interruption. As illustrated in Figure 3-2, nonopacifica-
tion may also be due to proximal tumor invasion. B, C: Target reconstructions at the level of the middle lobe bronchus imaged with lung and
mediastinal windows, respectively, show an appearance of intra-arterial filling defects (arrows) that, although more likely due to acute emboli,
may also be seen in patients with chronic thromboembolic disease.
A B
Figure 3-20 Acute pulmonary embolism. A, B: Axial CT section at the level of the inferior pulmonary veins imaged with mediastinal and
lung windows, respectively. In this case, the finding of subtle subsegmental emboli is easily masked by the presence of bibasilar subseg-
mental atelectasis. The presence of ancillary parenchymal and pleural abnormalities should not represent a contraindication for performing
contrast CT pulmonary angiography.
Figure 3-21 Pulmonary infarction. CT image at the level of the

right main pulmonary artery imaged with lung windows in a patient
with documented right main pulmonary artery emboli (not shown)
demonstrates a subpleural triangular wedge-shaped consolidation
and ground-glass attenuation associated with a small right-sided effu-
sion. Typical of pulmonary infarcts, this appearance likely reflects a
combination of infarction and hemorrhage as well as possible tran-
sient edema associated with a reperfusion injury.
Identification of decreased pulmonary vascularity, the

CT equivalent of a radiographic Westermark sign, is an
infrequent occurrence in acute pulmonary embolism and
is usually the product of a significant embolic load. Mosaic
attenuation of the lung parenchyma does not occur in the
acute phase of disease (Fig. 3-23).
Pleural effusions occur in association with 30% of acute
PE. Effusions may be simple transudates or hemorrhagic
exudates (184) (Fig. 3-24). Multiloculated pleural effu-
sions may also occur and have been reported to resolve
entirely with anticoagulation without the need for percuta-
neous drainage (185) (Fig. 3-25).
A B Massive Pulmonary Embolism and Computed

Figure 3-22 Resolving pulmonary infarction. A, B: Magnified CT Tomography Prediction of Clinical Outcome
images through the left lower lobe obtained at the same level in a
patient 3 months apart demonstrate the initial presentation of a pul- Massive pulmonary embolism is not synonymous with
monary infarct (A) showing the typical appearance of a subpleural anatomically descriptive terms such as “central pulmonary
triangular wedge-shaped density that subsequently resolves to
embolism” or “saddle embolism.” Rather, classification of
form a linear scar (B). This latter appearance is nonspecific, similar
to most other postinflammatory or infectious sequela and highlights embolism into “massive,” “submassive,” and “low risk” is
the difficulty in retrospectively identifying antecedent acute pul- based on physiology and is related to the presence or
monary embolism in the absence of prior documentation.
absence of associated echocardiographically detectable
right ventricular dysfunction and systemic hypotension
be present in 62% of patients with PE compared with only (Table 3-6). There are significant differences in mortality
27% without PE. It is likely that the finding of subpleural among these groups that highlight the fact that right
triangular consolidation is the result of a number of mech-
anisms occurring at various times, including pulmonary
hemorrhage, inflammatory exudates due to evolving
infarction, and/or reperfusion injury. Although infarcts
usually completely resolve, in our experience they may
also persist over a period of weeks, ultimately resulting in a
residual scar that in select cases may even be mistaken for
a pulmonary nodule or mass (Fig. 3-22).
Figure 3-24 Acute pulmonary embolism: pulmonary effusion.

Figure 3-23 Acute pulmonary embolism: global reduction of pul- Magnified CT section through the right lower lobe basilar arteries in
monary vascularity. CT section at the level of the left interlobar a patient with a dependent right-sided pleural effusion clearly
pulmonary artery imaged with lung windows. Markedly attenu- demonstrates the presence of right basilar pulmonary emboli.
ated, threadlike small pulmonary arterial and venous structures Unlike this case, effusions associated with acute pulmonary
are present in this patient with previously documented acute embolism are often hemorrhagic and frequently multiloculated. As
pulmonary embolism. Note that there is no evidence of mosaic previously illustrated (see Fig. 3-20), the presence of parenchymal or
attenuation. This finding is the CT equivalent of Westermark sign pleural disease in itself should not be a reason to avoid performing
and has been associated with increased patient morbidity. CT pulmonary angiography.
A B
C
Figure 3-25 Acute pulmonary embolism: pulmonary infarction. A, B: Non–contrast-enhanced axial CT images through the mid thorax in a
human immunodeficiency virus (HIV)–positive patient initially obtained at presentation (A) and subsequently following placement of a right-
sided pleural tube (B) show evidence of complex pleuroparenchymal cavitary disease that failed empirical therapy with both broad-spectrum
antibiotics and pigtail catheter drainage. During this period of hospitalization, a diagnosis of acute pulmonary embolism was not considered.
C: CT pulmonary angiography study at a later date confirms the presence of central pulmonary emboli as an unsuspected etiology for periph-
eral cavitary infarcts. Unlike in this case, cavitary infarcts are most frequently the result of superimposed, secondary infection.
ventricular dysfunction with associated right ventricular Patients with massive pulmonary embolism benefit
ischemia and/or infarction and not hypoxia, per se, is the from additional systemic thrombolysis as opposed to low-
key determinant of adverse clinical outcomes requiring risk pulmonary embolism patients in whom the additional
different modes of therapy (186–190). hemorrhagic risks of thrombolysis are not justified and
TABLE 3-6
CLASSIFICATION OF PULMONARY EMBOLISM GRADE
Defining Features
Pulmonary
Embolism Pulmonary Right Ventricular Systemic
Classification Embolism Dysfunction Hypotension Mortality
Massive 30%
Submassive 5%–10%
Low risk 4%
who are better treated by anticoagulation alone (191–198). techniques to rapidly dissolve emboli has been established,
In distinction, there is debate over the best treatment of they have yet to be evaluated in controlled clinical trials
submassive pulmonary embolism, with some studies sug- (190,211–217).
gesting that, although mortality may not be significantly Although right ventricular dysfunction may be present
improved by thrombolysis, progression to more significant in up to 30% to 40% of normotensive patients with pul-
disease and/or morbidity in select cases may be mitigated monary embolism, echocardiographic evaluation is not
(188,199–206). performed in all patients, in part because of limited avail-
Thrombolysis results in the dissolution of established ability. The presence of right ventricular dysfunction in
emboli, whereas anticoagulation is primarily directed at one study was associated with a 10% incidence of subse-
prevention of further clot formation or embolization. quent hypotension and a 5% incidence of death (189).
Thrombolysis results in a rapid dissolution of emboli, Not surprisingly, CT features of right cardiac strain have
with rapid pulmonary reperfusion and increased capillary been correlated with both echocardiographic features of
blood volume, and significantly reduces the incidence of right ventricular dysfunction (218–221) and increased
re-embolization. As important, there is also more rapid mortality and morbidity manifested as systemic hypoten-
improvement of right ventricular function with normaliza- sion requiring pressor therapy or embolectomy resulting
tion of right ventricular strain and systemic hypotension in a prolonged intensive care unit stay and/or hospitaliza-
(191,194–198,207,208). Thrombolysis is clearly contra- tion (221,222). Araoz et al. (223), arriving at a slightly
indicated in patients with a hemorrhagic diathesis, as there different conclusion, noted that, although CT features of
is a 2% to 3% estimated risk of intracranial hemorrhage right ventricular dysfunction were predictors of increased
as well as a 22% risk of extracerebral hemorrhage, poten- morbidity, they did not correlate with mortality.
tially as great as two units of blood or more (209,210)
(Figs. 3-26 and 3-27).
Surgical embolectomy, although considerably safer than Computed Tomography Features of Right
in previous decades, is still reserved for patients with Cardiac Strain
massive or submassive emboli for whom thrombolysis is Pulmonary arteries may increase in size owing to increased
contraindicated. The indications for mechanical disruption flow or pressure and may be definitely considered
and/or catheter-delivered localized intra-arterial thrombo- enlarged either when greater than 3.5 cm in diameter or
lysis, although poorly defined, are typically similar to indi- when between 3.0 and 3.5 cm if they exceed the size of
cations for embolectomy. Although the feasibility of these the ascending aorta at the same cross-sectional level
A B
Figure 3-26 Massive pulmonary embolism. A: Enlarged CT section through the right and left main pulmonary arteries following a bolus
of intravenous contrast media shows large bilateral pulmonary artery emboli. Note that the main pulmonary artery is enlarged. B: CT section
through the right atrium in the same patient as in A. Note that hemodynamically significant massive pulmonary embolism is demonstrated in
this case by the finding of marked dilatation of the right-sided cardiac chambers and the nearly reversed orientation of the interventricular
septum. The patient died a few hours after this examination, despite the institution of thrombolytic therapy.
A B
C D
Figure 3-27 Acute pulmonary embolism: thrombolysis. A, B: Contrast-enhanced CT sections through the right main pulmonary artery and
proximal basilar pulmonary arteries, respectively, in a young adult with hypotension following a long-haul flight show massive bilateral pul-
monary emboli. C, D: Contrast-enhanced CT images at the same levels as shown in A and B, respectively, following thrombolysis demonstrate
complete resolution of emboli, highlighting the more rapid anatomic and physiologic reversal of disease following therapy.
(221) (Fig. 3-1). It should be emphasized, however, that

the main pulmonary artery may remain unchanged in size
despite acute elevation of right heart pressures and is more
consistently enlarged as a result of chronic right-sided
heart failure (224) (Fig. 3-28).
A more specific feature of acute right cardiac strain is
dilation of the right-sided cardiac chambers when com-
pared with the left heart chambers. Optimally, chamber
size is measured at the site of maximum diameter, usually
near the base of the heart at the level of the atrioventricu-
lar valves. This is often associated with dilatation of the
right atrium with bowing of the interatrial septum toward
the left atrium or straightening of the normally convex an-
terior interventricular septum (Fig. 3-29). Compression of
the left ventricle results in reduction in cardiac output,
coronary perfusion, and eventually systemic hypotension
(Fig. 3-30). As a guide, the maximum lumen of the right
ventricle should not exceed that of the left ventricle
(218–220). Quiroz et al. (222) demonstrated that adverse
clinical outcome better correlated with measurements of
the ratio of right to left ventricular size when these were
calculated from true four-chamber projections recon- Figure 3-28 Massive pulmonary embolism. Contrast-enhanced
structed from volumetric datasets. Additional features of CT section shows enlargement of the main pulmonary artery.
right cardiac strain include reflux of contrast into the IVC: Although this represents another sign of massive pulmonary
embolism, it is a less sensitive indicator than right chamber enlarge-
Although this may be seen in normal patients, elevated ment and in fact is more frequently seen in patients with chronic
right heart pressure or tricuspid valve regurgitation should thromboembolic pulmonary hypertension.
A B
C D
Figure 3-29 Acute pulmonary embolism: cardiac evaluation. A–D: Images obtained from two patients, one with low-risk pulmonary
embolus (A and C), the other with massive pulmonary embolus (B and D). A, C: Contrast-enhanced axial and sagittal images, respectively,
through the heart in a patient with low-risk pulmonary embolus show the normal convex anterior orientation of the interventricular septum.
B, D: Axial and sagittal images, respectively, in the patient with massive pulmonary embolism corresponding in location to those shown in
A and C. Note that in distinction to the appearance of the septum in patients with low-risk pulmonary embolus, there is both straightening
and bowing of the septum resulting from increased right cardiac strain (compare A and B). Short axis reconstructions through the ventricles
(D) demonstrate increased size of the anteriorly situated right ventricle and straightening of the interventricular septum (compare B and D,
arrows). Evaluation of the right-sided cardiac chambers should be a routine component of the evaluation of all patients with suspected pul-
monary embolus.
be suspected if there is IVC reflux or dilatation inferior to The presence of right atrial or ventricular thrombus is
the diaphragm or into the hepatic veins (225) (Fig. 3-31). rare, occurring in only approximately 4% of patients evalu-
Pulmonary oligemia, a relatively rare finding in acute ated by echocardiography, as documented in a retrospec-
pulmonary embolism, should be considered an ominous tive study of the International Cooperative Pulmonary
sign frequently resulting in the need for intubation (223) Embolism Registry (ICOPER) database (226) (Fig. 3-32).
(Fig. 3-23). This finding likely reflects thrombus in transit from the
A B
C D
Figure 3-30 Massive pulmonary embolism: cardiac evaluation. A–D: Contrast-enhanced images obtained at presentation (A and C) and
following thrombolytic therapy (B and D) in a patient with massive pulmonary embolism. Contrast-enhanced CT sections through the main
pulmonary arteries (A) and right heart (C) at presentation demonstrate a saddle pulmonary embolus and abnormal configuration of the in-
terventricular septum, respectively, the latter causing a flattened appearance of the left ventricular chamber indicative of massive pul-
monary embolism. Following thrombolysis, clinical hemodynamic improvement is mirrored by normalization of CT pulmonary angiography
findings (compare A and C with B and D, respectively).
A B
Figure 3-31 Acute pulmonary embolism: evaluation of right cardiac strain. A, B: Coronal maximum intensity projections (MIPs) obtained
in a patient in whom acute pulmonary embolism was excluded show a normal appearance of the relationship between the right-sided car-
diac chambers and the inferior vena cava (IVC) and hepatic veins. C, D: Coronal MIP and coned-down axial CT image through the IVC and
liver in a different patient than in A and B with documented massive pulmonary embolus. Note that there is only minimal or absent reflux of
contrast into a normal-sized IVC (A and B). By comparison, note significant infradiaphragmatic reflux into a dilated IVC (C) and hepatic vein
reflux (D) secondary to massive pulmonary embolism consistent with right heart strain and/or tricuspid valve dysfunction (continued).
C D
lower limbs to the pulmonary arteries and is more likely to are associated with an increased incidence of both massive
be identified in patients with pre-existing cardiac failure and submassive embolism and hence were associated with
resulting in relative right heart stasis. Right heart thrombi an approximate doubling of both 2-week and 3-month
mortality (21% and 29%, respectively). Mortality in this
subset of patients in some series, despite treatment, has
been reported to be as high as 44% (227). The presence
of right heart thrombi is therefore often considered an
indication for thrombolysis even in the absence of right
ventricular dysfunction (228–230).
Gradation of Severity: Pulmonary Emboli

Obstruction Index, Myocardial Stratification
Occlusion of greater than 40% to 50% of the central pul-
monary arterial tree has been correlated with echocardio-
graphically elevated pulmonary arterial pressure (224,231).
Qanadli et al. (231) have suggested a simplified CT
obstruction index that scores 10 segmental arteries bilater-
ally as either patent (0 points), partially occluded (1 point),
or completely occluded (2 points), expressing the total as
a percentage derived from the sum of all 40 sites. This index
appears reproducible and correlates well with classic
angiographic indices (208). In separate studies, Collomb
et al. (221) validated the use of high CT obstruction
indices calculated by this method as positively associated
with the severity of PE, whereas Wu et al. (232) demon-
strated that patients with an obstruction index greater
than 60% suffered increased mortality (83%) and those
with an obstruction index less than 60% had increased
survival (98%).
In distinction, Araoz et al. (223), in their study of
patients with lesser degrees of arterial obstruction overall,
Figure 3-32 Intracardiac thrombi. Coned-down contrast- found no significant correlation between vascular obstruc-
enhanced CT section through the heart demonstrates serpiginous tion indices and either morbidity or mortality. This may in
filling defects in the right atrium suggestive of clot in transit. This
appearance is associated with higher morbidity and mortality and part be because in the absence of “overwhelming” vascular
is an indication for thrombolysis. occlusion, progression from vascular obstruction to elevated
right ventricular pressure, right ventricular dilatation, inter- Use of bolus-tracking techniques has led to more pre-
ventricular septal deviation, and finally left ventricular dictable contrast enhancement, with suboptimal opacifica-
compression may vary from individual to individual, with tion most often resulting from low or delayed peak arterial
outcomes largely influenced by comorbid conditions such opacification and only rarely occurring as a result of too
as pre-existing congestive failure. early opacification, with both factors usually related to
It appears likely that indices of vascular obstruction, poor cardiac output or restricted venous access (82).
especially when markedly elevated, will ultimately gain Motion artifacts are predominantly due to respiratory
widespread acceptance as a potential indicator for throm- or, to a lesser degree, cardiac pulsation. Image quality is
bolysis or as a means to more precisely monitor therapy. improved by ensuring that patients complete their breath-
However, it may be anticipated that this will necessitate hold before initiation of the examination or by reducing
the development of reliable automated calculations of CT the volume of coverage for dyspneic patients. A slight
obstruction indices to facilitate use of these techniques. improvement in resolution theoretically may be obtained
Increasingly, biochemical indicators other than D-dimers by adjusting the detector configuration and pitch to reduce
are being evaluated not only for the detection or exclusion the required breath-hold period, enhancing overall image
of PE but also for myocardial risk stratification in cases of quality by reducing motion. Respiratory motion artifacts
established normotensive PE (233). Troponin T, for exam- due to partial volume average effects are most apparent in
ple, is released from cardiac myocytes and is elevated in vessels oriented in the axial plane but can potentially be
approximately one third of PE patients, correlating with seen in all vessels as areas of ill-defined hypodensity. These
indices of right cardiac strain or ischemia, and is a docu- effects, which may be present on only a few contiguous
mented independent prognostic factor for fatal PE (234, sections, are best identified by simultaneously reviewing
235). Whereas troponin T is not myocyte specific and may lung window settings at the same levels, highlighting
be elevated in other conditions, such as chronic renal motion blurring (Fig. 3-33).
failure, troponin I is myocyte specific. It is envisaged that Although the number of technically suboptimal studies
a combination of increased troponin I, demonstrating has likely improved with multidetector technology, advan-
increased cardiac strain; normal creatinine kinase, demon- tages may be finite. A 64-detector scanner can acquire an
strating the absence of myocardial infarction; and elevated entire thoracic dataset in 5 seconds. As a consequence,
D-dimers consistent with pulmonary embolism may be an even minimal respiratory motion or suboptimal timing of
initial method of identifying normotensive patients with a contrast administration could affect all images, rendering
high risk of adverse outcome (236,237). High levels of pro- the entire study uninterpretable. Studies performed with a
brain natriuretic peptide (238,239) and myoglobin (240) 16-detector configuration are slightly slower (8 to 10 sec-
have also been suggested as predictors of poor outcome. In onds), during which time there is a greater probability that
combination with CT obstructive indices, these serum pro- a portion of the examination will be technically optimal.
teins may be used to guide more aggressive therapy in cases Therefore, although we await studies of the comparative
with more likely adverse outcomes. performance of 64-detector CTPA versus earlier-generation
scanners, our personal preference has been to use a sub-
maximal pitch for performing these studies (Table 3-3).
Pitfalls in Interpretation
Accurate interpretation of CTPA studies requires knowl-
Patient Factors—Anatomic
edge of a number of well-described pitfalls that may result
The presence of nodal tissue, either in hilar regions or more
in incorrect diagnoses either confirming or excluding the
peripherally at vessel branch points, may cause potential
presence of PE. These pitfalls may be classified as due to
misinterpretations (59). This pitfall is greatest in the right
technical, patient, or observer-related factors.
hilum at the level of the right main pulmonary artery as it
changes course to become the interlobar pulmonary artery.
Technical Factors Familiarity with the appearance of normal-sized nodes in
Despite the proven accuracy of CTPA for the evaluation of the right hilum, in conjunction with overlapping or MPRs,
PE, it has been estimated that approximately 5% to 10% is usually sufficient to avoid misinterpretation (Fig. 3-34).
of cases are nondiagnostic owing to technical factors In patients with interstitial edema, thickening of the perivas-
(96,241). The usual causes of uninterpretable examina- cular interstitium may cause an appearance of peripheral
tions are suboptimal contrast opacification and motion circumferential clot, which may be misinterpreted as due to
artifacts, although obesity and the presence of metallic acute or, more frequently, chronic emboli. In these patients
hardware with resultant streak artifacts may further differential enhancement of segmental pulmonary arteries
degrade studies. In our opinion, overinterpretation of due to congestive cardiac failure, pulmonary arterial hyper-
peripheral emboli in patients with suboptimal examina- tension, or alterations of pulmonary venous resistance by
tions is likely the most important cause of interobserver atelectasis and effusions may contribute to the spurious
variation and false-positive interpretations. appearance of low intravascular density (82).
A B
Figure 3-33 Acute pulmonary embolism: pitfalls in interpretation. A: Coned-down contrast-enhanced CT image through the left lower
lobe shows apparent eccentric filling defects in several left lower lobe basilar arteries. B: Identical image as in A, imaged with lung windows,
shows that the appearance of apparent emboli in A is in fact artifactual secondary to respiratory motion and cardiac pulsation. Unless
unequivocal, the presence of emboli should be confirmed by routine inspection of lung windows to obviate this frequent pitfall, especially
when evaluating peripheral subsegmental pulmonary arteries.
A B
Figure 3-34 Acute pulmonary embolism: pitfalls in interpretation.

A: Contrast-enhanced CT section through the distal right main pulmonary
artery (RMPA) shows an apparent filling defect (arrow). Hilar lymph nodes
may superficially mimic pulmonary embolism, especially at the point where
the RMPA changes course to become the right interlobar pulmonary
artery. In difficult cases, this problem may be solved either by selectively
obtaining contiguous thin sections to demonstrate normal patency of the
artery inferiorly (B) or by obtaining coronal reconstructions that demon-
C strate the extravascular nature of the low-density regions (arrows) (C).
A B
Figure 3-35 Acute pulmonary embolism: pitfalls in interpretation. A: Coned-down contrast-enhanced CT section through the proximal
left lower lobe basilar arteries shows apparent intra-arterial filling defects (arrow). B: Identical section as shown in A imaged with lung win-
dows confirms the presence of retained secretions in the basilar airways, mimicking pulmonary pitfalls in interpretation.
Although opacification of pulmonary veins may be toward the mediastinum more challenging, particularly
reduced by optimizing bolus timing, nonetheless, careful when there is associated respiratory motion (Fig. 3-20).
tracking of vessels back toward the mediastinum, espe- The presence of suspected filling defects in structures
cially on contiguous thin sections, is essential to avoid adjacent to opacified pulmonary arteries should raise the
misinterpreting early venous filling as PE. The presence of suspicion of mucoid impaction of accompanying bronchi
atelectasis may not only make PE more inconspicuous but and is readily confirmed by evaluation of corresponding
may also make tracking of peripheral venous vessels lung windows (242) (Figs. 3-35 and 3-36).
A B
Figure 3-36 Acute pulmonary embolism. A, B: Coned-down axial and coronal multiplanar reconstructions through the right lung base
and right hemithorax, respectively, demonstrate several suspected low-density filling defects (white arrows), which prove to represent mu-
coid impacted airways within a completely atelectatic right lung. Note that in this case, there is also evidence of peripheral subsegmental
emboli adjacent to these same mucoid impacted airways (black arrows).
Patient Factors—Physiologic Gosselin et al. (242) identified this artifact in 36.7% of

The presence of suspected emboli involving all central studies. Although the effect was most often mild, in at least
vessels in both lungs at the same level associated with good three cases it was severe enough to result in false-positive
proximal and distal pulmonary arterial opacification interpretations. This artifact is thought to arise because of a
should raise the possibility of a physiologic artifact termed vigorous inspiratory effort, particularly by younger patients,
contrast interruption (Fig. 3-37). In a study of 129 patients, at the initiation of the examination. This causes sudden
A, B C
D, E F
G, H I
Figure 3-37 Acute pulmonary embolism: pitfalls in interpretations. A–C: Coned-down views of sequential contrast-enhanced CT
images through the upper lobes show numerous apparent filling defects bilaterally at approximately the same levels (arrows). D–I:
Sequential contrast-enhanced images through the main and interlobar pulmonary arteries demonstrate poor contrast opacification of the
central pulmonary arteries but much improved opacification of the more peripheral lower lobe arteries, the more proximal pulmonary
outflow tract, and the right-sided cardiac chambers, respectively. This constellation of findings should suggest the presence of artifacts
due to contrast interruption.
negative intrathoracic pressure resulting in suction of a performed with MDCT scanners, especially in the lower
bolus of unopacified blood from the IVC into the right side lobes, as the velocity of arterial flow is approximately equal
of the heart, which subsequently may travel as an unopaci- to the table feed. On a single-detector scanner the bolus of
fied bolus through the right side of the heart (242,243) unopacified blood usually appears on one to two sections
(Fig. 3-38). If images are acquired as the bolus traverses the only before it moves ahead of the scanner or dissipates.
pulmonary arteries, the result is apparent filling defects in Recommendations for improving scan quality in patients
select central pulmonary arteries (Figs. 3-39 and 3-40). The with this artifact include instructing patients to take a
same physiologic effect can result from a patent foramen slower breath, ensuring a slight delay following completion
ovale or atrial septal defect (244). This finding is more of inspiration before scanning, and scanning caudocra-
likely to be encountered and more apparent on studies nially to avoid the artifact occurring in the lower lobes,
C
Figure 3-38 Acute pulmonary embolism: demonstration of the mechanism of contrast interruption. A–C: Sequential contrast-enhanced
axial images through the inferior vena cava (IVC) and right atrium demonstrate a bolus of unopacified IVC blood traveling through the right
atrium. If imaging is acquired as this bolus of unopacified blood passes through the pulmonary arterial tree, appearances mimicking
pulmonary embolism may be generated, as illustrated in Figure 3-37.
where emboli are more typically suspected. In select cases

we have found performing a repeat examination entirely in
expiration beneficial (Fig. 3-41).
Patient Factors—Tumor Mimicking Embolic Disease

As discussed in greater detail later in this chapter, direct
invasion by lung carcinoma is the commonest etiology of
tumor located within the pulmonary arteries, generally
causing little interpretative confusion owing to the presence
of extensive extravascular tumor. However, although far less
frequent, intravascular tumor may be indistinguishable
from bland acute or chronic PE (Fig. 3-42 and Figs. 3-69
to 3-72). Additional parenchymal features—in particular,
mosaic attenuation—have been described in sarcomas orig-
inating from the pulmonary arterial intima, mimicking
chronic pulmonary thromboembolic disease (245). Purely
intravascular tumor may be due either to rare primary sarco-
mas or, more commonly, to metastatic disease.
In the absence of extraluminal extension, differentiation
between tumor emboli and bland thrombus may be
exceedingly challenging, with the diagnosis frequently
established postmortem (246). The presence of parenchy-
mal metastases, particularly when large and asymmetric,
should raise suspicion of tumor embolism. A filling defect
occupying the entire lumen of a central artery, expansion of
the artery, and heterogeneous density due to enhancement
or necrosis within emboli have been described as features
Figure 3-39 Acute pulmonary embolism: pitfalls in interpreta-
tion—contrast interruption. Off-axis multiplanar reconstruction suggestive of tumor emboli (247). Peripheral microem-
along the main and left main pulmonary arteries demonstrates to bolic disease may be associated with dilated and beaded
good advantage a bolus of unopacified blood traveling through small peripheral arteries (248). Definitive evaluation
the arterial circulation as a discrete bolus. It is this phenomenon
that accounts for the appearance of apparent upper lobe emboli requires demonstration of intravascular tumor enhance-
illustrated in Figure 3-37. ment either by CT or by MRI (249) or asymmetric increased
Figure 3-40 Acute pulmonary embolism: pitfalls in interpretation—contrast interruption. Sequential coned-down coronal maximum
intensity projections through the main pulmonary arteries demonstrate the passage of unopacified blood through the central arterial tree
(arrows). Typically, contrast interruption leads to apparent bilateral filling defects generally at the same level, often appearing slightly ill de-
fined along their edges.
A B
Figure 3-41 Acute pulmonary embolism: pitfalls in interpretation—countering contrast interruption. A: Contrast-enhanced CT section
through the right main pulmonary artery demonstrates suspected contrast interruption with poor right main pulmonary artery opacification
and a suspected filling defect in the left interlobar pulmonary artery (arrow). Note the ill-defined nature of this filling defect. B: Contrast-
enhanced CT at approximately the same level as shown in A, obtained with the patient in deep expiration, demonstrates the previously
suspicious-appearing left interlobar pulmonary artery to be normal.
Figure 3-42 Acute pulmonary embolism: pitfalls in interpretation—tumor mimicking pulmonary emboli (PE). Coned-down contrast-
enhanced CT image through the right hilum demonstrates tumor infiltrating the adjacent right lower lobe pulmonary artery. This can be
distinguished from acute PE by the presence of extensive extravascular soft tissue.
metabolic activity within the pulmonary artery at fluo- absence of limited cardiopulmonary reserve or peripheral
rodeoxyglucose position emission tomography (FDG-PET) thrombus, may be of limited clinical impact. One possible
(250). Histologic confirmation using endovascular suction remedy to reduce potential false-negative studies currently
biopsy has been suggested as an alternative to surgical being evaluated is automatic detection by CAD. Particular
biopsy (98,251). areas in which CAD may prove useful include emboli in
the truncus anterior artery, which is often subject to signifi-
cant beam hardening artifacts as it passes behind the
Observational Errors densely contrast-filled SVC to supply the right upper lobe.
Because of the large number of images acquired during Fortunately, isolated upper lobe PE are unusual. Emboli in
routine MDCT pulmonary angiography, resulting in the this vessel may be diagnosed if low-density abnormalities
visualization of seemingly innumerable peripheral subseg- are entirely endoluminal and are relatively well defined or
mental pulmonary arteries, it is almost inevitable that appear to branch. The use of MPRs in this instance may be
small PE, usually isolated, may occasionally be missed by of particular assistance (Fig. 3-43).
even the most careful or experienced observer. As discussed Observational errors also result from so-called satisfac-
earlier, although controversial, it is likely that the finding tion of search. The presence of PE, for example, should
of an isolated subsegmental embolus, especially in the prompt a search for potential sources that may also be
A, B, C
Figure 3-43 Acute pulmonary embolism. A–C: Sequential

contrast-enhanced axial CT sections through the carina demon-
strate an apparent isolated pulmonary embolus in the truncus ante-
rior, the right upper lobe pulmonary artery. This artery, the first
branch of the right main pulmonary artery arising within the peri-
cardium, generally follows a steeply oblique course superolaterally
to supply the right upper lobe. On suboptimal studies, apparent
filling defects in the truncus anterior may be seen owing to beam-
hardening artifacts resulting from dense contrast in the superior
vena cava. Although initially a problem, especially with single-
detector CT studies, the use of multidetector CT scanners with
faster acquisition times has essentially obviated this pitfall. In this
case, the finding of a relatively well-defined branching low-density
filling defect is consistent with a true embolus. In select cases in
which suspicion persists, findings are easily confirmed on a corre-
D sponding coronal multiplanar reconstruction (D).
A, B, C
Figure 3-44 Acute pulmonary embolism. A–C: Sequential contrast-enhanced axial images through the superior vena cava in a patient
with documented left main pulmonary artery emboli (B and C, shorter arrows) show that the etiology of these is the presence of thrombus
forming adjacent to an indwelling venous catheter, identifiable as low-density material surrounding the distal catheter tip (A–C, longer
arrows). Although diagnostic in this case, caution should be exercised when only a small quantity of low-density material can be identified
around central venous access catheters, as this may be a normal finding due to catheter motion, beam hardening, or contrast mixing.
recognized. Central venous catheters, in particular, repre- on examinations performed for indications other than
sent a potential source of thrombi within the chest and pulmonary thromboembolism. In these patients, even with
should be closely scrutinized for the presence of adherent thicker 5- to 7-mm images, PE have been noted as inciden-
clot (Fig. 3-44). Similarly, the presence of a recognized tal findings in as many as 1% to 2% of patients (98,252).
pitfall such as hilar nodal tissue should not divert atten- In fact, the incidence is likely far higher, reported in as
tion from the need to search for concomitant PE in many as 9% of cancer patients (253,254) (Fig. 3-47). When
adjacent vessels (Figs. 3-45 and 3-46). necessary, the presence of PE incidentally suspected on thick
Due to the increasing use of MDCT, providing faster sections may subsequently be confirmed by retrospectively
image data acquisition following contrast administration, reconstructing select overlapping thin (2 to 3 mm) sections
the pulmonary arteries are frequently well visualized even (Fig. 3-48).
A B
Figure 3-45 Acute pulmonary emboli: pitfalls in interpretation. A: Contrast-enhanced axial CT image through the right and left main pul-
monary arteries in a human immunodeficiency virus (HIV)–positive patient show the combined presence of diffuse adenopathy and a left-sided
main pulmonary artery embolus. B: Contrast-enhanced axial CT image through the left main pulmonary artery in a different patient than shown
in A. In this case non–small cell lung cancer has resulted in extensive bilateral hilar adenopathy as well as direct tumor invasion of the left pul-
monary artery. Note that there is also evidence of an acute pulmonary embolus in the left main pulmonary artery. This latter case demonstrates
that tumor can mimic the appearance of thrombi due to involvement of hilar nodes, cause direct pulmonary arterial invasion, and simultane-
ously be the underlying cause of acute pulmonary thromboembolism, all potentially present in the same image.
A B
Figure 3-46 Acute pulmonary embolism: pitfalls in interpreta-

tion. A: Contrast-enhanced CT section through the distal right
main pulmonary artery raises the suspicion of a distal filling defect.
B: CT section through the right interlobar pulmonary artery shows
the presence of a more convincing intravascular filling defect.
C: Coronal reconstruction confirms both the presence of enlarged
right hilar nodes and the presence of an acute thrombus to better
C advantage.
Figure 3-47 Acute pulmonary embolism: association with un-

derlying malignancy. Off-axis coronal multiplanar reconstruction
shows a large filling defect (short arrow) in the left main pulmonary
artery in a patient with a documented tumor arising in the tail of
the pancreas (long arrow) associated with numerous hepatic
metastases. It is not uncommon to find incidental emboli in
patients evaluated for known or suspected underlying malignancy,
even when less than optimal CT technique is used (see, in addition,
Figs. 3-45B and 3-48).
B, C, D E, F, G
Figure 3-48 Unsuspected pulmonary emboli. A: Coned-down

contrast-enhanced 5-mm section through the left lower lobe
obtained as part of a routine thoracic CT study using a multidetector
CT scanner demonstrates an apparent unexpected low-density
filling defect (arrow)—in this case identifiable on only this single
prospectively reconstructed image. B–G: Select contiguous 1-mm-
thick sections retrospectively reconstructed through the same
suspect pulmonary artery convincingly confirm the presence of an
acute pulmonary embolus (arrows). H: Multiplanar reconstructions
through this same region showing the true extent of the clot to
H better advantage (arrow).
A B
Figure 3-49 Acute pulmonary embolism: identification of non–contrast-enhanced examinations. A: Coned-down non–contrast-enhanced
section through the proximal right lower lobe pulmonary artery shows the presence of a high-attenuation filling defect suspicious for a
possible pulmonary embolus. B: Contrast-enhanced CT image at the same level as shown in A confirms the presence of acute clot. Rarely,
pulmonary emboli may be identified even without contrast administration, especially in anemic patients due to abnormally decreased blood
density. Their presence, however, should always be confirmed with a follow-up contrast-enhanced examination.
An analogous situation arises when proximal PE are toms of CTEPH, a progressive increase in the incidence of
visualized on noncontrast examinations, their conspicuity angiographically diagnosed CTEPH was documented in up
depending on recognizing a difference between the density to 3.8% of patients at 2 years, beyond which no further
of the pulmonary embolus and arterial blood (253,254) cases occurred (258). Unfortunately, in a large proportion
(Fig. 3-49). Consequently, identification of emboli on of patients presenting with characteristic symptoms of
noncontrast examinations will depend not only on the dyspnea and fatigue, an original symptomatic episode
chronicity of the embolus but also on the patient’s hemat- of venous thrombosis or pulmonary embolism is typically
ocrit (255,256). The precise sensitivity for detecting PE on absent, suggesting that the incidence of CTEPH following
non–contrast-enhanced studies is unknown but undoubt- an episode of acute pulmonary embolism may be consider-
edly low. Therefore, in patients in whom intravenous CT ably higher than previously thought (257).
contrast material administration is contraindicated, At present, prognostic factors leading to the develop-
additional confirmation may be obtained by V/Q imaging, ment of chronic PE are poorly understood. In a study of
MRI, or CT using gadolinium contrast media. 62 patients, Remy-Jardin et al. (259) followed patients with
massive pulmonary embolism for a mean of 11 months
after intensive care unit admission. In this group with se-
CHRONIC THROMBOEMBOLIC vere PE, complete resolution was documented in 48%, in-
PULMONARY HYPERTENSION complete recovery in 39%, and chronic PE evolution in
13%, the only prognostic factor being a larger embolic load
Although chronic thromboembolic pulmonary hyperten- at the time of presentation. Although 10% of patients with
sion (CTEPH) is a far less common disease than acute CTEPH have lupus anticoagulant, 20% have anticardiolipin
thromboembolic disease, the precise incidence following or lupus anticoagulant, and 39% have elevated factor VIII,
an episode of acute thromboembolism is unknown. no causative defects in fibrinolysis have been identified
Pulmonary angiographic data suggest that 90% of PE either (257,260).
entirely resolve or result in only minimal residual disease, CTEPH is not merely caused by repeated episodes of em-
with normal pulmonary hemodynamics noted within bolic disease but rather is the sequela of a failure of anatomic
30 days (Fig. 3-50). Similarly, in most patients undergoing and physiologic resolution of an initial embolic event. There
cardiac catheterization, data suggest normalization of is a localized fibrotic reaction mediated by the local release
pulmonary arterial pressure within 10 to 21 days (257). In of cytokines that results in a spiral of events leading to vessel
one follow-up study of 223 patients initially treated for narrowing, hypertensive arteriopathy, vascular remodeling,
pulmonary embolism who subsequently developed symp- and recurrent in situ thrombosis. Eventually, pulmonary
A B
Figure 3-50 Acute pulmonary embolism: incomplete resolution. A: Coned-down contrast-enhanced image through the medial aspect of
the left upper lobe at the level of the aortic arch shows two discrete filling defects consistent with acute pulmonary embolism (arrows).
B: Follow-up contrast study obtained 30 days later shows that, although one of the emboli has completely resolved, there has been incom-
plete resolution of the more posterior embolus with retraction to the vessel wall (arrows). The precise rate of incomplete resolution of
pulmonary emboli or the likelihood of progression to chronic thromboembolic pulmonary hypertension is unknown.
hypertension, right-sided heart failure, and early death occur subsequently re-endothelialized. The result is broad-based
(257). It is worth emphasizing that these patients are emboli usually forming oblique angles with the intimal sur-
frequently initially referred for evaluation to exclude other face (263). The retracted nature of the clot is often easier to
parenchymal causes of symptoms that may only indirectly appreciate in larger vessels, further aided by MPRs along the
be related to cor pulmonale, necessitating a high index of long axis of the involved vessel (101) (Figs. 3-51 and 3-52).
suspicion, especially when a pattern of diffuse mosaic atten- If the thrombus is large and completely occludes the artery,
uation is encountered.
Imaging of Chronic Thromboembolic

Hypertension
CT remains the primary method of investigation for CTEPH
not only because of the exquisite demonstration of charac-
teristic pulmonary arterial abnormalities in these patients
but also because CT not infrequently allows the diagnosis of
alternative causes of pulmonary arterial hypertension, such
as chronic obstructive pulmonary disease and interstitial
fibrosis (261–263). Even with 5-mm collimation, CT proves
slightly more accurate even than pulmonary angiography for
the depiction of main and lobar arterial involvement in pa-
tients for whom surgical correlation is available (126,264).
This may in part reflect the documented insensitivity of
pulmonary angiography to detect well-endothelialized,
smoothly contoured chronic central PE (265).
Peripheral Vascular Changes Figure 3-51 Chronic pulmonary embolism. Contrast-enhanced

axial CT image in a patient with known pulmonary hypertension
In contrast to acute pulmonary thromboembolic disease, demonstrates chronic emboli in both the truncus anterior and a
large left main pulmonary artery embolus. These may be distin-
the hallmark of CTEPH is the finding of emboli incorpo- guished from acute emboli by their forming broad-based oblique
rated into the wall of pulmonary arteries, where they are margins with adjacent arterial walls.
A B
Figure 3-52 Chronic pulmonary embolism. A: Coned-down

contrast-enhanced axial CT section shows a well-defined filling
defect (arrow) within a segmental left lower lobe artery. In this
image it is unclear whether this is acute or chronic. B: A multipla-
nar reconstruction along the plane of the vessel shown in
A demonstrates that the embolus (arrow) appears to be longitudi-
nally retracted along the vessel surface, suggesting chronicity.
C: An axial image retrospectively reconstructed through the true
axis of the vessel confirms that this is a chronic re-endothelialized
C embolus (arrow).
distinguishing acute from chronic thrombus may be diffi- intraluminal webs or bands. Vessel walls may become
cult. Interpretation may be further complicated by the fact irregular and tortuous, develop focal outpouchings, or
that episodes of both recurrent acute embolism and, more demonstrate abrupt cut-off. As vascular involvement is
commonly, in situ, thrombosis may occur. In these cases, the patchy, asymmetric bilateral arterial enlargement is more
density of emboli may be of assistance. In one study, acute common than the symmetric bilateral enlargement that
pulmonary thrombi occurring in patients with underlying results from other causes of pulmonary arterial hyperten-
CTEPH proved generally lower in density (mean 33 HU) sion (263) (Figs. 3-54 and 3-55).
than chronic thrombi (mean 87 HU), likely because of clot The presence of enlarged bronchial collaterals, most
retraction, early calcification, and possibly mild enhance- notable in the subcarinal space, is another indirect
ment occurring in organizing thrombus (266). diagnostic sign far more commonly seen in CTEPH
With time, peripheral clot may become calcified or (50% to 68%) than in either acute pulmonary embolism
recanalized (Fig. 3-53). Incomplete anatomic resolution (7%) or primary pulmonary arterial hypertension (14%)
may result in the formation of synechiae appearing as (126,267,268) (Fig. 3-56). Enlarged systemic collaterals
A B
Figure 3-53 Chronic pulmonary embolism. A: Coned-down contrast-enhanced axial CT image through the left main pulmonary artery
shows the presence of a large acute pulmonary embolus. B: Contrast-enhanced CT image at the same level as shown in A obtained
9 months later demonstrates partial recanalization of an eccentric thrombus, an additional finding confirming the diagnosis of chronic
pulmonary embolism.
A B
Figure 3-54 Chronic pulmonary embolism. A–F: Sequential coned-down contrast-enhanced CT sections through the central pulmonary
arteries demonstrate multiple findings consistent with chronic thromboembolic disease, including smoothly contoured eccentric thrombus
(A and B), especially well seen along the posterior wall of the right main pulmonary artery (arrows in B); complex intravascular webs, espe-
cially prominent in the left interlobar pulmonary artery (arrow in C); bandlike indentations in the right interlobar pulmonary artery (arrow in
D); eccentric mural thrombus and wall irregularity in the proximal left lower lobe pulmonary artery (arrow in E); and focal outpouchings of
the right lower lobe pulmonary artery (arrow in F). Note that in F there is also evidence of a peripheral infarct. These findings both sepa-
rately and together are sufficient to definitively diagnose chronic embolic disease (continued).
C D
E F
Figure 3-55 Chronic pulmonary embolism. A, B: Coned-down

contrast-enhanced CT images through the left lower lobe show
marked tortuosity of left lower lobe arteries, another finding in pa-
tients with chronic thromboembolic pulmonary hypertension, in
A, B this case associated with focal infarcts (see Fig. 3-58A).
A B
Figure 3-56 Chronic pulmonary embolism: bronchial artery

collaterals. A: Coned-down contrast-enhanced CT image through
the subcarinal space shows enhancement of enlarged bronchial
arteries (arrows). B, C: Axial and coronal maximum intensity
projections, respectively, show the size and course of the hyper-
C trophied bronchial arteries to good advantage.
originating from the inferior phrenic, intercostal, and in- with either CTEPH or primary pulmonary arterial hyper-
ternal mammary arteries, in particular, may also be seen, tension and in situ thrombosis.
in one study identified in 45% of patients with chronic
thromboembolic disease but in no patient with primary
Cardiac and Main Pulmonary Artery Findings
pulmonary hypertension (268). These radiologic findings
corroborate earlier MR perfusion and angiographic dye- A prominent feature of chronic thromboembolic disease
dilution studies in which patients with CTEPH demon- is the demonstration of pulmonary arterial hypertension
strated a mean of approximately 30% of systemic blood with features of right-sided heart failure. These findings are
flow to the lung parenchyma from bronchopulmonary comparable to those in the minority of patients with acute
shunts, in distinction to patients with primary pulmonary pulmonary embolism that demonstrates features of right
hypertension, in whom enlarged systemic collaterals were cardiac strain but are more chronic. As a consequence, not
either absent or measured less than 2% of total paren- only are findings of enlargement of pulmonary arteries
chymal perfusion (269,270). Identification of enlarged and the right-sided cardiac chambers and reflux of contrast
systemic collaterals, therefore, should further assist in dif- into the dilated IVC and hepatic veins more common and
ferentiating three radically different groups of patients: advanced in patients with CTEPH, but features of right
those with acute thromboembolic disease versus those ventricular hypertrophy are also apparent. Specifically,
pathologic thickening of the anterior free wall of the Parenchymal Findings

right ventricle (4 mm), the moderator band, or right
ventricle trabeculae becomes apparent. The latter feature is Mosaic attenuation of the lung parenchyma is one of
most prominent at the base of the heart, where the the key features of CTEPH, seen in up to 75% of patients
additional trabeculae may superficially resemble cardiac but in only 5% to 10% of patients with cardiac or primary
motion (Fig. 3-57). Within these hypertrophied and dila- pulmonary causes of pulmonary arterial hypertension
ted right-sided chambers, adherent thrombus is also more (271–274). When combined with a notable disparity in
commonly seen than in acute PE (263). the size of segmental arteries, this finding is considered
A B
C D
Figure 3-57 Chronic thromboembolic pulmonary hypertension: cardiac evaluation. A: Coned-down contrast-enhanced CT image shows
marked dilatation of the main pulmonary artery when compared with the adjacent ascending aorta. B–D: Sequential contrast-enhanced
images through the right atrium and ventricle, respectively, show that the anterior free wall of the right ventricle is thickened, measuring
more than 4 mm (arrow in B), that the moderator band is also thickened in the dilated right ventricle (arrow in C), and that, inferiorly, there is
marked trabeculation of the right ventricle (D), identifiable as crisscrossing linear densities.
highly specific for CTEPH (273) (Fig. 3-58). Mosaic atten- pulmonary arterial hypertension and right-sided heart fail-
uation is usually not accentuated in expiration, suggesting ure, leads through ventricular interdependence to improved
that it occurs as a result of reduced perfusion in areas of left atrial filling, interventricular septal motion, and left
embolic disease and redistribution of arterial perfusion to ventricular function (285).
nonobstructed vasculature (275–277). However, Arakawa Because of advances in surgical technique and improved
et al. (278,279) have recently suggested that mosaic atten- preoperative selection of potential candidates, there has
uation may be contributed to by air trapping that is most been a significant reduction in the mortality associated with
pronounced in areas of hypoperfusion. This finding may this procedure, from 17% in the period from 1970 to 1990
be related to the radiologic observation of mild bronchial to 4.4% in the period from 1998 to 2002 (283,284). Not
dilatation in areas of vascular occlusion and hypoperfu- surprisingly, CT plays an important role in characterizing
sion, a finding only rarely seen in patients with acute suitable surgical candidates (273,274,286). As surgery is
pulmonary embolism (280). technically more demanding for smaller caliber vessels, the
Peripheral scarring is also common in CTEPH as a result best surgical results are achieved in patients with isolated
of prior infarction or inflammation, but its frequent occur- central nonocclusive disease associated with only a limited
rence in other causes of pulmonary arterial hypertension number of peripheral stenotic vessels. Although the absence
makes it a poor discriminator (273). It must be noted, of central disease does not preclude a good surgical out-
however, that in the investigation of a patient with pul- come, the presence of isolated peripheral stenotic vessels,
monary arterial hypertension, focal scarring in association isolated mosaic attenuation, or significant comorbid dis-
with CTEPH may result in restrictive pulmonary physiology ease, such as emphysema or interstitial fibrosis, adversely
in the absence of other underlying restrictive lung disease affects prognosis (273,286–288). Surgical planning requires
(281) (Fig. 3-58). mapping the true extent of disease, as well as an evaluation
of the extent of arterial occlusion and the location and
extent of bronchopulmonary shunts (283,284). In this
Preoperative Evaluation and Postoperative
regard, the potential for virtual CT angioscopy to replace
Changes presurgical angioscopic planning offers considerable future
Unlike causes of pulmonary arterial hypertension, chronic promise (Fig. 3-59).
thromboembolic disease is potentially curable by throm- Although the role of CT in depicting CTEPH is well
boendarterectomy, a procedure in which the thrombus and established, the role of MRI is less clear. In a comparison
pulmonary artery intima are resected (282–284). Resultant of early studies comparing CT and MR, CT demonstrated
improvement in pulmonary perfusion, with diminished significantly higher accuracy for the evaluation of chronic
A B
Figure 3-58 Chronic thromboembolic pulmonary hypertension (CTEPH): pulmonary parenchymal evaluation. A: Coned-down axial CT
section through the left lower lobe in a patient with documented CTEPH imaged with lung windows (corresponding to Fig. 3-55) confirms
the presence of infarcts in the periphery of the left lower lobe associated with markedly tortuous proximal pulmonary arteries. B: Coned-
down axial CT section through the right lower lobe in the same patient shown in A imaged with lung windows shows the characteristic
appearance of mosaic attenuation. Note that, in distinction to patterns of mosaic attenuation identified in patients with either small airway
disease or infiltrative lung disease, patients with CTEPH demonstrate significantly smaller vessels in geographic areas of lower attenuation
and enlarged arteries in areas of higher lung attenuation. It is important to recognize pulmonary parenchymal findings in these cases, as not
infrequently high resolution non–contrast-enhanced studies are first ordered to evaluate otherwise unexplained dyspnea.
A B
C D
Figure 3-59 Chronic pulmonary embolism: virtual endoluminal angioscopy. See Color Figure 3-59B,D. A, B: Multiplanar reconstruction
(MPR) and corresponding virtual angioscopic image through the distal left main pulmonary artery demonstrate that, unlike the MPR showing
apparent complete arterial occlusion, the virtual angioscopic view documents that clot is actually eccentric (arrows). C, D: Contrast-enhanced
axial CT image and corresponding virtual angioscopic image through the lower lobes in the same patient as in A and B show that the irregular
nature of more peripheral nonocclusive thrombi is highlighted to good advantage on the accompanying endoluminal projection. Such render-
ings may assist in surgical planning for thromboendarterectomy.
thrombus in the central and segmental arteries (79% well with reduction of mean pulmonary arterial pressure
and 76%, respectively) when compared with contrast- and pulmonary vascular resistance following thrombo-
enhanced MR using a gradient-echo recall technique (43% endarterectomy (290).
and 59%, respectively) (126). Although current MR Following thromboendarterectomy, the pulmonary
techniques are far superior for demonstrating the central arterial intima may demonstrate only subtle evidence of
pulmonary arteries (289), conventional angiography and prior surgical intervention. However, the volume of clot is
CT remain superior for depicting the patency of sub- reduced, and previously identified webs and bands will
segmental vessels (290,291). However, although CT and have been resected. Features of secondary pulmonary arte-
MRA are both effective at morphologically demonstrating rial hypertension, right heart dilatation, and functional
successful outcome following thromboendarterectomy tricuspid regurgitation are gradually reversed (294). In
(290,292,293), MR has the advantage of also providing our experience, there may be a reversal of the pattern of
functional analysis. Quantitative MR measurements of mosaic attenuation, suggesting reactive hyperperfusion to
improved right ventricular ejection fraction and reduced within areas of the lung parenchyma that were previously
pulmonary artery velocity have been shown to correlate underperfused (Fig. 3-60).
A B
C D
E F
G H
Figure 3-60 Chronic pulmonary thromboembolism: thromboendarterectomy. A, C: Contrast-enhanced axial CT images through the
central pulmonary arteries in the same patient shown in Figure 3-54 show characteristic findings of chronic pulmonary emboli. B, D:
Contrast-enhanced axial images corresponding to A and C, respectively, following thromboendarterectomy show that previously identified
eccentric filling defects, webs, and bands have been resected and are no longer apparent. E–H: Axial images through the lungs imaged with
lung windows before (E and G) and after (F and H) surgery in the same patient. In addition to reversal of the features of right cardiac failure,
mosaic attenuation is significantly improved with new foci of ground-glass density appearing in areas of prior hypoperfusion (arrows),
suggesting that there has been postsurgical reperfusion of these areas.
MISCELLANEOUS ABNORMALITIES absence, as the pulmonary arterial tree is characteristically

reconstituted and patent within the corresponding
lung (295). Resulting from embryologic disappearance
Congenital Anomalies of the proximal portions of either the right or the
A number of congenital abnormalities may affect the main left sixth arch, respectively, pulmonary artery interruption
and the right and left pulmonary arteries, with and without may occur either as an isolated event or, especially when
associated cardiac abnormalities (6). These need to be left sided, in association with a right aortic arch or other
distinguished from pulmonary venous anomalies. congenital cardiac anomalies, most often tetralogy of
Fallot. These patients are prone to several complications,
including dyspnea, recurrent infections, and hemorrhage,
Unilateral Interruption of a Pulmonary Artery
likely a result of hypertrophied bronchial arteries
The radiologic features of interruption of the right or (296,297).
left pulmonary artery are frequently characteristic and in- Both interrupted and hypoplastic pulmonary arteries
clude a small hemithorax, ipsilateral displacement of the are easily identified with both CT and MR (Fig. 3-61)
mediastinum, absence of the corresponding pulmonary (298,299). In most cases, not only can the diagnosis be
artery, and reticular densities along the pleura and within established but additional information concerning the
the lung, usually attributed to systemic collateral circula- presence and extent of collateral circulation also may
tion. The term interruption is preferable to the term be obtained.
A B
C
Figure 3-61 Congenital absence of the right pulmonary artery. A, B: Contrast-enhanced, 1.5-mm sections through the carina and sub-
carinal space, respectively, show the lack of an identifiable right pulmonary artery. Note the presence of markedly enlarged intercostal ar-
teries serving as collateral vessels to the right lung (arrows). C: A 1.5-mm section obtained through the lower lobes imaged with wide
windows shows marked volume loss in the right lung. The pleural surface appears beaded because of the presence of prominent intercostal
arteries serving as collateral vessels. Note the presence of smaller than usual branches of the right inferior pulmonary vein compared with
the normal-sized left inferior pulmonary vein (asterisk).
Hypogenetic Lung (Scimitar) Syndrome and exertional dyspnea most commonly reported. Con-
and Partial Anomalous Pulmonary genital heart disease is also common, identified in approx-
Venous Drainage imately 25% of cases, most often septal defects or patent
ductus arteriosus. Still rarer is an association with this syn-
Anomalies of the right pulmonary artery may also be seen drome and “horseshoe lung,” in which the right and left
in association with hypoplasia of the right lung and lower lobes are connected by an isthmus of lung tissue
partial anomalous pulmonary venous return to the IVC, crossing behind the heart (301).
the so-called scimitar syndrome (Fig. 3-62) (300). This Various forms of partial pulmonary venous return have
rare anomaly is characterized by the following features: been reported to occur in approximately 5% of the popu-
(a) hypoplasia of the lung usually associated with abnor- lation and are usually asymptomatic. Anomalous veins
mal lobar segmentation, (b) hypoplasia of the ipsilateral drain into a variety of vessels: on the right side typically
pulmonary artery, (c) anomalous pulmonary venous into either the superior or inferior vena cavae, the azygos
return—either partial or complete—to the IVC, and (d) vein, or directly into the right atrium; on the left side typi-
anomalous systemic arterial supply to the hypoplastic cally into the left brachiocephalic vein or, less commonly,
lung. In any individual case, these features are variably a persistent left-sided SVC or coronary sinus. Both the CT
expressed. Typically seen in younger individuals, nearly (302) and MR appearances of anomalous pulmonary veins
half are symptomatic, with episodes of recurrent infection have been well described (Figs. 3-62 to 3-64) (303–306).
A B
C D
Figure 3-62 Partial anomalous pulmonary venous return: scimitar or hypogenetic lung syndrome. A–C: Sequential axial images obtained
following a bolus of intravenous contrast media beginning at the level of the carina and extending to the lung base show a markedly dilated
aberrant right inferior pulmonary vein coursing vertically (arrow in B) to connect inferiorly with the inferior vena cava (arrow in C). Note that
the right heart chambers are markedly dilated, consistent with a left to right shunt. D: Coronal multiplanar reconstruction showing the course
of the anomalous vein to better advantage.
B C
D E
Figure 3-63 Scimitar (hypogenetic lung) syndrome. A: Posteroanterior chest radiograph shows volume loss on the right side, with shift of
the heart and mediastinum. An ill-defined linear density is apparent in the right lower lobe, medially, suggesting aberrant venous drainage
(arrow). B–D: Every-beat gated axial MR images through the carina and the middle and lower thorax, respectively. Marked volume loss is
apparent on the right. The right main pulmonary artery is well defined and is normal in caliber (arrow in B). The right inferior pulmonary vein
cannot be identified in its normal location (compare with the normal left inferior pulmonary vein, arrow in C). Instead, anomalous drainage of
the right inferior pulmonary vein into the inferior vena cava can be seen (arrow in D). E: CT scan through the right lower lobe, imaged with
lung windows, shows enlarged inferior pulmonary veins coalescing within the right lower lobe (arrow in E). These are easily defined with CT,
unlike MR (compare with D). (From Naidich DP, Rumancik WM, Ettenger NA, et al. Congenital anomalies of the lungs in adults: MR diagno-
sis. AJR Am J Roentgenol. 1988;151:13–19, with permission.)
A B
Figure 3-64 Partial anomalous venous drainage. A, B: Axial and coronal maximum intensity projections, respectively, show an anomalous
left superior pulmonary vein draining into the left brachiocephalic vein.
Anomalous Origin of the Left Pulmonary Artery artery with diffuse long segment tracheal stenosis. In these
cases, tracheal stenosis is due to complete cartilage rings
Another rare pulmonary arterial anomaly is the so-called with absence of any significant pars membranacea. Unlike
pulmonary artery sling, or anomalous origin of the left cases of isolated aberrant left pulmonary arteries, in cases
pulmonary artery (307–309). In this condition, an anom- with the ring-sling complex, repair of the anomalous pul-
alous, retrotracheal left pulmonary artery arises from the monary artery does not lead to symptomatic relief of stri-
extrapericardial segment of the right pulmonary artery and dor. Identification of anomalies of the pulmonary arteries
wraps around the junction of the trachea and right main- is especially important in the evaluation of patients with
stem bronchus to pass in front of the esophagus on its way cyanotic heart disease (310,311). In these cases, informa-
to the left lung (Fig. 3-65). En route, the anomalous artery tion concerning the presence and size of the pulmonary
results in a sling that may cause compression of the right arteries is frequently of vital importance in determining
mainstem bronchus, leading to chronic stridor. operative strategy. This evaluation may be accomplished
Berden et al. (309) coined the term “ring-sling” to using either volumetric CT with or without multiplanar
describe the association of an aberrant left pulmonary reformations (312) or MR (313).
Pulmonary Artery Aneurysms

The role of CT in the evaluation of pulmonary artery
aneurysms has been well documented (314). Although
rare, pulmonary artery aneurysms may be idiopathic or
associated with other conditions, especially patent ductus
arteriosus and both atrial and ventricular septal defects.
They also occur in association with congenital abnormali-
ties of the pulmonary valve, with congenital and acquired
primary vascular diseases, including both cystic medial
necrosis and atherosclerosis, with hereditary telangiectasia,
and with penetrating and nonpenetrating trauma (314–
316). It should be emphasized that aneurysms involving
Figure 3-65 Anomalous left pulmonary artery. Contrast-
enhanced scan outlining the anomalous course of the left pul- the pulmonary arteries should be distinguished from
monary artery swinging behind the trachea to reach the left idiopathic dilatation of the pulmonary trunk. This benign
hilum (arrow). (From Moncada R, Demos TC, Churchill R, Reyes C. condition is a rare congenital anomaly resulting in abnor-
Case report. Chronic stridor in a child: CT diagnosis of pulmonary
vascular sling. J Comput Assist Tomogr. 1983;7:713–715, with mal dilatation of the main or right and left pulmonary
permission.) arteries in the absence of pulmonary hypertension or
A B
Figure 3-66 Idiopathic pulmonary artery aneurysm. See Color Figure 3-66B. A, B: Three-dimensional volumetrically rendered reconstruc-
tions showing the aneurysm from the left side and from above, respectively.
other possible etiologies for pulmonary arterial aneurysms prognosis, with death reported to occur in up to 30% of
(Fig. 3-66) (6). patients within 2 years (321).
Pulmonary artery aneurysms also are associated with Although the leading cause of pulmonary artery
generalized vasculitis, including Behçet disease, Hughes- aneurysms is usually thought to be cystic medial necrosis,
Stovin syndrome, and, less commonly, giant cell arteritis. In in association with both congenital and acquired cardio-
patients with Behçet disease, pulmonary artery aneurysms vascular diseases accompanied by pulmonary hyperten-
occur in approximately 65% of cases (with occlusions sion, investigators are increasingly aware of the problem
occurring in the remaining 35%) and are seen in associa- of false aneurysms of the pulmonary artery developing as
tion with a variety of vascular abnormalities, including a consequence of Swan-Ganz catheterization (Fig. 3-68).
venous occlusions and arterial aneurysms (Fig. 3-67) (317). Ferretti et al. (322), in a review of seven false aneurysms in
These patients also develop aphthous stomatitis, genital five patients identified over a 4-year period, noted that all
ulcerations, and uveitis. In most cases, aneurysms are easily these lesions occurred in either segmental or subsegmental
identified with either CT or MR (318–320). Unfortunately, arteries, in some cases as late as 19 days after removal of
pulmonary artery aneurysms are associated with a poor the catheter.
A B
Figure 3-67 Behçet disease. A, B: Axial images obtained following a bolus of intravenous contrast media show bilateral pulmonary artery
aneurysms (arrows) in a patient with documented Behçet disease. C: Axial section corresponding to B imaged with lung windows shows the
presence of ill-defined ground-glass infiltrates in both lower lobes due to hemorrhage. D: Sagittal multiplanar reconstruction showing the
appearance of the left-sided aneurysm in relation to the left main pulmonary artery.
C D
A B
C D
Figure 3-68 Posttraumatic pulmonary artery aneurysm. A: Posteroanterior radiograph shows the presence of a malpositioned Swan-Ganz
catheter with its tip directed toward the base of the right lower lobe. Pneumoperitoneum is due to abdominal surgery. B, C: Sequential CT
sections through the right main pulmonary artery (arrow in B) and right lower lobe, respectively. A large pleural effusion is present, causing
compression atelectasis of both the middle and right lower lobes. Within the collapsed middle lobe, marked contrast enhancement can
be identified within an apparent vascular structure (arrow in C). D: Coned-down view from a selective right pulmonary artery arteriogram shows
a peripheral pulmonary artery aneurysm corresponding in location to the tip of the Swan-Ganz catheter shown in A and the large peripheral
vascular structure identified within the middle lobe in C. (Case courtesy of Deborah Reede, MD, Long Island College Hospital, New York,
New York.)
A, B
Figure 3-69 A–D: Pulmonary artery sarcoma.
Select 5-mm axial sections following a bolus
of intravenous contrast media shows the pres-
ence of an irregular filing defect in the main
(arrowhead in C) and both the right (arrows in
A and B) and left pulmonary arteries. At biopsy,
this lesion proved to be a primary pulmonary
artery sarcoma. At least one of the nodules
identified in the right lower lobe has the same
density as the tumor noted throughout the right
and left main pulmonary arteries (arrow in D),
a finding suggesting that this may actually
represent tumor within a focally distended pul-
monary artery branch. (From Naidich DP.
Volumetric CT in the evaluation of focal pul-
monary disease. In: Remy-Jardin M, Remy J,
eds. Spiral CT of the chest. Berlin: Springer;
C, D 1997:129–151, with permission.)
Radiologic diagnosis is frequently problematic (314).

Although pulmonary artery aneurysms are generally
visible, their appearance is rarely specific. Peripheral
aneurysms may be mistaken for parenchymal nodules or
metastases, unless accompanying vessels are identified;
central lesions are easily confused with hilar adenopathy,
masses, or even aortic aneurysms. Although definitive
diagnosis traditionally required pulmonary angiography,
both contrast-enhanced CT and MR have effectively
replaced diagnostic pulmonary angiography in all cases
save those for which therapeutic intervention is planned
(323–325).
Neoplastic Involvement of the Pulmonary

Arteries
In addition to PE, intravascular filling defects may also
result from both primary and secondary (metastatic)
tumors (Figs. 3-43 and 3-69 to 3-72). Although primary
vascular tumors arising in the pulmonary arteries are rare,
they are not unknown (Figs. 3-69 to 3-71) (20,326).
Typically sarcomatous, these tumors may originate within Figure 3-70 Pulmonary artery neoplasm: MR evaluation. Oblique
the right ventricle and extend through the valve into the sagittal source image from breath-hold gadolinium-enhanced
three-dimensional MR angiogram (TR/TE 5/2 ms) demonstrates
main pulmonary arteries, or, alternatively, they may arise an enhancing mass within the pulmonary artery from metastatic
directly within the pulmonary arteries themselves. melanoma (arrow).
A B
C D
Figure 3-71 A: CT scan at the level of the carina after a bolus injection of intravenous contrast medium shows a focal filling defect within
the right main pulmonary artery (arrow). This appearance is nonspecific and may result from either tumor or thrombus. B: Every-beat gated,
non–contrast-enhanced spin-echo MR section at the level of the carina shows an eccentric, slightly nodular intravascular filling defect within
the right main pulmonary artery (compare with A). Additionally, a subtle increase in signal intensity can be seen within the right lung as com-
pared with the left lung, presumably reflecting vascular stasis and congestion. C: Every-beat gated, non–contrast-enhanced spin-echo MR
image at the same level as A and B, 3 months later. A lobular mass of intermediate signal intensity is now clearly identifiable filling most of
the lumen of the right main pulmonary artery, extending medially to partially obstruct the lumen of the left main pulmonary artery (arrow).
A significant increase in the size of this lesion is apparent. D: Every-beat gated spin-echo MR image at the same level as shown in C
obtained 1 minute after intravenous injection of gadolinium-diethylenetriamine penta-acetic acid (Gd-DTPA). Note the marked increase in
the signal intensity of the intraluminal filling defect within the right main pulmonary artery as compared with the precontrast-enhanced
image (compare with C). This finding is consistent with the presence of vascularized tissue, as occurs within tumor, as distinct from a nonvas-
cularized intraluminal thrombus. Again, note the presence of increased signal within the right lung, which also shows considerable signal
enhancement after the intravenous injection of Gd-DTPA (compare with A). This lesion is a surgically verified pulmonary artery sarcoma.
(From Weinreb JC, Davis SD, Berkman YM, et al. Pulmonary artery sarcoma: evaluation using Gd-DTPA. J Comput Assist Tomogr.
Histologically, pulmonary artery sarcomas are most fre- MR appearance of intravascular tumor and chronic organ-
quently undifferentiated; specific subtypes include, most izing PE, in particular; in both cases, the result is often
commonly, leiomyosarcomas and myxosarcomas, with irregularly marginated filling defects. This problem may
rhabdomyosarcomas, fibrosarcomas, chondrosarcomas, be solved, however, by noting contrast enhancement
and malignant mesenchymomas accounting for the within tumor using either CT or gadolinium-enhanced MR
remainder. Considerable overlap exists in both the CT and (249,327,328).
In addition to primary vascular tumors, tumor within overinterpretation should be avoided. Specifically, the find-
pulmonary arteries may be caused by tumor arising within ing of tumor adjacent to the main pulmonary arteries,
the lung, hilum, or mediastinum. Although most cases are regardless of extent, should not be interpreted as definitive
secondary to bronchogenic carcinoma (Fig. 3-42), both evidence of unresectability. Although tumor obstruction of
extrathoracic metastases and primary intracardiac tumors hilar vessels is most often secondary to bronchogenic carci-
may invade or cause obstruction of both pulmonary arter- noma, other lesions, both benign and malignant, have
ies and veins. Although rare, this type of tumor extension is been reported to cause pulmonary artery obstruction
easily identifiable with both contrast-enhanced CT and MR (330–332).
(Fig. 3-72) (329,330). Encasement of the main pulmonary In addition to finding clots in patients with underly-
arteries is an absolute contraindication to resection in pa- ing malignancies, on occasion it is also possible to iden-
tients with bronchogenic carcinoma (Fig. 3-45B). Although tify intravascular pulmonary metastases with CT. First
recognition of this condition is usually not problematic, described by Shepard et al. (247), these typically occur in
A B
C D
Figure 3-72 Pulmonary artery metastases: melanoma. A, B: Sequential, electrocardiogram (ECG)-gated, sagittal spin-echo images
through the right ventricular outflow tract and main pulmonary artery, respectively, show intermediate signal throughout the right ventricle
(arrow in A) and within the proximal main pulmonary artery (arrow in B). C, D: ECG-gated axial spin-echo image and the corresponding
contrast-enhanced axial CT section through the right and left ventricles confirm extensive intracardiac tumor. Biopsy confirmed metastatic
melanoma.
patients with adenocarcinomas of the breast, gastrointesti- ADDITIONAL CAUSES OF PULMONARY

nal tract, lungs, prostate, and kidneys and are only rarely ARTERY OBSTRUCTION
identified premortem. As these lodge in peripheral sub-
segmental arteries, CT findings include peripheral vessels The CT appearance of pulmonary artery obstruction sec-
with subtle nodular contours (Fig. 3-73). Intravascular ondary to fibrosing mediastinitis, in particular, is well
metastases have also been reported to result in a tree- documented (Fig. 3-74) (334–336). This disorder causes a
in-bud appearance, although in our experience this is wide spectrum of clinical and radiographic changes,
exceedingly rare (333). from incidental, asymptomatic disease to life-threatening
A B
C, D E
F G
Figure 3-73 Intravascular tumor emboli. See Color Figure 3-73F,G. A, B: Sequential axial images through the lower lobes show the ap-
pearance of markedly tortuous and slightly beaded right lower lobe pulmonary arteries. C: Axial image through the right interlobar pul-
monary artery following a bolus of intravenous contrast media shows extensive clot within the pulmonary arteries. In this case, filling de-
fects in both the central and peripheral pulmonary arteries were the result of tumor emboli from known osteogenic carcinoma. D, E:
Sequential axial images through a different patient than in A through C show dilated peripheral arteries in the right lower lobe and thick-
ened interlobular septa in the left lower lobe. F, G: Histologic sections obtained from an open lung biopsy performed in the right lower
lobe show evidence of tumor both in lymphatics (F) and within the lumen of a peripheral pulmonary artery due to metastatic breast
cancer. (D–G courtesy of Jane Ko, MD, New York, New York.)
A B
C D
Figure 3-74 Fibrosing mediastinitis. A: A posteroanterior radiograph shows markedly diminished ventilation in the right lung, associ-
ated with a shift of the heart and mediastinum to the right. Right hilar structures are difficult to identify. B: A pulmonary angiogram
shows abrupt cutoff of the right main pulmonary artery, raising the possibility of obstruction secondary to tumor. C: Every-beat gated,
axial spin-echo MR scan obtained just below the carina. The left pulmonary artery is slightly enlarged (straight arrow). The right main pul-
monary artery is markedly attenuated (curved arrow). Although a slight increase in the signal within the mediastinum adjacent to the right
main pulmonary artery can be seen, the appearance is nonspecific. No obvious mediastinal mass is seen. D: Nonenhanced CT scan at the
same level as C shows extensive mediastinal calcification in the region surrounding the right main pulmonary artery. Extensive mediasti-
nal and hilar calcifications were present as well at numerous other levels (not shown). (Courtesy of Jerry Patt, MD, Union Memorial
Hospital, Baltimore, Maryland.)
hypertension. Usually the result of histoplasmosis, a char- ing aorta (338). In addition to direct compression, the
acteristic appearance of widespread mediastinal and main pulmonary arteries may also be compromised when
hilar calcifications causing narrowing or obliteration of aneurysms rupture. As the ascending aorta and main pul-
both pulmonary arteries and veins should suggest the monary arteries share a common adventitia, hemorrhage
diagnosis (337). Although alterations in the configuration may track directly into the pulmonary arterial wall, result-
of mediastinal and hilar vessels secondary to fibrosing ing in obstruction. In these cases, it is necessary to closely
mediastinitis have been documented with MR, the evaluate the aorta, as the cause of the hemorrhage may be
inability of MR to detect calcifications is a clear disadvan- subtle, as may be seen in patients with focal penetrating
tage (336). atherosclerotic ulcers (Fig. 3-75). Unfortunately, when the
Other nontumorous causes of pulmonary artery com- cause is other than vascular, accurate diagnosis usually ne-
pression include aneurysms of the ascending and descend- cessitates histologic evaluation.
A C
Figure 3-75 Extrinsic compression of the right main pulmonary

artery (RMPA). A: Axial image through the RMPA obtained following
a bolus of intravenous contrast media shows marked narrowing of
the length of the RMPA by an extrinsic soft tissue mass (asterisk).
Note that there is also complete occlusion of the left interlobar pul-
monary artery (arrow). B, C: Corresponding axial and sagittal multi-
planar reconstructions through the ascending aorta, respectively,
show the presence of a small penetrating atherosclerotic ulcer
(arrows) causing an intrapericardial hematoma. Due to a shared
adventitia, hemorrhage has extended into the wall of the RMPA,
resulting in near complete occlusion with consequent obstruction of
the left interlobar pulmonary artery. (Case courtesy of David Levin,
B MD, Los Angeles, California.)
REFERENCES 6. Castaner E, Gallardo X, Ramola J, et al. Congenital and acquired

pulmonary artery anomalies in the adult: radiologic overview.
Radiographics. 2006;26:349–371.
1. Kuriyama K, Gamsu G, Stern RG, et al. CT-determined pul- 7. Murata K, Itoh H, Todo G, et al. Centrilobular lesions of the
monary artery diameters in predicting pulmonary hypertension. lung: demonstration by high-resolution CT and pathologic
Invest Radiol. 1984;19:16–22. correlation. Radiology. 1986;161:641–645.
2. Frazier AA, Galvin JR, Franks JJ, et al. From the archives of the 8. Horlander KT, Mannino DM, Leeper KP. Pulmonary embolism
AFIP: pulmonary vasculature—hypertension and infarction. mortality in the United States, 1979–1998: an analysis
Radiographics. 2000;20:491–524. using multiple-cause mortality data. Arch Intern Med.
3. Haimovici JB, Trotman-Dickenson B, Halpern EF, et al. 2003;163:1711– 1717.
Relationship between pulmonary artery diameter at computed 9. Morgenthaler TI, Ryu JH. Clinical characteristics of fatal
tomography and pulmonary artery pressures at right-sided heart pulmonary embolism in a referral hospital. Mayo Clin Proc.
catheterization. Acad Radiol. 1997;4:327–334. 1995;70:417.
4. Jefferson KE. The normal pulmonary angiogram and some 10. Moser KM. Fatal pulmonary embolism: old pitfalls, new chal-
changes seen in chronic nonspecific lung disease. I. The pul- lenges. Mayo Clin Proc. 1995;70:501–502.
monary vessels seen in the normal pulmonary angiogram. Proc 11. Pineda LA, Hathwar VS, Grant BJ. Clinical suspicion of fatal
R Soc Med. 1965;58:677. pulmonary embolism. Chest. 2001;120:791–795.
5. Elliot FM, Reid L. Some new facts about the pulmonary artery 12. Goldhaber SZ, Visani L, De RM. Acute pulmonary embolism:
and its branching pattern. Clin Radiol. 1965;16:193. clinical outcomes in the International Cooperative Pulmonary
Embolism Registry (ICOPER) [see comment]. Lancet. 1999; 36. Kelly J, Hunt BJ. A clinical probability assessment and
353:1386. D-dimer measurement should be the initial step in the investi-
13. Egermayer P. The mortality of untreated pulmonary embolism. gation of suspected venous thromboembolism. Chest. 2003;
[comment]. Chest. 1996;110:303. 124:1116.
14. Egermayer P, Town GI. The mortality of untreated pulmonary 37. van Belle A, Buller HR, Huisman MV, et al. Effectiveness of
embolism in patients with intermediate probability lung scans. managing suspected pulmonary embolism using an algorithm
Chest. 1998;114:1497. combining clinical probability, D-dimer testing, and computed
15. Rodger MA, et al. Diagnostic value of arterial blood gas meas- tomography [see comment]. JAMA. 2006;295:172–179.
urement in suspected pulmonary embolism. Am J Respir Crit 38. Coche E, Verschuren F, Hainaut P, et al. Pulmonary embolism
Care Med. 2000;162:2105–2108. findings on chest radiographs and multislice spiral CT. Eur Radiol.
16. Egermayer P. Normoxemia and pulmonary embolism [com- 2004;14:1241–1248.
ment]. Chest. 2001;120:1756. 39. Elliott CG, Goldhaber SZ, Visani L, et al. Chest radiographs in
17. Stein PD, Goldhaber SZ, Henry JW. Alveolar-arterial oxygen gra- acute pulmonary embolism. Results from the International
dient in the assessment of acute pulmonary embolism [see Cooperative Pulmonary Embolism Registry. Chest. 2000;118:
comment]. Chest. 1995;107:139–143. 33–38.
18. Stein PD, Goldhaber SZ, Henry JW, et al. Arterial blood gas 40. Coche EE, Muller NL, Kim KI, et al. Acute pulmonary embolism:
analysis in the assessment of suspected acute pulmonary ancillary findings at spiral CT. Radiology. 1998;207:753–758.
embolism [see comment]. Chest. 1996;109:78–81. 41. van Strijen MJ, de MW, Schiereck J, et al. Single-detector helical
19. Wells PS, Ginsberg JS, Anderson DR, et al. Use of a clinical computed tomography as the primary diagnostic test in suspec-
model for safe management of patients with suspected pul- ted pulmonary embolism: a multicenter clinical management
monary embolism. Ann Intern Med. 1998;129:997–1005. study of 510 patients. Ann Intern Med. 2003;138:307–314.
20. Anderson DR, Wells PS. Improvements in the diagnostic 42. Tsai KL, Gupta E, Haramati LB. Pulmonary atelectasis: a frequent
approach for patients with suspected deep vein thrombosis or alternative diagnosis in patients undergoing CT-PA for suspected
pulmonary embolism. Thromb Haemost. 1999;82:878–886. pulmonary embolism. Emerg Radiol. 2004;10:282–286.
21. Stein PD, Kayali F, Olson RE. Estimated case fatality rate of pul- 43. Anonymous. Value of the ventilation/perfusion scan in acute
monary embolism, 1979 to 1998. Am J Cardiol. 2004;93:1197. pulmonary embolism. Results of the Prospective Investigation
22. Frost SD, Brotman DJ, Michota FA. Rational use of D-dimer of Pulmonary Embolism Diagnosis (PIOPED). The PIOPED
measurement to exclude acute venous thromboembolic disease. Investigators. JAMA. 1990;263:2753–2759.
Mayo Clin Proc. 2003;78:1385. 44. Siegel A, Holtzman SR, Bettmann MA, et al. Clinicians’ percep-
22a. Stein PD, Hull RD, Patel KC, et al. D-dimer for the exclusion of tions of the value of ventilation-perfusion scans. Clin Nucl Med.
acute venous thrombosis and pulmonary embolism: a system- 2004;29:419–425.
atic review. Ann Intern Med. 2004;140:589–602. 45. Gottschalk A, Sostman HD, Coleman RE, et al. Ventilation-
23. Goldstein NM, Kollef MH, Ward S, et al. The impact of the perfusion scintigraphy in the PIOPED study. Part II. Evaluation
introduction of a rapid D-dimer assay on the diagnostic eval- of the scintigraphic criteria and interpretations. J Nucl Med.
uation of suspected pulmonary embolism [see comment]. Arch 1993;34:1119–1126.
Intern Med. 2001;161:567–571. 46. Sostman HD, Coleman RE, DeLong DM, et al. Evaluation of
24. Perrier A, MP de, Bounameaux H, Introduction of D-dimer revised criteria for ventilation-perfusion scintigraphy in patients
assay: does it increase resource utilization or correct identifica- with suspected pulmonary embolism. Radiology. 1994;193:
tion of pulmonary embolism? [comment]. Arch Intern Med. 103–107.
2001;161:2049. 47. van Rossum AB, Pattynama PM, Ton ER, et al. Pulmonary
25. Propp DA. Caveat emptor: the positive D-dimer slippery slope embolism: validation of spiral CT angiography in 149 patients.
[comment]. Ann Emerg Med. 2003;42:309. Radiology. 1996;201: 467–470.
26. Mavromatis BH, Kessler CM. D-dimer testing: the role of the 48. Blachere H, Latrabe V, Montaudon M, et al. Pulmonary
clinical laboratory in the diagnosis of pulmonary embolism. embolism revealed on helical CT angiography: comparison with
J Clin Pathol. 2001; 54:664–668. ventilation-perfusion radionuclide lung scanning. AJR Am J
27. Irwin GA, Luchs JS, Donovan V, et al. Can a state-of-the-art Roentgenol. 2000;174:1041–1047.
D-dimer test be used to determine the need for CT imaging in 49. Padley SP. Commentary. Lung scintigraphy vs spiral CT in the
patients suspected of having pulmonary embolism? Acad assessment of pulmonary emboli. Br J Radiol. 2002;75:5–8.
Radiol. 2002;9:1013. 50. Coche E, Verschuren F, Keyeux A, et al. Diagnosis of acute
28. Brotman DJ, Segal JB, Jani JT, et al. Limitations of D-dimer pulmonary embolism in outpatients: comparison of thin-
testing in unselected inpatients with suspected venous throm- collimation multi-detector row spiral CT and planar ventilation-
boembolism. Am J Med. 2003;114:276. perfusion scintigraphy. Radiology. 2003;229:757–765.
29. De Monye W, Sanson BJ, Mac Gillavry MR, et al. Embolus loca- 51. Breslaw BH, Dorfman GS. Ventilation/perfusion scanning for
tion affects the sensitivity of a rapid quantitative D-dimer assay prediction of pulmonary emboli location: correlation with
in the diagnosis of pulmonary embolism. Am J Respir Crit Care angiographic findings. Radiology. 1992;185:180–185.
Med. 2002;165:345–348. 52. Remy-Jardin M, Remy J, Deschildre F, et al. Diagnosis of pul-
30. Chabloz P, Reber G, Boehlen F, et al. TAFI antigen and D-dimer monary embolism with spiral CT: comparison with pulmonary
levels during normal pregnancy and at delivery. Br J Haematol. angiography and scintigraphy. Radiology. 1996;200:699–706.
115:150–152. 53. Ruiz Y, Caballero P, Caniego JL, et al. Prospective comparison of
31. Francalanci I, Comeglio P, Liotta AA, et al. D-dimer plasma helical CT with angiography in pulmonary embolism: global
levels during normal pregnancy measured by specific ELISA. and selective vascular territory analysis. Interobserver agree-
Int J Clin Lab Res. 1997;27:65–67. ment. Eur Radiol. 2003; 13:823–829.
32. Francalanci I, Comeglio P, Liotta AA, et al. D-dimer concentra- 54. Baile EM, King GG, Muller NL, et al. Spiral computed tomogra-
tions during normal pregnancy, as measured by ELISA. Thromb phy is comparable to angiography for the diagnosis of
Res. 1995;78:399–405. pulmonary embolism. Am J Respir Crit Care Med. 2000;
33. Abcarian PW, Sweet JD, Watabe JT, et al. Role of a quantitative 161(3 Pt 1):1010.
D-dimer assay in determining the need for CT angiography of 55. Zuckerman DA, Sterling KM, Oser RF. Safety of pulmonary
acute pulmonary embolism. AJR Am J Roentgenol. 2004;182:1377. angiography in the 1990s. J Vasc Intervent Radiol. 1996;7:199–205.
34. Burkill GJ, Bell JR, Chinn RJ, et al. The use of a D-dimer assay in 56. Diffin DC, Leyendecker JR, Johnson SP, et al. Effect of anatomic
patients undergoing CT pulmonary angiography for suspected distribution of pulmonary emboli on interobserver agreement
pulmonary embolus. Clin Radiol. 2002;57:41. in the interpretation of pulmonary angiography. AJR Am J
35. Dunn KL, Wolf JP, Dorfman DM, et al. Normal D-dimer levels Roentgenol. 1998;171:1085.
in emergency department patients suspected of acute pul- 57. Stein PD, Henry JW, Gottschalk A. Reassessment of pulmonary
monary embolism [see comment]. J Am Coll Cardiol 2002; angiography for the diagnosis of pulmonary embolism:
40:1475–1478. relation of interpreter agreement to the order of the involved
pulmonary arterial branch [see comment]. Radiology. 1999; 80. Stein PD, Fowler SE, Goodman LR, et al. Multidetector
210:689–691. computed tomography for acute pulmonary embolism. N Engl J
58. Safriel Y, Zinn H. CT pulmonary angiography in the detection Med. 2006;354:2317–2327.
of pulmonary emboli: a meta-analysis of sensitivities and 81. Johnson JC, Brown MD, Mc Cullough N, et al. CT lower extrem-
specificities. Clin Imaging. 2002;26(2):101. ity venography in suspected pulmonary embolism in the ED.
59. Goodman LR, Curtin JJ, Mewissen MW, et al. Detection of pul- Emerg Radiol. 2006; 12:160–163.
monary embolism in patients with unresolved clinical and 82. Kuzo RS, Goodman LR. CT evaluation of pulmonary embolism:
scintigraphic diagnosis: helical CT versus angiography. AJR Am J technique and interpretation. AJR Am J Roentgenol. 1997;
Roentgenol. 1995; 164:1369–1374. 169:959–965.
60. Rathbun SW, Raskob GE, Whitsett TL. Sensitivity and specificity 83. Remy-Jardin M, Remy J, Artaud D, et al. Spiral CT of pulmonary
of helical computed tomography in the diagnosis of pulmonary embolism: technical considerations and interpretive pitfalls. J
embolism: a systematic review [see comment]. Ann Intern Med. Thorac Imaging. 1997;12:103–117.
2000;132:227. 84. Marten K, Engelke C, Funke M, et al. ECG-gated multislice spiral
61. Patel S, Kazerooni EA, Cascade PN. Pulmonary embolism: CT for diagnosis of acute pulmonary embolism. Clin Radiol.
optimization of small pulmonary artery visualization at multi- 2003;58:862–868.
detector row CT. Radiology. 2003;227:455–460. 85. Loud PA, Katz DS, Klippenstein DL, et al. Combined CT venog-
62. Anonymous. Opinions regarding the diagnosis and manage- raphy and pulmonary angiography in suspected thromboem-
ment of venous thromboembolic disease. ACCP Consensus bolic disease: diagnostic accuracy for deep venous evaluation.
Committee on Pulmonary Embolism. Chest. 1998;113:499–504. AJR Am J Roentgenol. 2000; 174:61.
63. Tapson VF, Carroll BA, Davidson BL, et al. The diagnostic 86. Loud PA, Katz DS, Bruce DA, et al. Deep venous thrombosis
approach to acute venous thromboembolism. Clinical practice with suspected pulmonary embolism: detection with combined
guideline. American Thoracic Society. Am J Respir Crit Care Med. CT venography and pulmonary angiography. Radiology. 2001;
1999;160:1043–1066. 219:498–502.
64. Remy-Jardin M, Remy J, Artaud D, et al. Peripheral pulmonary 87. Bruce D, Loud PA, Klippenstein DL, et al. Combined CT ven-
arteries: optimization of the spiral CT acquisition protocol [see ography and pulmonary angiography: how much venous
comment]. Radiology. 1997;204:157–163. enhancement is routinely obtained? AJR Am J Roentgenol. 2001;
65. Remy-Jardin M, Baghaie F, Bonnel F, et al. Thoracic helical CT: 176:1281–1285.
influence of subsecond scan time and thin collimation on 88. Rademaker J, Griesshaber V, Hidajat N, et al. Combined CT
evaluation of peripheral pulmonary arteries. Eur Radiol. 2000; pulmonary angiography and venography for diagnosis of pul-
10:1297–1303. monary embolism and deep vein thrombosis: radiation dose. J
66. Raptopoulos V, Boiselle PM. Multi-detector row spiral CT pul- Thorac Imaging. 2001; 16:297–299.
monary angiography: comparison with single-detector row 89. Walsh G, Redmond S. Does addition of CT pelvic venography to
spiral CT. Radiology. 2001;221:606. CT pulmonary angiography protocols contribute to the diagno-
67. Schoepf UJ, Holzknecht N, Helmberger TK, et al. Subsegmental sis of thromboembolic disease? Clin Radiol. 2002;57:462–465.
pulmonary emboli: improved detection with thin-collimation 90. Remy-Jardin M, Bahepar J, Lafitte JJ, et al. Multi-detector row
multi-detector row spiral CT. Radiology. 2002;222:483–490. CT angiography of pulmonary circulation with Gadolinium-
68. Schoepf UJ, Costello P. Multidetector-row CT imaging of pul- based contrast agents: prospective evaluation in 60 patients.
monary embolism. Semin Roentgenol. 2003;38:106. Radiology. 2006;238:1022–1035.
69. Schoepf, UJ, Costello P. CT angiography for diagnosis of pul- 90a. Morcos S, Remy-Jardin M. Gadolinium-based contrast media
monary embolism: state of the art. Radiology. 2004;230:329. for multi-detector row spiral CT pulmonary angiography
70. Qanadli SD, Hajjam ME, Mesurolle B, et al. Pulmonary embolism in patients with renal insufficiency [comment]. Radiology.
detection: prospective evaluation of dual-section helical CT versus 2006;238:1077–1078.
selective pulmonary arteriography in 157 patients [see comment]. 91. Haage P, Schmitz-Rode T, Hubner D, et al. Reduction of contrast
Radiology. 2000; 217:447–455. material dose and artifacts by a saline flush using a double
71. Oser RF, Zuckerman DA, Gutierrez FR, et al. Anatomic dis- power injector in helical CT of the thorax. AJR Am J Roentgenol.
tribution of pulmonary emboli at pulmonary angiography: 2000;174:1049–1053.
implications for cross-sectional imaging [see comment]. 92. Irie T, Kajitani M, Yamaguchi M, et al. Contrast-enhanced CT
Radiology. 1996;199:31. with saline flush technique using two automated injectors: how
72. Gurney JW. No fooling around: direct visualization of pul- much contrast medium does it save? J Comput Assist Tomogr.
monary embolus. Radiology. 1993;188:618–619. 2002;26:287–291.
73. Goodman LR. Small pulmonary emboli: what do we know? 93. Schoellnast H, Tillich M, Deutschmann HA, et al. Abdominal
Radiology. 2005;234:654–658. multidetector row computed tomography: reduction of cost and
74. Stein PD, Henry JW, Relyea B. Untreated patients with pulmonary contrast material dose using saline flush. J Comput Assist Tomogr.
embolism. Outcome, clinical, and laboratory assessment [see 2003;27:847–853.
comment]. Chest. 1995;107:931–935. 94. Schoellnast H, Tillich M, Deutschmann HA, et al. Improvement
75. Swensen SJ, Sheedy PF, Ryu JH, et al. Outcomes after withholding of parenchymal and vascular enhancement using saline flush
anticoagulation from patients with suspected acute pulmonary and power injection for multiple-detector-row abdominal CT.
embolism and negative computed tomographic findings: a Eur Radiol. 2004;14: 659–664.
cohort study. Mayo Clin Proc. 2002;77:130–138. 95. Hartmann IJ, Lo RT, Bakker J, et al. Optimal scan delay in spiral
76. Garg K, Sieler H, Welsh CH, et al. Clinical validity of helical CT CT for the diagnosis of acute pulmonary embolism. J Comput
being interpreted as negative for pulmonary embolism: implica- Assist Tomogr. 2002;26:21.
tions for patient treatment [see comment]. AJR Am J Roentgenol. 96. Francois CJ, Shors SM, Bonow RO, et al. Analysis of cardiopul-
1999; 172:1627. monary transit times at contrast material-enhanced MR imaging
77. Tillie-Leblond I, Mastora I, Radenne F, et al. Risk of pulmonary in patients with heart disease. Radiology. 2003;227:447.
embolism after a negative spiral CT angiogram in patients with 97. Ghaye B, Szapiro D, Mastora I, et al. Peripheral pulmonary
pulmonary disease: 1-year clinical follow-up study. Radiology. arteries: how far in the lung does multi-detector row spiral CT
2002;223:461–467. allow analysis? Radiology. 2001;219:629–636.
78. Donato AA, Scheirer JJ, Atwell MS, et al. Clinical outcomes in 98. Gosselin MV, Rubin GD, Leung AN, et al. Unsuspected pul-
patients with suspected acute pulmonary embolism and nega- monary embolism: prospective detection on routine helical CT
tive helical computed tomographic results in whom anticoagu- scans. Radiology. 1998;208:209–215.
lation was withheld. Arch Intern Med. 2003;163:2033–2038. 99. Brink JA, Woodard PK, Horesh L, et al. Depiction of pulmonary
79. Remy-Jardin M, Tillie-Leblond I, Szapiro D, et al. CT angiogra- emboli with spiral CT: optimization of display window settings
phy of pulmonary embolism in patients with underlying respi- in a porcine model. Radiology. 1997;204:703.
ratory disease: impact of multislice CT on image quality and 100. Remy-Jardin M, Remy J, Cauvain O, et al. Diagnosis of central
negative predictive value. Eur Radiol. 2002;12:1971. pulmonary embolism with helical CT: role of two-dimensional
multiplanar reformations. AJR Am J Roentgenol. 1995;165: 124. Revel D, Loubeyre P, Delignette A, et al. Contrast-enhanced
1131–1138. magnetic resonance tomoangiography: a new imaging tech-
101. Fleischmann D, Scholten C, Klepetko W, et al. Three-dimensional nique for studying thoracic great vessels. Magn Reson Imaging.
visualization of pulmonary thromboemboli in chronic throm- 1993;11:1101–1105.
boembolic pulmonary hypertension with multiple detector-row 125. Rubin GD, Herfkens RJ, Pelc NJ, et al. Single breath-hold pul-
spiral computed tomography. Circulation. 2001;103:2993. monary magnetic resonance angiography. Optimization and
102. Coche E, Pawlak S, Dechambre S, et al. Peripheral pulmonary comparison of three imaging strategies. Invest Radiol.
arteries: identification at multi-slice spiral CT with 3D recon- 1994;29:766–792.
struction. Eur Radiol. 2003; 13:815. 126. Bergin CJ, Sirlin CB, Hauschildt JP, et al. Chronic thromboem-
103. Lawler LP, Fishman EK. Multi-detector row CT of thoracic disease bolism: diagnosis with helical CT and MR imaging with
with emphasis on 3D volume rendering and CT angiography. angiographic and surgical correlation. Radiology. 1997;204:
Radiographics. 2001;21:1257–1273. 695–702.
104. Jeong YJ, Lee KS, Yoon YC, et al. Evaluation of small pul- 127. Laissy JP, Assayag P, Henry-Feugeas MC, et al. Pulmonary time-
monary arteries by 16-slice multidetector computed tomogra- of-flight MR angiography at 1.0 T: comparison between 2D and
phy: optimum slab thickness in condensing transaxial images 3D tone acquisitions. Magn Reson Imaging. 1995;13:949–957.
converted into maximum intensity projection images. J 128. Prince MR, Narasimham DL, Stanley JC, et al. Breath-hold
Comput Assist Tomogr. 2004;28:195. gadolinium-enhanced MR angiography of the abdominal aorta
105. Simon M, Boiselle PM, Choi JR, et al. Paddle-wheel CT display and its major branches. Radiology. 1995;197:785–792.
of pulmonary arteries and other lung structures: a new imaging 129. Gefter WB, Hatabu H, Holland GA, et al. Pulmonary MR
approach. AJR Am J Roentgenol. 2001;177:195. angiography. Semin Ultrasound CT MR. 1996;17:316–323.
106. Simon M, Chiang EE, Boiselle PM. Paddle-wheel multislice hel- 130. Isoda H, Masui T, Hasegawa S, et al. Pulmonary MR angiogra-
ical CT display of pulmonary vessels and other lung structures. phy: a comparison of 2D and 3D time-of-flight. J Comput Assist
Radiol Clin North Am. 2003;41:617. Tomogr. 1994;18:402–407.
107. Chiang EE, Boiselle PM, Raptopoulos V, et al. Detection of 131. Leung DA, McKinnon GC, Davis CP, et al. Breath-hold, contrast-
pulmonary embolism: comparison of paddlewheel and coronal enhanced, three-dimensional MR angiography. Radiology. 1996;
CT reformations—initial experience. Radiology. 2003;228:577. 200:569–571.
108. Greer IA. Thrombosis in pregnancy: maternal and fetal issues 132. Li KC, Pelc LR, Napel SA, et al. MRI of pulmonary embolism
[see comment]. Lancet. 1999;353:1258–1265. using Gd-DTPA-polyethylene glycol polymer enhanced 3D fast
109. Winer-Muram HT, Boone JM, Brown HL, et al. Pulmonary gradient echo technique in a canine model. Magn Reson Imaging.
embolism in pregnant patients: fetal radiation dose with helical 1997;15: 543–550.
CT. Radiology. 2002; 224:487–492. 133. Meaney JF, Weg JG, Chenevert TL, et al. Diagnosis of pulmonary
110. Forsted DH, Kalbhen CL. CT of pregnant women for urinary embolism with magnetic resonance angiography. N Engl J Med.
tract calculi, pulmonary thromboembolism, and acute appen- 1997;336:1422–1427.
dicitis. AJR Am J Roentgenol. 2002;178:1285. 134. Steiner P, McKinnon GC, Romanowski B, et al. Contrast-
111. Schuster ME, Fishman JE, Copeland JF, et al. Pulmonary enhanced, ultrafast 3D pulmonary MR angiography in a single
embolism in pregnant patients: a survey of practices and policies breath-hold: initial assessment of imaging performance. J Magn
for CT pulmonary angiography. AJR Am J Roentgenol. 2003;181: Reson Imaging. 1997;7:177–182.
1495–1498. 135. Hatabu H, Gefter WB, Listerud J, et al. Pulmonary MR angiogra-
112. Posteraro RH, Sostman HD, Spritzer CE, et al. Cine-gradient- phy utilizing phased-array surface coils. J Comput Assist Tomogr.
refocused MR imaging of central pulmonary emboli. AJR Am 1992;16:410–417.
J Roentgenol. 1989;152:465–468. 136. Foo TK, MacFall JR, Hayes CE, et al. Pulmonary vasculature:
113. Erdman WA, Peshock RM, Redman HC, et al. Pulmonary single breath-hold MR imaging with phased-array coils.
embolism: comparison of MR images with radionuclide and Radiology. 1992;183:473–477.
angiographic studies. Radiology. 1994;190:499–508. 137. Oudkerk M, van Beek EJ, Wielopolski P, et al. Comparison of
114. Schiebler ML, Holland GA, Hatabu H, et al. Suspected pul- contrast-enhanced magnetic resonance angiography and con-
monary embolism: prospective evaluation with pulmonary MR ventional pulmonary angiography for the diagnosis of pul-
angiography. Radiology. 1993;189:125–131. monary embolism: a prospective study. Lancet. 2002;359:
115. Gefter WB, Hatabu H, Dinsmore BJ, et al. Pulmonary vascular 1643–1647.
cine MR imaging: a noninvasive approach to dynamic imaging 138. Gupta A, Frazer CK, Ferguson JM, et al. Acute pulmonary
of the pulmonary circulation. Radiology. 1990;176:761–770. embolism: diagnosis with MR angiography [see comment].
116. Grist TM, Sostman HD, MacFall JR, et al. Pulmonary angiogra- Radiology. 1999;210:353.
phy with MR imaging: preliminary clinical experience. 139. Sostman HD, Layish DT, Tapson VF, et al. Prospective compari-
Radiology. 1993;189:523–530. son of helical CT and MR imaging in clinically suspected acute
117. MacFall, JR, Sostman HD, Foo TK. Thick-section, single breath- pulmonary embolism. J Magn Reson Imaging. 1996;6:275–281.
hold magnetic resonance pulmonary angiography. Invest Radiol. 140. Haage P, Piroth W, Krombach G, et al. Pulmonary embolism:
1992;27:318–322. comparison of angiography with spiral computed tomography,
118. Hatabu H, Gefter WB, Kressel HY, et al. Pulmonary vasculature: magnetic resonance angiography, and real-time magnetic reso-
high-resolution MR imaging. Work in progress. Radiology. 1989; nance imaging. Am J Respir Crit Care Med. 2003;167:729–734.
171:391–395. 141. Hurst DR, Kazerooni EA, Stafford-Johnson D, et al. Diagnosis of
119. Wielopolski PA, Haacke EM, Adler LP. Three-dimensional MR pulmonary embolism: comparison of CT angiography and MR
imaging of the pulmonary vasculature: preliminary experience. angiography in canines. J Vasc Intervent Radiol. 1999;10:
Radiology. 1992;183:465–472. 309–318.
120. Wielopolski PA. Pulmonary arteriography. Magn Reson Imaging 142. Ohno Y, Kawamitsu H, Higashino T, et al. Time-resolved
Clin North Am. 1993;1:295–313. contrast-enhanced pulmonary MR angiography using sensitivity
121. Friedli JL, Paschal CB. Evaluation of variable-angle uniform encoding (SENSE). J Magn Reson Imaging. 2003;17:330–336.
signal excitation, tilted optimized nonsaturating excitation, and 143. Reittner P, Coxson HO, Nakano Y, et al. Pulmonary embolism:
flat radio-frequency pulses in free-breathing non-contrast- comparison of gadolinium-enhanced MR angiography with
enhanced pulmonary MR angiography. Radiology. 1997;202: contrast-enhanced spiral CT in a porcine model. Acad Radiol.
863–867. 2001;8:343–350.
122. Friedli JL, Paschal CB, Loyd JE, et al. Quantitative 3D VUSE pul- 144. Seo JB, Im JG, Goo JM, et al. Comparison of contrast-enhanced
monary MRA. Magn Reson Imaging. 1999;17:363–370. CT angiography and gadolinium-enhanced MR angiography in
123. Loubeyre P, Revel D, Douek P, et al. Dynamic contrast-enhanced the detection of subsegmental-sized pulmonary embolism. An
MR angiography of pulmonary embolism: comparison with pul- experimental study in a pig model. Acta Radiol. 2003;44:
monary angiography. AJR Am J Roentgenol. 1994;162:1035–1039. 403–410.
145. Goyen M, Laub G, Ladd ME, et al. Dynamic 3D MR angiogra- 169. Mai VM, Berr SS. MR perfusion imaging of pulmonary
phy of the pulmonary arteries in under four seconds. J Magn parenchyma using pulsed arterial spin labeling techniques:
Reson Imaging. 2001; 13:372–377. FAIRER and FAIR. J Magn Reson Imaging. 1999;9:483–487.
146. Schoenberg SO, Essig M, Bock M, et al. Comprehensive MR 170. Wildberger JE, Klotz E, Ditt H, et al. Multislice computed to-
evaluation of renovascular disease in five breath holds. J Magn mography perfusion imaging for visualization of acute pul-
Reson Imaging. 1999;10:347–356. monary embolism: animal experience. Eur Radiol. 2005;15:
147. Amundsen T, Kvaerness J, Jones RA, et al. Pulmonary embolism: 1378–1386.
detection with MR perfusion imaging of lung—a feasibility 171. Wildberger JE, Schoepf UJ, Mahnken AH, et al. Approaches to CT
study. Radiology. 1997;203:181–185. perfusion imaging in pulmonary embolism. Semin Roentgenol.
148. Berthezene Y, Croisille P, Bertocchi M, et al. Lung perfusion 2005;40(1):64–73.
demonstrated by contrast-enhanced dynamic magnetic reso- 172. Wildberger JE, Klotz E, Ditt H, et al. Multi-slice CT for visuali-
nance imaging. Application to unilateral lung transplantation. zation of acute pulmonary embolism: single breath-hold
Invest Radiol. 1997;32:351–356. subtraction technique. Rofo Fortschritte auf dem Gebiete der
149. Hennig J, Scheffler K, Laubenberger J, et al. Time-resolved Rontgenstrahlen und der Nuklearmedizin. 2005;177:17–23.
projection angiography after bolus injection of contrast agent. 173. Hatabu H, Tadamura E, Prasad PV, et al. Noninvasive pul-
Magn Reson Med. 1997;37:341–345. monary perfusion imaging by STAR-HASTE sequence. Magn
150. Korosec FR, Frayne R, Grist TM, et al. Time-resolved contrast- Reson Med. 2000;44:808–812.
enhanced 3D MR angiography. Magn Reson Med. 1996;36: 174. Iwasawa T, Saito K, Ogawa N, et al. Prediction of postoperative
345–351. pulmonary function using perfusion magnetic resonance imag-
151. Kluge A, Muller C, Hansel J, et al. Real-time MR with TrueFISP ing of the lung. J Magn Reson Imaging. 2002;15:685–692.
for the detection of acute pulmonary embolism: initial clinical 175. Hatabu H, Tadamura E, Levin DL, et al. Quantitative assessment
experience. Eur Radiol. 2004;14:709–718. of pulmonary perfusion with dynamic contrast-enhanced MRI.
152. Sodickson DK, Manning WJ. Simultaneous acquisition of spa- Magn Reson Med. 1999;42:1033–1038.
tial harmonics (SMASH): fast imaging with radiofrequency coil 176. Amundsen T, Torheim G, Kvistad KA, et al. Perfusion abnormal-
arrays. Magn Reson Med. 1997;38:591–603. ities in pulmonary embolism studied with perfusion MRI
153. Pruessmann KP, Weiger M, Scheidegger MB, et al. SENSE: sensi- and ventilation-perfusion scintigraphy: an intra-modality and
tivity encoding for fast MRI. Magn Reson Med. 1999;42:952–962. inter-modality agreement study. J Magn Reson Imaging. 2002;
154. Weiger M, Pruessmann KP, Kassner A, et al. Contrast-enhanced 15:386–394.
3D MRA using SENSE. J Magn Reson Imaging. 2000;12:671–677. 177. Suga K, Ogasawara N, Okada M, et al. Regional lung functional
155. van Beek EJ, Wild JM, Fink C, et al. MRI for the diagnosis of pul- impairment in acute airway obstruction and pulmonary
monary embolism. J Magn Reson Imaging. 2003;18:627–640. embolic dog models assessed with gadolinium-based aerosol
156. Ohno Y, Higashino T, Takenaka D, et al. MR angiography with ventilation and perfusion magnetic resonance imaging. Invest
sensitivity encoding (SENSE) for suspected pulmonary Radiol. 2002;37:281–291.
embolism: comparison with MDCT and ventilation-perfusion 178. Ogasawara N, Suga K, Kawakami Y, et al. Assessment of regional
scintigraphy. AJR Am J Roentgenol. 2004;183:91–98. lung function impairment in airway obstruction and pulmonary
157. Hany TF, Schmidt M, Hilfiker PR, et al. Optimization of contrast embolic dogs with combined noncontrast electrocardiogram-
dosage for gadolinium-enhanced 3D MRA of the pulmonary gated perfusion and gadolinium diethylenetriaminepentaacetic
and renal arteries. Magn Reson Imaging. 1998;16:901–906. acid aerosol magnetic resonance images. J Magn Reson Imaging.
158. Kauczor HU. Contrast-enhanced magnetic resonance angiogra- 2004;20:46–55.
phy of the pulmonary vasculature. A review. Invest Radiol. 1998; 179. Jalali A, Ishii M, Edvinsson JM, et al. Detection of simulated
33:606–617. pulmonary embolism in a porcine model using hyperpolarized
159. Fink C, Ley S, Puderbach M, et al. 3D pulmonary perfusion 3He MRI. Magn Reson Med. 2004;51:291–298.
MRI and MR angiography of pulmonary embolism in pigs 180. Altes TA, Mai VM, Munger TM, et al. Pulmonary embolism: com-
after a single injection of a blood pool MR contrast agent. Eur prehensive evaluation with MR ventilation and perfusion scan-
Radiol. 2004;14:1291–1296. ning with hyperpolarized helium-3, arterial spin tagging, and
160. Hoffmann U, Loewe C, Bernhard C, et al. MRA of the lower contrast-enhanced MRA. J Vasc Intervent Radiol. 2005;16:
extremities in patients with pulmonary embolism using a blood 999–1005.
pool contrast agent: initial experience. J Magn Reson Imaging. 181. Remy-Jardin M, Mastora I, Remy J. Pulmonary embolus imaging
2002;15:429–437. with multislice CT. Radiol Clin North Am. 2003;41:507.
161. Davis CP, Hany TF, Wildermuth S, et al. Postprocessing 182. Schoepf UJ, Goldhaber SZ, Costello P. Spiral computed tomogra-
techniques for gadolinium-enhanced three-dimensional MR phy for acute pulmonary embolism. Circulation. 2004;109:2160.
angiography. Radiographics. 1997;17:1061–1077. 183. Schoepf UJ, Kessler MA, Rieger CT, et al. Multislice CT imaging
162. Ladd ME, Gohde SC, Steiner P, et al. Virtual MR angioscopy of of pulmonary embolism. Eur Radiol. 2001;11:2278.
the pulmonary artery tree. J Comput Assist Tomogr. 1996;20: 184. Romero CS, Hernandez BL, Soler MJ, et al. Biochemical and
782–785. cytologic characteristics of pleural effusions secondary to pul-
163. Moody AR. Magnetic resonance direct thrombus imaging. J monary embolism. Chest. 2002;121:465–469.
Thromb Haemost. 2003;1:1403–1409. 185. Erkan L, Fyndyk S, Uzun O, et al. A new radiologic appearance
164. Kelly J, Hunt BJ, Moody A. Magnetic resonance direct thrombus of pulmonary thromboembolism: multiloculated pleural
imaging: a novel technique for imaging venous thromboem- effusions. Chest. 2004;126:298–302.
boli. Thromb Haemost. 2003;89:773–782. 186. Kasper W, Konstantinides S, Geibel A, et al. Prognostic signifi-
165. Fraser DG, Moody AR, Morgan PS, et al. Diagnosis of lower- cance of right ventricular afterload stress detected by echocar-
limb deep venous thrombosis: a prospective blinded study of diography in patients with clinically suspected pulmonary
magnetic resonance direct thrombus imaging. Ann Intern Med. embolism. Heart (Brit Cardiac Soc). 1997;77:346.
2002;136:89–98. 187. Lualdi JC, Goldhaber SZ. Right ventricular dysfunction after
166. Moody AR, Pollock JG, O’Connor AR, et al. Lower-limb deep acute pulmonary embolism: pathophysiologic factors, detec-
venous thrombosis: direct MR imaging of the thrombus. tion, and therapeutic implications. Am Heart J. 1995;130:1276.
Radiology. 1998;209:349–355. 188. Konstantinides S, Geibel A, Heusel G, et al. Heparin plus
167. Moody AR, Liddicoat A, Krarup K. Magnetic resonance pul- alteplase compared with heparin alone in patients with sub-
monary angiography and direct imaging of embolus for the det- massive pulmonary embolism [see comment]. N Engl J Med.
ection of pulmonary emboli. Invest Radiol. 1997;32:431–440. 2002;347:1143.
168. Keilholz SD, Mai VM, Berr SS, et al. Comparison of first-pass 189. Grifoni S, Olivotto I, Cecchini P, et al. Short-term clinical out-
Gd-DOTA and FAIRER MR perfusion imaging in a rabbit model come of patients with acute pulmonary embolism, normal
of pulmonary embolism. J Magn Reson Imaging. 2002;16: blood pressure, and echocardiographic right ventricular
168–171. dysfunction. Circulation. 2000;101:2817.
190. Wood KE. Major pulmonary embolism: review of a pathophys- 216. Aklog L, Williams CS, Byrne JG, et al. Acute pulmonary embolec-
iologic approach to the golden hour of hemodynamically tomy: a contemporary approach. Circulation. 2002;105:1416.
significant pulmonary embolism [see comment]. Chest. 2002; 217. Yalamanchili K, Fleisher AG, Lehrman SG, et al. Open pulmonary
121:877. embolectomy for treatment of major pulmonary embolism. Ann
191. Konstantinides S. The case for thrombolysis in acute major Thorac Surg. 2004;77:819.
pulmonary embolism: hemodynamic benefits and beyond 218. Contractor S, Maldjian PD, Sharma VK, et al. Role of helical CT
[comment]. Intensive Care Med. 2002;28:1547. in detecting right ventricular dysfunction secondary to acute
192. Konstantinides S. Should thrombolytic therapy be used in pulmonary embolism. J Comput Assist Tomogr. 2002;26:587.
patients with pulmonary embolism? Am J Cardiovasc Drugs. 219. Oliver TB, Reid JH, Murchison JT. Interventricular septal shift
2004;4(2):69. due to massive pulmonary embolism shown by CT pulmonary
193. Konstantinides S, Tiede N, Geibel A, et al. Comparison of angiography: an old sign revisited. Thorax. 1998,53:1092–1094.
alteplase versus heparin for resolution of major pulmonary 220. Reid JH, Murchison JT. Acute right ventricular dilatation: a new
embolism. Am J Cardiol. 1998;82:966–970 helical CT sign of massive pulmonary embolism [see com-
194. Goldhaber SZ, Haire WD, Feldstein ML, et al. Alteplase versus ment]. Clin Radiol. 1998;53:694–698.
heparin in acute pulmonary embolism: randomised trial assess- 221. Collomb D, Paramelle PJ, Calaque O, et al. Severity assessment
ing right-ventricular function and pulmonary perfusion. Lancet. of acute pulmonary embolism: evaluation using helical CT. Eur
1993;341:507–511. Radiol. 2003;13:1508.
195. Goldhaber SZ, Bounameaux H. Thrombolytic therapy in pul- 222. Quiroz R, Kucher N, Schoepf UJ, et al. Right ventricular enlarge-
monary embolism. Semin Vasc Med. 2001;1:213. ment on chest computed tomography: prognostic role in acute
196. Goldhaber SZ. Thrombolysis in pulmonary embolism: a large- pulmonary embolism [see comment]. Circulation. 2004;109:
scale clinical trial is overdue. Circulation. 2001;104:2876. 2401.
197. Goldhaber SZ. Thrombolysis for pulmonary embolism [see 223. Araoz PA, Gotway MB, Trowbridge RL, et al. Helical CT pul-
comment]. N Engl J Med. 2002;347:1131. monary angiography predictors of in-hospital morbidity and
198. Goldhaber SZ. Thrombolysis in pulmonary embolism: a large- mortality in patients with acute pulmonary embolism. J Thorac
scale clinical trial is overdue. Circulation. 2001;104:2876–2878. Imaging. 2003;18:207.
199. Meneveau N, Ming LP, Seronde MF, et al. In-hospital and long- 224. Mastora I, Remy-Jardin M, Masson P, et al. Severity of acute pul-
term outcome after sub-massive and massive pulmonary monary embolism: evaluation of a new spiral CT angiographic
embolism submitted to thrombolytic therapy. Eur Heart J. 2003; score in correlation with echocardiographic data. Eur Radiol.
24:1447. 2003;13:29.
200. Ajani AE, Waksman R. Thrombolytic therapy in patients with 225. Groves AM, Win T, Charman SC, et al. Semi-quantitative assess-
submassive pulmonary embolism [comment]. N Engl J Med. ment of tricuspid regurgitation on contrast-enhanced multide-
2003;348:357. tector CT [see comment]. Clin Radiol. 2004;59:715.
201. Ashton RW, Daniels CE, Ryu JH. Thrombolytic therapy in 226. Torbicki A, Galie N, Covezzoli A, et al. Right heart thrombi in
patients with submassive pulmonary embolism [comment]. pulmonary embolism: results from the International Cooperative
N Engl J Med. 2003;348:357. Pulmonary Embolism Registry. J Am Coll Cardiol. 2003;41:2245.
202. Berlin DA. Thrombolytic therapy in patients with submassive 227. Chartier L, Bera J, Delomez M, et al. Free-floating thrombi in
pulmonary embolism [comment]. N Engl J Med. 2003;348:357. the right heart: diagnosis, management, and prognostic indexes
203. Dalen JE. Thrombolysis in submassive pulmonary embolism? in 38 consecutive patients. Circulation. 1999;99:2779.
No [comment]. J Thromb Haemost. 2003;1:1130. 228. Rose PS, Punjabi NM, Pearse DB. Treatment of right heart
204. Dalen JE. Thrombolytic therapy in patients with submassive thromboemboli. Chest. 2002;121:806–814.
pulmonary embolism [comment]. N Engl J Med 2003;348: 229. Procopiou M, Perrier A. Treatment of right heart thromboem-
357. boli with IV recombinant tissue-type plasminogen activator
205. Konstantinides S. Thrombolysis in submassive pulmonary [comment]. Chest. 2000;117:920–921.
embolism? Yes [see comment]. J Thromb Haemost. 2003;1:1127. 230. Greco F, Bisignani G, Serafini O, et al. Successful treatment of
206. Mascitelli L, Pezzetta F. Thrombolysis in the treatment of sub- right heart thromboemboli with IV recombinant tissue-type
massive pulmonary embolism [comment]. Arch Intern Med. plasminogen activator during continuous echocardiographic
2003;163:1238. monitoring: a case series report [see comment]. Chest. 1999;116:
207. Mofredj A, Baraka D, Habki R. The approach to massive pul- 78–82.
monary embolism [comment]. Intensive Care Med. 2002; 28:995. 231. Qanadli SD, El HM, Vieillard-Baron A, et al. New CT index to
208. Miller GA, Sutton GC, Kerr IH, et al. Comparison of streptoki- quantify arterial obstruction in pulmonary embolism: compar-
nase and heparin in treatment of isolated acute massive ison with angiographic index and echocardiography. AJR Am J
pulmonary embolism. Br Med J. 1971;2:681–684. Roentgenol. 2001; 176:1415.
209. Dalen JE, Alpert JS, Hirsch J. Thrombolytic therapy for pul- 232. Wu AS, Pezzullo JA, Cronan JJ, et al. CT pulmonary angiogra-
monary embolism: is it effective? Is it safe? When is it indicated? phy: quantification of pulmonary embolus as a predictor of pa-
[see comment]. Arch Intern Med. 1997;157:2550. tient outcome—initial experience. Radiology. 2004;230:831.
210. Kanter DS, Mikkola KM, Patel SR, et al. Thrombolytic therapy 233. Kucher N, Goldhaber SZ. Cardiac biomarkers for risk stratifica-
for pulmonary embolism. Frequency of intracranial hemor- tion of patients with acute pulmonary embolism [see com-
rhage and associated risk factors [comment]. Chest. ment]. Circulation. 2003;108:2191.
1997;111:1241–1245. 234. Giannitsis E, Muller-Bardorff M, Kurowski V, et al. Independent
211. Schmitz-Rode T, Janssens U, Duda SH, et al. Massive pulmonary prognostic value of cardiac troponin T in patients with con-
embolism: percutaneous emergency treatment by pigtail rota- firmed pulmonary embolism. Circulation. 2000;102:211.
tion catheter. J Am Coll Cardiol. 2000;36:375. 235. Pruszczyk P, Bochowicz A, Torbicki A, et al. Cardiac troponin T
212. Schmitz-Rode T, Janssens U, Schild HH, et al. Fragmentation of monitoring identifies high-risk group of normotensive patients
massive pulmonary embolism using a pigtail rotation catheter with acute pulmonary embolism [see comment]. Chest.
[see comment]. Chest. 1998;114:1427. 2003;123:1947.
213. Schmitz-Rode T, Kilbinger M, Gunther RW. Simulated flow pat- 236. Nunes JP, Macedo F. An analytical triad for the diagnosis of pul-
tern in massive pulmonary embolism: significance for selective monary embolism. Cardiology. 2000;94:264.
intrapulmonary thrombolysis. Cardiovasc Intervent Radiol. 1998; 237. Konstantinides S, Geibel A, Olschewski M, et al. Importance of
21:199. cardiac troponins I and T in risk stratification of patients with
214. Reekers JA, Baarslag HJ, Koolen MG, et al. Mechanical thrombec- acute pulmonary embolism. Circulation. 2002;106:1263.
tomy for early treatment of massive pulmonary embolism. 238. Kucher N, Goldhaber SZ. Cardiac biomarkers for risk stratifica-
Cardiovasc Intervent Radiol. 2003;26:246. tion of patients with acute pulmonary embolism. Circulation.
215. De Gregorio MA, Gimeno MJ, Mainar A, et al. Mechanical and 2003;108:2191.
enzymatic thrombolysis for massive pulmonary embolism. J Vasc 239. Kucher N, Printzen G, Doernhoefer T, et al. Low pro-brain natri-
Intervent Radiol. 2002;13(2 Pt 1):163. uretic peptide levels predict benign clinical outcome in acute
pulmonary embolism. [see comment]. Circulation. 2003;107: 263. King MA, Ysrael M, Bergin CJ. Chronic thromboembolic pul-
1576–1578. monary hypertension: CT findings. AJR Am J Roentgenol. 1998;
240. Pruszczyk P, Bochowicz A, Kostrubiec M, et al. Myoglobin strati- 170:955–960.
fies short-term risk in acute major pulmonary embolism. Clin 264. Bergin CJ, Hauschildt JP, Brown MA, et al. Identifying the cause
Chim Acta. 2003; 338(1–2):53. of unilateral hypoperfusion in patients suspected to have
241. Steiner P, Lund GK, Debatin JF, et al. Acute pulmonary chronic pulmonary thromboembolism: diagnostic accuracy of
embolism: value of transthoracic and transesophageal echocar- helical CT and conventional angiography. Radiology. 1999;213:
diography in comparison with helical CT. AJR Am J Roentgenol. 743–749.
1996;167:931–936. 265. Strunk HM. Chronic thromboembolism. Radiology. 1998;207:551.
242. Gosselin MV, Rassner UA, Thieszen SL, et al. Contrast dynamics 266. Wittram C, Maher MM, Halpern EF, et al. Attenuation of acute
during CT pulmonary angiogram: analysis of an inspiration and chronic pulmonary emboli. Radiology. 2005;235:1050–1054.
associated artifact. J Thorac Imaging. 2004;19:1–7. 267. Hasegawa I, Boiselle PM, Hatabu H. Bronchial artery dilatation
243. Henk CB, Grampp S, Linnau KF, et al. Suspected pulmonary on MDCT scans of patients with acute pulmonary embolism:
embolism: enhancement of pulmonary arteries at deep-inspira- comparison with chronic or recurrent pulmonary embolism.
tion CT angiography—influence of patent foramen ovale and AJR Am J Roentgenol. 2004;182:67–72.
atrial-septal defect. Radiology. 2003;226:749. 268. Remy-Jardin M, Duhamel A, Deken V, et al. Systemic collateral
244. Dennie CJ, Veinot JP, McCormack DG, et al. Intimal sarcoma of supply in patients with chronic thromboembolic and primary
the pulmonary arteries seen as a mosaic pattern of lung attenua- pulmonary hypertension: assessment with multi-detector row
tion on high-resolution CT. AJR Am J Roentgenol. 2002;178: helical CT angiography. Radiology. 2005;235:274–281.
1208. 269. Endrys J, Hayat N, Cherian G. Comparison of bronchopul-
245. Roberts KE, Hamele-Bena D, Saqi A, et al. Pulmonary tumor monary collaterals and collateral blood flow in patients with
embolism: a review of the literature. Am J Med. 2003;115: chronic thromboembolic and primary pulmonary hyperten-
228–232. sion. Heart. 1997;78:171–176.
246. Yi CA, Lee KS, Choe YH, et al. Computed tomography in 270. Ley S, Kreitner KF, Morgenstern I, et al. Bronchopulmonary
pulmonary artery sarcoma: distinguishing features from pul- shunts in patients with chronic thromboembolic pulmonary
monary embolic disease. J Comput Assist Tomogr. 2004;28:34. hypertension: evaluation with helical CT and MR imaging. AJR
247. Shepard JA, Moore EH, Templeton PA, et al. Pulmonary intravas- Am J Roentgenol. 2002;179:1209–1215.
cular tumor emboli: dilated and beaded peripheral pulmonary 271. King MA, Bergin CJ, Yeung DW, et al. Chronic pulmonary
arteries at CT [see comment]. Radiology. 1993;187:797–801. thromboembolism: detection of regional hypoperfusion with
248. Kauczor HU, Schwickert HC, Mayer E, et al. Pulmonary artery CT. Radiology. 1994;191:359–363.
sarcoma mimicking chronic thromboembolic disease: com- 272. Sherrick AD, Swensen SJ, Hartman TE. Mosaic pattern of lung
puted tomography and magnetic resonance imaging findings. attenuation on CT scans: frequency among patients with
Cardiovasc Intervent Radiol. 1994;17:185–189. pulmonary artery hypertension of different causes. AJR Am J
249. Kim JH, Gutierrez FR, Lee EY, et al. Primary leiomyosarcoma of Roentgenol. 1997;169:79–82.
the pulmonary artery: a diagnostic dilemma. Clin Imaging. 273. Bergin CJ, Rios G, King MA, et al. Accuracy of high-resolution
2003;27:206. CT in identifying chronic pulmonary thromboembolic disease.
250. Yamada N, Kamei S, Yasuda F, et al. Primary leiomyosarcoma of AJR Am J Roentgenol. 1996;166:1371–1377.
the pulmonary artery confirmed by catheter suction biopsy. Chest. 274. Schwickert HC, Schweden F, Schild HH, et al. Pulmonary arter-
1998;113:555. ies and lung parenchyma in chronic pulmonary embolism:
251. Winston CB, Wechsler RJ, Salazar AM, et al. Incidental pul- preoperative and postoperative CT findings. Radiology. 1994;
monary emboli detected at helical CT: effect on patient care. 191:351–357.
Radiology. 1996;201:23. 275. Im JG, Choi YW, Kim HD, et al. Thin-section CT findings of the
252. Bick RL. Cancer-associated thrombosis [see comment]. N Engl J lungs: experimentally induced bronchial and pulmonary artery
Med. 2003;349:109. obstruction in pigs. AJR Am J Roentgenol. 1996;167:631.
253. Kanne JP, Gotway MB, Thoongsuwan N, et al. Six cases of acute 276. Kim TK, Im JG, Kim SH, et al. Experimentally induced pul-
central pulmonary embolism revealed on unenhanced multi- monary arterial occlusion with detachable balloon in pigs: thin-
detector CT of the chest. AJR Am J Roentgenol. 2003;180: section CT findings. Acad Radiol. 1998;5:822.
1661–1664. 277. Worthy SA, Muller NL, Hartman TE, et al. Mosaic attenuation
254. Kanne JP, Thoongsuwan N, Stern EJ. Detection of central pul- pattern on thin-section CT scans of the lung: differentiation
monary embolism on computed tomography densitometry among infiltrative lung, airway, and vascular diseases as a cause.
images before computed tomography pulmonary angiography. Radiology. 1997;205:465.
J Comput Assist Tomogr. 2003;27:907. 278. Arakawa H, Kurihara Y, Sasaka K, et al. Air trapping on CT of pa-
255. Cobelli R, Zompatori M, De Luca G, et al. Clinical usefulness of tients with pulmonary embolism. AJR Am J Roentgenol.
computed tomography study without contrast injection in the 2002;178:1201–1207.
evaluation of acute pulmonary embolism. J Comput Assist 279. Arakawa H, Stern EJ, Nakamoto T, et al. Chronic pulmonary
Tomogr. 2005;29:6–12. thromboembolism. Air trapping on computed tomography and
256. Cobelli R, Zompatori M, Sverzellati N, et al. Attenuation of correlation with pulmonary function tests. J Comput Assist
acute and chronic clots. Radiology. 2006;238:376–377; author Tomogr. 2003;27:735.
reply 377–378. 280. Remy-Jardin M, Remy J, Louvegny S, et al. Airway changes in
257. Lang IM. Chronic thromboembolic pulmonary hypertension— chronic pulmonary embolism: CT findings in 33 patients.
not so rare after all. N Engl J Med. 2004;350:2236–2238. Radiology. 1997;203: 355–360.
258. Pengo V, Lensing AW, Prins MH, et al. Incidence of chronic 281. Morris TA, Auger WR, Ysrael MZ, et al. Parenchymal scarring is
thromboembolic pulmonary hypertension after pulmonary em- associated with restrictive spirometric defects in patients with
bolism. N Engl J Med. 2004;350:2257–2264. chronic thromboembolic pulmonary hypertension. Chest.
259. Remy-Jardin M, Louvegny S, Remy J, et al. Acute central throm- 1996;110:399–403.
boembolic disease: posttherapeutic follow-up with spiral CT 282. Archibald CJ, Auger WR, Fedullo PF, et al. Long-term outcome
angiography. Radiology. 1997;203:173. after pulmonary thromboendarterectomy. Am J Respir Crit Care
260. Fedullo PF, Auger WR, Kerr KM, et al. Chronic thromboembolic Med. 1999;160:523.
pulmonary hypertension. N Engl J Med. 2001;345:1465–1472. 283. Jamieson SW. Pulmonary thromboendarterectomy. Heart (Brit
261. Han D, Lee KS, Franquet T, et al. Thrombotic and nonthrom- Cardiac Soc). 1998;79:118.
botic pulmonary arterial embolism: spectrum of imaging 284. Jamieson SW, Kapelanski DP, Sakakibara N, et al. Pulmonary
findings. Radiographics. 2003;23:1521. endarterectomy: experience and lessons learned in 1,500 cases.
262. Kereiakes DJ, Herfkens RJ, Brundage BH, et al. Computerized Ann Thorac Surg. 2003;76:1457.
tomography in chronic thromboembolic pulmonary hyperten- 285. Menzel T, Wagner S, Kramm T, et al. Pathophysiology of im-
sion. Am Heart J. 1983;106:1432–1436. paired right and left ventricular function in chronic embolic
pulmonary hypertension: changes after pulmonary thromboend- 308. Moncada R, Demos TC, Churchill R, et al. Case report. Chronic
arterectomy. Chest. 2000;118:897–903. stridor in a child: CT diagnosis of pulmonary vascular sling. J
286. Bergin CJ, Sirlin C, Deutsch R, et al. Predictors of patient Comput Assist Tomogr. 1983;7:713–715.
response to pulmonary thromboendarterectomy. AJR Am J 309. Berden WE, Baker DH, Wung JT. Complete cartilage-ring tra-
Roentgenol. 2000;174:509. cheal stenosis associated with anomalous left pulmonary artery:
287. Heinrich M, Uder M, Tscholl D, et al. CT scan findings in the ring-sling complex. Radiology. 1984;152:57–64.
chronic thromboembolic pulmonary hypertension: predictors 310. Hernandez RJ, Bank ER, Shaffer EM, et al. Comparative evalua-
of hemodynamic improvement after pulmonary thromboend- tion of the pulmonary arteries in patients with right ventricular
arterectomy. Chest. 2005;127:1606–1613. outflow tract obstructive lesions. AJR Am J Roentgenol. 1987;148:
288. Oikonomou A, Dennie CJ, Muller NL, et al. Chronic throm- 1189–1194.
boembolic pulmonary arterial hypertension: correlation of 311. Jacobstein MD, Fletcher BD, Nelson AD, et al. Magnetic reso-
postoperative results of thromboendarterectomy with preopera- nance imaging: evaluation of palliative systemic-pulmonary
tive helical contrast-enhanced computed tomography. J Thorac artery shunts. Circulation. 1984;70:650–656.
Imaging. 2004;19:67. 312. Park HS, Im JG, Jung GW, et al. Anomalous left pulmonary
289. Kruger S, Haage P, Hoffmann R, et al. Diagnosis of pulmonary artery with complete cartilaginous ring. J Comput Assist Tomogr.
arterial hypertension and pulmonary embolism with magnetic 1997;21:478–480.
resonance angiography. Chest. 2001;120:1556–1561. 313. Sotomora RF, Edwards JE. Anatomic identification of so-called
290. Kreitner KF, Ley S, Kauczor HU, et al. Chronic thromboembolic absent pulmonary artery. Circulation. 1978;57:624–633.
pulmonary hypertension: pre- and postoperative assessment 314. Bartter T, Irwin RS, Nash G. Review. Aneurysms of the pul-
with breath-hold MR imaging techniques. Radiology. 2004;232: monary arteries. Chest. 1988;94:1065–1079.
535–543. 315. Butto F, Lucas RV, Edwards JE. Pulmonary arterial aneurysm.
291. Ley S, Kauczor HU, Heussel CP, et al. Value of contrast- A pathologic study of five cases. Chest. 1987;91:237–241.
enhanced MR angiography and helical CT angiography in 316. SanDretto MA, Scanlon GT. Multiple mycotic pulmonary artery
chronic thromboembolic pulmonary hypertension. Eur Radiol. aneurysms secondary to intravenous drug use. AJR Am J
2003;13:2365. Roentgenol. 1984;143:89–90.
292. Kreitner KF, Mayer E, Voigtlaender T, et al. Three-dimensional 317. Koc Y, Gullu I, Akpek G, et al. Vascular involvement in Behçet’s
contrast-enhanced magnetic resonance angiography in a patient disease. J Rheumatol. 1992;19:402–410.
with chronic thromboembolic pulmonary hypertension before 318. Tunaci A, Berkmen YM, Gokmen E. Thoracic involvement in
and after thromboendarterectomy. Circulation. 1999;99:1101. Behçet’s disease: pathologic, clinical and imaging features. AJR
293. Prince MR, Alderson PO, Sostman HD. Chronic pulmonary Am J Roentgenol. 1995;164:51–56.
embolism: combining MR angiography with functional assess- 319. Numan F, Islak C, Berkmen T, et al. Behçet disease: pul-
ment. Radiology. 2004;232:325–326. monary arterial involvement in 15 cases. Radiology. 1994;192:
294. Thistlethwaite PA, Jamieson SW. Tricuspid valvular disease in 465–468.
the patient with chronic pulmonary thromboembolic disease. 320. Berkmen T. MR angiography of aneurysms in Behçet disease:
Curr Opin Cardiol. 2003;18:111. a report of four cases. J Comput Assist Tomogr. 1998;22:202–206.
295. Kieffer SA, Amplatz K, Anderson RC, et al. Proximal interrup- 321. Erkan F, Gul A, Tasali E. Pulmonary manifestations of Behçet’s
tion of a pulmonary artery. AJR Am J Roentgenol. 1965;95: disease. Thorax. 2001;56:572–578.
592–597. 322. Ferretti GR, Thony F, Link RM, et al. False aneurysm of the pul-
296. Bouros D, Pare P, Panagou P, et al. The varied manifestations of monary artery induced by a Swan-Ganz catheter: clinical presen-
pulmonary artery agenesis in adulthood. Chest. 1995; tation and radiologic management. AJR Am J Roentgenol.
108:670–676. 1996;167:941–945.
297. Rene M, Sans T, Dominguez J, et al. Unilateral pulmonary artery 323. Shin MS, Ceballus R, Bini RM, et al. Case report. CT diagnosis of
agenesis presenting with hemoptysis: treatment by emboliza- false aneurysm of the pulmonary artery not demonstrated by
tion of systemic collaterals. Cardiovasc Intervent Radiol. angiography. J Computed Assist Tomogr. 1983;7:524–526.
1995;18:251–254. 324. Daykin EL, Irwin GAL, Harrison DA. Case report. CT demon-
298. Naidich DP, Rumancik WM, Ettenger NA, et al. Congenital stration of a traumatic aneurysm of the pulmonary artery.
anomalies of the lungs in adults: MR diagnosis. AJR Am J J Computed Assist Tomogr. 1986;10:323–324.
Roentgenol. 1988;151:13–19. 325. Jeang MK, Adyanthaya A, Kuo L, et al. Multiple pulmonary ar-
299. Ryu DS, Spirn PW, Trotman-Dickenson BEA. HRCT findings of tery aneurysms: new use for magnetic resonance imaging. Am J
proximal interruption of the right pulmonary artery. J Thorac Med. 1986;81:1001–1004.
Imaging. 2004;19:171–175. 326. Cox JE, Chiles C, Aquino SL, et al. Pulmonary artery sarcomas: a
300. Webb WR. Congenital bronchopulmonary lesions. In Thoracic review of clinical and radiologic features. J Comput Assist Tomogr.
imaging. Philadelphia: Lippincott Williams & Wilkins; 2005:1–29. 1997;21:750–755.
301. Dikensoy O, Kervancioglu R, Bayram NG, et al. Horseshoe lung 327. Weinreb JC, Davis SD, Berkmen YM, et al. Pulmonary artery sar-
associated with scimitar syndrome and pleural lipoma. J Thorac coma: evaluation using Gd-DTPA. J Comput Assist Tomogr.
Imaging. 2006;21:73–75. 1990;14:647–649.
302. Dillon EH, Camputaro C. Partial anomalous pulmonary venous 328. Winchester PA, Kilnani NM, Trost DW, et al. Technical note.
drainage of the left upper lobe vs duplication of the superior Endovascular catheter biopsy of a pulmonary artery sarcoma.
vena cava: distinction based on CT findings. AJR Am J AJR Am J Roentgenol. 1996;167:657–659.
Roentgenol. 1993;160:375–379. 329. Dore R, Alerci M, D’Andrea F, et al. Intracardiac extension of
303. White CS, Baffa JM, Haney PJ, et al. Anomalies of pulmonary lung cancer via pulmonary veins: CT diagnosis. J Computed
veins: usefulness of spin-echo and gradient-echo MR images. Assisted Tomogr. 1988;12:565–568.
AJR Am J Roentgenol. 1998;170:1365–1368. 330. Shields JJ, Cho KJ, Geisinger KR. Pulmonary artery constriction
304. White CS, Baffa JM, Haney PJ, et al. Congenital anomalies of the by mediastinal lymphoma simulating pulmonary embolus. AJR
thoracic veins: MR findings. Radiographics. 1997;17:595–608. Am J Roentgenol. 1980;135:147–150.
305. Masui T, Seelos KC, Kersting-Sommerhoff BA, et al. 331. Marshall ME, Trump DL. Acquired extrinsic pulmonic stenosis
Abnormalities of the pulmonary veins: evaluation with MR caused by mediastinal tumors. Cancer. 1982;49:1496–1499.
imaging and comparison with cardiac angiography and 332. Hamper UM, Fishman EK, Khouri NF, et al. Typical and atypical
echocardiography. Radiology. 1991;181:645–649. CT manifestations of pulmonary sarcoidosis. J Comput Assist
306. Choe YH, Lee HJ, Kim HS, et al. MRI of total anomalous pul- Tomogr. 1986;10:928–936.
monary venous connections. J Comput Assist Tomogr. 1994;18: 333. Tack D, Nollevaux MC, Gevenois PA. Tree-in-bud pattern in neo-
243–249. plastic pulmonary emboli. AJR Am J Roentgenol. 2001;176:1421.
307. Stone DN, Bein ME, Garris JB. Anomalous left pulmonary 334. Weinstein JB, Aronberg DJ, Sagel SS. CT of fibrosing medias-
artery: two new adult cases. AJR Am J Roentgenol. 1980;135: tinitis: findings and their utility. AJR Am J Roentgenol. 1983;
1259–1263. 141:247–251.
335. Berry DE, Buccigrossi D, Peabody J, et al. Review. Pulmonary 337. Sherrick AD, Brown LL, Harms GF, et al. The radiographic find-
vascular occlusion and fibrosing mediastinitis. Chest. 1986;89: ings of fibrosing mediastinitis. Chest. 1994;106:484–489.
296–301. 338. Cramer M, Foley WD, Palmer TE. Compression of the right pul-
336. Farmer DW, Moore E, Amparo E, et al. Calcific fibrosing medi- monary artery by aortic aneurysms: CT demonstration. J Comput
astinitis: demonstration of pulmonary vascular obstruction by Assist Tomogr. 1985;9:310–314.
magnetic resonance imaging. AJR Am J Roentgenol. 1984;143:
1189–1191.
Mediastinum 4
COMPUTED TOMOGRAPHY TECHNIQUES 291 PARATHYROID GLANDS 349

Computed Tomography Evaluation of Parathyroid
MAGNETIC RESONANCE IMAGING Disease 350
TECHNIQUES 293 Magnetic Resonance Evaluation of Parathyroid
Disease 351
NORMAL MEDIASTINUM: REGIONAL
ANATOMY 293 MEDIASTINAL LYMPH NODES AND LYMPH
Transverse Plane 294 NODE MASSES 352
Sagittal and Coronal Planes 301 Lymph Node Groups 352
Computed Tomography Appearance of Normal
DIAGNOSIS OF MEDIASTINAL MASS 303 Lymph Nodes 357
Computed Tomography Diagnosis of Lymph Node
MASSES PRIMARILY INVOLVING THE Abnormalities 358
PREVASCULAR SPACE 306 Lymph Node Enlargement 360
Computed Tomography Assessment of Lymph Node
NORMAL THYMUS 308 Morphology 362
Appearance of the Thymus in Children 309 Computed Tomography Assessment of Lymph Node
Appearance of the Thymus in Adults 309 Attenuation 362
Location of Enlarged Mediastinal Lymph
THYMIC ENLARGEMENT AND MASSES 313 Nodes 366
Thymic Enlargement 313 Magnetic Resonance Evaluation of Mediastinal
Thymic Hyperplasia 314 Lymph Nodes 368
Thymic Rebound 315
Thymoma and Thymic Epithelial Tumors 315 LYMPHOMA 368
Imaging in Myasthenia Gravis 330 Hodgkin Lymphoma 369
Thymic Carcinoma 331 Non-Hodgkin Lymphoma 371
Thymic Neuroendocrine Tumor 332 Primary Mediastinal Non-Hodgkin Lymphoma 373
Thymolipoma 333 Magnetic Resonance Imaging Diagnosis
Thymic Cyst 334 of Lymphoma 374
Thymic Lymphoma and Metastases 335 Diagnosis of Residual Tumor After Treatment 376
GERM-CELL TUMORS 336 LEUKEMIA 382

Teratoma 337
Seminoma 341 CASTLEMAN DISEASE 384
Nonseminomatous Germ-Cell Tumors 341
OTHER LYMPHOPROLIFERATIVE DISORDERS 385
THYROID GLAND 343
Computed Tomography Evaluation of Thyroid METASTATIC TUMOR 385
Disease 343
Magnetic Resonance Evaluation of Thyroid Disease 346 SARCOIDOSIS 386
INFECTIONS 388 sensitivity and ability to delineate the extent of mediastinal

Tuberculosis 388 abnormalities and the relationship of masses to specific
Fungal Infections 389 mediastinal structures.
Because of its excellent density resolution and tomo-
FIBROSING MEDIASTINITIS 390 graphic format, CT is able to identify normal mediastinal
structures, opacified vessels, and vascular abnormalities
MEDIASTINAL CYSTS 392 and is often able to characterize masses based on their
Bronchogenic Cyst 392 attenuation and precisely localize them as to their extent
Esophageal Duplication Cyst 394 and site of origin.
Neurenteric Cyst 394 The indications for mediastinal CT can be divided into
Pericardial Cyst 395 two broad categories. First, CT is commonly used to define
Thoracic Duct Dilatation and Thoracic Duct Cyst 396 and characterize a mediastinal abnormality suspected or
Differential Diagnosis of Cystic Masses 396 diagnosed on plain chest radiographs. When a mediastinal
contour abnormality is visible on plain films, CT should
TUMORS AND MASSES OF MESENCHYMAL generally be the next imaging procedure performed.
ORIGIN 398 Second, CT is often used to evaluate the mediastinum in
Tumors of Adipose Tissue and Fatty Masses 398 patients who have normal chest radiographs yet a clinical
Tumors and Masses of Vascular Origin 403 reason to suspect mediastinal disease. For example, CT has
Tumors and Masses of Fibroblastic Origin 408 been shown to be of significant value in the detection and
Tumors of Muscle Origin 410 assessment of mediastinal masses or lymph node abnor-
Tumors of Bone and Cartilage 411 malities in patients with neoplasm or granulomatous
disease and in the detection of a pathologic process in
ESOPHAGUS 412 patients with conditions such as myasthenia gravis or
Esophageal Carcinoma 413 surgically resistant hyperparathyroidism, which can be
Differentiation of Intrinsic and Extrinsic Esophageal associated with a mediastinal mass (thymoma and ectopic
Lesions 418 parathyroid adenoma, respectively) (1–4).
Benign Esophageal Tumors 419 The clinical indications for mediastinal magnetic reso-
Esophagitis 422 nance imaging (MRI) are somewhat more limited (5–12).
Esophageal Dilatation 422 In certain situations, MRI provides diagnostic information
Esophageal Varices 422 superior to that available from CT, and MRI can be used as
Hiatal Hernia 423 the primary imaging modality. Primary uses of MRI
Esophageal Perforation 427 include (a) the diagnosis of mediastinal abnormalities
that are suspected to be vascular, (b) the evaluation of
MEDIASTINITIS AND MEDIASTINAL ABSCESS 428 posterior mediastinal or paravertebral masses and neuro-
genic tumors, and (c) distinguishing mass and fibrous
PARASPINAL ABNORMALITIES 429
tissue in a patient with treated lymphoma or carcinoma in
whom tumor recurrence may be present. These primary
NEUROGENIC TUMORS 430
uses of MRI generally relate to its ability (a) to image ves-
Peripheral Nerve Sheath Tumors 430
sels, (b) to distinguish different tissues, (c) to image in
Tumors Originating from Sympathetic Ganglia 432
non-transaxial planes, or (d) to obviate iodinated contrast
Magnetic Resonance Evaluation of Neurogenic
material. In many other instances, however, MRI is best
Tumors 435
reserved as a secondary or problem-solving tool, in
patients who have equivocal or confusing CT findings, or
PARAGANGLIOMA 435
in whom specific anatomic information is required,
which is less clearly demonstrated using CT (5,9,10,13).
ANTERIOR OR LATERAL THORACIC
In such cases, a limited number of MR images can be
MENINGOCELE 436
helpful.
EXTRAMEDULLARY HEMATOPOIESIS 436 It is important to understand, however, that if MRI pro-
vides diagnostic information that is merely equivalent to
MEDIASTINAL PSEUDOCYST 437 that of CT, then CT is usually the more appropriate clinical
study (5). CT is less expensive, is more available, takes less
time, is easier for debilitated patients to tolerate, often pro-
Computed tomography (CT) is indispensable in imaging vides more ancillary information, and has an established
the mediastinum. Although conventional radiographs can role in the clinical diagnosis of a number of entities.
show recognizable abnormalities in many patients with However, it is a mistake to think of CT and MRI as necessar-
mediastinal pathology, radiographs are limited in their ily competitive. Rather, CT and MRI can be complementary
Chapter 4: Mediastinum 291
in many instances—each has its own particular strengths kVp, mAs, field of view, and reconstruction algorithm,
that may be advantageous in specific circumstances. these include contrast administration (intravenous or oral),
detector width, pitch, reconstruction interval, and recon-
struction plane. In addition, important choices regarding
COMPUTED TOMOGRAPHY contrast administration must be made, including type of
TECHNIQUES contrast agent, rate and method of infusion, and optimal
scan delay. These decisions are further complicated by the
The use of helical CT to acquire a volumetric dataset during frequent need to integrate chest imaging with neck and/or
a single breath-hold period is fundamental to imaging the abdominal examinations.
mediastinum and thorax. This technique eliminates misreg- Using multidetector CT, scanning the entire thorax
istration between adjacent slices, and because of the rapid during suspended respiration is easily achieved. It is opti-
scan time, minimizes respiratory and, to a lesser degree, car- mal to acquire the scan data using a narrow detector width
diac motion artifacts. As important, rapid acquisition of scan (e.g., 1 mm), with the volumetric data reconstructed as
data ensures optimal delivery and utilization of intravenous thicker images (e.g., 5 mm) for routine viewing. Thinner
contrast material, not only ensuring that all mediastinal ves- slices may also be reconstructed and are sometimes required
sels will be opacified on all sections but allowing a more for (a) small lesions, (b) imaging regions in which
accurate assessment of normal and abnormal vessels. anatomic structures are not oriented perpendicular to the
It is best to think of helical CT as a method for acquiring scan plane [such as in the aortopulmonary window (APW)],
a volumetric dataset, which may then be interrogated by the (c) when mediastinal fat is lacking, or (d) if hilar structures
viewer to create two-dimensional (2D) or three- and bronchi also need to be assessed. Non-transaxial image
dimensional (3D) displays of varying thickness and in reconstructions may be obtained as needed for diagnosis
various planes anywhere within the scan volume; examples but are not generally considered routine.
of 3D reconstructions include shaded surface displays, 3D Although adequate images can often be obtained without
displays from an endoluminal perspective (e.g., virtual bron- the use of intravenous contrast media, especially if the
choscopy, endoscopy, or aortography), and maximum and primary indication for CT is to determine the presence or
minimum intensity projection imaging (13–25). absence of a mediastinal mass, in our experience whenever
However, despite the ease of performing multiplanar or possible the use of contrast administration should be
3D reconstructions with current CT scanners, routine encouraged (31). Contrast infusion can be indispensable in
transaxial images are well suited to the assessment of a num- identifying and differentiating between vascular and nonvas-
ber of mediastinal structures, such as the trachea, esophagus, cular structures. Contrast media should be used routinely if
aorta, and superior vena cava, which are oriented perpendi- clinical or radiographic findings suggest the possibility of (a)
cular to the plane of scan and are thus imaged in cross sec- a vascular abnormality (Fig. 4-1), (b) a hilar mass or hilar
tion; provide a good display of most lymph node–bearing lymph node enlargement (as in patients with lung cancer)
regions in the mediastinum; and are usually sufficient for (Fig. 4-2), or (c) a mediastinal mass that may be invading or
the diagnosis of mediastinal abnormalities. The use of image compressing vessels (as in the presence of superior vena cava
reconstruction in non-transaxial planes is most useful in syndrome) (Fig. 4-2). Also, the enhancement characteristics
demonstrating the aorta, great vessels, pulmonary arteries, or of some mediastinal lesions can be helpful in diagnosis or
trachea along their longitudinal axes (25–29). In selected differential diagnosis. For example, in patients with active
cases, reconstructions may be of value in showing the rela- tuberculosis (TB), enlarged mediastinal lymph nodes often
tionship of a mediastinal mass to structures such as the show a low-attenuation center and an enhancing rim on
heart, aorta, pleura, or diaphragm, but generally their use is contrast-enhanced CT, findings that can strongly suggest the
of limited value. For example, the role of multiplanar image diagnosis (32–34). Similarly, contrast enhancement may be
reconstruction in staging lung cancer was assessed in 41 con- an invaluable aid in assessing the presence and extent of
secutive patients with lung cancer who underwent multislice tumor necrosis as occurs in some patients with lymphomas
helical CT (30). Although a significant increase in confi- or germ-cell tumors (35,36). Use of contrast enhancement
dence was found when diagnosing various features of the also allows differentiation of highly vascular lymph nodes
primary tumor by using multiplanar reconstructions, no sig- and tumors from patients with avascular masses or cysts. In
nificant difference was demonstrated in the diagnosis of me- patients with a contraindication to administration of iodi-
diastinal lymphadenopathy (30). A meta-analysis of virtual nated contrast agents, gadolinium may be used as an alter-
bronchoscopy, including 17 studies and 459 patients, has nate contrast agent with rapid multidetector scanners (37).
indicated a sensitivity of 84% and a specificity of 75% in the Obtaining both precontrast and postcontrast images is
diagnosis of central airway lesions (20). The use of virtual not usually necessary but may be of value in some cases.
bronchoscopy may also be of value in guiding the broncho- On enhanced images it may be impossible to distinguish a
scopic biopsy of mediastinal lesions (24). high-attenuation mass (e.g., containing calcium or blood)
Performing a helical examination requires a number of from an enhancing mass. Unenhanced images are also
technical choices. In addition to standard options such as valuable in the diagnosis of aortic intramural hematoma.
A B
Figure 4-1 Vascular disease: value of contrast-enhanced helical CT. A: Helical CT sections following the bolus injection of 120 mL 60%
iodinated contrast media at a rate of 2 mL/second in a patient with a left mediastinal mass visible on chest radiographs shows a focal saccular
aneurysm of the aortic arch. Note that in B the contrast-opacified lumen of the aneurysm (An) can be clearly distinguished from clot lining its wall.
Delayed images may also be obtained if unenhanced current scanners, the use of reduced amounts of contrast is
images were not obtained prospectively. possible, even when injection at a high rate. It has been
If contrast injection is required, it is best administered in suggested that the use of reduced iodine concentration,
the form of a rapid bolus, preferably utilizing a power achieved by diluting full-strength contrast medium with
injector. Because of the rapid scan acquisition possible with saline, may be useful by reducing venous flow artifacts and
A B
Figure 4-2 Superior vena caval obstruction: bronchogenic carcinoma. A: Contrast-enhanced helical CT in a patient with radiographic
evidence of a right hilar mass obtained following a bolus of 120 mL of 60% iodinated contrast media injected at a rate of 2 mL/second con-
firms the presence of tumor invading the mediastinum (arrow) anterior to the right main bronchus and carina. The aorta (Ao) and superior
vena cava are densely opacified at this rate of contrast infusion. B: At a lower level, tumor surrounds and narrows the right pulmonary artery
(white arrow) and compresses the superior vena cava (black arrow). PA, main pulmonary artery.
improving image quality (38). However, this technique is Although, as noted previously, evaluation of mediasti-
not in wide use. Contrast administration at the rate of 3 to nal masses with MRI has traditionally relied on acquisition
4 mL/second is typical. of conventional spin-echo (SE) T1- and T2-weighted
It should be noted that regardless of technique, images with ECG triggering supplemented with cine gradi-
although it is common to see small bubbles of air in the ent-echo images as needed, image quality is heavily
most anterior part of the left brachiocephalic vein or main dependent on a good ECG signal. As a consequence, exam-
pulmonary artery after contrast infusion, clinically signifi- inations are often suboptimal or even nondiagnostic in
cant venous air embolism is an unusual complication of patients with abnormal cardiac rhythms. Even with
intravenous contrast injection (39,40). optimal triggering, examination times are lengthy, which
The use of retrospective electrocardiographic (ECG) and predisposes to degradation from patient motion. Finally,
respiratory gating has been used to advantage with current respiratory motion may also degrade image quality.
multidetector scanners to significantly reduce cardiac and Technologic advances, which include the widespread
respiratory motion artifacts (41–44). It likely will be use of echo-train fast spin-echo (ETFSE) pulse sequences,
widely used, particularly when imaging small vessels such specialized phased-array multicoils and magnets with
as the coronary arteries (43,45–47). high performance gradient systems now provide for a
comprehensive evaluation of the mediastinum with short
acquisition times. Because of these advances, many
MAGNETIC RESONANCE IMAGING sequences can be performed during a comfortable breath-
TECHNIQUES hold, thus eliminating respiratory artifacts.
A typical imaging protocol for evaluation of a medi-
In our experience, MRI is best utilized to evaluate vascular astinal mass now includes ECG-triggered ETFSE T1- and
structures in patients who cannot tolerate radiographic con- T2-weighted images. These can be acquired with breath-
trast agents or as a problem-solving modality when CT has holding if the TR is sufficiently shortened at the expense
proven inconclusive (5–7,12,48,49). MRI is useful in con- of less slice coverage. Alternatively an ETFSE short T1
firming the cystic nature of mediastinal lesions that appear inversion recovery sequences could be used instead of
of soft tissue attenuation on CT and, by revealing small T2-weighted images. In patients with abnormal cardiac
amounts of intralesional fat, can suggest the diagnosis of rhythms and those who cannot breath-hold or are other-
hemangioma, teratoma, or extramedullary hematopoiesis. wise uncooperative, a diagnostic quality examination can
MRI is the preferred modality for imaging neurogenic still be performed with single shot ETFSE sequences with
tumors, because of its multiplanar capability and high half Fourier reconstruction (HASTE). Because each slice is
contrast resolution; it is optimal for delineating the acquired in less than 1 second, these sequences are imper-
intraspinal extension and craniocaudad extent of neuro- vious to respiratory motion.
genic lesions (48). Finally, a comprehensive vascular evaluation can be
Each case is individualized according to the information performed with breath-hold 3D gadolinium-enhanced
needed, and no standard algorithm is universally MRI. The ultrashort TR and echo time (TE) inherent to
applicable. As a rule, following acquisition of initial coronal these pulse sequences allows for imaging the great vessels
scout images, cardiac-gated, T1-weighted axial sections are with a temporal resolution of as little as 10 seconds. This
obtained through regions of interest and are most valuable allows clear distinction between arteries and veins. In
in demonstrating mediastinal structures and delineating addition the 3D technique is amenable to high-quality
the presence of a mediastinal mass. With gating, the multiplanar reformations and can be postprocessed with
effective repetition time (TR) is determined by the heart maximum intensity projection (MIP) algorithms. These
rate. Coronal and sagittal images are generally best reserved MIP images resemble conventional angiographic images.
for specific indications, such as analyzing the chest wall, the Gadolinium-enhanced 3D MRI represents a significant
APW, or subcarinal space (5–7). advance from the traditional time-of-flight techniques
T1-weighted images obtained following contrast admin- because saturation effects and pulsatility artifacts are
istration or T2-weighted images are usually required only minimized.
in those cases for which tissue characterization is deemed
necessary. T2-weighted images may be acquired either uti-
lizing a prolonged TR (2,000 ms) without cardiac gating or NORMAL MEDIASTINUM: REGIONAL
preferably utilizing cardiac gating (especially if the region ANATOMY
of interest is near the heart or hila) with images obtained
every second or third cardiac cycle. Gradient-refocused The accurate interpretation of mediastinal images, whether
sequences with breath-holding also may be of value to obtained using CT or MRI, requires a detailed knowledge
reduce scan time and optimize visualization of vascular of normal cross-sectional anatomy. Transverse imaging is
pathology. With this technique, flowing blood results in an fundamental to most imaging studies. In addition, sagittal
intense signal, whereas soft tissue contrast is reduced. and coronal images can be valuable in selected cases and
may be obtained using MRI or by reformation of spiral CT bronchi serve as reliable guides to localizing other impor-
data; some knowledge of mediastinal anatomy in these tant mediastinal structures. Furthermore, the relationship
planes is also essential. between vessels and other mediastinal structures is impor-
tant to know if the potential effects of masses on vascular
structures is to be assessed.
Transverse Plane
As an aid to understanding regional mediastinal ana-
Cross-sectional anatomy is best learned using CT. In most tomy, the mediastinum can be conceptualized as divided
respects, the anatomy of mediastinal structures shown into several compartments, respectively, from superior to
using MRI is identical to that shown on CT, although the inferior, (a) the superior or supra-aortic mediastinum;
gray-scale representation of various tissues is different with (b) the region of the aortic arch, subaortic mediastinum,
the two techniques. and APW; (c) the subcarinal space and azygoesophageal
The anatomy of the mediastinal great vessels and air- recess; and (d) the paracardiac mediastinum.
ways is discussed in detail in subsequent chapters, but it
also is reviewed briefly in this section, as a knowledge of
Superior or Supra-aortic Mediastinum
this anatomy is important in understanding the medi-
astinum as a whole. The aorta and its branches, the great At or near the thoracic inlet (Figs. 4-3 and 4-4), the medi-
veins, the pulmonary arteries, and the trachea and main astinum is relatively narrow from anterior to posterior.
A B
C D
Figure 4-3 Thoracic inlet: CT and MRI evaluation. A: CT section through the lower neck, just above the thoracic inlet. Note the close
proximity of the thyroid gland to the trachea. B–D: Sequential MR images from above-downward through the thoracic inlet. The anatomy of
the thoracic inlet is especially well visualized with MRI. Note the relationship between the anterior scalene muscle and the subclavian artery
that lies immediately posteriorly. The brachial plexus usually can be identified just above the subclavian artery as it passes behind the
anterior scalene. AJV, anterior jugular veins; ASm, anterior scalene muscle; AXA, axillary artery; AXV, axillary vein; BP, brachial plexus;
E, esophagus; EJV, external jugular vein; IJV, internal jugular vein; JV, jugular vein; LCCA, left common carotid artery; LJV, left jugular
vein; LVA, left vertebral artery; MSm, middle scalene muscle; RCCA, right common carotid artery; RJV, right jugular vein; SA, subclavian
artery; SCMm, sternocleidomastoid muscle; Th, Thy, thyroid; Tr, trachea; VA, vertebral artery.
A B
C D
E
Figure 4-4 Supra-aortic mediastinum: normal CT anatomy. Sections were obtained using 1.25-mm slice thickness during contrast injec-
tion at a rate of 4.5 mL/sec. A: At the level of the lung apices, the trachea (T) is clearly seen, with the air-filled esophagus posterior and to
the left of it. The right (RIJV) and left internal jugular (LIJV) veins are located lateral to the trachea, being seen medial to the clavicles (C). The
common carotid arteries (CCA) are medial to the jugular veins. The vertebral arteries (VA) and the right (RSCA) and left (LSCA) subclavian
arteries are also visible at this level. B: At 2.5 mm below A, the right (RBCV) and left (LBCV) brachiocephalic veins are visible posterior to the
clavicular heads (C). The large arterial branches of the aorta are situated anterior and lateral to the trachea (T). The left subclavian artery
(LSCA) is most posterior and is situated lateral to the trachea and esophagus (E), contacting the mediastinal pleura. The left common carotid
artery (LCCA) is anterior to the left subclavian artery and is somewhat variable in position. The right common carotid artery (RCCA) is usually
anterior and to the right of the tracheal midline. The right subclavian artery (RSCA) lies to the right of the trachea and is posterolateral to
the right common carotid. The right subclavian vein (SCV) is densely opacified because of contrast injection into the right arm. C: At 5 mm
below B, the right (RBCV) and left (LBCV) brachiocephalic veins are visible in the anterior mediastinum, posterior to the clavicles (C). The left
subclavian (LSCA) and common (LCCA) carotid arteries maintain the same positions relative to the trachea (T) and esophagus (E) as in B. The
innominate artery (IA) is visible at this level, anterior and to the right of the trachea. D: At 2.5 mm below C, the left (LBCV) brachiocephalic
vein begins to cross the mediastinum and is located posterior to the manubrium (M). The right brachiocephalic vein (RBCV) is anterior and to
the right of the trachea (T), contacting the mediastinal pleura. The innominate artery (IA) is anterior to the trachea at this level. LSCA, left
subclavian artery; LCCA, left common carotid artery; E, esophagus. E: At 2.5 mm below D, the left (LBCV) brachiocephalic vein joins the
right (RBCV) to form the superior vena cava. The right internal mammary vein (RIMV) is visible arising from the anterior right brachiocephalic
vein. The left internal mammary artery and vein are also visible (LIMA&V). The innominate artery (IA) and the left common carotid (LCCA) ar-
teries are closely opposed near their origins from the aortic arch. Mediastinal fat and normal lymph nodes occupy the pretracheal space
(PreTrSp), anterior to the trachea (T) and posterior to the vascular structures.
The trachea appears central in the supra-aortic medi- constant. The great arterial branches lie posterior to the
astinum. The esophagus lies posterior to the trachea at veins and adjacent to the anterior and lateral walls of the
this level, but depending on the position of the trachea trachea. These can be reliably identified by their consistent
relative to the spine, the esophagus can be displaced to positions.
right or left. It is usually collapsed and appears as a Below the thoracic inlet, the left brachiocephalic vein
flattened structure of soft tissue attenuation, but small crosses the mediastinum from left to right, anterior to the
amounts of air or air and fluid may sometimes be seen in arterial branches of the aorta (Figs. 4-4D and E). The two
its lumen (Fig. 4-4A). brachiocephalic veins have very different configurations.
Above the level of the aortic arch, the large arterial The right brachiocephalic vein has a nearly vertical course
branches of the aorta (i.e., the innominate artery, the left throughout its length; the left brachiocephalic vein is
common carotid artery, and the left subclavian artery) and longer and courses horizontally as it crosses the medi-
the brachiocephalic veins are the most apparent normal astinum. The horizontal segment of the left brachio-
structures, other than the trachea and esophagus (Figs. 4-4 cephalic vein is a convenient anatomic landmark, being
and 4-5). At or near the thoracic inlet, the brachiocephalic the line of demarcation between the anterior medi-
veins are the most anterior and lateral vascular branches astinum (the prevascular space) that is anterior to it and
visible, lying immediately behind the clavicular heads. the middle mediastinum posterior to it. The precise
Although they vary in size, their positions are relatively cephalocaudal location of this vein is quite variable;
A B
C D
Figure 4-5 Cross-sectional MRI anatomy. A–D: Cardiac-gated MR images through the mediastinum from above-downward. Note that signal
is present within the azygos vein, posterior to the SVC (arrow in C), presumably secondary to slow flow. Tr, trachea; LBV, left brachiocephalic
vein; LCCA, left common carotid artery; LSA, left subclavian artery; E, esophagus; RBV, right brachiocephalic vein; BA, brachiocephalic artery;
Cl, clavicles; Ao, aorta; SVC, superior vena cava; MPA, main pulmonary artery; LMB, left mainstem bronchus; LPA, left interlobar pulmonary
artery; DAo, descending aorta; BI, bronchus intermedius; AAo, ascending aorta; RSPV, right superior pulmonary vein; RMPA, right main
pulmonary artery.
although it is most frequently visible at the level of the position to the left of the spine. At this level, the superior
great vessels, the horizontal portion of the left brachio- vena cava is visible anterior and to the right of the
cephalic vein also can be seen at the level of the aortic trachea, usually being elliptical (Figs. 4-5C and 4-6). The
arch (Fig. 4-5B). Because of the central location of the left esophagus appears the same as at higher levels, being
brachiocephalic vein in the supra-aortic mediastinum, posterior to the trachea but variable in position (50).
when a mass is suspected in this region, intravenous con- Often it lies somewhat to the left of the midline of the
trast medium can be injected using a left-sided arm vein. trachea.
This maximizes visualization of the full length of the left The aortic arch on the left, the right brachiocephalic
brachiocephalic vein. vein or superior vena cava and mediastinal pleura on the
The innominate artery is located in close proximity to right, and the trachea posteriorly serve to define the pretra-
the anterior tracheal wall, near its midline or slightly to the cheal or anterior paratracheal space (Fig. 4-6A) (51,52).
right of midline in most normal patients; it is the most This fat-filled space contains middle mediastinal lymph
variable of all the great arteries (Figs. 4-4C–E and 4-5A). nodes in the pretracheal or anterior paratracheal chain.
The left common carotid artery lies to the left and slightly Other mediastinal node groups are closely related to this
posterolateral to the innominate artery; generally it has group both spatially and in regard to lymphatic drainage.
the smallest diameter of the three major arterial branches. It is not uncommon to see a few normal-sized lymph
The left subclavian artery is a relatively posterior structure nodes in this compartment.
throughout most of its course, lying to the left of and Anterior to the great vessels (aorta and superior vena
frequently directly lateral to the trachea. The lateral border cava) at this level is another triangular space called the
of the left subclavian artery typically indents the mediasti- prevascular space. This compartment, which is anterior
nal surface of the left upper lobe. mediastinal, primarily contains lymph nodes, the thymus,
Other than the great vessels, trachea, normal lymph and fat (Figs. 4-5C and 4-6). The apex of this triangular
nodes, and esophagus, little is usually seen in the supra- space represents the anterior junction line, which can
aortic mediastinum. Small vascular branches, particularly sometimes be seen on chest radiographs. At higher levels
the internal mammary veins, can sometimes be seen in this in the mediastinum (Fig. 4-4), the prevascular space is
part of the mediastinum (Fig. 4-4D). In some patients, the anterior to the great arterial branches of the aorta and the
thyroid gland may extend into the superior mediastinum, brachiocephalic veins.
and the right and left thyroid lobes may be visible on each In young patients, usually in their teens or early twen-
side of the trachea. This appearance is not abnormal and ties, CT shows the thymus to be of soft tissue attenuation
does not imply thyroid enlargement. On CT, the thyroid (39,53–58). The thymus has two lobes, the right and the
can be distinguished from other tissues or masses; because left, but the left thymic lobe often predominates. On CT,
of its iodine content, its attenuation is greater than that of the thymus often appears bilobed or arrowhead shaped,
soft tissue. with each of the two thymic lobes contacting the mediasti-
nal pleura. Each lobe usually measures 1 to 2 cm in thick-
ness (measured perpendicular to the pleura), but this is
variable. In patients in whom the left thymic lobe predom-
Aortic Arch, Subaortic Mediastinum,
inates, the thymus appears to be predominantly left sided,
and Aortopulmonary Window
paralleling the aortic arch. In adulthood, the thymus
Although the supra-aortic region largely contains arterial gradually involutes, and soft tissue attenuation elements
and venous branches of the aorta and vena cava, this com- are replaced by fat (53–55,57). Thus, in patients older
partment contains the undivided mediastinal great vessels, than 30 years, the prevascular space appears to be prima-
including the aorta, superior vena cava, and pulmonary rily fat filled, with thin strands of wispy tissue passing
arteries. This compartment also contains a number of through the fat. Most of this, including the fat, actually
important lymph node groups. represents the thymus. At higher levels, the thymus is
The aortic arch has a characteristic but somewhat sometimes visible anterior to the brachiocephalic arteries
variable appearance. The anterior aspect of the arch is and veins, within the prevascular space.
located anterior and to the right of the trachea, with the At a level slightly below the aortic arch, the ascending
arch passing to the left and posteriorly (Figs. 4-5B and C aorta and descending aorta are visible as separate struc-
and 4-6). The posterior arch usually lies anterior and tures (Fig. 4-6B and C). Characteristically, the descending
lateral to the spine. The aortic arch tapers slightly along aorta is slightly smaller (mean diameter 2.6 cm) than the
its length, from anterior to posterior. The position of the ascending aorta (mean diameter 3.5 cm) (40).
anterior and posterior aspects of the arch can vary in At or near this level, the trachea bifurcates into the
the presence of atherosclerosis and aortic tortuosity; typi- right and left main bronchi (Fig. 4-6C and D). Near the
cally the anterior arch is located more anteriorly and to carina, the trachea commonly assumes a somewhat trian-
the right in patients with a tortuous aorta, although the gular shape (51,52). The carina itself is usually visible on
posterior aorta lies more laterally and posteriorly, in a CT. On the right side, the arch of the azygos vein enters
A B
C D
E
Figure 4-6 Aortic arch, subaortic mediastinum, and aortopulmonary window: normal CT anatomy. This is the same scan series as in
Figure 4-4. A: Aortic arch level, 2.5 mm below Figure 4-5E. The superior vena cava (SVC) contacts the right mediastinal pleura, and along
with the aortic arch (Ao) delineates the anterior aspect of the pretracheal space (PreTrSp). The trachea (T) delineates the posterior aspect of
the pretracheal space. The prevascular space (PreVasSp) is anterior to the great vessels, and contains the thymus, which is of soft tissue
attenuation in this patient. B: Subaortic mediastinum. Aorticopulmonary window level, 2.5 mm below A. At the level of the tracheal carina
(C), the aorticopulmonary window (APW) is visible under the aortic arch, and between the ascending aorta (AAo) and proximal descending
aorta (DAo). On the right side of the mediastinum, the azygos arch (AzA) is visible arising from the posterior aspect of the superior vena
cava (SVC), contacting the right mediastinal pleura, and forming the lateral margin of the pretracheal (or in this instance, the precarinal)
space. PVS, prevascular space; Th, thymus; IMA&V, internal mammary arteries and veins. C: Subaortic mediastinum. Left pulmonary artery
level, 2.5 mm below B. The left pulmonary artery (LPA) marginates the caudal aspect of the aortopulmonary window. At this level the right
and left main bronchi are visible. The precarinal space (PreCarSp) is continuous with the pretracheal space seen at higher levels. The trian-
gular prevascular space (PVS) containing thymus (Th) remains visible. SVC, superior vena cava; Ao, aorta; AzV, posterior azygos vein; HAzV,
hemiazygos vein; E, esophagus. D: Subcarinal space and azygoesophageal recess level, 2.5 mm below C. The superior aspect of the
azygoesophageal recess (AzEsRec) is visible, in close relationship to the posterior azygos vein (AzV), esophagus (E), and subcarinal space
(SubCarSp). PVS, prevascular space, Th, thymus; SVC, superior vena cava; Ao, aorta; PA, main pulmonary artery; LPA, left pulmonary artery.
E: Right pulmonary artery and azygoesophageal recess, 3.75 mm below D. The right pulmonary artery (RPA) arises from the main
pulmonary artery (PA) and crosses the mediastinum anterior to the subcarinal space, azygoesophageal recess, azygos vein (AzV), and esoph-
agus (E). It is closely associated with the superior vena cava (SVC) and ascending aorta (Ao) along its anterior wall. PVS, prevascular space,
Th, thymus; LPA, left pulmonary artery; LMB, left main bronchus; BI, bronchus intermedius; LSPV, left superior pulmonary vein.
the posterior wall of the superior vena cava, passes or azygos vein and is particularly common in patients with
over the right main bronchus, and continues posteriorly a narrow mediastinum and in children (64).
along the mediastinum, to lie to the right and anterior In many subjects, the azygoesophageal recess is some-
of the spine (59). Below the level of the azygos arch, the what posterior to the node-bearing subcarinal space,
azygos vein is consistently seen in this position. The which lies between the main bronchi (65). Normal nodes
azygos arch is often visible on one or two adjacent slices are commonly visible in this space, being larger than
and sometimes appears nodular (60). However, its charac- normal nodes in other parts of the mediastinum and up to
teristic location is usually sufficient to allow its correct 1.2 cm in short axis diameter (66,67). The esophagus
identification. When the azygos arch is visible, it mar- usually is seen immediately behind the subcarinal space,
ginates the right border of the pretracheal space. and distinguishing nodes and esophagus may be difficult,
On the left side of the mediastinum, caudal to the unless the esophagus contains air or contrast material, or
aortic arch, but cephalad to the main pulmonary artery, is its course is traced on adjacent scans. At levels below the
the region termed the aortopulmonary window (APW) subcarinal space, the appearance of the azygoesophageal
(61). The APW contains fat, lymph nodes, the left recurrent recess is relatively constant, although it narrows anteropos-
laryngeal nerve, and the ligamentum arteriosum; the latter teriorly in the retrocardiac region (Figs. 4-6D and E and
two are usually invisible. These lymph nodes freely com- 4-7A–C).
municate with those in the pretracheal space, and, in fact, Also, at or near this level, the main pulmonary artery
a distinction between nodes in the medial APW from divides into its right and left branches. The left pulmonary
those in the left pretracheal space is rather arbitrary and artery is somewhat higher than the right, usually being
generally unnecessary. In some patients, the APW is not seen 1 cm above it, and appears as the continuation of the
well seen, with the main pulmonary artery lying immedi- main pulmonary artery, directed posterolaterally and to
ately below the aortic arch. In such patients, it is usually the left (Figs. 4-6C–E and 4-7A). The right pulmonary
difficult to distinguish APW lymph nodes from volume artery arises at an angle of nearly 90 degrees to the main
averaging of the adjacent aorta and pulmonary artery, and left pulmonary arteries and crosses the mediastinum
unless thin collimation is used. from left to right, anterior to the carina or main bronchi
At or slightly below the APW, the level of the ascending (Fig. 4-6E). The right pulmonary artery limits the most
aorta is first clearly seen in cross section (i.e., it is round or caudal extent of the pretracheal space.
nearly round); a portion of the pericardial space, usually The azygos vein parallels the esophagus along the right
containing a small amount of pericardial fluid, extends side of the mediastinum and laterally contacts the medial
cephalad, into the pretracheal space, immediately behind pleural reflections of the right lower lobe, defining the
the ascending aorta. This part of the pericardium is called medial and posterior border of the azygoesophageal
the superior pericardial recess (62,63). Although it can recess (Fig. 4-6D and E and 4-7A–C) (59). On the left side,
sometimes be confused with a lymph node, its typical the hemiazygos vein parallels the descending aorta, lying
location, immediately behind and contacting the aortic posterior to it; it is not always visible. The hemiazygos vein
wall, its oval or crescentic shape, and its relatively low generally drains into the azygos vein via communicating
(water) attenuation allow it to be distinguished from a sig- branches that cross the midline, behind the aorta, in the
nificant abnormality (see Fig. 4-81). Another part of the vicinity of the T8 vertebral body. The communicating vein
pericardial recess can sometimes be seen anterior to the or veins are sometimes seen on CT in normal individuals
ascending aorta and pulmonary artery. (68,69).
Subcarinal Space and Azygoesophageal Paracardiac Mediastinum

Recess
The heart occupies most of the inferior mediastinum, but
Below the level of the tracheal carina and azygos arch, the other important structures at this level include the descend-
medial aspect of the right lung contacts the posterior ing aorta, esophagus, azygos and hemiazygos veins, thoracic
aspect of the middle mediastinum, in close association to duct, vagus and inferior phrenic nerves, and the pericardium
the azygos vein and esophagus (Figs. 4-6D and E). This (Fig. 4-7). Important lymph node groups are found in rela-
part of the mediastinum, called the azygoesophageal tion to the azygoesophageal recess and esophagus at this
recess, is important because of adjacent subcarinal lymph level, the inferior pulmonary ligaments, the paravertebral
nodes and its close relationship to the esophagus and regions, the diaphragm, and the pericardium.
main bronchi. The contour of the azygoesophageal recess The descending aorta and esophagus maintain the same
is concave laterally in the large majority of normal sub- relative positions as at higher levels. The azygos and hemi-
jects, and a convexity in this region should be regarded as azygos veins are often but not always visible at this level;
suspicious of mass, and the scan examined closely for a the azygos vein is larger and more easily seen.
pathologic process. However, a convexity in this region The thoracic duct enters the thorax through the aortic
may also be produced by a prominent normal esophagus hiatus, along the right anterior aspect of the spine. It
A B
C D
E
Figure 4-7 Paracardiac mediastinum: normal CT anatomy. This is the same scan series as in Figures 4-4 and 4-6. A: Left atrium and main
pulmonary artery, 3.75 mm below E. The main pulmonary artery (PA) and ascending aorta (Ao) are anterior. The superior aspect of the left
atrium (LA) is visible. The subcarinal space is no longer visible at this level, and the prevascular space becomes inconspicuous, although the
inferior thymus is present anteriorly. The azygoesophageal recess remains visible. SVC, superior vena cava; SPV, superior pulmonary vein;
RMLA, right middle lobe artery; RLLPA, right lower lobe pulmonary artery; LAA, left atrial appendage; LIPA, left interlobar pulmonary artery;
E, esophagus. B: Left atrium, aortic root, and right ventricular outflow tract, 2.5 mm below Figure 4-8A. The right ventricular outflow tract
(RVOT) and is anterior and the left atrium (LA) is posterior. The azygoesophageal recess remains visible. RAA, right atrial appendage; SVC, supe-
rior vena cava; RSPV, right superior pulmonary vein; RLLPA, right lower lobe pulmonary artery; SSPA, superior segment lower lobe pulmonary
artery; Ao, aortic root; LSPV, left superior pulmonary vein; LIPV, left inferior pulmonary vein; E, esophagus.
F G
Figure 4-7 (continued) C: Cardiac chambers and paracardiac mediastinum, 2.5 mm below 8A. The prevascular space is nearly obliterated,
as the heart fills the anterior mediastinum. The esophagus (E) is visible in the azygoesophageal recess, also containing lymph nodes. Similarly,
the internal mammary artery and vein (IMA&V) indicate the location of the internal mammary chain of lymph nodes. The hemiazygos
vein (HAzV) is associated with paravertebral lymph nodes. RV, right ventricle; LV, left ventricle; RA, right atrium; LA, left atrium; RIPV, right
inferior pulmonary vein; LIPV, left inferior pulmonary vein. D: Cardiac chambers and paracardiac mediastinum, 2.5 mm below C. RV, right
ventricle; LV, left ventricle; VS, ventricular septum; RA, right atrium; LA, left atrium; RIPV, right inferior pulmonary vein; LIPV, left inferior
pulmonary vein; E, esophagus. E: Cardiac chambers and paracardiac mediastinum, 2.5 mm below D. RV, right ventricle; LV, left ventricle; VS,
ventricular septum; RA, right atrium; CS, coronary sinus; E, esophagus; AzV, azygos vein; HAzV, hemiazygos vein. F: Cardiac chambers
and paracardiac mediastinum, 5 mm below E. RV, right ventricle; LV, left ventricle; IVC, inferior vena cava; AzV, azygos vein; HazV, hemiazygos
vein. Lymph nodes may be seen in relation to the azygos and hemiazygos veins, but are not visible in this normal subject. G: Cardiac
chambers and paracardiac mediastinum, 3.75 mm below E. RV, right ventricle; LV, left ventricle; Per, pericardium; IVC, inferior vena cava; Ao,
descending aorta; E, esophagus; L, liver dome. Lymph nodes may be seen in relation to the pericardium and diaphragm, but are not visible in
this normal subject. The pericardium is usually visible as a thin line, a few millimeters in thickness and separated from the myocardium by
a layer of fat.
crosses to the left side near the T6 or T7 vertebral body, and descending aorta. Its mean longitudinal, anteropos-
lying along the left lateral wall of the esophagus, adjacent terior, and transverse diameters measured in one study
to the descending aorta, and terminates near the junction were 33.5 1.7 (SD) mm, 5.2 0.13 mm, and 5.2
of the left internal jugular and subclavian veins (see 0.15 mm, respectively (74).
Fig. 4-117) (70–72). The thoracic duct measures a few The pericardium is often visible anterior and to the left of
millimeters in diameter in normal subjects. It is difficult the heart, and normally measures 3 mm or less, although its
to recognize on CT and distinguish from other vascular visibility and appearance vary with slice thickness. Using
structures unless it is dilated (71). When visible on CT, it rapid scan times to reduce cardiac pulsation artifacts, the
is low in attenuation. The thoracic duct is commonly normal pericardium appears 2 mm or less in thickness in
visible on unenhanced MRI obtained with respiratory 95% of subjects (76). Using 1-mm high-resolution CT
gating and using a 3D, half-Fourier, fast SE sequence. (HRCT) slices, the upper limit of the thinnest portion of
Using this technique, the thoracic duct has been reported the normal pericardium (mean value 2 SD) measured
to have a maximum diameter of 3.7 mm (SD 0.81 mm) 0.7 mm (77). Focal thickening or a localized fluid collection
in normal subjects (73). is often visible in normal persons posterior to the lower
Caudally, the thoracic duct arises from the cisterna sternum (78,79). Pericardial recesses and sinuses are also
chili or a confluence of lymphatic vessels in the upper commonly visible using CT or MRI (80).
abdomen. Using T2-weighted images obtained with a
single-shot fast SE technique, an abdominal confluence
Sagittal and Coronal Planes
of lymphatics was visible anterior to the spine in 96% of
subjects as a very intense structure (74). On routine MRI, Although cross-sectional imaging of the mediastinum
it is less often seen (75). It was located at the level of (Figs. 4-4 to 4-7) is generally adequate to visualize most
L1–2 in one third of cases and in the midline in 70%. It pathologic processes, sagittal and coronal imaging using
most often appears tubular or sausage shaped, but it may MRI (Figs. 4-8 to 4-11) or reformatted spiral CT can some-
also be recognized as parallel or converging tubular struc- times add to our understanding of the origin and extent of
tures, as a focal round or oval fluid collection, or as a disease. Utilization of multiplanar images, however,
plexus of lymphatic channels (75). It is typically visible requires a detailed knowledge of normal relationships
in a retrocrural location, lying between the azygos vein between mediastinal structures in both standard coronal
A B
C D
Figure 4-8 Coronal reformations in the same normal subject as shown in Figures 4-4, 4-6, and 4-7. A: Oblique coronal reformation in
the plane of the trachea and main bronchi. T, trachea; C, carina; RSCV, right subclavian vein; RSCA, right subclavian artery; LIJV, left inter-
nal jugular vein; LCCA, left common carotid artery; AzA, azygos arch; RUL, right upper lobe bronchus; BI, bronchus intermedius; RLLPA,
right lower lobe pulmonary artery; RIPV, right inferior pulmonary vein; LA; left atrium; LIPV, left inferior pulmonary vein; LUL, left upper
lobe bronchus; LPA, left pulmonary artery; RPA, right pulmonary artery; APW, aortopulmonary window; Ao, aorta. The fat-filled subcarinal
space lies between the carina and right pulmonary artery. B: Coronal reformation anterior to the trachea (T) primarily shows great vessels.
C, clavicle; IJ, internal jugular vein; RSCV, right subclavian vein; LSCV, left subclavian vein; RBCV, left brachiocephalic vein; LBCV, left
brachiocephalic vein; IA, innominate artery; SVC, superior vena cava; Ao, ascending aorta; AR, aortic root; PA, pulmonary artery; LAA, left
atrial appendage; RA, right atrium; RV, right ventricle; LV, left ventricle. Serration of the cardiac and vascular margins is due to pulsation.
C: Coronal reformation through the distal trachea (T) and carina. E, esophagus; LSCA, left subclavian artery; Ao, aorta; AzA, azygos arch;
APW, aortopulmonary window; LPA, left pulmonary artery; RPA, right pulmonary artery; Ao, aortic arch; RSPV, right superior pulmonary
vein; LSPV, left superior pulmonary vein; LA, left atrium; RA, right atrium; RV, right ventricle; LV, left ventricle. D: Coronal reformation
through the descending aorta (Ao). E, esophagus; LSCA, left subclavian artery; LIPA, left interlobar pulmonary artery; LIPV, left inferior
pulmonary vein.
A, B C
D E
Figure 4-9 Sagittal reformations, from right to left, in the same normal subject as shown in Figures 4-5, 4-7, and 4-8. A: Reformation
through the right mediastinum and superior vena cava (SVC). C, clavicle; IA, innominate artery; RBCV, right brachiocephalic vein; RPA, right
pulmonary artery; RPV, right pulmonary veins; RAA, right atrial appendage; RA, right atrium; RV, right ventricle; IVC, inferior vena cava. B:
Reformation through the trachea (T), slightly to the left of A. IA, innominate artery; LBCV, left brachiocephalic vein; PA, pulmonary artery;
RPA, right pulmonary artery; RV, right ventricle; Ao, ascending aorta; RV, right ventricle; LV, left ventricle; LA, left atrium. Anterior medi-
astinal soft tissues are visible anterior to the aorta. The subcarinal space (SCC) is visible inferior to the carina (C) and the pretracheal space
(*) is anterior to the distal trachea. C: Reformation to the left of B. LC, left common carotid artery; LBCV, left brachiocephalic vein; Ao,
aorta; E, esophagus; LMB, left main bronchus; PA, pulmonary artery; RV, right ventricle; RV, right ventricle; LA, left atrium. D: Reformation
to the left of C. C, clavicle. LIJV, left internal jugular vein; LSA, left subclavian artery; Ao, aorta; LMB, left main bronchus; PA, pulmonary
artery; RV, right ventricle; LA, left atrium; LV, left ventricle. E: Oblique reformation through the aorta and its branches, 7 mm maximum
intensity projection image. Ao, aorta; LSA, left subclavian artery; LCCA, left common carotid artery; IA, innominate artery; RV, right ventri-
cle; LA, left atrium.
and sagittal planes. Once familiar, these images can be and identification of the structure from which it is arising;
augmented by use of select parasagittal planes that maxi- whether it is single, multifocal (involving several different
mize visualization of select regions, such as the APW. areas or lymph node groups), or diffuse; its size and shape
and its attenuation (fatty, fluid, soft tissue, or a combina-
tion of these); the presence of calcification and its character
DIAGNOSIS OF MEDIASTINAL MASS and amount; and its opacification following administra-
tion of contrast agents (8).
The differential diagnosis of a mediastinal mass on CT is The mediastinum is divided anatomically into supe-
usually based on several findings, including its location rior, anterior, middle, and posterior compartments, and
A B
C D
E F
Figure 4-10 Mediastinal anatomy: coronal MR images. A–H: Sequential coronal MR images from anterior to posterior. Identification of
individual structures is easiest when comparison is made to the sections immediately adjacent, limiting potential problems of interpretation
caused by partial volume averaging. Note that the thyroid gland is easily identified adjacent to the trachea in A and B. Ao, aorta; RA, right
atrium; LV, left ventricle; MPA, main pulmonary artery; BA, brachiocephalic artery; RBV, right brachiocephalic vein; LBV, left brachiocephalic
vein; Tr, trachea; LCCA, left common carotid artery; SVC, superior vena cava; RSPV, right superior pulmonary vein; RIPA, right interlobar
pulmonary artery; RSA, right subclavian artery; LSA, left subclavian artery; TA, truncus anterior; RMPA, right main pulmonary artery; LMPA,
left main pulmonary artery; LSPV, left superior pulmonary vein; APW, aorticopulmonary window; RULB, right upper lobe bronchus; LMB, left
mainstem bronchus; RMB, right mainstem bronchus; LA, left atrium; RMPA, right main pulmonary artery; LULB, left upper lobe bronchus;
A-PSB, apical-posterior segmental bronchus; DAo, descending aorta.
G H
A B
C D
Figure 4-11 Mediastinal anatomy: parasagittal MR images. A–E: Sequential parasagittal images through the mediastinum, from the level
of the aortic arch to the right hilum, respectively. Familiarity with this particular plane is especially important, as it intersects both the
ascending and descending aorta, thus allowing visualization of the entire aorta in a single plane (see A). Ao, ascending aorta; BCA, brachio-
cephalic artery; RA, right atrium; LA, left atrium; RMPA, right main pulmonary artery; PV, left superior pulmonary vein; DAo, descending
aorta; Tr, trachea; SVC, superior vena cava; RSPV, right superior pulmonary vein; TA, truncus anterior; RMB, right main-stem bronchus.
F: True sagittal MR image through the main pulmonary artery (MPA). Note that in this plane, the relationship between the MPA and the right
ventricular outflow tract is especially well seen. As shown in this example, occasionally a section at this level allows visualization of the pul-
monic valves (continued ).
E F
localizing a mediastinal mass to one of these divisions can cell tumors, and lymph node metastases; lymphangioma
facilitate their differential diagnosis on conventional radi- and hemangioma; neurogenic tumors; thoracic meningo-
ographs (81). On CT, however, it is more appropriate to cele; fluid collections such as mediastinal abscess, media-
base the differential diagnosis of a mediastinal mass on a stinitis, hematoma, seroma, or pseudocyst; cystic goiter; a
direct observation of the tissue or structure from which the dilated, fluid-filled esophagus; and fluid localized in the
mass is arising (e.g., lymph nodes, veins and arteries, superior pericardial recess (32,84,92–94).
thymus, thyroid, parathyroid, trachea, esophagus, vertebral High-attenuation masses generally contain calcium, have
column) rather than its location in the anterior, middle, or high iodine content, or are related to the presence of fresh
posterior mediastinum. If this is not possible, then localiz- blood (Table 4-4). They include calcified lymph nodes sec-
ing the mass to specific regions of the mediastinum ondary to granulomatous infection, sarcoidosis, inhalational
(e.g., prevascular space, pretracheal space, subcarinal space, diseases such as silicosis, and Pneumocystis carinii (now called
APW, anterior cardiophrenic angle, paraspinal region) can Pneumocystis jiroveci) infection; lymph node calcification
be valuable in differential diagnosis (Table 4-1). However, caused by metastatic tumor such as adenocarcinoma or sar-
it is also important to keep in mind that, although many coma; lymphoma with calcification usually after treatment;
pathologic processes have a typical location or locations, partially calcified primary neoplasms including germ-cell
most can be seen in any part of the mediastinum (82). tumors, thymoma, and neurogenic tumors; calcified goiter;
The attenuation of mediastinal masses is also of pri- calcified vascular lesions; calcium within a cyst; goiter with
mary importance in their differential diagnosis (83–86). high iodine content; and mediastinal hematoma (83).
Masses can be categorized as (a) fat attenuation; (b) low Enhancing masses are highly vascular (Table 4-5) (95)
attenuation, having density greater than that of fat, but and include substernal thyroid and parathyroid glands
less than that of muscle; (c) high attenuation, with a (96,97); carcinoid tumors; lymphangiomas and heman-
density greater than that of muscle; and (d) enhancing, giomas that may contain vascular elements (98,99);
showing a significant increase in attenuation following paraganglioma (100,101); Castleman disease (35,102);
the injection of contrast. and a variety of partially necrotic, inflammatory, and
Fat attenuation masses, composed primarily of or par- neoplastic lesions in which some rim enhancement is
tially containing fat or lipid-rich tissues, include lipomato- visible (32,33).
sis and fat pads; thymolipoma; teratoma; lymphangioma
and hemangioma; lipoma, liposarcoma, and other tumors
derived from fat; fatty hernias; and, rarely, lymph node MASSES PRIMARILY INVOLVING
enlargement in Whipple disease or extramedullary THE PREVASCULAR SPACE
hematopoiesis (Table 4-2) (85,87–91).
Low-attenuation masses are usually cystic or fluid filled Mediastinal masses result in alterations of normal medi-
or contain some lipid (Table 4-3). These include a variety astinal contours and displacement or compression of
of congenital or acquired cysts (bronchogenic, esophageal mediastinal structures; recognizing these findings can be
duplication, neurenteric, pericardial, and thymic); necrotic valuable in diagnosis and in suggesting the site of origin of
neoplasms, including lymphoma, cystic thymoma, germ- the mass.
TABLE 4-1
DIFFERENTIAL DIAGNOSIS OF MEDIASTINAL MASSES BASED ON COMMON SITES OF ORIGIN
Prevascular Space Sarcoidosis
Thymic masses Lymphoma
Hyperplasia Metastases
Thymoma Infections (e.g., TB)
Thymic carcinoma Mesenchymal masses (e.g., lipomatosis, lipoma)
Thymic carcinoid tumor Vascular abnormalities (aorta or pulmonary artery)
Thymolipoma Chemodectoma
Thymic cyst Foregut cyst
Thymic lymphoma and metastases Subcarinal Space, Azygoesophageal Recess
Germ-cell tumors Lymph node masses
Teratoma Lung carcinoma
Seminoma Sarcoidosis
Nonseminomatous germ-cell tumors Lymphoma
Thyroid abnormalities Metastases
Parathyroid tumor Infections (e.g., TB)
Lymph node masses (particularly Hodgkin lymphoma) Foregut cyst
Vascular abnormalities (aorta and great vessels) Dilated azygos vein
Fatty masses (lipomatosis, lipoma, liposarcoma) Esophageal masses
Foregut cyst Varices
Vascular tumors and masses (lymphangioma, hemangioma, Hernia
angiosarcoma) Paravertebral Region
Tumors and masses of fibroblastic origin (solitary fibrous tumor, Neurogenic tumor
fibrosarcoma) Nerve sheath tumors
Tumors of muscle origin (leiomyosarcoma) Sympathetic ganglia tumors
Tumors of bone and cartilage (e.g., osteosarcoma, chondroma, Paraganglioma
chondrosarcoma) Foregut cyst
Cardiophrenic Angle Meningocele
Lymph node masses (particularly lymphoma and metastases) Extramedullary hematopoiesis
Pericardial cyst Pseudocyst
Morgagni hernia Thoracic spine abnormalities
Thymic masses Hernias
Germ-cell tumors Esophageal masses
Pretracheal Space Varices
Lymph node masses Lipomatosis
Lung carcinoma Lymph node masses
Sarcoidosis Lymphoma (particularly non-Hodgkin)
Lymphoma (particularly Hodgkin disease) Metastases
Metastases Dilated azygos or hemiazygos vein
Infections (e.g., TB) Dilated thoracic duct or thoracic duct cyst
Foregut cyst Hernia
Tracheal tumor Lymphangioma and hemangioma
Mesenchymal masses (e.g., lipomatosis, lipoma) Thymic mass or germ-cell tumor
Thyroid abnormalities Tumors and masses of fibroblastic origin (mediastinal fibromatosis)
Vascular abnormalities (aorta and great vessels) Tumors of muscle origin (leiomyoma, leiomyosarcoma)
Lymphangioma and hemangioma Tumors of bone and cartilage (e.g., osteosarcoma, chondroma,
Aorticopulmonary Window chondrosarcoma)
Lymph node masses
Lung carcinoma
TB, tuberculosis.
TABLE 4-2 TABLE 4-4

FAT-CONTAINING MASSES: DIFFERENTIAL HIGH-ATTENUATION MASSES: DIFFERENTIAL
DIAGNOSIS DIAGNOSIS
Lipomatosis Calcified lymph nodes Metastatic tumor
Fat pads Granulomatous infection Adenocarcinoma (mucinous)
Thymolipoma Sarcoidosis Thyroid carcinoma
Teratoma Silicosis and coal workers’ Sarcoma
Lymphangioma and hemangioma pneumoconiosis Lymphoma with calcification,
Lipoma, liposarcoma, and other tumors derived from fat Pneumocystis carinii infection usually after treatment
Fat-containing hernias Calcified primary neoplasms Amyloidosis
Treated extramedullary hematopoiesis Germ-cell tumors Calcified vascular lesions
Lymph node enlargement in Whipple disease Thymoma Calcium within a cyst
Neurogenic tumors Goiter with high iodine
Goiter or thyroid tumor content
Hematoma
Masses in the prevascular space, when large, tend to dis-
place the aorta and great arterial branches posteriorly, but
distinct compression or narrowing of these relatively thick
masses, fatty masses, and lymphangioma (hygroma)
walled structures is unusual. Within the supra-aortic medi-
(81,93,103,104).
astinum, displacement, compression, or obstruction of the
CT is advantageous in the diagnosis of masses in the
brachiocephalic veins is common. Posterior displacement
prevascular mediastinum (92,105). In a study of 128 pa-
or compression of the superior vena cava is typical only
tients, the correct first choice diagnosis was made in
with large or right-sided masses. On the left, compression
48% of cases, based on CT, although this was possible
of the main pulmonary artery can be seen. Recognizing
in only 36% of cases based on plain radiographs
abnormalities of mediastinal contours is not particularly
(105). Those masses most accurately diagnosed were
helpful in diagnosing masses in this region. The anterior
cystic in nature, contained* fat, or had characteristic
junction line is rather variable in thickness and cannot
enhancement patterns, including benign germ-cell
be relied on as being abnormal unless grossly thickened.
tumors, thymolipoma, omental hernia, and tuberculous
The mediastinal pleural reflections bordering the prevascu-
lymphadenopathy.
lar space are often convex laterally, although a marked
convexity may suggest a mass. The differential diagnosis of
masses arising in this area include thymoma and other
NORMAL THYMUS
thymic tumors, cysts, lymphoma and other lymph node
masses, germinal-cell tumors, thyroid masses, parathyroid
The thymus has two lobes that are fused superiorly near
the thyroid gland and smoothly molded to the anterior
aspect of the great vessels at lower levels. It occupies
the thyropericardic space of the anterior mediastinum
TABLE 4-3 and extends inferiorly to the base of the heart. The
LOW-ATTENUATION MASSES: DIFFERENTIAL thymus is very rarely found in an ectopic location, usu-
ally the neck (106).
DIAGNOSIS
Congenital or acquired cysts Neurogenic tumors
Bronchogenic Meningocele
Esophageal duplication Fluid collections
Neurenteric Abscess
Pericardial Mediastinitis TABLE 4-5
Thymic Hematoma ENHANCING MASSES: DIFFERENTIAL
Thoracic duct cyst or dilated Seroma DIAGNOSIS
thoracic duct Pseudocyst
Necrotic neoplasms Dilated esophagus Thyroid and parathyroid glands
Lymphoma Achalasia Tumors, hyperplasia, enlargement
Thymoma Scleroderma Carcinoid and atypical carcinoid tumors
Germ-cell tumors Stricture Paraganglioma
Lymph node metastases Carcinoma with Lymphangiomas and hemangiomas
Goiter or thyroid tumor obstruction Castleman disease
Lymphangioma and Pericardial recesses and Inflammatory disease with enhancing lymph nodes
hemangioma pericardial effusion Metastatic tumor
The thymus weighs an average of 22 13 g at birth and patients, thymus may appear similar in intensity to fat.
increases in size progressively to reach a maximum weight at Homogeneous enhancement occurs after administration
puberty of approximately 34 g (107). Beginning at puberty, of gadolinium.
the thymus begins to involute, and this process continues Thymic size can be quantitated using its length (meas-
for a period of 5 to 15 years. During thymic involution, ured in the cephalocaudal dimension), width (measured in
atrophied thymic follicles are progressively replaced by fat, the transverse dimension), and thickness (measured per-
and the relative proportion of fat to thymic tissue increases pendicular to its length and anteroposterior dimension)
progressively, until after age 60. At this age, remaining (Figs. 4-14 and 4-15). Although thymic size increases with
thymic tissue is negligible (108,109). For example, in an age in young subjects, its width shows little change. On CT,
autopsy series of 20 patients older than 60 years, no thymus the average thickness of the thymus decreases with age, from
could be recognized with gross examination and only 11 an average of 1.4 cm in children aged 5 years or less, to an
had any thymic tissue visible histologically (108). average of 1.0 cm in children aged 10 to 19 years (39).
CT allows imaging of the thymus with much greater In children, CT (39,56) and MRI (54,58) provide some-
clarity than is possible using plain radiographs. The CT what different normal values for thymic size, a finding
appearance of the normal thymus has been well described that likely reflects the fact that MRI scans are obtained
in both adults and children (39,53,55–57). In interpreting during quiet breathing, whereas CT, especially in older
CT, one must be aware of the significant changes that children, is commonly performed at full inspiration (54).
occur in thymic morphology with age and the significant As compared to CT, MRI in children usually shows a
variation in the normal size and weight of the thymus seen slightly greater thymic width but a marked decrease in the
in younger individuals, particularly those younger than the ratio of the size of the thymus relative to both the trans-
age of 25 years. verse and cephalocaudal dimensions of the thorax. On
MRI, the thymus appears to have a slightly greater
thickness than noted on CT, with an average thickness of
1.8 cm for the right lobe and 2.1 cm for the left lobe.
Appearance of the Thymus in Children
Occasionally the signal intensity characteristics of normal
In children, the normal thymus occupies the retrosternal thymic tissue may be of value in identifying unusual con-
space, draping itself over the great vessels and cardiac figurations of the thymus, especially in children in whom
margins (56). Its left lobe usually extends laterally, adja- posterior mediastinal extension between the superior vena
cent to the arch of the aorta, and posteriorly to the level of cava and trachea is not uncommon (113).
the descending aorta, but distinct lobes are not generally
seen (110). Cephalad, the thymus extends an average of
1.7 cm superior to the innominate vein and can be seen as
Appearance of the Thymus in Adults
high as the thyroid gland, where it may simulate lymph
node enlargement (111). Its inferior extent is variable. In From puberty to the age of about 25 years, the thymus is
infancy, it is commonly seen at the level of the pulmonary recognizable as a distinct triangular or bilobed structure,
arteries or below, but its inferior extent decreases with age. usually outlined by mediastinal fat (Fig. 4-14). Typically
In infants and younger children, the thymus appears its borders are flat or concave laterally, but on occasion
quadrilateral in shape in the axial plane but assumes the thymus may show convex margins, a finding more
a more triangular shape as the child grows (Figs. 4-12 to typical in children (Fig. 4-12). Its CT attenuation usually
4-14). Its margins are sharp, smooth, and convex in infants decreases to less than that of muscle because of fatty
and often become straight in older children (Fig. 4-14) replacement. In persons older than the age of 25 years, the
(53,56,112). On CT, the thymus is of soft tissue attenua- thymus is no longer recognizable as a soft tissue structure,
tion, slightly denser than vessels but approximately the because of progressive fatty involution. Islands or strands
same attenuation as muscle. The mean attenuation of the of soft tissue density within a background of more abun-
thymus has been found to be 36 Hounsfield units (HU) dant fat are typically noted on CT (Fig. 4-16). The rapidity
(112). It is often possible to distinguish between thymus and degree of thymic involution are variable in different
and vessels even without contrast enhancement, but con- subjects, and occasionally the thymus may still be recog-
trast is required to optimally delineate the thymus, which nized as a discrete structure up to the age of 40. With com-
shows homogeneous enhancement of 20 to 30 HU after plete thymic involution, the anterior mediastinum appears
bolus contrast injection. to be entirely filled with fat. It should be realized, however,
MRI clearly distinguishes the thymus from mediastinal that most of the anterior mediastinal fat is contained
vascular structures (Figs. 4-12C–E and 4-13) (54,58). On within the fibrous skeleton of the thymus and may have
T1-weighted images, the signal intensity of the thymus is a CT density slightly higher than that of subcutaneous fat.
slightly greater than that of muscle; on T2-weighted Normal measurements for the thymus as shown on CT
images, the thymus increases in intensity relative to fat in adults have been reported (Fig. 4-15) (53,55). Although
but usually remains of intermediate intensity. In some values for the length (measured in the cephalocaudal
A B
C D
E
Figure 4-12 Normal thymus in a 12-month-old boy. A: Enhanced CT at the level of the carina. The thymus (arrows) is large and has con-
vex margins. It typically assumes a more triangular shape in older children. Its attenuation is similar to that of muscle. As is typical in young
children, the thymus occupies the retrosternal space, draping itself over the great vessels and cardiac margins. Its left lobe usually extends
laterally, adjacent to the arch of the aorta, and posteriorly to the level of the descending aorta. Distinct lobes are not generally seen at this
age. B: CT at a lower level, the thymus (arrows) extends posteriorly to the hila. C: T1-weighted MRI obtained at the same level as A follow-
ing administration of gadolinium shows the thymus to be slightly more intense than muscle. Enhancement is homogeneous. D: T2-weighted
image at the same level as C. The thymus remains of intermediate intensity. E: T1-weighted coronal image shows the inferior extent of the
thymus (arrows) adjacent to the superior vena cava (V) and right atrium (RA) on the right, and the pulmonary artery (P), left atrial appendage,
and left ventricle (LV) on the left.
A B
Figure 4-13 Normal thymus in a 12-year-old boy. A: T1-weighted transaxial

image (TR/TE 750/20 msec) of the mediastinum shows the homogeneous
intensity thymus (arrow) anterior to the great vessels. The right mediastinal
pleura is convex laterally. In this patient, the right thymic lobe predominates.
B: T2-weighted transaxial image (TR/TE 2,250/70 msec) shows an increase in
intensity of the thymus, as compared with mediastinal fat. C: T1-weighted
transaxial image (TR/TE 750/20 msec) in the coronal plane shows the thymus
occupying most of the superior mediastinum, extending inferiorly to the level
of the right atrium. The right thymic lobe is larger (large arrow), but the left
C thymic lobe (small arrow) is also visible.
dimension) and width (measured in the transverse dimen- Nicolaou et al. (114). Francis showed that, although
sion) of the thymus have been defined, the most helpful thymic thickness is a sensitive indicator of thymic abnor-
value for thymic size is “thickness,” measured perpendicu- mality, thymic shape was just as reliable in separating
lar to the length. In subjects younger than the age 20 years, normal from abnormal thymus, especially when the
a thickness of 1.8 cm is considered the maximum allow- thymus was multilobulated (55). In fact, qualitative assess-
able normal value, with 1.3 cm being the maximum ment of thymic shape alone should prove sufficient for
normal value in older subjects. The accuracy of these diagnosing mass lesions of the thymus. Nicolaou et al.
dimensions has been confirmed by Francis et al. (55) and found a sensitivity of 100% for diagnosing thymic hyper-
A B
Figure 4-14 Normal thymus in a teenager. A: Unenhanced CT at the level of the aortic arch (A) shows a triangular thymus (arrows) of soft
tissue attenuation. The thymic margins are straight rather than convex (see Fig. 4-12). B: At the level of the aortopulmonary window, the
thymus (large arrows) also appears triangular, with slight convexity of its left lobe. The small arrows indicate the thickness of the left thymic
lobe (see Fig. 4-15).
intensity on T1-weighted images, being less intense than

surrounding mediastinal fat but greater intensity than
muscle (Figs. 4-12C–E and 4-13). However, because of pro-
gressive thymic involution, its appearance is dependent on
the age of the patient. In patients older than 30 years, dif-
ferentiation between the thymus and adjacent mediastinal
fat may be difficult because of thymic involution and
replacement by fat. The T2 relaxation times of the thymus
are similar to those of fat at all ages (Fig. 4-17), making
visualization of the thymus more difficult than on T1-
weighted images (54,115). On MRI, thymic thickness can
appear considerably greater than that seen on CT, with a
mean value of up to 20 mm (54).
Positron emission tomography using fluorine-18 fluo-
rodeoxyglucose (FDG-PET) may show uptake by normal
Figure 4-15 Normal thymus, assessment of size. The thymus thymus (116,117). In young adults, the normal thymus
is measured using its thickness, its width (transverse dimension),
and its anteroposterior dimension. Its cephalocaudal extent is may appear metabolically active on FDG-PET, and its
termed length. (From Frances IR, Glazer GM, Shapiro B, et al. degree of activity correlates with its CT attenuation (118).
Complementary roles of CT and 131-I-MIBG scintigraphy in In a study by Nakahara et al. (118), 32 (34%) of 94 normal
diagnosing pheochromocytoma. AJR Am J Roentgenol. 1983;
141:719–725, with permission.) young adults (mean age 25.4 years, range 18 to 29 years)
showed thymic uptake of FDG. In these 32 cases, the count
ratio between the thymus and the lung (T/L ratio) was
2.86 0.49 (range 2.02 to 3.99). The CT attenuation of the
plasia or thymoma based on a thymic thickness of thymus averaged 17.5 45.7 HU (range 103.6 HU to
greater than 1.3 cm or a focal soft tissue density greater 79.9 HU) correlated with the T/L ratio (r 0.58). Also, CT
than 7 mm (114). thymic attenuation values in subjects with positive PET
On MRI, the normal thymus (54,58,115) characteristi- findings were significantly higher than CT values in sub-
cally appears homogeneous and of intermediate signal jects with negative PET findings (p .001).
A B
Figure 4-16 Normal thymus in a 51-year-old woman. A: Unenhanced CT at the level of the aortic arch shows a triangular thymus (arrows),
which is largely of fat attenuation, but contains small islands or strands of soft tissue. The thymic margins are slightly concave. B: At the level
of the aortopulmonary window and pulmonary artery, the thymus (arrows) is largely of fat attenuation, with strands of residual thymic tissue
within it.
cyst (n 1), or Hodgkin lymphoma (HL) (n 1). Similar

results have been reported by others (120). It must be kept
in mind, however, that thymic hyperplasia can also result
in a focal thymic mass, up to 5 cm in diameter (114). The
role of MRI in diagnosing thymic masses is more limited
(92). One important exception is in the diagnosis of vascu-
lar invasion, especially in patients for whom intravenous
contrast cannot be administered (121). In these cases, MRI
exquisitely delineates the extent of vascular involvement,
including identification of intracardiac disease extension.
Thymic masses and tumors include thymic hyperplasia,
thymoma, thymic carcinoma, thymic neuroendocrine
(carcinoid) tumor, thymic cyst, thymolipoma, and lym-
phoma (Table 4-7) (93,122). Overall, thymic tumors
account for approximately 20% of primary mediastinal
Figure 4-17 Relative T1 values of thymus and fat in different age tumors (104,123–125).
groups. The T1 of thymus decreases in older patients because of
FDG-PET has emerged as an important diagnostic tool
fatty replacement. The T1 of fat remains constant relative to age.
for the diagnosis and staging of thymic masses, but normal
thymic uptake of FDG complicates the assessment of
THYMIC ENLARGEMENT AND MASSES thymic lesions in children and young adults (117).
Increased thymic FDG uptake may be physiologic or may
The left thymic lobe is normally larger than the right and be seen in the presence of thymic hyperplasia, lymphoma-
the thymus is almost always slightly asymmetric. However, tous infiltration, primary thymic neoplasm, or metastatic
if enlargement of the thymus is grossly asymmetric or if tumor. In equivocal cases, correlation with morphologic
the thymus has a lobular contour, a thymic mass should data from CT or MRI is important in diagnosis (117).
be suspected. Using these criteria, CT has proven extremely
accurate in diagnosing thymoma and other thymic
neoplasms (Table 4-6). In a study reported by Chen et al. Thymic Enlargement
(119), CT proved 91% sensitive and 97% specific in estab- The thymus can weigh as much as 45 to 50 g in normal
lishing the presence of a mass within the thymus. Of 34 individuals, and it is difficult to diagnose mild degrees of
patients with a CT diagnosis of a mass or neoplasm, all but thymic enlargement. However, when the thymus appears as
one proved surgically to have a thymoma (n 31), thymic a mediastinal mass readily detectable on plain chest radi-
ographs in older children or young adults, it should be
considered enlarged.
TABLE 4-6 “Thymic enlargement” should be taken to mean that
DIFFERENTIATION OF THYMIC TUMOR the thymus is increased in size but maintains a normal
shape and appearance. This occurrence may be described
FROM NORMAL THYMUS: COMPUTED
using the term “hyperplasia,” but it is best to reserve this
TOMOGRAPHY CRITERIA
Thymic Tumor
1. Soft tissue mass in a patient older than 30 years.
2. Soft tissue mass is spherical or lobulated, not triangular or TABLE 4-7
arrowhead-shaped.
3. Unilateral or midline. THYMIC MASSES: DIFFERENTIAL DIAGNOSIS
4. Attenuation of mass equals or exceeds that of muscle
Thymic enlargement and hyperplasia
of chest wall.
Thymic rebound
5. Surrounded by fat (i.e., involuted thymus).
Thymoma (thymic epithelial tumor, WHO types A, AB, B1–3)
6. Calcification.
Thymic carcinoma (thymic epithelial tumor, WHO type C)
Normal Thymus Thymic neuroendocrine tumor
1. Patient younger than 20 years. Carcinoid
2. “Mass” is elongated with length width and triangular or Atypical carcinoid
arrowhead-shaped, as typical for a normal thymus. Small cell neuroendocrine carcinoma
3. “Mass” is bilateral, draped over the great vessels. Thymic cyst
4. “Mass” contains a large amount of fat in a patient older than 30 Thymolipoma
years. Lymphoma and leukemia
5. The “mass” occupies the thymic bed, with little additional fat Metastases
visible. Ectopic parathyroid adenoma or carcinoma
6. No calcification.
WHO, World Health Organization.
term for follicular thymic hyperplasia (FTH), described Thymic enlargement associated with LFH may be seen
later. There is overlap between thymic enlargement and in patients with human immunodeficiency virus (HIV)
FTH, and for practical purposes this distinction is often infection (129). LFH and multiloculated thymic cysts have
arbitrary. been reported in children with HIV infection (130,131).
Hyperthyroidism, Graves disease, and thyroid carcinoma In one study (130), eight HIV-infected children, 2.1 to
may be associated with thymic enlargement (107,126,127). 12.1 years of age, had an anterior mediastinal mass discov-
Less common associations with thymic enlargement are ered incidentally on chest radiography; these children had
acromegaly, Addison disease, and recovery from burns or CD4 cell counts between 102 and 733 cells/mm3. In all
chemotherapy (128). eight cases, CT revealed thymic enlargement with a multi-
cystic appearance. Histologic examination demonstrated
cystic changes, LFH, diffuse plasmacytosis, and multinu-
cleated giant cells. Follow-up showed a decrease in size of
Thymic Hyperplasia
the mass or resolution in five of the eight.
Thymic germinal or lymphoid follicular hyperplasia (LFH) Although thymic enlargement can be identified on CT
is a term used by pathologists to describe the presence of in some patients with LFH, the usefulness of CT in detect-
hyperplastic lymphoid germinal centers in the thymic ing this abnormality is limited (114,119,132). In one large
medulla, associated with a lymphocytic and plasma cell study correlating CT with pathologic findings, the thymus
infiltrate (107). The presence of LFH is commonly associ- typically appeared larger in patients with LFH than in age-
ated with myasthenia gravis, although lymphoid follicles matched controls. However, based on diffuse enlargement
can also be present in the medulla of the normal thymus, of the thymus, CT proved only 71% sensitive for diagnos-
particularly in young subjects. ing lymphoid hyperplasia (119). In the study by Nicolaou
On CT, 30% to 50% of patients with LFH can have et al. (114), only 55% of patients with LFH showed an
a normal-appearing thymus. In many of the remaining abnormal thymus on CT. Furthermore, Castleman and
patients, the thymus appears abnormally enlarged but gen- Norris (133) have shown that, in patients with myasthenia
erally has a normal shape (Fig. 4-18). In some patients with gravis, the weight of thymus glands showing LFH does not
LFH, the thymus appears nodular, masslike (Fig. 4-19), or differ significantly from that of normal controls. Up to
inhomogeneous in attenuation. In a study of 22 patients 50% of the thymus glands called normal on CT may prove
with myasthenia gravis and LFH (114), 10 (45%) had a to contain LFH after surgical removal (134).
normal-appearing thymus on CT, 7 (32%) had an enlarged On MRI, thymic hyperplasia is associated with thymic
thymus, and 5 (23%) had a focal thymic mass. Focal enlargement, but the signal intensity is the same as for
thymic masses ranged from 1.7 to 5 cm in diameter. normal thymus (115).
A B
Figure 4-18 Thymic enlargement and hyperplasia in hyperthyroidism. A, B: Enhanced CT at two levels in a 29-year-old woman shows
marked thymic enlargement, without evidence of a focal mass. Thyroid enlargement was also visible. Biopsy confirmed the diagnosis of
thymic hyperplasia.
(136), thymic volume decreased an average of 43% in

response to chemotherapy. On follow-up studies, however,
the thymus recovered its pretreatment volume in most
cases and, in 25% of patients, thymic rebound occurred,
with the volume of the thymus exceeding the baseline vol-
ume by 50%. Thymic rebound as visualized by CT has also
been reported after treatment of Cushing syndrome (137).
In most patients with extrathoracic malignancies,
thymic rebound is unlikely to cause concern or diagnostic
confusion. As a practical matter, thymic rebound is most
problematic when it is observed in patients with malignant
lymphoma who have been treated using chemotherapy.
When an enlarged thymus is seen in such patients, it can be
difficult to distinguish thymic rebound from recurrent
lymphoma. In our experience, however, when a patient
with lymphoma shows thymic enlargement as an isolated
finding after treatment (Fig. 4-20), with no adenopathy
being visible, it is reasonable to follow the patient with the
presumption that thymic rebound is responsible.
In a study by Kissin et al. (138) of adults receiving
chemotherapy, there was a suggestion that thymic rebound
Figure 4-19 Thymic mass in thymic hyperplasia. A sharply
marginated nodule (arrow) is visible within a thymus that is largely
may be a favorable prognostic factor, in that 13 of 14
replaced by fat in a 68-year-old woman. This represented thymic patients with thymic enlargement after chemotherapy were
hyperplasia. It is indistinguishable from thymoma. free of disease after a mean follow-up of 45 months (138).
On the other hand, thymic hyperplasia, manifested on CT by
an increase in thymic size or attenuation, has been reported
Thymic Rebound after high-dose chemotherapy and autologous stem cell
The thymus involutes during periods of stress, sustaining transplantation in patients with breast cancer but is uncom-
a decrease in volume that variably depends on the age of mon and does not appear to correlate with survival (139).
the patient and the severity and duration of the stress. This On MRI, patients with thymic rebound may show
phenomenon is most marked in children, but it has also enlargement of the thymus, but its signal intensity is the
been observed in young adults. Following involution, the same as for normal thymus (115).
thymus will generally reacquire its premorbid size within
several months of the stressful episode. It may also exhibit Thymoma and Thymic Epithelial Tumors
“rebound,” or growth to a size significantly larger than its
original size (Figs. 4-20 and 4-21). For example, Gelfand The most common primary thymic tumor is thymoma (140).
et al. (128) found enlargement of the thymus, as deter- It accounts for about 15% of primary mediastinal masses
mined using plain chest radiographs, in five children aged (140). Thymomas are derived from thymic epithelium, as are
5 to 12 who were recovering from burns. Also, largely thymic carcinomas (141). Thymoma is rare before age 20 and
using chest radiographs, Cohen et al. (135) illustrated is most common in patients aged 50 to 60 years. From 30%
rebound thymic growth in seven children aged 3 to 11 who to 54% of patients with a thymoma develop myasthenia
had received chemotherapy for 5 months to 3 years for gravis. Other syndromes associated with thymoma include
Wilms tumor, malignant lymphoma, osteosarcoma, malig- red cell aplasia and hypogammaglobulinemia.
nant teratoma, or acute lymphatic leukemia. In three cases, A number of systems have been used to classify thy-
a prominent thymus was observed during the course of moma and thymic epithelial tumors (141–143). Many of
chemotherapy; in the remaining four patients the changes these have fallen from favor because of their inability to
were detected 1 to 9 months after cessation of therapy. predict tumor behavior or prognosis.
Because CT is more sensitive than are plain chest radi- The simplest classification of thymic epithelial tumors
ographs for detecting thymic enlargement, thymic rebound considers three groups, based on a combination of histo-
is detected with greater frequency when patients are logic features and behavior: noninvasive thymoma, invasive
monitored with CT. Using CT, Choyke et al. (136) made thymoma, and thymic carcinoma. These three types, respec-
serial observations of the mediastinum in a group of tively, have also been termed benign thymoma (being
patients aged 2 to 35, who were receiving chemotherapy for noninvasive and having benign histologic features), type 1
various malignancies, including HL (n 6), osteosarcoma malignant thymoma (characterized by invasion but having
(n 5), testicular neoplasms (n 4), Wilms tumor benign histologic features), and type 2 malignant thymoma
(n 3), and rhabdomyosarcoma (n 2). In this study (appearing malignant histologically) (143).
A B
C D
E F
Figure 4-20 Thymic rebound in a 9-year-old following chemotherapy for non-Hodgkin lymphoma. A–C: CT at three levels following the
completion of chemotherapy. The thymus is small and difficult to see. D–F: CT 4 months later at the same levels shown in A to C. The thy-
mus has increased markedly in size due to thymic rebound. G–I: CT 1 year after scans D to F. No further chemotherapy has been used. The
thymus has decreased in size and appears normal for a patient of this age.
Suster and Moran (144) described a classification of tains some normal features of thymic morphology and lacks
thymic epithelial tumors based on histology and the degree the obvious malignant appearance of thymic carcinoma.
of cellular atypia, similar to that described previously. It Atypical thymoma is associated with local invasion (143).
permits the classification of thymic epithelial tumors Müller-Hermelink extended the histologic description
into three simple diagnostic categories: thymoma, atypical of thymic epithelial tumors to emphasize their morpho-
thymoma, and thymic carcinoma (144). Atypical thymoma logic features as cortical, medullary, or mixed, based on
is characterized histologically by features of low-grade their resemblance to cortical or medullary regions of the
malignancy, including cellular atypia, large cell size, occa- thymus (145–147). Cortical thymomas, as classified by
sional mitotic figures, and aggressive behavior, but main- Müller-Hermelink, tend to behave more aggressively than
G H
I
medullary or mixed tumors. The Müller-Hermelink histo- with a greater likelihood of complete resection and have a
logic classification has been shown to correlate with tumor better survival than types B2 and B3. For example, in a study
stage and survival (146). of 273 patients with thymoma (149) classified using the
Recently, the World Health Organization (WHO) has pro- WHO system, there was a significant relationship between
posed a histologic classification of thymomas based on the the tumor type, the frequency of invasion, and patient sur-
Müller-Hermelink system but using alpha-numeric des- vival. In patients with type A, AB, B1, B2, and B3 tumors,
ignations to describe the various tumor types (Table 4-8). the respective proportions of invasive tumors were 11.1%,
The WHO classification system is in wide use and has 41.6%, 47.3%, 69.1%, and 84.6%; the respective propor-
been adopted by most investigators. In the WHO system, tions of tumors with involvement of the great vessels were
thymomas are divided into two groups (type A and B) 0%, 3.9%, 7.3%, 17.5%, and 19.2%; and the respective 20-
depending on whether the neoplastic epithelial cells year survival rates were 100%, 87%, 91%, 59%, and 36%.
and their nuclei are spindle or oval shaped (type A) or have a Thymomas are staged at the time of surgery, based on
dendritic or epithelioid appearance (type B); type AB tumors the presence and extent of invasion (146,150,152). The
have mixed features. Type B tumors are further divided into presence or absence of growth through the tumor capsule
three subtypes (B1 to B3) according to the proportional and the extent of local invasion are fundamental in deter-
increase in their epithelial component and the presence of mining the stage and resultant prognosis.
atypical cells. All types of thymic carcinoma are classified as Overall, approximately 30% of all thymomas are inva-
type C. The relationship between the WHO classification sive, although this varies with the histologic classification
and the Suster and Moran and Müller-Hermelink classifica- of the tumor. Characteristically, invasive thymomas infil-
tion systems is summarized in Table 4-9 (145–147). trate adjacent structures or result in pleural or pericardial
This WHO classification of thymic epithelial tumors has implants; thymoma rarely metastasizes outside the thorax.
been shown to correlate significantly with tumor stage and Typically, thymomas (a) invade mediastinal fat or local
prognosis (148–151). Types A, AB, and B1 thymoma are less mediastinal structures, including the superior vena cava,
likely to be invasive or recur after surgery and are associated great vessels, and even the airways; (b) invade the adjacent
A B
Figure 4-21 Thymic rebound in a 30-year-old with colon cancer.

A: CT following chemotherapy, at the level of the left pulmonary
artery, shows a normal-appearing thymus with some fatty replace-
ment. B: Five months later, following completion of treatment, CT
at the same level shows a significant increase in thymic size and an
increase in its attenuation. The thymic margins are more convex
than on the prior study. C: CT at the same time as B, at a lower
C level, shows a prominent thymus.
lungs or chest wall; or (c) spread by contiguity along pleu- thymoma, invasion of adjacent lung was present in 9 of
ral reflections, usually on one side of the chest cavity, and 27 (33%) cases, and invasion of adjacent chest wall was
may seed the diaphragmatic surfaces with consequent present in 2 of 27 (8%) cases.
direct extension into the abdomen (Fig. 4-22) (141,142, The thymoma staging system proposed by Masaoka
153–157). In one study (158), invasion of mediastinal (152) is in general use. The principal features of the
structures was present in 10 of 27 (37%) patients with Masaoka staging system are as follows:
invasive thymoma, including 5 cases (19%) with invasion
of the pericardium, 1 with vascular invasion, and 4 (15%) Stage I: the tumor capsule is intact.
with both pericardial and vascular invasion. Invasion of Stage II: invasion of the tumor capsule or surrounding
the pleura was found in 24 of 27 (89%) cases of invasive fat or mediastinal pleura.
TABLE 4-8
WORLD HEALTH ORGANIZATION CLASSIFICATION OF THYMIC EPITHELIEAL TUMORS
Type Definition
A A tumor comprised of a homogenous population of neoplastic epithelial cells with spindle/oval shape, lacking nuclear
atypia, and accompanied by few or no nonneoplastic lymphocytes
AB A tumor in which foci with the features of type A thymoma are admixed with foci rich in lymphocytes; the segregation
of two patterns can be sharp or indistinct
B1 A tumor that resembles the normal functional thymus in that it combines large expanses with an appearance practically
indistinguishable from that of normal thymic cortex with areas resembling thymic medulla
B2 A tumor in which the neoplastic epithelial component appears as scattered plump cells with vesicular nuclei and
distinct nucleoli among a heavy population of lymphocytes; perivascular spaces are common
B3 A tumor comprised predominantly of epithelial cells with a round or polygonal shape and exhibiting mild atypia
admixed with a minor component of lymphocytes; foci of squamous metaplasia and perivascular spaces are common
C Tumor exhibiting clear-cut cytologic atypia and lacking a significant number of immature interepithelial thymocytes.
Mature lymphocytes and plasma cells are present in the septa between tumor lobules and in the tumor periphery.
This subtype is usually indistinguishable from extrathymic carcinomas
Stage III: invasion of surrounding organs (pericardium, apy may be used in patients with incomplete resection or
great vessels, or lung). distant metastases (159–164).
Stage IVa: pleural or pericardial implants.
Stage IVb: hematogenous or lymphogenous metastases.
Computed Tomography Findings
There is a relationship between WHO tumor type and
of Thymoma
stage at diagnosis. WHO tumor types A, AB, and B1 are less
likely to be invasive (i.e., stages II to IV) than types B2 and The CT appearance of thymomas has been well described
B3. Recent studies (146,149–151) have shown that, even (55,114,119,120,143,158,165–172). Approximately 80%
though the tumor stage and WHO classification are related, of thymomas occur at the base of the heart (140).
they are both important and independent factors in deter- Small thymomas are often subtle or invisible on chest
mining patient survival (Table 4-10). In a recent study radiographs (173).
(149), the 20-year survival rates for thymoma were 89%, On CT, thymomas appear as homogeneous soft tissue
91%, 49%, 0%, and 0%, respectively, in patients with attenuation masses, which are usually sharply demarcated
Masaoka Stage I, II, III, IVa, and IVb disease. and oval, round, or lobulated and do not conform to the
Resection is indicated in all stages, with supplemental normal shape of the thymus (Figs. 4-23 to 4-25) (140).
radiotherapy in patients with invasion through the tumor Most often the tumor grows asymmetrically to one side of
capsule or incomplete resection. For bulky or unresectable the anterior mediastinum and prevascular space (Figs. 4-23
tumors, preoperative chemotherapy and postoperative to 4-25). Because ectopic thymic tissue in the neck is
radiotherapy may be used. Radiotherapy and chemother- found in up to 21% of normal subjects, thymoma can
TABLE 4-9
RELATIONSHIPS AMONG HISTOLOGIC CLASSIFICATIONS OF THYMIC EPITHELIAL TUMORS
WHO Classification Müller-Hermelink Classification Suster-Moran Classification
Type A thymoma Medullary thymoma Thymoma

Type AB thymoma Mixed thymoma Thymoma
Type B1 thymoma Predominantly cortical thymoma Thymoma
Type B2 thymoma Cortical thymoma Thymoma
Type B3 thymoma Well-differentiated thymic carcinoma Atypical thymoma
Type C thymoma (thymic carcinoma) Malignant thymoma, category II Thymic carcinoma

A B
Figure 4-22 Invasive thymoma: pattern of spread. A, B: The pattern of spread of invasive thymomas is graphically represented. In cross-
section, the tumor can directly invade surrounding structures (broken arrows) or insinuate itself between the chest wall or mediastinum and
the parietal pleura (solid arrows). Once the para-aortic—paraspinal area is reached, inferior extension through the aortic or esophageal hiatus
can occur. With CT, such extension can be seen reliably and the study should always be extended to the upper abdomen. Ao, aorta; PA, pul-
monary artery; PV, pulmonary vein; Br, bronchus; E, esophagus; SVC, superior vena cava. (From Zerhouni EA, Scott WW, Baker RR, et al.
Invasive thymomas: diagnosis and evaluation by computed tomography. J Comput Assist Tomogr. 1982;6:92–100, with permission.)
occur in the neck or at the thoracic inlet, thus mimicking The presence of ipsilateral pleural effusion or pleural
a thyroid mass (Fig. 4-26) (174). Rarely, thymomas may nodules or masses may also be used to predict invasion
appear cystic, with discrete nodular components or in asso- (Fig. 4-29). Pleural nodules, when present, are often
ciation with a thick cyst wall or discrete mass (Figs. 4-27 unilateral and, unlike other tumors resulting in pleural
and 4-28). Except in patients with cystic masses, tumors metastases, may be unassociated with pleural effusion.
usually enhance homogeneously after contrast medium Thus, CT can show focal, well-defined pleural masses, unob-
injection. Focal calcifications are seen in 10% to 40% scured by pleural fluid.
(Figs. 4-27 and 4-29). As a general rule, caution should be used when diag-
In a patient with thymoma, CT findings that may indi- nosing invasion based on CT findings. In one study (158),
cate invasion include, in order of increasing sensitivity and a correct first choice diagnosis of invasive or noninvasive
decreasing specificity, (a) invasion of mediastinal structures, thymoma was made in only 38 of 50 (76%) cases. These
(b) infiltration of mediastinal fat, (c) poor definition of the included 21 of 27 (78%) cases of invasive thymoma and
tumor margin, (d) distortion or compression of mediastinal 17 of 23 (74%) cases of noninvasive thymoma.
structures, (e) obliteration of fat planes between the tumor The absence of fat planes between the thymoma and
and mediastinal structures, and (f) extensive contact mediastinal structures or direct contact between the tumor
between the tumor and mediastinal structures. and adjacent structures is often used to predict invasion
TABLE 4-10
RELATIONSHIP OF THYMOMA HISTOLOGY TO STAGE, SURGICAL RESULTS, AND SURVIVAL
WHO Classification % Stage I Average Stage Surgical Failure (%) 20-Year Survival (%)
Type A thymoma 89 1.17 0 100

Type AB thymoma 58 1.53 4.5 87
Type B1 thymoma 53 1.71 8.5 91
Type B2 thymoma 31 2.33 18.1 59
Type B3 thymoma 15 2.46 28.6 36

Modified from Okumura M, Ohta M, Tateyama H, et al. The World Health Organization histologic classification system reflects the oncologic behavior
of thymoma: a clinical study of 273 patients. Cancer. 2002;94:624–632, with permission.
A B
Figure 4-23 Noninvasive thymoma. A: A large homogeneous mass (arrows in A and B) is visible in the prevascular space on a contrast-
enhanced helical CT scan. It extends preferentially to the right side. B: It is separated from the ascending aorta by a well-defined fat plane,
suggesting that it is noninvasive.
but are not reliable criteria. In a study by Tomiyama et al.

(158), the CT appearance of partial or complete obliteration
of fat planes adjacent to the tumor was not helpful in the
diagnosis of invasion. Although infiltration of mediastinal
structures was suggested on CT in all patients with surgi-
cally proven mediastinal infiltration, it was also thought to
be present in many patients with noninvasive thymoma.
On the other hand, clear delineation of fat planes
surrounding a thymoma on CT should be interpreted as
indicating an absence of extensive local invasion, espe-
cially when the study is performed with thin collimation
following a bolus of intravenous contrast medium (120).
However, capsular invasion may be present without
disruption of adjacent fat planes.
The presence of additional CT findings may be used to
predict invasion (Figs. 4-22, 4-29 to 4-33). These include
large tumor size, lobulation, irregular contours, heteroge-
neous opacification, areas of low attenuation, and calcifica-
tion. In a study by Tomiyama et al. (158), the CT scans of
27 patients with invasive thymoma and 23 patients with
noninvasive thymoma were compared. Invasive thymomas
were significantly larger (mean long-axis diameter 5.50 cm)
than noninvasive thymomas (long axis diameter 3.95 cm)
(p .05). Also, invasive thymomas were more likely to
have lobulated (16/27, 59%) or irregular (6/27, 22%) con-
tours than noninvasive thymomas (8/23, 35% and 1.5/23,
6%, respectively) (p .05). Homogeneous enhancement
was found in 11 of 27 (44%) cases of invasive thymoma
Figure 4-24 Noninvasive thymoma. A focal, sharply marginated, and 15 of 23 (79%) cases of noninvasive thymoma
rounded mass is visible in the prevascular space. The right and left
brachiocephalic veins are displaced posteriorly, but a fat plane (p .05). Invasive thymomas had a higher prevalence of
separating the thymoma and vessels is visible. focal low-attenuation areas seen on CT (16/27, 60%) than
Figure 4-25 Noninvasive thymoma: CT and MR findings. A: CT section

at the level of the right main pulmonary artery following a bolus of intra-
venous contrast media. A well-defined rounded mass of soft tissue atten-
uation is present on the right side (arrow). B, C: T1- and T2-weighted
coronal images through the mass, respectively. On the T1-weighted
image, fat clearly marginates both the superior and the inferior borders
of the mass (arrows in B), indicative of a mediastinal origin. The margins
of the mass are sharply marginated. On the T2-weighted image (C), there
A is marked increase in the relative signal of the lesion.
B C
noninvasive thymomas (5/23, 22%) (p .001) as well as tour, areas of low attenuation, and multifocal calcification
foci of calcification (14.5/27, 54% vs. 6/23, 26%; p .01). was seen in 5 of 27 (19%) cases of invasive thymoma
In this study, the combination of lobulated or irregular and no cases of noninvasive thymoma. The combination of
tumor contour and areas of low attenuation was seen in smooth contour and homogeneous enhancement was
15 of 27 (56%) cases of invasive thymoma and 2.5 of
23 (11%) cases of noninvasive thymoma. The combination
of lobulated or irregular contour, areas of low attenuation,
and calcification was seen in 7.5 of 27 (28%) cases of
invasive thymoma and 1.5 of 23 (7%) cases of noninvasive
thymoma. The combination of lobulated or irregular con-
Figure 4-27 Cystic thymoma. Thymoma with cystic components

(arrows) and calcification is visible in the right paracardiac medi-
Figure 4-26 Cervical thymoma. A noninvasive thymoma (arrow) is astinum. A thick wall (small arrow) distinguishes this abnormality
visible anterior and lateral to the trachea in the neck. from a simple cyst.
A B
Figure 4-28 Cystic thymoma. (A) Contrast-enhanced CT scan shows a well-defined partially cystic mass in the anterior mediastinum
within which distinct soft tissue nodules can be seen. Biopsy proved cystic thymoma. (B) MR image in a patient different from the patient in
A shows a heterogeneous high signal intensity mass in the anterior mediastinum within which discrete nodules of intermediate signal
intensity can be identified (arrows). Although most thymomas are solid, occasional thymomas prove to be at least partially cystic. High signal
intensity within the mass proved to be secondary to complex cystic fluid and not fat. Cystic thymoma was surgically documented.
found in 3 of 25 (12%) cases of invasive thymoma and limited value in differentiating among types AB, B1, B2, and
9.5 of 19 (50%) cases of noninvasive thymoma (158). B3 (171). An irregular contour strongly suggests a type C
The CT appearance of a thymoma is also related to its tumor (thymic carcinoma). Also, type C tumors are signifi-
histology determined using the WHO classification system cantly larger and more commonly associated with lym-
(Figs. 4-34 to 4-36; see Fig. 4-42) (143,171,172). On CT, phadenopathy than type B3. For example, in a CT study of
a smooth contour and round shape are most suggestive of a 53 patients with a thymic epithelial tumor classified using
type A tumor, and pleural seeding is rare in type A and WHO criteria (171), type A tumors were more likely to have
AB tumors. Calcification is suggestive of type B, but CT is of smooth contours on CT (100%) and round shapes (88%)
Figure 4-30 Invasive thymoma: CT findings. Contrast-enhanced

CT section through the carina shows a lobulated, slightly hetero-
geneous soft tissue mass that is invading the adjacent superior
Figure 4-29 Thymoma with calcification. A bulky, lobulated mass vena cava (straight arrow). Enlarged collateral vessels can be iden-
in the prevascular space contains dense areas of calcification. The tified in the region of the aortopulmonary window draining into
presence of pleural effusion on the side of the tumor suggests that the hemiazygos and azygos veins. Pleural implants are apparent
the tumor is invasive. on the right side (curved arrow).
A B
C D
Figure 4-31 Invasive thymoma: CT findings. A–D: Sequential contrast-enhanced CT sections through the mediastinum show a bulky,
heterogeneous soft tissue mass deeply invading the mediastinum with encasement of all the great vessels as well as the left main pulmonary
artery (arrow in D).
A B
Figure 4-32 Invasive thymoma in a patient with myasthenia gravis: CT findings. A: Enhanced CT at the level of the left pulmonary artery
shows a large lobulated mass with compression of the superior vena cava (small arrow). A fat plane separating the superior vena cava and
mass is not visible. A pleural nodule is present anteriorly (large arrow). B: At a lower level, the mass appears somewhat heterogeneous in at-
tenuation and calcification is visible within the mass.
A B
Figure 4-33 Invasive thymoma: CT findings. Invasive thymomas may involve the diaphragm by contiguous spread. A: CT at the level of
the diaphragm shows a normal right crus (black arrows) and thickening of the left crus (white arrows) because of tumor invasion. B: At a
lower level, thickening of the left crus (white arrows) is again seen. A nodule of tumor (black arrow) extends through the aortic hiatus to
involve the retroperitoneum.
than any other type of thymic epithelial tumor (p .05). tumors, having little potential for distant metastases, and,
Calcification was more frequently seen in type B1 (44%), if in doubt, a follow-up CT examination can be safely
type B2 (61%), and type B3 (75%) tumors than in type AB recommended. The clinical information and CT features
(14%) and type C (6%) tumors (all p .05). Type C tumors that are most helpful in distinguishing between a thy-
had a higher prevalence of irregular contours (75%) than moma and a prominent but normal thymus are summa-
any other type of thymic epithelial tumor (p .05). rized in Table 4-6.
Invasive thymomas growing along pleural surfaces can
reach the posterior mediastinum and extend downward
along the aorta to involve the crus of the diaphragm and
the retroperitoneum (Figs. 4-22 and 4-33) (140). These
areas are “hidden” on chest radiographs, but CT excels at
demonstrating these sites of involvement (153,154). A full
CT examination of the thorax extended to include the
upper abdomen should be performed in these patients.
Metastases are apt to be silent because only a minority of
patients with invasive thymomas present with symptoms
from intrathoracic spread. CT provides invaluable guid-
ance for the radiotherapist and chemotherapist to adjust
their treatment plans (162).
Because of the large size of the normal thymus in
patients younger than 25 years, diagnosing thymoma
in this age group can be difficult. However, thymomas are
extremely rare in this age group, and, unless very obvious,
we would avoid diagnosing thymoma in a patient younger
than 25 years (134,165,169,170,175). In patients older
than the age of 40 years, the diagnosis of thymoma on CT
usually poses no problem, as the normal thymus is largely
replaced by fat. In patients ranging from 25 to 40 years, the
thymus may be well seen, and, unless a definite mass is
visible, as is often the case in the presence of thymoma, a Figure 4-34 Thymoma: histology and CT findings. World Health
Organization (WHO) type A thymoma. A rounded, smoothly
definite diagnosis cannot be made. It should be noted that marginated mass (arrow) is visible in the anterior mediastinum. At
invasive thymomas, our main concern, are slow-growing surgery, the tumor capsule was intact, without evidence of invasion.
A B
Figure 4-35 Thymoma: histology and CT findings. World Health Organization (WHO) type B1 thymoma. A, B: A large mass is visible in
the prevascular space. A fat plane separating the mass from aorta is obliterated (arrow, A). It has a slightly lobulated contour best shown in
B. Pericardial invasion was present at surgery.
A B
Figure 4-36 Thymoma: histology and CT findings. World Health Organization (WHO) type B3 thymoma. A: A large rounded mass (large
arrows) is visible in the left anterior mediastinum. Focal areas of pleural thickening within the posterior left hemithorax represent pleural
implants. B, C: At two levels, the tumor is irregular in contour (arrows, B and C). D: After resection, the pleural implants (arrows) appear in-
creased in size.
C D
Magnetic Resonance Imaging Findings demonstrate signal intensity higher than that of muscle and
of Thymoma similar to that of fat. Also, with T2-weighting, thymomas
may appear homogeneous in intensity, may appear inhomo-
Thymomas typically have a signal intensity similar to or geneous because of cystic areas or regions of hemorrhage, or
slightly greater than that of muscle on T1-weighted MRI, may show nodules or lobules of tumor separated by thin,
and signal intensity increases with T2 weighting (Figs. 4-37 relatively low intensity septations (Fig. 4-38) (177,178).
to 4-40) (92,115,176). On T2-weighted images thymomas With gadolinium–diethylenetriamine penta-acetic acid
A B
Figure 4-37 Thymoma: MRI assessment. A: T1-weighted MR image at the level of the carina shows a homogeneous soft tissue mass of
intermediate signal intensity within the anterior mediastinum. B: T2-weighted MR image at the same level as shown in A. There is consider-
able signal enhancement within the mass that is now more difficult to differentiate from adjacent mediastinal fat. This appearance is entirely
nonspecific. Biopsy proved thymoma.
A B
Figure 4-38 Noninvasive thymoma: MRI assessment. A: Contrast-enhanced CT at the level of the aortic root shows a large thymic mass
within which there is a suggestion of lobulation. B: T1-weighted axial MRI at the same level as in A shows the thymic mass is clearly distin-
guished from the normal fluid filled pericardium (white arrow). There no evidence of pericardial invasion or effusion or vascular invasion.
Relatively low intensity septations (black arrows) delineate lobules of tumor.
(Gd-DTPA)–enhanced MR imaging, homogeneous and B3 thymomas show inhomogeneous signal intensity with
moderate enhancement is often observed. scattered high-intensity areas. At pathologic examination
The appearance of a thymoma on MRI is partially these areas correspond to cystic spaces, with or without the
related to its WHO histologic type. The overall signal presence of hemorrhage (172,178). Also, a lobulated inter-
intensity of types B2 and B3 thymoma is similar to that of nal architecture, with tumor lobules being separated by
types A, AB, and B1 tumors on both T1- and T2-weighted low-intensity bands, is more frequent in types B1, B2, and
images. However, with T2-weighted imaging, most B2 and B3 tumors than in other histologic types (172).
A B
Figure 4-39 Invasive thymoma: MRI evaluation. A, B: Sequential MR images through the heart show a poorly marginated mass of
intermediate signal intensity within the anterior mediastinum (curved arrow in A) that is directly invading the superior mediastinum (arrow-
head in A). The mass extends into the right atrium and can be seen crossing the tricuspid valve plane to enter the right ventricle
(arrowheads in B). A moderately sized right pleural fluid collection is present as well. Invasive thymoma was surgically documented.
A B
Figure 4-40 Invasive thymoma. A: Contrast-enhanced

CT shows a large, lobulated thymoma (large arrow).
There is marked displacement of the mediastinal vessels.
Right pleural masses (small arrows) represent pleural
metastases occurring because of invasion of the pleural
space. B: TR/TE 1500/30 MRI clearly shows the relation-
ship of the mass to great vessels. The pleural masses are
visible as in A. C: MRI at a lower level shows large
lobulated right pleural metastases These are typical of
C pleural metastases from thymoma.
In a study comparing the MRI appearances of invasive Extension to other sites (Fig. 4-40) and recurrent tumor
and noninvasive thymomas with pathologic findings, after resection or radiation (Fig. 4-41) may also be shown
Sakai et al. (178) found that 6 of 12 invasive thymomas using MRI.
showed a multinodular or lobulated internal architecture The MRI enhancement may also be of value in diagnosis
on T2-weighted images, with tumor lobules separated and staging of thymoma. In a study by Sakai et al. (179),
by 1- to 2-mm low-intensity fibrous septa; this lobular 59 patients with anterior mediastinal tumors were examined
appearance was not seen in any of 5 patients with nonin- using contrast-enhanced dynamic MRI. Thirty-one had
vasive lesions. The lobulated appearance described by thymomas and 28 had other lesions, including carcinomas,
Sakai et al. is identical to that seen in patients with WHO lymphoma, germ-cell tumor, and thymic carcinoid tumor.
type B2 or B3 tumors, which are more likely invasive than Sequential images were obtained at 30-second intervals for
types A, AB or B1 thymoma. Although lobular internal 5 minutes after bolus contrast injection. They found signifi-
architecture may be regarded as suggestive of invasive cant differences in the mean time of peak enhancement,
thymoma, it must be considered nonspecific, as it can also depending on the tumor type. The mean peak enhancement
be seen with noninvasive tumors (Fig. 4-38). time was 1.5 minutes for thymoma and 3.2 minutes for
MRI has proven valuable in identifying the presence or other tumors. Using a peak enhancement time of 2 minutes
absence of vascular invasion in patients with thymoma, or less as the threshold, dynamic MRI had a sensitivity of
especially patients in whom intravenous contrast cannot 79%, specificity of 84%, and accuracy of 81% in distinguish-
be administered. In these cases, MRI exquisitely delineates ing thymoma (enhancement time of 2 minutes of less) from
the extent of vascular involvement, including identifi- other tumors (peak enhancement time of 2.5 minutes or
cation of intracardiac disease extension (Fig. 4-39). more) (179). Furthermore, stages I and II thymomas
A B
Figure 4-41 Recurrent thymoma after irradiation. A: T1-weighted image. Collapsed and fibrotic lung (black arrow) secondary to radiation
fibrosis is visible medially. Tumor within the right pleural space (white arrow) is of similar intensity with this imaging sequence. B: With T2
weighting, high-intensity areas (arrows) within the fibrotic lung and pleural space represent recurrent tumor. This was confirmed at biopsy.
showed a mean peak enhancement time of 1.3 minutes showing thymic enlargement or focal thymic mass on CT
as compared to 2.5 minutes for stage III tumors. improved following thymectomy, although improvement
was seen in 80% of those having a normal thymus on CT.
The presence of an abnormality on CT is not necessary in
Imaging in Myasthenia Gravis
choosing patients for thymectomy, as a normal CT does
Myasthenia gravis is commonly associated with thymic not exclude FTH, and poor clinical response determines
pathology. Sixty-five percent of patients with myasthenia the need for thymic resection regardless of the presence or
gravis have thymic hyperplasia, thymoma is present in absence of thymic hyperplasia.
from 10% to 28% of patients with myasthenia gravis, and On the other hand, because thymoma is invasive in
the remainder have a normal or involuted thymus. about 30% of cases, thymectomy is indicated in all patients
Although there is no convincing evidence to suggest who have thymoma, regardless of the effect of surgery on
that thymectomy improves the outcome of myasthenia the course of the disease. The role of the radiologist in
patients who are well controlled medically, it is generally patients with myasthenia, most appropriately, is to diag-
agreed that thymectomy is indicated, regardless of thymic nose patients in whom thymoma is likely, because of the
status, when medical therapy fails. In this setting, effective presence of a focal thymic mass, and to distinguish them
surgical treatment necessitates complete removal of all from patients who show thymic enlargement indicative of
thymic tissue. Because surgical-anatomic studies have hyperplasia; such patients require surgery regardless of their
shown that both gross and microscopic thymic tissue is likelihood of responding to thymectomy, based on their
widely distributed throughout the mediastinum, it has age and the duration of their symptoms.
been suggested that en bloc transcervical-transsternal In patients with myasthenia gravis, CT is more accurate
“maximal” thymectomy be performed (180). Myasthenia in making the diagnosis of thymoma than thymic hyper-
patients most likely to benefit from thymectomy, in terms plasia. In a study of 104 patients with myasthenia who had
of disease control, are young females with a disease of thymectomy, CT was considered to show thymoma in
short duration. Older patients and patients with long- 46 of 52 patients (sensitivity 88.5% and specificity 77%).
standing myasthenia rarely benefit from thymectomy. In 44 patients with hyperplasia at histology, thymic hyper-
The role of CT in patients with myasthenia remains plasia was diagnosed on CT in only 16 cases (sensitivity
somewhat unclear, at least partially because the role of 36%, specificity 95%).
surgery in patients with myasthenia gravis, in relation to The MRI appearance of the thymus in patients with
thymic histology, is not clearly defined (181). Although myasthenia gravis has been described, but its utility in diag-
the presence of thymic enlargement or a focal mass on CT nosing thymic abnormalities in this setting appears limited.
reliably indicates the presence of FTH or thymoma (114), In a study of 16 patients with myasthenia gravis, imaged
these findings do not have a strong correlation with with both CT and MRI, who subsequently underwent
surgical outcome. In one study (114) of 45 patients with thymectomy, MRI provided little, if any, distinctive informa-
myasthenia gravis having thymectomy, 17 of 19 patients tion as compared with CT. Specifically, no distinctive MRI
features could be identified in patients with thymomas attributable to the mediastinal mass, and superior vena
apart from the presence of a mass (182). In seven patients cava syndrome may be present. Although paraneoplastic
with a histologic diagnosis of thymic hyperplasia, both syndromes such as myasthenia gravis, pure red cell
CT and MRI displayed normal thymic morphology in five aplasia, and hypogammaglobulinemia are common
patients and an enlarged and a small thymus in one patient with thymoma, they are rare with thymic carcinoma
each, respectively. Unfortunately, MRI signal intensities (184,185).
were of no value in differentiating thymic hyperplasia from Thymic carcinoma cannot be distinguished from thy-
normal thymus. moma on CT unless enlarged lymph nodes are visible in the
mediastinum or distant metastases are evident (105,
122,143,186). A large mass with or without areas of low
Thymic Carcinoma
attenuation is typical but not specific (105,143,171,183). In
Thymic carcinoma, like thymoma, arises from thymic one study, thymic carcinoma averaged 7.2 cm in diameter
epithelial cells (122,183–185), but it is much less common. (143). Calcification is uncommon, being seen on CT in 6%
Thymic carcinoma accounts for about 20% of thymic epi- of patients in one study (171).
thelial tumors. Unlike thymoma, thymic carcinoma can be An irregular contour strongly suggests thymic carcinoma
diagnosed as malignant on the basis of histologic criteria. In rather than thymoma. In one study, thymic carcinomas
the WHO classification system of thymic epithelial tumors showed an irregular contour in 75% of cases (Fig. 4-42)
reviewed previously (Tables 4-8 and 4-9), thymic carcinoma (171). However, thymoma is more likely than thymic
is type C. carcinoma to result in local invasion (186). In a study by Do
Thymic carcinoma is aggressive and more likely to result et al. (186), irregular infiltration of adjacent structures was
in distant metastases than invasive thymoma; although dis- seen in 92% of patients with invasive thymoma and 80% of
tant metastases are present in only about 5% of patients those with thymic carcinoma. Pleural implants were ob-
with invasive thymoma, they are present at diagnosis in served in 33% of patients with invasive thymoma and 10%
50% to 65% of patients with thymic carcinoma (184). of those with thymic carcinoma. On the other hand, medi-
Frequent sites of metastasis include the lungs (105), liver, astinal lymphadenopathy was seen in 8% with invasive thy-
brain, and bone (184). Complete surgical resection is the moma and 40% with thymic carcinoma. Metastases to the
desired treatment, but relapse rates and mortality are high lung, adrenal glands, or liver were observed in 40% with
(185). It has a poor prognosis. thymic carcinoma but none with invasive thymoma (186).
Thymic carcinoma is most frequently encountered in In a study by Jung et al., the CT appearances in
the fifth and sixth decades of life. Symptoms are usually 9 patients with invasive thymoma and 18 with thymic
A B
Figure 4-42 Thymic carcinoma, World Health Organization (WHO) type C. A: A large, poorly marginated, and irregular mass in the ante-
rior mediastinum shows stippled calcification. Fat planes separating the mass from aorta and main pulmonary artery are poorly seen. B: At
a different level, a cystic region or thymic cyst (arrow) is visible in association with the tumor.
carcinoma were compared. Most tumors had a lobulated docrine carcinoma, depending on their histologic charac-
contour (24/27, 89%). Thymic carcinomas were signifi- teristics and in order of increasing malignancy. The bio-
cantly larger (mean, 7.2 cm) than invasive thymomas logic behavior of these tumors shows a direct relation to
(mean, 4.7 cm) (p 0.041). Findings of invasion of the their degree of differentiation, but each of these tumor
great vessels, lymph node enlargement, extrathymic metas- types behaves in an aggressive and malignant fashion,
tases, and phrenic nerve palsy were seen only in patients with a tendency for local recurrence or distant metastases
with thymic carcinoma (143). following resection (190,191). A male preponderance
On MRI, thymic carcinoma appears higher in signal (3:1) has been reported and the mean age at presentation
intensity than muscle on T1-weighted MR images, with an is 54 years (189). Most patients present with symptoms
increase in signal on T2-weighted images (176). Hetero- related to compression of mediastinal structures.
geneous signal may reflect the presence of necrosis, cystic Approximately 40% of patients have Cushing syndrome
regions within the tumor, or hemorrhage. It has been sug- as a result of tumor secretion of adrenocorticotrophic hor-
gested by Endo et al. (177) that thymoma has a greater mone (ACTH) (Fig. 4-44) (192), and up to 20% have
tendency to show a multinodular appearance on MRI been associated with multiple endocrine neoplasia
than thymic carcinoma. Of the 15 cases of thymoma they (MEN) syndromes I and II (193). This tumor is rarely
reported (177), 79% showed lobulation, with areas of associated with the typical carcinoid syndrome. Whenever
tumor separated by thick fibrous septa of lower intensity; possible, surgical resection is the treatment of choice
this appearance was not seen in their 5 cases of thymic (189). Radiotherapy or adjuvant chemotherapy may also
carcinoma. be employed (164,194).
Using FDG-PET, Sasaki et al. (187) found a mean Although thymic neuroendocrine tumor does not
standardized uptake value (SUV) of 7.2 2.9 for thymic differ significantly from thymoma in radiographic or CT
carcinoma, higher than that in invasive thymoma appearance (92,122,190), the diagnosis can be suspected
(3.8 1.3), noninvasive thymoma (3.0 1.0), and on clinical grounds if associated with Cushing syndrome or
thymic cysts (0.9) (p .01). By using an SUV of 5.0 as MEN. In some patients, a mediastinal mass may not be
a threshold, the authors achieved a sensitivity of 84.6%, a visible on CT despite the presence of endocrine abnormali-
specificity of 92.3%, and an accuracy of 88.5% in distin- ties (188). In some patients, the tumor may show marked
guishing thymic carcinoma from thymoma. enhancement (Figs. 4-43 and 4-44). This tumor is more
aggressive than thymoma, often presenting in an advanced
stage, and superior vena cava obstruction is much more
Thymic Neuroendocrine Tumor
common with thymic carcinoid than with thymoma
Thymic neuroendocrine tumors are believed to arise from (Fig. 4-43). Regional lymph node metastases may be seen
thymic cells of neural crest origin (81,104,125,188,189). (Fig. 4-44), and distant metastases, which may include
They may be classified as carcinoid tumor (Fig. 4-43), blastic bone lesions, are seen in many cases. MRI findings
atypical carcinoid (Fig. 4-44), or small cell neuroen- are nonspecific and identical to those of thymoma (92).
A B
Figure 4-43 Thymic carcinoid tumor with superior vena cava obstruction. A: Contrast-enhanced CT section at the level of the carina
shows a bulky, somewhat heterogeneous, enhancing soft tissue mass within the prevascular space, obliterating the left brachiocephalic
vein and invading the superior vena cava (arrow). There is dense opacification of the azygos vein (curved arrow) serving as a collateral.
B: Coned-down view from simultaneous bilateral antecubital fossa venous injections shows complete obstruction of the brachiocephalic
veins bilaterally. At surgery this proved to be a primary thymic carcinoid tumor.
A B
Figure 4-44 Atypical thymic carcinoid tumor in a patient with Cushing syndrome. A: A mass in the prevascular space shows areas of
dense enhancement (arrows). B: At a different level, an enlarged lymph node (arrow) is associated with the tumor.
Octreotide radionuclide imaging or PET may be valuable wisps of soft tissue (Fig. 4-45); in 8 of 11 patients having
in imaging these tumors (195). CT, roughly equal amounts of fat and soft tissue were
visible, with the soft tissue appearing as linear whorls
intermixed with fat (seven cases) or as rounded opacities
Thymolipoma embedded within the fat (one case). The second pattern,
Thymolipoma is a rare, benign, well-encapsulated thymic seen in two cases, was that of predominant fat attenuation,
tumor, consisting primarily of fat but also containing vari- with tiny internal foci of soft tissue. The final pattern was
able amounts of thymic tissue; it can arise within the seen in a single case in which a mass appeared to be pri-
thymus or be connected to the thymus by a pedicle marily of soft tissue attenuation. In all cases, CT showed a
(104,125,196). It can be seen at any age, but is most com- connection between the mass and the thymic bed, which
mon in children and young adults. In a recent review of was narrow in seven cases and broad in four. When fat is
27 cases, 81% presented in the first four decades (196).
In a majority of cases, thymolipoma is unaccompanied
by symptoms and detected incidentally on chest radi-
ograph. However, in some patients thymolipoma may
present with symptoms resulting from compression of
mediastinal structures, such as dyspnea or chest pain. An
association with myasthenia gravis has been reported but
is rare (197).
Thymolipomas are often large, ranging up to 36 cm in
diameter (mean 18 cm) at the time of diagnosis, and may
project into both hemithoraces (196,198). Because of its
fatty content and pliability, it tends to drape over the heart,
extending inferiorly into the cardiophrenic angles, and can
simulate cardiac enlargement. Radiographically they have
also been mistaken for pleural or pericardial tumors, basal
atelectasis, and even pulmonary sequestration (199).
On CT, thymolipoma usually appears to contain a
Figure 4-45 Thymolipoma. A large mass is visible anterior and
significant amount of fat, but three patterns have been to the left of the heart (large arrow). It is composed primarily of fat,
reported (196). Most common is a combination of fat and although some soft tissue (small arrow) is visible within the mass.
visible on CT, a preoperative diagnosis can often be made.

As would be expected from their fat content, MRI shows
areas of high signal intensity on T1-weighted SE images,
similar to the intensity of subcutaneous fat, with areas
of intermediate signal intensity reflecting the presence of
soft tissue (115,196,200). Despite attaining a large size,
thymolipomas do not invade surrounding structures
(201). However, some compression of mediastinal struc-
tures is visible in half of cases (196).
Thymic Cyst
Thymic cyst can be either congenital or acquired. Thymic
cyst accounts for 1% of anterior mediastinal masses (93).
Acquired thymic cysts are generally associated with inflam-
mation and have been reported following radiation therapy
for HL (202,203), in association with thymic tumors
(Fig. 4-42) (204–206), and following thoracotomy (207).
Congenital thymic cysts are usually unilocular, have a
thin wall, and contain clear fluid (Fig. 4-46) (93). They
may occur in the neck but are quite rare in this location
(208). Acquired thymic cysts are often multilocular (93). Figure 4-46 Thymic cyst. A low-attenuation, sharply marginated
Multilocular thymic cysts have been seen in children with cyst is visible in the region of the thymus (arrow). A cyst wall is not
HIV infection (130,131). The attenuation of thymic cysts is visible.
usually that of water but can be higher or lower depending
on the presence of hemorrhage or lipid deposits. Calci- CT can suggest the diagnosis if (a) the cyst appears thin
fication of the cyst wall can also be seen. walled, (b) it is unassociated with a mass lesion, and (c)
One should be conservative in making the diagnosis of it contains fluid with a density close to that of water or
thymic cyst, as cystic regions can be seen in a variety remains unopacified following contrast infusion. MRI char-
of thymic tumors, including thymoma and lymphoma (93). acteristics are similar to those of other cystic lesions (115).
A B
Figure 4-47 Hodgkin disease with thymic involvement. A, B: Contrast-enhanced CT scans at the level of the aorticopulmonary window
and main pulmonary artery, respectively, show thymic enlargement (Th) and enlarged lymph nodes in the pretracheal space (black arrow in
A) and aorticopulmonary window (white arrow in A). Lymph node enlargement in conjunction with thymic enlargement is typical of lym-
phoma involving the thymus.
Thymic Lymphoma and Metastases

HL has a predilection for involvement of the thymus in
conjunction with involvement of mediastinal lymph nodes
(Figs. 4-47 and 4-48) (209). In a study of adults with newly
diagnosed HL involving the thorax, thymic enlargement
was seen in 30% and in all, mediastinal lymph nodes were
also enlarged (Figs. 4-47 to 4-49). In a study of 60 pediatric
patients with HL, 17 (28%) had thymic enlargement; this
represented 49% of patients with a mediastinal abnormal-
ity (210). In this study, 73% also showed mediastinal
lymph node enlargement. Thymic enlargement was seen in
38% patients with intrathoracic recurrence (209). Thus,
HL, particularly the nodular sclerosing type, should be
considered in the differential diagnosis of a thymic mass,
but lymph node enlargement is typically present, at least in
Figure 4-48 Hodgkin disease with thymic involvement. Anterior
adults, and should suggest the correct diagnosis. Non- mediastinal mass (arrows) likely represents thymic involvement
Hodgkin lymphoma (NHL) much less commonly involves because of its typical shape. Lymph node enlargement was visible
the thymus. at other levels.
Thymic involvement in Hodgkin or other lymphomas
is usually indistinguishable from thymoma or other head (83%) or bilobed (17%) appearance, but appears
causes of anterior mediastinal mass. Lobulation or a enlarged and has convex borders where it contacts lung
nodular appearance is common (211). In some cases, the (209). In adults, the thickness of the thymus in patients
enlarged thymus retains its normal shape, with an arrow- with Hodgkin disease measured 1.5 to 5 cm (209); in
A B
C D
Figure 4-49 Hodgkin lymphoma with thymic involvement: MR findings. A: T1-weighted transaxial image shows a large thymic mass
(arrow), associated with right hilar and subcarinal lymphadenopathy. B: With T2 weighting the mass appears inhomogeneous, a finding
common with lymphomas. C: T1-weighted coronal image shows the thymic mediastinal mass (arrow), associated with a large pericardial
effusion. D: After radiation, a T2-weighted image at the same level as C shows the thymic mass (arrow) to be significantly reduced in size,
as are the hilar lymph nodes.
A B
Figure 4-50 Thymic lymphoma. A: A T1-weighted image

(800/20) shows thymic enlargement appearing homogeneous in
intensity. B: With T2 weighting (2400/70), the mass appears inho-
mogeneous. (C) At a lower level, the relationship of the mass
to the fluid filled pericardium (arrow) is visible on a coronal image
C obtained with T1-weighting.
children, thickness of the larger thymic lobe ranged from present. CT and MRI appearances of thymic metastases are
2.5 to 8.6 cm (210). nonspecific.
In patients with lymphoma, the thymus usually appe-
ars homogeneous in attenuation, but an inhomogeneous
nodular internal architecture can be seen (210,211). Cystic GERM-CELL TUMORS
areas of necrosis may be visible at CT; this has been reported
in 21% to 23% of adult patients (209,211) but is less Primary germ-cell tumors account for about 10% to 15%
common in children (210). Except in occasional cases, calcifi- of primary mediastinal masses and a higher proportion of
cation does not occur in the absence of radiation (210). anterior mediastinal masses. They are histologically identi-
On MRI, thymic lymphoma shows low intensity on cal to their gonadal counterparts (81,104). Presumably
T1-weighted images, with a variable change on T2- they arise from primitive germ cells that have arrested their
weighted images (115). In one study (58), two of seven pa- embryologic migration in the mediastinum, frequently
tients with thymic lymphoma had a homogeneous appear- within the thymus (123,213). They are most common in
ing thymus, whereas in the remaining five, areas of the anterior mediastinum; only about 5% to 8% originate
increased or decreased intensity were visible on T2- in the posterior mediastinum (81,123,124). Most germ-
weighted images (Fig. 4-50). The low-intensity areas may cell tumors present during the second to fourth decades of
represent fibrosis, whereas the high-intensity areas could life. Germ-cell tumors include benign and malignant
reflect hemorrhage or cystic degeneration. Although the teratoma, seminoma, embryonal carcinoma, endodermal
MRI characteristics of lymphoma are not characteristic, the sinus (yolk sac) tumor, choriocarcinoma, and mixed types
combination of a thymic mass with mass or lymph node (92,214,215). Generally, these tumors are considered in
enlargement in other areas of the mediastinum is very sug- three categories: (a) teratoma, (b) seminoma, and (c) non-
gestive of the diagnosis. seminomatous germ-cell tumors.
Lung and breast carcinomas, and other metastatic Overall, more than 80% of germ-cell tumors are be-
tumors, can also involve the thymus (212). In the case of nign (Table 4-11) (124,216), with the large majority of
lung cancer this is typically the result of direct extension, benign tumors being teratomas. Although the sex distri-
although hematogenous metastases may also occur. bution of benign germ-cell tumors is about equal, there
Involvement of mediastinal lymph nodes is also typically is a strong preponderance of males among patients with
TABLE 4-11 TABLE 4-12

GERM-CELL TUMORS TERATOMA
10% to 15% of primary mediastinal masses Contain elements of all germinal layers
Classified as mature, immature, or malignant
Benign tumors
More than 80% of germ-cell tumors are benign Mature teratoma
Most are teratomas 60% to 70% of all mediastinal germ-cell tumors
Equal sex distribution Well-differentiated benign tissue
Usually asymptomatic Ectodermal elements such as skin and hair predominate
Mature cystic teratoma: dermoid cyst
Malignant tumors
Anterior mediastinum
Male predominance
Often cystic
Symptoms common
CT: combination of fluid, fat, soft tissue, calcification
Classified as malignant teratoma (10%), seminoma (30%–40%,)
MRI: variable appearance with T1 and T2 weighting depending
or nonseminomatous germ-cell tumor (50%–60%)
on contents
Immature teratoma
Tissues typical of fetal development
Presentation in infancy and childhood: benign course typical
malignant germ-cell tumors (123,124). A striking male
Presentation in adults: often aggressive and malignant
predominance has been noted recently in a series of three
reports from the Armed Forces Institute of Pathology Malignant teratoma
Frankly malignant tissues
(217–219). Of 322 patients with primary mediastinal Poor prognosis
germ-cell tumors in this study, 320 were men, including Seen almost entirely in men
all 120 patients with mediastinal seminomas and all 64
cases with yolk sac tumors, embryonal carcinomas, and
nonteratomatous germ-cell tumors of the mediastinum,
respectively. tissues by bronchial and gastrointestinal epithelium or
Among patients with malignant tumors, seminoma is pancreatic tissue.
most common, representing between 30% and 40% of Immature teratomas contain less well developed tissues
cases (104), with embryonal carcinoma and malignant more typical of those present during fetal development; in
teratoma each responsible for about 10% and choriocarci- infancy or early childhood, these tumors often have a
noma and endodermal sinus tumor responsible for about benign course, whereas in adults they usually behave in an
5% each; the remainder of malignancies, approximately aggressive and malignant fashion (123). Malignant
40% of cases, represent mixed tumors (124,220). teratomas contain frankly malignant tissues and have a
Benign tumors are often asymptomatic, whereas malig- very poor prognosis; they are seen almost entirely in men.
nant tumors are more likely to cause symptoms (123). Teratomas are usually found in the prevascular space,
Confirmation that these lesions are primary to the medi- although it is important to realize that these lesions may
astinum requires that there be no evidence of a testicular occur in other locations in up to 20% of cases, including
or retroperitoneal tumor. Although of some value in diag- both the middle and posterior mediastinum as well as in
nosis, the primary role of CT in evaluating patients with multiple compartments (221–225). Regardless of their
malignant germ-cell tumors is defining disease extent and histology, CT often shows a combination of fluid-filled
monitoring response to therapy. cysts, fat, soft tissue, and areas of calcification (Figs. 4-51 to
4-55) (93,215,223,226–228). This pleomorphic appear-
ance is an important clue to the diagnosis and often makes
Teratoma
it possible to distinguish these lesions from thymoma and
Teratomas contain elements of all germinal layers. They are lymphoma (92,105). Calcification is seen in from 20% to
classified as mature, immature, and malignant (Table 4-12) 80% of cases (Figs. 4-51, 4-54, and 4-55), being focal,
(104). Mature teratomas account for the vast majority of rimlike, or rarely representing teeth or bone (123). Fat is
cases and represent between 60% and 70% of all visible on CT in half of cases (Figs. 4-51 to 4-53), a finding
mediastinal germ-cell tumors (104). They are composed of that is highly suggestive of this diagnosis. A fat-fluid level
well-differentiated benign tissue, with ectodermal elements within the mass is diagnostic (229,230).
such as skin and hair predominating. Mature cystic teratomas (dermoid cysts) are usually
Mature cystic teratoma, often referred to as dermoid found in the anterior mediastinum (Figs. 4-53 and 4-54);
cyst, is a cystic tumor composed of well-differentiated they occasionally occur in the posterior mediastinum or
elements from at least two of the three germ-cell layers lung. A large, predominantly cystic, anterior mediastinal
(ectoderm, mesoderm, and endoderm). Ectodermal tissues mass with a thin and well-defined wall is highly suggestive
may be represented by skin, teeth, and hair; mesodermal of mature cystic teratoma (93). Most cystic teratomas are
tissues by bone, cartilage, and muscle; and endodermal multilocular, but unilocular cystic lesions also occur.
Figure 4-52 Teratoma. CT section through the carina shows

a well-encapsulated mass anterior to the left pulmonary artery,
within which fat elements are clearly discernible. This was con-
Figure 4-51 Mature teratoma. Enhanced CT in an adult with a firmed by measurements obtained within the region of interest
mature teratoma. A large mass is visible, containing fat (arrows) demarcated by cursor number 1. The remainder of the lesion was
and dense calcifications. The mass originates in the anterior medi- composed of soft tissue elements.
astinum but extends posteriorly and displaces the heart to the left.
Rarely, the mature cystic teratoma has an imperceptible in 53%. All these elements were present in the same lesion in
wall (94). 39%, whereas the combination of soft tissue, fluid, and fat
Moeller et al. (223) reported findings in a total of was seen in 24%. Importantly, nonspecific cystic lesions
66 cases of mature teratoma located in the mediastinum. without fat or calcium were seen in a total of 15% of cases.
Although soft tissue attenuation was identified in virtually Wu et al. (231) have reported the sonographic findings
all cases, fluid was present in 88%, fat in 76%, and calcium associated with 28 mediastinal teratomas. These findings
A B
Figure 4-53 Mature teratoma. A: A sharply marginated anterior mediastinal mass with a well-defined wall (arrow) contains a large
amount of fat. B: At a different level, some soft tissue attenuation (arrow) is seen within the fat. At surgery, the mass was cystic and
contained both fat and white hair.
A B
Figure 4-54 Mature cystic teratoma. A: Posteroanterior radiograph shows extensive opacification of the mid and lower right hemithorax
in a teenage girl whose chief complaint was moderate dyspnea. The chest radiograph is nonspecific. B: Contrast-enhanced CT section
through the midthorax shows a complex, cystic mass within which discrete septations can be recognized (curved white arrow), causing
compression and posterior displacement of the adjacent lung (curved black arrows) as well as mediastinal shift to the left. Faint curvilinear
calcifications can be identified posteriorly (straight black arrow). Although a pleural etiology was initially considered, the finding of multiple
septations within this mass occurring in a young girl suggested the proper diagnosis. At surgery, the mass was found to extend posterior to
the lung as well, accounting for the area of fluid density seen posteriorly (straight white arrow), otherwise mimicking the appearance of
loculated pleural fluid.
A B
Figure 4-55 Teratoma: MRI findings. A: Unenhanced CT just

below the carina shows a well-encapsulated mass anterior to the
right main pulmonary artery (curved arrows). The mass is strikingly
heterogeneous with fluid, fat (arrow), and calcific elements clearly
discernible. B, C: T1- and T2-weighted MR images obtained at the
same level as shown in A, respectively. On the T1-weighted image,
regions of high signal intensity correspond precisely to the fatty el-
ements seen in A (arrow in B), whereas regions corresponding to
calcium demonstrate signal void (curved arrow in B). Intermediate
signal is present within the remainder of the mass, in this case pre-
sumably secondary to complex fluid within the tumor. Note that
there is considerable signal enhancement within these regions on
the T2-weighted image, consistent with the presence of fluid
within the tumor. (Case courtesy of Jeffrey Weinreb, MD, New
C Haven, Connecticut.)
A B
Figure 4-56 Teratoma: MR findings. A: Coronal non–breath-

hold, ECG-gated proton density, fast spin-echo [100/43/150
(TR/TE/flip angle)] image performed in phased-array coil demon-
strates a complex cystic anterior mediastinal mass. B, C: Axial and
sagittal breath-hold, fat-suppressed T1-weighted gradient-echo
[183/4/80 (TR/TE/Flip angle)] images, respectively, performed in
the body phased-array coil, after the administration of gadolinium,
demonstrate an irregular, thick-walled cystic anterior mediastinal
mass with both fluidlike (arrow in B) and nodular components
C (curved arrow in B).
generally fall into three patterns. Most (18 of the 28) hyperechoic mass; this correlated with the presence of hair
patients showed a complex mass of heterogeneous and sebaceous material found on pathologic study. Two
echogenicity; this pattern was associated with varying com- patients showed floating spherules within a cystic mass.
binations of fat, sebaceous and mucinous materials, hair, These were composed of sebaceous material associated with
and calcification. Eight patients showed a homogeneous fluid (231).
Figure 4-57 Malignant teratoma. Contrast-enhanced CT scan

through the carina shows a large, poorly demarcated tumor within
the anterior mediastinum causing compression and displacement Figure 4-58 Mediastinal growing teratoma syndrome. An
of adjacent mediastinal structures. The tumor is markedly hetero- enlarging multicystic mass (arrows) was visible in a patient previ-
geneous with areas of low density, presumably caused by exten- ously treated for embryonal carcinoma that contained elements of
sive tissue necrosis, as well as some punctate areas of calcification teratoma and was metastatic to the mediastinum. Surgery showed
(arrow). These findings suggest a malignant etiology. mature cystic teratoma.
Mature teratomas are predisposed to rupture. This has

been reported in as many as a third of cases (232). This
TABLE 4-13
phenomenon has been attributed to the presence of
SEMINOMA
digestive enzymes secreted by pancreatic or intestinal
mucosa within these tumors, leading to rupture into ad- Occurs almost entirely in men
jacent structures including bronchi, pleura, lung, and Mean age at presentation of 29 years
even the pericardium. Rarely, a fat-fluid level may be Large, smooth or lobulated, homogeneous soft tissue mass
Obliteration of fat planes
identifiable within a pleural effusion caused by rupture Sensitive to radiation and chemotherapy
of the primary tumor into the pleural space (233). The
fluids within the cystic parts of the tumors also vary in
their CT density and may reach soft tissue density.
Teratomas are typically encapsulated and well demar- Seminomas are very sensitive to both radiation and
cated; rim enhancement can be seen. chemotherapy, although chemotherapy is associated with
MRI can show various appearances, depending on the a better survival and is generally preferred (236). Surgical
composition of the tumor (Figs. 4-55 and 4-56) (92,115). resection may be performed in patients with a residual
They commonly contain fat, which is intense on mediastinal mass, although the likelihood of viable tumor
T1-weighted images, and cystic areas, which are low in inten- is low (236). Residual mediastinal mass may also reflect
sity on T1-weighted images, but increase with T2 weighting. the presence of associated mature teratoma. Long-term
Malignant teratoma typically appears nodular or poorly survival may be anticipated in up to 80% to 90% of cases
defined, and the tumor molds and compresses surround- (104,236).
ing structures (Fig. 4-57), whereas benign teratomas are
well defined and smooth (228). Malignant teratomas are
more likely to appear solid and less often contain fat
Nonseminomatous Germ-Cell Tumors
(40%) than benign lesions (123), but they can be cystic as Nonseminomatous tumors, including embryonal carci-
well. Following contrast infusion, malignant teratoma can noma, endodermal sinus (yolk sac) tumor, choriocarci-
show a thick enhancing capsule (123,214). noma, and mixed types, are often grouped together
An unusual phenomenon associated with the treatment because of their similar appearance and aggressive
of malignant teratoma or seminoma has been reported; behavior (Table 4-14) (123,214,215,227,228,238–240). As
this is the so-called mediastinal growing teratoma syn- a group, nonseminomatous tumors are more common
drome (234). Teratomas and other germ-cell tumors may than seminoma; in a recent study of mediastinal germ-cell
be mixed, with a combination of malignant and benign tumors, 86% were nonseminomatous and 14% were semi-
tissues being present. Benign parts of the tumor can nomas (236). An association with Klinefelter syndrome
continue to grow after therapy, despite sterilization of has been reported in as many as 18% of patients (104),
malignant elements of the tumor (Fig. 4-58). Surgery may but in a recent review of 286 mediastinal nonseminoma-
be required if growth is sufficient to impinge on adjacent tous tumors, only 3 (1%) had Klinefelter syndrome.
mediastinal organs or if the diagnosis is in doubt. On CT, these tumors usually show heterogeneous opac-
ity, including ill-defined areas of low attenuation secondary
Seminoma
Seminoma occurs almost entirely in men, with a mean age
at presentation of 29 years (Table 4-13) (235). They repre-
sent 40% of malignant germ-cell tumors of single histol- TABLE 4-14
ogy (104). Approximately 10% with pure seminoma have NONSEMINOMATOUS GERM-CELL TUMORS
evidence of elevated b-human chorionic gonadotropin
Tumors grouped together because of appearance and aggressive
(HCG) levels, but never elevated a-fetoprotein (AFP)
behavior
levels. Presenting symptoms include dyspnea (25%), chest Embryonal carcinoma
pain (23%), cough (17%), fever (13%), weight loss Endodermal sinus (yolk sac) tumor
(11%), superior vena cava syndrome (6%), or fatigue and Choriocarcinoma
weakness (6%) (236). Mixed tumors
Typically, primary mediastinal seminoma appears as More common than seminoma
a large, smooth or lobulated, homogeneous soft tissue CT findings
mass, although small areas of low attenuation can be seen Mass with heterogeneous attenuation
(Fig. 4-59) (92,123,214,215,227,228,237). Obliteration Ill-defined
of fat planes is common, and pleural or pericardial effu- Cystic areas of low attenuation
Infiltrative and poorly marginated
sion may be present, although direct invasion of adjacent
Obliteration of fat planes
structures has been reported to be rare (104,214).
B C
Figure 4-59 Seminoma. A: Posteroanterior chest radiograph in a 35-year-old man shows poor definition of the right heart border and
elevation of the right hemidiaphragm, findings initially thought to be secondary to middle lobe volume loss. B, C: Contrast-enhanced CT sec-
tions through the mid and lower thorax, respectively, show a homogeneous soft tissue mass to be present in the anterior mediastinum
(arrows in B and C), marginating the right heart border. This appearance is entirely nonspecific. At surgery, this proved to be a seminoma.
A B
Figure 4-60 Nonseminomatous germ-cell tumor. A: Posteroanterior chest radiograph shows a mediastinal mass deforming the right pleu-
romediastinal interface. B: Contrast-enhanced CT section at the level of the aortic arch shows an irregular, heterogeneous tumor mass that
is partially obstructing the superior vena cava (arrow). Although most germ-cell tumors arise in the prevascular or anterior mediastinal com-
partment, these tumors may originate anywhere within the thorax. Surgery confirmed a nonseminomatous malignant germ-cell tumor.
piration biopsy performed as indicated (241,242). CT has

generally been reserved for evaluation of patients with sus-
pected intrathoracic extension of thyroid abnormalities.
Although intrathoracic extension occurs in only a few per-
cent of patients with thyroid disease, in some series it repre-
sents nearly 10% of mediastinal masses resected at thoraco-
tomy (242–244). Intrathoracic thyroid masses nearly always
represent direct contiguous growth into the mediastinum of
a goiter or other thyroid lesion. They are almost always con-
nected to the thyroid gland, even though radionuclide stud-
ies may suggest that the mass is separated from the thyroid
(96,245). Truly ectopic mediastinal thyroid tissue in the me-
diastinum is rare. The differential diagnosis of intrathoracic
thyroid lesions includes goiter, thyroid enlargement associ-
ated with thyroiditis, and thyroid carcinoma,
Mediastinal involvement by thyroid masses is most
often anterior (96). In 80% of cases, an enlarged thyroid
extends into the thyropericardiac space anterior to the
recurrent laryngeal nerve and the subclavian and innomi-
nate vessels. Posterior mediastinal goiters constitute
approximately 10% to 25% of cases (96,245). Posteriorly
located goiters typically arise from the posterolateral por-
tion of the gland and descend behind the brachiocephalic
vessels; they are most commonly found on the right side
in close proximity to the trachea and bounded inferiorly
Figure 4-61 MR of nonseminomatous germ-cell tumor:
embryonal carcinoma. T1-weighted sagittal image (TR 600 msec; by the arch of the azygos vein (245). Less often, thyroid
TE 28 msec) shows a large inhomogeneous mass with areas tissue may either extend between the esophagus and tra-
of both high and low signal intensity. Low-intensity regions chea or even come to lie posterior to the esophagus (245).
are likely cystic, whereas high-intensity regions may represent
hemorrhage.
to necrosis and hemorrhage or cystic areas (123,214,238)
of Thyroid Disease
(Fig. 4-60). They often appear infiltrative, with obliteration The CT appearances of both normal and abnormal thyr-
of fat planes, and may be spiculated. Calcification may be oid gland have been extensively reviewed (96,242–244,
seen. MRI findings also reflect the inhomogeneous nature 246–256). The appearance of normal thyroid tissue is char-
of these lesions (Fig. 4-61). acteristic. On precontrast scans, thyroid tissue is high in at-
It is worth emphasizing that over the past decade there tenuation, relative to adjacent soft tissues, because of its
has been a remarkable improvement in the therapy of high iodine content; thus, it is easily identified (243,249).
nonseminomatous mediastinal germ-cell tumors. With the Attenuation values of normal thyroid tissue average 112
introduction of cisplatin-based chemotherapeutic regi- 10 HU. Thyroid attenuation measures 57 11 HU in pa-
mens, the likelihood of long-term survival has risen from tients with chronic thyroiditis, and in hypothyroid patients,
3%–10% to 45%–80% (234,236). Surgery is reserved for thyroid attenuation is only slightly greater than soft tissue.
those patients with radiologic evidence of persistent Thyroid tissue typically enhances 25 HU or more following
masses and may improve survival (236). In one study, intravenous administration of contrast, depending on the
nearly half of patients with mediastinal nonseminoma rapidity of contrast infusion (Fig. 4-62A) (249).
had resection of residual masses, which represented viable Recognizing that a mediastinal mass originates from the
tumor in more than 30%. As with seminoma, teratoma thyroid gland on CT is contingent on the following obser-
may be responsible for residual mass in some patients; vations: (a) demonstration of a communication with the
mature teratoma was responsible for the residual mass in cervical portion of the thyroid gland when contiguous sec-
26% of cases in a study by Bokemeyer et al. (236). tions are extended to include the neck (Fig. 4-62); (b) high
attenuation of at least of portion of the mass; (c) marked
enhancement after contrast medium injection, sometimes
THYROID GLAND to the point of simulating a vascular lesion; and (d) pro-
longed contrast enhancement (243) presumably caused
Usually, diseases of the thyroid gland are evaluated using ra- by the thyroid actively trapping iodine contained in the
dionuclide scintigraphy or ultrasonography, with needle as- contrast medium.
A B
Figure 4-62 Thyroid gland and mediastinal goiter. A: The cervical thyroid gland is densely enhanced following contrast administration.
A portion of the right thyroid lobe (arrow) extends posteriorly and is contiguous with the mediastinal goiter shown in B. B: An enhancing
heterogeneous mass (arrow) is present posterior and lateral to the trachea, representing a goiter.
Generally speaking, the CT appearance of intrathoracic tate, or ringlike calcifications can also be seen in both
thyroid masses is nonspecific. Mediastinal thyroid masses benign and malignant lesions (Fig. 4-65); fat has not been
commonly appear inhomogeneous and cystic on CT reported in thyroid masses and its presence can help in
regardless of their cause (Figs. 4-62 to 4-66). Although the distinguishing a teratoma or dermoid cyst from a thyroid
appearance of low-density cystic areas shown on CT corre- lesion (257). CT is of greatest value in defining the morpho-
lates with areas of decreased isotope uptake on radionuclide logic extent of a thyroid mass and its relation to other
studies, this appearance is nonspecific. Limitations in histo- structures (Figs. 4-62 to 4-65). Marked irregularity of the
logic specificity have been noted when CT is compared with gland contour, loss of distinct mediastinal fascial planes,
high-resolution sonography (252–256). Curvilinear, punc- and/or the presence of cervical or mediastinal adenopathy
A B
Figure 4-63 Posterior mediastinal goiter. A, B: Sequential CT sections at the level of the great vessels shows a well-defined, inhomoge-
neous mass posterior to the trachea and esophagus, which are displaced anteriorly. Note that in this case, thyroid tissue is limited posteri-
orly by an aberrant left subclavian artery (arrows).
A B
Figure 4-64 Anterior mediastinal goiter. A, B: Large, inhomogeneous anterior mediastinal mass is visible on both sections following con-
trast enhancement, with areas of low attenuation consistent with the presence of colloid cysts. At higher levels, the mass was contiguous
with the inferior aspect of the thyroid gland (not shown).
A C
Figure 4-65 Middle mediastinal goiter. A: CT section through the

great vessels immediately following contrast administration shows
the right lobe of the thyroid gland (arrows) to be markedly
enlarged and cystic in appearance. The left lobe of the thyroid
(arrowhead) is slightly prominent. B: Several seconds later, CT
section at a slightly lower level than in A. There is inhomogeneous
enhancement within the goiter (arrow), which is still clearly distinct
from adjacent mediastinal vessels. Note that in this case the mass
lies in the pretracheal space (middle mediastinum), posterior to
the great vessels. C: CT section in a different patient from the one
in A or B also shows a heterogeneous substernal thyroid gland
extending along the right side of the trachea. Note the presence of
coarse punctate calcifications within the thyroid, a common finding
B in goiters.
A B
Figure 4-66 Thyroid cancer. A: Contrast-enhanced CT section at the level of the thoracic inlet. Thyroid tissue is easily identifiable because of
marked contrast enhancement that is characteristic following intravenous administration of contrast. Within the thyroid, areas of low tissue
attenuation and foci of calcification can be identified. Although the thyroid is markedly enlarged, causing the trachea to deviate to the right,
the outlines of the thyroid are still identifiable. Biopsy proved follicular carcinoma of the thyroid. B: T1-weighted MRI section in a patient differ-
ent from the one in A. The thyroid is markedly enlarged and is poorly marginated, causing loss of visualization of normally discrete fascial
planes (arrows). Note that with T1 weighting the thyroid appears relatively homogeneous. Biopsy proved follicular carcinoma of the thyroid.
should signal potential malignancy (Fig. 4-66). However, their markedly prolonged T2 values or on enhanced
carcinomas may also be sharply marginated. sequences. Functioning thyroid nodules, however, are an
CT plays an especially important role in the preopera- exception; these have been reported to be isointense with
tive assessment of substernal goiters (96). It has been normal thyroid tissue on both T1- and T2-weighted scans
shown that the surgical approach to these lesions depends (264).
on precise anatomic localization. Intrathoracic goiters Multinodular goiters are relatively hypointense as com-
should be removed for symptomatic relief as well as to re- pared with normal thyroid tissue on T1-weighted images,
duce the risk consequent to acute hemorrhage or inflam- except when there are foci of either hemorrhage or cysts, in
mation. Although anterior substernal goiters usually can which case focal areas of high signal intensity may be
be removed through a routine cervical incision, posterior visualized (Figs. 4-68 and 4-69). They generally remain
mediastinal goiters require a selective approach. In these more intense than muscle. On T2-weighted images,
cases, a decision as to the need for a thoracotomy or a multinodular goiters are typically heterogeneous, with
combined cervicothoracic incision generally depends on high signal intensity noted throughout most of the gland
the size of the lesion or whether or not the mass is mainly (Figs. 4-68 and 4-69) (261,262,264–267). Although it
intrathoracic without a significant cervical component has been suggested that benign adenomas can be differen-
(245). tiated from follicular carcinomas based on the presence
of an intact pseudocapsule surrounding adenomas, this
finding has been insufficiently documented (261,268). In
most series, MRI has proven no more specific histologi-
of Thyroid Disease cally than CT (259,262–265,267).
MRI is useful in evaluating thyroid masses (242,258–260). In patients with Graves disease, Noma et al. (261) have
Characteristically, on T1-weighted images, the signal inten- noted typical morphologic features, including the pres-
sity of the normal thyroid is equal to or slightly greater ence of numerous coarse, bandlike structures traversing
than that seen in the adjacent sternocleidomastoid muscle; the thyroid gland and dilated vascular structures within
on T2-weighted scans or on T1-weighted gadolinium- the thyroid parenchyma. Also, in patients with Graves dis-
enhanced images, the signal intensity of the thyroid gland ease, MRI has the capability to provide physiologic insight
is significantly greater (Fig. 4-67) (261–263). Most focal into the functional status of the thyroid gland (269).
pathologic processes, including adenomas, cysts, and can- Typically, patients with Graves disease have moderate to
cer, are easily identified on T2-weighted images because of marked increased signal intensity on studies performed
A B
Figure 4-67 Normal thyroid gland: MR assessment. A: T1-

weighted. B: T2-weighted fat saturated. C: Gadolinium-enhanced,
T1-weighted, fat saturated. On T1-weighted images the signal
intensity in the normal gland is equal to or slightly greater than is
seen in adjacent muscles (arrows in A). Note that the signal intensity
of the thyroid is significantly greater with T2 weighting (arrows in B)
C and following enhancement (C).
with both short and long TRs. Significantly, in these as cervical and/or mediastinal adenopathy; displacement of
patients, there is a linear relationship between the thyroid- mediastinal structures such as the esophagus, trachea, and
muscle signal intensity contrast ratio and both the serum great vessels; and loss of normal cervical and intrathoracic
thyroxine (T4) level and the 24-hour radioactive iodine fascial planes (261–268,270–272). Unfortunately, MRI has
uptake. Furthermore, these changes have been shown to proven of only limited value as a means for tissue charac-
normalize following therapy (269). Descriptions of the terization.
thyroid gland in patients with Hashimoto thyroiditis have A potential role for MRI in the evaluation of patients
also been published. Although this disorder is character- following surgery has also been proposed (264,273). In
ized by diffuse enlargement, no distinct pattern has been one series of 24 patients with primary thyroid carcinoma
identified with MRI (261). evaluated postoperatively with MRI, MRI correctly diag-
MRI has been shown to be a sensitive means of defining nosed or excluded disease in 20 patients. However, MRI
the extent of thyroid enlargement (Figs. 4-69 and 4-70). provided a false-positive diagnosis in one case and a false-
This includes the presence of associated abnormalities such negative diagnosis in three cases (273).
A B
Figure 4-68 Goiter: MR evaluation. A: T1-weighted MR image at the thoracic inlet shows typical appearance of a goiter. Note that the
enlarged left thyroid lobe (arrow) is easily identified, separate from adjacent mediastinal fat and vessels, displacing the trachea to the right.
It appears slightly more intense than adjacent muscle. B: T2-weighted MR image. There is considerable increase in signal intensity within the
goiter, which appears heterogeneous. On T2-weighted images, multinodular goiters are typically heterogeneous, with high signal intensity
identifiable throughout the gland.
A B
Figure 4-69 Substernal goiter, MR evaluation. A: T1-weighted

image through the neck shows a large mass (arrows), containing
some areas of high intensity, which may represent hemorrhage.
The mass is similar to muscle in intensity. B: T2-weighted image
near the same level shows the mass to be heterogeneous in inten-
sity and brighter than muscle. C: T1-weighted image through the
mediastinum at the level of the aortic arch shows the goiter to
extend into the pretracheal space (arrow). D: T1-weighted coronal
image shows the goiter (arrows). E: T1-weighted coronal image
shows the goiter (arrows) lying anterior to the trachea and extend-
ing from the neck into the superior mediastinum. The trachea is
C displaced posteriorly.
D E
PARATHYROID GLANDS Approximately 10% of parathyroid glands are ectopic

(Table 4-15). In one study, 62% of the ectopic glands were
Ninety percent of parathyroid glands are located near the located in the anterior mediastinum, 30% were embedded
thyroid gland. Although four glands are usually present, within thyroid tissue, and 8% were found in the posterior-
their precise localization and number are variable. The superior mediastinum, in the region of the tracheoe-
upper pair is typically located dorsal to the superior poles sophageal groove (274). In a study of ectopic parathyroid
of the thyroid gland, whereas the lower pair lies just below tumors in the mediastinum, 81% were in the anterior
the lower thyroid poles, in the region of the minor mediastinum and 19% were in the posterior mediastinum
neurovascular bundle. The lower pair of glands is most (275); in another study, ectopic mediastinal glands were
variable in location. Most parathyroid adenomas are most often within the thymus or in the paraesophageal
found in the lower group of parathyroid glands. regions (276). Anterior mediastinal parathyroid glands are
A B
Figure 4-70 Substernal thyroid carcinoma, MR evaluation. A: T1-weighted image through the upper mediastinum shows a large mass,
which displaces and narrows the trachea (arrow). The mass is slightly more intense than muscle. B: T1-weighted coronal image shows the
extensive tumor involving the right neck and mediastinum (arrows).
performed for primary hyperparathyroidism, Stark et al.

TABLE 4-15 (253) found that 40% of these patients had nonpalpable
MEDIASTINAL PARATHYROID MASS thyroid nodules detected either by CT or high-resolution
sonography.
Primary hyperparathyroidism is due to adenoma (85%) or
hyperplasia
The ability of CT to detect parathyroid abnormalities is
Adenomas, hyperplastic glands range in size from 0.3 to 3 cm clearly related to scan technique. Utilizing contiguous
10% of parathyroid glands are ectopic 5-mm sections from the hyoid bone to the carina,
62%–81% in the anterior mediastinum prospectively reconstructed with small fields of view, a
30% within thyroid tissue 512 matrix, and maximum contrast enhancement using
8% in the tracheoesophageal groove
Likelihood of an ectopic adenoma increased in postoperative
50 g of iodine injected through a power injector, Cates
patients with persistent hyperparathyroidism et al. (292) documented that CT correctly identified
Differentiation of hyperplasia, adenoma, carcinoma is difficult parathyroid adenomas preoperatively in 81% of patients,
using CT or MRI although somewhat lower sensitivities have been reported
Intense on T2-weighted images; enhancement using gadolinium by others (293). The accuracy of CT compares favorably
MRI and sestamibi imaging are complementary
with reported sensitivities of both high-resolution ultra-
sonography and thallium radionuclide imaging (290,291,
294). However, CT obtained with substandard technique
has a much lower accuracy (281). It should be empha-
thought to result from islands of parathyroid tissue that sized that in patients with primary hyperparathyroidism,
are carried into the anterior mediastinum by the descend- surgical neck exploration with resection of parathyroid
ing thymus during embryologic development. Anterior tissues is curative in about 90% to 95% of cases (11,295).
mediastinal parathyroid adenomas are intimately con- As a consequence, in most institutions, no imaging is
nected with the thymus. done prior to surgery. However, the persistence of hyper-
Primary hyperparathyroidism results from a solitary parathyroidism following surgical resection of the cervical
adenoma in approximately 85% of cases. Other causes glands suggests the presence of an ectopic parathyroid
include diffuse hyperplasia (10%), multiple adenomas adenoma or hyperfunctioning gland. Almost 50% of these
(5%), and rarely, carcinoma (1%) (263). Various studies patients will have mediastinal parathyroid glands; two
are available for detecting parathyroid disease. These thirds of these glands will be located in the superior medi-
include high-resolution ultrasonography, radionuclide astinum, particularly in the posterior aspect near the
imaging using thallium or sestamibi, high-resolution tracheoesophageal groove (285), and one third will be in
contrast-enhanced CT, MRI, and selective venous the lower anterior mediastinum, often in relation to the
catheterization (242). thymus, although a wide variety of locations can be seen
(276). Parathyroid adenomas in the APW are rare, but
their appearance has recently been described (97). In a
review of 285 consecutive patients treated surgically for
of Parathyroid Disease
hyperparathyroidism, mediastinal parathyroid tumors
The CT appearance of parathyroid adenomas has been well were present in 22% and were present in 38% of the
described (277–292). Normal glands cannot be identified patients requiring reoperation for persistent or recurrent
on CT. Parathyroid adenomas and hyperplastic glands hyperparathyroidism (275). Preoperative localization
are usually small but vary in size from 0.3 to 3 cm; rarely has also been advocated for patients at a high risk for
are they large enough to be detected on plain radiographs. surgery (293).
When visible on CT, they usually appear homogeneous in Reports comparing the accuracy of CT to competing
density. Adenomas having their vascular supply compro- modalities in postoperative patients vary significantly (281).
mised at surgery may appear cystic (287,288,292). Rarely, Miller et al. (294) compared sonography, thallium scintigra-
parathyroid adenomas appear calcified (289). phy, CT, and MRI in 53 postoperative patients and found
No CT criteria reliably differentiate an adenoma from that no technique detected more than 50% of abnormal
hyperplasia or carcinoma. In the anterior mediastinum glands. Similar results have been reported by others, includ-
parathyroid lesions are usually found in the expected loca- ing false-positive rates of nearly 20%. However, in a recent
tion of the thymus and may be indistinguishable from study (97), CT obtained with 5-mm collimation during the
small thymic remnants, small thymomas, or small lymph infusion of contrast material was positive in eight (89%)
nodes. An important potential source of error is to mistake of nine patients with a parathyroid adenoma in the APW,
a parathyroid adenoma for a thyroid adenoma (253,292). whereas MRI was positive in 63%, arteriography was
An association between thyroid and parathyroid disease is positive in 56%, and sestamibi radionuclide images were
well known; as many as 30% of patients with parathyroid positive in all six patients in whom they were obtained.
disease prove to have synchronous thyroid abnormalities. Also, higher accuracy rates have recently been reported for
In a review of 65 patients in whom parathyroid surgery was MRI and sestamibi scintigraphy (293,296,297).

of Parathyroid Disease
The appearance of abnormal parathyroid glands on MRI
has been well documented (263,265,266,270,271,
276,294,298–306). Similar to thyroid adenomas, most
parathyroid adenomas appear intense on T2-weighted
images (Fig. 4-71), increasing significantly in intensity as
compared with T1-weighted images (Fig. 4-72). Similar
findings have been documented for parathyroid hyperpla-
sia and carcinomas. In a small but significant percentage
of cases, parathyroid adenomas fail to show increased
signal intensity with T2 weighting. As shown by
Auffermann et al. (303), in a study of 30 patients with
recurrent hyperparathyroidism, 13% of the abnormal
glands failed to display high intensity on T2-weighted
images. Enhancement following gadolinium infusion is
typical, and fat suppression images can be valuable in
demonstrating parathyroid adenomas (260,305,306).
Reports of the accuracy of MRI, although variable,
largely show that this modality is superior to other nonin-
vasive techniques for detecting parathyroid pathology,
with the exception of sestamibi radionuclide imaging Figure 4-72 Ectopic parathyroid adenoma. T1-weighted MRI in
a patient with persistent hyperparathyroidism after surgery. A small
(263,265,266,271,293,294,296–303). Spritzer et al. (301), mass is visible in the anterior mediastinum (arrow), anterior to the
using both T1- and T2-weighted scans to evaluate main pulmonary artery. Mediastinal parathyroid adenomas are
23 patients with suspected parathyroid adenomas reported often found in the usual location of the thymus. They are difficult to
distinguish from small nodes or residual thymic tissue.
an overall sensitivity of 78% and a specificity of 95%, with
a resulting overall accuracy of 90%. Kneeland et al. (299),
evaluating 22 patients with hyperparathyroidism, reported In postoperative patients and those with mediastinal
a 74% and 88% sensitivity and specificity of MR imaging, parathyroid glands, MRI has proven more sensitive than
respectively. thallium/technetium scintigraphy or sonography (276). In
A B
Figure 4-71 Parathyroid adenoma: MR evaluation. A, B: T1- and T2-weighted images through the lower poles of the thyroid gland show
a well-defined mass in the region of the tracheoesophageal groove (arrows). Note the marked signal enhancement within the mass on T2-
weighted images. These findings are characteristic of parathyroid adenomas. Unfortunately, this same appearance can be seen in patients
with thyroid adenomas, differentiation of which can be problematic, especially as a significant percentage of patients with hyperparathy-
roidism have concomitant thyroid disease.
a study of 25 patients with mediastinal lesions (276), MRI of thoracic diseases, including bronchogenic carcinoma,
had a sensitivity of 88%, whereas thallium/technetium lymphoma, and granulomatous diseases. The assessment
scintigraphy and sonography had sensitivities of 58% and of mediastinal lymph nodes in patients with bronchogenic
12%, respectively. However, in interpreting these results, it carcinoma is discussed in detail in Chapter 5.
should be kept in mind that (a) false-positive MRI studies
are not uncommon, because of lymph node enlargement
Lymph Node Groups
or thyroid nodules, and (b) the sensitivity of sestamibi
radionuclide imaging is higher than that of thallium Mediastinal lymph nodes are generally classified by loca-
(97,296,297). tion, and most descriptive systems are based on Rouvière’s
It has recently been suggested that a complementary classification of lymph node groups (309). Patterns of
role for MRI and sestamibi imaging is appropriate lymphatic drainage are also considered in defining groups
(293,306). In 25 patients with hyperparathyroidism, of related lymph nodes, an approach that can be helpful
17 of whom were postoperative, MRI and sestamibi imag- in understanding the involvement of particular nodes in
ing had sensitivities of 84% and 79%, respectively, with patients with diseases involving various sites in the chest
specificities of 75% and 94%; when both studies were (70,310).
interpreted in conjunction, their sensitivity was 89% with Intrathoracic lymph nodes are described as belonging to
a specificity of 95% (293). Gotway et al. (306) compared specific node groups, although they freely communicate
the sensitivity and positive predictive value of MRI with each other via numerous lymphatic vessels (71,72).
and technetium 99m 2-methoxyisobutyl-isonitrile (MIBI) Different methods of classification of intrathoracic lymph
scintigraphy for the detection of hyperfunctioning nodes have been used by different authors, although most
parathyroid tissue in 98 consecutive patients with recur- divide lymph nodes into parietal and visceral groups, based
rent or persistent hyperparathyroidism after surgery. In on their location and structures they drain. The parietal
these patients, 130 abnormal parathyroid glands were lymph nodes are related to the chest wall and lie outside the
identified at surgery. No significant difference was found parietal pleura; they primarily drain structures of the chest
between the sensitivity and positive predictive values of wall and are classified as internal mammary, diaphragmatic,
MR imaging (82% and 89%, respectively) and those for or paracardiac, and intercostal. Visceral node groups are
99mTc MIBI scintigraphy (85% and 89%). However, the located within the mediastinum or are related to the lung
sensitivity and positive predictive value for the detection or hila; they drain the lungs and mediastinal structures,
of abnormal parathyroid tissue on a per-gland basis and include intrapulmonary, bronchopulmonary, tracheo-
increased to 94% and 98%, respectively, when only one of bronchial, paratracheal, paraesophageal, and anterior
the two tests was required to be positive. mediastinal groups.
Given the limited accuracy rates of the various nonin- The following classification is based on a modification
vasive imaging modalities in patients with hyperparathy- of well-recognized anatomic descriptions of lymphatic
roidism, each case must be individualized to select the anatomy (309,311), although some terms have been
most appropriate study. To some extent, availability, cost, modified to be consistent with current usage, and
and familiarity with these procedures will determine their anatomic descriptions emphasize the localization of
usage. In our experience, MRI has come to replace CT in lymph nodes on imaging studies, rather than considering
the investigation of patients with recurrent hyperparathy- nodes as parietal or visceral. Anterior, tracheobronchial,
roidism following surgery, and sestamibi radionuclide and posterior nodes are considered.
imaging is assuming an increased role. CT is reserved for
those patients with suspected recurrent parathyroid carci-
Anterior Lymph Nodes
noma, especially to evaluate the lungs and the liver to rule
out metastatic disease (307). Another specialized use of Internal mammary lymph nodes are located in a retroster-
CT is to evaluate patients for whom angiographic ablation nal position, at the anterior ends of the intercostal spaces,
of mediastinal parathyroid adenomas has been performed near the internal mammary artery and veins; they are con-
(Fig. 4-73) (308). sidered to be part of the parietal lymph node group. They
drain the anterior chest wall, anterior diaphragm, and
medial breasts and freely communicate with prevascular
MEDIASTINAL LYMPH NODES lymph nodes and paracardiac or diaphragmatic lymph
AND LYMPH NODE MASSES nodes (312,313). They are most often enlarged as a result of
lymphoma (Fig. 4-74) or metastatic breast cancer (314,315).
Mediastinal lymph node abnormalities can be seen in any Prevascular lymph nodes lie anterior to the aorta and
mediastinal compartment, although they most commonly in relation to the great vessels (Fig. 4-75). These nodes
involve middle mediastinal regions such as the pretracheal drain most anterior mediastinal structures including the
space, APW, and subcarinal space (82). Their detection pericardium, thymus, thyroid, pleura, and the anterior
and diagnosis are important in the evaluation of a number hila. They represent visceral nodes. They communicate
A, B C
Figure 4-73 Mediastinal parathyroid adenomas: angio-

graphic ablation. A: Enlargement of a contrast-enhanced
CT scan through the aortic arch shows a well-defined, ho-
mogeneous soft tissue mass, otherwise indistinguishable in
appearance from a thymoma (arrow). B: Subtraction
angiogram demonstrates an anterior mediastinal parathy-
roid adenoma supplied by a descending branch of the right
inferior thyroid artery. C: Appearance of the adenoma after
selective intra-arterial injection of 180 mL of contrast mate-
rial. D: CT section obtained 24 hours later in the same
patient without additional contrast material shows dense,
persistent staining of the adenoma. (From Miller DL,
Doppman JL, Chang R, et al. Angiographic ablation of
parathyroid adenomas: lessons from a 10-year experience.
D Radiology. 1987;165:601–607, with permission.)
with the internal mammary chain of nodes anteriorly and nodes, paracardiac nodes are most commonly enlarged in
paratracheal and aortopulmonary lymph nodes posteri- patients with lymphoma and metastatic carcinoma, partic-
orly. They may be involved in a variety of diseases, ularly breast cancer (312,313,319,320). Paracardiac lymph
notably lymphoma and granulomatous diseases, but their nodes correspond to the anterior (prepericardiac) and
involvement in lung cancer is relatively uncommon. middle (juxtaphrenic) subgroups of the diaphragmatic
Paracardiac or cardiophrenic angle lymph nodes parietal lymph node group (309,311). Prepericardiac
(316–318) lie anterior to or lateral to the heart and peri- nodes are located posterior to the xiphoid process and
cardium, on the surface of the diaphragm (Fig. 4-74B slightly lateral to it. Juxtaphrenic lymph nodes are situated
and C). They communicate with the lower internal mam- adjacent to the pericardium, where the phrenic nerves
mary chain and drain the lower intercostal spaces, peri- meet the diaphragm. From a clinical standpoint, there is
cardium, diaphragm, and liver. As with internal mammary little reason to distinguish between them.
B C
Figure 4-74 Anterior lymph node enlargement in non-Hodgkin lymphoma. A: A right internal mammary lymph node is abnormally
enlarged (arrow). B, C: At lower levels, multiple paracardiac lymph nodes are visible (arrows) anterior to the heart and pericardium.
Tracheobronchial Lymph Nodes lymph nodes are subdivided into a number of important
node groups, which are all closely related. Paratracheal
Tracheobronchial lymph nodes generally serve to drain nodes lie anterior to, and on either side of, the trachea,
the lungs. Lung diseases (e.g., lung cancer, sarcoidosis, TB, thus occupying the pretracheal (or anterior paratracheal)
fungal infections) that secondarily involve lymph nodes space (Figs. 4-47, 4-75 to 4-77) (321,322). Retrotracheal
typically involve these lymph nodes. Tracheobronchial nodes may also be seen (Fig. 4-75). The most inferior
node in this region is the so-called azygos node, medial to
the azygos arch. These nodes form the final pathway for
lymphatic drainage from most of both lungs, excepting
the left upper lobe (310). Because of this, they are com-
monly abnormal regardless of the location of the lung
disease.
Aortopulmonary nodes are grouped by Rouvière with
prevascular nodes, but because they serve the same function
on the left as paratracheal nodes on the right, and freely
communicate with paratracheal nodes, it is most
appropriate to group them together. They lie in the APW,
lateral to the left main bronchus and between the aorta and
pulmonary artery (Fig. 4-77). The left upper lobe drains via
this node group.
Peribronchial nodes surround the main bronchi on
Figure 4-75 Anterior mediastinal lymph node enlargement in each side and lie between the main bronchi in the subcari-
Hodgkin disease. Bulky mediastinal lymph nodes involve the nal space. These drain the lungs. Bronchopulmonary
prevascular space (arrows) anterior to the great vessels. Right
paratracheal and retrotracheal lymph nodes (large arrow) are also nodes are located distal to the main bronchi and are
markedly enlarged. usually considered to be hilar.
Figure 4-76 Paratracheal lymph node enlargement in metasta-

tic carcinoma. A lymph node in the pretracheal or paratracheal
space is markedly enlarged (arrows), associated with anterior
displacement of the superior vena cava.
Figure 4-78 Subcarinal and lymph node enlargement with non-
Hodgkin lymphoma. Subcarinal and paraesophageal lymph node
enlargement is visible (arrow).
Subcarinal nodes represent peribronchial nodes lying
between the main bronchi in the subcarinal space (Fig. 4-78)
(65,323). These nodes drain the inferior hila and lower posterior mediastinal lymph nodes. They are part of
lobes on both the right and left and communicate in turn the parietal group.
with the right paratracheal chain. Retrocrural lymph nodes lie posterior to the diaphrag-
matic crura (Fig. 4-80). They communicate with lumbar
nodes and posterior mediastinal lymph nodes and drain
Posterior Lymph Nodes
the diaphragm and liver, and represent the posterior group
Paraesophageal and inferior pulmonary ligament nodes of diaphragmatic parietal lymph nodes.
are associated with the esophagus and descending aorta
and lie medial to the inferior pulmonary ligament (Figs.
Lymph Node Stations: American Thoracic
4-78 and 4-79). They represent visceral nodes and drain
the medial lower lobes, esophagus, pericardium, and pos-
Society and American Joint Committee on
terior diaphragm. On the right, they are impossible to dis- Cancer/Union Internationale Contre le Cancer
tinguish from subcarinal nodes, unless they are near the Classifications
diaphragm. In the 1970s, the American Joint Committee on Cancer
Intercostal and paravertebral lymph nodes are found in (AJCC) and the Union Internationale Contre le Cancer
the posterior intercostal spaces and adjacent to thoracic (UICC) introduced a numeric system for localization of in-
vertebral bodes (Fig. 4-79). These drain the posterior trathoracic lymph nodes for the purpose of lung cancer
pleura, chest wall, and spine and communicate with other staging. Lymph nodes were described relative to regions in
the mediastinum termed “lymph node stations.” The
AJCC/UICC node mapping system was modified in 1983
by the American Thoracic Society (ATS) to more precisely
define anatomic and CT criteria for each station (324), and
the ATS classification system has been in common use since
then. In 1997, the AJCC/UICC published a further revision
intended to be a compromise between the AJCC and ATS
classifications (325). In this system, nodes are classified as
belonging to one of 14 node stations, for which medi-
astinoscopic, surgical, and CT criteria (326) have been es-
tablished (see Chapter 7 for an in-depth discussion).
This system is not generally used for the clinical descrip-
tion of lymph node abnormalities in diseases other than
lung cancer. Nonetheless, familiarity with this classifica-
tion is encouraged (see Chapter 7), and specific patterns of
thoracic lymphatic spread have been described relative to
these node stations for various thoracic tumors and a
Figure 4-77 Paratracheal and aortopulmonary lymph node number of diseases affecting mediastinal lymph nodes
enlargement in non-Hodgkin lymphoma. Pretracheal lymph node
enlargement (large arrow) and enlargement of the aortopulmonary (71,72). Table 4-16 may be used for reference and provides
lymph node (small arrow) are both visible. a comparison of ATS and AJCC/UICC criteria.
A
B
Figure 4-79 Posterior lymph node enlargement in metastat-

ic carcinoma. A: An enlarged lymph node (arrow) is visible in
the right paraesophageal region. Pleural effusions are also
seen. B: At a lower level, paraesophageal and para-aortic
lymph nodes (large arrows) and paravertebral lymph nodes
(small arrows) are enlarged. C: At a level below B, enlarged
paravertebral lymph nodes (arrows) are visible. They are higher
C in attenuation than the pleural fluid.
apply. Subcarinal nodes can be quite large in normal

patients. Pretracheal nodes are also commonly visible,
but these nodes are typically smaller than normal subcari-
nal nodes. Nodes in the supra-aortic mediastinum are
usually smaller than lower pretracheal nodes, and left
paratracheal nodes are usually smaller than right paratra-
cheal nodes.
Numerous reports have addressed the issue of what
constitutes the normal size, number, and appearance of
mediastinal lymph nodes (66,67,321,330–332). The short
axis or least diameter (i.e., the smallest node diameter seen
in cross section) of a lymph node should generally be used
when measuring its size (Fig. 4-82). This measurement
Figure 4-80 Retrocrural lymph node enlargement with non- more closely reflects actual node diameter when nodes are
Hodgkin lymphoma. A right retrocrural mass (arrow) is visible in
the para-aortic and paravertebral region.
obliquely oriented relative to the scan plane and shows
less variation among normal subjects than does the long
axis or greatest diameter.
Computed Tomography Appearance As determined both in vivo and by autopsy evaluation,
of Normal Lymph Nodes the most useful upper limit of normal for mediastinal
lymph nodes is 1.0 cm, as measured in its short axis
In autopsy series, the average number of mediastinal lymph (67,319,330), except in the subcarinal region. In studies by
nodes is 64 (327). Almost 80% of mediastinal lymph Glazer et al. (66) and Genereux and Howie (332), 95% or
nodes are located in relation to the trachea and main more of lymph nodes in normal subjects measured 1.0 cm
bronchi and serve to drain the lungs. or less in the lower paratracheal (4R), right paratracheal
On CT, lymph nodes are generally visible as (a) dis- (10R), and APW (5) regions.
crete and surrounded by mediastinal fat; (b) round, el- However, there are significant variations in normal node
liptical, or triangular; and (c) of soft tissue attenuation size, depending on the precise location of the node. In an
(Figs. 4-74 to 4-79). The node hilum can sometimes evaluation of 56 normal patients by Glazer et al. (66), the
be seen to contain a small quantity of fat; this is largest lymph nodes visible using CT were subcarinal
most apparent when scans are obtained with thin (least diameter 6.2 mm, SD 2.2 mm) and lower right tra-
collimation. cheobronchial (least diameter 5.9 mm, SD 2.1 mm)
Lymph nodes can usually be distinguished from vessels
by their location. However, the ability to recognize and
correctly identify lymph nodes is directly related to the
amount of mediastinal fat surrounding them; in patients
having little mediastinal fat, lymph nodes can be difficult
to distinguish from vessels without contrast infusion. In
many parts of the mediastinum, lymph nodes often occur
in clusters of a few nodes of similar size.
A number of pitfalls in diagnosing lymphadenopathy
have been described. Structures that can be mistaken
for enlarged lymph nodes include both normal and anom-
alous vascular structures, such as an aberrant right subcla-
vian artery or high-riding left pulmonary artery (328,329)
and prominent pericardial recesses, especially the superior
recess of the pericardium (Fig. 4-81) (62). It should be
emphasized that viewing contiguous scans often can
be helpful in differentiating lymph nodes from vessels;
mediastinal vessels are usually visible on several adjacent
scans, whereas lymph nodes are not, unless they are quite
large. In difficult cases, the use of bolus intravenous con-
trast material injection will prove diagnostic.
Internal mammary nodes, paracardiac nodes, and
paravertebral nodes are not commonly seen on CT in nor- Figure 4-81 Normal superior pericardial recess. A low-attenua-
mal persons, but in other areas of the mediastinum, tion arcuate opacity contiguous with the posterior wall of the
ascending aorta (arrows) represents the superior pericardial recess.
normal lymph nodes are often visible. The size of normal Its appearance, location, and low attenuation are characteristic,
nodes varies with their location, and a few general rules allowing it to be distinguished from lymph node enlargement.
TABLE 4-16
COMPARISION OF ATS AND AJCC LYMPH NODE STATIONS
Node ATS Node ATS ATS AJCC/UICC AJCC/UICC AJCC/UICC
Group Station Designation Anatomic Criteria Station Designation Anatomic Criteria
Paratracheal 1 Highest Cranial to the superior aspect

mediastinal of L brachiocephalic vein
Paratracheal 2R R upper R of the tracheal midline, 2 Upper Below station 1 and cranial
paratracheal between the lung apex paratracheal to superior aspect of the
and the caudal margin aortic arch
of the innominate artery
(or for radiologists, the
superior aspect of the
aortic arch as with 2L)
2L L upper L of the tracheal midline,
paratracheal between the lung apex and
the superior aortic arch
Prevascular 3 Prevascular Anterior to the great vessel

branches and cranial to
aortic arch
Paraesophageal Retrotracheal Posterior to the trachea and

cranial to the inferior aspect
of azygos arch
Paratracheal 4R R lower R of the tracheal midline, 4R R lower R of tracheal midline, below 2,

paratracheal below 2R and above the paratracheal and cranial to the RUL
azygos arch bronchus (this equals
ATS 4R 10R)
4L L lower L of the tracheal midline, 4L L lower L of the tracheal midline, below
paratracheal below 2L, cephalad to the paratracheal 2, and cranial to the LUL
carina, and medial to the bronchus (this equals ATS
ligamentum arteriosum 4L 10L)
Aortopulmonary 5 Aorto- Subaortic and para-aortic 5 Subaortic or Lateral to the ligamentum

pulmonary nodes lateral to the aorto- arteriosum, aorta, or LPA,
ligamentum arteriosum, pulmonary proximal to the first branch
aorta, or LPA, proximal to of the LPA within the
the first branch of the LPA mediastinal pleural envelope
Prevascular 6 Anterior Anterior to aortic arch or 6 Para-aortic Anterior and lateral to the
innominate artery (including (ascending ascending aorta and the
some pretracheal and aortic or aortic arch and innominate
preaortic nodes) phrenic) artery, caudal to the superior
aspect of aortic arch
(corresponding to AJCC/UICC nodal stations 4R and 10R, Computed Tomography Diagnosis of Lymph
respectively), whereas upper paratracheal nodes (AJCC/ Node Abnormalities
UICC 2) are smaller than lower paratracheal nodes (4R),
and right-sided nodes in general are smaller than those on CT is restricted to providing the following information re-
the left side (Table 4-17). Kiyono et al. (67) reported garding mediastinal lymph nodes: (a) precise measurements
very similar findings in an evaluation of 40 adult cadavers of lymph node diameter, (b) delineation of lymph node
(Table 4-18). In their series, mean short transverse nodal morphology, and (c) characterization of lymph node density
diameters ranged from 2.4 to 5.6 mm, with substantial and the internal characteristics, both before and after intra-
variations noted depending on nodal station, subcarinal venous contrast enhancement. Using CT, lymph nodes can
nodes again being largest. Based on these findings, these be characterized as homogeneous in density, calcified, low-
authors recommended using short axis measurements of density and necrotic, or enhancing after contrast infusion.
12 mm as upper limits of normal for subcarinal lymph CT can also be valuable in localizing abnormal lymph
nodes, 10 mm for right tracheobronchial and low paratra- nodes and in determining their extent and relationship to
cheal lymph nodes, and 8 mm for all other nodal groups mediastinal structures such as the superior vena cava,
(67). Receiver operating characteristic curve analysis of pulmonary arteries, airways, and the esophagus. In combi-
node size performed by Glazer et al. (66) suggests that the nation, these findings frequently limit the range of differ-
optimal size for upper limits of normal for mediastinal ential diagnoses; they also can allow a determination of
lymph nodes in patients suspected of malignancy is the best surgical or CT-guided approach, when necessary,
1.0 cm, except in the subcarinal region. to establish a histologic diagnosis.
TABLE 4-16
(continued)
Node ATS Node ATS ATS AJCC/UICC AJCC/UICC AJCC/UICC
Group Station Designation Anatomic Criteria Station Designation Anatomic Criteria
Subcarinal 7 Subcarinal Caudal to the carina but 7 Subcarinal Caudal to the carina but not
not associated with the associated with the lower lobe
lower lobe bronchi or bronchi or pulmonary arteries
arteries within the lung within the lung
Paraesophageal 8 Paraesopha- Dorsal to the posterior 8 Paraesopha- Adjacent to the wall of the
geal wall of trachea, and on geal esophagus and to the right
either side of the or left of esophagus
esophagus (not subcarinal
nodes)
Inferior 9 R or L In relation to right or 9 Pulmonary Within the pulmonary ligament,

pulmonary pulmonary left inferior pulmonary ligament including those in the
ligament ligament ligaments posterior wall and lower
part of the inferior pulmonary
vein
Peribronchial 10R Tracheo- R of the tracheal midline, 10R R hilar Caudal to the superior RUL
(hilar) bronchial caudal to 4R and above bronchus, adjacent to the R
the origin of the RUL main bronchus or proximal
bronchus bronchus intermedius
10L Peribronchial L of the tracheal midline, 10L L hilar Caudal to the superior LUL
between the carina and bronchus, adjacent to L main
the LUL bronchus, bronchus
medial to the ligamentum
arteriosum
Broncho- 11 Intrapulmonary Nodes removed at 11 Interlobar Between lobar bronchi and

pulmonary pneumonectomy or distal adjacent to proximal lobar
(hilar) to the mainstem bronchi bronchi
or secondary carina
(includes interlobar, lobar,
and segmental nodes)
Lobar 12 May be determined post- 12 Lobar Adjacent to distal lobar bronchi

thoracotomy
Segmental 13 May be determined post- 13 Segmental Adjacent to segmental bronchi

thoracotomy
14 Subsegmental Adjacent to subsegmental

bronchi
ATS, American Thoracic Society; AJCC, American Joint Committee on Cancer; UICC, Union Internationale Contre le Cancer; R, right; L, left; RUL,
right upper lobe; LPA, left pulmonary artery; LUL, left upper lobe.
Figure 4-82 Short axis lymph node

diameter in a patient with metastatic lung
carcinoma. A: Pretracheal and prevascular
lymph node enlargement is present. B:
The short axis or least diameter of the
enlarged pretracheal lymph node is indi-
cated by the arrows. This measurement is
generally used for assessing lymph node
size. In this case, the short axis lymph
A, B node diameter is 2 cm.
TABLE 4-17
LYMPH NODE SIZE BY ATS LYMPH NODE STATION, MEASURED USING COMPUTED TOMOGRAPHY
IN 56 NORMAL SUBJECTS
Maximum Short Axis Upper Limits
Patients with Number of Number of Diameter of Normal
ATS Station Nodes (%) Nodes SD Nodes SD (mm) (mm) (mean 2 SD)
2R 95 2.1 1.3 6 3.5 1.3 6.1

2L 75 1.9 1.6 6 3.3 1.6 6.5
4R 100 3.2 2.0 10 5.0 2.0 9.0
4L 84 2.1 1.6 7 4.7 1.9 8.5
5 59 1.2 1.1 3 4.7 2.1 8.9
6 86 4.8 3.5 12 4.1 1.7 7.5
7 95 1.7 1.1 6 6.2 2.2 10.6
8R 57 1.0 1.1 4 4.4 2.6 9.6
8L 45 0.8 1.2 6 3.8 1.7 7.2
10R 100 2.8 1.3 7 5.9 2.1 10.1
10L 70 1.0 0.8 3 4.0 1.2 6.4
ATS, American Thoracic Society; R, right; L, left.

Modified from Glazer GM, Gross BH, Quint LE, et al. Normal mediastinal lymph nodes: number and size according to American Thoracic Society
mapping. AJR Am J Roentgenol. 1985;144:261–265, with permission.
It has been emphasized recently that there are different Lymph Node Enlargement
thoracic lymphatic drainage pathways in the thorax rele-
vant in the staging of lung cancer, breast cancer, lym- Enlargement of lymph nodes is the most common CT find-
phoma, esophageal cancer, and malignant mesothelioma. ing in a number of neoplastic and inflammatory diseases
To properly search for metastatic spread, it is important (Figs. 4-74 to 4-79 and 4-82). As indicated previously,
to carefully evaluate the specific nodal stations that drain a short axis node diameter exceeding 1 cm is generally
the thoracic structures from which a primary tumor origi- thought to indicate abnormal node enlargement, except in
nates (70–72). the subcarinal space. However, it is important to understand
TABLE 4-18
LYMPH NODE SIZE BY ATS LYMPH NODE STATION, MEASURED IN 40 CADAVERS
Maximum Short Axis Upper Limits
Patients with Number of Number of Diameter of Normal (mm)
ATS Station Nodes (%) Nodes SD Nodes (mm) (mean 2 SD)
2R 80 2.5 2.2 11 3.4 7.8

2L 68 2.1 2.2 7 2.8 5.6
4R 98 4.8 2.8 11 3.4 9.2
4L 98 4.5 2.9 16 3.6 9.2
5 58 1.1 1.4 6 3.3 8.5
6 85 4.7 3.9 15 3.0 7.2
7 100 2.9 1.4 6 5.5 12.3
8R 58 1.2 1.4 6 3.6 8.2
8L 50 1.1 1.4 5 2.4 6.1
9R 10 0.1 0.4 2 2.3 3.9
9L 35 0.5 0.8 3 3.1 6.5
10R 95 3.5 2.3 10 4.0 10.8
10L 90 2.4 1.9 7 3.2 6.8

Modified from Kiyono K, Sone S, Sakai F, et al. The number and size of normal mediastinal lymph nodes: a postmortem study. AJR Am J Roentgenol.
1988;150:771–776, with permission.
that a simple assessment of lymph node size has a limited in patients with bronchogenic carcinoma; in this setting
accuracy in determining whether mediastinal nodes are the likelihood that a mediastinal node is involved by
normal or abnormal. Early in the course of a number of tumor is directly proportional to its size, and no node
diseases, significant nodal pathology may be present in the diameter is clearly able to separate normal from abnormal
absence of lymph node enlargement, and normal-sized nodes (Table 4-19) (338,339).
lymph nodes do not rule out a pathologic process. For The significance given to the presence of a minimally
example, as many as 40% of patients with lymph node enlarged lymph node must be tempered by a knowledge of
metastases from bronchogenic carcinoma show no evidence the patient’s clinical situation. For example, if the patient is
of lymph node enlargement on CT (333–339). Further- known to have lung cancer, then an enlarged lymph node
more, it must also be recognized that mediastinal lymph has a significant likelihood of being involved by tumor.
nodes can be enlarged in the absence of significant disease, However, the same node in a patient without lung cancer
usually as a result of hyperplasia. In patients with bron- is much less likely to be of clinical significance. In the
chogenic carcinoma, approximately 30% of patients with- absence of a known disease, an enlarged node must be
out evidence of lymph node metastases at surgery show regarded as likely to be hyperplastic or postinflammatory.
some evidence of lymph node enlargement on CT. Lymph nodes having a short axis of 2 cm or more often
In general, enlarged mediastinal lymph nodes are well reflect the presence of neoplasm, such as metastatic tumor
seen using both contrast-enhanced and unenhanced CT, or lymphoma, sarcoidosis, or infection and should always
although contrast-enhanced CT may reveal more enlarged be treated as potentially significant. Although mediastinal
lymph nodes in specific regions (31). In one study (31), lymph nodes become enlarged in a variety of noninfec-
50 patients with known or suspected bronchogenic tious and nongranulomatous inflammatory diseases, they
carcinoma had both unenhanced and contrast-enhanced are usually smaller than 2 cm (340–342).
thoracic CT. Three observers identified enlarged lymph It should also be recognized that although the least node
nodes (short axis 10 mm) and assigned the enlarged diameter is most frequently used to assess node diameter
nodes to ATS nodal stations. When all lymph node and lymph node enlargement, it has been shown in several
stations were taken together, intraobserver agreement studies that the long axis or greatest node diameter can also
between contrast-enhanced and unenhanced studies was be used, with very similar accuracy rates. In patients with
almost perfect (k range, .85 to .94), and no difference was bronchogenic carcinoma, using a least node diameter of
found for any observer in the proportion of patients with 10 mm and a greatest node diameter of 15 mm had identi-
at least one enlarged lymph node. However, the number of cal sensitivities and specificities for diagnosis of tumor
enlarged lymph nodes identified using contrast-enhanced involvement of nodes (335). Generally speaking, however,
CT was 11% higher than on unenhanced studies (418 vs. using greatest node diameter should be avoided if a node
377; p .044). In particular, significantly more enlarged appears much longer than it is wide (i.e., its greatest node
nodes were recognized in the right upper paratracheal diameter is much longer than its least diameter).
lymph node station (2R) using contrast-enhanced CT Lymph node enlargement may be seen on CT in patients
(i.e., 68 vs. 44 on unenhanced scans; p .014). with congestive heart failure (343,344). In a study of
In general, the diameter of an enlarged lymph node cor- 46 patients who had thoracic CT during a bout of con-
relates with its likelihood of harboring significant or active gestive heart failure (344), CT showed typical findings
disease, and the larger the node, the more likely it indi- of septal thickening, bilateral pleural effusions, vascular
cates a significant abnormality. This has best been studied redistribution, and cardiac enlargement. In addition,
TABLE 4-19
RELATIONSHIP OF LEAST NODE DIAMETER TO LIKELIHOOD OF
INVOLVEMENT BY TUMOR IN 143 PATIENTS WITH LUNG CANCER
Number of Nodes Least Node Diameter (cm) Percent Malignant
336 1 13
57 1.0–1.9 25
13 2.0–2.9 62
6 3.0–3.9 67
2 4.0 or more 100
Modified from McLoud TC, Bourgouin PM, Greenberg RW, et al. Bronchogenic carcinoma: analysis of staging
in the mediastinum with CT by correlative lymph node mapping and sampling. Radiology. 1992;182:319–323,
with permission.
enlarged mediastinal lymph nodes and hazy mediastinal inflammatory reaction. This appearance is most typical
fat were seen in 55% and 33% of cases, respectively. In a of infection, granulomatous disease, and neoplasm.
cohort of 17 patients with elevated pressures in the pul- 3. Diffuse mediastinal involvement. Mediastinal connective
monary capillary wedge documented within 24 hours of tissue and fat are diffusely infiltrated, with no recogniza-
CT, CT scans revealed lymphadenopathy in 14 patients ble nodes or node masses. Typically, mediastinal fat
(82%) and inhomogeneous fat in 10 patients (59%) appears to be replaced by soft tissue density (Fig. 4-83C).
(344). This diagnosis may be difficult to make unless the atten-
uation of mediastinal soft tissues is compared to the
attenuation of subcutaneous fat on the same scan—they
Computed Tomography Assessment
should be similar. This pattern suggests lymphoma,
of Lymph Node Morphology undifferentiated carcinoma, generalized infection, or
In patients with lymph node abnormalities, three patterns granulomatous mediastinitis.
of involvement can be seen on CT. This pattern may be of
some value in differential diagnosis. These patterns are as
Computed Tomography Assessment
follows:
of Lymph Node Attenuation
1. Discrete enlarged nodes. Nodes are enlarged, as
CT can also be used to define the density of lymph nodes,
determined by measurement of least node diameter,
both before and after the injection of intravenous contrast
and individual nodes are surrounded by fat and appear
media. Although enlarged nodes most often have a
distinct (Figs. 4-74 to 4-77 and 4-82). Although adja-
nonspecific appearance, being of soft tissue attenuation,
cent nodes may contact each other, with focal obscura-
nodes can be calcified or low in density and necrotic in
tion of their margins, they remain well defined. This
appearance or can opacify following the administration of
pattern can be seen in association with all causes of
a bolus of intravenous contrast media.
mediastinal lymph node enlargement.
2. Coalescence of enlarged nodes. With progression of dis-
Calcified Lymph Nodes
ease, a pathologic process involving several contiguous
nodes may also involve surrounding mediastinal fat, Calcification of lymph nodes is most frequently seen
and several adjacent nodes may fuse to form a single in patients with granulomatous diseases, including TB,
larger mass. Poor definition of node margins or poor histoplasmosis and other fungal infections, and sarcoido-
definition of the margins of the coalescent node mass sis (Figs. 4-84 and 4-85) but can also be seen in other
can indicate extension of the disease process through diseases (Table 4-20) (345–348). Calcification can be
the node capsule (Fig. 4-83) or an associated fibrotic or dense, involving the node in a homogeneous fashion,
A, B C
Figure 4-83 Mediastinal lymph node metastases with extranodal involvement and mediastinal infiltration in small cell lung carcinoma.
A: A right hilar mass is present. Pretracheal lymph nodes (arrows) are enlarged, but their margins are poorly defined. B: At the level of the
carina, several contiguous enlarged pretracheal and aorticopulmonary lymph nodes are visible, but their margins are poorly defined. Note
that the superior vena cava is narrowed and its posterior wall is irregular (arrow), suggestive of direct invasion. C: At the level of the right
upper lobe bronchus, the precarinal mediastinal fat is infiltrated by tumor. Individual lymph nodes are not visible.
A B
C D
Figure 4-84 Calcified lymph nodes: CT assessment. A: Nonenhanced CT section through the carina shows scattered calcified hilar and
pretracheal lymph nodes in a patient with transbronchial, biopsy-documented sarcoidosis. Although these nodes frequently have been
described as containing egg-shell calcifications, central or bull’s-eye calcifications can be identified within enlarged nodes (arrows) as
frequently. B: Nonenhanced CT section through the aortic arch in a patient with documented silicosis. Calcified paratracheal nodes are
easily identified, in this case associated with dense foci of parenchymal calcifications in the posterior segments of both upper lobes caused
by early progressive massive fibrosis (arrows). In addition, air can be seen in both enlarged left hilar and prevascular nodes (curved arrows).
This unusual appearance is secondary to superimposed tuberculosis (silicotuberculosis). Gas within these nodes is presumably secondary to
necrosis of peribronchial nodes, with resultant fistulization to adjacent airways. C: Contrast-enhanced CT section just below the carina.
Amorphous calcifications can be identified within markedly enlarged pretracheal and prevascular lymph nodes (arrows). A large, partially
loculated pleural fluid collection is present on the right, within which foci of soft tissue density can be identified (curved arrows). This patient
has known mucin-producing colon carcinoma, metastatic to both the pleura and mediastinal nodes. Diagnosis was confirmed by medi-
astinoscopy. D: Nonenhanced CT section through the great vessels shows diffuse calcifications throughout pre- and paratracheal nodes.
This patient had AIDS and was being treated prophylactically with aerosolized pentamidine. Mediastinoscopy revealed these nodes to be
filled with Pneumocystis carinii organisms, a finding increasingly being observed in patients treated with aerosolized pentamidine.
stippled, “egg-shell” in appearance, or faint and cloudlike. radiation, and has occasionally been described in patients
The abnormal nodes are often enlarged but can also be with blastomycosis, histoplasmosis, amyloidosis, sclero-
of normal size. Multiple calcified lymph nodes are often derma, and Castleman disease (349–351).
visible, usually in contiguity, and hilar lymph node calcifi- Rarely, node calcification is seen in untreated lymphoma
cations are often associated. (346,348) or as a result of metastatic carcinoma, typically
Dense calcification involving all or most of an abnormal adenocarcinoma, and most frequently from a colonic
node is typical of previous granulomatous infection or sar- primary tumor (Figs. 4-84C and 4-85). Lymph node calci-
coidosis. Egg-shell calcification is defined by the presence fication has also been reported with metastatic bron-
of calcium in the node periphery, which is often ringlike, choalveolar carcinoma, in relation to psammoma bodies
but central calcification can be present as well. Egg-shell (347). Usually calcification of tumors is stippled or faint
calcification is most often seen in patients with either and cloudlike. Calcification may also be seen in patients
silicosis or coal workers’ pneumoconiosis, sarcoidosis, and with metastatic osteogenic sarcoma, as well as in patients
TB, but it also occurs in patients with HL, usually following with primary intrathoracic extraosseous osteogenic sarco-
A B
Figure 4-85 Egg-shell lymph node calcification. A, B: CT sections

through the right paratracheal and subcarinal spaces, respectively,
show enlarged nodes with peripheral curvilinear calcifications, con-
sistent with history of treatment for known metastatic mucinous
adenocarcinoma. C: CT section at the level of the aorticopulmonary
window in a different patient than in A and B again shows character-
istic egg-shell calcifications within nodes, in this case in a patient un-
dergoing therapy for known metastatic gallbladder carcinoma.
C
mas (352,353). Calcified hilar and mediastinal lymph calcified nodes can remain functional and can become in-
nodes have been observed in acquired immune deficiency volved by other processes, such as metastatic neoplasm.
syndrome (AIDS) patients with Pneumocystis carinii infec-
tion, especially in patients receiving prophylaxis with
Low-Attenuation or Necrotic Lymph Nodes
aerosolized pentamidine (Fig. 4-84D), possibly as a result
of a necrotizing vasculitis with resultant tissue infarction After administration of intravenous contrast medium, low-
and dystrophic calcification (354). attenuation lymph nodes, with or without an enhancing
Although calcified lymph nodes generally indicate the rim, have been described as characteristic of a number of
presence of old disease, it should be kept in mind that densely different pathologic entities (Table 4-21) (84). Typically,
low-density nodes reflect the presence of necrosis and are
TABLE 4-20
CALCIFIED LYMPH NODES: DIFFERENTIAL TABLE 4-21
DIAGNOSIS LOW-ATTENUATION AND NECROTIC LYMPH
Common NODES: DIFFERENTIAL DIAGNOSIS
Infectious granulomatous diseases
Tuberculosis Common
Fungal infections (histoplasmosis) Infectious granulomatous diseases
Sarcoidosis Tuberculosis
Silicosis Fungal infections (histoplasmosis)
Hodgkin lymphoma (almost always following treatment) Metastases
Lung cancer
Rare
Germ-cell tumors
Pneumocystis carinii pneumonia
Lymphoma
Metastases (e.g., mucinous adenocarcinoma, thyroid carcinoma)
Amyloidosis Rare
Scleroderma Whipple disease
Castleman disease Sarcoidosis
commonly seen in patients with tuberculous, fungal infec- Also, in pediatric patients with primary TB, the finding of
tions, and neoplasms, such as metastatic carcinoma and “ghostlike” ring enhancement has been reported as very
lymphoma (Figs. 4-86 to 4-88) (32,33,355,356). common (34).
Necrotic lymph nodes are common in patients with In patients with germ-cell tumors, including semi-
active TB. In a study by Im et al. (32), the CT scans of noma, low-attenuation areas usually result from exten-
23 patients with tuberculous lymphadenitis were revi- sive tissue necrosis, although low density may also corre-
ewed. On CT, right paratracheal and tracheobronchial late with the presence of numerous small epithelial-lined
node enlargement predominated. After injection of intra- cystic spaces (Fig. 4-88) (357,358). Low-density nodes
venous contrast medium, nodes larger than 2 cm in diam- occur in patients with metastatic lung cancer (Figs. 4-86C
eter invariably showed central areas of low attenuation and 4-87), when the primary tumor also appears
with peripheral rim enhancement (Fig. 4-86A and B), and necrotic, and in metastatic ovarian, thyroid, and gastric
the enhancing walls were usually irregular in thickness. carcinomas. Low-attenuation or necrotic lymph nodes
Smaller lymph nodes showed varying degrees of enhance- are common in patients with lymphoma, both before
ment. In our experience, this appearance is particularly and after treatment. An incidence of 10% to 21% has
common in patients with infection associated with AIDS. been reported (359,360). They have also been described
A B
C D
Figure 4-86 Low-density masses and necrotic lymph nodes: CT assessment. A: Contrast-enhanced CT section through the great vessels.
Massively enlarged low-density paratracheal and prevascular lymph nodes are present, many of which have a distinct enhancing capsule
(arrow). This appearance is especially suggestive of granulomatous infections such as tuberculous or cryptococcal infection, particularly in
patients with AIDS. Tuberculous adenopathy in this case was established by mediastinoscopy. B: Contrast-enhanced CT section at the level
of the great vessels shows several low-attenuation lymph nodes in the left paratracheal space with distinct rim enhancement. Tuberculous
lymphadenitis was verified in this HIV patient by transbronchial needle aspiration. C: Contrast-enhanced CT section in a patient with malig-
nant obstruction of the left upper lobe bronchus with associated left upper lobe atelectasis. An enlarged, low-density, necrotic subcarinal
lymph node is present (arrows), surrounded by a vascular capsule. We have seen this type of necrotic lymph node in association with lung
carcinomas of all histologic types. Transbronchial biopsy documented squamous cell carcinoma. D: Contrast-enhanced CT shows massively
enlarged and necrotic subcarinal lymph nodes in this patient with documented non-Hodgkin lymphoma.
in patients with a variety of other entities, including sar-

coidosis and Whipple disease (361).
Lymph Node Enhancement

Lymph nodes can be defined as enhancing and thus
vascular, when they substantially increase in attenuation
following a bolus of intravenous contrast medium. The dis-
tribution of intravenous contrast medium within tumors
and/or nodes is a reflection of several factors, including
blood flow, the total quantity and concentration of contrast
medium injected, the distribution between the vascular
A and extravascular spaces, and renal function. The differen-
tial diagnosis of enhancing mediastinal nodes is limited,
and includes Castleman disease (Table 4-22) (35,102,362)
and angioimmunoblastic lymphadenopathy (AIL), as well
as vascular metastases, in particular, from renal cell carci-
noma, papillary thyroid carcinoma, and small cell lung
carcinoma (Fig. 4-89) (95). We have also observed consid-
erable contrast enhancement within enlarged mediastinal
and hilar nodes in some patients with sarcoidosis. Also, as
indicated previously, nodes in patients with tuberculous
lymphadenopathy can show significant enhancement,
although inhomogeneous in appearance.
Differentiation between enhancing lymph nodes and
B enhancing mediastinal masses may be problematic. Enhan-
Figure 4-87 Low-attenuation mediastinal lymph node metas- cing masses, unrelated to lymph nodes, include substernal
tases in lung cancer. A, B: On a contrast-enhanced CT, right upper thyroid and parathyroid glands or masses, carcinoid tumors,
lobe consolidation reflects bronchial obstruction. Enlarged pretra-
cheal lymph nodes (arrows in A and B) are low in attenuation be- lymphangioma, hemangioma, and paraganglioma (100),
cause of necrosis. In patients with lung cancer, necrotic lymph both intracardiac and mediastinal (98,99,363).
nodes are often associated with a necrotic primary tumor.
Location of Enlarged Mediastinal

Lymph Nodes
The differential diagnosis of lymph node enlargement at
least partially depends on whether nodes are predominantly
middle mediastinal (including paratracheal, subcarinal, and
APW nodes) and/or hilar, prevascular, internal mammary,
paracardiac, or posterior mediastinal.
Approximately 30% of patients with lung cancer have
mediastinal node metastases at the time of diagnosis. Lung
cancer most often involves middle mediastinal node groups
(70,310,335,337–339). Left upper lobe cancers typically
TABLE 4-22
ENHANCING LYMPH NODES: DIFFERENTIAL
DIAGNOSIS
Common
Metastases
Castleman disease (multicentric or localized)
Figure 4-88 Low-attenuation and necrotic lymph node metas- Rare
tases from testicular seminoma. Lymph nodes metastases (arrows) Sarcoidosis
in the azygoesophageal recess, posterior to the left atrium, are Angioimmunoblastic lymphadenopathy
heterogeneously low in attenuation.
A B
Figure 4-89 Enhancing lymph nodes; CT assessment. A:

Contrast-enhanced section through the aortic arch in a patient with
obstruction of the left upper lobe bronchus caused by non–small cell
carcinoma. Note that there is considerable enhancement within
several enlarged prevascular lymph nodes (arrows). A large pleural
effusion is present on the left side within which a portion of the
collapsed left upper lobe can be identified, also caused by contrast
enhancement of the still-vascularized parenchyma (curved arrow). B:
Contrast-enhanced CT section through the aorta shows markedly
enlarged, enhancing pre- and paratracheal and prevascular lymph
nodes in a patient with documented sarcoidosis. C: Enhancing
paravertebral lymph nodes (arrows) in a patient with metastatic
paraganglioma. D: Multicentric Castleman disease with dense
C enhancement of axillary and mediastinal lymph nodes (arrows).
metastasize to APW nodes or anterior mediastinal (prevas- finding on chest radiograph in a patient with mediastinal
cular) lymph nodes, whereas tumors involving the lower lymphadenopathy should lead one to question the diag-
lobes on either side tend to metastasize to the subcarinal nosis. Isolated enlarged nodes, or mediastinal adenopathy
and right paratracheal chains (70,310). Right upper lobe in the absence of hilar adenopathy, is distinctly unusual
tumors typically involve paratracheal nodes. In patients (364–366).
with lung cancer, the presence of enlarged lymph nodes On CT, lymph node enlargement is visible in more than
only in an atypical location suggests that the node enlarge- 80% to 95% of patients with sarcoidosis, with the great
ment is unrelated to metastasis. majority showing both hilar and mediastinal lymph
Sixty percent to 90% of patients with sarcoidosis node enlargement (367). In patients with sarcoidosis
develop lymphadenopathy at some time in the course of and enlarged nodes, CT shows abnormalities, in order of
their disease. Chest radiographs show lymph node enlar- decreasing frequency, in the right paratracheal space, APW,
gement, in order of decreasing frequency, in the hila (85% hila, subcarinal space, prevascular space, and posterior me-
to 95%), right paratracheal region (75%), APW (50% to diastinum (367). Symmetry, or bilateral node involvement,
75%), subcarinal space (20%), and anterior mediastinum is visible more often on CT than on radiographs.
(10% to 15%). The presence of hilar lymph node enlarge- It is estimated in some series that more than 85%
ment is so typical of sarcoidosis that the absence of this of patients with HL eventually develop intrathoracic
abnormalities, typically involving the superior mediasti- signal intensity has been noted in calcified nodes on
nal (prevascular, pretracheal, and aortopulmonary) T2-weighted images in patients with fibrosing mediastini-
lymph nodes (368–370). In fact, it has been suggested tis and other causes of node calcification, but this appear-
that the absence of involvement of these mediastinal ance is nonspecific (377). The recognition of fibrosing
nodes in patients with intrathoracic adenopathy should mediastinitis using MRI is often dependent on identifying
prompt one to question the diagnosis of HL. secondary consequences such as vascular displacement
Similarly, the finding of enlarged paracardiac nodes has and occlusion (377,378).
diagnostic significance, as these generally prove malignant. Although there are some differences in the MRI charac-
As shown by Vock and Hodler (320), in their study of teristics of benign lymph nodes and lymph nodes involved
21 cases of cardiophrenic angle adenopathy, all proved to by tumor (379–381), their differentiation in individual
be caused by malignancy, including 12 patients with malig- cases has been difficult. However, recent studies (382,383)
nant lymphomas, 7 with metastatic carcinomas, and 2 with suggest that short inversion time inversion-recovery (STIR)
metastatic malignant mesothelioma. Similar findings have turbo spin-echo (TSE) MRI may be of value in the detec-
been reported by Sussman et al. (319) who found that of tion of lymph node metastases in lung cancer. In a study
45 patients with paracardiac adenopathy only 2 were found by Ohno et al. (383), 110 patients with non–small cell
to have benign lymphadenopathy; the remaining patients lung cancer were examined using respiratory-triggered
had a lymphoma (40%), carcinoma, or sarcoma. STIR TSE MRI. Ratios of lymph node signal intensity to
The ability to localize enlarged nodes is especially that of a saline phantom (LSR) were calculated for 92 of
important in the preoperative assessment of patients 802 lymph nodes that were pathologically proven to con-
with lung cancer. As indicated previously, node location tain metastases and 710 lymph nodes that did not contain
is predictive of the likelihood of malignancy. More metastases. Mean LSR in the lymph nodes containing
importantly, accurate localization is usually critical in tumor were higher than that in the group without metasta-
determining the most efficacious method for definitive sis (p .05), with the mean values depending on lymph
staging. Enlarged nodes in close association to the tra- node size. For lymph nodes measuring less than 6 mm in
chea suggest that bronchoscopy with transbronchial nee- short axis, mean LSR measured 0.62 for nodes containing
dle biopsy may be diagnostic. Nodes in the pretracheal tumor and 0.24 for normal nodes; for lymph nodes meas-
space, anterior subcarinal region, and anterior to the uring 6 to 10 mm, these values were 0.67 and 0.31, and for
right main bronchus are accessible to routine supraster- larger nodes the measurements were 0.71 and 0.41. When
nal mediastinoscopy, whereas nodes anterior to the left an LSR of 0.6 was used as a threshold for making the diag-
main bronchus, posterior subcarinal space, APW, and an- nosis of node involvement by tumor on a per-patient
terior mediastinum are not. Left anterior mediastinal basis, the sensitivity of STIR MRI was 93% and specificity
nodes and APW nodes are accessible to left parasternal was 87%. Qualitative analysis of STIR images had sensitiv-
mediastinotomy. ity and specificity values of 88% and 86%, respectively. In
In patients with lung cancer, CT may also be valuable the same study, CT was found to have a sensitivity of 53%
in assessing node morphology. CT may allow differentiation and a specificity of 83%.
between intra- and extranodal disease in those cases in As discussed in greater detail later, a role for MRI in
which tumor transgresses the nodes capsule. CT also may assessing residual or recurrent tumor in patients with
provide invaluable information concerning the effects of lymphoma has been extensively studied, although FDG-
enlarged mediastinal nodes on adjacent structures. For PET and PET/CT will likely assume a more important role
example, in patients presenting with signs of recurrent in making this diagnosis.
laryngeal nerve paralysis or airway compression, the In some specific regions of the mediastinum, MRI can
demonstration of appropriate anatomic abnormalities fre- be advantageous in demonstrating mediastinal nodes,
quently obviates more invasive diagnostic procedures (371). because of its ability to image in non-transaxial planes,
although this advantage is uncommonly of clinical utility
(323,384,385) and has become less important with the
advent of volumetric CT scanners capable of multiplanar or
of Mediastinal Lymph Nodes 3D reconstructions.
MRI is comparable to CT in identifying mediastinal and
hilar lymph nodes, even though its spatial resolution is
inferior to that of CT (37,372–376). This probably reflects LYMPHOMA
the greater contrast resolution of MRI, which makes identi-
fication of mediastinal lymph nodes especially easy when Lymphoma accounts for about 4% of newly diagnosed
they are embedded in mediastinal fat or adjacent to medi- malignancies in the United States (386). Lymphomas are
astinal or hilar blood vessels. primary neoplasms of the lymphoreticular system and are
MRI is unable to detect calcification, rendering identifi- classified in two main types: Hodgkin lymphoma and non-
cation of granulomatous nodes exceedingly difficult. Low Hodgkin lymphoma. Although HL is the less common of
TABLE 4-23
ANN ARBOR STAGING CLASSIFICATION FOR LYMPHOMA
Stage Definition
I Involvement of a single lymph node region (I) or a single extralymphatic organ

or site (IE)
II Involvement of two or more lymph node regions on the same side of the diaphragm
(II) or localized involvement of an extralymphatic organ or site and of one or more
lymph node regions of the same side of the diaphragm (IIE)
III Involvement of lymph node regions on both sides of the diaphragm (III), which may
also be accompanied by involvement of the spleen (IIIS) or by localized involvement
of an extralymphatic organ or site (IIIE) or both (IIISE)
IV Diffuse or disseminated involvement of one or more extralymphatic organs or tissues,
with or without associated lymph node involvement
The absence or presence of fever, night sweats, and/or unexplained loss of 10% or more of body
weight in 6 months is denoted by the suffix A or B, respectively
Modified from Castellino RA. Hodgkin disease: practical concepts for the diagnostic radiologist.
Radiology. 1986;157:305–310; and Castellino RA. The non-Hodgkin lymphomas: practical concepts for the
diagnostic radiologist. Radiology. 1991;178:315–321, with permission.
the two types, representing about 25% to 30% of cases, it is with prognosis (Table 4-23). Radiation is used for stages I
more common as a cause of mediastinal mass or lymph and II, with a combination of radiation and chemotherapy
node enlargement (386). or chemotherapy alone being employed in stages III and IV
(387). There is a 75% to 80% cure rate for adult HL, and in
children, the cure rate is about 95% (370).
Hodgkin Lymphoma
HL has a predilection for thoracic involvement, and up to
HL occurs at all ages, but its peak incidence is in the third 85% of patients with HL present with mediastinal adenopa-
and eighth decades; it accounts for about 0.5% to 1% of all thy (210,316,369). HL most often involves prevascular and
newly diagnosed malignancies in the United States paratracheal lymph nodes (210). In one study, these node
(369,387). It is more prevalent in males, with a male to groups were abnormal in 170 (84%) of 203 patients with
female ratio of from 1.38 to 1.94 (369). The Ann Arbor HL; this represented 98% of the 173 patients with intratho-
Staging Classification is used to describe the anatomic racic disease (368) (Table 4-24); if these nodes appear nor-
extent of disease at the time of diagnosis and correlates well mal on CT, intrathoracic adenopathy is unlikely to represent
TABLE 4-24
PLAIN RADIOGRAPH AND COMPUTED TOMOGRAPHY ABNORMALITIES
IN PATIENTS WITH HODGKIN DISEASE
Radiograph Change in Treatment
Abnormal (%) CT Abnormal (%) Based on CT Findings (%)
Superior mediastinal nodes 77 84 0.5

Hilar nodes 21 38 1
Subcarinal nodes 7 22 3.5
Internal mammary nodes 1 5 0
Posterior mediastinal nodes 2 5 1
Paracardiac nodes 1 8 1.5
Lung parenchyma 8 8 0
Pleura 10 13 0
Pericardium 2 6 1
Chest wall 1 6 2.5
All sites 78 87 9.4
Modified from Castellino RA, Blank N, Hoppe RT, Cho C. Hodgkin disease: contributions of chest CT in the initial
staging evaluation. Radiology. 1986;160:603–605, with permission.
HL (369). Multiple node groups are commonly involved in In patients with HL, enlarged lymph nodes can be
patients with HL (Figs. 4-90 and 4-91). Other sites of lymph variable in appearance. Nodes are of homogeneous soft tis-
node enlargement include hilar nodes (28% to 44% of all sue attenuation in the majority of cases (210). Multiple
cases), subcarinal nodes (22% to 44%), cardiophrenic angle enlarged lymph nodes are often seen, and they can be well
(paracardiac) lymph nodes (8% to 10%) (312,313), internal defined and discrete, matted, or associated with diffuse
mammary nodes (5% to 37%), and posterior mediastinal mediastinal infiltration. It is not uncommon for lymph
nodes (5% to 12%) (368,388). Enlargement of a single node node masses to show areas of low attenuation or necrosis
group can be seen in some patients with HL, most com- following contrast enhancement; cystic- and necrotic-
monly in the prevascular mediastinum. This often indicates appearing nodes have been reported in 10% to 21% of
the presence of nodular sclerosing histology that accounts cases (359,360). Rarely, nodes show fine flecks of calcifica-
for 50% to 80% of adult HL (369). Direct extension of lym- tion in untreated patients (348).
phoma from the mediastinum to the lung or the chest wall is The significance of low-density nodes in patients present-
common with large mediastinal masses (389). ing with newly diagnosed lymphoma has been studied by
CT is advantageous in showing mediastinal lymph node Hopper et al. (360). A review of 76 patients with newly dia-
abnormalities in patients with lymphoma. For example, gnosed HL found that 16 patients (21%) had low-density
paracardiac lymph node enlargement was seen on CT in 14 nodes at presentation. However, no difference was found
(6.6%) of 274 patients with HL or NHL, but in only 3 of between those patients with and without necrotic nodes
these was the abnormality visible on plain radiographs with respect to stage, disease distribution, cell type, disease
(317). In a study by Castellino et al. (368), of 203 patients, extent, the presence of bulk disease, or, most importantly,
findings shown only on CT changed treatment in more than prognosis (360).
9% of patients (Table 4-24). In another study, however, it HL also has a predilection for involvement of the thymus
was uncommon for CT to show evidence of mediastinal in association with mediastinal lymph node enlargement
adenopathy if the chest radiograph was normal, but if the (209,210). In a study of 50 patients with newly diagnosed
radiograph was abnormal, CT detected additional sites of disease and evidence of thoracic involvement, the thymus
adenopathy in many cases (388). In this study, CT was most was enlarged in 15 (30%); in all, mediastinal lymph nodes
helpful in diagnosing subcarinal, internal mammary, and were also enlarged. In other studies (210,211), thymic
APW node enlargement not visible on radiographs. In a sig- enlargement visible on CT has been reported in 17 (40%)
nificant percentage of patients, the additional node involve- of 43 adult patients and 17 (28%) of 60 pediatric patients.
ment shown by CT changed therapy (370,388). In a review of 50 episodes of recurrence of HL, 18 patients
A B
Figure 4-90 Mediastinal lymph node enlargement in Hodgkin dis-

ease. A: Multiple discrete lymph nodes and masses are visible in the
anterior mediastinum (large white arrow), aortopulmonary window
(small white arrow), and pretracheal space (black arrow). B:
Extensive lymph node enlargement is visible in all parts of the me-
diastinum, with the largest nodes being anterior. C: Paracardiac
lymph node enlargement (arrows) is visible near the level of the
C diaphragm.
A B
Figure 4-91 Involvement of multiple lymph node groups in

Hodgkin disease. A, B: Involvement of pretracheal (small arrows)
and prevascular (large arrows) is common in Hodgkin disease. C:
At a lower level, paracardiac lymph node enlargement is also seen
C (arrows).
had an intrathoracic relapse, and thymic enlargement HL. Fewer than 30% of HL patients with hepatic involve-
was present in 7 (38%) of the 18; in all but one patient who ment at autopsy had detectable hepatomegaly at clinical
had a thymic cyst, mediastinal lymph nodes were also examination. Thus, CT or MRI demonstration of liver
involved (209). enlargement is of no diagnostic value. Likewise, spleen size
HL is believed by most to be unifocal in origin and typi- evaluation is of limited value; solitary or multiple splenic
cally spreads by contiguity to involve adjacent lymph node lesions are the most reliable evidence of splenic involve-
groups (369,390). It is unusual for HL to skip lymph node ment with either CT or MRI. In one series, almost two
groups, and, if areas contiguous with the mediastinum, thirds of HL nodules involving the spleen were less than
such as the lower neck or upper abdomen, are not involved 1 cm in diameter, rendering their detection difficult
by HL, it is not generally necessary to scan regions, such as on both CT and MRI, although marked splenomegaly is
the pelvis, which are more distant from the involved nodes. almost always related to tumor involvement (369,392).
However, in patients with mediastinal HL, scanning should
always be extended to include the upper abdomen. Intra-
Non-Hodgkin Lymphoma
abdominal periaortic adenopathy can be found in 25% of
patients with HL, and the spleen and liver are involved in The term non-Hodgkin lymphoma refers to a diverse
37% and 8%, respectively (391). group of diseases, varying in radiologic findings, clinical
It is important to note that there is no strong correla- presentation, course, and prognosis (386,393–397). The
tion between liver or spleen size and their involvement by incidence of NHL is rising, particularly in the industrial-
ized world. Several risk factors have been postulated to be CT can show evidence of intrathoracic disease when it is
associated with NHL, including exposure to chemicals, unrecognizable on plain radiographs and may be helpful in
viral infections, organ transplantation, blood transfusion, diagnosis (393,394). In a recent study, both CT and chest ra-
family history, and lifestyle (398). In comparison to HL, diographic findings were positive in 65 (36%) of 181 patients
NHL tends to occur in an older group (median 55 years) with NHL, whereas CT findings were positive and chest radi-
(386) and is more common in children (4). ographs were negative in 17 (9%) (393). In the 65 patients
The classification of NHL has been controversial and is with abnormal CT and chest radiographs, CT showed more
the subject of many publications. The National Cancer extensive intrathoracic disease in 49 (75%), and this resulted
Institute Working Formulation initially classified NHL into in an increase in stage in 16. In another study (394), chest CT
low, intermediate, and high grades on the basis of histology, showed intrathoracic disease in 28% of patients with NHL
including a number of tumor subtypes (397). In 1994, the thought to have normal chest films, and 45% of patients with
International Lymphoma Study Group developed a con- equivocal plain radiographic findings.
sensus list of lymphoid neoplasms, which was published An understanding of the use of CT in patients with NHL
in 1994 as the Revised European-American Classification of requires some knowledge of how it is treated. The WHO
Lymphoid Neoplasms (i.e., the R.E.A.L. classification of Clinical Advisory Committee has concluded that treatment
lymphoma). The classification was based on the principle is best determined by the specific type of lymphoma
that there is a “real” list of NHL types, which should be present (using the WHO classification), with the addition
reproducibly defined by a combination of morphology, of grade within the specific tumor type (397,399,400).
immunophenotype, genetic features, and clinical features. NHL is usually treated using chemotherapy in patients
Starting in 1995, the European Association of Pathologists with intermediate- or high-grade lymphomas regardless of
and the Society for Hematopathology have participated in their anatomic stage and in the majority of patients with
developing a WHO classification of lymphoid malignan- low-grade lymphoma (393,395). The use of primary radio-
cies, using an updated R.E.A.L. Classification. New entities therapy is usually limited to a small subgroup of patients
not specifically recognized in the Working Formulation with early stage (I or II) low-grade lymphoma; only 20%
accounted for more than one fourth of the cases classified to 25% of patients with low-grade lymphoma are stage I or
using this system, and a distinction of cases as low grade or II at diagnosis (395). Although NHL is staged using the
intermediate/high grade, as in the Working Formulation, same system as HL, staging of NHL is much less important.
was shown to be of limited usefulness (397,399,400). With HL, the anatomic extent of the tumor strongly
Thoracic involvement is about half as common with predicts outcome; with NHL, the histopathologic classifi-
NHL (40% to 50%) as with HL (85%) (316,386,393). In cation is more predictive (386).
a study of 181 newly diagnosed patients with NHL, only The role of chest CT in patients with NHL is continuing
82 (45%) had intrathoracic disease (393). As with HL, the to evolve in conjunction with developing techniques such
most common thoracic abnormality in patients with NHL as PET (386,393,398). Although CT is commonly obtained
is mediastinal lymph node involvement, although the pat- in clinical practice, its use has been called into question
tern of lymph node abnormalities is somewhat different. by Castellino et al. (393); although CT often shows more
Involvement of one node group is much more common in extensive thoracic disease than do plain radiographs, this
patients with NHL; in one study, 40% of patients with information does not necessarily alter treatment. None-
NHL and thoracic involvement had involvement of only theless, some reasonable recommendations for CT can
one node group, whereas this was seen in only 15% of probably be made, based on a consideration of tumor
patients with HL (316). Enlargement of anterior mediasti- type, clinical stage, and plain radiographic findings (394).
nal, internal mammary, paratracheal, and hilar nodes is In patients with intermediate- or high-grade NHL or stage
much less common with NHL than with HL. Nonetheless, III or IV disease, chest CT does not seem to have a distinct
superior mediastinal node involvement remains the most role to play, as patients are treated using chemotherapy
frequent abnormality seen. In one study (393), 34% of (393,394). In fact, it should be pointed out that the lack of
patients had superior mediastinal (prevascular and pretra- utility for staging NHL found for CT by Castellino et al.
cheal) node involvement, and this represented 74% of the (393) is at least partially explained by the preponderance
patients with an intrathoracic abnormality. Subcarinal of intermediate- and high-grade NHL in their study. On
lymph node enlargement was present in 13% of cases, the other hand, in patients thought to have stage I or II
hilar nodes in 9%, and cardiophrenic nodes in 7%. low-grade NHL, CT seems appropriate for determining
Involvement of posterior mediastinal nodal groups is whether intrathoracic disease is present and, if localized, in
much more common with NHL; this was seen in 10% of helping to plan radiation therapy (394).
cases (316,393,401). Rarely, calcification of node masses In contrast to patients with HL, the abdomen, pelvis,
can be seen (346). Extranodal involvement is much more and neck must be scanned in all patients with NHL, as
common with NHL (386). Extranodal sites of involvement noncontiguous spread is common; NHL is assumed to be
include lung (13%), pleura (20%), pericardial space (8%), multifocal in origin (395). Abdominal involvement is
and chest wall (5%) (393). more common than in patients with HL, and a variety of
findings may be present (402); intra-abdominal periaortic NHL (404,405). It is thought to arise from thymic
adenopathy is found in 49% of patients with NHL (391), medullary B cells. Both clinically and radiographically, this
the spleen is involved in 41%, and the liver in 14%. It is entity resembles HL (Table 4-25). Affected patients are
important to note that, as with HL, there is no strong usually younger than other patients with NHL, with a mean
correlation between liver size and liver involvement; only age of 35 to 45 years. Typically Med-DLBCL is confined to
57% of patients with NHL and hepatomegaly demonstrate the mediastinum at presentation. A large, smooth or lobu-
liver lymphoma at histologic examination. In one study lated, anterior mediastinal mass is the predominant finding
(403), CT of the abdomen and pelvis resulted in an in nearly all patients, Masses average about 10 cm in diame-
upgrading of the clinical stage in 33% of patients with ter (Fig. 4-92). In a study of 43 patients with Med-DLBCL
NHL and detected unsuspected active disease in 43% of (405), an anterior mediastinal mass was the predominant
patients thought to be in remission. This resulted in clini- finding in 42 (98%) of 43 patients studied using chest
cally important upstaging in 15% of all cases. radiographs and 24 (96%) of 25 patients having CT. Low-
The incidence of NHL, particularly of the intermediate attenuation areas within the mass may represent necrosis,
and high grades, is significantly higher in immunodeficient hemorrhage, or cystic degeneration and were seen on CT in
patients, being associated with congenital immunodefi- 50% of cases. Calcification is uncommon but may occur
ciency syndromes, HIV infection, and immunosuppressive (405). Heterogeneous enhancement is seen in about 40%
therapy (386,395). These tumors differ somewhat from of cases after contrast infusion (404). Lymph node enlarge-
those occurring spontaneously in immunocompetent pati- ment may also be seen in the subcarinal space and posterior
ents, usually being polyclonal rather than monoclonal and mediastinum but is less common. In one study, 14 (67%)
usually involving extranodal sites (e.g., central nervous of 21 patients showed mediastinal lymph node enlargement
system, lung, gastrointestinal tract). separate from the primary mass, but extrathoracic lymph
node enlargement is unusual (404). Pleural and pericardial
effusions were present in about one third. Superior vena
Primary Mediastinal Non-Hodgkin Lymphoma
cava obstruction and chest wall invasion were uncommon
NHL may occur primarily in the mediastinum in a (405). CT is valuable in determining tumor extent, as radia-
small proportion of cases. The most common cell types tion is usually used. Splenomegaly is uncommon, being
presenting in this fashion are diffuse large B-cell lym- seen in only 1 of 21 patients (404). Med-DLBCL may spread
phoma and precursor T-cell lymphoblastic lymphoma to involve organs such as kidney, liver, ovary, adrenal gland,
(Table 4-25) (404). gastrointestinal tract, and central nervous system.
CT has become an integral part of the evaluation of
patients with diffuse mediastinal large B-cell lymphoma
Primary Mediastinal Diffuse Large B-Cell
at presentation and after completion of therapy (406).
Lymphoma A large residual tumor volume after completion of treatment
Primary mediastinal diffuse large B-cell lymphoma (Med- predicts an increased risk of relapse. In one study (406), a
DLBCL) has recently been described as a distinct subtype of tumor volume of greater than 100 mL after completion of
TABLE 4-25
COMPUTED TOMOGRAPHY FINDINGS IN PATIENTS WITH PRIMARY
MEDIASTINAL LYMPHOMA
Hodgkin L Med-DLBCL T-LBL
Well-defined margins 66% 48% 62%

Mean tumor diameter 9.7 2.8 cm 9.7 2.5 cm 10.2 3.3 cm
Lobulation 69% 33% 25%
Vascular encasement 7% 62% 38%
Cervical lymphadenopathy 31% 0% 56%
Mediastinal lymphadenopathy 97% 67% 50%
Axillary lymphadenopathy 14% 0% 50%
Splenomegaly 3% 5% 60%
Pleural effusion 21% 57% 50%
Hodgkin L, Hodgkin lymphoma; Med-DLBCL, mediastinal diffuse large B-cell lymphoma; T-LBL, precursor
T-cell lymphoblastic lymphoma.
Modified from Tateishi U, Muller NL, Johkoh T, et al. Primary mediastinal lymphoma: characteristic features
of the various histological subtypes on CT. J Comput Assist Tomogr. 2004;28:782–789, with permission.
A study by Tateishi et al. (404) assessed 66 patients with

primary mediastinal lymphoma, including 29 patients with
HL, 21 with Med-DLBCL, and 16 with T-LBL. Although each
of these tumors typically resulted in a large, heterogeneous,
mediastinal mass, they showed significant differences in
their CT appearances (Table 4-25). HL characteristically
showed an anterior mediastinal mass with a lobulated
contour (20 of 29 cases, 69%) and high prevalence of asso-
ciated mediastinal lymph node enlargement (28 of 29,
97%). Characteristic features of Med-DLBCL included a
mass having a smooth contour (14 of 21, 67%), associated
with vascular encasement (13 of 21, 62%) but without
extrathoracic (i.e., cervical and abdominal) lymphadenopa-
thy (0 of 21 cases). T-LBL resulted in a mass with a smooth
Figure 4-92 Primary mediastinal diffuse large B-cell lymphoma
in a 36-year-old man. CT shows a large inhomogeneous anterior contour (12 of 16, 75%) associated with a high prevalence
mediastinal mass, with regions of low attenuation likely due to of cervical (9 of 16, 56%) and abdominal (6 of 16, 38%)
necrosis. This appearance mimics Hodgkin disease. lymphadenopathy and splenomegaly (11 of 16, 69%)
(404). CT findings independently associated with an
therapy, measured using CT, was a statistically significant increased likelihood of HL were lobulation (p .01),
predictor of increased risk of relapse (p .02). the presence of pleural effusion (p .05), and the ab-
sence of vascular encasement (p .01) (404). The presence
of vascular encasement was associated with incre-
Precursor T-Cell Lymphoblastic Lymphoma ased likelihood of Med-DLBCL (p .001). Other CT
Precursor T-cell lymphoblastic lymphoma (T-LBL) is also a findings such as the presence of cervical lymph nodes or -
common cause of primary mediastinal NHL. Most patients inguinal lymph nodes (p .001), the presence of
are children or young adults. In one study, the mean age at pericardial effusion (p .05), and the absence of lobulation
presentation was 31 years (404). A large mediastinal mass (p .05) were significantly associated with the likelihood
representing thymic or lymph node enlargement is typically of T-LBL.
visible on CT (Fig. 4-93), and heterogeneous enhancement
is common (Table 4-25) (404). Extrathoracic lymph- Magnetic Resonance Imaging Diagnosis
adenopathy with involvement of superficial cervical (44%
of Lymphoma
of cases), deep cervical (56%), supraclavicular (50%), axil-
lary (50%), submandibular, submental, parotid, mesen- Histologically, lymphomas contain variable proportions of
teric, and inguinal groups is seen in the majority of patients cells and connective tissue. For example, nodular sclerosing
with T-LBL. Splenomegaly was visible in 11 (61%) of 16 pa- HL is composed of malignant cells interspersed with large
tients in one study (404). If bone marrow and hematologic amounts of fibrous tissue (Fig. 4-94). Conversely, in diffuse
involvement are predominant features of this disease, it is NHL, tumors may consist almost entirely of malignant cells
termed lymphoblastic leukemia. Presenting symptoms are without significant stroma. Based on these histologic differ-
usually related to compression of mediastinal structures. ences, lymphomas can have varying MRI appearances,
showing as homogeneous or heterogeneous on T2-weighted
images or following contrast enhancement.
In patients with homogeneous-appearing lymph nodes
or masses, lesions demonstrate high signal intensity,
greater than that of muscle and similar to that of fat, on
T2-weighted images (407,408). This appearance can be
seen with HL or NHL. Typically, patients with a
heterogeneous pattern show mixed signal intensity on
T2-weighted images, with interspersed areas of high and
low signal intensity. This pattern is often seen in untreated
nodular sclerosing HL, with the low signal intensity areas
on T2-weighted images related to regions of fibrosis
in the tumor and high-intensity regions representing
tumor tissue or cystic regions. Furthermore, it has been
Figure 4-93 Precursor T-cell lymphoblastic lymphoma in a suggested that MRI signal characteristics may be of value in
14-year-old. CT shows an anterior mediastinal mass (arrows) prob-
ably representing thymic enlargement. This appearance in a young differentiating low-grade from high-grade lymphomas
patient is typical. (409). The likelihood of recurrence has been found to be
A B
Figure 4-94 Histologic changes before and after therapy of nodu-

lar sclerosing Hodgkin lymphoma. A: Islands of highly cellular tissue
are interspersed among more fibrotic, relatively hypocellular areas.
Low-power photomicrograph of nodular sclerosing Hodgkin lym-
phoma demonstrates typical biphasic histology with areas of predom-
inant sclerosis with paucity of malignant cells and areas of malignant
cell infiltrations. B: Higher magnification of A shows predominance of
malignant cells with interspersed connective tissue components. C:
After therapy, repeat surgical biopsy of residual mass demonstrates
cell necrosis with marked hypocellularity and extensive sclerosis, pre-
C sumably representing the residual fibrous component of the tumor.
higher if the signal intensity of the initial mass is high on

T2-weighted MR images (410). Following injection of
gadolinium, masses in patients with lymphoma enhance
more than muscle, with enhancement values ranging from
41% to 124% (407).
Diagnosis of Residual Tumor After Treatment

Imaging studies are frequently obtained to judge the com-
pleteness of tumor response to treatment, and periodic
surveillance imaging is commonly used to determine the
presence or absence of relapse (369,387). Many patients
with recurrence will eventually be cured.
Computed Tomography Diagnosis of Residual

Tumor and Recurrence
CT has been used most frequently to image patients
after treatment. It is usually obtained 2 to 4 months after
the completion of therapy to assess tumor response and Figure 4-95 Residual mediastinal mass in treated Hodgkin
to provide a baseline for subsequent studies (369,387). lymphoma. Enlarged lymph nodes (arrows) persist in the anterior
mediastinum years after treatment of Hodgkin lymphoma.
FDG-PET or PET/CT imaging is also used when available, Calcification is visible within the lymph nodes. These were stable
and MRI may be used in some cases. over years.
Reduction of tumor bulk is universally present in
patients with adequately treated tumor, and this finding is
commonly used when assessing the efficacy of treatment. However, the thymus remained enlarged following treat-
Reduction in tumor size after treatment is satisfactorily ment in 9 (38%) of 24 adult patients in one study
monitored using CT. Patients showing complete resolution (211) and 3 (27%) of 11 pediatric patients in another
of the masses or enlarged nodes on CT are generally con- (210). Although it is acknowledged that in patients with
sidered to have had a satisfactory response. HL, these masses are usually benign and of no conseque-
However, monitoring the decrease in tumor size is not nce, close monitoring of residual masses is recommended.
always an accurate measure of treatment success, as the In some patients, masses remaining after treatment
rate of decrease in tumor size may be different in different continue to decrease in size or resolve over a period rang-
patients and varies with the size of the initial mass, its site, ing from 3 to 11 months (411,418).
histology, and the type of treatment (410,411). Tumors Recurrent HL does not commonly involve previously
containing a significant fibrous component, such as nodu- irradiated intrathoracic lymph nodes. However, so-called
lar sclerosing HL, will generally show less decrease in size in-field recurrence does occur. In one study (411), 1 of
than tumors composed primarily of cellular elements, and 20 intrathoracic relapses was in the treatment field, and in
residual mediastinal masses are common after treatment another (419), 13 of 60 patients with stage I or II mediasti-
in patients with HL, even when a cure has been achieved. nal disease had an intrathoracic relapse, and six of these
Residual mediastinal masses following treatment often were in previously treated mediastinal nodes. All of the six
reflect so-called sterilized lymphoma—residual fibrous patients with a mediastinal recurrence had large masses
tissue components of the tumor itself or post-treatment (greater than one third of the chest diameter) before
fibrosis (412). Pathologic studies of residual mediastinal treatment. Large masses and masses in the anterior medi-
masses in patients with treated HL have shown them to be astinum are generally considered to carry an increased risk
nearly acellular, hyalinized fibrous tissue (Fig. 4-94) (411, of recurrence (369,411,415,419–421).
413,414). In a study by North et al. (411), intrathoracic relapse
Based on the interpretation of chest radiographs, resid- was more than twice as frequent in HL patients who had
ual masses were thought to be present in a minority of residual masses after treatment. Other studies, however,
patients after treatment of lymphoma. However, studies have not confirmed this finding, perhaps because of differ-
using CT indicate that residual masses can be seen in as ing patient populations and different findings used to
many as 88% of patients with HL (Figs. 4-95 and 4-96) diagnose mediastinal mass (415,418). On the other hand,
and 40% of patients with NHL (415–417). In patients both residual mediastinal masses and recurrence are more
with thymic enlargement resulting from HL, most show common in HL patients who have a large mass initially. In
a return to normal thymic size following treatment. a series of NHL patients restaged surgically following
A B
Figure 4-96 Residual masses after treatment for Hodgkin

lymphoma: CT appearance. A: Pretreatment CT shows ex-
tensive mediastinal lymph node enlargement typical of
Hodgkin disease. B: Post-treatment radiograph shows per-
sistent widening of the superior mediastinum. C: Post-treat-
ment CT, obtained at the same time of the radiograph in B,
shows residual lymph node enlargement. Such residual
masses are very common in patients with Hodgkin disease
and, if stable, do not imply recurrence.
initial therapy of NHL, residual masses were present in Magnetic Resonance Imaging Diagnosis
20%, and persistent disease and relapses occurred mainly of Residual Tumor and Recurrence
in areas of previous involvement (417). In patients with
recurrent tumor, CT can show an increase in size of one or MRI can provide important information in distinguishing
more masses. CT attenuation is identical for both active active tumor from residual benign masses (408,410,423,
and inactive residual masses in patients treated for lym- 424). The basis for the MRI signal intensity differences
phoma and is of limited value in making this diagnosis. exhibited by different tissues is fundamentally different from
Despite the common use of CT for monitoring tumor that resulting in density differences in CT (379,425). In sim-
response, its cost effectiveness has been questioned (422). plified terms, relaxation times are primarily dependent on
In a retrospective review of 107 treated patients with the proportion of water and proteins as well as the type of
HL (1,209 total follow-up visits including a total of 283 proteins within a given tissue. Lymphoma cells contain a
CT examinations), there were 109 suspected relapses of large amount of water, with a relatively low proportion of
which 22 proved to be true relapses. Fourteen of the 22 proteins and thus appear intense on T2-weighted images.
true relapses were diagnosed clinically, 6 radiologically, Conversely, fibrous tissue contains much less water and a
and 2 using lab tests. The routine CT scan detected only high proportion of highly polymerized proteins and shows
2 of the 22 relapses (9%) but accounted for 29% of low intensity with T2 weighting (425).
the total follow-up costs. The authors conclude that Because residual mediastinal masses in sterilized lym-
routine CT is an expensive and inefficient mode of routine phoma are composed primarily of fibrous tissue, it is
follow-up (422). reasonable to expect that some changes in the MRI signal in-
A B
Figure 4-97 Mediastinal lymphoma, MR assessment pretreatment. A: T1-weighted image through the mediastinum shows a mass of
homogeneously low signal intensity. B: T2-weighted image at the same level as A shows homogeneously high signal intensity.
tensity pattern may be seen during the post-treatment evolu- geneous low signal intensity on T1-weighted images
tion of lymphoma. This hypothesis has been the basis of re- and heterogeneous signal intensity on T2-weighted
search into the potential utilization of MRI to monitor lym- images, with areas of both high and low signal inten-
phoma patients showing residual masses after treatment sity. This pattern can be seen in untreated nodular scle-
(426). rosing HL, with the low signal intensity areas on
By systematically studying a series of patients with T2-weighted images presumably related to regions of
treated lymphoma, four characteristic MRI signal patterns fibrosis in the tumor (Fig. 4-98). Also, it is commonly
have been defined on the basis of T1-weighted and seen during the post-treatment response phase of most
T2-weighted images (408,427): lymphomas, with the high signal intensity regions
presumably representing residual active lymphoma-
1. Homogeneous, hyperintense (active) pattern. This is char- tous masses, areas of necrosis, or inflammation. The
acteristic of untreated lymphoma (Fig. 4-97). The lesion low signal areas likely represent fibrotic tissue. This
demonstrates homogeneous low signal intensity similar pattern is seen in about 20% of patients with residual
to that of muscle on T1-weighted images, and a high sig- masses after treatment.
nal intensity similar to that of fat on T2-weighted images. 3. Heterogeneous inactive pattern with mixed areas of low
This pattern is never seen in inactive residual lymphoma- and high signal intensity on both T1- and T2-weighted
tous masses. images. Areas of high intensity on the T1-weighted im-
2. Heterogeneous active pattern with mixed hypo- and ages should correspond to those seen with T2 weighting.
hyperintensity. This pattern is characterized by homo- This pattern is seen in lesions containing mixed fat (high-
A B
Figure 4-98 A 48-year-old patient with newly diagnosed Hodgkin disease. A: Proton density image shows mass in superior mediastinum.
B: T2-weighted image demonstrates a heterogeneous pattern with mixed areas of high and low signal intensity. This is typical of nodular
sclerosing Hodgkin disease, with its biphasic histology comprising fibrotic areas interspersed with more cellular areas.
intensity regions) and fibrous tissue (low-intensity rect diagnoses of tumor recurrence were made. Four were in
regions), and is seen after treatment in patients with ster- the first 4 months after treatment, caused by the presence
ilized tumors. of necrosis or inflammation in the tumor or increased
4. Homogeneous, hypointense (inactive) pattern. This intensity in adjacent lung, two were caused by misinterpre-
appearance is characteristic of inactive residual fibrotic tation of a heterogeneous inactive pattern as abnormal, and
masses in patients having successful therapy for lym- an additional patient showed a heterogeneous active
phoma. These lesions are characterized by homoge- pattern 6 months after treatment caused by the presence of
neously low signal intensity on both T1-weighted and chronic inflammation.
T2-weighted images. Approximately 80% of masses The use of gadolinium-enhanced MRI may also be of
show a homogeneous hypointense pattern within 6 to value. Gadolinium enhancement of mediastinal masses in
8 weeks of initiation of therapy. lymphoma decreases markedly after treatment in patients in
complete remission but not in patients with relapse. In a
It is important to emphasize that both T1-weighted and study by Rahmouni et al. (407), 31 patients with bulky
T2-weighted images are necessary in evaluating patients mediastinal lymphoma (17 with HL, 14 with NHL) under-
with lymphoma. Active residual areas of lymphoma, went serial MRI before and up to 50 months after treatment,
necrosis, or inflammation can only be diagnosed if high with routine follow-up (including CT). Signal intensity
signal intensity is seen on T2-weighted images, but low sig- ratios between masses and muscle were calculated on
nal intensity is seen on T1-weighted images in the same T1-weighted, T2-weighted, and contrast material–enhanced
region. Another solution to this problem would be the T1-weighted SE MR images. Twenty-one patients with com-
use pulse sequences with selective fat suppression. plete remission had a mean percentage enhancement of
In patients with lymphoma, complete regression of the residual masses that was significantly lower (mean 4%)
tumor mass will often be observed following treatment; in than that of initial masses (mean 78%). In seven patients
such patients there is no need for MRI (423). However, in with relapse, the percentage enhancement value of the resid-
those patients having a residual mass after one or two ual mass was as high as that of the initial mass (407).
cycles of chemotherapy, MRI may be used to assess the In our experience, monitoring the MRI signal intensity
potential activity of the remaining lesion. A divergence of of residual masses can provide a qualitative measure of
lesion size change and MRI signal intensity is common in tumor response and can benefit patient management in
the first 6 months after treatment. This was seen in 44 several ways.
(47%) of 93 cases in one study (408). Generally in this First, in patients showing a decrease in mass size and
situation, signal intensity changes were more reliable than a homogeneous decrease in T2-weighted signal intensity, a
size in assessing tumor status. favorable response can be expected. Of course, it must be
In particular, the T2-weighted intensity of a residual mass recognized that microscopic residual tumor cannot be
shown on MRI has a significant relationship to the likeli- detected by MRI, given the low spatial resolution and the
hood of residual or recurrent tumor. Typically, the intensity variability of signal intensities intrinsic to the method. In
of HL on T2-weighted images significantly decreases follow- all such cases, continued surveillance is warranted until
ing treatment (410). In a study by Nyman et al. (410), there is complete resolution of the residual mass. Residual
a prompt and persistent decrease in intensity occurred after masses may resolve in 12 to 18 months or persist without
treatment in all but three cases. In two of these three, the change in size. Reappearance of foci of high signal inten-
persisting high intensity may have reflected necrosis of the sity in such masses is a strong indicator of early recurrence
mass, cysts within the treated tumor, or volume averaging and foci are usually noted before clinical symptoms.
with fat, but in one, an increase in intensity, accompanied Second, decrease in size of the mass with a persistent
by an increase in size of the mass, indicated recurrence. homogeneous hyperintense or heterogeneous hyperin-
Also, Zerhouni et al. (426) found that inactive (treated) tense pattern suggests partial response. However, inflam-
lymphomas were significantly lower in intensity on mation and necrosis may result in a similar appearance,
T2-weighted images (mass/fat intensity ratio 0.3) than mimicking active tumor, in the early post-therapy phase
active lymphomas (mass/fat intensity ratio 0.9). The (424). In a series of 115 patients we have studied over
average mass/fat intensity ratio on the T2-weighted 3 years, all instances of a false-positive MRI diagnosis of
sequence in Nyman’s study was approximately 0.8 prior to residual mass caused by necrosis or inflammation of the
treatment, and slightly less than 0.4 after treatment. Thus, tumoral mass occurred within 4 months of the initiation
a high relative intensity with T2 weighting, although not of therapy. One should be cautious in not overinterpreting
specific for persistent or recurrent tumor, might indicate these patterns as definite evidence of lack of response
the need for biopsy or close follow-up of a residual medi- within the first few months following therapy.
astinal mass. Change in intensity from low to high during In some patients a marked size reduction occurs, but a
follow-up should be regarded as very significant. small residual mass or masses show heterogeneous or
In a study of 93 intervals between MRI examinations in homogeneous high signal intensity (Fig. 4-99). In such
34 patients with treated lymphoma (408), using the four patients, although the size decrease would suggest tumor
signal intensity patterns described previously, seven incor- response (428–430), the high signal intensity shown on
A B
C D
E F
Figure 4-99 A 27-year-old patient with Hodgkin lymphoma in the anterior mediastinum. A, B: Proton density and T2-weighted images,
respectively, obtained in a patient prior to the initiation of therapy demonstrate a large mass in the anterior mediastinum with large right
pleural effusion. The mass appears mostly hyperintense with a signal intensity almost equal to that of surrounding fat. Slight heterogeneity
is noted within regions of the mass. C, D: T1- and T2-weighted images, respectively, 2 months after therapy initiation. A marked decrease in
the bulk of the tumor is noted. However, a small amount of residual mass is seen immediately anterior to the arch of the aorta. Note the
heterogeneous, predominantly high signal intensity seen on the T2-weighted image and the correspondingly low signal intensity on the
T1-weighted scan. E, F: Scans obtained 5 months after start of therapy. Again there is a persistent residual mass anterior to the aorta,
essentially unchanged in size from the previous examination. Note the heterogeneous high signal intensity pattern on the T2-weighted
image. Even though the initial mass has regressed markedly, the lack of change in size between months 2 and 5 and the persistently “active”
MR signal pattern strongly suggest residual active disease. The patient was asymptomatic at this point.
G H
I J
Figure 4-99 (continued ) G, H: On T1-and T2-weighted images obtained 9 months after therapy initiation, there has been a definite in-
crease in the mass size compared to the previous examination. Note again the low homogeneous signal intensity on the T1-weighted image
compared to the heterogeneous, mostly high signal intensity seen on the T2-weighted image. Additionally, an enlarged right paratracheal
lymph node can now be identified. On the basis of these findings, therapy was reinitiated. I, J: Twelve months following the initial diagnosis,
the mass has decreased somewhat in size since the previous examination. However, a pattern of high signal intensity on T2-weighted im-
ages persists, suggesting a lack of response. Note also the appearance of pulmonary lesions. The patient died 2 months following the last
study. This case represents a typical example of the dissociation between the evolution of size and that of the MR signal intensity pattern,
thus demonstrating the unique information MR can provide in this category of patients.
MRI justifies close monitoring for recurrence. In a series of tion should be the rule within the first 4 to 6 months
34 patients prospectively followed by serial MRI studies following the initiation of therapy. T1- and T2-weighted
over 30 months, this combination of findings occurred in images should always be interpreted in combination, or fat
23% of cases (415). suppression techniques should be used, to exclude the
MRI signal intensity increases during follow-up of resid- possibility of fibrofatty masses mimicking the appearance of
ual masses previously considered inactive. In this group recurrent tumor (408). Motion artifacts in the phase direc-
of patients, “islands” of high signal intensity can be seen tion of the image or flow artifacts may, on occasion, render
to reappear in low intensity residual masses. This appear- interpretation difficult. Good technique with cardiac gating
ance indicates recurrence. In our experience, about 20% of and presaturation of incoming blood is usually sufficient to
patients will demonstrate this MRI finding, and it can resolve this problem. Another important caveat is that MRI
precede clinical symptoms by 8 to 12 weeks, allowing signal intensity analysis is not necessarily valid in assessing
earlier detection than otherwise possible (408,427). lung abnormalities in these patients (408,425). In our expe-
Certain caveats, however, should be emphasized in rience, high signal intensity can often be seen in areas of
regard to interpretation of MRI in patients with treated pulmonary radiation fibrosis, in the absence of any residual
lymphoma. As mentioned previously, cautious interpreta- tumor. Clearly, bronchial secretions accumulating in areas
of previously irradiated lung can increase signal intensity on negative FDG-PET is most likely in patients who are
T2-weighted images. In areas of irradiated lung, the ciliated imaged near the end of chemotherapy (437,439).
respiratory epithelium is usually damaged and mucociliary In a study by Zinzani et al. (436), the utility of PET was
clearance of secretions from the lung is reduced. Secretions compared to that of CT after induction treatment
accumulate in damaged areas of lung, and in bronchi (chemotherapy with or without radiation) in 75 patients
dilated because of surrounding lung fibrosis. MRI signal with HL and aggressive NHL. PET positivity after induction
intensity analysis should be limited to purely mediastinal treatment was highly predictive for the presence of residual
and hilar masses and exclusive of their intrapulmonary disease, whereas CT findings were much less helpful. After
manifestations. treatment, four out of five (80%) patients who were
PET() and CT() relapsed. Among the 41 CT()
patients, 10 out of 11 (91%) who were PET() relapsed, as
Fluorine-18 Fluorodeoxyglucose Positron
compared with 0 out of 30 who were PET(). None of
Emission Tomography Diagnosis of Residual 29 patients who were PET()/CT() relapsed (436).
Tumor and Recurrence
Gallium-67 scintigraphy has been shown to be a valuable
adjunct in HL and NHL, detecting residual tumor follow- LEUKEMIA
ing chemotherapy and tumor relapse (431–433). However,
this technique has been largely supplanted by FDG-PET, Mediastinal lymph nodes can be involved in patients
where available, because of its improved accuracy (434). with leukemia, particularly lymphocytic leukemia. In one
FDG-PET is more sensitive and specific than CT in diag- study, involvement of thoracic lymph nodes was present
nosing tumor recurrence in patients with lymphoma at autopsy in 50% and 67% of patients, respectively, with
(433–436). In a study by Guay et al. (437), 48 patients acute and chronic lymphocytic leukemia (Fig. 4-100) and
with HL were followed using both FDG-PET and CT after 35% and 36% of patients with acute and chronic myel-
the completion of chemotherapy. Thirty-four patients were ogenous leukemia (Figs. 4-101 and 4-102), although
disease free during a mean follow-up of 605 days, and lymph node enlargement was present in only 41% of
14 relapsed during a mean follow-up of 197 days. The these; on chest radiographs, enlarged mediastinal lymph
sensitivity and specificity of FDG-PET for predicting nodes were visible in 17% (440). Mediastinal lymph
relapse were 79% and 97%, respectively. The diagnostic node enlargement is more common than hilar node
accuracy of FDG-PET (92%) was significantly higher than enlargement (440).
the accuracy of CT (56%) (p .0005) (437). Similar Approximately 15% of children with acute lympho-
results have been reported in another study (438), in blastic leukemia have so-called lymphoma syndrome, with
which coregistered (combined) PET/CT was compared mediastinal mass present in more than 50% of cases; differ-
with contrast-enhanced CT alone, for the diagnosis of entiation of T-LBL with marrow involvement from leukemia
lymph node involvement after treatment in patients with with lymphoma syndrome is difficult (4).
HL and high-grade NHL. In 41 patients studied, PET/CT In patients with acute or chronic myelogenous
had a sensitivity of 85% and a specificity of 100% in diag- leukemia, masses of malignant myeloid precursor cells
nosing mediastinal node involvement; in contrast, CT had may be found in an extramedullary location; these masses
a sensitivity of 69% and a specificity of 86% (438). False- are termed “granulocytic sarcoma” (441). They most
A B
Figure 4-100 Chronic lymphocytic leukemia with mediastinal and axillary lymph node enlargement. A: Contrast-enhanced CT at the aor-
tic arch level shows large pretracheal and retrotracheal lymph nodes (arrows). Axillary lymph node enlargement (arrowheads) is also visible.
B: At a lower level, large pretracheal lymph nodes (arrow) are visible. A right pleural effusion is also present.
A B
C D
Figure 4-101 Chronic myelogenous leukemia with extramed-

ullary involvement of mediastinal lymph nodes (granulocytic sar-
coma). A, B: Posteroanterior and lateral chest radiographs show a
superior and anterior mediastinal mass. C: CT at the level of the
aortic arch shows enlarged pretracheal and prevascular lymph
nodes with evidence of superior vena cava obstruction (black
arrow). Numerous enhanced veins in the chest wall, including the
internal mammary vein (white arrow), serve as collateral pathways.
D: At a lower level, enhancement of the azygos and hemiazygos
veins (arrows) reflects their function as collaterals. E: The azygos
arch (arrow) and left superior intercostal vein (arrowheads) are
E densely opacified and represent significant collaterals.
and the plasma cell type (442). From a clinical standpoint,

Castleman disease is also classified as localized or multicen-
tric (Table 4-26) (35).
The hyaline-vascular type of Castleman disease occurs
in up to 90% of cases, and is characterized histologically
by lymph nodes having hypervascular hyaline germinal
centers marked by extensive capillary proliferation.
Patients with the hyaline-vascular type are usually children
or young adults and are usually asymptomatic, and their
disease behaves in a benign fashion, with cure following
complete surgical resection (35,444). Up to 70% of
patients have an asymptomatic, localized mediastinal
mass on radiographs and CT.
Figure 4-102 Chronic myelogenous leukemia with extramed- Unlike the hyaline-vascular type of Castleman disease,
ullary involvement of mediastinal lymph nodes (granulocytic
sarcoma): MRI appearance. T1-weighted image shows a large lob- the plasma cell variety often presents as a multicentric
ulated mass involving the anterior mediastinum. This appearance process, associated with generalized lymphadenopathy
is nonspecific.
and hepatosplenomegaly; most cases of multicentric
Castleman disease are of the plasma cell variant, with
some cases having mixed histology. Clinically, multicen-
commonly are present at the time of first diagnosis of tric disease occurs in an older population than localized
leukemia. In about 50% of cases, masses involve the medi- Castleman disease, with most patients being in their
astinum (Figs. 4-101 and 4-102), either as a focal mass or fifth or sixth decade (444). It often results in a systemic
generalized widening. Lymph node enlargement or medi- illness, associated with anemia, infections, malignan-
astinal infiltration can be seen. Any portion of the medi- cies such as lymphoma or Kaposi sarcoma (KS), fever,
astinum can be affected (441). hypergammaglobulinemia, and the POEMS syndrome
(polyneuropathy, organomegaly, endocrinopathy, mono-
clonal protein abnormality, and skin abnormalities)
CASTLEMAN DISEASE (362). When associated with localized node involve-
ment, such systemic findings usually disappear following
Castleman disease (also referred to by a number of other total resection; however, the multicentric form of the
terms, including angiofollicular mediastinal lymph node disease is difficult to treat and usually progressive, even
hyperplasia, angiomatous lymphoid hamartoma, and giant with the use of steroids and chemotherapeutic agents.
mediastinal lymph node hyperplasia) is a disease of On contrast-enhanced CT, Castleman disease localized
unknown etiology (362,442,443). Histologically, two forms to the mediastinum typically shows dense contrast
of the disease have been described, the hyaline-vascular type enhancement (Fig. 4-103) (36,95,102,445); any mediasti-
TABLE 4-26
CASTLEMAN DISEASE: CHARACTERISTICS OF LOCALIZED
AND MULTICENTRIC TYPES
Localized Diffuse
Typical histology Hyaline-vascular type Plasma cell type

Age Children and young adults 40–60 years
Symptoms Often asymptomatic Systemic illness common
Systemic illness uncommon
Associations Uncommon Anemia, infections, malignancy,
hypergammaglobulinemia,
POEMS syndrome
CT findings Localized enhancing Generalized adenopathy, enhancing
mediastinal mass nodes, hepatosplenomegaly
Treatment Excision Chemotherapy, steroids
Prognosis Good Poor
POEMS syndrome, polyneuropathy, organomegaly, endocrinopathy, monoclonal protein abnormality,

and skin abnormalities.
Castleman disease. Hypervascular follicular hyperplasia

has also been seen in association with KS. Similar features
have also been described in patients with AIL. This is espe-
cially significant given the propensity of AIL to evolve into
malignancy.
OTHER LYMPHOPROLIFERATIVE
DISORDERS
In addition to those diseases listed previously, a variety of

uncommon lymphoproliferative diseases affecting the lung
can be associated with hilar or mediastinal lymph node
enlargement (442,449). These include post-transplantation
Figure 4-103 Localized Castleman disease. Contrast-enhanced lymphoproliferative disorders, AIL, lymphoid interstitial
CT at the level of the aortic arch shows a lobulated mass in the right pneumonitis, and lymphomatoid granulomatosis (442,
paratracheal space with marked, uniform enhancement. Biopsy 449,450). Among these, lymph node enlargement is most
proved Castleman disease.
common with AIL, a disease often occurring in patients
older than 50 years, characterized by enlarged, hypervascu-
nal compartment can be involved, as can abdominal lar lymph nodes, constitutional symptoms, and infections
lymph nodes (Figs. 4-89D and 4-104) (446). Central, (442,449).
dense, or flocculent lymph node calcifications can
occasionally be seen (36,362,447). In a report of CT
findings in six patients with multicentric Castleman dis- METASTATIC TUMOR
ease, all had retroperitoneal lymphadenopathy and
splenomegaly; axillary, inguinal, mesenteric, peripancre- Metastases to mediastinal lymph nodes from extrathoracic
atic, and mediastinal lymphadenopathy were also seen malignancies are uncommon. In a review of 1,071 cases of
(36,448). Patients with multicentric Castleman disease extrathoracic neoplasms, only 25 (2.3%) had evidence of
evaluated using dynamic contrast-enhanced CT have hilar or mediastinal adenopathy (451). The extrathoracic
shown early, dense, uniform enhancement of enlarged tumors most likely to metastasize to the mediastinum are
mediastinal lymph nodes (35). Further imaging findings carcinomas of the head and neck, genitourinary tract, and
reported in patients with multicentric disease include breast, and malignant melanoma.
skin thickening, pleural and pericardial effusions, and Most metastatic tumors cause lymph node enlargement
ascites (35,444,448). without distinguishing characteristics. Hypervascularity is
Similar findings have more recently been identified in the exception, most often secondary to metastatic leio-
patients with AIDS, including hypervascular follicular myosarcoma, neurofibrosarcoma, melanoma, renal cell
hyperplasia, indistinguishable from the plasma cell type of carcinoma, and papillary thyroid carcinoma. Enhancing
lymph nodes also occur with some frequency in pa-
tients with primary lung cancer, metastatic to mediastinal
nodes. Calcification may also occur, especially in patients
with metastatic mucinous adenocarcinomas (Figs. 4-84C
and 4-85).
The location of enlarged nodes is sometimes suggestive
of the primary tumor site. Lymph node enlargement involv-
ing posterior mediastinal and paravertebral lymph nodes
suggests an abdominal location for the primary tumor, and
superior mediastinal lymph node involvement suggests a
head and neck tumor. Internal mammary lymph node
metastases are most likely caused by breast carcinoma
(314). In a review of 219 women with breast cancer (314)
who had CT, 20.5% had internal mammary lymph node
enlargement. Of these, 71% had unilateral involvement,
Figure 4-104 Multicentric Castleman disease. Extensive, enhanc- and multiple intercostal spaces were involved in 57%; the
ing mediastinal lymphadenopathy (arrows) in a patient with the upper intercostal spaces were most frequently involved.
diffuse form of Castleman disease. The enhancement is typical.
Enlarged axillary, abdominal, and inguinal lymph nodes were also Average node diameter was 1.95 cm (range 0.6 to 6.0 cm).
visible. Bilateral pleural effusions are also present. Paracardiac lymph node enlargement can occur as a result
of metastasis from abdominal or thoracic tumors in approx- Typically, lymph node enlargement involves the hilar as
imately equal numbers. In two studies (319,320) review- well as mediastinal lymph node groups, and masses
ing the causes of paracardiac lymph node enlargement, appear bilateral and symmetrical in the large majority of
although a variety of metastatic tumors were responsible, patients (365,453); this combination usually allows the
the most common were colon carcinoma, lung carcinoma, differentiation of sarcoidosis from lymphoma. On plain
ovarian carcinoma, and breast carcinoma. radiographs, in order of decreasing frequency, hilar, para-
tracheal, aortopulmonary, subcarinal, and prevascular
lymph nodes most commonly appear abnormal; internal
SARCOIDOSIS mammary, paravertebral, and retrocrural lymph node
enlargement can also be seen but is much less common
Mediastinal lymph node enlargement is very common (366). The presence of hilar lymph node enlargement is so
with sarcoidosis, occurring in 60% to 90% of patients at typical of patients with sarcoidosis who show lymph node
some stage in their disease. About half of these will also enlargement on chest radiographs that the absence of this
show findings of lung disease on plain radiographs. In a finding in a patient with mediastinal lymphadenopathy
study of plain radiographs by Kirks et al. (365), lymph should lead one to question the diagnosis.
node enlargement was seen in association with lung In CT studies of patients with sarcoidosis, lymph node
disease in 41% of 150 patients, and in 43%, it appeared as enlargement has been reported in more than 80% of cases
an isolated abnormality. However, it must be recognized (364,366,452,454), with the great majority showing both
that a greater percentage of patients with lymph node hilar and mediastinal lymph node enlargement (Fig. 4-105).
enlargement caused by sarcoidosis will show lung abnor- However, because CT is more sensitive than chest radio-
malities on high-resolution CT (452). graphy in demonstrating lymph node enlargement, the
A B
C D
Figure 4-105 Mediastinal lymphadenopathy in sarcoidosis. Lymph node enlargement is extensive and tends to be symmetrical.
A: Enhanced CT at the level of the aortic arch reveals marked enlargement of pretracheal lymph nodes (small arrow) and anterior mediasti-
nal lymph nodes (large arrows). The lymph nodes appear homogeneous in attenuation. B: At a lower level, enlarged lymph nodes are visible
in the aortopulmonary window (large white arrow), precarinal and peribronchial (tracheobronchial) regions (black arrows), and the subcarinal
space (small white arrow). C: Below the level shown in B, precarinal (black arrow), subcarinal (small white arrow), and hilar lymph node
enlargement (large white arrows) is visible. D: Below C, extensive subcarinal lymph node enlargement (arrows) and bilateral hilar lymph node
enlargement are visible.
frequency of involvement of specific node-bearing regions

is somewhat different when CT rather than radiographs
is used for analysis. Specifically, hilar lymph node enlarge-
ment, although commonly seen on CT, is not invari-
ably present in patients with mediastinal adenopathy (Table
4-27) (367). Bilateral node involvement is visible more
often on CT than on radiographs, but nonetheless, asym-
metry with a right-sided predominance of lymph node en-
largement is also identified using CT (367,454). Among
mediastinal lymph node groups, paratracheal, tracheo-
bronchial, and peribronchial groups are most commonly
involved (Fig. 4-106) (Table 4-27).
In many patients with sarcoidosis, chest radiographic and
clinical findings are sufficient for diagnosis and clinical man-
agement, but approximately 25% of cases are associated with
atypical findings and CT is obtained for further evaluation
(364,455). Even in these patients, however, CT can often
show typical lymph node abnormalities. In a group of 25 pa-
tients with sarcoidosis having CT because plain radiographs
suggested malignancy or other disease or were not diagnostic
of sarcoidosis, 84% had right paratracheal (56%) or bilateral
(28%) paratracheal lymph node enlargement and 60% had
subcarinal node enlargement (364). Anterior mediastinal
lymph node enlargement was present in 56%; although ante-
Figure 4-106 Anterior mediastinal lymph node enlargement in
rior lymphadenopathy has been regarded as unusual based sarcoidosis. Enlarged lymph nodes are visible involving the internal
on plain radiographic studies, it is very common on CT. mammary chain (small white arrows), prevascular region (large
white arrows), and pretracheal space (black arrow). A, aorta; V,
Although 81% of patients had hilar lymph node enlarge-
superior vena cava.
ment, it was bilateral in only 56% and unilateral in 25%; this
atypical distribution likely reflects the groups studied.
In patients with sarcoidosis, lymph nodes can be
several centimeters in diameter (Fig. 4-105), but sarcoid is A variety of patterns of pulmonary involvement from
not generally associated with large localized masses as is small nodules to large ill-defined masses or pulmonary
lymphoma. Lymph nodes can calcify densely (Fig. 4-107) fibrosis can also be seen in patients with sarcoidosis.
or show a stippled or egg-shell appearance (Fig. 4-84A) However, it should be kept in mind that not all patients
and rarely enhance (Fig. 4-89B) or appear necrotic on with active pulmonary sarcoidosis show lymph node
contrast-enhanced scans (455). enlargement on CT. It is common to see typical features of
TABLE 4-27
LYMPH NODE ENLARGEMENT IN SARCOIDOSIS BY ATS NODE STATION
ATS Node Station ATS Designation Lymph Node Enlargement (%)
2R R upper paratracheal 50
2L L upper paratracheal 32.5
4R R lower paratracheal 57
4L L lower paratracheal 50
5 Aortopulmonary 52.5
6 Anterior 17.5
7 Subcarinal 52.5
8 Paraesophageal 47.5
10R R Tracheobronchial 72.5
10L L Peribronchial 35
11R R Intrapulmonary (hilar) 57.5
11L L Intrapulmonary (hilar) 65

Modified from Patil SN, Levin DL. Distribution of thoracic lymphadenopathy in sarcoidosis using computed
tomography. J Thorac Imaging. 1999;14:114–117, with permission.
A B
Figure 4-107 Lymph node calcification in two patients with sarcoidosis. A, B: Coronal reconstructions in two patients with sarcoidosis
show extensive, symmetrical, bilateral hilar (large arrows) and mediastinal (small arrows) lymph node enlargement. Many nodes are densely
calcified. A, aorta; T, trachea.
sarcoid lung disease on high-resolution CT without lymph and B) (Table 4-28). However, it should be noted that in
node enlargement being visible. the study by Im et al. right-sided pulmonary TB was twice
as common as disease on the left.
In a study of 100 pediatric patients (34), the most com-
INFECTIONS mon locations for lymphadenopathy were the subcarinal
space (n 90), pulmonary hila (n 85; left 74, right 72,
A variety of infectious agents can cause mediastinal lymph bilateral 61), anterior mediastinum (n 79), the precari-
node enlargement during the acute phase of disease. These nal space (n 64), and the right paratracheal region
include TB, fungal infections including histoplasmosis and (n 63). Multiple sites of involvement were present in
coccidioidomycosis, bacterial infections, and viral infec- 88 patients. Bronchial compression was identified in
tions. Typically, there will be symptoms and signs of acute 29 patients; most commonly, the left main bronchus was
infection, and chest radiographs will show evidence of involved (n 21), followed by the right main bronchus
lung disease, although this is not always the case. In pati- (n 14) and the bronchus intermedius (n 8) (34).
ents with prior granulomatous infections, lymph node In a study by Im et al. (32) of 23 patients with active TB,
calcification is common (345), with such nodes appearing nodes larger than 2 cm in diameter invariably showed
normal in size or enlarged. central areas of low attenuation on contrast-enhanced CT,
with peripheral rim enhancement (Fig. 4-86A and B). It
Tuberculosis
Hilar and mediastinal lymph node enlargement is com-
monly seen on CT in patients with active TB (356,456), TABLE 4-28
although it is more frequently seen in children than SITES OF LYMPH NODE ENLARGEMENT
adults (32). In a study by Im et al. (33), mediastinal lymph IN 23 PATIENTS WITH TUBERCULOSIS
node enlargement was seen on CT in 9 of 29 patients
ATS Percent
with newly diagnosed disease and in 2 of 12 patients with Site Designation Abnormal
reactivation.
Lymph node enlargement is usually seen on the side of Right paratracheal 2R 83
lung disease, but involvement of contralateral nodes can Left paratracheal 2L 4
Right paratracheal 4R 87
sometimes be present. Although Lee et al. (356) indicate Left paratracheal 4L 9
that lymphadenopathy without lung disease is unusual in Right tracheobronchial 10R 65
TB patients, except in the presence of AIDS, in a study by Subcarinal 7 52
Im et al. (32), only 61% of patients with tuberculous Aorticopulmonary 5 35
lymphadenopathy had evidence of pulmonary TB on chest Right hilar 11R 30
Left hilar 11L 9
radiographs. Right-sided adenopathy usually predominates
(32,356); right paratracheal lymph node enlargement was ATS, American Thoracic Society; R, right; L, left.
seen in 87% of the cases reported by Im et al., whereas right Modified from Im J-G, Song KS, Kang HS, et al. Mediastinal tubercu-
lous lymphadenitis: CT manifestations. Radiology. 1987;164:115–119,
tracheobronchial nodes were abnormal in 65% (Fig. 4-86A with permission.
should be noted, however, that these areas of low attenua- the remaining three patients (13%), nodes were homoge-
tion are not of water density but range from about 40 to neous and hypointense on both T1- and T2-weighted
60 HU; they are visible only on contrast-enhanced scans. images and did not enhance after contrast infusion. Such
Although smaller nodes did not typically show low attenu- nodes were fibrocalcific and patients had no symptoms.
ation, only 1 of the 23 patients had nodes less than 2 cm Similar findings have been reported by Kushihashi
in diameter. In a small percentage of cases, areas of low et al. (457).
attenuation can be seen extending outside nodes, with
obliteration of mediastinal fat, representing tuberculous
Fungal Infections
mediastinitis (457) or a cold abscess (32). In patients with
large areas of low attenuation seen on CT, constitutional Although some degree of lymph node enlargement may
symptoms were frequent. Rim-enhancing lymph nodes be seen with a variety of fungal infections, lymph node
have also been reported in nearly 85% of AIDS patients enlargement is most typical of histoplasmosis and coccid-
and 67% of HIV(+) patients with culture or histologically ioidomycosis.
verified TB (458). In a study of 100 pediatric patients with Infection with Histoplasma capsulatum is a well-
clinically suspected primary TB, nodes greater than 1 cm recognized cause of hilar and mediastinal lymph node
were present in 46. Enhancement of lymph nodes was enlargement. In reported patients with acute or chronic
present in 67 and was almost always ghostlike or poorly histoplasmosis, CT has shown paratracheal, subcarinal, and
defined ring enhancement (34). Calcification was present hilar lymph node enlargement with irregular enhancement
in only 9 patients. and necrotic regions, as in patients with TB (463,464).
CT more accurately defines the presence and extent of Masses ranged up to several centimeters in diameter and
lymph node enlargement than does routine chest radio- may consist of a solitary nodal mass or a matted group of
graphy in patients with TB. In particular, the finding of lymph nodes. Rim enhancement or apparent enhancing
low-attenuation necrotic lymph nodes, with rim enhance- septa within these masses have been described. These
ment following contrast infusion, strongly suggests a masses often displace mediastinal structures such as the
diagnosis of mycobacterial infection, both in immuno- superior vena cava or esophagus. Calcification is common.
competent patients and in patients who are HIV positive If enlarged lymph nodes in histoplasmosis are localized,
(32,458–460). Although a similar appearance may be seen they may be termed “mediastinal granuloma.” More exten-
in patients with fungal infections, especially cryptococcosis sive mediastinal abnormalities may be seen in patients
and histoplasmosis, this appearance is less frequent in with histoplasmosis, termed fibrosing, sclerosing, or granu-
patients with lymphadenopathy resulting from metastatic lomatous mediastinitis. This entity is described in the next
neoplasm, KS, or lymphoma, and in our experience almost section.
always indicates the presence of a treatable infection (461). Hilar or mediastinal lymph node enlargement may be
In some cases CT can serve as a guide for determining the seen in patients with acute or chronic coccidioidomycosis
best sites for node biopsy and can help determine whether (465). Hilar lymph node enlargement is common in
mediastinoscopy or parasternal mediastinotomy is most patients with acute disease, being seen in about 20% of
appropriate. cases. Mediastinal lymph node enlargement is less com-
Three patterns of lymph node enlargement have been mon, and is said to portend dissemination (Fig. 4-108)
reported on MRI in patients with TB (462), correlating
with clinical symptoms and pathologic findings, and par-
alleling the expected CT findings in such patients (32). In
14 (61%) of 23 cases, nodes were inhomogeneous with
marked peripheral enhancement after injection of contrast
material (462). Areas of enhancement were intermediate
in intensity on T1-weighted images and hypointense on
T2-weighted images, corresponding pathologically to pe-
ripheral granulation tissue within nodes. Unenhanced
areas were hypointense on T1-weighted images and hyper-
intense on T2-weighted images, corresponding to areas of
central caseation or liquefaction. Almost all patients with
this pattern had clinical symptoms of active TB.
In six (26%) patients with the second pattern of node
involvement, nodes were relatively homogeneous and
hyperintense to muscle on both T1- and T2-weighted
images and enhanced homogeneously after contrast
infusion (462). This pattern correlated with the presence
Figure 4-108 Coccidioidomycosis with lymph node enlarge-
of granulomas without necrosis or with minimal necrosis; ment. A right upper lobe pneumonia is associated with paratra-
symptoms were minimal or absent in these patients. In cheal lymph node enlargement (arrow).
(465). Enlarged lymph nodes may appear homogeneous the United States (Fig. 4-109) (467). Sherrick et al. (473)
in attenuation or may show areas of necrosis with rim also described a diffuse pattern of involvement seen in 18%
enhancement. Calcification of lymph nodes is quite of cases, manifested as a diffusely infiltrating, noncalcified
uncommon. mass, affecting multiple mediastinal compartments. The
diffuse pattern may be related to causes other than histo-
plasmosis or may be idiopathic (Fig. 4-110).
FIBROSING MEDIASTINITIS Typically, enlarged lymph nodes in patients with
granulomatous disease, such as histoplasmosis, TB, and
In some patients with granulomatous disease involving sarcoidosis are sharply defined. Fibrosing mediastinitis is
mediastinal lymph nodes, extension of the disease process manifested on CT by replacement of low-density mediasti-
to involve surrounding mediastinal tissues results in exten- nal fat by higher density fibrous tissue, often associated
sive acellular fibrosis with infiltration of mediastinal fat with calcification (Fig. 4-109) (467). In the presence of
(466,467). This is termed fibrosing, sclerosing, or granulo- fibrosing mediastinitis, discrete enlarged lymph nodes
matous mediastinitis. Symptomatic encasement and/or cannot be identified.
compression of a number of mediastinal structures, partic- Manifestations on CT include hilar and mediastinal
ularly vessels, and the trachea or esophagus can result masses (Fig. 4-110); stippled or diffuse calcification; and
(468). The most common causes are histoplasmosis compression and/or encasement of the trachea, main
(Fig. 4-109), TB, and sarcoidosis, but fibrosing mediastini- bronchi, mediastinal vessels, or esophagus (Fig. 4-109)
tis can also be associated with autoimmune diseases, (467,469). Those structures most often involved are those
retroperitoneal fibrosis, sclerosing cholangitis, Behçet that have the thinnest walls (e.g., superior vena cava) or the
disease, Riedel thyroiditis, pseudotumor of the orbit, or longest mediastinal course (e.g., trachea and left main
drugs (i.e., methysergide) or may be idiopathic (467, bronchus and right pulmonary artery). In a study of patients
469–472). In patients with histoplasmosis, it has been with fibrosing mediastinitis, the most common complica-
suggested that fibrosis may result from seepage of fungal tion was narrowing or obstruction of the superior vena
antigen from adjacent lymph nodes, and genetic suscepti- cava (39%), bronchi (33%), pulmonary artery (18%), and
bility is likely involved. In patients with TB, extension of esophagus (9%) (473). Rarely, these findings primarily
the infectious process into the mediastinum may also be affect the posterior mediastinum, with esophageal encase-
associated (32,457). Most patients are young adults at the ment and dysphagia predominating (471).
time of presentation. Contrast-enhanced CT is useful in patients with suspected
Two distinct patterns of mediastinal involvement may be vascular abnormalities such as superior vena cava obstruc-
seen in patients with fibrosing mediastinitis (Table 4-29) tion or encasement or obstruction of pulmonary arteries and
(467,473). Sherrick et al. (473) described a focal pattern, veins. Both arterial and venous obstruction can cause in-
seen in 82% of cases, manifested by a localized mass or farcts, which are manifested on CT as peripheral areas of ho-
masses of soft tissue attenuation, usually in the right para- mogeneous opacity. These infarcts are typically wedge
tracheal or subcarinal regions or hila, and often calcified shaped and do not contain air bronchograms. Pulmonary
(63% of cases). This localized form of fibrosing mediastini- venous obstruction may also result in areas of ground-glass
tis is most likely caused by histoplasmosis in patients from attenuation or thickening of interlobular septa, which may
A B
Figure 4-109 Fibrosing (granulomatous) mediastinitis caused by histoplasmosis. After diffuse mediastinal involvement, healing by fibrosis
and diffuse calcification may be seen. In this patient, extensive mediastinal calcification is visible. The superior vena cava is likely obstructed.
TABLE 4-29
CHARACTERISTICS OF FOCAL AND DIFFUSE FIBROSING MEDIASTINITIS
Focal Diffuse
Frequency 80% 20%

Causes and Usually histoplasmosis in Histoplasmosis less frequent;
associations the United States autoimmune diseases, retroperitoneal
fibrosis, sclerosing cholangitis, Behçet
disease, Riedel thyroiditis, pseudotumor
of the orbit, drugs (i.e., methysergide),
idiopathic
CT findings Localized mass or masses, Diffuse infiltrating mass; multiple sites
often right paratracheal, involved; noncalcified
hilar, subcarinal; calcification
may be present
A B
C D
Figure 4-110 Fibrosing mediastinitis in a 24-year-old woman. A: Right hilar and mediastinal lymph node enlargement (arrows) is visible
on contrast-enhanced CT. The enlarged lymph nodes are uncalcified. B: Focal ground-glass opacity in the right upper lobe is associated with
right superior pulmonary vein obstruction. She had previously suffered a pulmonary infarct in this region. C: T1-weighted MRI shows hilar
and subcarinal lymph node enlargement (arrows). The right superior pulmonary vein appears obstructed. D: T2-weighted MRI with fat satu-
ration shows high-intensity right hilar lymph nodes (arrows) indicative of active inflammation. Mediastinal biopsy yielded dense fibrosis.
There was no evidence of histoplasmosis.
indicate the presence of pulmonary edema (Fig. 4-110B). CT MEDIASTINAL CYSTS

is also useful for assessing the site, length, and severity of air-
way stenosis, particularly when thin-collimation and 2D or Cysts account for about 9% of primary mediastinal
3D reconstruction techniques are used. masses in adults; foregut duplication cysts account for
MRI findings in patients with fibrosing and granuloma- 11% of mediastinal masses in children (4). Most medi-
tous mediastinitis have also been reported (467,474). astinal cysts are congenital in origin. These include
Although MRI and CT are equivalent in defining the extent foregut duplication cysts (i.e., bronchogenic, esophageal
of adenopathy or fibrosis, CT is superior in demonstrating duplication, and neurenteric cysts) and pleuropericardial
node calcification, a finding that is often important in cysts (Table 4-30) (93,94,476,477). In a study of 105 pa-
making the diagnosis. On T1-weighted MR images, fibros- tients with mediastinal cysts (478), there were 47 bron-
ing mediastinitis is typically manifested by a heteroge- chogenic cysts, 30 thymic cysts, 12 pericardial cysts, 7
neous, infiltrative mass of intermediate signal intensity pleural cysts, 4 esophageal duplications, 1 thoracic duct
(Fig. 4-110C). Its appearance on T2-weighted MR images cyst, 2 meningoceles, and 2 other types. Thymic cyst, de-
is variable; regions of both increased and markedly scribed previously, may be congenital or acquired. Rarely,
decreased signal intensity are frequently seen in the same hydatid cysts may occur in the mediastinum (479,480).
or in different parts of the mediastinum. It is likely that
areas of decreased signal intensity on T2-weighted images
Bronchogenic Cyst
indicate the presence of calcification or fibrous tissue
and that areas of increased signal intensity indicate more Bronchogenic cyst is most common, representing approxi-
active inflammation (Fig. 4-110D). After administration of mately 60% of foregut duplication cysts (Table 4-30). It
Gd-DTPA, the mediastinal masses usually show heteroge- probably results from defective growth of the lung bud dur-
neous enhancement. MRI is particularly valuable in assessing fetal development (84,94,481). Bronchogenic cysts are
ing vascular patency without the need for intravenous lined by pseudostratified ciliated columnar epithelium, typi-
contrast media (Fig. 4-111). FDG-PET may show uptake in cal of the respiratory system, and frequently associated with
areas of inflammation (475). smooth muscle, mucous glands, or cartilage in the cyst wall.
A B
Figure 4-111 Fibrosing (granulomatous) mediastinitis: MR

imaging. A–C: On T1-weighted images, low-intensity tissue in the
right paratracheal mediastinum (arrows, A and B) reflects fibrosis
secondary to histoplasmosis. Obstruction of the right pulmonary
artery is present (arrow in C). The patient has had prior bypass of
C the superior vena cava because of obstruction.
TABLE 4-30
CHARACTERISTICS OF MEDIASTINAL CYSTS
Bronchogenic Cyst Cause pain, often diagnosed in children
60% of foregut duplication cysts Rarely contain air
Lined by pseudostratified ciliated columnar epithelium
Pericardial Cyst
Cyst fluid often proteinaceous
Coelomic in origin
Symptoms may result from compression of structures
Usually asymptomatic
CT findings
90% contact the diaphragm
Rounded or elliptical
65% in the right cardiophrenic angle
Attenuation of contents similar to soft tissue in half
25% in the left cardiophrenic angle
Half occur near the carina
Usually water attenuation on CT
May occur in any mediastinal compartment
Wall may calcify Thoracic Duct Cyst
Rare
Esophageal Duplication Cyst Compression of structures may result in symptoms
Less common than bronchogenic cyst Mediastinal or cervical, contiguous with thoracic duct
Lined by gastrointestinal tract mucosa Sharply marginated, wall thin
Often connected to the esophagus Low attenuation on CT
60% occur in the lower mediastinum Size may vary with lymph flow
Usually adjacent to the esophagus
Indistinguishable from bronchogenic cysts on CT, except for Thymic Cyst
location Occur within the thymus, rarely cervical
Cysts may contain ectopic gastric mucosa Congenital
Usually unilocular
Neurenteric Cyst Acquired
Rare Associated with inflammation, radiation, surgery, thymic
Composed of both neural and gastrointestinal elements tumors
Connected to the meninges Often multilocular
Associated with a defect in one or more vertebral bodies in half CT attenuation similar to water in most
Located posteriorly, in relation to the spine Calcification of wall may occur
Bronchogenic cysts contain fluid, which ranges in color from On CT, bronchogenic cysts usually appear rounded or el-
clear to milky white to brown; the fluid can contain variable liptical, smooth in contour, and sharply marginated. The
amounts of protein and can be either serous in nature, hem- wall of a bronchogenic cyst appears thin or is imperceptible
orrhagic, or highly viscous and gelatinous. Cyst fluid can (Fig. 4-112). Rarely, calcification of the cyst wall is present.
rarely contain milk of calcium (calcium oxalate crystals). Because of the variable composition of fluid contained
A B
Figure 4-112 Subcarinal bronchogenic cyst. A, B: A low-attenuation bronchogenic cyst is visible (arrows) on a contrast-enhanced CT. Its
wall is imperceptible.
within bronchogenic cysts, their attenuation on CT is highly aspiration has been used for the diagnosis and treatment
variable (4,482–488). Half of bronchogenic cysts are of of bronchogenic cysts and other benign mediastinal fluid
water attenuation. In the other half, the CT density can vary collections, including esophageal duplication cysts and
from near that of water to more than 100 HU, if milk of cal- mediastinal pseudocysts (492–497). In select cases, these
cium is present. When dense, bronchogenic cysts may be procedures may obviate more invasive procedures such as
difficult to distinguish from solid lesions. The cyst wall may mediastinoscopy or thoracotomy.
appear to be denser than the cyst contents; when this is the
case, the wall appears very thin. An important clue to the di- Esophageal Duplication Cyst
agnosis can be their lack of enhancement on scans obtained
Esophageal duplication cysts are lined by gastrointestinal
following intravenous contrast enhancement. MRI may also
tract mucosa and are often connected to the esophagus
be used to determine their cystic nature; bronchogenic cysts
(Table 4-30) (94). Sixty percent are found in the lower
invariably have a high intensity signal on T2-weighted
posterior mediastinum, adjacent to the esophagus, and
images (94). Mediastinal bronchogenic cysts rarely contain
are sometimes found within its wall (Fig. 4-114)
air or become infected, although this is common in patients
(84,496,498). Their appearance on CT is indistinguishable
with pulmonary cysts.
from that of bronchogenic cysts, except for their location
Bronchogenic cysts can be seen in any part of the medi-
(498,499). Rarely, these cysts may calcify (500).
astinum, but most are located in the middle or posterior
They may result in dysphagia, pain, or other symptoms
mediastinum, near the carina (52%) (Fig. 4-112), in the
owing to compression of adjacent structures, but many are
paratracheal region (19%) (Fig. 4-113), adjacent to the
asymptomatic. The majority are detected in infants or
esophagus (14%), or in a retrocardiac location (9%) (489).
children. In 50% of pediatric patients, these cysts contain
They rarely occur in the anterior mediastinum or the infe-
ectopic gastric mucosa and radionuclide imaging with the
rior aspect of the posterior mediastinum (93). Large bron-
use of technetium-99m sodium pertechnetate may be
chogenic cysts may be associated with symptoms because
helpful in diagnosis.
of compression of adjacent structures, such as the trachea
and carina, mediastinal vessels, and the left atrium
(490,491).
Neurenteric Cyst
Small, asymptomatic cysts can be followed (492). These rare lesions are connected to the meninges through
However, enlargement over years is typical, and rapid a midline defect in one or more vertebral bodies, and are
enlargement can indicate hemorrhage or infection and be composed of both neural and gastrointestinal elements
associated with pain (94). Because of their tendency to (Table 4-30) (84). A connection with the esophagus is
increase in size, surgical management has been traditional often present. The appearance of neurenteric cyst on CT is
(493). Recently, percutaneous and transbronchial needle the same as that of other duplication cysts, but the pres-
A B
Figure 4-113 Bronchogenic cyst: CT/MRI correlation. A: Coned-down view from a posteroanterior radiograph of the chest shows a mass
(arrow) of uncertain etiology, initially thought possibly to be bronchogenic carcinoma. B: Nonenhanced CT section through the superior me-
diastinum shows a well-demarcated mass (arrow) with a density measuring within the range of soft tissue. This patient could not receive in-
travenous contrast media because of an allergy history. C: T1-weighted MR image at the same level as in B shows the mass to be of very low
signal intensity (arrows). The mass is clearly distinct from the adjacent vessels and mediastinal fat. D–G: Single-level, multi-echo sequence
obtained at the same levels as shown in B and C, using TEs of 30, 60, 90, and 120 ms, respectively. This technique allows acquisition of heav-
ily T2-weighted images. Note that the signal intensity of the lesion steadily increases relative to all other tissues, including fat. This combi-
nation of findings is diagnostic of a cystic mass, in this location almost certainly a bronchogenic cyst. TE, echo time. (From Naidich DP,
Rumancik WM, Ettenger NA, et al. Congenital anomalies of the lungs in adults: MR diagnosis. AJR Am J Roentgenol. 1988;151:13–19, with
permission.)
D E
F G
ence of the vertebral abnormality points to the diagnosis; diaphragm, with 65% occurring the right cardiophrenic
vertebral anomalies are present in about half of cases angle (Figs. 4-115 and 4-116) and 25% in the left cardio-
(84). The cysts rarely contain air, because of communica- phrenic angle; 10% are seen at higher levels. They can be
tion with the viscera. They frequently cause pain and are seen as high as the pericardial recesses at the level of the
generally diagnosed at a young age. They are typically proximal aorta and pulmonary arteries (501–503). Their
posterior and paravertebral in location. appearance is usually diagnostic of their cystic nature. They
are sharply marginated and have low CT numbers (near
the attenuation of water), although pericardial cysts with
Pericardial Cyst
high CT numbers have been reported (504,505).
Pericardial or pleuropericardial cyst results from a defect in Pericardial cysts are not always round; they may assume
the embryogenesis of the coelomic cavity (Table 4-30). different shapes when studied at different times (506). In
Their walls are composed of connective tissue and a single some cases, cysts can extend into the major fissure. When
layer of mesothelial cells (94). Most patients are asympto- small, they can sometimes be confused with enlarged
matic. Approximately 90% of pericardial cysts contact the cardiophrenic angle lymph nodes.
Figure 4-114 Esophageal duplication cysts. A: Contrast-

enhanced CT section through the lower chest shows a well-
defined fluid-containing mass adjacent to the esophagus, findings
characteristic of esophageal duplication cysts. B, C: Coronal
T1-weighted (B) and transverse T2-weighted MRI (C) in a different
C patient shows an intense mass in contiguity with the esophagus.
Thoracic Duct Dilatation and Thoracic have symptoms, usually due to compression of adjacent
Duct Cyst structures and related to cysts located above the aortic arch;
symptoms include retrosternal pain, dysphagia, dyspnea,
Thoracic duct dilatation (Fig. 4-117) is uncommon, but and symptoms of superior vena cava compression. Cyst fluid
may be seen in patients with lymphangiomyomatosis is chylous and contains T lymphocytes and triglycerides.
(507), in patients with portal hypertension and hepatic Thoracic duct cysts may occur anywhere along the course of
cirrhosis (508), in some patients with thoracic duct the duct, and although most are mediastinal, there also
obstruction, and in patients with venous obstruction near seems to be a propensity for them to occur in the left neck
its termination in the neck. In patients with portal hyper- near the duct termination (512). They appear sharply mar-
tension and cirrhosis, it is thought to be due to increased ginated and low in attenuation on CT and the cyst wall is
hepatic lymph flow. In one study, using MRI, the maxi- very thin or invisible (509,512,513). MRI shows typical find-
mum thoracic duct diameter measured 3.74 0.81 mm in ings of a cyst (511). They may be up to 15 cm in diameter,
all healthy volunteers and 6.98 2.77 mm in patients but reported cases average 7 cm (511). Cysts may change size
with alcoholic cirrhosis (73). over time depending on the volume of lymph flow within
Thoracic duct cysts are rare, representing less than 1% of the thoracic duct. Treatment consists of surgical removal of
mediastinal cysts (Table 4-30) (478). They may be related to the cyst and ligation of all lymphatics connected to the cyst.
congenital or degenerative weakness of the thoracic duct wall
(509). In a review of 29 cases (510), 19 thoracic duct cysts
Differential Diagnosis of Cystic Masses
were found to be mediastinal, 9 were cervical, and 1 was ab-
dominal. The mean age of patients at diagnosis is about 50 Differentiation of an uncomplicated congenital cyst from
years (range 17 to 86 years) (510,511). About half of patients other cystic encapsulated lesions, such as abscess, chronic
A, B C
Figure 4-115 Pericardial cyst. A–C: Pericardial cyst. A well-defined low-attenuation pericardial cyst (arrows) is visible in the right cardio-
phrenic angle on an unenhanced scan with attenuation clearly lower than that of the adjacent right atrium (RA). The key to the correct diag-
nosis of a pericardial cyst versus enlarged paracardiac nodes in this case is demonstration of water density within this lesion. Alternatively,
images may be obtained both prior to and following administration of contrast to further confirm the avascular or cystic nature of this le-
sion. This approach is of particular value, as mediastinal cysts frequently measure greater than water density prior to enhancement caused
by the presence of prior hemorrhage, or elevated protein or debris within.
hematoma, cystic lymphangioma or hemangioma, cystic Cystic masses of the anterior mediastinum rarely repre-
teratoma, or other cystic tumor relies on the clinical presen- sent a congenital cyst but are more likely related to the
tation, the location and appearance of the cyst on CT, and presence of a cystic tumor, thymic cyst, or other abnormality
findings on additional imaging studies (84,93,94,514). (93,514). Mediastinal parathyroid cyst is a rare occurrence
(288). Abscess, hematoma, and cystic tumors commonly
demonstrate thick walls, with septa sometimes visible
within the cystic space, and mixed-density fluids (94).
Anterior thoracic meningocele may closely resemble a con-
genital cyst, but careful examination of adjacent sections
should demonstrate the associated vertebral anomalies or
enlargement of the adjacent neural foramen. Occasionally,
loculated fluid in the superior pericardial recess can simu-
late a cystic mediastinal mass (515).
The potential of MRI to identify cystic or fluid-filled
masses is well established (94,516). MRI has been used
successfully to diagnose a wide range of lesions, including
bronchogenic cyst, pericardial cyst, thymic cyst, colloid
cysts within a goiter, cystic hygroma, and mediastinal
pseudocyst (270,271,517–521). MRI is especially valuable
in assessing complex cysts that do not appear fluid-filled
on CT (483–487), especially when intravenous contrast
cannot be administered (Fig. 4-113) (517).
High signal intensity is characteristically seen within
cystic lesions on T2-weighted sequences regardless of the
nature of the cyst contents, but variable patterns of signal
intensity have been noted on T1-weighted sequences, pre-
sumably because of variable cyst contents and the presence
of protein and/or hemorrhage or mucoid material (522).
Figure 4-116 Pericardial cyst. Enhanced CT shows a lobulated
pericardial CT (large arrow) continuous with the pericardium (small This variable appearance can be of some value in distin-
arrows). The cyst is low in attenuation. guishing the type of cyst or its contents. In patients with
A B
Figure 4-117 A, B: Thoracic duct dilatation. The thoracic duct (arrow) is low in attenuation, and is located to the right of the aorta.
pericardial cysts containing serous fluid, the cysts always tissue, myoblastic tissue, or from osseous or cartilaginous
appeared less intense than muscle on T1-weighted images, tissues. Their prevalence and malignant potential are great-
whereas in patients with thymic cysts, the signal intensity est in children or young adults. Presenting symptoms
ratio (SIR) of cyst to muscle ranged from 0.48 to 1.65 (522). depend primarily on the location, size, propensity for
However, in patients with bronchogenic cysts or cystic invasion, histology, and the patient’s age. Except when
teratomas, the SIRs were considerably higher, reflecting the they contain fat or large vascular structures, their imaging
complex nature of the cyst contents in these lesions; in all appearance is nonspecific.
patients with these lesions, cysts appeared more intense
than muscle and the SIR ranged from about 1.4 to more
Tumors of Adipose Tissue and Fatty Masses
than 3 (522). A fluid-fluid level within a bronchogenic cyst
has been reported on T1-weighted MRI (523). Fat is recognized on CT by its low attenuation, which
varies from 40 to 130 HU (85). Fat is normally present
in the mediastinum, and its amount may increase with
TUMORS AND MASSES OF age. In patients older than 25 years, most of the fat visible
MESENCHYMAL ORIGIN in the anterior mediastinum is contained within the
fibrous skeleton of the involuted thymus. In patients with
Mesenchymal tumors or masses constitute 5% to 10% of a mediastinal mass, MRI may be useful in revealing small
adult primary mediastinal tumors, and many types are amounts of fat, a finding that may be of value in differen-
quite rare. These arise from (a) mediastinal tissues such as tial diagnosis (48).
fat, (b) mesenchymal components of recognizable medi- Normal fat is unencapsulated and equally distributed
astinal structures (e.g., thymus, aorta and pulmonary artery, throughout the connective tissue matrix of the medi-
trachea), (c) mesenchymal components of unrecognizable astinum. The contours of the mediastinum as seen on chest
mediastinal structures such as small vessels or nerves, or radiographs are not generally affected by the accumulation
(d) undifferentiated cells located within mediastinal of excessive amounts of normal fat. However, asymmetrical
tissues. Mesenchymal sarcomas originating in the medi- accumulations of fat in the anterior cardiophrenic angles
astinum may occur spontaneously, may be related to (i.e., asymmetrical epicardiac fat pads) are common and
sarcomatous degeneration of germ-cell or neurogenic can suggest the presence of a mediastinal mass on chest
tumors, or may occur because of prior radiotherapy. radiographs (524).
Mesenchymal tumors and masses include those arising Abnormalities of fat distribution can be diffuse, as in
from adipose tissue, vascular or lymphatic tissues, fibrous mediastinal lipomatosis, or focal, as in the presence of
A B
Figure 4-118 Morgagni hernia. A, B: Sequential sections through the lower thorax and upper portion of the right diaphragm, respec-
tively, show the presence of omental fat anteriorly, recognizable by the finding of visible mesenteric vessels (arrows in B) within the fat. Note
that there is some compression on the adjacent right side of the heart, most pronounced in A.
lipoma or fat-containing transdiaphragmatic hernia. In fat accumulate in the mediastinum (Table 4-31).
our experience, most fatty masses are seen in the peridi- Lipomatosis may be associated with Cushing syndrome,
aphragmatic areas and they most often represent hernia- steroid treatment, or obesity, but these factors are absent
tion of abdominal fat (Figs. 4-118 and 4-119). in up to one half of cases (525). It is unassociated with
Nonetheless, a variety of pathologic processes may be symptoms. Lipomatosis is relatively common and is
associated with fatty masses (Table 4-31). In the large often detected incidentally in patients having chest CT.
majority of cases, discovery of the fatty nature of a mass The excess fat deposition is most prominent in the
indicates its benignancy, but mediastinal liposarcomas upper mediastinum, resulting in smooth mediastinal
rarely occur (353). widening as shown on chest radiographs and convex or
bulging mediastinal pleural surfaces on CT (Fig. 4-120).
Tracheal compression or displacement is absent (89,
Mediastinal Lipomatosis
526,527). Less commonly, fat also accumulates in the car-
Lipomatosis is a benign condition in which overabun- diophrenic angles and paraspinal regions (Figs. 4-121 and
dant amounts of histologically normal, unencapsulated 4-122) (528–530). In patients with lipomatosis, the fat
A B
Figure 4-119 Large paraesophageal omental hernia in a 39-year-old woman with dysphagia. A: CT scan at level of an apparent soft tissue
mass identified in the lower thorax overlying the left heart border on the corresponding chest radiograph. Note the presence of a large,
homogeneous fatty mass in the posterior mediastinum within which faint linear densities can be identified compatible with omental vessels.
B: Oblique reconstruction through the axis of the line seen in A confirms the true intra-abdominal origin of the mass found at surgery to rep-
resent omental fat extending into the mediastinum via a large esophageal hiatus. Peridiaphragmatic fatty masses are much more likely to
represent hernias containing fat than lipomas.
TABLE 4-31
TUMORS OF ADIPOSE TISSUE AND FATTY MASSES
Mediastinal Lipomatosis Symptoms of compression in two thirds
Normal, unencapsulated fat CT and MRI
Half associated with Cushing syndrome, steroid treatment, or Mass with variable amounts of fat and soft tissue
obesity Sharply marginated or invasive
Most prominent in the upper mediastinum
Lipoblastoma
Less often in cardiophrenic angles and paraspinal regions
Benign but infiltrating tumor
CT: homogeneous fat attenuation
Arises from embryonal fat
Multiple Symmetrical Lipomatosis Usually diagnosed in children younger than 3 years
Similar to mediastinal lipomatosis in appearance Involvement of chest wall may occur
Compression of mediastinal structures, often trachea CT and MRI
Mass consisting primarily of fat
Lipoma Strands of soft tissue within the mass
Most common in the prevascular space
Compression of structures uncommon Hibernoma
CT: homogeneous fat attenuation Benign neoplasm derived from brown adipose tissue
Usually in locations where normal brown fat is found in infants,
Liposarcoma e.g., periscapular or interscapular region, neck, axilla,
Adults mediastinum
Most common in the prevascular space CT: fat attenuation
May arise within the thymus Metabolically active, may mimic malignancy on FDG-PET
FDG-PET, fluorine-18 fluorodeoxyglucose positron emission tomography.
should appear homogeneously low in attenuation, sharply patients with a suitable clinical history, usually requires no
outlining mediastinal vessels and lymph nodes. If the fat further evaluation; however, if confirmation of this condi-
appears inhomogeneous or the margins of mediastinal tion is desired, CT is diagnostic.
structures are ill defined, superimposed processes such as A rare condition, termed multiple symmetrical lipo-
mediastinitis, hemorrhage, tumor infiltration, fibrosis, or matosis, may mimic the appearance of simple lipomatosis
postsurgical abnormalities should be considered. The on CT; however, this entity often results in compression of
appearance of smooth mediastinal widening on plain radi- mediastinal structures, the trachea in particular, and does
ographs is characteristic of mediastinal lipomatosis and, in not involve the anterior mediastinum, cardiophrenic
angles, or paraspinal regions. In addition, periscapular
lipomatous masses are almost always present (88).
Lipoma
Mediastinal lipoma is uncommon, constituting approxi-
mately 2% of all mediastinal tumors (531,532). As with
other mesenchymal tumors, lipomas can occur in any part
in the mediastinum but are most common in the prevas-
cular space (Fig. 4-123) (Table 4-31) (533). Lipomas are
soft and pliable and do not result in symptomatic com-
pression of adjacent structures unless they are very large.
They may or may not be encapsulated. Although they
variably contain fibrous septa, lipomas appear to have
homogeneously low CT numbers (534).
Liposarcoma
Mediastinal liposarcoma is rare. It is malignant and
composed largely of fat. Liposarcoma often occurs in the
Figure 4-120 Mediastinal lipomatosis. CT image at the level of anterior mediastinum (Fig. 4-124) and may arise in the
the pulmonary artery shows abundant fat throughout the medi- thymus or contain thymic tissue (Table 4-31) (535).
astinum (arrows). Note that the fat is similar in density to subcuta-
neous fat, and is exceedingly homogeneous. This patient had no Histologic differentiation between a lipoma and well-
history of either steroid therapy or Cushing syndrome. differentiated liposarcoma depends on the presence of
A B
Figure 4-121 Mediastinal lipomatosis. A: Coned-down anteroposterior radiograph of the lower thoracic spine in a patient undergoing a
barium enema. The paraspinal interfaces are abnormally convex bilaterally (arrows), the appearance of which is nonspecific. B: CT section at
the level of the distal esophagus unequivocally confirms that the abnormality seen in A is caused by excessive fat in the posterior medi-
astinum (arrows). Note again the homogeneous nature of the fat, confirming this as lipomatosis.
mitotic activity, cellular atypia, fibrosis, neovasculariza- CT and MRI show a mass containing variable amounts
tion, and tumor infiltration. This tumor most typically of fat and soft tissue (Fig. 4-124). The more fat and the less
occurs in adults, with an average age at presentation of soft tissue the mass contains, the better differentiated it
43 years (535). Symptoms may include cough, chest pain, tends to be (536). Masses often appear large and sharply
dyspnea, or superior vena cava syndrome, but more than defined. However, invasive tumors may appear poorly
one third are detected incidentally on chest radiographs marginated or may be associated with poor definition of
(535). The majority of mediastinal liposarcomas are well adjacent mediastinal structures or obvious findings of in-
differentiated, and although more than 30% demonstrate filtration or invasion (537–542). Calcification of septa-
local recurrence within a few years of resection, distant tions and some enhancement may be seen (534). Despite
metastases are unusual (535). suggestive CT findings in some cases, a specific diagnosis
A B
Figure 4-122 Cardiophrenic lipomatosis. A: Contrast-enhanced CT image through the heart demonstrates a large accumulation of fat
adjacent to the right heart border (arrow). Note the similarity in appearance between paracardiac and subcutaneous fat. B: T1-weighted MR
image at the same level as shown in A. On T1-weighted sequences fat appears bright (compare with the signal intensity of adjacent sub-
cutaneous fat).
Figure 4-124 Mediastinal liposarcoma. An anterior mediastinal

mass is composed of both soft tissue and fat (arrow).
(544). Lipoblastoma arises from embryonal fat and is

Figure 4-123 Mediastinal lipoma. A focal rounded fatty mass usually diagnosed in children younger than 3 years (545).
(arrow) is visible in the anterior mediastinum. Most present as a painless soft tissue mass. Invasion or
involvement of adjacent structures including the chest wall
(Fig. 4-125) and spinal canal may be seen. Most are less
may be difficult to make. As documented by DeSantos et than 5 cm in diameter, but they may be quite large. CT and
al. (543), a preoperative CT diagnosis of liposarcomas was MRI show a mass consisting primarily of fat but contain-
made in only 4 (22%) of 17 cases. ing strands of soft tissue (Fig. 4-125) (546).
Lipoblastoma Hibernoma
Lipoblastoma is a benign but infiltrating tumor that may Hibernoma represents a neoplasm derived from brown
involve the extremities, mediastinum (less than 10% of adipose tissue (Table 4-31). It is usually seen in locations
cases), neck, diaphragm, and chest wall (Table 4-31) where normal brown fat is found in infants, such as the
A B
Figure 4-125 Lipoblastoma. An 8-month-old boy with a large mass discovered on chest radiograph. A, B: CT scans show a large mass
with low CT numbers compatible with a fatty tumor. The suggestion of chest wall invasion anteriorly in B, the patient’s age, and the minimal
inhomogeneity would be unusual for a simple lipoma. Surgery was therefore recommended. A lipoblastoma, a rare benign tumor of child-
hood made of immature adipose tissue with rapid growth and local recurrence potential, was removed.
periscapular or interscapular region, the neck, the axilla, or herniated omental fat and probably represent omental
within the thorax and mediastinum (547). It appears as a vessels. When seen within a fatty mass, these linear densi-
fatty mass, which can be large (548). Total excision is advo- ties should suggest fat herniation rather than a lipoma.
cated for cure, as there is no known malignant potential. Fat herniation through the foramen of Bochdalek occurs
Focal collections of brown fat may be present in normal most often on the left side because the liver limits the fre-
subjects and generally go unrecognized. However, because quency of its occurrence on the right (557,558). Although
brown fat is metabolically active, it may show uptake on such hernias are most often located in the posteromedial
FDG-PET, mimicking neoplasm (549–551). In a study by diaphragm, they can occur anywhere along the posterior
Truong et al. (549), 15 (1.8%) of 845 oncologic patients costodiaphragmatic margin. Bochdalek hernias in adults
having PET/CT scans showed focal hypermetabolic medi- usually contain retroperitoneal fat, although kidney can
astinal brown fat. Hypermetabolic mediastinal deposits of occasionally be present. CT has shown that small Bochdalek
fat (mean SUV 5.7) were more common in children (4/8) hernias may occur in as many as 5% of normal individuals.
than in adults (11/837) and more common in women Characteristically, a thinning or defect in the diaphragm is
(9/372) than in men (2/465). Foci of hypermetabolic visible on CT, marginating the collection of fat.
brown fat were localized to the paratracheal, paraesoph- Herniation of perigastric fat through the phreni-
ageal, prevascular, and pericardial regions; interatrial sep- coesophageal membrane surrounding and fixating the
tum; and azygoesophageal recess. Five patients had focal esophagus to the diaphragm is the first step in the patho-
hypermetabolic brown fat isolated to the mediastinum. Ten genesis of hiatus hernias (559). The herniated fat can extend
patients also had extramediastinal hypermetabolic brown along the aorta and widen the paraspinal line or it can
fat in the neck, thorax, and abdomen. There was no differ- appear as a retrocardiac mass (Fig. 4-119). Although multi-
ence in the body weight or body mass index of patients planar reconstructions are sometimes helpful for demon-
with hypermetabolic brown fat as compared with age- and strating the connections of the fatty hernia with abdominal
sex-matched control subjects. FDG uptake in supraclavicu- fat, because of its multiplanar imaging capabilities, MRI
lar fat is also common (552). may be better suited to the evaluation of such lesions (560).
Magnetic Resonance Evaluation of Fatty Lesions

Other Fatty Masses
The MR appearance of fatty tissue has been well described
Other rare, fatty lesions have been reported to involve the
(92,200,561,562). Typically, fat has high signal intensity on
mediastinum. Thymolipoma usually appears on CT as a
both T1- and T2-weighted sequences, appearing identical
large fatty mass, in the anterior mediastinum, containing
to subcutaneous fat (Figs. 4-55 and 4-56). Similar findings
wisps or strands of fibrous tissue (87) (Fig. 4-45). Fat is
have been described, however, in patients with hematoma
also commonly identifiable as a component of mediasti-
or hemorrhagic tumors. In an analysis of 17 patients with
nal germ-cell tumors (Figs. 4-51 to 4-53). In the posterior
18 lipomatous tumors, Dooms et al. (562) reported that of
mediastinum, spinal lipomas rarely may present as
16 benign lesions, including 12 lipomas, the fatty compo-
primary mediastinal masses (553). Other fat-containing
nents of the tumors were equally well visualized with both
masses that have been identified by CT in the posterior
CT and MRI. As reported by London et al. (561) in an eval-
mediastinum include angiolipomas (554,555) and fatty
uation of 15 patients with liposarcomas, MRI correctly
transformation of thoracic extramedullary hematopoiesis
identified the presence of fat in all 8 cases in which it was
following splenectomy (91). Angiolipoma is a very rare
present pathologically. Unfortunately, MRI has proven no
benign mediastinal tumor composed of mature fat and
more accurate than CT in differentiating liposarcomas from
blood vessels, which appears encapsulated on CT and con-
benign lipomas.
tains both fat and soft tissue; it can mimic the appearance
of liposarcoma (555).
Tumors and Masses of Vascular Origin
Mediastinal tumors of vascular origin account for 2% to
Hernias Containing Fat
5% of all mediastinal tumors and include a heterogeneous
There are several direct connections between the abdomen group of neoplasms of endothelial-cell composition
and mediastinum that permit passage of intra-abdominal arising from either blood vessels or lymphatic vessels,
fat into the thorax. most typically in the anterior or posterior mediastinum
Omental fat is freely mobile and can herniate through (Table 4-32) (81).
the foramen of Morgagni to create the appearance of a
cardiophrenic angle mass, almost always located on the
Lymphangioma and Cystic Hygroma
right side (Fig. 4-118) (90,556). The transverse colon may
accompany the omentum in patients with a Morgagni Lymphangioma is a rare benign lesion of lymphatic origin
hernia. Fine linear densities can sometimes be seen within and represents 0.7% to 4.5% of mediastinal tumors; most
TABLE 4-32
TUMORS AND MASSES OF VASCULAR ORIGIN
Lymphangioma and Cystic Hygroma May extend into the neck
Most present at birth and detected in the first 2 years of life CT findings
Most common in the neck (75%) and axilla (20%) Heterogeneous in attenuation on unenhanced scans
10% of cervical lymphangiomas extend into the mediastinum May contain fat
Less than 5% of lymphangiomas are limited to the mediastinum Heterogeneous enhancement is typical following contrast
In adults, mediastinal lymphangioma common infusion
Most common in the anterior and superior mediastinum Opacified vascular channels
Classified as capillary, cavernous, or cystic (hygroma) Phleboliths
Cystic most common
CT findings Angiosarcoma
Often homogeneous and near to water in attenuation Usually arise in the heart or pulmonary artery
May contain soft tissue, fat, enhancing vascular Primary mediastinal tumors rare
malformations Large mass, usually anterior mediastinum
May appear to envelop mediastinal structures CT findings
Some may appear as solid masses Heterogeneous attenuation due to hemorrhage, necrosis, cysts
Enhancement may be seen
Lymphangiomatosis
Diffuse proliferation of lymphatic vessels in multiple sites Epithelioid Hemangioendothelioma
Most common in children Low-grade malignancy derived from endothelial cells
CT findings Intermediate between hemangioma and angiosarcoma
Generalized mediastinal widening Local infiltration common; 20% metastasize
Increased attenuation of mediastinal fat (similar to water) CT findings
Interlobular septal thickening with lung involvement Large and may be smooth or lobulated
Pleural thickening or pleural effusion in nearly all Calcification may be seen
Enhancement may be seen
Hemangioma
Benign vascular tumor Hemangiopericytoma
Large interconnecting vascular channels Tumor composed of undifferentiated cells associated with
Varying amounts of fat and fibrous tissue prominent, thin-walled, “staghorn” vessels
75% present before the age of 35 years May metastasize
Most common in the anterior and posterior mediastinum CT: large, well-circumscribed, heterogeneously enhancing mass
lymphangiomas are present at birth and are detected in the posterior (16% to 28%) mediastinum. They can be quite
first 2 years of life (Table 4-32) (563,564). Lymphangiomas large, ranging up to 30 cm in diameter (563,566).
are most common in the neck (75%) and axilla (20%). The imaging appearance of a lymphangioma is related to
Although 10% of cervical lymphangiomas extend into the its type (i.e., cystic, cavernous, or capillary). On CT, cystic
mediastinum, less than 5% of lymphangiomas are limited lymphangioma usually appears homogeneous and near to
to the mediastinum itself (Fig. 4-126) (94,564). Patients
may be asymptomatic because the mass is soft and yield-
ing, but compression of mediastinal structures can result in
chest pain, cough, or dyspnea (94). Because of a tendency
for local growth, surgery is recommended.
Lymphangioma can be seen in adults, with or without a
history of incomplete resection as a child. In adults, it is
common for lymphangioma to be localized to the medi-
astinum. Lymphangioma may be associated with Gorham
disease, Klippel-Trenaunay syndrome, Servelle-Noques
disease, and lymphangiomyomatosis (81).
Lymphangiomas are classified as capillary, cavernous,
or cystic (hygroma), depending on the size of the lymph
channels they contain (94). Cystic lymphangiomas are
most common, and were seen in 63% of adult patients in
one study (563); they may be either unilocular or multi-
locular and contain either serous or chylous fluid (565);
thin septations within the mass can sometimes be seen
(565). In adults, lymphangiomas are most common in the Figure 4-126 Low-attenuation mediastinal lymphangioma.
Contrast-enhanced CT scan shows a heterogeneous, low-density,
anterior (28% to 37%) or superior (16% to 36%) medi- right paratracheal mass. Lymphangioma limited to the medi-
astinum but are also seen in the middle (20% to 26%) and astinum is unusual.
water in attenuation (Figs. 4-126 to 4-129), but they can be ing vascular channels with regions of thrombosis and
of higher attenuation. Lymphangioma may be variably varying amounts of interposed stroma, such as fat and
composed of a combination of fluid, solid tissue, and fat; fibrous tissue. Tumors are categorized according to the
calcification is rare (563,565,567–569). Although they are size and nature of their vascular spaces as capillary,
usually well circumscribed and localized, they may appear cavernous, or venous; cavernous hemangiomas make up
to envelop mediastinal structures (Figs. 4-127 and 4-129) about 75% of cases (98). They are well defined and rarely
(565). Lymphangioma may be associated with vascular are invasive (98).
malformations that are easily identifiable following the Mediastinal hemangiomas are most common in young
administration of intravenous contrast (Fig. 4-129) patients, and about 75% present before the age of 35
(99,570,571). Lymphangiomas composed of capillary-sized, (98,363). One third to one half of cases are asymptomatic,
thin-walled channels may appear as solid masses (563). but some patients present with symptoms of compression
On MRI, heterogeneous signal is typical, with increased of mediastinal structures. Occasional cases are associated
signal on T2-weighted images reflecting their fluid content with peripheral hemangiomas or Osler-Weber-Rendu
(Figs. 4-128 and 4-129). MR may also show an enhancing, syndrome (98).
multicystic, and multiseptated appearance; this suggests Hemangiomas most commonly arise in the anterior
the presence of cavernous lymphangioma (94,568). and posterior mediastinum and may extend into the neck
(Fig. 4-131). Of 14 cases studied by McAdams et al. (98),
6 (43%) were posterior, 5 (36%) were anterior, 1 was
Lymphangiomatosis middle mediastinal, and 2 were multicompartmental. In a
review of 77 cases (363), 68% were primarily located in
This rare condition is characterized by the diffuse prolifer-
the anterior mediastinum, whereas 22% were posterior,
ation of lymphatic vessels in multiple sites (Table 4-32)
and in 7%, extension into the neck was present. Masses
(572). It is most common in children. The mediastinum
may appear well marginated on CT (71%), inseparable
may be involved, as may the lung and pleura. Chylothorax
from adjacent mediastinal structures (21%) or diffusely
may result. Generalized mediastinal widening is typical on
infiltrative (7%) (98). Apparent infiltration may or may
CT, with increased attenuation of mediastinal fat (similar
not predict unresectability.
to that of water) (Fig. 4-130) (572). A focal mass is not
On CT, tumors are often heterogeneous in attenuation on
visible. Pulmonary involvement may be associated with
unenhanced scans, and fat may occasionally be seen within
interlobular septal thickening. Pleural thickening or pleu-
them. Heterogeneous enhancement is typical following
ral effusion is seen in nearly all patients.
contrast infusion (Fig. 4-131) but is not always present
(98,574,575). Enhancement may be dense, multifocal or
diffuse, and central or peripheral. Central enhancement
Hemangioma
is thought to be suggestive of this tumor. Opacified
Hemangiomas are rare benign vascular tumors, account- vascular channels can be seen within the mass, with rapid
ing for less than 0.5% of mediastinal masses (Table 4-32) enhancement similar to that of normal mediastinal vessels.
(98,573,574). They are composed of large interconnect- Phleboliths, thought to be pathognomonic, are visible in up
A B
Figure 4-127 Cystic hygroma. A, B: Sections through the lower neck and superior mediastinum, respectively, show a low-attenuation mass
(arrows) visible in the neck (A) and upper mediastinum (B). In B, the mass appears to envelop the left carotid artery.
B C
D E
Figure 4-128 Cystic hygroma. A: Contrast-enhanced CT section through the lower neck shows a well-defined cystic mass lying behind
the carotid sheath; adjacent to the trachea, esophagus, and cervical spine; and extending laterally to lie behind the sternocleidomastoid
muscle (arrows). This lesion also was visualized with CT extending inferiorly into the thoracic inlet and mediastinum (not shown). B–E: T2-
weighted MR images through the lower neck and superior mediastinum in the same patient confirm the presence of an elongated cystic
mass extending from the neck into the posterior mediastinum (arrows). Surgery confirmed cystic hygroma.
A B
C D
Figure 4-129 Mediastinal lymphangioma in a 13-year-old boy. A: Contrast-enhanced CT shows a large low-attenuation anterior mediasti-
nal mass, containing enhancing vascular structures (arrows). B: T1-weighted MRI shows a complex mass, containing a fluid-fluid level (arrow).
C: T2-weighted image shows a large cystic component. D: T1-weighted coronal image shows a complex mass. High-intensity regions likely
reflect hemorrhagic cystic areas.
Figure 4-130 Mediastinal lymphangiomatosis in a 13-year-old

boy. Generalized mediastinal widening is visible, with increased Figure 4-131 Mediastinal hemangioma in a 3-month-old child
attenuation of mediastinal fat. This appearance, along with pleural with stridor. A large enhancing mass (arrows) is visible surrounding
thickening, is typical. the trachea. The mass resulted in tracheal obstruction.
to 10% of cases on plain radiographs (363). On CT, punctate Hemangiopericytoma

calcifications are visible in 21% and phleboliths are visible
in 7% (98). Hemangiopericytoma is a tumor composed of spindle
shaped or oval undifferentiated cells that proliferate around,
and are intimately associated with, prominent, thin-walled,
Angiosarcoma often branching (“staghorn”) vessels (Table 4-32) (81). It
occurs in adults and is sometimes associated with hypo-
Angiosarcoma represents the most commonly primary car- glycemia. Hemangiopericytoma can metastasize even
diac sarcoma, and the majority of primary right atrial when histologically benign (15% to 20%), most often
masses are angiosarcomas (81,576). Angiosarcoma may within 5 years after initial excision. Treatment of choice is
also arise within central pulmonary arteries (i.e., pul- surgery; radiation therapy and chemotherapy have been
monary artery sarcoma). Primary angiosarcoma may also shown to be ineffective. It may appear on CT as a large, well-
occur in the mediastinum without an obvious vascular circumscribed, enhancing-enhancing mass (575). It may
origin, but such primary mediastinal tumors are rare have a predominance in the superior mediastinum
(Table 4-32). They have been associated with radiation (575,582). It has been reported in association with sponta-
therapy and with occupational exposure to vinyl chloride neous hemothorax (582).
and arsenic (353,577).
Angiosarcoma typically occurs in middle-aged adults.
Symptoms associated with the tumor depend on its loca- Tumors and Masses of Fibroblastic Origin
tion. Primary mediastinal tumors result in symptoms as
a result of local mass effect. Cardiac tumors are manifested Mediastinal Fibromatosis
with symptoms of congestive heart failure or arrhythmia. Mediastinal fibromatosis is a rare entity that affects patients
Pulmonary artery sarcoma may be asymptomatic or pre- of all ages. It consists of dense and unencapsulated
sent with hemoptysis or other clinical findings similar to collagenous tissue and highly differentiated fibroblasts
those of pulmonary embolism. (Table 4-33). Despite its benign nature, it infiltrates
Extracardiac mediastinal angiosarcoma typically pre- surrounding tissues and may surround or compress medi-
sents as a large mass, usually in the anterior mediastinum. astinal structures, such as the aorta, trachea, esophagus, or
At CT, they most often appear heterogeneous in attenua- heart (583–587). Treatment using surgery may be difficult
tion due to areas of hemorrhage, necrosis, and cyst and is directed at relieving the compression or obstruction
formation. Cardiac angiosarcomas appear as diffuse wall of vital mediastinal structures. Although fibromatosis tends
thickening or focal masses that typically involve the right to recur locally, metastases do not occur. Differential diag-
atrium and enhance heterogeneously after intravenous nosis includes fibrosing (sclerosing) mediastinitis and well-
administration of contrast material. differentiated fibrosarcoma. Mediastinal fibromatosis is
The prognosis is poor, with few patients surviving also termed aggressive fibromatosis and desmoid tumor.
beyond 3 years. When possible, treatment includes resec- Imaging features are nonspecific, and biopsy is neces-
tion and radiation therapy. Chemotherapy has been used sary for diagnosis. Fibromatosis usually appears on CT as
to decrease the size of the tumor prior to surgery to facili- a large, slowly growing, infiltrative, soft tissue mass. It is
tate resection, as well as in the treatment of metastatic most frequent in the middle or posterior mediastinum but
disease (353,577,578). may occur in any location (583,585–588). Patients with
esophageal involvement may present with dysphagia; a
barium swallow may show a regular or eccentric stenosis
Epithelioid Hemangioendothelioma
without dilatation of the proximal esophagus (585). In
Epithelioid hemangioendothelioma is thought to be a such patients, CT typically demonstrates a periesophageal
low-grade malignancy derived from endothelial cells; its mass and allows evaluation of the extent of the lesion. In
histologic features are intermediate between those of some patients, invasion of the chest wall, spinal canal, or
hemangioma and angiosarcoma (Table 4-32) (81). The abdomen is present (586,588).
mediastinum is an uncommon site of origin (579–581). On T1-weighted MRI, the mass may appear isointense
The majority of these tumors have a benign histologic or slightly hyperintense relative to muscle. On fast SE
appearance, but local infiltration is common and about T2-weighted images, it appears hyperintense relative to
20% metastasize (579). Treatment is by complete resection. muscle. Enhancement occurs after intravenous administra-
Recurrence and metastases may benefit from radiation and tion of gadolinium (586,588).
chemotherapy. These tumors tend to be large and may be
smooth or lobulated. Calcification may be seen. In one
Solitary Fibrous Tumor
case, CT showed a well-defined low-attenuation mass with
peripheral nodular enhancement similar to that seen with Solitary fibrous tumor of the mediastinum is similar to
hepatic hemangiomas (580). solitary or localized fibrous tumor of the pleura but is
Inflammatory Myofibroblastic Tumor

TABLE 4-33
The etiology of inflammatory myofibroblastic tumor
TUMORS AND MASSES OF FIBROBLASTIC
(IMT) is controversial. Although this lesion has been
ORIGIN
regarded as reactive or postinflammatory in nature, recent
Mediastinal Fibromatosis evidence favors its being a neoplasm. This evidence
Benign; consists of dense unencapsulated collagenous tissue includes the presence of clonal chromosomal abnormali-
and highly differentiated fibroblasts
ties and the frequent expression of anaplastic lymphoma
Also termed aggressive fibromatosis and desmoid tumor
Infiltrates surrounding tissues kinase (ALK), a marker also found in large cell lymphomas
May surround and compress mediastinal structures (592–594). IMT has also been described as inflammatory
Differential diagnosis includes fibrosing (sclerosing) mediastinitis pseudotumor, plasma cell granuloma, histiocytoma,
and well-differentiated fibrosarcoma xanthoma, cellular inflammatory pseudotumor, inflam-
CT: large, slowly growing, infiltrative, soft tissue mass
matory fibrosarcoma, and various combinations of these
Most frequent in the middle or posterior mediastinum
Esophageal compression common terms (595).
IMT most frequently arises in the lung; its incidence is
Solitary Fibrous Tumor
reported to be 0.04% to 1% of all lung tumors (592,595).
Similar to localized fibrous tumor of the pleura
Often anterior mediastinum Less often, this tumor originates in the mediastinum or
involves the mediastinum by direct extension from a
Inflammatory Myofibroblastic Tumor
Likely a neoplasm rather than being inflammatory pulmonary site of origin (Table 4-33). It is frequently
Composed of spindle cells with a variable inflammatory diagnosed in children or young adults but occurs in older
infiltrate adults as well (592).
May arise in mediastinum or adjacent lung Pathologically, IMT is composed of spindle cells
Slow growing but locally invasive
(myofibroblasts and fibroblasts) with a variable inflam-
CT findings
Sharply marginated, lobulated mass matory infiltrate, including plasma cells, histiocytes, and
May be large lymphocytes. IMT usually grows slowly. Local invasion
May surround mediastinal structures may occur, but metastases are rare. Surgical resection is
Homogeneous or heterogeneous in attenuation recommended, and prognosis after complete resection
May enhance
is excellent. However, local recurrence may result if resec-
Fibrosarcoma tion is incomplete.
Rare CT generally shows a sharply marginated, lobulated
Usually anterior mediastinum
mass, ranging up to 13 cm or more in diameter (Fig. 4-132)
Malignant Fibrous Histiocytoma (592–595). The mass may be seen to surround medi-
Mediastinal origin rare
astinal structures and has been reported to involve all
CT: localized mass with heterogeneous attenuation and
enhancement mediastinal compartments (593,594); superior vena cava
obstruction has been reported in association with a large
superior mediastinal mass (593). IMT may appear homo-
geneous or heterogeneous in attenuation on CT and may or
may not enhance with contrast infusion. Calcification may
much less common (81,589). It is characterized histologi- be seen on CT. T1-weighted MR in a few cases has shown
cally by varying proportions of fibroblastlike cells and heterogeneous signal with an intensity slightly greater than
connective tissue (Table 4-33) (590). As with solitary that of muscle. T2-weighed imaging showed high signal
fibrous tumor of the pleura, it may be associated with intensity.
hypoglycemia (81).
Witkin et al. (591) reported 14 cases of mediastinal
Fibrosarcoma
solitary fibrous tumor. The lesions presented with cough,
chest pain, and dyspnea or as asymptomatic masses Fibrosarcoma usually arises in the soft tissues of the chest
detected radiographically. A localized well-circumscribed wall but may also occur in the mediastinum, heart, and
mass is typically seen. Two patients had associated hypo- lungs and as an endobronchial mass in the main or
glycemia. Eleven of the 14 tumors were located in lobar bronchi, in children and young adults (Table 4-33).
the anterior mediastinum. One arose on a pedicle from The left atrium is the most common location of cardiac
the thymus, and another had entrapped thymic ele- fibrosarcoma.
ments. Features associated with a poor prognosis in- Mediastinal fibrosarcoma is rare. It usually occurs in the
cluded large size, marked cellularity, mitotic activity, and anterior mediastinum (81,353,596). It is locally invasive
anaplasia (591). Surgical resection is curative for benign and may recur following resection but uncommonly
lesions, but aggressive lesions may recur locally or metastasizes. Despite surgical resection, radiation therapy,
metastasize. and chemotherapy, its prognosis remains poor. The differ-
C D
Figure 4-132 Inflammatory myofibroblastic tumor in a child. A: A complex mass involves the left mediastinum and adjacent lung (arrow).
A left chest tube is in place for pleural fluid drainage. B: T1-weighted MRI shows the mass to be homogeneous, but containing an area of
high intensity likely due to hemorrhage. C: Coronal MRI has a similar appearance. D: Coronal positron emission tomography (PET) shows the
mass to be highly active (arrow).
ential diagnosis of low-grade tumors includes mediastinal and radiation therapy for local control, supplemented by
fibromatosis or solitary fibrous tumor. chemotherapy for prevention of recurrence and treatment
of metastatic disease. However, the prognosis is poor, with
a high incidence of local recurrence and distant metastasis.
Malignant Fibrous Histiocytoma
Calcification is uncommon prior to treatment.
Malignant fibrous histiocytoma is the most common soft
tissue sarcoma in adults, although a thoracic origin is
Tumors of Muscle Origin
infrequent, and most cases arise in relation to the chest
wall musculature. Mediastinal malignant fibrous histiocy-
Leiomyoma
toma is extraordinarily rare (Table 4-33) (81,597). CT
usually shows a localized mass with heterogeneous attenu- Leiomyoma is among the least common mesenchymal
ation and enhancement. Treatment is usually resection tumors of the mediastinum (81). It develops most
frequently in middle-aged women, and arises from struc- Although the frequency of metastases is not related to
tures containing smooth muscle, including small vessels tumor size, the prognosis of leiomyosarcoma is poorer
(Table 4-34). The posterior mediastinum is the most com- with larger tumors because of a higher frequency of local
mon location. They are slow-growing, well-capsulated invasion and recurrence. Pulmonary artery lesions have a
tumors that are cured surgically. Their CT appearance is particularly poor prognosis, with few patients alive at 6 to
nonspecific, although they generally appear well defined 12 months. Treatment consists, when possible, of local
and homogeneous in attenuation. excision, radiation, and chemotherapy.
Leiomyosarcoma Rhabdomyosarcoma
Leiomyosarcomas are rare malignant tumors of smooth Rhabdomyosarcoma is a rare thoracic neoplasm that may
muscle origin that may arise from (a) small vessels within arise in the lung, bronchi, mediastinum, heart, and chest
the mediastinum, (b) heterotopic smooth muscle cells wall (81,353,596). They most typically present in childhood
derived from displaced splanchnic mesoderm, and (c) the or in the fifth to seventh decades of life and predominate
esophagus (81). Most are located in the anterior or poste- in males (Table 4-34). Less than 2% of childhood
rior mediastinum (Table 4-34) (598). rhabdomyosarcomas occur in the mediastinum, but rhab-
Primary leiomyosarcoma occurs in the mediastinum, domyosarcoma is the most common cardiac sarcoma in
heart, and lung. Although it is reported to arise from medi- children. In adults, rhabdomyosarcoma most often involves
astinal structures such as the esophagus or mediastinal the chest wall or diaphragm, but it may also arise in relation
vasculature, including the pulmonary artery and superior to the thymus or in association with a teratoma.
vena cava, most mediastinal tumors do not involve these Rhabdomyosarcomas appear as large masses of variable
structures (353). attenuation due to necrosis or cystic regions within the
Leiomyosarcoma usually occurs in the sixth decade or tumor. Invasion of adjacent structures may occur. MRI sig-
later. There tends to be predominance in men. However, nal intensity is variable; the tumors may be isointense to
pulmonary artery leiomyosarcomas occur with an equal muscle or have heterogeneous signal intensity because of
frequency in men and women, usually a decade or so necrosis or cystic regions (353).
younger (mean age at diagnosis of 50 years) (353). Rhabdomyosarcomas are usually treated with chemother-
Leiomyosarcomas located in the mediastinum typically apy; radiation therapy and surgery are used in some patients,
present with symptoms of mass (e.g., pain, cough, superior depending on tumor location and the extent of local
vena cava syndrome). invasion.
Mediastinal leiomyosarcomas are usually large neo-
plasms and often appear heterogeneous in appearance due
to necrosis or hemorrhage (353). Because leiomyosarcomas Tumors of Bone and Cartilage
of the pulmonary artery frequently grow within the lumen, Osteosarcoma, chondroma, chondrosarcoma, and chor-
they can be difficult to distinguish from pulmonary artery doma (Fig. 4-133) usually arise from the thoracic skeleton.
thrombi. MRI can often be useful in establishing the diagno-
sis because leiomyosarcomas, unlike thrombi, enhance after
intravenous administration of gadolinium contrast material.
TABLE 4-34
TUMORS OF MUSCLE ORIGIN
Leiomyoma
Most often occurs in middle-aged women
Posterior mediastinum most common
Slow growing
CT: well-defined mass, homogeneous attenuation
Leiomyosarcoma
Predominance in men
Anterior or posterior mediastinum
CT: large, heterogeneous mass
Figure 4-133 Mediastinal chordoma. A section through the
Rhabdomyosarcoma superior mediastinum following the bolus injection of contrast.
Predominate in males A bulky mediastinal mass is of homogeneous low attenuation, but
Present in childhood or in the fifth to seventh decades discrete vessels are visible traversing the mass. At surgery, biopsy
CT: large masses of variable attenuation due to necrosis or cysts led to considerable hemorrhage requiring transfusion of several
units of blood. Surgically proved mediastinal chordoma.
Primary mediastinal extraosseous occurrences may be seen mediastinum, the esophagus is in intimate contact anteri-
but are very rare (353,599). Usually located in the anterior orly with the posterior or posterolateral trachea, the left
or posterior mediastinum, their origin may be attributed mainstem bronchus, and the left atrium. The esophagus
to metaplasia of the connective tissues or malignant trans- lies between the aorta on the left and the azygos vein on
formation of embryonic remnants (81). the right. Intraluminal air or a small amount of fluid is
Suster et al. (599) reviewed six cases of primary common and a normal finding with CT.
mediastinal chondrosarcoma. The patients ranged in age Evaluation of esophageal disease is limited if the esoph-
from 11 to 63 years. In all cases, the lesions presented agus is incompletely distended (601). In those cases for
as well-circumscribed masses centered in the soft tissues in which assessment of the esophagus is the primary indica-
the posterior mediastinum without radiographic evidence tion for the study, or in patients with suspected esophageal
of origin from bone. They were small, well circumscribed, perforation, 500 mL of oral contrast material may be
and showed focal areas of calcification, mimicking the administered approximately 30 minutes prior to the exam-
appearance of a neurogenic tumor. These tumors tend to ination to distend the stomach and proximal small bowel,
be locally aggressive, with high recurrence rate after resec- with another 250 mL given just prior to placing the patient
tion, and may have distant metastases (599). Chordomas supine. With the patient in the scanner, a final swallow of
may also arise in the posterior mediastinum, appearing as oral contrast may then be timed to coincide with helical
relatively well circumscribed, encapsulated soft tissue acquisition. This has the advantage of ensuring that the
masses not clearly related to the thoracic spine (600). entire length of the esophagus is opacified during scan
acquisition (Fig. 4-134).
Mazzeo et al. (19) recently assessed the diagnostic
ESOPHAGUS utility of CT virtual endoscopy in various esophageal
diseases, including leiomyoma, squamous cell and
When sufficient mediastinal fat is present, the esophagus adenocarcinoma, esophageal infiltration by thyroid
is easily visualized with CT. At successive levels in the cancer, benign polyposis, chronic esophagitis, and
A B
C D
Figure 4-134 Esophagogastrectomy: CT evaluation. A–H: Sequential CT sections through the thorax in a patient with status postesoph-
agogastrectomy. During administration of oral contrast, 7-mm sections were obtained with a pitch of 2. There is uniform distention of the
intrathoracic portion of the stomach, allowing precise evaluation of the anastomosis in A. Note the incidental presence of a small pericardial
effusion in H.
E F
G H
postsclerotherapy stenosis. Patients had multidetector CT intrinsic or extrinsic masses, and (c) to evaluate patients
following esophageal distention by means of an efferves- with suspected esophageal perforation and to assess the
cent powder administered after induction of pharmaco- extent of associated pleural and mediastinal fluid collec-
logic esophageal hypotonia, and data were postprocessed tions (602). CT is less important in the evaluation of most
with 2D and 3D software reconstruction tools. This benign esophageal diseases, including benign strictures,
technique allowed virtual endoluminal visualization, esophagitis, and disorders of esophageal motility. CT also
accurate longitudinal and axial evaluations of disease may be employed in patients with a mediastinal mass
extent, and simultaneous evaluation of T and N parame- detected on chest radiography, resulting from an unsus-
ters in patients with neoplasm. Esophageal distention in pected esophageal abnormality.
the upper and middle thirds was classified as “good” in
32 of 33 cases (97%); in the lower third, esophageal
Esophageal Carcinoma
distention was “good” in 21 of 33 cases (64%). When
good distention was achieved, the thickness of normal Esophageal carcinoma represents approximately 10%
esophageal wall was less than 3 mm (range, 1.5 to of all cancers of the gastrointestinal tract. Most
2.4 mm; mean, 1.9 mm). Pathologic wall thickening was esophageal carcinomas are of squamous cell origin, al-
observed in 25 of 33 cases (76%), with values ranging though the incidence of distal esophageal adenocarci-
between 3.6 and 36 mm (mean, 9.6 mm). The sensitivity noma has increased dramatically, now representing
of this technique was 84%, its specificity was 87%, and nearly 20% of cases of primary esophageal carcinoma
diagnostic accuracy was 85%. (603). Esophageal carcinoma usually presents in an
In our experience, CT has several distinct indications in advanced stage, with 5-year survival rates varying between
patients with suspected esophageal disease. It is indicated 3% and 20% (604). This poor prognosis results from
(a) to evaluate and stage patients with esophageal carci- rapid submucosal extension of the tumor and early trans-
noma and to provide a means for assessing response to mural invasion; invasion of the mediastinum is common
therapy and resultant complications, (b) to evaluate and and is facilitated by lack of an esophageal serosa. Tumor
characterize esophageal contour abnormalities detected at may involve the trachea, bronchi, or aorta by direct
esophagography or endoscopy and their relationship to extension, and early spread to regional lymph nodes
is often present. Distant metastases may involve cervical

or abdominal lymph nodes or the liver, adrenal glands, TABLE 4-35
and lungs. AMERICAN JOINT COMMITTEE ON CANCER
Esophageal carcinoma is staged using a TNM classi- (AJCC) STAGING OF ESOPHAGEAL CARCINOMA
fication system devised by the AJCC. This system deter-
Primary tumor (T)
mines stage based on the extent of primary tumor invasion
TX: Primary tumor cannot be assessed
and the presence or absence of nodal involvement or T0: No evidence of primary tumor
distal metastases (Table 4-35) (605). Several authors Tis: Carcinoma in situ
(606–609) have proposed modifications of the AJCC T1: Tumor invades lamina propria or submucosa
staging system to reflect CT findings such as the esopha- T2: Tumor invades muscularis propria
T3: Tumor invades adventitia
geal wall thickness. However, these are not in common
T4: Tumor invades adjacent structures
clinical use.
The CT manifestations of esophageal carcinoma Regional lymph nodes (N)
NX: Regional lymph nodes cannot be assessed
include (a) thickening of the esophageal wall (usually 5 N0: No regional lymph node metastasis
mm), either symmetric or asymmetric; (b) an intraluminal N1: Regional lymph node metastasis
mass; (c) narrowing of the esophageal lumen; (d) dilata-
Distant metastasis (M)
tion of the esophagus above an obstructing tumor; (e) loss MX: Distant metastasis cannot be assessed
of periesophageal fat planes, with or without evidence M0: No distant metastasis
of invasion of surrounding structures or organs; and M1: Distant metastasis
(f) periesophageal, mediastinal, upper abdominal, or Tumors of the lower thoracic esophagus:
M1a: Metastasis in celiac lymph nodes
cervical lymph node enlargement (Figs. 4-135 to 4-138)
M1b: Other distant metastasis
(603,610–613). Tumors of the midthoracic esophagus:
Wall thickening is an early finding of esophageal carci- M1a: Not applicable
noma but is nonspecific. A study of 200 consecutive M1b: Nonregional lymph nodes and/or other distant
CT examinations revealed thickened esophageal walls metastasis
Tumors of the upper thoracic esophagus:
(3 mm) in 35% of patients. In only half of the cases
M1a: Metastasis in cervical nodes
were the thickened walls due to esophageal carcinoma. M1b: Other distant metastasis
Other mediastinal malignancies, as well as benign inflam-
AJCC stage groupings
matory, vascular, and fibrotic diseases such as reflux and
esophagitis, esophageal varices, and postirradiation scar- Stage 0
Tis, N0, M0
ring, were found to cause thickening of the esophageal
wall (607). Stage I
Wall thickening in esophageal carcinoma is most often T1, N0, M0
asymmetrical and irregular in contour. More than 2 cm of Stage IIA
the esophageal wall is usually involved at presentation. T2, N0, M0
T3, N0, M0
On CT, the proximal extent of the tumor is often
well shown because of esophageal dilatation above the Stage IIB
area of narrowing. The distal extent of tumor is less well T1, N1, M0
T2, N1, M0
shown, and CT may underestimate the tumor length
(606,614). Stage III
T3, N1, M0
T4, any N, M0
Assessment and Staging of Esophageal Stage IV
Carcinoma Any T, any N, M1
The depth of esophageal wall invasion is important in Stage IVA

Any T, any N, M1a
the assessment of esophageal carcinoma, and is used to
determine the T classification in the AJCC staging system Stage IVB
Any T, any N, M1b
(Table 4-35). CT is unable to accurately delineate layers of
the esophageal wall (i.e., lamina propria, submucosa, From Esophagus. In: American Joint Committee on Cancer. AJCC
muscularis propria, and adventitia) and is thus inaccurate cancer staging manual, 6th ed. New York: Springer; 2002:91–98, with
permission.
in distinguishing T1 to T3 tumors and in the diagnosis
of early or microscopic mediastinal fat invasion (609,
615–618). Endoscopic sonography is more accurate than endoscopic sonography and CT in the diagnosis of the
CT in the determination of the T1 to T3 classifications, depth of esophageal infiltration by tumor, the accuracy
unless esophageal narrowing precludes the use of this of sonography has been reported to be 84% to 92%, as
technique (619). In studies comparing the accuracy of compared with 42% to 68% for CT (609,613,618,620).
A B
C D
E F
G H
Figure 4-135 Squamous cell carcinoma of the midesophagus. A–H: Enlargements of sequential contrast-enhanced CT sections through
the mediastinum from above-downward following the oral administration of 2 tablespoons of 3% barium paste. An irregular mass is present,
most easily identified as irregular thickening of the esophageal wall posterior to the distal trachea and left mainstem bronchus (arrows in G
and H) extending at least 5 cm in length. Note that at the level of the tumor, the esophageal lumen is markedly irregular; superiorly, the
esophagus is dilated because of partial obstruction (curved arrow in A). In addition, mediastinal adenopathy is apparent superiorly (straight
arrows in A, B, and C). Note that the distal trachea and left mainstem bronchi are bowed forward. Although this appearance suggests direct
extension of tumor into the airways, in our experience this finding is nonspecific. By CT criteria, this is at least a stage 3 lesion.
A B
C D
Figure 4-136 Adenocarcinoma of the gastroesophageal junction. A–D: Enlargements of sequential CT sections through the esophageal
hiatus with the patient in a left lateral decubitus position to facilitate getting air into the distal esophagus. There is marked, irregular thick-
ening of the distal esophageal wall (arrows in B and C). The immediate surrounding fat is intact. At surgery this proved to be a gastric
adenocarcinoma with extension into the distal esophagus.
MRI has an accuracy comparable to that of CT (60%) in poorly marginated soft tissue mass. Usually, periesoph-
determining the T classification (618). ageal invasion is easily identified in patients with bulky
CT is more accurate in the diagnosis of macroscopic tumors (Figs. 4-136 to 4-139). Although mediastinal fat
invasion of mediastinal fat. In this instance, CT usually invasion affects prognosis, it does not preclude en bloc
shows obliteration of periesophageal fat planes or a resection.
Figure 4-138 Esophageal carcinoma: mediastinal extension.

Contrast-enhanced CT section through the carina shows a bulky
posterior mediastinal soft tissue mass confirmed as esophageal
Figure 4-137 Esophageal carcinoma: mediastinal extension. carcinoma at biopsy. Note that posteriorly to the left the mass
Section at the level of the carina shows a posterior mediastinal mass extends to contact most of the medial wall of the aorta (arrows).
with poorly defined borders and indistinct foci of air. Note that the Evaluation of aortic invasion is limited by CT, as even this degree
mass extends to the posterior aspect of the left main pulmonary of contact does not necessarily preclude resection. In this case,
artery. Biopsy proved esophageal carcinoma with perforation and palliative therapy only was administered because of documented
mediastinal invasion. metastatic disease.
45 degrees and 90 degrees, findings are considered indeter-

minate (625). Some investigators have chosen to regard
cases with less than 90 degrees of contact to be negative,
to reduce the number of indeterminate studies (619,626).
Although some have reported that these criteria are highly
accurate in predicting aortic invasion (625), results are
quite variable, with sensitivity values ranging from 6% to
100%, reported specificities of 52% to 96%, and accuracies
reported to be 55% to 96% (613). The obliteration of a
small triangular collection of fat separating the esophagus,
aorta, and spine has also been used to predict aortic
invasion (627).
Esophageal carcinoma most commonly metastasizes
to lymph nodes in the mediastinum, neck, or upper
Figure 4-139 Esophageal carcinoma: nodal metastases.
Contrast-enhanced CT section shows a bulky posterior mediastinal abdomen. In general, the lymphatics of the upper two
mass consistent with esophageal carcinoma. Note the presence of thirds of the esophagus drain cephalad; tumors in this
enlarged left paratracheal nodes. Although with tumors this large location often involve paraesophageal, subcarinal, para-
the likelihood of nodal metastases is great, enlarged nodes are
nonspecific and may reflect hyperplastic changes only. tracheal, or cervical lymph nodes. Lymphatics in the
lower third of the esophagus drain toward the abdomen,
and lower esophageal tumors may spread to paradi-
aphragmatic, gastrohepatic, and celiac lymph nodes (70).
The most important use of CT in the determination of Most studies consider paraesophageal and mediastinal
T is in the diagnosis of T4 tumors invading structures lymph nodes with a short axis diameter of more than
such as the trachea, bronchi, or aorta; this prevents resec- 1 cm to be abnormal, for the purposes of staging.
tion (613,615). If CT shows preservation of fat planes However, as with lung cancer, lymph node enlargement
adjacent to mediastinal structures, then resection is likely may occur as a result of hyperplasia and normal-sized
possible (615). Unfortunately, definitive assessment nodes may harbor metastases.
of mediastinal invasion may be difficult in cachectic In general, CT is considered to be of limited accuracy
patients in whom a paucity of mediastinal fat makes visu- in the diagnosis of regional lymph node metastases in
alization of the entire length of the esophagus difficult patients with esophageal carcinoma. In various studies, the
(607,621,622). sensitivity of CT for the diagnosis of regional mediastinal
CT has been reported to be an accurate means for lymph node metastases has varied considerably, ranging
detecting invasion of the carina and the mainstem from 34% to 77% with a specificity of 79% to 97% and an
bronchi, although the accuracy of CT is limited in subtle accuracy of 51% to 78% (613,618,619). Furthermore,
cases (Figs. 4-135 to 4-139) (608,612,623). CT findings paraesophageal lymph nodes may be difficult to distin-
that suggest the presence of tracheal or bronchial invasion guish from the primary tumor. The accuracy of CT in the
include an endoluminal mass, bronchial obstruction, diagnosis of abdominal lymph node metastases is some-
displacement of the airway by tumor, or posterior inden- what higher than for the diagnosis of mediastinal lymph
tation or thickening of the airway wall. However, such node metastases, but enlarged abdominal lymph nodes
findings are nonspecific and tumors that show these find- are also nonspecific (Fig. 4-139).
ings may still be resectable (624). Absence of a visible fat Additional techniques may be helpful in the diagnosis
plane between the tumor and airway wall cannot be used of regional lymph node metastases (628). MRI has been
to predict invasion, as the fat plane may be absent in shown to be comparable to CT in its accuracy (618,627).
normal persons. The sensitivity of CT in predicting the Endoscopic ultrasound has an accuracy similar that of CT.
presence of airway invasion varies in different series and In a study by Wu et al. (618), CT was more accurate in
ranges from 31% to 100%, with a specificity of 86% to staging regional lymph nodes (78%) than endoscopic
98% and an accuracy of 74% to 97% (613). sonography and MRI (71% and 64%, respectively), but the
Similarly, the CT diagnosis of aortic invasion may be difference was not statistically significant. The specificity
difficult (Fig. 4-138) (612,614,621,625). It has been and sensitivity were 79% and 77% for spiral CT, 75% and
suggested that the likelihood of aortic invasion may be 68% for sonography, and 68% and 62% for MRI, respec-
predicted by observing the extent of contact between tively (618). However, the sensitivity and accuracy of
the tumor and aorta. This is measured by using the arc sonography are improved if this technique is combined
through which tumor contacts the aortic circumference. with endoscopic needle aspiration of identified lymph
According to Picus (625), if this arc measures less than nodes (617,629,630). The use of FDG-PET has also been
45 degrees (of 360 degrees), then invasion is unlikely; on recommended for nodal staging and has a higher sensitiv-
the other hand, contact of more than 90 degrees is consid- ity and accuracy than CT (631,632). In a study by Choi
ered to predict invasion. If the extent of contact is between et al. (631), 48 patients with esophageal carcinoma under-
went esophagectomy and lymph node dissection follow- unusual postoperative findings have been reported, includ-
ing FDG-PET, CT, and endoscopic ultrasound. FDG-PET ing identifying recurrent tumor within thoracotomy
had a sensitivity of 57%, specificity of 97%, and accuracy incisions, as well as postoperative esophageal mucoceles
of 86% in detecting individual abnormal nodes, whereas (647,648).
CT had a sensitivity of 18% (p .0001), specificity of 99% The comparative utility of CT and MRI in the diagnosis
(p .033), and accuracy of 78% (p .003). For N staging, of tumor recurrence after resection of esophageal and
FDG-PET was correct in 83% of the patients, whereas CT gastroesophageal carcinomas was studied by Kantarci et al.
and ultrasound were correct in 60% (p .006) and 58% (649). Twenty-three patients who developed recurrent
(p .003), respectively. tumors after transthoracic esophagogastrectomy for esoph-
CT is considered advantageous in the diagnosis of dis- ageal carcinoma were analyzed retrospectively. Primary
tant metastases in patients with esophageal carcinoma, tumor recurrence was detected at the anastomosis side in
particularly those involving the liver, adrenal glands, and 19 patients (13 with intraluminal mass; 6 with diffuse or
lungs, or direct extension of tumor into the pleura, focal wall thickening). Distant recurrence was seen in the
lung, or adjacent vertebral bodies (613,618,633,634). liver (n 5), lung (n 4), bone (n 3), or abdominal
Laparoscopic ultrasound may also be useful for assessing lymph nodes (n 4) or manifested as pleural (n 2) or
distant metastases in the upper abdomen (617). FDG-PET pericardial (n 1) effusion. CT and MRI were found equal
is also useful in the diagnosis of distant metastases in showing intraluminal masses, liver metastasis, and pleu-
(615,616,635). ral and pericardial effusion. Thickening of the esophageal
The role of CT in preoperative staging of esophageal wall was seen in nine patients using CT, but in two of these,
carcinoma has proven controversial with excellent staging the thickening was due to fibrosis; MRI correctly identified
accuracy reported by some authors and others presenting the seven cases of tumor recurrence. MRI was falsely nega-
less favorable results (606,608,611,612,614,619,621–625, tive in two (50%) of four patients with lung metastases
636–641). Recent studies suggest that CT is most valuable detected using CT, and CT was falsely negative in one
in the exclusion of extensive mediastinal invasion, in the patient with bone metastasis detected using MRI (649).
detection of lymph node enlargement, and in the detec- Rarely, CT may detect unusual complications of eso-
tion of distant metastases (615,628). It has been further phageal carcinoma, including the presence of perforation
suggested that a combination of CT, endoscopic ultra- (Fig. 4-137) or fistula between the esophagus and the
sound, and FDG-PET provides the most comprehensive spinal canal (633,650). CT has proven of little value in
evaluation of tumor stage and resectability (613,615,616, patients with unusual esophageal tumors, although these
628,642,643). are occasionally encountered (Fig. 4-140).
The relationship between CT findings and prognosis in
esophageal carcinoma has also been investigated. In a
Differentiation of Intrinsic and Extrinsic
retrospective study of 89 patients, Halvorsen et al. (623)
found a significant correlation between decreased survival
Esophageal Lesions
and CT findings of tracheal, aortic, or pericardial invasion CT plays a major role in identifying extramural abnormali-
(623). Evidence of mediastinal invasion and enlarged ties that secondarily affect the esophagus. There are many
upper abdominal lymph nodes was ominous with mean causes of extrinsic esophageal compression. In the upper
survival in this group of only 90 days. These findings are mediastinum CT is especially helpful in diagnosing vascular
consistent with data linking various parameters to 5- and abnormalities. Detection of aortic aneurysms obviates aor-
10-year survival in patients with esophageal carcinoma in a tography. Aneurysmal dilatation of other arteries causing
large series of patients in Japan (644). compression of the esophagus also may occasionally be
identified. Rarely, these may rupture into the esophagus
with resultant exsanguination (Fig. 4-141) (22). In addition,
Postoperative Assessment of Esophageal
CT is useful in diagnosing a variety of vascular anomalies
Carcinoma that cause compression or displacement of the esophagus,
There is little dispute concerning the value of CT in postsur- including aberrant right and left subclavian arteries, double
gical follow-up (Fig. 4-134) (645,646). As documented by aortic arch anomalies, and pulmonary vascular slings (651,
Heiken et al. (645), CT may be especially useful in detecting 652). CT is also helpful in diagnosing substernal thyroid
early postoperative complications, including anastomotic glands, especially when they extend posterior to the esopha-
leaks with or without associated mediastinitis, parenchymal gus (Fig. 4-63) (249–251). Throughout the length of the
consolidation, empyema, and subphrenic abscesses. Equally mediastinum the esophagus may be displaced by enlarged
important, CT is an accurate means for detecting early lymph nodes. In select cases, CT may allow a presumptive
tumor recurrence. Gross et al. (646) have reported detecting diagnosis. In particular, secondary involvement of the
locally recurrent disease with CT in 7 of 21 patients follow- esophagus caused by low-density tuberculous lymph nodes
ing transhiatal esophagectomies, whereas corresponding with resultant esophagomediastinal fistulas has been
barium studies proved positive in only 4 cases. Other described (Fig. 4-142) (653,654). The esophagus may
Figure 4-140 Oat cell carcinoma of the esophagus. A: Coned-down view from an
esophagram shows a lobular mass with central ulceration (arrow). B: Section through the
distal esophagus shows a lobular soft tissue mass filling most of the esophageal lumen
outlined by oral contrast media (black arrows). Note the presence of contrast media within
the mass because of central ulceration (white arrow). Biopsy proved oat cell carcinoma.
(From Naidich DP. Esophagus. In: Megibow AJ, Balthazar EJ, eds. Computed tomography
A of the gastrointestinal tract. St. Louis: CV Mosby; 1986:33–98, with permission.)
become secondarily involved by paraesophageal neoplasms On CT, leiomyoma appears as a sharply marginated
that either obstruct the esophagus mechanically or invade mass, occurring in contact with the esophagus (Fig. 4-143).
the esophageal wall (655,656). It usually involves one esophageal wall but occasionally
may be circumferential. Often the bulk of the mass appears
extrinsic and is sharply outlined by mediastinal fat (Fig. 4-
Benign Esophageal Tumors
143). In a recent report (658), the radiographic findings of
Benign tumors of the esophagus can be classified as 12 patients with leiomyoma were reported. These included
mucosal or submucosal in origin. The most common muc- 10 men and two women, aged 34 to 47 years. The tumors
osal lesions include squamous papillomas, adenomas aris- ranged from 9 to 90 mm in diameter, and were located in
ing in Barrett mucosa, and inflammatory polyps; these are the upper (n 1), middle (n 5), or lower esophagus (n
not typically diagnosed using CT. Submucosal or intramural 6). In 10 of the 12 patients, the tumor was visible on chest
tumors include leiomyomas, leiomyomatosis, fibrovascular radiograph as a mediastinal mass. Esophagography showed
polyps, granular cell tumors, lipoma, and fibroma. Despite them as eccentric, smoothly elevated filling defects in 11 pa-
their infrequency, these lesions may be diagnosed using CT tients and a multilobulated encircling filling defect in one.
based on their characteristic appearances (657). They are CT scans revealed a smooth (n 9) or lobulated (n 2)
often asymptomatic unless large, but may result in a medi- tumor margin. On unenhanced scans, the tumors appeared
astinal abnormality recognizable on chest radiographs. homogeneous and similar to muscle in attenuation (n 5)
or were relatively low in attenuation (n 4); their mean
attenuation was 28 HU. Enhanced CT scans showed
Leiomyoma
homogeneous enhancement (mean 37 HU), with the
Leiomyoma is the most common benign esophageal lesion, tumor appearing low in attenuation relative to chest wall
representing about 50% of such cases, and is the most com- muscle (n 7) or similar in attenuation (n 4). T1- and
mon benign tumor to present as an asymptomatic mediasti- T2-weighted MRI may show a well-defined mass isointense
nal mass (603). This tumor is encapsulated and slow grow- or slightly hyperintense as compared with muscle (658).
ing; when large, dysphagia may result. Most occur in Leiomyosarcoma of the esophagus is rare. Dysphagia
patients 20 to 60 years of age, and there is a male predomi- is present in most patients. In a review of 10 cases (659),
nance (658). Approximately 60% arise in the distal esopha- chest radiographs revealed a mediastinal mass in half.
gus and 30% arise in the middle third. These tumors are On esophagram, the tumor may appear intramural, intra-
usually solitary and measure up to 8 cm in most patients. luminal, or infiltrative. Intramural tumors all had large
B, C
D, E
F, G
J H, I
Figure 4-141 Ruptured aortic aneurysm: CT evaluation. A: A coned-down view from an esophagram obtained in a patient who
presented with hematemesis. A large filling defect is conspicuous within the esophageal lumen. B–I: Sequential contrast-enhanced CT
sections from above-downward show a large aortic aneurysm involving the proximal portion of the descending aorta. At the level of the
distal trachea and carina, disruption in the wall of the aneurysm can be identified (arrows in F and G) with resultant mediastinal
hematoma and associated left pleural effusion. A blood-fluid level can also be seen within the esophageal lumen (curved arrow in H).
J: Corresponding aortogram showing a large aneurysm of the aorta without apparent leak. This patient died shortly thereafter despite
attempts at surgical repair.
Figure 4-142 Tuberculous esophageal fistula. A: CT section at the

level of the carina shows oral contrast within the esophageal lumen. Note
the presence of a small pocket of paraspinal air on the right. This finding
is suspicious for a possible esophageal fistula. B: Coned-down lateral view
from subsequent esophagram confirms the presence of an esophageal fis-
tula extending posteriorly (arrow). Patient was subsequently verified to
be sputum positive. B
exophytic components. One of the intraluminal lesions muscle on T1-weighted images and hyperintense on
appeared as a polypoid expansile mass and the other, as T2-weighted images (659).
a smooth expansile sausage-shaped mass mimicking a
fibrovascular polyp. CT typically shows a mass, which
Esophageal Leiomyomatosis
may appear homogeneous or heterogeneous in attenua-
tion, with a large exophytic component. Areas of low Esophageal leiomyomatosis (multiple leiomyomas) may
attenuation or extraluminal collections of gas or oral occur in children and in association with the Alport syn-
contrast material may be seen within the tumor. MRI drome. In a review of six cases (660), the average age was
revealed large masses that were isointense with skeletal 10.8 years (range, 6 to 18 years). Five patients presented
A B
Figure 4-143 Esophageal leiomyoma. A: An unenhanced CT shows a smoothly marginated, rounded mass of homogenous attenuation,
visible in the subcarinal space, and contiguous with the esophagus. The most common cause of this abnormality would be a bronchogenic
cyst. B: After contrast infusion, enhancement of the mass is visible. This is inconsistent with bronchogenic cyst, but typical of esophageal
leiomyoma, which commonly occurs in this location.
with typical symptoms of slowly progressive dysphagia. with fistula formation, perforation, mediastinitis, or medi-
Esophagrams revealed smooth, tapered narrowing of the astinal abscess. TB involving mediastinal lymph nodes may
distal esophagus and defects in the superomedial aspect secondarily affect the esophagus, with the development of
of the gastric fundus, presumably due to bulging of esophagomediastinal fistulas. CT shows enlarged mediastinal
the thickened mass of muscle into the stomach. In two lymph nodes and paraesophageal air collections (654).
patients, CT revealed marked thickening of the distal The CT findings in esophagitis have been reviewed by
esophageal wall (660). Berkovich in 29 patients (662). The cause was gastro-
esophageal reflux disease in six, radiation esophagitis in 2,
Candida esophagitis in 1, herpes esophagitis in 1, com-
Fibrovascular Polyps bined reflux and radiation esophagitis in 2, combined
Fibrovascular polyps of the esophagus are rare benign non- reflux and Candida esophagitis in 1, and unspecified
neoplastic intraluminal masses. In a review of 16 cases esophagitis in 12 (662). Patients were scanned using intra-
(661), all patients were symptomatic. Fourteen (87%) had venous contrast, but no esophageal opacification was
dysphagia and four (25%) had respiratory symptoms. The used. In this study, 16 (55%) patients showed esophageal
average duration of symptoms was 17 months, but seven wall thickening (5 mm) (Fig. 4-144). Although this find-
patients (44%) had symptoms for 6 or fewer months. Two ing was present in only 4% of 85 normal controls, the
patients (12%) had a history of regurgitating the tumor mean esophageal wall thickness was 4.7 mm (range, 2.3 to
into the pharynx or mouth, but none had the known 8.7 mm) in the 29 patients with esophagitis versus
complication of asphyxiation due to occlusion of the larynx. 2.9 mm (range, 1.6 to 5.0 mm) in the 85 controls
Chest radiographs revealed a right-sided superior mediasti- (p .001). If 3 mm is used as the threshold for esophageal
nal mass and/or anterior tracheal bowing in seven patients wall thickening, 23 (79%) of the 29 patients with
(44%). Esophagrams revealed a smooth but variably lobu- esophagitis had a thickened esophageal wall, as did 36
lated intraluminal mass that originated in the lower cervical (42%) of 85 controls, limiting the value of this finding. All
esophagus, with an average length of 15 cm. Depending on 16 patients with a thickened esophageal wall had concen-
the amount of fat and fibrovascular tissue in the lesion, CT tric circumferential wall thickening. Twelve of these
revealed a heterogeneous appearance in four patients, patients (75%) had wall thickening greater than 3.5 mm in
lesions of predominantly fat density in two, and lesions of length, and the entire thoracic esophagus was affected in
predominantly soft tissue density in two (661). half. Five patients (17%) with a thickened esophageal wall
also had evidence of a “target sign” on CT, with enhance-
ment of the mucosa and a relatively hypodense sub-
Esophagitis mucosa. None of the controls exhibited a target sign of
Esophageal inflammation or infection is termed esophagitis. mucosal enhancement and a hypodense submucosa (662).
Inflammation is typically related to gastroesophageal reflux.
Esophageal infection occurs most commonly in immuno- Esophageal Dilatation
suppressed patients with neoplasm, chemotherapy, organ
transplantation, or AIDS (Fig. 4-144). Organisms most com- Esophageal dilatation may be associated with esophagitis
monly involved include Candida albicans, herpes simplex and stricture, esophageal carcinoma or other tumors,
virus, cytomegalovirus, and TB (603,662). Circumferential fibrosing mediastinitis, scleroderma, achalasia, or leiomy-
esophageal wall thickening is typically present but mediasti- omatosis (603). Marked dilatation may result in an appar-
nal fat and tissue planes are preserved. In severe cases, ulcera- ent mediastinal mass on chest radiographs. In patients with
tion of the esophageal wall occurs, sometimes associated known esophageal dilatation, CT may be used in an attempt
to identify the cause (e.g., esophageal carcinoma) (603).
Esophageal dilatation in achalasia (Fig. 4-145) and sclero-
derma are usually associated with normal wall thickness
(663). An air-fluid level and retained food may be visible
in patients with achalasia (Fig. 4-145), stricture, or carci-
noma but is less common with scleroderma (Fig. 4-146).
Esophageal dilatation is present in as many as 80% of
patients with scleroderma and often asymptomatic (664).
Esophageal Varices
CT is able to detect both esophageal and paraesophageal
varices (Figs. 4-147 and 4-148) (665–667). Although endo-
scopy and esophagography are able to detect esophageal
Figure 4-144 Esophageal wall thickening with esophagitis.
Candida esophagitis in a patient with AIDS is associated with varices, paraesophageal varices previously required angiog-
esophageal wall thickening (arrows). raphy for definite diagnosis (665). Typically these appear as
C D
Figure 4-145 Esophageal dilatation with achalasia. A: Chest radiograph shows an abnormal convexity involving the right inferior medi-
astinum (white arrows). The right heart border (black arrows) is superimposed on this abnormality. B: CT through the upper mediastinum
shows a dilated esophagus (arrows) associated with an air-fluid-fluid level. C: CT through the lower esophagus shows a dilated fluid-filled
esophagus (arrows). D: Esophagram shows dilatation of the distal esophagus, typical of achalasia.
either nonspecific right- or left-sided mediastinal soft tissue following CT findings: (a) esophageal wall thickening
masses on chest radiographs, necessitating differentiation within which areas of low density could be identified;
from enlarged periesophageal lymph nodes or other poste- (b) obliteration of mediastinal fat planes, often associated
rior mediastinal masses. Using CT during the administra- with a focal fluid collection; (c) thickening of the
tion of a bolus of intravenous contrast, paraesophageal diaphragmatic crura; and (d) associated pleural effusions
varices are easily identified with CT (666). In fact, with a and subsegmental atelectasis. Similar findings have been
sufficient contrast, varices within the wall of the esophagus reported by Saks et al. (670).
itself can often be seen. Varices may be associated with
portal hypertension or venous obstruction with collateral
Hiatal Hernia
flow. So-called down-hill varices may result from superior
vena cava obstruction. Thorough familiarity with the normal cross-sectional
CT may be of value in patients following endoscopic appearance of the gastroesophageal region is a necessary
sclerotherapy (668,669). In an assessment of nine patients prerequisite for accurate identification of hiatal hernias
evaluated by CT following otherwise uncomplicated (Fig. 4-149). The esophageal hiatus is formed by the
esophageal sclerotherapy, Mauro et al. (669) noted the decussation of muscle fibers originating from the dia-
phragm around the lower esophagus. The esophagus is

fixed at the level of the hiatus by the phrenicoesophageal
ligament, which is not routinely visible on CT scans. The
esophageal hiatus is an elliptical opening just to the left of
the midline, corresponding superiorly to the level of the
tenth thoracic vertebral body. The margins of the hiatus are
formed by the arms of the diaphragmatic crura, which are
easily identified in cross section. Variation in the normal
appearance of the crura is common, especially nodular
thickening that is particularly common and may be mis-
taken for either abnormally enlarged lymph nodes or
rarely, crural invasion by adjacent tumor (671,672). As the
esophagus passes through the upper margin of the hiatus,
it assumes an oblique orientation, coursing in a posterior-
to-anterior and right-to-left direction. The gastroesopha-
geal junction itself lies just below the diaphragm. As it
courses through the upper abdomen, the distal esophagus
is enveloped by the most cranial portion of the gastrohe-
patic ligament, which originates from a deep cleft in the
liver, separating the left lobe from the caudate. This land-
mark serves as a convenient reference point for identifying
the esophagogastric junction (673). On cross section the
abdominal or submerged portion of the esophagus fre-
quently appears cone shaped with its base at the junction
with the gastric fundus (674). This segment is only rarely
distended by either air or barium, and hence is easily
Figure 4-146 Dilated esophagus in scleroderma. The esopha- mistaken as abnormal (675). This problem may be solved
gus (arrow) is dilated and fluid filled. The esophagus more often
appears dilated and air filled in patients with scleroderma. by scanning patients in the left lateral decubitus position
following ingestion of at least 200 mL of standard oral
contrast material (676).
A B
Figure 4-147 Paraesophageal varices. A: Unenhanced CT shows a large mass in relation to the distal esophagus. B: Enhanced CT shows
opacification of multiple paraesophageal varices (arrow), associated with cirrhosis and portal hypertension.
A B
Figure 4-148 Sliding hiatal hernia and esophageal varices. A, B: Sequential CT sections through the distal esophagus and esophageal
hiatus, respectively, show a large sliding hiatal hernia. Note elevated medial margins of the right crus (arrow in B). In addition to the stom-
ach, a portion of the peritoneal cavity has also herniated, within which a small quantity of ascites can be identified within the hernia sac
(arrow in A). Posteriorly, numerous dilated vascular structures can also be identified (curved arrows in A and B), because of paraesophageal
varices.
In a patient with sliding hiatal hernia, the most associated with an apparent increase in mediastinal fat
common abnormalities identified are dehiscence of the surrounding the distal esophagus, secondary to herniation
diaphragmatic crura and stretching of the phrenicoesoph- of omentum through the phrenicoesophageal ligament.
ageal ligament, which ceases to exist for all practical In the presence of massive ascites, it may be possible to
purposes in most adults. These findings manifest as widen- actually identify fluid within herniated peritoneum ante-
ing of the esophageal hiatus on cross section, identifiable rior to the contrast-filled stomach (677).
whenever the medial margins of the diaphragmatic crura Identification of sliding hiatal hernias rarely pres-
are not tightly opposed (Fig. 4-150) (559). Actual measure- ents much difficulty when they are seen on CT (671,673,
ments of the standard width of the esophageal hiatus, 674,678). Paraesophageal herniation is easily differentiated
defined as the distance between the medial margins of because in these cases, although the stomach is herniated,
the crura, have been reported. This distance has a mean the esophagogastric junction remains in a normal position
measurement of 10.7 mm (SD 2.4 mm) with a maximum (Fig. 4-151). Occasionally tumors arise in hernias; this
width of 15 mm (624). Sliding hiatal hernias are frequently appearance may also be mimicked by incomplete filling
A B
Figure 4-149 Gastroesophageal junction: schematic representations of the gastroesophageal junction in the coronal and transverse
planes, respectively. A: Note the relationship between the gastrohepatic ligament, the caudate lobe of the liver, and the medial wall of the
gastroesophageal junction. B: A cross section of the gastroesophageal region, corresponding to the level of the dotted line in A. Note that
laterally the gastrohepatic ligament can be identified as a line separating the caudate lobe posteriorly from the lateral segment of the left
lobe anteriorly. (From Imran MB, Kubota K, Yoshioka S, et al. Sclerosing mediastinitis: findings on fluorine-18 fluorodeoxyglucose positron
emission tomography. Clin Nucl Med. 1999;24:305–308, with permission.)
A B
Figure 4-150 Hiatal hernia. A–C: Sequential CT sections from

above-downward following the administration of oral contrast in a
patient with a moderate-sized sliding hiatal hernia. Note that the
medial margins of the crura are widely separated (curved arrows in
B and C). A portion of the contrast-filled stomach as well as peri-
toneal fat (straight arrow in B) can be identified between the
widened margins of the crura. A few linear densities can be identi-
fied within the fat, presumably representing peritoneal vessels. The
C open arrow in B points to the fissure of the gastrohepatic ligament.
A B
Figure 4-151 Paraesophageal hernia. A, B: Contrast-enhanced CT sections in a surgically documented paraesophageal hernia show that
a considerable portion of the stomach (S) lies anterior to and alongside the distal esophagus (arrows), which is normal in caliber. Note the
presence of fluid within a portion of herniated peritoneum to the right of the spine (curved arrow in A).
A B
Figure 4-152 Boerhaave syndrome. A: CT following oral contrast administration shows opacification of the esophagus, extravasation of
contrast into the mediastinum (black arrow), pneumomediastinum (small white arrows), and left pleural effusion (large white arrow). A right-
sided chest tube is in place. B: At a lower level, extensive contrast extravasation is visible (arrow).
of the herniated stomach (679). Accurate differentiation Although the diagnosis of esophageal perforation
usually requires esophagography. is generally made using esophagography, CT may pro-
vide valuable information concerning the extent of
associated mediastinal, pleural, and parenchymal dis-
Esophageal Perforation
ease or may first suggest the diagnosis in patients with
Esophageal perforation is a life-threatening condition nonspecific symptoms or chest pain (602,683–685).
which may be associated with esophageal carcinoma, Also, it is important to consider that routine radiographs
violent retching (Boerhaave syndrome) (Figs. 4-152 to may be normal in as many as 12% of patients with per-
4-154), or trauma or may be iatrogenic, complicating foration (686), and esophagrams performed using
endoscopy, esophageal dilatation, attempted intubation, water-soluble contrast agent may not show esophageal
or surgery (602,624,680–682). This condition is frequently leaks (687).
complicated by the rapid onset of mediastinitis, empyema, Findings on CT obtained without oral contrast admi-
and sepsis. nistration include paraesophageal mediastinal gas or
Figure 4-154 Esophageal perforation with esophageal-

Figure 4-153 Esophageal perforation: Boerhaave syndrome. CT pleural-cutaneous fistula. Helical CT section at the level of the
section through the distal esophagus following administration of distal esophagus following oral contrast administration shows
oral contrast media. There is perforation of the esophagus with free the presence of an esophageal-pleural-cutaneous fistula in a
extravasation of contrast into the mediastinum, associated with a patient with a history of trauma following placement of a right-
moderate-sized left pleural fluid collection. (Case courtesy of Robert sided pleural tube. In this case, administration of oral contrast
Meisell, MD, Booth Memorial Hospital, New York, New York.) confirms the continuity of these spaces.
fluid collections, esophageal wall thickening, or pleural

effusion (Fig. 4-152). For example, White et al. (682) re-
viewed the CT scans of 12 patients with esophageal perfo-
ration. CT abnormalities included esophageal wall thick-
ening in 9 patients, periesophageal fluid in 11 patients,
extraluminal air in 11, and pleural effusion in 9. The site
of the perforation was visible on the CT scan in two pa-
tients. In four patients (33%), CT findings were the first
indication of esophageal perforation (682). Furthermore,
in select cases (682), CT can be used to determine which
patients require immediate surgical intervention and
which patients can be managed conservatively; the pres-
ence of mediastinal fluid collections suggests the need for
intervention.
The use of CT esophagography has recently been
reported for the diagnosis of esophageal perforation or Figure 4-155 Mediastinitis secondary to retropharyngeal
rupture (Figs. 4-152 to 4-154) and may obviate fluoro- abscess. Mediastinal soft tissues are of water density because of
scopic esophagram by demonstrating contrast extravasa- edema, and small gas bubbles are visible.
tion (602). In a recent study (602), following scans
obtained without an oral contrast agent, patients with
clinically suspected esophageal perforation, or those with MEDIASTINITIS AND MEDIASTINAL
CT findings suggestive of this diagnosis, received approxi- ABSCESS
mately 50 mL of an aqueous solution consisting of 10%
iodinated low-osmolar intravenous contrast material and Acute mediastinal infections are uncommon and usually
effervescent granules (sodium bicarbonate and tartaric related to surgery, esophageal perforation, or spread of
acid), by injection through a nasogastric tube or rapid infection from adjacent regions (94,602,691–693). Infec-
ingestion. Contrast extravasation was shown in relation to tions may be classified as diffuse mediastinitis or mediasti-
paraesophageal air collections. nal abscess, depending on their extent. Diffuse mediastinitis
Although uncommon, the CT appearance of acquired has a relatively poor prognosis.
tracheoesophageal fistulas has also been reported (688). CT findings in mediastinitis include diffuse or streaky infil-
In patients having esophageal surgery, a localized fluid tration of mediastinal fat (greater than 25 HU), mediastinal
collection without air bubbles, may represent seroma widening, localized fluid collections, pleural or pericardial
(689). Aspiration (690) may be necessary for differentia- effusion, lymph node enlargement, and compression of medi-
tion from abscess (84). astinal structures (Figs. 4-155 to 4-157). Gas bubbles in the
A B
Figure 4-156 Mediastinitis 4 weeks after median sternotomy. A: Mediastinal fat is increased in attenuation and a focal fluid collection in
the right mediastinum (arrow) represents an abscess. A small amount of gas is visible in relation to the sternotomy incision. B: At a lower
level, a small gas bubble (arrow) is visible in the mediastinum and in relation to the sternotomy. This appearance is distinctly abnormal this
long after surgery.
sternotomy, findings can closely mimic the appearance of

mediastinitis, and abnormal findings can persist for up to
3 weeks (691,694,695). In a study by Kay et al. (694),
more than 75% of patients having median sternotomy
showed retrosternal fluid collections, air, hematoma, or
a combination of these in the early postoperative period.
In a study assessing the accuracy of CT in diagnosing medi-
astinitis after sternotomy (695), a diagnosis based on the
presence of mediastinal fluid or air collections had a sensi-
tivity of 100%. In the first 14 days after surgery, these
findings had a specificity of only 33%, but after 14 days,
their specificity was 100% (Fig. 4-156).
Figure 4-157 Mediastinitis following esophageal perforation.

PARASPINAL ABNORMALITIES
The mediastinum is widened, multiple air collections are visible,
and mediastinal fat is increased in attenuation. A wide variety of pathologic processes may involve the
paraspinal regions (477,696). Most frequently, masses
affecting these regions are neural in origin, including neu-
mediastinum, with or without associated fluid collections, are rogenic tumors, neurenteric cyst, and anterior or lateral
seen in up to half of cases, and are an important finding thoracic meningocele, or are related to the spine. Infec-
(691,692). In patients with an abscess, a localized fluid-filled tions involving the spine may lead to the development of
space is visible, often containing air (84,94). Abscess may have paraspinal abscess. This is most common with bacterial
a distinct wall, which enhances following contrast infusion. osteomyelitis or TB (Fig. 4-158). In addition, lymphoma
Because many cases of mediastinitis occur after and other causes of lymph node enlargement may result in
median sternotomy, it is important to consider the normal abnormalities in this region; paraspinal lymph nodes
postoperative appearance of the mediastinum in such freely communicate with lymph nodes in the upper abdo-
patients. In subjects having CT after an uncomplicated men, and contiguous involvement is common. Because
Figure 4-158 Tuberculous mediastinal abscess. A: Lateral radi-

ograph of the cervical spine shows destruction of the T1 vertebral
body (arrow). B: Contrast-enhanced CT section confirms lytic
destruction of T1 (arrow) associated with a large, poorly mar-
ginated posterior mediastinal fluid collection (curved arrows).
A Surgery confirmed tuberculous abscess.
the mediastinum communicates with the retroperitoneal

space via the esophageal hiatus, aortic hiatus, and other TABLE 4-36
defects in the diaphragm, diseases can spread between the PERIPHERAL NERVE SHEATH TUMORS
abdomen and thorax by direct extension; inflammatory
Neurilemoma (schwannoma), neurofibroma, malignant
masses such as pancreatic pseudocyst can involve the schwannoma
paraspinal mediastinal regions. Rare entities such as
Usually diagnosed in adults
extramedullary hematopoiesis, primary myolipoma of
the mediastinum, hemangioendothelioma, mediastinal CT findings
fibromatosis, and fibrosing mediastinitis have been rep- Well-marginated mass
Smooth, rounded, or elliptical
orted in the paravertebral regions (471,554,583,697). Occur in the paravertebral regions or along the courses
of nerves
Enlargement of neural foramina may occur
NEUROGENIC TUMORS Low attenuation in up to 70% (lipid-rich Schwann cells or
cystic regions)
Variable enhancement; peripheral enhancement is common
A high proportion of posterior mediastinal masses are of
neurogenic origin, particularly in children (4). Neurogenic MRI findings
T1-weighted: intensity slightly greater than muscle
tumors account for about 9% of primary mediastinal
T2-weighted: markedly increased signal intensity, often
masses in adults (698), although they are more prevalent heterogeneous
in children, representing 29% of mediastinal tumors (698). Schwannoma: high-intensity center and low-intensity rim
Tumors may arise from peripheral nerves and nerve in half
sheath (neurofibroma, schwannoma, malignant peripheral Neurofibroma: central region less intense than periphery
nerve sheath tumors) or sympathetic ganglia (ganglio-
neuroma, ganglioneuroblastoma, neuroblastoma) (699).
In two reviews of thoracic neurogenic tumors in 160 and
134 patients (698,700), the most common lesions were or organized more loosely in association with a myx-
schwannoma (31% to 32%), ganglioneuroma (25% to oid stroma (Antoni B pattern); areas of infarction are
26%), neuroblastoma (15% to 21%), ganglioneuroblas- common. Neurofibromas are also variable in appearance,
tomas (7% to 14%), and neurofibroma (5% to 10%), but consisting of spindle cells, a loose myxoid matrix, neu-
the incidence varies with the patient’s age. Nearly 85% of rofibrils, and collagen.
tumors in children are of ganglionic origin, whereas in The CT findings in patients with peripheral nerve
adults, more than 75% are nerve sheath tumors (698). sheath tumors have been well described (Table 4-36)
Specifically, schwannoma and neurofibroma are more (702–710). They typically appear as well-marginated,
common in adults, whereas ganglioneuroblastoma and smooth, rounded, or elliptical masses in the paravertebral
neuroblastoma are more common in children. The mean regions or along the courses of the vagus, phrenic, recur-
ages at diagnosis for neurogenic tumors in a study by Reed rent laryngeal, or intercostal nerves. Enlargement of neural
et al. (700) was 5.8 years for neuroblastoma, 8.4 years foramina, with or without extension into the spinal canal,
for ganglioneuroblastoma, 19.6 years for ganglioneuroma, may be associated with paravertebral tumors.
29.7 years for neurofibroma, and 38 years for schwan- Although peripheral nerve or nerve sheath tumors can
noma. Similar results were reported by Ribet and Cardot be of soft tissue attenuation, low attenuation is characteris-
(698). In their study, 23 of 28 neuroblastomas occurred tic and is seen in up to 73% of cases (Figs. 4-159 and
before the age of one. Patients with ganglioneuroblas- 4-160) (702,703,705,709). Low-attenuation areas within
tomas were slightly older, but all were diagnosed before nerve sheath tumors can be caused by the presence of
the age of 10. (a) lipid-rich Schwann cells, (b) adipocytes, (c) perineural
Malignant tumors are most common in children. In a adipose tissue entrapped by plexiform neurofibromas,
review of 60 patients with neurogenic tumors, there were (d) cystic spaces caused by the coalescence of interstitial
48 benign (80%) and 12 (20%) malignant tumors when fluid in schwannomas with Antoni B tissue, and (e) cystic
all age groups were considered; the malignancy rate was degeneration secondary to infarction (703,711). Variable
61.5% (8 of 13) in children and 8.5% (4 of 47, p .05) in enhancement of the tumor may be seen following contrast
adults (701). infusion; peripheral enhancement is common. In some
cases, these lesions may closely mimic the appearance of
a congenital cyst (94).
Peripheral Nerve Sheath Tumors
Neurofibromas may be associated with von Reckling-
Nerve sheath tumors included neurilemoma (schwan- hausen disease, which can result in vertebral abnormalities
noma), neurofibroma, and sarcoma (malignant schwan- including kyphoscoliosis, scalloped vertebrae, and lateral
noma) (Table 4-36). Schwannomas are composed of meningocele (712). In one study (700) more than one third
spindle cells densely packed together (Antoni A pattern) of patients with neurofibromas had neurofibromatosis.
C D
Figure 4-159 Neurogenic tumors. A: Contrast-enhanced CT scan shows a well-defined, slightly heterogeneous posterior paravertebral
mass. The appearance and location are typical. Biopsy proved neurofibroma. B, C: Contrast-enhanced CT sections in a different patient with
neurofibromatosis. The mediastinum is infiltrated by a relatively homogeneous soft tissue mass, resulting in marked displacement of
mediastinal vessels, including the azygos vein and aortic arch (arrow in B) as well as marked narrowing of the right and especially left main-
stem bronchi (arrow in C). D: Malignant neurofibrosarcoma. The tumor has infiltrated the chest wall. (B, C, Case courtesy Israel Gary,
Montefiore Medical Center, New York, New York.)
Malignant nerve sheath tumors, termed malignant On MRI (709,714–716), neurogenic tumors typically
schwannoma or neurofibrosarcoma, are relatively uncom- have slightly greater signal intensity than muscle on
mon, but represent up to 15% of nerve sheath tumors T1-weighted images, and markedly increased signal inten-
(698,713). The CT diagnosis of malignancy can be difficult. sity on T2-weighted images, although often in a heteroge-
Although benign tumors tend to be small, sharply mar- neous fashion (Figs. 4-160 to 4-162) (714,717). In half
ginated, and fairly homogeneous in attenuation, and of the cases in one study (717), schwannomas had a high-
malignant nerve sheath tumors tend to be large, infiltrat- intensity center and a relatively low-intensity outer wall;
ing, irregular, and inhomogeneous, these findings are not this appearance correlated with the presence of central
sufficiently reliable to obviate histologic evaluation cystic degeneration. Neurofibromas, on the other hand,
(708,709,713). In a study of eight cases of malignant showed the central region to be less intense than the
schwannoma arising in the thorax (713), tumors averaged periphery on T2-weighted images, with central contrast
9 cm in diameter (range 6 to 13 cm) and six showed areas enhancement; this correlated with the presence of tumor
of low attenuation caused by necrosis, but only three had tissue centrally and peripheral myxoid degeneration.
an irregular edge and five had a smooth margin. Both be- Plexiform neurofibroma represents an extensive fusiform
nign and malignant lesions may be symptomatic, render- or infiltrating mass along the course of the sympathetic
ing clinical differentiation of limited utility. Calcification chains, mediastinal, or intercostal nerves (Figs. 4-159B and
and some degree of contrast opacification may be present C and 4-160) (705). It is considered pathognomonic of von
with either benign or malignant tumors. Lung metastases Recklinghausen disease. As with localized nerve sheath
may be present (713). tumors, plexiform neurofibromas usually appear low in
Figure 4-160 Extensive mediastinal and intercostal neurogenic

tumors in neurofibromatosis. CT and MR correlation. A: CT ob-
tained with a bone window setting shows masses in the left medi-
astinum (arrow), the left paravertebral region, and in relation to
intercostal nerves. B: T1-weighted MRI (TR/TE 600/8 msec)
shows multiple neural tumors in the neck and left mediastinum.
These appear low in intensity. C: T1-weighted MRI (TR/TE 550/20
msec) following the injection of contrast material shows
inhomogeneous enhancement of the multiple tumors. Inhomogen-
C eous enhancement can be seen with neurofibroma or schwannoma.
attenuation as compared to muscle, with CT values ranging to determine a prognostic classification. The POG system is
from 15 to 20 HU on unenhanced scans. They are often based solely on the degree of differentiation of the different
multiple and lobulated, have ill-defined margins, and tend histologic elements. The Shimada classification combines
to surround mediastinal vessels with loss of normally histologic features and patient age at diagnosis (699).
visible fat planes (705). They can closely mimic the appear-
ance of extensive mediastinal lymph node enlargement.
Ganglioneuroma
Calcification and contrast enhancement can be seen.
Ganglioneuroma, a benign tumor made up of Schwann
cells, collagen, and ganglion cells, is a common tumor
in teenagers and young adults (Table 4-37). It cannot be
Tumors Originating from Sympathetic
accurately distinguished from schwannoma or neuro-
Ganglia fibroma on the basis of its appearance, although a
In children, 80% of posterior mediastinal masses are somewhat different whorled MRI appearance has been
derived from sympathetic ganglia (4). Included in this described (717).
group are ganglioneuroma, neuroblastoma, and ganglio-
neuroblastoma. Neuroblastoma and ganglioneuroblas-
Neuroblastoma
toma are usually grouped together for the purposes of
cancer reporting, staging, and survival statistics, because Neuroblastoma represents 8% to 10% of all childhood
both share a malignant potential (699). malignancies, but, because of its aggressive nature, it
Two histologic classification systems are commonly accounts for close to 15% of fatalities (699). The median
used to stratify these tumors into risk groups: the Shimada age at diagnosis is 22 months, and more than 95% of cases
classification and the Pediatric Oncology Group (POG) are diagnosed by 10 years of age (Table 4-37). Approxi-
classification (699). Both systems assess histologic features mately 15% of neuroblastomas arise in the mediastinum
A B
Figure 4-161 Neurogenic tumor: MR evaluation. A, B:

Sequential MR images through the lung apex from above-
downward show a sharply defined, homogeneous mass in the right
paravertebral space. A thin mantle of fat surrounds the lesion
(arrows in A), confirming that this lesion arises extrapleurally. Note
that the lesion clearly extends into the T1–2 foramen (arrow in B). C:
Coronal MR image confirms extension of the lesion through the in-
tervertebral foramen. This lesion has remained stable for 2 years,
compatible with the clinical and morphologic diagnosis of a neurofi-
broma. (Case courtesy of Andrew Litt, MD, New York University
C Medical Center, New York, New York.)
(Fig. 4-162), and almost all are located posteriorly. and invasion of the neural foramen into the epidural space
Mediastinal neuroblastoma is seen almost exclusively in is also frequent. Adenopathy of the renal hilum, porta
young children, younger than the age of 5 years (698,700). hepatis, and retroperitoneum may be seen. Invasion of the
It may present with a wide variety of signs and symptoms extradural spinal canal is frequent, even in the absence of
including chest pain, fever, malaise, anemia, Horner syn- neurologic signs and symptoms (718).
drome (ptosis, pupillary constriction, and ipsilateral Using MRI, neuroblastoma is typically heterogeneous,
facial anhidrosis and flushing), and extremity weakness variably enhancing, and of relatively low signal intensity
(699). Less common presentations include opsoclonus- on T1-weighted images (699). On T2-weighted images,
myoclonus (jerking movements of the extremities and neuroblastoma appears intense (Fig. 4-162). MRI has the
eyes, sometimes with cerebellar ataxia). advantage of allowing tumor and surrounding soft tissues
On CT, neuroblastomas appear as soft tissue attenua- to be distinguished more readily than on CT and is more
tion masses, but up to 80% to 90% contain speckled or accurate in the recognition of bone marrow involvement.
curvilinear calcifications (699). Low-attenuation areas of Neuroblastoma may extend from a primary site in the
necrosis or hemorrhage are frequently noted at CT. They abdomen into the thorax. Most commonly, this occurs by
are most common in the paravertebral regions and may direct invasion through the retrocrural space and into the
extend superiorly and inferiorly for several centimeters. lower paravertebral regions, often on both sides. In the
Neuroblastoma often shows heterogeneous enhancement rare instance of adult neuroblastoma, differentiation from
following contrast injection. lymphoma may be difficult (710).
CT and MRI may be used to help determine the Neuroblastoma is an uncommon malignancy in adults,
presence and extent of mediastinal or vertebral column and often pursues an aggressive clinical course, involves
invasion. In the abdomen, vascular encasement and com- multiple sites, and has a poor prognosis. Tumors usually
pression of the renal vessels, splenic vein, inferior vena involve the retroperitoneum but may occur in the medi-
cava, aorta, celiac artery, and superior mesenteric artery astinum (719). The most common CT finding is a poorly
may occur, although vascular invasion is rare. Regional marginated and heterogeneous mass without calcification.
invasion of psoas and paraspinal musculature may occur, Cystic components may be present. On contrast-enhanced
A B
Figure 4-162 Mediastinal neuroblastoma in a 9-month-old child. A:

T1-weighted coronal MRI (TR/TE 722/20 msec) shows a large right
mediastinal mass. B: T1-weighted transaxial MRI (TR/TE 689/20
msec) shows the mass to be relatively low in intensity. C: T1-weighted
transaxial MRI (TR/TE 714/20 msec) following the injection of con-
C trast agent shows significant enhancement of the mass.
CT or MR images, tumors showed heterogeneous enhance- 50% of patients with recurrence (720). Indium-111 pente-
ment. On T2-weighted images, all masses demonstrated treotide scintigraphy may also demonstrate neuroblastic
predominantly hyperintense signal relative to skeletal tumors, although it does not appear to be superior to
muscle and the images showed heterogeneous appearance MIBG scintigraphy (699). As documented by Levine et al.
with focal areas of high intensity interspersed with septa- (709), in these cases, gallium scintigraphy appears to be a
tions of low signal intensity. promising screening technique, as gallium uptake appears
Radionuclide imaging may be helpful in the diagnosis to occur only in malignant lesions.
of these tumors or in the detection of metastases. Identifi-
cation of the primary tumor is performed with either a cat-
Ganglioneuroblastoma
echolamine analog (metaiodobenzylguanidine iodine-123
or MIBG) or a somatostatin analog (pentetreotide indium- Ganglioneuroblastoma is found in somewhat older child-
111). MIBG is taken up by catecholamine-producing ren than is neuroblastoma and is less common. Ganglio-
tumors such as neuroblastoma, ganglioneuroblastoma, neuroblastomas are regarded by some as partially matured
ganglioneuroma, pheochromocytoma, carcinoid tumor, malignant neuroblastomas (Table 4-37) (698) and are
and medullary thyroid cancer. Although 90% to 95% of somewhat intermediate in histology between neuro-
neuroblastoma and ganglioneuroblastoma secrete cate- blastoma and ganglioneuroma. Their imaging characteris-
cholamines, only about 70% of these tumors are MIBG tics are indistinguishable from those of neuroblastoma
positive. The use of MIBG scintigraphy in routine patient (699). They can present either as a large, smooth spherical
assessment is controversial; in a recent study (720), MIBG mass or as a small, elongated sausage-shaped mass. Some of
results did not alter staging or treatment. In the same the large ones are of low, homogeneous density and show
study, recurrent neuroblastoma failed to show uptake in little, if any, contrast enhancement.

TABLE 4-37 of Neurogenic Tumors
TUMORS ORIGINATING FROM
Compared with other imaging modalities including CT,
SYMPATHETIC GANGLIA
MR has several distinct advantages for imaging neuro-
Ganglioneuroma, neuroblastoma, and ganglioneuroblastoma genic tumors. In addition to identifying the tumor, MR is
Usually diagnosed in children and young adults especially helpful in assessing intraspinal extension, as
well as the presence of associated spinal cord pathology
Ganglioneuroma
Mean age at diagnosis about 20 years (Figs. 4-160 to 4-162) (48). Its ability to obtain multipla-
Benign tumor of Schwann cells, collagen, and ganglion cells nar images is particularly helpful as many of these tu-
Cannot be distinguished from schwannoma or neurofibroma mors demonstrate longitudinal extension into structures
Neuroblastoma along the axis of the spine. Recently, MRI findings have
Mean age at diagnosis about 2 years been shown to accurately reflect underlying pathologic
CT findings findings in patients with neurogenic tumors (717). For
Soft tissue attenuation masses these reasons, MRI has largely replaced CT as an initial
Low-attenuation areas of necrosis or hemorrhage common
80%–90% contain speckled or curvilinear calcifications
imaging modality in the evaluation of patients with sus-
Heterogeneous enhancement pected neurogenic tumors. Unfortunately, MRI is no
Paravertebral, extend for several centimeters more specific than CT in differentiating benign from ma-
Invasion of the extradural spinal canal is frequent lignant lesions.
MRI findings
T1-weighted: heterogeneous, relatively low signal intensity
T2-weighted: markedly increased signal intensity
Variably enhancing
PARAGANGLIOMA
Ganglioneuroblastoma
Mean age at presentation about 8 years Paraganglioma (chemodectoma) is a rare tumor originat-
Intermediate in histology between neuroblastoma and
ganglioneuroma
ing from neuroectodermal cells located in relation to the
CT and MRI findings autonomic nervous system, especially in the region of the
Large, smooth spherical mass APW and the posterior mediastinum (Fig. 4-163) (100,
Small, elongated sausage-shaped mass 721); it has also been identified within the atria. In the
Homogeneous study of thoracic neurogenic tumors reported by Reed
Little enhancement
et al. (700), only 4% were paragangliomas.
A B
Figure 4-163 Mediastinal paraganglioma. A, B: CT scans obtained at the level of the left atrium, both pre– and post–intravenous contrast
enhancement, respectively. A large, homogeneous, low-density mass causing marked compression of the left atrium can be identified
(arrows in A and B). Note that following intravenous contrast administration there has been marked contrast enhancement within the mass.
Without benefit of the precontrast image, this mass could easily be overlooked as representing a normal left atrium. (Case courtesy of Barry
Gross, MD, Detroit, Michigan.)
TABLE 4-38
PARAGANGLIOMA
Neuroectodermal tumor located in relation to the autonomic
nervous system
Occurs in the posterior mediastinum and aortopulmonary
window
Half of paravertebral masses secrete catecholamines
Catecholamine secretion rare with aortopulmonary window
tumors
CT findings
Nonspecific appearance on unenhanced scans
Dense enhancement following contrast injection
On unenhanced CT, paraganglioma has no characteris-

tic features. However, scanning during bolus contrast infu- Figure 4-164 Anterior thoracic meningocele associated with
sion shows dense enhancement (Table 4-38) (721), and neurofibromatosis: CT myelography. Vertebral defects and scolio-
sis are evident. The meningocele (arrow) is opacified.
marked hypervascularity is seen on angiography (100).
Patients are often asymptomatic and the tumor is detected
incidentally, although compression of mediastinal struc-
tures may result in symptoms. The tumor may involve the
spinal canal (721). Approximately 10% of these lesions fluid (Figs. 4-164 and 4-165). Findings that suggest the di-
are malignant or invasive. Paravertebral paragangliomas agnosis include rib or vertebral anomalies at the same level
are more easily resectable than aortopulmonary tumors or an association with scoliosis. The mass is often visible at
and have a better prognosis. Local recurrence is common the apex of the scoliotic curve (725). MRI is diagnostic, as
following surgery in patients with aortopulmonary para- is filling of the meningocele with contrast at CT myelogra-
ganglioma. MRI may be valuable in planning a surgical phy (Figs. 4-164 and 4-165). As with meningocele,
approach (722). neurenteric cysts are frequently associated with vertebral
About half of patients with paravertebral paraganglioma anomalies or scoliosis. They rarely fill with myelographic
have symptoms of catecholamine secretion by the tumor or contrast.
associated tumors in other locations, but catecholamine
secretion is rare in patients with aortopulmonary tumors.
Identification of hormonally active lesions has been greatly
aided recently by use of 131I metaiodobenzylguanidine EXTRAMEDULLARY HEMATOPOIESIS
(131I-MIBG) scintigraphy, which localizes to catecholamine-
producing tumors, including neuroblastoma and carcinoid Extramedullary hematopoiesis can result in paravertebral
tumor (101,723,724). masses in patients with severe hemolytic anemia caused by
excessive destruction of blood cells. It can be seen in the
presence of thalassemia, hereditary spherocytosis, and
sickle-cell anemia (85,726,727). These masses are of
ANTERIOR OR LATERAL THORACIC unknown origin but may arise from herniations of verte-
MENINGOCELE bral or rib marrow through small cortical defects or may
arise from lymph nodes or elements of the reticuloen-
This entity represents anomalous herniation of the spinal dothelial system.
meninges through an intervertebral foramen or a defect in Lobulated paravertebral masses, usually multiple and
the vertebral body. It results in a soft tissue mass visible on bilateral and caudad to the sixth thoracic vertebra, are
chest radiographs. In many patients, this abnormality is typically seen. They appear well marginated and of ho-
associated with neurofibromatosis; most are detected in mogeneous soft tissue attenuation (30 to 65 HU) (726)
adults (725). or may show areas of fat attenuation (50 HU) that
Meningoceles are described as lateral or anterior, may increase after treatment (85,91,728). In some pa-
depending on their relationship to the spine. They are tients, the masses may appear to be composed primarily
slightly more common on the right (725). On CT they ap- of fat. Fat may be seen within the mass using MRI (48).
pear low in attenuation, as they contain cerebrospinal The diagnosis can be suggested by the presence of
chronic anemia and radiographic or CT findings in the

skeleton suggesting a bone marrow abnormality.
Coarsening of the trabecular pattern, rib expansion, and
periosteal new bone may be seen (726). Although not
always present, splenomegaly is common (Figs. 4-166
and 4-167).
MEDIASTINAL PSEUDOCYST
Pancreatic pseudocyst represents an encapsulated collec-

tion of pancreatic secretions, blood, and necrotic mate-
rial (729). Mediastinal extension of a pancreatic pseudo-
A
cyst is rare (497) but can occur via the aortic hiatus or
esophageal hiatus or through a defect in the diaphragm.
Symptoms are generally those of pancreatitis. CT shows a
cystic and low-attenuation mass in the posterior medi-
astinum or adjacent thoracic cavity, associated with com-
pression or displacement of the esophagus or splaying of
the diaphragmatic crura (497,498). Mediastinal pseudo-
cysts are commonly located under and posterior to the
heart, anterior to the aorta and esophagus, and medial to
the inferior vena cava, in close relationship to the
esophageal or aortic hiatus (94). The fluid can be of
water density or higher, depending on the presence of
blood or infection. The presence of an abdominal com-
ponent is common, but not invariably present (521). CT
may also be helpful in percutaneous catheter drainage of
a mediastinal pseudocyst (94,497). Reported MRI find-
B
ings are compatible with the cystic nature of this lesion
(521).
C
Figure 4-165 Anterior thoracic meningocele associated with
neurofibromatosis: CT and MR correlation. A: CT shows a large,
low-attenuation meningocele occupying the apex of the right
thoracic cavity, associated with vertebral anomalies. A small left
meningocele is also present. B: Transaxial T1-weighted (TR/TE Figure 4-166 Extramedullary hematopoiesis. Contrast-enhanced
731/20 msec) MRI at nearly the same level as A shows the right CT section through the mid thorax shows marked expansion of the
meningocele to very low in intensity because of its fluid contents. medial aspects of the ribs bilaterally with evidence of bony trabecu-
Communication with the spinal canal and the left meningocele are lae extending through the cortex to form paraspinal masses. This
visible. C: Coronal MRI also shows the bilateral meningoceles. patient had longstanding thalassemia.
A B
Figure 4-167 Extramedullary hematopoiesis: MR evaluation. A: Unenhanced T1-weighted gradient-echo MR image (echo time of 4 msec)
in this patient with thalassemia shows evidence of iron overload within the liver (lower signal than skeletal muscle) and prior splenectomy.
Note bilateral small paravertebral masses representing extramedullary hematopoiesis (arrows). B: The masses (arrows) enhance strongly
after the administration of gadolinium.
REFERENCES 19. Mazzeo S, Caramella D, Gennai A, et al. Multidetector CT and

virtual endoscopy in the evaluation of the esophagus. Abdom
1. Baron RL, Levitt RG, Sagel SS, Stanley RJ. Computed tomogra- Imaging. 2004;29:2–8.
phy in the evaluation of mediastinal widening. Radiology. 20. Jones CM, Athanasiou T. Is virtual bronchoscopy an efficient
1981;138:107–113. diagnostic tool for the thoracic surgeon? Ann Thorac Surg.
2. Webb WR. Advances in computed tomography of the thorax. 2005;79:365–374.
Radiol Clin North Am. 1983;21:723–739. 21. Ueno J, Murase T, Yoneda K, et al. Three-dimensional imaging
3. Sones PJ, Torres WE, Colvin RS, et al. Effectiveness of CT in eval- of thoracic diseases with multi-detector row CT. J Med Invest.
uating intrathoracic masses. AJR Am J Roentgenol. 1982;139: 2004;51:163–170.
469–475. 22. Shaer AH, Bashist B. Case report. Computed tomography of
4. Merten DF. Diagnostic imaging of mediastinal masses is children. bronchial artery aneurysm with erosion into the esophagus.
AJR Am J Roentgenol. 1992;158:825–832. J Comput Assist Tomogr. 1989;13:1069–1071.
5. Webb WR, Sostman HD. MR imaging of thoracic diseases: clini- 23. Finkelstein SE, Schrump DS, Nguyen DM, et al. Comparative
cal uses. Radiology. 1992;182:621–630. evaluation of super high-resolution CT scan and virtual bron-
6. Gefter W. Chest applications of magnetic resonance imaging: an choscopy for the detection of tracheobronchial malignancies.
update. Radiol Clin North Am. 1988;26:573–588. Chest. 2003;124:1834–1840.
7. Weinreb JC, Naidich DP. Thoracic magnetic resonance imaging. 24. McAdams HP, Goodman PC, Kussin P. Virtual bronchoscopy for
Clin Chest Med. 1991;12:33–54. directing transbronchial needle aspiration of hilar and medi-
8. Laurent F, Latrabe V, Lecesne R, et al. Mediastinal masses: diag- astinal lymph nodes: a pilot study. AJR Am J Roentgenol.
nostic approach. Eur Radiol. 1998; 8:1148–1159. 1998;170:1361–1364.
9. Boiselle PM, Patz EF Jr, Vining DJ, et al. Imaging of mediastinal 25. Gruden JF, Tigges S, Baron MG, Pearlman H. MDCT pulmonary
lymph nodes: CT, MR, and FDG PET. Radiographics. 1998;18: angiography: image processing tools. Semin Roentgenol. 2005;
1061–1069. 40:48–63.
10. Boiselle PM. MR imaging of thoracic lymph nodes. A comparison 26. Yoon YC, Lee KS, Jeong YJ, et al. Hemoptysis: bronchial and
of computed tomography and positron emission tomography. nonbronchial systemic arteries at 16-detector row CT. Radiology.
Magn Reson Imaging Clin North Am. 2000;8:33–41. 2005;234:292–298.
11. Shaha AR, Jaffe BM. Cervical exploration for primary hyper- 27. Siegel MJ. Multiplanar and three-dimensional multi-detector
parathyroidism. J Surg Oncol. 1993;52:14–17. row CT of thoracic vessels and airways in the pediatric popula-
12. Reddy GP, Higgins CB. MR imaging of the thoracic aorta. Magn tion. Radiology. 2003;229:641–650.
Reson Imaging Clin North Am. 2000;8:1–15, vii. 28. Sharma U, Ghai S, Paul SB, et al. Helical CT evaluation of aortic
13. Muller NL. Computed tomography and magnetic resonance im- aneurysms and dissection: a pictorial essay. Clin Imaging. 2003;
aging: past, present and future. Eur Respir J Suppl. 2002;35: 27:273–280.
3s–12s. 29. Rubin GD. CT angiography of the thoracic aorta. Semin
14. Kalender WA. Technical foundations of spiral CT. Semin Roentgenol. 2003;38:115–134.
Ultrasound CT MR. 1994;15:81–89. 30. Chooi WK, Matthews S, Bull MJ, Morcos SK. Multislice
15. Costello P, Dupuy DE, Ecker CP, Tello R. Spiral CT of the thorax computed tomography in staging lung cancer: the role of multi-
with reduced volume of contrast material: a comparative study. planar image reconstruction. J Comput Assist Tomogr. 2005;29:
Radiology. 1992;185:663–666. 357–360.
16. Heiken JP, Brink JA, Vannier MW. Spiral (helical) CT. Radiology. 31. Cascade PN, Gross BH, Kazerooni EA, et al. Variability in the
1993;189:647–656. detection of enlarged mediastinal lymph nodes in staging lung
17. Rubin GD. Helical CT angiography of the thoracic aorta. cancer: a comparison of contrast-enhanced and unenhanced CT.
J Thorac Imaging. 1997;12:128–149. AJR Am J Roentgenol. 1998;170:927–931.
18. Rubin GD, Napel S, Leung AN. Volumetric analysis of volumetric 32. Im J-G, Song KS, Kang HS, et al. Mediastinal tuberculous lym-
data: achieving a paradigm shift. Radiology. 1996;200:312–317. phadenitis: CT manifestations. Radiology. 1987;164:115–119.
33. Im JG, Itoh H, Shim YS, et al. Pulmonary tuberculosis: CT 58. Siegel MJ, Glazer HS, Weiner JI, Molina PL. Normal and abnor-
findings—early active disease and sequential change with mal thymus in childhood: MR imaging. Radiology. 1989;172:
antituberculous therapy. Radiology. 1993;186:653–660. 367–371.
34. Andronikou S, Joseph E, Lucas S, et al. CT scanning for the 59. Smathers RL, Buschi AJ, Pope TL, et al. The azygos arch: normal
detection of tuberculous mediastinal and hilar lymphadenopa- and pathologic CT appearance. AJR Am J Roentgenol. 1982;139:
thy in children. Pediatr Radiol. 2004;34:232–236. 477–483.
35. Kirsch CFE, Webb EM, Webb WR. Multicentric Castleman’s 60. Rockoff SD, Druy EM. Tortuous azygos arch simulating a pul-
disease and POEMS syndrome: CT findings. J Thorac Imaging. monary lesion. AJR Am J Roentgenol. 1982;138:577–579.
1997;12:75–77. 61. Jolles PR, Shin MS, Jones WP. Aortopulmonary window
36. Moon WK, Im JG, Kim JS, et al. Mediastinal Castleman’s disease: lesions: detection with chest radiography. Radiology. 1986;159:
CT findings. J Comput Assist Tomogr. 1994;18:43–46. 647–651.
37. Remy-Jardin M, Dequiedt P, Ertzbischoff O, et al. Safety and 62. Aronberg DJ, Peterson RR, Glazer HS, Sagel SS. The superior
effectiveness of gadolinium-enhanced multi-detector row spiral CT sinus of the pericardium: CT appearance. Radiology. 1984;153:
angiography of the chest: preliminary results in 37 patients with 489–492.
contraindications to iodinated contrast agents. Radiology. 63. McMurdo KK, Webb WR, von SG, Gamsu G. Magnetic reso-
2005;235:819–826. nance imaging of the superior pericardial recesses. AJR Am
38. Rubin GD, Lane MJ, Bloch DA, et al. Optimization of thoracic J Roentgenol. 1985;145:985–988.
spiral CT: effects of iodinated contrast medium concentration. 64. Fitzgerald SW, Donaldson JS. Azygoesophageal recess: normal
Radiology. 1996;201:785–791. CT appearance in children. AJR Am J Roentgenol. 1991;158:
39. St Amour TE, Siegel M, Glazer HS, Nadel SN. CT appearances of 1101–1104.
the normal and abnormal thymus in childhood. J Comput Assist 65. Müller NL, Webb WR, Gamsu G. Subcarinal lymph node
Tomogr. 1987;11:645–650. enlargement: radiographic findings and CT correlation. AJR Am
40. Aronberg DJ, Glazer HS, Madsen K, Sagel SS. Normal thoracic J Roentgenol. 1985;145:15–19.
aortic diameters by computed tomography. J Comput Assist 66. Glazer GM, Gross BH, Quint LE, et al. Normal mediastinal
Tomogr. 1984;8:247–250. lymph nodes: number and size according to American Thoracic
41. Flohr T, Prokop M, Becker C, et al. A retrospectively ECG-gated Society mapping. AJR Am J Roentgenol. 1985;144:261–265.
multislice spiral CT scan and reconstruction technique with sup- 67. Kiyono K, Sone S, Sakai F, et al. The number and size of normal
pression of heart pulsation artifacts for cardio-thoracic imaging mediastinal lymph nodes: a postmortem study. AJR Am
with extended volume coverage. Eur Radiol. 2002;12:1497–1503. J Roentgenol. 1988;150:771–776.
42. Keall PJ, Starkschall G, Shukla H, et al. Acquiring 4D thoracic 68. Takasugi JE, Godwin JD. CT appearance of the retroaortic anas-
CT scans using a multislice helical method. Phys Med Biol. tomoses of the azygos system. AJR Am J Roentgenol. 1990;154:
2004;49:2053–2067. 41–44.
43. Nikolaou K, Flohr T, Knez A, et al. Advances in cardiac CT imag- 69. Smathers RL, Lee JKT, Heiken JP. Clinical image: anomalous
ing: 64-slice scanner. Int J Cardiovasc Imaging. 2004;20:535–540. preaortic interazygous vein. J Comput Assist Tomogr. 1983;7:
44. Hofmann LK, Zou KH, Costello P, Schoepf UJ. Electrocardi- 732–733.
ographically gated 16-section CT of the thorax: cardiac motion 70. Sharma A, Fidias P, Hayman LA, et al. Patterns of lymphadenopa-
suppression. Radiology. 2004;233:927–933. thy in thoracic malignancies. Radiographics. 2004;24:419–434.
45. Heuschmid M, Kuettner A, Schroeder S, et al. ECG-gated 71. Aquino SL, Hayman LA, Loomis SL, Taber KH. The source and
16-MDCT of the coronary arteries: assessment of image quality direction of thoracic lymphatics, part II: The lower thorax.
and accuracy in detecting stenoses. AJR Am J Roentgenol. 2005; J Comput Assist Tomogr. 2003;27:657–661.
184:1413–1419. 72. Aquino SL, Hayman LA, Loomis SL, Taber KH. Source and direc-
46. Hamoir XL, Flohr T, Hamoir V, et al. Coronary arteries: assess- tion of thoracic lymphatics: part I—upper thorax. J Comput
ment of image quality and optimal reconstruction window in Assist Tomogr. 2003;27:292–296.
retrospective ECG-gated multislice CT at 375-ms gantry rotation 73. Takahashi H, Kuboyama S, Abe H, et al. Clinical feasibility of
time. Eur Radiol. 2005;15:296–304. noncontrast-enhanced magnetic resonance lymphography of
47. Hong C, Chrysant GS, Woodard PK, Bae KT. Coronary artery the thoracic duct. Chest. 2003;124:2136–2142.
stent patency assessed with in-stent contrast enhancement 74. Erden A, Fitoz S, Yagmurlu B, Erden I. Abdominal confluence
measured at multi-detector row CT angiography: initial experi- of lymph trunks: detectability and morphology on heavily
ence. Radiology. 2004;233:286–291. T2-weighted images. AJR Am J Roentgenol. 2005;184:35–40.
48. Erasmus JJ, McAdams HP, Donnelly LF, Spritzer CE. MR imaging 75. Pinto PS, Sirlin CB, Andrade-Barreto OA, et al. Cisterna chyli at
of mediastinal masses. Magn Reson Imaging Clin North Am. routine abdominal MR imaging: a normal anatomic structure in
2000;8:59–89. the retrocrural space. Radiographics. 2004;24:809–817.
49. Thompson BH, Stanford W. MR imaging of pulmonary and 76. Delille JP, Hernigou A, Sene V, et al. Maximal thickness of the
mediastinal malignancies. Magn Reson Imaging Clin North Am. normal human pericardium assessed by electron-beam com-
2000;8:729–739. puted tomography. Eur Radiol. 1999;9:1183–1189.
50. Franquet T, Erasmus JJ, Gimenez A, et al. The retrotracheal 77. Bull RK, Edwards PD, Dixon AK. CT dimensions of the normal
space: normal anatomic and pathologic appearances. Radiogra- pericardium. Br J Radiol. 1998;71:923–925.
phics. 2002;22(spec no):S231–246. 78. Broderick LS, Brooks GN, Kuhlman JE. Anatomic pitfalls of the
51. Gamsu G, Webb WR. Computed tomography of the trachea: heart and pericardium. Radiographics. 2005;25:441–453.
normal and abnormal. AJR Am J Roentgenol. 1982;139: 79. Axel L. Assessment of pericardial disease by magnetic resonance
321–326. and computed tomography. J Magn Reson Imaging. 2004;19:
52. Gamsu G, Webb WR. Computed tomography of the trachea and 816–826.
mainstem bronchi. Semin Roentgenol. 1983;18:51–60. 80. Groell R, Schaffler GJ, Rienmueller R. Pericardial sinuses and
53. Baron RL, Lee JK, Sagel SS, Peterson RR. Computed tomography recesses: findings at electrocardiographically triggered electron-
of the normal thymus. Radiology. 1982;142:121–125. beam CT. Radiology. 1999;212:69–73.
54. de Geer G, Webb WR, Gamsu G. Normal thymus: assessment 81. Macchiarini P, Ostertag H. Uncommon primary mediastinal
with MR and CT. Radiology. 1986;158:313–317. tumours. Lancet Oncol. 2004;5:107–118.
55. Francis IR, Glazer GM, Bookstein FL, Gross BH. The thymus: 82. Kim Y, Lee KS, Yoo JH, et al. Middle mediastinal lesions: imag-
reexamination of age-related changes in size and shape. AJR Am ing findings and pathologic correlation. Eur J Radiol. 2000;35:
J Roentgenol. 1985;145:249–254. 30–38.
56. Heiberg E, Wolverson MK, Sundaram M, Nouri S. Normal 83. Glazer HS, Molina PL, Siege MJ, Sagel SS. Pictorial essay. High-
thymus: CT characteristics in subjects under age 20. AJR Am J attenuation mediastinal masses on unenhanced CT. AJR Am J
Roentgenol. 1982;138:491–494. Roentgenol. 1991;156:45–50.
57. Moore AV, Korobkin M, Olanow W, et al. Age-related changes in 84. Glazer HS, Siege MJ, Sagel SS. Pictorial essay. Low-attenuation
the thymus gland: CT-pathologic correlation. AJR Am J Roent- mediastinal masses on CT. AJR Am J Roentgenol. 1989;152:
genol. 1983;141:241–246. 1173–1177.
85. Glazer HS, Wick MR, Anderson DJ, et al. CT of fatty thoracic 114. Nicolaou S, Müller NL, Li DKB, Oger JJF. Thymus in myasthenia
masses. AJR Am J Roentgenol. 1992;159:1181–1187. gravis: comparison of CT and pathologic findings and clinical
86. Buckley JA, Vaughn DD, Jabra AA, et al. CT evaluation of medi- outcome after thymectomy. Radiology. 1996;20:471–474.
astinal masses in children: spectrum of disease with pathologic 115. Molina PL, Siegel MJ, Glazer HS. Thymic masses on MR imag-
correlation. Crit Rev Diagn Imaging. 1998;39:365–392. ing. AJR Am J Roentgenol. 1990;155:495–500.
87. Mendez G, Isikoff MB, Isikiff SK, Sinner WN. Fatty tumors of 116. Wittram C, Fischman AJ, Mark E, et al. Thymic enlargement and
the thorax demonstrated by CT. AJR Am J Roentgenol. 1979; FDG uptake in three patients: CT and FDG positron emission
133:207–212. tomography correlated with pathology. AJR Am J Roentgenol.
88. Enzi G, Biondetti PR, Fiore D, Mazzoleni FD. Computed tomog- 2003;180:519–522.
raphy of deep fat masses in multiple symmetrical lipomastosis. 117. Ferdinand B, Gupta P, Kramer EL. Spectrum of thymic uptake at
Radiology. 1982;144:122–124. 18F-FDG PET. Radiographics. 2004;24:1611–1616.
89. Lee WJ, Fatal G. Mediastinal lipomatosis in simple obesity. 118. Nakahara T, Fujii H, Ide M, et al. FDG uptake in the morpholog-
Chest. 1976;70:308–309. ically normal thymus: comparison of FDG positron emission
90. Rohlfing BM, Korobkin N, Hall AD. Computed tomography of tomography and CT. Br J Radiol. 2001;74:821–824.
intrathoracic omental herniation and other mediastinal fatty 119. Chen J, Weisbrod GL, Herman SJ. Computed tomography and
masses. J Comput Assist Tomogr. 1977;1:181–183. pathologic correlations of thymic lesions. J Thorac Imaging.
91. Martin J, Palacio A, Petit J, Martin C. Fatty transformation of 1988;3:61–65.
extramedullary hematopoiesis following splenectomy: CT 120. Keen SJ, Libshitz HI. Thymic lesions. Experience with computed
features. J Comput Assist Tomogr. 1990;14:477–478. tomography in 24 patients. Cancer. 1987;59:1520–1523.
92. Brown LR, Aughenbaugh GL. Masses of the anterior mediastinum: 121. Weinreb JC, Mootz A, Cohen JM. MRI evaluation of mediastinal
CT and MR imaging. AJR Am J Roentgenol. 1991;157: 1171–1180. and thoracic inlet venous obstruction. AJR Am J Roentgenol.
93. Kim JH, Goo JM, Lee HJ, et al. Cystic tumors in the anterior 1986;146:679–684.
mediastinum. Radiologic-pathological correlation. J Comput 122. Freundlich IM, McGavran MH. Abnormalities of the thymus. J
Assist Tomogr. 2003;27:714–723. Thorac Imaging. 1996;11:58–65.
94. Jeung MY, Gasser B, Gangi A, et al. Imaging of cystic masses of 123. Hoffman OA, Gillespie DJ, Aughenbaugh GL, Brown LR.
the mediastinum. Radiographics. 2002;22(spec no):S79–93. Primary mediastinal neoplasms (other than thymoma). Mayo
95. Spizarny DL, Rebner M, Gross BH. CT of enhancing mediastinal Clin Proc. 1993;68:880–891.
masses. J Comput Assist Tomogr. 1987;11:990–993. 124. Wychulis AR, Payne WS, Clagett OT, Wollner LB. Surgical treat-
96. Katlic MR, Wang C, Grillo HC. Substernal goiter. Ann Thorac ment of mediastinal tumors. a 40-year experience. J Thorac
Surg. 1985;39:391–399. Cardiovasc Surg. 1971;62:379–391.
97. Doppman JL, Skarulis MC, Chen CC, et al. Parathyroid adenomas 125. Strollo DC, Rosado-de-Christenson ML. Tumors of the thymus.
in the aortopulmonary window. Radiology. 1996;201:456–462. J Thorac Imaging. 1999;14:152–171.
98. McAdams HP, Rosado-de-Christenson ML, Moran CA. 126. Inoue K, Sugio K, Inoue T, et al. Hyperplasia of the thymic gland
Mediastinal hemangioma: radiographic and CT features in 4 in a patient with Graves’ disease. Ann Thorac Cardiovasc Surg.
patients. Radiology. 1994;193:399–402. 2000;6:397–400.
99. Joseph AE, Donaldson JS, Reynolds M. Neck and thorax venous 127. Niendorf ER, Parker JA, Yechoor V, et al. Thymic hyperplasia in
aneurysm: association with cystic hygroma. Radiology. 1989;170: thyroid cancer patients. J Thorac Imaging. 2005;20:1–4.
109–112. 128. Gelfand DW, Goldman AS, Law EJ, et al. Thymic hyperplasia in
100. Drucker EA, McLoud TC, Dedrick CG, et al. Mediastinal para- children recovering from thermal burns. J Trauma. 1972;12:
ganglioma: radiologic evaluation of an unusual vascular tumor. 813–817.
AJR Am J Roentgenol. 1987;148:521–522. 129. Prevot S, Audouin J, Andre-Bougaran J, et al. Thymic pseudotu-
101. Sheps SG, Brown ML. Localization of mediastinal paragan- morous enlargement due to follicular hyperplasia in a human
gliomas (pheochromocytoma). Chest. 1985;87:807–809. immunodeficiency virus sero-positive patient. Immunohis-
102. Onik G, Goodman PC. CT of Castleman Disease. AJR Am J tochemical and molecular biological study of viral infected
Roentgenol. 1983;140:691–692. cells. Am J Clin Pathol. 1992;97:420–425.
103. Mullen B, Richardson JD. Primary anterior mediastinal tumors 130. Kontny HU, Sleasman JW, Kingma DW, et al. Multilocular
in children and adults. Ann Thorac Surg. 1986;42:338–345. thymic cysts in children with human immunodeficiency virus
104. Strollo DC, Rosado de Christenson ML, Jett JR. Primary medi- infection: clinical and pathologic aspects. J Pediatr. 1997;131:
astinal tumors. Part 1: tumors of the anterior mediastinum. 264–270.
Chest. 1997;112:511–522. 131. Mishalani SH, Lones MA, Said JW. Multilocular thymic cyst.
105. Ahn JM, Lee KS, Goo JM, et al. Predicting the histology of ante- A novel thymic lesion associated with human immuno-
rior mediastinal masses: comparison of chest radiography and deficiency virus infection. Arch Pathol Lab Med. 1995;119:
CT. J Thorac Imaging. 1996;11:265–271. 467–470.
106. Martin KW, McAliste WH. Case report. Intratracheal thymus: a 132. Goldberg RE, Haaga JR, Yulish BS. Case report. Serial CT scans
rare cause of airway obstruction. AJR Am J Roentgenol. 1987;149: in thymic hyperplasia. J Comput Assist Tomogr. 1987;11:539–540.
1217–1218. 133. Castleman B, Norris EH. The pathology of the thymus in myas-
107. Goldstein G, Mackey IR. The human thymus. St. Louis: Warren thenia gravis. Medicine. 1949;28:27–58.
H. Green; 1969. 134. Brown LR, Muhm JR, Sheedy PF, et al. The value of computed
108. Perlo VP, Arnason B, Castleman B. The thymus gland in elderly tomography in myasthenia gravis. AJR Am J Roentgenol. 1983;
patients with myasthenia gravis. Neurology. 1975;25:294–295. 140:31–35.
109. Dixon AK, Hilton CJ, Williams GT. Computed tomography and 135. Cohen M, Hill CA, Cangir A, Sullivan MP. Thymic rebound after
histologic correlation of the thymic remnant. Clin Radiol. 1981; treatment of childhood tumors. AJR Am J Roentgenol. 1980;
32:255–257. 135:151–156.
110. Day DL, Gedgaudas E. The thymus. Radiol Clin North Am. 1984; 136. Choyke PL, Zeman RK, Gootenberg JE, et al. Thymic atrophy
22:519–538. and regrowth in response to chemotherapy: CT evaluation. AJR
111. Cory DA, Cohen MD, Smith JA. Thymus in the superior medi- Am J Roentgenol. 1987;149:269–272.
astinum simulating adenopathy: appearance on CT. Radiology. 137. Doppman JL, Oldfield EH, Chrousos GP. Rebound thymic
1987;162:457–459. hyperplasia after treatment of Cushing’s syndrome. AJR Am J
112. Salonen OLM, Kivisaari ML, Somer JK. Computed tomography Roentgenol. 1986;147:1145–1147.
of the thymus of children under 10 years. Pediatr Radiol. 138. Kissin CM, Husband JE, Nicholas D, Eversman W. Benign
1984;14:373–375. thymic enlargement in adults after chemotherapy: CT demon-
113. Rollins NK, Currarino G. Case report. MR imaging of poste- stration. Radiology. 1987;163:67–70.
rior mediastinal thymus. J Comput Assist Tomogr. 1988;12: 139. Hara M, McAdams HP, Vredenburgh JJ, et al. Thymic hyperpla-
518–520. sia after high-dose chemotherapy and autologous stem cell
transplantation: incidence and significance in patients with 164. Kondo K, Monden Y. Therapy for thymic epithelial tumors:
breast cancer. AJR Am J Roentgenol. 1999;173:1341–1344. a clinical study of 1,320 patients from Japan. Ann Thorac Surg.
140. Morgenthaler TI, Brown LR, Colby TV, et al. Thymoma. Mayo 2003;76:878–884; discussion 884–875.
Clin Proc. 1993;68:1110–1123. 165. Mink JH, Bein ME, Sukov R, et al. Computed tomography of the
141. LeGolvan DP, Abell MR. Thymomas. Cancer. 1977;39:2142–2157. anterior mediastinum in patients with myasthenia gravis and
142. Bergh N, Gatzinsky P, Larson S, et al. Tumors of the thymus and suspected thymoma. AJR Am J Roentgenol. 1978;130:239–246.
thymic region: 1. clinicopathological studies on thymomas. Ann 166. McLoud TC, Wittenberg J, Ferrucci JT. Computed tomography of
Thorac Surg. 1978;25:91–98. the thorax and standard radiographic evaluation of the chest:
143. Jung KJ, Lee KS, Han J, et al. Malignant thymic epithelial A comparative study. J Comput Assist Tomogr. 1979;3:170–180.
tumors: CT-pathologic correlation. AJR Am J Roentgenol. 2001; 167. Aita JF, Wannamaker WM. Body computerized tomography and
176:433–439. the thymus. Arch Neurol. 1979;36:20–21.
144. Suster S, Moran CA. Thymoma, atypical thymoma, and thymic 168. Baron RL, Lee JKT, Sagel SS, Levitt RG. Computed tomography
carcinoma. A novel conceptual approach to the classification of the abnormal thymus. Radiology. 1982;142:127–134.
of thymic epithelial neoplasms. Am J Clin Pathol. 1999;111: 169. Keesey J, Bein M, Mink J. Detection of thymoma in myasthenia
826–833. gravis. Neurology. 1980;30:233–239.
145. Marino M, Muller-Hermelink HK. Thymoma and thymic carci- 170. Fon GT, Bein ME, Mancuso AA, et al. Computed tomography of
noma. Relation of thymoma epithelial cells to the cortical and the anterior mediastinum in myasthenia gravis. A radiologic-
medullary differentiation of the thymus. Virchows Arch. 1985; pathologic correlative study. Radiology. 1982;142:135–141.
407:119–149. 171. Tomiyama N, Johkoh T, Mihara N, et al. Using the World
146. Lardinois D, Rechsteiner R, Lang RH, et al. Prognostic relevance Health Organization Classification of thymic epithelial neo-
of Masaoka and Muller-Hermelink classification in patients plasms to describe CT findings. AJR Am J Roentgenol. 2002;
with thymic tumors. Ann Thorac Surg. 2000;69:1550–1555. 179:881–886.
147. Ricci C, Rendina EA, Pescarmona EO, et al. Correlations 172. Han J, Lee KS, Yi CA, et al. Thymic epithelial tumors classified
between histological type, clinical behaviour, and prognosis in according to a newly established WHO scheme: CT and MR
thymoma. Thorax. 1989;44:455–460. findings. Korean J Radiol. 2003;4:46–53.
148. Pescarmona E, Rendina EA, Venuta F, Ricci C, Ruco LP, Baroni 173. Ellis K, Austin JHM, Jaretzki A. Radiologic detection of thymoma
CD. The prognostic implication of thymoma histologic subtyp- in patients with myasthenia gravis. AJR Am J Roentgenol. 1988;
ing: a study of 80 consecutive cases. Am J Clin Pathol. 1990;93: 151:873–881.
190–195. 174. Miller WT Jr. Thymoma mimicking a thyroid mass. Radiology.
149. Okumura M, Ohta M, Tateyama H, et al. The World Health 1992;184:75–76.
Organization histologic classification system reflects the onco- 175. Kaye AD, Janssen R, Arger PH, et al. Mediastinal computed
logic behavior of thymoma: a clinical study of 273 patients. tomography in myasthenia gravis. J Comp Assist Tomogr. 1983;7:
Cancer. 2002;94:624–632. 273–279.
150. Rea F, Marulli G, Girardi R, et al. Long-term survival and prog- 176. Kushihashi T, Fujisawa H, Munechika H. Magnetic resonance
nostic factors in thymic epithelial tumours. Eur J Cardiothorac imaging of thymic epithelial tumors. Crit Rev Diagn Imaging.
Surg. 2004;26:412–418. 1996;37:191–259.
151. Kim DJ, Yang WI, Choi SS, et al. Prognostic and clinical 177. Endo M, Adachi S, Kusumoto M, et al. A study of the utility of
relevance of the world health organization schema for the the MR image for the diagnosis of thymic tumors: imaging and
classification of thymic epithelial tumors: a clinicopathologic pathologic correlation [Japanese]. Nippon Igaku Hoshasen Gakkai
study of 108 patients and literature review. Chest. 2005; Zasshi. 1993;53:1–10.
127:755–761. 178. Sakai F, Sone S, Kiyono K, et al. MR imaging of thymoma:
152. Masaoka A, Monden Y, Nakahara K, Tanioka T. Follow-up study radiologic-pathologic correlation. AJR Am J Roentgenol. 1992;
of thymomas with special reference to their clinical stages. 158:751–756.
Cancer. 1981;48:2485–2492. 179. Sakai S, Murayama S, Soeda H, et al. Differential diagnosis
153. Zerhouni EA, Scott WW, Baker RR, et al. Invasive thymomas: between thymoma and non-thymoma by dynamic MR imaging.
diagnosis and evaluation by computed tomography. J Comput Acta Radiol. 2002;43:262–268.
Assist Tomogr. 1982;6:92–100. 180. Jaretzki A, Penn AS, Younger DS, et al. “Maximal” thymectomy
154. Scatarige JC, Fishman EK, Zerhouni EA, Siegelman SS. for myasthenia gravis. J Thorac Cardiovasc Surg. 1988;
Transdiaphragmatic extension of invasive thymoma. AJR Am J 95:747–757.
Roentgenol. 1985;144:31–35. 181. Pirronti T, Rinaldi P, Batocchi AP, et al. Thymic lesions and
155. Kaplan I, Swayne LC, Widmann WD, Wolff M. Case report. CT myasthenia gravis. Diagnosis based on mediastinal imaging and
demonstration of “ectopic” thymoma. J Comput Assist Tomogr. pathological findings. Acta Radiol. 2002;43:380–384.
1988;12:1037–1038. 182. Batra P, Herrmann C, Mulder D. Mediastinal imaging in myas-
156. Korobkin M, Casano VA. Case report. Intracaval and intracar- thenia gravis: correlation of chest radiography, CT, MR, and
diac extension of malignant thymoma: CT diagnosis. J Comput surgical findings. AJR Am J Roentgenol. 1987;148:515–519.
Assist Tomogr. 1989;13:348–350. 183. Quagliano PV. Thymic carcinoma: case reports and review.
157. Asamura H, Morinaga S, Shimosato Y, et al. Thymoma display- J Thorac Imaging. 1996;11:66–74.
ing endobronchial polypoid growth. Chest. 1988;94:647–649. 184. Hartman CA, Roth C, Minck C, Niedobitek G. Thymic carci-
158. Tomiyama N, Muller NL, Ellis SJ, et al. Invasive and noninvasive noma: report of five cases and review of the literature. J Cancer
thymoma: distinctive CT features. J Comput Assist Tomogr. 2001; Res Clin Oncol. 1990;116:69–82.
25:388–393. 185. Chung DA. Thymic carcinoma—analysis of nineteen clinico-
159. Maggi G, Giaccone G, Donadio M, et al. Thymomas. A review of pathological studies. Thorac Cardiovasc Surg. 2000;48:114–119.
169 cases, with particular reference to results of surgical treat- 186. Do YS, Im JG, Lee BH, et al. CT findings in malignant tumors of
ment. Cancer. 1986;58:756–776. thymic epithelium. J Comput Assist Tomogr. 1995;19:192–197.
160. Fujimura S, Kondo T, Handa M, et al. Results of surgical treat- 187. Sasaki M, Kuwabara Y, Ichiya Y, et al. Differential diagnosis of
ment for thymoma based on 66 patients. J Thorac Cardiovasc thymic tumors using a combination of 11C-methionine PET
Surg. 1987;93:708–714. and FDG PET. J Nucl Med. 1999;40:1595–1601.
161. Krueger JB, Sagerman RH, King GA. Stage III thymoma: results 188. Doppman JL, Pass HI, Nieman LK, et al. Corticotropin-secreting
of postoperative radiation therapy. Radiology. 1988;168: carcinoid tumors of the thymus: diagnostic unreliability of
855–858. thymic venous sampling. Radiology. 1992;184:71–74.
162. Dziuba SJ, Curran WJ Jr. The radiotherapeutic management of 189. Chaer R, Massad MG, Evans A, et al. Primary neuroendocrine
invasive thymomas. Chest Surg Clin North Am. 2001;11:457–466. tumors of the thymus. Ann Thorac Surg. 2002;74:1733–1740.
163. Froudarakis ME, Tiffet O, Fournel P, et al. Invasive thymoma: a 190. Fujiwara K, Segawa Y, Takigawa N, et al. Two cases of atypical
clinical study of 23 cases. Respiration. 2001;68:376–381. carcinoid of the thymus. Intern Med. 2000;39:834–838.
191. Gal AA, Kornstein MJ, Cohen C, et al. Neuroendocrine tumors ocarcinoma, and combined nonteratomatous germ cell tumors
of the thymus: a clinicopathological and prognostic study. Ann of the mediastinum—a clinicopathologic and immunohisto-
Thorac Surg. 2001;72:1179–1182. chemical study of 64 cases. Cancer. 1997;80:699–707.
192. Doppman JL, Nieman L, Miller DL, et al. Ectopic adrenocorti- 219. Moran CA, Suster S. Primary germ cell tumors of the medi-
cotropic hormone syndrome: localization studies in 28 patients. astinum: I. Analysis of 322 cases with special emphasis on
Radiology. 1989;172:115–124. teratomatous lesions and a proposal for histopathologic classi-
193. Ferolla P, Falchetti A, Filosso P, et al. Thymic neuroendocrine fication and clinical staging. Cancer. 1997;80:681–690.
carcinoma (carcinoid) in MEN1 syndrome: the Italian series. 220. Knapp RH, Hurt RD, Payne WS, et al. Malignant germ cell tumors
J Clin Endocrinol Metab. 2005;90:2603–2609. of the mediastinum. J Thorac Cardiovasc Surg. 1985;89:82–89.
194. Tiffet O, Nicholson AG, Ladas G, et al. A clinicopathologic study 221. Dobranowski J, Martin LFW, Bennett WF. Case report. CT evalu-
of 12 neuroendocrine tumors arising in the thymus. Chest. ation of posterior mediastinal teratoma. J Comput Assist Tomogr.
2003;124:141–146. 1987;11:156–157.
195. Groves AM, Mohan HK, Wegner EA, Hain SF, Bingham JB, 222. Weinberg B, Rose JS, Efremidis SC, et al. Posterior mediastinal
Clarke SE. Positron emission tomography with FDG to show teratoma (cystic dermoid): diagnosis by computerized tomogra-
thymic carcinoid. AJR Am J Roentgenol. 2004;182:511–513. phy. Chest. 1980;77:694–695.
196. Rosado-de-Christenson ML, Pugatch RD, Moran CA, Galobardes 223. Moeller KH, Rosado-de-Christenson ML, Templeton PA.
J. Thymolipoma: analysis of 27 cases. Radiology. 1994;193: Mediastinal mature teratoma: imaging features. AJR Am J
121–126. Roentgenol. 1997;169:985–990.
197. Cekirdekci A, Akpolat N, Ayan E, Duran M. Giant thymolipoma 224. Kurosaki Y, Tanaka YO, Itai Y. Mature teratoma of the posterior
in association with myasthenia gravis. Eur J Cardiothorac Surg. mediastinum. Eur Radiol. 1998;8:100–102.
2003;23:422. 225. Sinclair DS, Bolen MA, King MA. Mature teratoma within the
198. Hirai S, Hamanaka Y, Mitsui N, Kumagai H, Kobayashi T. Gigan- posterior mediastinum. J Thorac Imaging. 2003;18:53–55.
tic thymolipoma. Jpn J Thorac Cardiovasc Surg. 2002;50:40–42. 226. Suzuki M, Takashima T, Itoh H, et al. Computed tomography of
199. Herrera L, Oz M, Lally J, Davies A. Thymolipoma simulating mediastinal teratomas. J Comput Assist Tomogr. 1983;7:74–76.
pulmonary sequestration. J Pediatr Surg. 1982;17:313–315. 227. Drevelegas A, Palladas P, Scordalaki A. Mediastinal germ cell
200. Shirkhoda A, Chasen MH, Eftekhari F, et al. MR imaging of tumors: a radiologic-pathologic review. Eur Radiol. 2001;11:
mediastinal thymolipoma. J Comput Assist Tomogr. 1987;11: 1925–1932.
364–365. 228. Strollo DC, Rosado-de-Christenson ML. Primary mediastinal
201. Yeh HC, Gordon A, Kirshner PA, Cohen BA. Computed tomog- malignant germ cell neoplasms: imaging features. Chest Surg
raphy and sonography of thymolipoma. AJR Am J Roentgenol. Clin North Am. 2002;12:645–658.
1983;140:1131–1133. 229. Fulcher AS, Proto AV, Jolles H. Cystic teratoma of the medi-
202. Baron RL, Sagel SS, Baglan RJ. Thymic cysts following radiation astinum: demonstration of fat-fluid level. AJR Am J Roentgenol.
therapy for Hodgkin disease. Radiology. 1981;141:593–597. 1990;154:259–260.
203. Lewis CR, Manoharan A. Benign thymic cysts in Hodgkin’s 230. Seltzer SE, Herman PG, Sagel SS. Differential diagnosis of medi-
disease: report of a case and review of published cases. Thorax. astinal fluid levels visualized on computed tomography.
1987;42:633–634. J Comput Assist Tomogr. 1984;8:244–246.
204. Lindfors KK, Meyer JE, Dedrick CG, et al. Thymic cysts in medi- 231. Wu TT, Wang HC, Chang YC, et al. Mature mediastinal teratoma:
astinal Hodgkin disease. Radiology. 1985;156:37–41. sonographic imaging patterns and pathologic correlation.
205. Gouliamos A, Striggaris K, Lolas C, et al. Case report. Thymic J Ultrasound Med. 2002;21:759–765.
cyst. J Comput Assist Tomogr. 1982;6:172–174. 232. Sasaka K, Kurihara Y, Nakajima Y, et al. Spontaneous rupture: a
206. Levine C. Case report. Cervical presentation of a large thymic complication of benign mature teratomas of the mediastinum.
cyst: CT appearance. J Comput Assist Tomogr. 1988;12:656–657. AJR Am J Roentgenol. 1998;170:323–328.
207. Jaramillo D, Perez-Atayde A, Griscom NT. Apparent association 233. Yeoman LJ, Dalton HR, Adam EJ. Case report. Fat-fluid level in
between thymic cysts and prior thoracotomy. Radiology. pleural effusion as a complication of a mediastinal dermoid: CT
1989;172:207–209. characteristics. J Comput Assist Tomogr. 1990;14:307–309.
208. De Caluwe D, Ahmed M, Puri P. Cervical thymic cysts. Pediatr 234. Afifi HY, Bosl GJ, Burt ME. Mediastinal growing teratoma
Surg Int. 2002;18:477–479. syndrome. Ann Thorac Surg. 1997;64:359–362.
209. Heron CW, Husband JE, Williams MP. Hodgkin disease: CT of 235. Polansky SM, Barwick KW, Ravin CE. Primary mediastinal semi-
the thymus. Radiology. 1988;167:647–651. noma. AJR Am J Roentgenol. 1979;132:17–21.
210. Luker G, Siegel MJ. Mediastinal Hodgkin disease in children: 236. Bokemeyer C, Nichols CR, Droz JP, et al. Extragonadal germ cell
response to therapy. Radiology. 1993;189:737–740. tumors of the mediastinum and retroperitoneum: results from
211. Wernecke K, Vassallo P, Rutsch F, et al. Thymic involvement in an international analysis. J Clin Oncol. 2002;20:1864–1873.
Hodgkin disease: CT and sonographic findings. Radiology. 237. Shin MS, Ho KJ. Computed tomography of primary mediastinal
1991;181:375–383. seminomas. J Comput Assist Tomogr. 1983;7:990–994.
212. Lagrange W, Dahm HM, Karstens J, et al. Melanocystic neuroen- 238. Levitt RG, Husband JE, Glazer HS. CT of primary germ-cell
docrine carcinoma of the thymus. Cancer. 1987;59:484–488. tumors of the mediastinum. AJR Am J Roentgenol. 1984;142:
213. Davis RD, Oldham HN, Sabiston DC. Primary cysts and 73–78.
neoplasms of the mediastinum: recent changes in clinical 239. Blomlie V, Lien HH, Fossa SD, et al. Computed tomography in
presentatin, methods of diagnosis, management, and results. primary non-seminomatous germ cell tumors of the medi-
Ann Thorac Surg. 1987;44:229–237. astinum. Acta Radiol. 1988;29:289–292.
214. Lee KS, Im JG, Han C, et al. Malignant primary germ cell tumors 240. Fox MA, Vix VA. Endodermal sinus (yolk sac) tumors of
of the mediastinum: CT features. AJR Am J Roentgenol. 1989; the anterior mediastinum. AJR Am J Roentgenol. 1980;135:
153:947–951. 291–294.
215. Rosado-de-Christenson ML, Templeton PA, Moran CA. Medi- 241. Park HM, Tarver RD, Siddiqui AR, et al. Efficacy of thyroid
astinal germ-cell tumors: radiologic and pathologic correlation. scintigraphy in the diagnosis of intrathoracic goiter. AJR Am J
Radiographics. 1992;12:1013–1030. Roentgenol. 1987;148:527–529.
216. Takeda S, Miyoshi S, Ohta M, et al. Primary germ cell tumors in 242. Weber AL, Randolph G, Aksoy FG. The thyroid and parathyroid
the mediastinum: a 50-year experience at a single Japanese insti- glands. CT and MR imaging and correlation with pathology and
tution. Cancer. 2003;97:367–376. clinical findings. Radiol Clin North Am. 2000;38:1105–1129.
217. Moran CA, Suster S, Przygodzki RM, Koss MN. Primary germ 243. Glazer GM, Axel L, Moss AA. CT diagnosis of mediastinal
cell tumors of the mediastinum: II. Mediastinal seminomas— thyroid. AJR Am J Roentgenol. 1982;138:495–498.
a clinicopathologic and immunohistochemical study of 120 244. Buckley JA, Stark P. Intrathoracic mediastinal thyroid goiter:
cases. Cancer. 1997;80:691–698. imaging manifestations. AJR Am J Roentgenol. 1999;173: 471–475.
218. Moran CA, Suster S, Koss MN. Primary germ cell tumors of the 245. Shahian DM, Rossi R. Posterior mediastinal goiter. Chest.
mediastinum: III. Yolk sac tumor, embryonal carcinoma, chori- 1988;94:599–602.
246. Sekiya T, Tada S, Kawakami K, et al. Clinical application of com- 273. Auffermann W, Clark OH, Thurnher S, et al. Recurrent thyroid
puted tomography to thyroid disease. Comput Tomogr. 1979;3: carcinoma: characteristics on MR images. Radiology. 1988;168:
185–193. 753–757.
247. Machida K, Yoshikawa K. Case report. Aberrant thyroid gland 274. Norris EH. The parathyroid adenoma: a study of 322 cases. Int
demonstrated by computed tomography. J Comput Assist Tomogr. Abst Surg. 1947;84:1–41.
1979;3:689–690. 275. Clark OH. Mediastinal parathyroid tumors. Arch Surg. 1988;
248. Silverman PM, Newman GE, Korobkin M, et al. Computed 123:1096–1100.
tomography in the evaluation of thyroid disease. AJR Am J 276. Kang YS, Rosen K, Clark OH, Higgins CB. Localization of abnor-
Roentgenol. 1984;141:897–902. mal parathyroid glands of the mediastinum with MR imaging.
249. Bashist B, Ellis K, Gold RP. Computed tomography of intratho- Radiology. 1993;189:137–141.
racic goiters. AJR Am J Roentgenol. 1983;140:455–460. 277. Doppman JL, Brennan MF, Koehler JO, Marx SJ. Computed
250. Binder RE, Pugatch RD, J FL, et al. Case report. Diagnosis of tomography for parathyroid localization. J Comput Assist Tomogr.
posterior mediastinal goiter by computed tomography. J Comput 1977;1:30–36.
Assist Tomogr. 1980;4:550–552. 278. Shimshak RR, Sundaram M, Eddelston B, Prendergast J.
251. Morris UL, Colletti PM, Ralls PW, et al. Case report. CT demon- Diagnosis of parathyroid adenoma by computed tomography.
stration of intrathoracic thyroid tissue. J Comput Assist Tomogr. J Comput Assist Tomogr. 1979;3:117–119.
1982;6:821–824. 279. Whitley NO, Bohlman M, Connor TB, et al. Computed tomog-
252. Radecki PD, Arger PH, Arenson RL, et al. Thyroid imaging: com- raphy for localization of parathyroid adenomas. J Comput Assist
parison of high-resolution real-time ultrasound and computed Tomogr. 1981;5:812–817.
tomography. Radiology. 1984;153:145–147. 280. Wolverson MK, Sundaram M, Eddelston B, Prendergast J.
253. Stark DD, Clark OH, Gooding GAW, Moss AA. High-resolution Diagnosis of parathyroid adenoma by computed tomography.
ultrasonography and computed tomography of thyroid lesions J Comput Assist Tomogr. 1981;5:818–821.
in patients with hyperparathyroidism. Surgery. 1983;94: 281. Krudy AG, Doppman JL, Brennan MF, et al. The detection of
863–868. mediastinal parathyroid glands by computed tomography,
254. Takashima S, Ikezoe J, Morimoto S, et al. Primary thyroid selective arteriography, and venous sampling. Radiology. 1981;
lymphoma: evaluation with CT. Radiology. 1988;168:765–768. 140:739–744.
255. Takashima S, Morimoto S, Ikezoe J, et al. Primary thyroid lym- 282. Adams JE, Adams PH, Mantora H. Computed tomography in
phoma: comparison of CT and US assessment. Radiology. localization of parathyroid tumors. Clin Radiol. 1981;32:251–254.
1989;171:439–443. 283. Ovenfors CO, Stark D, Moss A, et al. Localization of parathyroid
256. Katz JF, Kane RA, Reyes J, et al. Thyroid nodules: sonographic- adenoma by computed tomography. J Comput Assist Tomogr.
pathologic correlation. Radiology. 1984;151:741–745. 1982;6:1094–1098.
257. Kier R, Silverman PM, Korobkin M, et al. Case report. Malignant 284. Sommer B, Welter HF, Spelsberg F, et al. Computed tomography
teratoma of the thyroid in an adult: CT appearance. J Comput for localizing enlarged parathyroid glands in primary hyper-
Assist Tomogr. 1985;9:174–176. parathyroidism. J Comput Assist Tomogr. 1982;6:521–526.
258. deCertaines J, Herry JY, Lancien G, et al. Evaluation of human 285. Doppman JL, Krudy AG, Brennan MF, et al. CT appearance of
thyroid tumors by proton nuclear magnetic resonance. J Nucl enlarged parathyroid glands in the posterior superior medi-
Med. 1982;23:48–51. astinum. J Comput Assist Tomogr. 1982;6:1099–1102.
259. Tennvall J, Biorklund A, Moller T, et al. Studies of NMR- 286. Takagi H, Tominaga Y, Uchida K, et al. Preoperative diagnosis of
relaxation-times in malignant tumours and normal tissues of secondary hyperparathyroidism using computed tomography.
the human thyroid gland. Prog Nucl Med. 1984;8:142–148. J Comput Assist Tomogr. 1982;6:527–528.
260. Gotway MB, Higgins CB. MR imaging of the thyroid and 287. Krudy AG, Doppman JL, Shawker TH, et al. Hyperfunctioning
parathyroid glands. Magn Reson Imaging Clin North Am. 2000;8: cystic parathyroid glands: CT and sonographic findings. AJR Am
163–182, ix. J Roentgenol. 1984;142:175–178.
261. Noma S, Nishimura K, Togashi K, et al. Thyroid gland: MR 288. College D, Rohatgi RK. Mediastinal parathyroid cyst. J Comput
imaging. Radiology. 1987;164:495–499. Assist Tomogr. 1983;7:140–142.
262. Higgins CB, McNamara MT, Fisher MR, Clark OH. MR imaging 289. Lineaweaver W, Clore F, Mancuso A, et al. Calcified parathyroid
of the thyroid. AJR Am J Roentgenol. 1986;147:1255–1261. glands detected by computed tomography. J Comput Assist
263. Higgins CB, Auffermann W. MR imaging of thyroid and Tomogr. 1984;8:975–977.
parathyroid glands: a review of current status. AJR Am J 290. Krubsack AJ, Wilson SD, Lawson TL, et al. Prospective compari-
Roentgenol. 1988;151:1095–1106. son of radionuclide, computed tomographic, and sonographic
264. Gefter WB, Sprizter CE, Eisenberg B, et al. Thyroid imaging with localization of parathyroid tumors. World J Surg. 1986;10:
high-field strength surface-coil MR. Radiology. 1987;164: 579–585.
483–490. 291. Stark DD, Gooding GAW, Moss AA, et al. Parathyroid imaging:
265. Stark DD, Moss AA, Gamsu G, et al. Magnetic resonance imag- comparison of high-resolution CT and high-resolution sonogra-
ing of the neck, part 2: pathologic findings. Radiology. 1984;150: phy. AJR Am J Roentgenol. 1983;141:633–638.
455–461. 292. Cates JD, Thorsen K, Lawson TL, et al. CT evaluation of parathy-
266. Stark DD, Clark OH, Moss AA. Magnetic resonance imaging of roid adenomas: diagnostic criteria and pitfalls. J Comput Assist
the thyroid, thymus, and parathyroid glands. Surgery. 1984;96: Tomogr. 1988;12:626–629.
1083–1091. 293. Lee VS, Spritzer CE, Coleman RE, et al. The complementary
267. Mountz JM, Glazer GM, Dmuchowski C, Sisson JC. MR imaging roles of fast spin-echo MR imaging and double-phase 99mTC-
of the thyroid: comparison with scintigraphy in the normal and sestamibi scintigraphy for localization of hyperfunctioning
diseased gland. J Comput Assist Tomogr. 1987;11:612–619. parathyroid glands. AJR Am J Roentgenol. 1996;167:1555–1562.
268. Noma S, Kanaoka M, Minami S, et al. Thyroid masses: MR 294. Miller DL, Doppman JL, Shawker TH, et al. Localization of
imaging and pathologic correlation. Radiology. 1988;168: parathyroid adenomas in patients who have undergone surgery.
759–764. Part 1. Noninvasive imaging methods. Radiology. 1987;162:
269. Charkes ND, Maurer AH, Siegel JA, et al. MR imaging in thyroid 133–137.
disorders: correlation of signal intensity with Graves disease 295. Satava RM, Beahrs OH, Scholz DA. Success rate of cervical
activity. Radiology. 1987;164:491–494. exploration for hyperparathyroidism. Arch Surg. 1975;110:
270. von Schulthess GK, McMurdo K, Tscholakoff D, et al. 625–627.
Mediastinal masses: MR imaging. Radiology. 1986;158:289–296. 296. Perez-Monte JE, Brown ML, Shah AN, et al. Parathyroid adeno-
271. Gamsu G, Stark DD, Webb WR, et al. Magnetic resonance imag- mas: accurate detection and localization with Tc-99m sestamibi
ing of benign mediastinal masses. Radiology. 1984;151:709–713. SPECT. Radiology. 1996;201:85–91.
272. Stark DD, Moss AA, Gamsu G, Clark OH, et al. Magnetic reso- 297. Weber CJ, Vansant J, Alazraki N, et al. Value of technetium 99m
nance imaging of the neck, part 1: normal anatomy. Radiology. sestamibi iodine 123 imaging in reoperative parathyroid sur-
1984;150:447–454. gery. Surgery. 1993;114:1011–1018.
298. Levin KE, Gooding GAW, Okerlund MD. Localizing studies in 323. Platt JF, Glazer GM, Orringer MB, et al. Radiologic evaluation of
patients with persistent or recurrent hyperparathyroidism. the subcarinal lymph nodes: a comparative study. AJR Am J
Surgery. 1987;102:917–925. Roentgenol. 1988;151:279–282.
299. Kneeland JB, Krubsack AJ, Lawson TL, et al. Enlarged parathy- 324. American Thoracic Society. Medical section of the American
roid glands: high-resolution local coil MR imaging. Radiology. Lung Association. Clinical staging of primary lung cancer. Am
1987;162:143–146. Rev Respir Dis. 1983;127:659–664.
300. Peck WW, Higgins CB, Fisher MR, et al. Hyperparathyroidism: 325. Mountain CF, Dresler CM. Regional lymph node classification
comparison of MR imaging with radionuclide scanning. for lung cancer staging. Chest. 1997;111:1718–1723.
Radiology. 1987;163:415–420. 326. Ko JP, Drucker EA, Shepard JA, et al. CT depiction of regional
301. Spritzer CE, Gefter WB, Hamilton R, et al. Abnormal parathy- nodal stations for lung cancer staging. AJR Am J Roentgenol.
roid glands: high-resolution MR imaging. Radiology. 1987;162: 2000;174:775–782.
487–491. 327. Beck E, Beattie EJ. The lymph nodes in the mediastinum. J Int
302. Kier R, Blinder RA, Herfkins RJ, et al. MR imaging with surface Coll Surgeons. 1958;29:247–251.
coils in primary hyperparathyroidism. J Comput Assist Tomogr. 328. Glazer HS, Aronberg DJ, Sagel SS. Pictorial essay. Pitfalls in CT
1987;11:863–868. recognition of mediastinal lymphadenopathy. AJR Am J
303. Auffermann W, Gooding GAW, Okerlund MD, et al. Diagnosis Roentgenol. 1985;144:267–274.
of recurrent hyperparathyroidism: comparison of MR imaging 329. Mencini RA, Proto AV. The high left and main pulmonary arter-
and other imaging techniques. AJR Am J Roentgenol. 1988; ies: a CT pitfall. J Comput Assist Tomogr. 1982;6:452–459.
150:1027–1033. 330. Quint LE, Glazer GM, Orringer MB, et al. Mediastinal lymph
304. Kang EH, Schiebler ML, Gefter WB, Kressel HY. Case report. MR node detection and sizing at CT and autopsy. AJR Am J
demonstration of bilateral intrathyroidal parathyroid glands. Roentgenol. 1986;147:469–472.
J Comput Assist Tomogr. 1988;12:349–350. 331. Moak GD, Cockerill EM, Farver MO, et al. Computed tomogra-
305. Gotway MB, Leung JW, Gooding GA, et al. Hyperfunctioning phy vs standard radiology in the evaluation of mediastinal
parathyroid tissue: spectrum of appearances on noninvasive adenopathy. Chest. 1982;82:69–75.
imaging. AJR Am J Roentgenol. 2002;179:495–502. 332. Genereux GP, Howie JL. Normal mediastinal lymph node size
306. Gotway MB, Reddy GP, Webb WR, et al. Comparison between and number: CT and anatomic study. AJR Am J Roentgenol.
MR imaging and 99mTc MIBI scintigraphy in the evaluation 1984;142:1095–1100.
of recurrent of persistent hyperparathyroidism. Radiology. 2001; 333. Klein JS, Webb WR. The radiologic staging of lung cancer. J
218:783–790. Thorac Imaging. 1991;7:29–47.
307. Krudy AG, Doppman JL, Marx SJ, et al. Radiographic findings in 334. Daly BD, Faling LJ, Bite G, et al. Mediastinal lymph node evalu-
recurrent parathyroid carcinoma. Radiology. 1982;142:625–629. ation by computed tomography in lung cancer. an analysis of
308. Miller DL, Doppman JL, Chang R, et al. Angiographic ablation 345 patients grouped by TNM staging, tumor size, and tumor
of parathyroid adenomas: lessons from a 10-year experience. location. J Thorac Cardiovasc Surg. 1987;94:664–672.
Radiology. 1987;165:601–607. 335. Staples CA, Müller NL, Miller RR, et al. Mediastinal nodes in
309. Rouviere H. Anatomie des lymphatiques de l’homme. Paris: Masson bronchogenic carcinoma: comparison between CT and medi-
et Cie; 1932. astinoscopy. Radiology. 1988;167:367–372.
310. Buy JN, Ghossain MA, Poirson F, et al. Computed tomography 336. Lewis JW, Pearlberg JL, Beute GH, et al. Can computed tomogra-
of mediastinal lymph nodes in nonsmall cell lung cancer: a new phy of the chest stage lung cancer? yes and no. Ann Thorac Surg.
approach based on the lymphatic pathway of tumor spread. 1990;49:591–596.
J Comput Assist Tomogr. 1988;12:545–552. 337. Webb WR, Gatsonis C, Zerhouni EA, et al. CT and MR imaging
311. Goss CM, ed. Gray’s anatomy. Philadelphia: Lea and Febiger; in staging non-small cell bronchogenic carcinoma: report of
1966. the Radiologic Diagnostic Oncology Group. Radiology. 1991;
312. Jochelson MS, Balikian JP, Mauch P, Liebman H. Peri- and 178:705–713.
paracardial involvement in lymphoma: a radiographic study in 338. Mori K, Yokoi K, Saito Y, et al. Diagnosis of mediastinal lymph
11 cases. AJR Am J Roentgenol. 1983;140:483–488. node metastases in lung cancer. Jpn J Clin Oncol. 1992; 22:35–40.
313. Meyer JE, McLoud TC, Lindfors KK. CT demonstration of 339. McLoud TC, Bourgouin PM, Greenberg RW, et al. Bronchogenic
cardiophrenic angle lymphadenopathy in Hodgkin disease. carcinoma: analysis of staging in the mediastinum with CT by
J Comput Assist Tomogr. 1985;9:485–488. correlative lymph node mapping and sampling. Radiology.
314. Scatarige JC, Fishman EK, Zinreich ES, et al. Internal mammary 1992;182:319–323.
lymphadenopathy in breast carcinoma: CT appraisal of 340. Andonopoulos AP, Karadanas AH, Drosis AA, et al. CT evalua-
anatomic distribution. Radiology. 1988;167:89–91. tion of mediastinal lymph nodes in primary Sjögren syndrome.
315. Scatarige JC, Boxen I, Smathers RL. Internal mammary lym- J Comput Assist Tomogr. 1988;12:199–201.
phadenopathy: imaging of a vital lymphatic pathway in breast 341. Aberle DR, Gamsu G, Lynch D. Thoracic manifestations of
cancer. Radiographics. 1990;10:857–870. Wegener granulomatosis: diagnosis and course. Radiology.
316. Filly R, Blank N, Castellino R. Radiographic distribution of 1990;174:703–709.
intrathoracic disease in previously untreated patients with 342. Bergin C, Castellino RA. Mediastinal lymph node enlargement
Hodgkin’s disease and non-Hodgkin’s lymphoma. Radiology. on CT scans in patients with usual interstitial pneumonitis. AJR
1976;120:277. Am J Roentgenol. 1990;154:251–254.
317. Cho CS, Blank N, Castellino RA. CT evaluation of cardiophrenic 343. Ngom A, Dumont P, Diot P, Lemarie E. Benign mediastinal
angle lymph nodes in patients with malignant lymphoma. AJR lymphadenopathy in congestive heart failure. Chest. 2001;119:
Am J Roentgenol. 1984;143:719–721. 653–656.
318. Aronberg DJ, Peterson RR, Glazer HS, Sagel SS. Superior 344. Slanetz PJ, Truong M, Shepard JA, et al. Mediastinal lymphadenopa-
diaphragmatic lymph nodes: CT assessment. J Comput Assist thy and hazy mediastinal fat: new CT findings of congestive heart
Tomogr. 1986;10:937–941. failure. AJR Am J Roentgenol. 1998;171:1307–1309.
319. Sussman SK, Halvorsen RA, Silverman PM, Saeed M. Paracar- 345. Glazer HS, Aronberg DJ, Sagel SS, Friedman PJ. CT demonstra-
diac adenopathy: CT evaluation. AJR Am J Roentgenol. 1987; tion of calcified mediastinal lymph nodes: a guide to the new
149:29–34. ATS classification. AJR Am J Roentgenol. 1986;147:17–20.
320. Vock P, Hodler J. Cardiophrenic angle adenopathy: update of 346. Lautin EM, Rosenblatt M, Friedman AC, et al. Calcification in
causes and significance. Radiology. 1986;159:395–399. non-Hodgkin lymphoma occurring before therapy: identifica-
321. Schnyder PA, Gamsu G. CT of the pretracheal retrocaval space. tion of plain films and CT. AJR Am J Roentgenol. 1990;
AJR Am J Roentgenol. 1981;136:303–308. 155:739–740.
322. Müller NL, Webb WR, Gamsu G. Paratracheal lymphadenopa- 347. Mallens WMC, Nijhuis-Heddes JMA, Bakker W. Calcified lymph
thy: radiographic findings and correlation with CT. Radiology. node metastases in bronchioloalveolar carcinoma. Radiology.
1985;156:761–765. 1986;161:103–104.
348. Panicek DM, Harty MP, Scicutella CJ, Carsky EW. Calcification in ing in evaluating the extent of primary cancers of the bronchi].
untreated mediastinal lymphoma. Radiology. 1988;166:735–736. Bull Cancer (Paris). 1988;75:903–916.
349. Breatnach E, Myers JD, McElvein RB, Zorn GL. Roentgenogram 376. Musset D, Grenier P, Carette MF, et al. Primary lung cancer stag-
of the month. Unusual case of a calcified anterior mediastinal ing: prospective comparative study of MR imaging with CT.
mass. Chest. 1986;89:113–115. Radiology. 1986;160:607–611.
350. Gross BH, Schneider HJ, Proto AV. Eggshell calcification of lymph 377. Rholl KS, Levitt RG, Glazer HS. Magnetic resonance imaging
nodes: an update. AJR Am J Roentgenol. 1980;135:1265–1268. of fibrosing mediastinitis. AJR Am J Roentgenol. 1985;145:
351. Mesisel S, Rozenman J, Yellin A, et al. Castleman’s disease. An 255–259.
uncommon computed tomographic feature. Chest. 1988;93: 378. Farmer DW, Moore E, Amparo E, et al. Calcific fibrosing medias-
1306–1307. tinitis: demonstration of pulmonary vascular obstruction by
352. Stark P, Smith DC, Watkins GE, Chun KE. Primary intrathoracic magnetic resonance imaging. AJR Am J Roentgenol. 1984;143:
extraosseous osteogenic sarcoma: report of three cases. 1189–1191.
Radiology. 1990;174:725–726. 379. Dooms GC, Hricak H, Moseley ME, et al. Characterization of
353. Gladish GW, Sabloff BM, Munden RF, et al. Primary thoracic lymphadenopathy by magnetic resonance relaxation times:
sarcomas. Radiographics. 2002;22:621–637. preliminary results. Radiology. 1985;155:691–697.
354. Groskin SA, Massi AF, Randall PA. Calcified hilar and mediasti- 380. de Geer G, Webb WR, Sollitto R, Golden J. MR characteristics
nal lymph nodes in an AIDS patient with Pneumocystis carinii of benign lymph node enlargement in sarcoidosis and
infection. Radiology. 1990;175:345–346. Castleman’s disease. Eur J Radiol. 1986;6:145–148.
355. Reede DL, Bergeron RT. Cervical tuberculous adenitis: CT mani- 381. Glazer GM, Orringer MB, Chenevert TL, et al. Mediastinal
festations. Radiology. 1985;154:701–704. lymph nodes: relaxation time/pathologic correlation and impli-
356. Lee KS, Song KS, Lim TH, et al. Adult-onset pulmonary tubercu- cations in staging of lung cancer with MR imaging. Radiology.
losis: findings on chest radiographs and CT scans. AJR Am J 1988;168:429–431.
Roentgenol. 1993;160:753–758. 382. Takenaka D, Ohno Y, Hatabu H, et al. Differentiation of
357. Scatarige JC, Fishman EK, Kuhajda FP, et al. Low attenuation metastatic versus non-metastatic mediastinal lymph nodes in
nodal metastases in testicular carcinoma. J Comput Assist Tomogr. patients with non-small cell lung cancer using respiratory-
1983;7:682–687. triggered short inversion time inversion recovery (STIR) turbo
358. Yousem DM, Scatariage JC, Fishman EK, Siegelman SS. Low- spin-echo MR imaging. Eur J Radiol. 2002;44:216–224.
attenuation thoracic metastases in testicular malignancy. AJR 383. Ohno Y, Hatabu H, Takenaka D, et al. Metastases in mediasti-
Am J Roentgenol. 1986;146:291–293. nal and hilar lymph nodes in patients with non-small cell
359. Pombo F, Rodriquez E, Caruncho MV, et al. CT attenuation lung cancer: quantitative and qualitative assessment with STIR
values and enhancing characteristics of thoracoabdominal lym- turbo spin-echo MR imaging. Radiology. 2004;231:872–879.
phomatous adenopathy. J Comput Assist Tomogr. 1994;18:59–64. 384. Batra P, Brown K, Steckel RJ, et al. MR imaging of the thorax:
360. Hopper KD, Diehl LF, Cole BA, et al. The significance of necrotic a comparison of axial, coronal, and sagittal imaging planes.
mediastinal lymph nodes on CT in patients with newly diag- J Comput Assist Tomogr. 1988;12:75–81.
nosed Hodgkin disease. AJR Am J Roentgenol. 1990;155:267–270. 385. Webb WR, Jensen BG, Gamsu G, et al. Coronal magnetic reso-
361. Samuels T, Hamilton P, Shaw P. Case report. Whipple disease of nance imaging of the chest: normal and abnormal. Radiology.
the mediastinum. AJR Am J Roentgenol. 1990;154:1187–1188. 1984;153:729–735.
362. Shahidi H, Myers JL, Kvale PA. Castleman’s disease. Mayo Clin 386. Castellino RA. The non-Hodgkin lymphomas: practical concepts
Proc. 1995;70:969–977. for the diagnostic radiologist. Radiology. 1991;178:315–321.
363. Davis JM, Mark GJ, R G. Benign blood vascular tumors of the 387. Hoppe RT. The contemporary management of Hodgkin disease.
mediastinum. Radiology. 1978;126:581–587. Radiology. 1988;169:297–304.
364. Hamper UM, Fishman EK, Khouri NF, et al. Typical and atypical 388. Hopper KD, Diehl LF, Lesar M, et al. Hodgkin disease: clinical
CT manifestations of pulmonary sarcoidosis. J Comput Assist utility of CT in initial staging and treatment. Radiology.
Tomogr. 1986;10:928–936. 1988;169:17–22.
365. Kirks DR, McCormick VD, Greenspan RH. Pulmonary sarcoido- 389. Schomberg TJ, Evans RG, O’Connell MJ. Prognostic significance
sis: roentgenographic analysis of 150 patients. AJR Am J of mediastinal mass in adult Hodgkin’s disease. Cancer.
Roentgenol. 1973;117:777–786. 1984;53:324–328.
366. Kuhlman JE, Fishman EK, Hamper WM, et al. The computed 390. Kaplan HS. Hodgkin’s disease: unfolding concepts concerning
tomographic spectrum of thoracic sarcoidosis. Radiographics. its nature, management and prognosis. Cancer. 1980;45:2439.
1989;9:449–466. 391. Kadin ME, Glatstein EJ, Dorfman RE. Clinical pathologic stud-
367. Patil SN, Levin DL. Distribution of thoracic lymphadenopathy ies in 177 untreated patients subjected to a laparotomy for the
in sarcoidosis using computed tomography. J Thorac Imaging. staging of Hodgkin’s disease. Cancer. 1977;27:1277.
1999;14:114–117. 392. Castellino RA, Hoppe RT, Blank N. Computed tomography,
368. Castellino RA, Blank N, Hoppe RT, Cho C. Hodgkin disease: lymphography and staging laparotomy: correlations with initial
contributions of chest CT in the initial staging evaluation. staging of Hodgkin’s disease. AJR Am J Roentgenol. 1984;143:37.
Radiology. 1986;160:603–605. 393. Castellino RA, Hilton S, O’Brien JP, Portlock CS. Non-Hodgkin
369. Castellino RA. Hodgkin disease: practical concepts for the diag- lymphoma: contribution of chest CT in the initial staging evalu-
nostic radiologist. Radiology. 1986;157:305–310. ation. Radiology. 1996;199:129–132.
370. Castellino RA. Hodgkin disease: imaging studies and patient 394. Khoury MB, Godwin JD, Halvorsen R, et al. Role of chest CT in
management. Radiology. 1988;169:269–270. non-Hodgkin’s lymphoma. Radiology. 1986;158:659–662.
371. Frija J, Bellin MF, Laval-Jeantet M. CT mediastinum examination 395. Bragg DG, Colby TV, Ward JH. New concepts in the non-
in recurrent nerve paralysis. J Comput Assist Tomogr. 1984; Hodgkin lymphomas: radiologic implications. Radiology. 1986;
8:901–905. 159:289–304.
372. Dooms GC, Hricak H, Crooks LE, Higgins CB. Magnetic reso- 396. Wang Y. Classification of non-Hodgkin’s lymphoma. AJR Am J
nance imaging of the lymph nodes: comparison with CT. Roentgenol. 1986;147:205–208.
Radiology. 1984;153:719–728. 397. Harris NL, Jaffe ES, Stein H, et al. A revised European-American
373. Dubray B, Grenier P, Carette MF, et al. [Mediastinal lymph node classification of lymphoid neoplasms: a proposal from the
involvement in bronchogenic carcinoma. CT and MRI evalua- International Lymphoma Study Group. Blood. 1994;84:
tion using ROC methodology]. Rev Im Med. 1991;3:249–256. 1361–1392.
374. Poon PY, Bronskill MJ, Henkelman RM, et al. Mediastinal 398. Zinzani PL. Lymphoma: Diagnosis, staging, natural history, and
lymph node metastases from bronchogenic carcinoma: detec- treatment strategies. Semin Oncol. 2005;32:4–10.
tion with MR imaging and CT. Radiology. 1987;162:651–656. 399. Harris NL, Jaffe ES, Diebold J, et al. The World Health
375. Laurent F, Drouillard J, Dorcier F, et al. [Contribution and limits Organization classification of neoplastic diseases of the
of x-ray computed tomography and magnetic resonance imag- hematopoietic and lymphoid tissues. Report of the Clinical
Advisory Committee meeting, Airlie House, Virginia, 424. Nyman R, Rehn S, Glimelius B, et al. Magnetic resonance imag-
November, 1997. Ann Oncol. 1999;10:1419–1432. ing for assessment of treatment effects in mediastinal Hodgkin’s
400. Harris NL, Jaffe ES, Diebold J, et al. Lymphoma classification— disease. Acta Radiol. 1987;28:145–151.
from controversy to consensus: the R.E.A.L. and WHO 425. Glazer HS, Lee JK, Levitt RG, et al. Radiation fibrosis: differenti-
Classification of lymphoid neoplasms. Ann Oncol. 2000; ation from recurrent tumor by MR imaging. Radiology. 1985;
11(suppl 1):3–10. 156:721–726.
401. Kazerooni EA, Williams DM, Deeb GM. Thoracic periaortic 426. Zerhouni EA, Fishman EK, Jones R, et al. MR imaging of steril-
lymphoma mimicking aortic dissection. AJR Am J Roentgenol. ized lymphoma (abstr). Radiology. 1986;161(P):207.
1992;159:705–707. 427. Rahmouni AD, Zerhouni EA, eds. Role of MRI in the manage-
402. Fishman EK, Kuhlman JE, Jones RJ. CT of lymphoma: spectrum ment of thoracic lymphoma. New York: Churchill-Livingstone;
of disease. Radiographics. 1991;11:647–669. 1990.
403. Neuman CH, Robert NJ, Canellos G, Rosenthal D. Computed 428. Weller SA, Glatstein E, Kaplan HS, Rosenberg SA. Initial relapses
tomography of the abdomen and pelvis in non-Hodgkin in previously treated Hodgkin’s disease: results of second treat-
lymphoma. J Comput Assist Tomogr. 1983;7:846–850. ment. Cancer. 1976;37:2840–2846.
404. Tateishi U, Muller NL, Johkoh T, et al. Primary mediastinal 429. Schein PS, Chabner BA, Canellos GP, et al. Non-Hodgkin’s
lymphoma: characteristic features of the various histological lymphoma: patterns of relapse from complete remission after
subtypes on CT. J Comput Assist Tomogr. 2004;28:782–789. combination chemotherapy. Cancer. 1975;35:334–357.
405. Shaffer K, Smith D, Kirn D, et al. Primary mediastinal large- 430. Rostock RA, Giangreco A, Wharam MD, et al. CT scan modifica-
B-cell lymphoma: radiologic findings at presentation. AJR Am J tion in the treatment of mediastinal Hodgkin’s disease. Cancer.
Roentgenol. 1996;167:425–430. 1982;49:2267–2275.
406. Smith D, Shaffer K, Kirn D, et al. Mediastinal large cell lym- 431. Tumeh SS, Rosenthal DS, Kaplan WD, et al. Lymphoma: evalua-
phoma: prognostic significance of CT findings at presentation tion with Ga-67 SPECT. Radiology. 1987;164:111–114.
and after treatment. Oncology. 1998;55:284–288. 432. Israel O, Front D, Lam M, et al. Gallium 67 imaging in monitor-
407. Rahmouni A, Divine M, Lepage E, et al. Mediastinal lymphoma: ing lymphoma response to treatment. Cancer. 1988;61:
quantitative changes in gadolinium enhancement at MR imag- 2439–2443.
ing after treatment. Radiology. 2001;219:621–628. 433. Becherer A, Jaeger U, Szabo M, Kletter K. Prognostic value of
408. Rahmouni A, Tempany C, Jones R, et al. Lymphoma: monitor- FDG-PET in malignant lymphoma. Q J Nucl Med. 2003;47:
ing tumor size and signal intensity with MR imaging. Radiology. 14–21.
1993;188:445–451. 434. Kazama T, Faria SC, Varavithya V, et al. FDG PET in the evalua-
409. Rehn SM, Nyman RS, Glimelius BL, et al. Non-Hodgkin lym- tion of treatment for lymphoma: clinical usefulness and pitfalls.
phoma: predicting prognostic grade with MR imaging. Radiology. Radiographics. 2005;25:191–207.
1990;176:249–253. 435. Jerusalem G, Beguin Y, Fassotte MF, et al. Early detection of
410. Nyman RS, Rehn SM, Glimelius BLG, et al. Residual mediastinal relapse by whole-body positron emission tomography in the
masses in Hodgkin disease: prediction of size with MR imaging. follow-up of patients with Hodgkin’s disease. Ann Oncol.
Radiology. 1989;170:435–440. 2003;14:123–130.
411. North LB, Fuller LM, Sullivan-Halley J, Hagemeister FB. Regre- 436. Zinzani PL, Fanti S, Battista G, et al. Predictive role of positron
ssion of mediastinal Hodgkin disease after therapy: evaluation emission tomography (PET) in the outcome of lymphoma
of time interval. Radiology. 1987;164:599–602. patients. Br J Cancer. 2004;91:850–854.
412. Libshitz HJ, Jing BS, Walace S, Logothetis CJ. Sterilized metas- 437. Guay C, Lepine M, Verreault J, Benard F. Prognostic value of PET
tases: a diagnostic and therapeutic dilemma. AJR Am J Roentgenol. using 18F-FDG in Hodgkin’s disease for posttreatment evalua-
1983;140:14–19. tion. J Nucl Med. 2003;44:1225–1231.
413. Chen JL, Osborne BM, Butler JJ. Residual fibrous masses in 438. Schaefer NG, Hany TF, Taverna C, et al. Non-Hodgkin lym-
treated Hodgkin’s disease. Cancer. 1987;60:407–413. phoma and Hodgkin disease: coregistered FDG PET and CT at
414. Durkin W, Durant J. Benign mass lesions after therapy for staging and restaging—do we need contrast-enhanced CT?
Hodgkin’s disease. Arch Intern Med. 1979;139:333–336. Radiology. 2004;232:823–829.
415. Jochelson M, Mauch P, Balikian J, et al. The significance of the 439. Lavely WC, Delbeke D, Greer JP, et al. FDG PET in the follow-up
residual mediastinal mass in treated Hodgkin’s disease. J Clin management of patients with newly diagnosed Hodgkin and
Oncol. 1985;3:637–640. non-Hodgkin lymphoma after first-line chemotherapy. Int J
416. Radford JA, Cowan RA, Flanagan M, et al. The significance of Radiat Oncol Biol Phys. 2003;57:307–315.
residual mediastinal abnormality on the chest radiograph follow- 440. Maile CW, Moore AV, Ulreich S, Putman CE. Chest radiographic-
ing treatment for Hodgkin’s disease. J Clin Oncol. 1988;6: pathologic correlation in adult leukemic patients. Invest Radiol.
940–946. 1983;18:495–499.
417. Stewart FM, Williamson BR, Innes DJ, Hess CE. Residual tumor 441. Takasugi JE, Godwin JD, Marglin SI, Petersdorf SH. Intrathoracic
masses following treatment for advanced histiocytic lymphoma. granulocytic sarcomas. J Thorac Imaging. 1996;11:223–230.
Cancer. 1985;55:620–623. 442. Bragg DG, Chor PJ, Murray KA, Kjeldsberg CR. Lympho-
418. Lee CKK, Bloomfield CD, Goldman AI, Levitt SH. Progno- proliferative disorders of the lung: histopathology, clinical
stic significance of mediastinal involvement in Hodgkin’s manifestations, and Imaging features. AJR Am J Roentgenol.
disease treated with curative radiotherapy. Cancer. 1980;46: 1994;163:273–281.
2403–2409. 443. Keller AR, Hochholzer L, Castleman B. Hyaline-vascular and
419. Mauch P, Goodman R, Hellman S. The significance of mediasti- plasma-cell types of giant lymph node hyperplasia of the medi-
nal involvement in early stage Hodgkin’s disease. Cancer. 1978; astinum and other locations. Cancer. 1972;29:670–683.
42:1039–1045. 444. Menke DM, Camoriano JK, Banks PM. Angiofollicular lymph
420. North LB, Fuller LM, Hagemeister FB, et al. Importance of initial node hyperplasia: a comparison of unicentric, multicentric,
mediastinal adenopathy in Hodgkin’s disease. AJR Am J hyaline vascular, and plasma cell types of disease by morpho-
Roentgenol. 1982;138:229–235. metric and clinical analysis. Mod Pathol. 1992;5:525–530.
421. Anderson H, Jenkins JPR, Brigg DJ, et al. The prognostic signifi- 445. Fiore D, Biondetti PR, Calabro F, Rea F. Case report. CT demon-
cance of mediastinal bulk in patients with stage IA-IVB stration of bilateral Castleman’s tumors in the mediastinum.
Hodgkin’s disease: A report from the Manchester Lymphoma J Comput Assist Tomogr. 1983;7:719–720.
Group. Clin Radiol. 1985;36:449–454. 446. Ferreiros J, Leon NG, Mata MI, et al. Computed tomography in
422. Dryver ET, Jernstrom H, Tompkins K, et al. Follow-up of patients abdominal Castleman’s disease. J Comput Assist Tomogr. 1989;13:
with Hodgkin’s disease following curative treatment: the routine 433–436.
CT scan is of little value. Br J Cancer. 2003;89:482–486. 447. Meisel S, Rozenman J, Yellin A, et al. Castleman’s disease. an
423. Webb WR. MR imaging of treated mediastinal Hodgkin disease uncommon computed tomographic feature. Chest. 1988;93:
[editorial]. Radiology. 1989;170:315–316. 1306–1307.
448. Libson E, Fields S, Strauss S, et al. Widespread Castleman 474. Rodriguez E, Soler R, Pombo F, et al. Fibrosing mediastinitis: CT
disease: CT and US findings. Radiology. 1988;166:753–755. and MR findings. Clin Radiol. 1998;53:907–910.
449. Thompson GP, Utz JP, Rosenow EC, et al. Pulmonary lympho- 475. Imran MB, Kubota K, Yoshioka S, et al. Sclerosing mediastinitis:
proliferative disorders. Mayo Clin Proc. 1993;68:804–817. findings on fluorine-18 fluorodeoxyglucose positron emission
450. Harris KM, Schwartz ML, Slasky BS, et al. Posttransplanta- tomography. Clin Nucl Med. 1999;24:305–308.
tion cyclosporine-induced lymphoproliferative disorders: 476. Fitch SJ, Tonkin ILD, Tonkin AK. Imaging of foregut duplication
clinical and radiological manifestations. Radiology. 1987;162: cysts. Radiographics. 1986;6:189–201.
697–700. 477. Strollo DC, Rosado-de-Christenson ML, Jett JR. Primary medi-
451. McLoud TC, Kalisher L, Stark P, Greene R. Intrathoracic lymph astinal tumors: part II. Tumors of the middle and posterior
node metastases from extrathoracic neoplasms. AJR Am J mediastinum. Chest. 1997;112:1344–1357.
Roentgenol. 1978;131:403–407. 478. Takeda S, Miyoshi S, Minami M, et al. Clinical spectrum of
452. Müller NL, Kullnig P, Miller RR. The CT findings of pulmonary mediastinal cysts. Chest. 2003;124:125–132.
sarcoidosis: analysis of 25 patients. AJR Am J Roentgenol. 1989; 479. Eroglu A, Kurkcuoglu C, Karaoglanoglu N, et al. Primary
152:1179–1182. hydatid cysts of the mediastinum. Eur J Cardiothorac Surg.
453. Bein ME, Putman CE, McLoud TC, Mink JH. A reevaluation of 2002;22:599–601.
intrathoracic lymphadenopathy in sarcoidosis. AJR Am J 480. Purohit M. Primary hydatid cysts of the mediastinum. Eur J
Roentgenol. 1978;131:409–415. Cardiothorac Surg. 2003;23:257–258; author reply 258.
454. Niimi H, Kang EY, Kwong JS, et al. CT of chronic infiltrative 481. Panicek DM, Heitzman ER, Randall PA, et al. The continuum of
lung disease: prevalence of mediastinal lymphadenopathy. pulmonary developmental anomalies. Radiographics. 1987;7:
J Comput Assist Tomogr. 1996;20:305–308. 747–772.
455. Rockoff SD, Rohatgi PK. Unusual manifestations of thoracic 482. Pugatch RD, Faling LJ, Spira R. CT diagnosis of benign
sarcoidosis. AJR Am J Roentgenol. 1985;144:513–528. mediastinal abnormalities. AJR Am J Roentgenol. 1980;134:
456. Lee JY, Lee KS, Jung KJ, et al. Pulmonary tuberculosis: CT and 685–694.
pathologic correlation. J Comput Assist Tomogr. 2000;24: 483. Marvasti MA, Mitchell GE, Burke WA, Meyer JA. Misleading
691–698. density of mediastinal cysts on computerized tomography. Ann
457. Kushihashi T, Munechika H, Motoya H, et al. CT and MR find- Thorac Surg. 1981;31:167–170.
ings in tuberculous mediastinitis. J Comput Assist Tomogr. 1995; 484. Nakata H, Nakayama C, Kimoto T, et al. Computed tomogra-
19:379–382. phy of mediastinal bronchogenic cysts. J Comput Assist Tomogr.
458. Pastores SM, Naidich DP, Aranda CP, et al. Intrathoracic 1982;6:733–738.
adenopathy associated with pulmonary tuberculosis in patients 485. Mendelson DS, Rose JS, Efremidis SC, et al. Bronchogenic cysts
with human immunodeficiency virus infection. Chest. 1993; with high CT numbers. AJR Am J Roentgenol. 1983;140:
103:1433–1437. 463–465.
459. Naidich DP, Garay SM, Goodman PC, et al. Pulmonary mani- 486. Nakata H, Sato Y, Nakayama T, et al. Bronchogenic cyst with
festations of AIDS. In: Federle M, Megibow A, Naidich DP, eds. high CT number: analysis of contents. J Comput Assist Tomogr.
Radiology of acquired immune deficiency syndrome. New York: 1986;10:360–362.
Raven; 1988:47–77. 487. Yemault JC, Kuhn G, Dumortier P, et al. “Solid” mediastinal
460. Hartman TE, Primack SL, Müller NL, Staples CA. Diagnosis of bronchogenic cyst: mineralogic analysis. J Comput Assist Tomogr.
thoracic complications in AIDS: accuracy of CT. AJR Am J 1986;146:73–74.
Roentgenol. 1994;162:547–553. 488. Dahmash NS, Chen JTT, Ravin CE, et al. Unusual radiologic
461. Naidich DP, Tarras M, Garay SM, et al. Kaposi sarcoma: manifestations of bronchogenic cyst. South Med J. 1984;77:
CT-radiographic correlation. Chest. 1989;96:723–728. 762–764.
462. Moon WK, Im JG, In KY, et al. Mediastinal tuberculous 489. Reed JC, Sobonya RE. Morphologic analysis of foregut cysts in
lymphadenitis: MR imaging appearance with clinicopathologic the thorax. AJR Am J Roentgenol. 1974;120:851–860.
correlation. AJR Am J Roentgenol. 1996;166:21–25. 490. Volpi A, Cavalli A, Maggioni AP, Pieri-Nerli F. Left atrial com-
463. Landay MJ, Rollins NK. Mediastinal histoplasmosis granuloma: pression by a mediastinal bronchogenic cyst presenting with
evaluation with CT. Radiology. 1989;172:657–659. paroxysmal atrial fibrillation. Thorax. 1988;43:216–217.
464. McAdams HP, Rosado-de-Christenson ML, Lesar M, et al. Thoracic 491. Bankoff MS, Daly BDT, Johnson HA, Carter BL. Case report.
mycoses from endemic fungi: radiologic-pathologic correlation. Bronchogenic cyst causing superior vena cava obstruction: CT
Radiographics. 1995;15:255–270. appearance. J Comput Assist Tomogr. 1985;9:951–952.
465. McGahan JP, Graves DS, Palmer PE, et al. Classic and contem- 492. Kuhlman JE, Fishman EK, Wang KP, et al. Mediastinal cysts:
porary imaging of coccidioidomycosis. AJR Am J Roentgenol. diagnosis by CT and needle aspiration. AJR Am J Roentgenol.
1981;136:393–404. 1988;150:75–78.
466. Loyd JE, Tillman BF, Atkinson JB, Des Prez RM. Mediastinal 493. Haller JA, Golladay ES, Pickard LR, et al. Surgical management
fibrosis complicating histoplasmosis. Medicine. 1988;67: of lung bud anomalies: lobar emphysema, bronchogenic cyst,
295–310. cystic adenomatoid malformation, and intralobar pulmonary
467. Rossi SE, McAdams HP, Rosado-de-Christenson ML, Franks TJ, sequestration. Ann Thorac Surg. 1979;28:33–43.
Galvin JR. Fibrosing mediastinitis. Radiographics. 2001;21: 494. Schwartz DB, Beals TF, Wimbish KJ, Hammersley JR. Trans-
737–757. bronchial fine needle aspiration of bronchogenic cysts. Chest.
468. Wieder S, Rabinowitz JG. Fibrous mediastinitis: a late manifes- 1985;88:573–575.
tation of mediastinal histoplasmosis. Radiology. 1977;125: 495. Zimmer WD, Kamida CB, McGough PF, Rosenow EC. Medi-
305–312. astinal duplication cyst. Percutaneous aspiration and cystogra-
469. Weinstein JB, Aronberg DJ, Sagel SS. CT of fibrosing mediastini- phy for diagnosis and treatment. Chest. 1986;90:772–773.
tis: findings and their utility. AJR Am J Roentgenol. 1983;141: 496. Kuhlman JE, Fishman EK, Wang KP, Siegelman SS. Esophageal
247–251. duplication cysts: CT and transesophageal aspiration. AJR Am
470. Goodwin RA, Nickell JA, Des Pres RM. Mediastinal fibrosis J Roentgenol. 1985;145:531–532.
complicating healed primary histoplasmosis and tuberculosis. 497. Wittich GR, Karnel F, Schurawitzki H, Jantsch H. Percutaneous
Medicine. 1972;51:227–246. drainage of mediastinal pseudocysts. Radiology. 1988;167:51–53.
471. Kountz PD, Molina PL, Sagel SS. Fibrosing mediastinitis in the 498. Kawashima A, Fishman EK, Kuhlman JE, Nixon MS. CT
posterior thorax. AJR Am J Roentgenol. 1989;153:489–490. of posterior mediastinal masses. Radiographics. 1991;11:
472. Harman M, Sayarlioglu M, Arslan H, et al. Fibrosing mediastinitis 1045–1067.
and thrombosis of superior vena cava associated with Behçet’s 499. Weiss LM, Fagelman D, Warhit JM. CT demonstration of an eso-
disease. Eur J Radiol. 2003;48:209–212. phageal duplication cyst. J Comput Assist Tomogr. 1983; 7:716–718.
473. Sherrick AD, Brown LR, Harms GF, et al. Radiographic findings 500. Maroko I, Hirsch M, Sharon N, Benharroch D. Calcified medi-
of fibrosing mediastinitis. Chest. 1994;106:484–489. astinal enterogenous cyst. Gastrointest Radiol. 1984;9:105–106.
501. Rogers CI, Seymour Q, Brock JG. Case report. Atypical pericar- 530. Streiter ML, Schneider HJ, Proto AV. Steroid-induced thoracic
dial cyst location: the value of computed tomography. J Comput lipomatosis: paraspinal involvement. AJR Am J Roentgenol.
Assist Tomogr. 1980;4:683–684. 1982;139:679–681.
502. Patel BK, Markivee CR, George EA. Pericardial cyst simulating 531. Keely JG, Vana AJ. Lipomas of the mediastinum, 1940–1955. Int
intracardiac mass. AJR Am J Roentgenol. 1983;141:292–294. Abst Surg. 1956;103:312–322.
503. Kutlay H, Yavuzer I, Han S, Cangir AK. Atypically located peri- 532. Truwit JD, Jacobs JK, Newman JH, Dyer EL. Roentgenogram of
cardial cysts. Ann Thorac Surg. 2001;72:2137–2139. the month. Anterior mediastinal mass following pneumonec-
504. Pugatch RD, Braver JH, Robbins AH, Faling J. CT diagnosis of tomy. Chest. 1988;94:173–174.
pericardial cysts. AJR Am J Roentgenol. 1978;131:515–516. 533. Lee KH, Song KS, Kwon Y, et al. Mesenchymal tumours of the
505. Brunner DR, Whitley NO. A pericardial cyst with high CT thorax: CT findings and pathological features. Clin Radiol. 2003;
numbers. AJR Am J Roentgenol. 1984;142:279–280. 58:934–944.
506. Parienty RA, Fontaine Y, Guillemette D. Transformation of a 534. Jung JI, Kim H, Kang SW, Park SH. Radiological findings in myx-
pericardial cyst observed on long-term follow-up. CT. 1984;8: oid liposarcoma of the anterior mediastinum. Br J Radiol. 1998;
125–128. 71:975–976.
507. Avila NA, Kelly JA, Chu SC, et al. Lymphangioleiomyomatosis: 535. Klimstra DS, Moran CA, Perino G, et al. Liposarcoma of the
abdominopelvic CT and US findings. Radiology. 2000;216: anterior mediastinum and thymus. A clinicopathologic study of
147–153. 28 cases. Am J Surg Pathol. 1995;19:782–791.
508. Parasher VK, Meroni E, Malesci A, et al. Observation of thoracic 536. Santamaria G, Serres X, Pruna X, Forcada P. Primary mediastinal
duct morphology in portal hypertension by endoscopic ultra- liposarcoma with moderately high CT attenuation. AJR Am J
sound. Gastrointest Endosc. 1998;48:588–592. Roentgenol. 1996;167:1064–1065.
509. Pramesh CS, Deshpande MS, Pantvaidya GH, et al. Thoracic 537. Cohen WN, Seidelmann FE, Bryan PJ. Computed tomography
duct cyst of the mediastinum. Ann Thorac Cardiovasc Surg. 2003; of localized adipose deposits presenting as tumor masses. AJR
9:264–265. Am J Roentgenol. 1977;128:1007–1011.
510. Mattila PS, Tarkkanen J, Mattila S. Thoracic duct cyst: a case 538. Rubin E. Case of the winter season. Semin Roentgenol. 1978;
report and review of 29 cases. Ann Otol Rhinol Laryngol. 1999; 13:5–6.
108:505–508. 539. Schweitzer DL, Aguam AS. Primary liposarcoma of the medi-
511. Chen F, Bando T, Hanaoka N, et al. Mediastinal thoracic duct astinum. J Thorac Cardiovasc Surg. 1977;741:83–97.
cyst. Chest. 1999;115:584–585. 540. Yang R, Elliston L, Peterson R, Sahmel R. Roentgenogram of the
512. Gupta M, Lovelace TD, Sukumar M, Gosselin MV. Cervical month. Dysphagia and cough in a patient with a posterior
thoracic duct cyst. J Thorac Imaging. 2005;20:107–109. mediastinal mass. Chest. 1987;92:529–530.
513. Lamers RJ, van Belle AF. Thoracic duct cyst in the middle part of 541. Eisenstat R, Bruce D, Williams LE, Katz DS. Primary liposarcoma
the mediastinum. AJR Am J Roentgenol. 1993;161:675. of the mediastinum with coexistent mediastinal lipomatosis.
514. Zambudio AR, Lanzas JT, Calvo MJ, et al. Non-neoplastic medi- AJR Am J Roentgenol. 2000;174:572–573.
astinal cysts. Eur J Cardiothorac Surg. 2002;22:712–716. 542. Munden RF, Nesbitt JC, Kemp BL, et al. Primary liposarcoma of
515. Winer-Muram HT, Gold RE. Effusion in the superior pericardial the mediastinum. AJR Am J Roentgenol. 2000;175:1340.
recess simulating a mediastinal mass. AJR Am J Roentgenol. 543. DeSantos LA, Ginaldi S, Wallace S. Computed tomography in
1990;154:69–71. liposarcoma. Cancer. 1981;47:46–54.
516. Terrier F, Revel D, Pajannen H, et al. MR imaging of body fluid 544. Cowling MG, Holmes SJ, Adam EJ. Benign chest wall lipoblas-
collections. J Comput Assist Tomogr. 1986;10:953–962. toma of infancy producing underlying bone enlargement.
517. Naidich DP, Rumancik WM, Ettenger NA, et al. Congenital Pediatr Radiol. 1995;25:54–55.
anomalies of the lungs in adults: MR diagnosis. AJR Am J 545. Broeders A, Smet MH, Breysem L, Marchal G. Lipoblastoma: a
Roentgenol. 1988;151:13–19. rare mediastinal tumour in a child. Pediatr Radiol. 2000;30:580.
518. Stark DD, Higgins CB, Lanzer P, et al. Magnetic resonance imag- 546. Raman Sharma R, Mahapatra AK, Pawar SJ, et al. An unusual
ing of the pericardium: normal and pathologic findings. posterior mediastinal lipoblastoma with spinal epidural exten-
Radiology. 1984;150:469–474. sion presenting as a painful suprascapular swelling: case report
519. Barakos JA, Brown JJ, Brescia RJ, Higgins CB. High signal inten- and a brief review of the literature. J Clin Neurosci. 2002;
sity lesions of the chest in MR imaging. J Comput Assist Tomogr. 9:204–207.
1989;13:797–802. 547. Cantisani V, Mortele KJ, Glickman JN, et al. Large retroperi-
520. Siegel MJ, Nadel SN, Glazer HS, Sagel SS. Mediastinal lesions toneal hibernoma in an adult male: CT imaging findings with
in children: comparison of CT and MR. Radiology. 1986;160: pathologic correlation. Abdom Imaging. 2003;28:721–724.
241–244. 548. Udwadia ZF, Kumar N, Bhaduri AS. Mediastinal hibernoma. Eur
521. Winsett MZ, Amparo EG, Fagan CJ, et al. MR imaging of J Cardiothorac Surg. 1999;15:533–535.
mediastinal pseudocyst. J Comput Assist Tomogr. 1988;12: 549. Truong MT, Erasmus JJ, Munden RF, et al. Focal FDG uptake in
320–322. mediastinal brown fat mimicking malignancy: a potential pitfall
522. Murayama S, Murakami J, Watanabe H, et al. Signal intensity resolved on PET/CT. AJR Am J Roentgenol. 2004;183: 1127–1132.
characteristics of mediastinal cystic masses on T1-weighted MRI. 550. Minotti AJ, Shah L, Keller K. Positron emission tomography/
J Comput Assist Tomogr. 1995;19:188–191. computed tomography fusion imaging in brown adipose tissue.
523. Lyon RD, McAdams HP. Mediastinal bronchogenic cyst: demon- Clin Nucl Med. 2004;29:5–11.
stration of a fluid-fluid level at MR imaging. Radiology. 1993; 551. Fan CM, Fischman AJ, Kwek BH, et al. Lipomatous hypertrophy
186:427–428. of the interatrial septum: increased uptake on FDG PET. AJR Am
524. Paling MR, Williamson BRJ. Epipericardial fat pad: CT findings. J Roentgenol. 2005;184:339–342.
Radiology. 1987;165:335–339. 552. Cohade C, Osman M, Pannu HK, Wahl RL. Uptake in supraclav-
525. Homer JM, Wechsler RJ, Carter BL. Mediastinal lipomatosis. icular area fat (“USA-Fat”): description on 18F-FDG PET/CT.
Radiology. 1978;128:657–661. J Nucl Med. 2003;44:170–176.
526. Koerner HF, Sun DIC. Mediastinal lipomatosis secondary to 553. Quinn SF, Monson M, Paling M. Case report. Spinal lipoma pre-
steroid therapy. Am J Rad Ther. 1966;98:461–464. senting as a mediastinal mass: diagnosis by CT. J Comput Assist
527. Price JE, Rigler LG. Widening of the mediastinum resulting from Tomogr. 1983;7:1087–1089.
fat accumulation. Radiology. 1970;96:497–500. 554. Kim K, Koo BC, Davis JT, Franco-Saenz R. Primary myelolipoma
528. Bein NE, Mancuso AA, Mink JH, Hansen GC. Computed tomog- of mediastinum. CT. 1984;8:119–123.
raphy in the evaluation of mediastinal lipomatosis. J Comput 555. Kline ME, Patel BU, Agosti SJ. Noninfiltrating angiolipoma of
Assist Tomogr. 1978;2:379–383. the mediastinum. Radiology. 1990;175:737–738.
529. Glickstein MF, Miller WT, Dalinka MK, Lally JF. Paraspinal lipo- 556. Fagelman D, Caridi JG. CT diagnosis of hernia of Morgagni.
matosis: a benign mass. Radiology. 1987;163:79–80. Gastrointest Radiol. 1984;9:153–155.
557. Gale ME. Bochdalek hernia: prevalence and CT characteristics. 583. Black WC, Armstrong P, Daniel TM, Cooper PH. Computed
Radiology. 1985;156:449–452. tomography of aggressive fibromatosis in the posterior medi-
558. DeMartini WJ, House AJS. Partial Bochdalek’s herniation: com- astinum. J Comput Assist Tomogr. 1987;11:153–155.
puted tomographic evaluation. Chest. 1980;77:702–704. 584. Dosios TJ, Angouras DC, Floros DG. Primary desmoid tumor
559. Ginalski JM, Schnuder P, Moss AA, Brasch RC. Incidence and of the posterior mediastinum. Ann Thorac Surg. 1998;66:
significance of a widened esophageal hiatus at CT scan. J Clin 2098–2099.
Gastroenterol. 1984;6:467–470. 585. Hoeffel C, Floquet J, Regent D, et al. Periesophageal mediastinal
560. Yeager BA, Guglielmi GE, Schiebler ML, et al. Magnetic reso- fibromatosis. Abdom Imaging. 2000;25:235–238.
nance imaging of Morgagni hernia. Gastrointest. Radiology. 586. Ko SF, Ng SH, Hsiao CC, et al. Juvenile fibromatosis of the pos-
1987;12:296–298. terior mediastinum with intraspinal extension. AJNR Am J
561. London J, Kim EE, Wallace S, et al. MR imaging of liposarco- Neuroradiol. 1996;17:522–524.
mas: correlation of MR features and histology. J Comput Assist 587. Tam CG, Broome DR, Shannon RL. Desmoid tumor of the ante-
Tomogr. 1989;13:832–835. rior mediastinum: CT and radiologic features. J Comput Assist
562. Dooms GC, Hricak H, Sollitto RA, Higgins CB. Lipomatous Tomogr. 1994;18:499–501.
tumors and tumors with fatty component: MR imaging poten- 588. Ciftci E, Erden I, Koral K, Akyar S. MR imaging in a desmoid
tial and comparison of MR and CT results. Radiology. 1985;157: tumor of the posterior mediastinum with extension into the
479–483. abdominal cavity. A case report. Acta Radiol. 1998;39:301–303.
563. Shaffer K, Rosado-de-Christenson ML, Patz EF Jr, et al. Thoracic 589. Hanau CA, Miettinen M. Solitary fibrous tumor: histological
lymphangioma in adults: CT and MR imaging features. AJR Am and immunohistochemical spectrum of benign and malignant
J Roentgenol. 1994;162:283–289. variants presenting at different sites. Hum Pathol. 1995;26:
564. Brown LR, Reiman HM, Rosenow EC, et al. Intrathoracic 440–449.
lymphangioma. Mayo Clin Proc. 1986;61:882–892. 590. Ali SZ, Hoon V, Hoda S, et al. Solitary fibrous tumor. A cytologic-
565. Miyake H, Shiga M, Takaki H, et al. Mediastinal lym- histologic study with clinical, radiologic, and immunohisto-
phangiomas in adults: CT findings. J Thorac Imaging. 1996; chemical correlations. Cancer. 1997;81:116–121.
1:83–85. 591. Witkin GB, Rosai J. Solitary fibrous tumor of the mediastinum.
566. Pilla TJ, Wolverson MK, Sundaram M, et al. CT evaluation of A report of 14 cases. Am J Surg Pathol. 1989;13:547–557.
cystic lymphangiomas of the mediastinum. Radiology. 1982; 592. Sakurai H, Hasegawa T, Watanabe S, et al. Inflammatory myofi-
144:841–842. broblastic tumor of the lung. Eur J Cardiothorac Surg. 2004;25:
567. Pannell TL, Jolles H. Adult cystic mediastinal lymphangioma 155–159.
simulating a thymic cyst. J Thorac Imaging. 1991;7:86–89. 593. Yamaguchi M, Yoshino I, Osoegawa A, et al. Inflammatory
568. Charruau L, Parrens M, Jougon J, et al. Mediastinal lymphan- myofibroblastic tumor of the mediastinum presenting as supe-
gioma in adults: CT and MR imaging features. Eur Radiol. 2000; rior vena cava syndrome. J Thorac Cardiovasc Surg. 2003;126:
10:1310–1314. 870–872.
569. Oshikiri T, Morikawa T, Jinushi E, et al. Five cases of the lym- 594. Crespo C, Navarro M, Gonzalez I, et al. Intracranial and medi-
phangioma of the mediastinum in adult. Ann Thorac Cardiovasc astinal inflammatory myofibroblastic tumour. Pediatr Radiol.
Surg. 2001;7:103–105. 2001;31:600–602.
570. Angtuaco EJC, Jiminez JF, Burrows P, Ferris E. Case report. 595. Agrons GA, Rosado-de-Christenson ML, Kirejczyk WM, et al.
Lymphatic-venous malformation (lymphangiohemangioma) of Pulmonary inflammatory pseudotumor: radiologic features.
the mediastinum. J Comput Assist Tomogr 1983;7:895–897. Radiology. 1998;206:511–518.
571. Shaked A, Raz I, Gottehrer A, Romanoff H. Asymptomatic, 596. Cakir O, Topal U, Bayram AS, Tolunay S. Sarcomas: rare primary
highly vascularized superior mediastinal mass. Chest. 1984;86: malignant tumors of the thorax. Diagn Interv Radiol. 2005;11:
621–622. 23–27.
572. Swensen SJ, Hartman TE, Mayo JR, et al. Diffuse pulmonary 597. Murakawa T, Nakajima J, Fukami T, et al. Malignant fibrous his-
lymphangiomatosis: CT findings. J Comput Assist Tomogr. 1995; tiocytoma in the anterior mediastinum. Jpn J Thorac Cardiovasc
19:348–352. Surg. 2001;49:722–727.
573. Cohen AJ, Sbaschnig RJ, Hochholzer L, et al. Mediastinal 598. Moran CA, Suster S, Perino G, et al. Malignant smooth muscle
hemangiomas. Ann Thorac Surg. 1987;43:656–659. tumors presenting as mediastinal soft tissue masses. A clinico-
574. Abe K, Akata S, Ohkubo Y, et al. Venous hemangioma of the pathologic study of 10 cases. Cancer. 1994;74:2251–2260.
mediastinum. Eur Radiol. 2001;11:73–75. 599. Suster S, Moran CA. Malignant cartilaginous tumors of the
575. Hayashi A, Takamori S, Tayama K, et al. Primary hemangioperi- mediastinum: clinicopathological study of six cases presenting as
cytoma of the superior mediastinum: a case report. Ann Thorac extraskeletal soft tissue masses. Hum Pathol. 1997;28:588–594.
Cardiovasc Surg. 1998;4:283–285. 600. Suster S, Moran CA. Chordomas of the mediastinum: clinico-
576. Araoz PA, Eklund HE, Welch TJ, Breen JF. CT and MR imaging of pathologic, immunohistochemical, and ultrastructural study of
primary cardiac malignancies. Radiographics. 1999;19:1421–1434. six cases presenting as posterior mediastinal masses. Hum
577. Killion MJ, Brodovsky HS, Schwarting R. Pericardial angiosar- Pathol. 1995;26:1354–1362.
coma after mediastinal irradiation for seminoma. A case report 601. Halber MD, Daffner RH, Thompson WM. CT of the esopha-
and a review of the literature. Cancer. 1996;78:912–917. gus: normal appearance. AJR Am J Roentgenol. 1979;133:
578. Kardamakis D, Bouboulis N, Ravazoula P, et al. Primary heman- 1047–1050.
giosarcoma of the mediastinum. Lung Cancer. 1996;16:81–86. 602. Fadoo F, Ruiz DE, Dawn SK, et al. Helical CT esophagography
579. Mentzel T, Beham A, Calonje E, et al. Epithelioid hemangioen- for the evaluation of suspected esophageal perforation or
dothelioma of skin and soft tissues: clinicopathologic and rupture. AJR Am J Roentgenol. 2004;182:1177–1179.
immunohistochemical study of 30 cases. Am J Surg Pathol. 1997; 603. Noh HM, Fishman EK, Forastiere AA, et al. CT of the esophagus:
21:363–374. spectrum of disease with emphasis on esophageal carcinoma.
580. Rubinowitz AN, Moreira AL, Naidich DP. Mediastinal heman- Radiographics. 1995;15:1113–1134.
gioendothelioma: radiologic-pathologic correlation. J Comput 604. Boring CC, Squires TS, Tong T, Montgomery S. Cancer statistics,
Assist Tomogr. 2000;24:721–723. 1994. CA Cancer J Clin. 1994;44:7–26.
581. Suster S, Moran CA, Koss MN. Epithelioid hemangioendothe- 605. Beahrs OH, Meyers MH, eds. Manual for staging of cancer, 2nd
lioma of the anterior mediastinum. Clinicopathologic, immuno- ed. Philadelphia: JB Lippincott; 1983.
histochemical, and ultrastructural analysis of 12 cases. Am J Surg 606. Moss AA, Schnyder P, Thoeni RF, Margulis AR. Esophageal carci-
Pathol. 1994;18:871–881. noma: pretherapy staging by computed tomography. AJR Am J
582. Mori M, Nakanishi N, Furuya K. Hemangiopericytoma of the Roentgenol. 1981;136:1051–1056.
mediastinum causing spontaneous hemothorax. Ann Thorac 607. Reinig JW, Stanley JH, Schabel SI. CT evaluation of thickened
Surg. 1994;58:1525–1527. esophageal walls. AJR Am J Roentgenol. 1983;140:951–958.
608. Halvorsen RA, Thompson WM. Computed tomographic staging 634. Reddy SC. Esophagopleural fistula. J Comput Assist Tomogr.
of gastrointestinal tract malignancies. Part 1. Esophagus and 1983;7:376–378.
stomach. Invest Radiol. 1987;22:2–16. 635. van Kouwen MC, Oyen WJ, Nagengast FM, et al. FDG-PET scan-
609. Tio TL, Cohen P, Coene PP, et al. Endosonography and computed ning in the diagnosis of gastrointestinal cancers. Scand J
tomography of esophageal carcinoma. Preoperative classification Gastroenterol Suppl. 2004;241:85–92.
compared to the new (1987) TNM system. Gastroenterology. 636. Terrier F, Schapira CL, Fuchs WA. CT assessment of operability
1989;96:1478–1486. in carcinoma of the esophagogastric junction. Eur J Radiol.
610. Kressel HY, Callen PW, Montagne JP, et al. Computed tomo- 1984;4:114–117.
graphic evaluation of disorders affecting the alimentary tract. 637. Daffner RH, Halber MD, Postlethwait RW, et al. CT of the
Radiology. 1978;129:451–455. esophagus. Part 2. Carcinoma. AJR Am J Roentgenol. 1979;133:
611. Quint LE, Glazer GM, Orringer MB, Gross BH. Esophageal 1051–1055.
imaging by MR and CT: study of normal anatomy and neo- 638. Kron IL, Cantrell RW, Johns ME, et al. Computerized axial
plasms. Radiology. 1985;156:727–731. tomography of the esophagus to determine the suitability for
612. Quint LE, Glazer GM, Orringer MB, Gross BH. Esophageal carci- blunt esophagectomy. Ann Surg. 1984;49:173–174.
noma: CT findings. Radiology. 1985;155:171–175. 639. Moss AA. Critical reviews. Esophageal carcinoma: CT findings.
613. Wolfman NT, Scharling ES, Chen MY. Esophageal squamous Invest Radiol. 1987;22:84–87.
carcinoma. Radiol Clin North Am. 1994;32:1183–1201. 640. Maerz LL, Deveney CW, Lopez RR, McConnell DB. Role of
614. Thompson WM, Halvorsen RA, Foster WL, et al. Computed computed tomographic scans in the staging of esophageal
tomography for staging esophageal and gastroesophageal and proximal gastric malignancies. Am J Surg. 1993;165:
cancer: reevaluation. AJR Am J Roentgenol. 1983;141:951–958. 558–560.
615. Rice TW. Clinical staging of esophageal carcinoma. CT, EUS, 641. Legmann P, Marmuse JP, Rjob S, Benacerraf R. Preoperative
and PET. Chest Surg Clin North Am. 2000;10:471–485. computed tomography for transhiatal esophagectomy. Invest
616. Stein HJ, Brucher BL, Sendler A, Siewert JR. Esophageal cancer: Radiol. 1991;26:987–991.
patient evaluation and pre-treatment staging. Surg Oncol. 2001; 642. Meltzer CC, Luketich JD, Friedman D, et al. Whole-body FDG
10:103–111. positron emission tomographic imaging for staging esophageal
617. Wakelin SJ, Deans C, Crofts TJ, et al. A comparison of comput- cancer comparison with computed tomography. Clin Nucl Med.
erised tomography, laparoscopic ultrasound and endoscopic 2000;25:882–887.
ultrasound in the preoperative staging of oesophago-gastric 643. Yusoff IF, Sahai AV. Staging esophageal adenocarcinoma. Semin
carcinoma. Eur J Radiol. 2002;41:161–167. Gastrointest Dis. 2003;14:146–154.
618. Wu LF, Wang BZ, Feng JL, et al. Preoperative TN staging of 644. Iizuka T, Isono K, Kakegawa T, Watanabe H. Parameters linked to
esophageal cancer: comparison of miniprobe ultrasonography, ten-year survival in Japan of resected esophageal carcinoma.
spiral CT and MRI. World J Gastroenterol. 2003;9:219–224. Report of the Japanese Committee for Registration of Esophageal
619. Vilgrain V, Mompoint D, Palazzo L, et al. Staging of esophageal Carcinoma Cases. Chest. 1989;96:1005–1011.
carcinoma: comparison of results with endoscopic sonography 645. Heiken JP, Balfe DM, Roper CL. CT evaluation after esopha-
and CT. AJR Am J Roentgenol. 1990;155:277–281. gogastrectomy. AJR Am J Roentgenol. 1984;143:555–560.
620. Botet JF, Lightdale CJ, Zauber AG, et al. Preoperative staging of 646. Gross BH, Agha FP, Glazer GM, Orringer MB. Gastric interposi-
gastric cancer: comparison of endoscopic US and dynamic CT. tion following transhiatal esophagectomy: CT evaluation.
Radiology. 1991;181:426–432. Radiology. 1985;155:177–179.
621. Samuelson L, Hambraeus GM, Mercke EC. CT staging of eso- 647. Recht MP, Coleman BG, Barbot DJ, et al. Recurrent esophageal
phageal carcinoma. Acta Radiol Diag. 1984;25:7–11. carcinoma at thoracotomy incisions: diagnostic contributions
622. Freeny PC, Marks WM. Adenocarcinoma of the gastroesophageal of CT. J Comput Assist Tomogr. 1989;13:58–60.
junction: barium and CT examination. AJR Am J Roentgenol. 648. Glickstein MF, Gefter WB, Low D, Stephenson LW. Case report.
1982;138:1077–1084. Esophageal mucocele after surgical isolation of the esophagus.
623. Halvorsen RA, Magruder-Habib K, Foster WL, et al. Esophageal AJR Am J Roentgenol. 1987;149:729–730.
cancer staging by CT: long-term follow-up study. Radiology. 649. Kantarci M, Polat P, Alper F, et al. Comparison of CT and MRI
1986;161:147–151. for the diagnosis recurrent esophageal carcinoma after opera-
624. Schneekloth G, Terrier F, Fuchs WA. Computed tomography in tion. Dis Esophagus. 2004;17:32–37.
carcinoma of esophagus and cardia. Gastrointest Radiol. 1983; 650. Cornwell J, Walden C, Ghahremani GG. Case report. CT
8:193–206. demonstration of fistula between esophageal carcinoma and
625. Picus D, Balfe DM, Koehler RE, et al. Computed tomography spinal canal. J Comput Assist Tomogr. 1986;10:871–873.
in the staging of esophageal carcinoma. Radiology. 1983;146: 651. Webb WR, Gamsu G, Speckman JM, et al. CT demonstration of
433–438. mediastinal aortic arch anomalies. J Comput Assist Tomogr. 1982;
626. Botet JF, Lightdale C. Endoscopic sonography of the upper 6:445–451.
gastrointestinal tract. AJR Am J Roentgenol. 1991;156:63–68. 652. Day DL. Aortic arch in neonates with esophageal atresia: preop-
627. Takashima S, Takeuchi N, Shiozaki H, et al. Carcinoma of the erative assessment using CT. Radiology. 1985;155:99–100.
esophagus: CT vs MR imaging in determining resectability. AJR 653. Williford ME, Thompson WM, Hamilton JD, Postlethwait RW.
Am J Roentgenol. 1991;156:297–302. Esophageal tuberculosis: findings on barium swallow and
628. Korst RJ, Altorki NK. Imaging for esophageal tumors. Thorac computed tomography. Gastrointest Radiol. 1983;8:119–122.
Surg Clin. 2004;14:61–69. 654. Im JG, Kim JH, Han MC, Kim CW. Computed tomography of
629. Vazquez-Sequeiros E, Wiersema MJ, Clain JE, et al. Impact of esophagomediastinal fistula in tuberculous mediastinal lym-
lymph node staging on therapy of esophageal carcinoma. phadenitis. J Comput Assist Tomogr. 1990;14:89–92.
Gastroenterology. 2003;125:1626–1635. 655. Pakter RL, Fishman EK. Metastatic osteosarcoma to the heart
630. Moreto M. Diagnosis of esophagogastric tumors. Endoscopy. and mediastinum presenting as esophageal obstruction.
2005;37:26–32. J Comput Assist Tomogr. 1983;7:1114–1115.
631. Choi JY, Lee KH, Shim YM, et al. Improved detection of individ- 656. Gale ME, Birnbaum SB, Gale DR, Vincent ME. Esophageal inva-
ual nodal involvement in squamous cell carcinoma of the sion by lung cancer: CT diagnosis. J Comput Assist Tomogr. 1984;
esophagus by FDG PET. J Nucl Med. 2000;41:808–815. 8:694–698.
632. Yoon YC, Lee KS, Shim YM, et al. Metastasis to regional lymph 657. Levine MS. Benign tumors of the esophagus: radiologic evalua-
nodes in patients with esophageal squamous cell carcinoma: CT tion. Semin Thorac Cardiovasc Surg. 2003;15:9–19.
versus FDG PET for presurgical detection prospective study. 658. Yang PS, Lee KS, Lee SJ, et al. Esophageal leiomyoma: radiologic
Radiology. 2003;227:764–770. findings in 12 patients. Korean J Radiol. 2001;2:132–137.
633. Wippold FJ, Schnapf D, Bennet LL, Friedman AC. Case report. 659. Levine MS, Buck JL, Pantongrag-Brown L, et al. Leiomyo-
Esophago-subarachnoidal fistula: an unusual complication of sarcoma of the esophagus: radiographic findings in 10 patients.
esophageal carcinoma. J Comput Assist Tomogr. 1982;6:147–149. AJR Am J Roentgenol. 1996;167:27–32.
660. Levine MS, Buck JL, Pantongrag-Brown L, et al. Esophageal 686. Han SY, McElvein RB, Aldrete JS. Perforation of the esophagus:
leiomyomatosis. Radiology. 1996;199:533–536. correlation of site and cause with plain film findings. AJR Am J
661. Levine MS, Buck JL, Pantongrag-Brown L, et al. Fibrovascular Roentgenol. 1985;145:537–540.
polyps of the esophagus: clinical, radiographic, and pathologic 687. Swanson JO, Levine MS, Redfern RO, Rubesin SE. Usefulness of
findings in 16 patients. AJR Am J Roentgenol. 1996;166: high-density barium for detection of leaks after esophagogas-
781–787. trectomy, total gastrectomy, and total laryngectomy. AJR Am J
662. Berkovich GY, Levine MS, Miller WT Jr. CT findings in patients Roentgenol. 2003;181:415–420.
with esophagitis. AJR Am J Roentgenol. 2000;175:1431–1434. 688. Berkmen YM, Auh YH. CT diagnosis of acquired tracheoe-
663. Rabushka LS, Fishman EK, Kuhlman JE. CT evaluation of acha- sophageal fistula in adults. J Comput Assist Tomogr. 1985;9:
lasia. J Comput Assist Tomogr. 1991;15:434–439. 302–304.
664. Bhalla M, Silver RM, Shepard JO, McLoud TC. Chest CT in 689. Kronthal AJ, Heitmiller RF, Fishman EK, Kuhlman JE. Medi-
patients with scleroderma: prevalence of asymptomatic esopha- astinal seroma after esophagogastrectomy. AJR Am J Roentgenol.
geal dilatation and mediastinal lymphadenopathy. AJR Am 1991;156:715–716.
J Roentgenol. 1993;161:269–272. 690. Gobien RP, Stanley JH, Gobien BS, et al. Percutaneous catheter
665. Clark KE, Foley WD, Lawson TL, et al. CT evaluation of aspiration and drainage of suspected mediastinal abscesses.
esophageal and upper abdominal varices. J Comput Assist Radiology. 1984;151:69–71.
Tomogr. 1980;4:510–515. 691. Carrol CL, Jeffrey RB, Federle MP, Vernacchia FS. CT evaluation of
666. Balthazar EJ, Naidich DP, Megibow AJ, Lefleur RS. CT evaluation mediastinal infections. J Comput Assist Tomogr. 1987;11:449–454.
of esophageal varices. AJR Am J Roentgenol. 1987;148:131–135. 692. Breatnach E, Nath PH, Delany DJ. The role of computed tomog-
667. Hirose WJ, Takashima T, Suzuki M, Matsui O. Case report. raphy in acute and subacute mediastinitis. Clin Radiol. 1986;
“Downhill” varices demonstrated by dynamic computed 37:139–145.
tomography. J Comput Assist Tomogr. 1984;8:1007–1009. 693. Morgan DE, Nath H, Sanders C, Hasson JH. Mediastinal actino-
668. Halden WJ, Harnsberger HR, Mancuso AA. Computed tomogra- mycosis. AJR Am J Roentgenol. 1990;155:735–737.
phy of esophageal varices after sclerotherapy. AJR Am J Roentgenol. 694. Kay HR, Goodman LR, Teplick SK, Mundth ED. Use of com-
1983;140:1195–1196. puted tomography to assess mediastinal complications after
669. Mauro MA, Jaques PF, Swantkowski TM, et al. CT after uncom- median sternotomy. Ann Thorac Surg. 1983;36:706–714.
plicated esophageal sclerotherapy. AJR Am J Roentgenol. 1986; 695. Jolles H, Henry DA, Roberson JP, et al. Mediastinitis following
147:57–60. median sternotomy: CT findings. Radiology. 1996;201:463–466.
670. Saks BJ, Kilby AE, Dietrich PA. Pleural and mediastinal changes 696. Efremidis SC, Dan SJ, Cohen BA, et al. Displaced paraspinal
following endoscopic injection sclerotherapy of esophageal line: role of CT and lymphography. AJR Am J Roentgenol. 1981;
varices. Radiology. 1983;149:639–642. 136:505–509.
671. Callen PWQ, Filly RA, Korobkin M. Computed tomographic 697. Tarr RW, Glick AG, Shaff MI. Case report. Benign hemangioen-
evaluation of the diaphragmatic crura. Radiology. 1978;126: dothelioma involving posterior mediastinum: CT findings.
413–416. J Comput Assist Tomogr. 1986;10:865–867.
672. Naidich DP, Megibow AJ, Ross CR, et al. Computed tomogra- 698. Ribet ME, Cardot GR. Neurogenic tumors of the thorax. Ann
phy of the diaphragm: normal anatomy and variants. J Comput Thorac Surg. 1994;58:1091–1095.
Assist Tomogr. 1983;7:633–640. 699. Lonergan GJ, Schwab CM, Suarez ES, Carlson CL. Neuro-
673. Thompson WM, Halvorsen RA, Williford ME, et al. Computed blastoma, ganglioneuroblastoma, and ganglioneuroma: radio-
tomography of the gastroesophageal junction. Radiographics. logic-pathologic correlation. Radiographics. 2002;22:911–934.
1982;2:179–193. 700. Reed JC, Hallet KK, Feigin DS. Neural tumors of the thorax:
674. Govoni AF, Whalen JP, Kazam E. Hiatal hernia: a relook. subject review from the AFIP. Radiology. 1978;126:9–17.
Radiographics. 1983;3:612–644. 701. Topcu S, Alper A, Gulhan E, et al. Neurogenic tumours of the me-
675. Marks WM, Callen PW, Moss AA. Gastroesophageal region: diastinum: a report of 60 cases. Can Respir J. 2000;7:261–265.
source of confusion on CT. AJR Am J Roentgenol. 1981;136: 702. Ross CR, McCauley DI, Naidich DP. Intrathoracic neurofibroma
359–362. of the vagus nerve associated with bronchial obstruction.
676. Thompson WM, Halvorsen RA, Foster W, et al. Computed J Comput Assist Tomogr. 1982;6:406–412.
tomography of the gastroesophageal junction: value of the left 703. Kumar AJ, Kuhajda FP, Martinez CR, et al. CT of extracranial
lateral decubitus view. J Comput Assist Tomogr. 1984;8:346–349. nerve sheath tumors. J Comput Assist Tomogr. 1983;7:857–865.
677. Godwin JD, MacGregor JM. Case report. Extension of ascites 704. Biondetti PR, Vigo M, Fiore D, et al. CT appearance of generalized
into the chest with hiatal hernia: visualization on CT. AJR Am J von Recklinghausen neurofibromatosis. J Comput Assist Tomogr.
Roentgenol. 1987;148:31–32. 1983;7:866–869.
678. Lindell MM, Bernadino ME. Diagnosis of hiatus hernia by com- 705. Bourgouin PM, Shepard JO, Moore EH, McLoud TC. Plexiform
puted tomography. Comput Radiol. 1981;5:16–19. neurofibromatosis of the mediastinum: CT appearance. AJR Am
679. Pupols A, Ruzicka FF. Hiatal hernia causing a cardia pseudo- J Roentgenol. 1988;151:461–463.
mass on computed tomography. J Comput Assist Tomogr. 1984;8: 706. Joseph SG, Tellis CJ. Posterior mediastinal mass with intraspinous
699–700. extension. Chest. 1988;93:1101–1103.
680. Wechsler RJ, Steiner RM, Goodman LR, et al. Iatrogenic 707. Aughenbaugh GL. Thoracic manifestations of neurocutaneous
esophageal-pleural fistula: subtlety of diagnosis in the absence diseases. Radiol Clin North Am. 1984;22:741–756.
of mediastinitis. Radiology. 1982;144:239–243. 708. Coleman BG, Arger PH, Dalinka MK, et al. CT of sarcomatous
681. Backer CL, LoCicero J 3rd, Hartz RS, et al. Computed tomogra- degeneration in neurofibromatosis. AJR Am J Roentgenol. 1983;
phy in patients with esophageal perforation. Chest. 1990; 140:383–387.
98:1078–1080. 709. Levine E, Huntrakoon M, Wetzel LH. Malignant nerve-sheath
682. White CS, Templeton PA, Attar S. Esophageal perforation: CT neoplasms in neurofibromatosis: distinction from benign
findings. AJR Am J Roentgenol. 1993;160:767–770. tumors by using imaging techniques. AJR Am J Roentgenol.
683. Glenny RW, Fulkerson WJ, Ravin CE. Occult spontaneous 1987;149:1059–1064.
esophageal perforation. Unusual clinical and radiographic pres- 710. Feinstein RS, Gatewood OMB, Fishman EK, et al. Computed
entation. Chest. 1987;92:562–565. tomography of adult neuroblastoma. J Comput Assist Tomogr.
684. Pezzulli FA, Aronson D, Goldberg N. Case report. Computed 1984;8:720–726.
tomography of mediastinal hematoma secondary to unusual 711. Cohen LA, Schwartz AM, Rockoff SD. Benign schwannomas:
esophageal laceration: a Boerhaave variant. J Comput Assist pathologic basis for CT inhomogeneities. AJR Am J Roentgenol.
Tomogr. 1989;13:129–131. 1986;147:141–143.
685. Brown BM. Case report. Computed tomography of mediastinal 712. Healy JF, Wells MV, Carlstom T, Rosenkrantz H. Lateral thoracic
abscess secondary to posttraumatic esophageal laceration. meningocele demonstrated by computerized tomography.
J Comput Assist Tomogr. 1984;8:765–767. Comput Tomogr. 1980;4:159–163.
713. Moon WK, G. IJ, Han MC. Malignant schwannomas of the report and review of the literature. J Pediatr Surg. 2003;
thorax: CT findings. J Comput Assist Tomogr. 1993;17:274–276. 38:1114–1116.
714. Burk DLJr, Brunberg JA, Kanal E, et al. Spinal and paraspinal 722. Tanaka F, Kitano M, Tatsumi A, et al. Paraganglioma of the
neurofibromatosis: surface coil MR imaging at 1.5 T. Radiology. posterior mediastinum: value of magnetic resonance imaging.
1987;162:797–801. Ann Thorac Surg. 1992;53:517–519.
715. Dietrich RB, Kangarloo H, Lenarsky C, Feig SA. Neuroblastoma: 723. Frances IR, Glazer GM, Shapiro B, et al. Complementary roles of
the role of MR imaging. AJR Am J Roentgenol. 1987;148: 927–942. CT and 131-I-MIBG scintigraphy in diagnosing pheochromocy-
716. Cohen MD, Weetman R, Provisor A. Magnetic resonance imag- toma. AJR Am J Roentgenol. 1983;141:719–725.
ing of neuroblastoma with 0.15-T magnet. AJR Am J Roentgenol. 724. Swensen SJ, Brown ML, Sheps SG, et al. Use of 13I-MIBG
1984;143:1241–1248. scintigraphy in the evaluation of suspected pheochromocy-
717. Sakai F, Sone S, Kiyono K, et al. Intrathoracic neurogenic tumors: toma. Mayo Clin Proc. 1985;60:299–304.
MR-pathologic correlation. AJR Am J Roentgenol. 1992;159: 725. Miles J, Pennybacker J, Sheldon P. Intrathoracic meningocele. its
279–283. development and association with neurofibromatosis. J Neurol
718. Armstrong EA, Harwood-Nash DCF, Ritz CR, et al. CT of neu- Neurosurg Psychiatry. 1969;32:99.
roblastomas and ganglioneuromas in children. AJR Am J 726. Gumbs RV, Higginbotham-Ford EA, Teal JS, et al. Thoracic
Roentgenol. 1982;139:571–576. extramedullary hematopoiesis in sickle-cell disease. AJR Am J
719. Tateishi U, Hasegawa T, Makimoto A, Moriyama N. Adult neu- Roentgenol. 1987;149:889–893.
roblastoma: radiologic and clinicopathologic features. J Comput 727. Long JA, Doppman JL, Nienhuis AW. Computed tomographic
Assist Tomogr. 2003;27:321–326. studies of thoracic extramedullary hematopoiesis. J Comput
720. Andrich MP, Shalaby-Rana E, Movassaghi N, Majd M. The role Assist Tomogr. 1980;4:67–70.
of 131 iodine-metaiodobenzylguanidine scanning in the correl- 728. Danza FM, Falappa P, Leone G, Pincelli G. Extramedullary
ative imaging of patients with neuroblastoma. Pediatrics. 1996; hematopoiesis [letter]. AJR Am J Roentgenol. 1982;139:
97:246–250. 837–838.
721. Spector JA, Willis DN, Ginsburg HB. Paraganglioma (phe- 729. Kirchner SG, Heller RM, Smith CW. Pancreatic pseudocyst of
ochromocytoma) of the posterior mediastinum: a case the mediastinum. Radiology. 1977;123:37–42.
Airways 5
CLINICAL INDICATIONS 453 wheezing, dyspnea, or hemoptysis, especially with normal

or nonlocalizing radiographs (2–5); (b) patients with
TECHNIQUE 453 radiographic abnormalities for whom better anatomic
Advanced Reconstruction Techniques 457 delineation is required—included in this category are
patients with radiographic evidence of chronic infiltrates
NORMAL AIRWAY ANATOMY 460 and/or atelectasis (6) or diffuse parenchymal disease in
Trachea and Mainstem Bronchi 460 which identification of airway abnormalities is pertinent
Lobar and Segmental Bronchi 462 (e.g., sarcoidosis) (7); and (c) candidates for interventional
Anatomic Variants 467 bronchoscopic procedures—included in this category
are patients for whom an ever-increasing number of meth-
CENTRAL AIRWAYS 471 ods for diagnosing and especially treating endobronchial
Computed Tomography/Bronchoscopic abnormalities are now available for which precise prepro-
Correlations 471 cedural anatomic correlation is of value (8–12). In addi-
Abnormalities Involving the Trachea and Mainstem tion to these broad indications, CT also frequently plays a
Bronchi 486 critical role in the follow-up of patients undergoing treat-
ment for the entire gamut of diseases affecting the airways.
LOBAR ATELECTASIS 503
Obstructive (Resorption) Atelectasis 508
Nonobstructive Atelectasis 510 TECHNIQUE
PERIPHERAL AIRWAYS 512 The introduction of multidetector computed tomography
Bronchiectasis 512 (MDCT) has profoundly altered our approach to scanning
Small Airway Disease/Bronchiolitis 529 the airways (13,14). The ability to acquire contiguous and/or
overlapping high-resolution 0.5-mm to 1- to 2-mm-thick
ASTHMA 543 sections throughout the thorax in a single breath-hold en-
ables even the smallest airways to be optimally visualized
HEMOPTYSIS 543 (Fig. 5-1). With this approach it is now possible with newer
scanners to routinely reconstruct at a minimum three
prospective datasets, for example, 5-mm sections every
CLINICAL INDICATIONS 5 mm using standard lung and mediastinal reconstruction
algorithms, as well as an additional high-resolution dataset
Computed tomography (CT) remains the most accurate with 1-mm sections reconstructed every 10 mm using an
noninvasive means for evaluating the airways and is of appropriate high-resolution reconstruction algorithm. Fol-
established value for imaging virtually the entire gamut of lowing this “generic” approach, all studies in effect become
diseases that affect both large and small airways (1). In a combination of routine and high-resolution examinations
fact, the most common indications generally fall into one resulting in improved detection of airway lesions, especially
of three broad categories: (a) patients in whom occult when compared with conventional CT techniques (15).
disease is suspected—included in this category are sympto- First, although never evaluated in a prospective randomized
matic patients, presenting with chronic cough, localized fashion, in our experience, the availability of high-resolution
A, B
Figure 5-1 Volumetric data acqui-

sition with retrospective high-reso-
lution imaging of the airways. A–D:
Sequential magnified 1-mm sections
show subtle thickening and then ob-
struction with distal mucoid im-
paction of a sixth- to seventh-order
airway in the apical segment of the
right upper lobe (arrows, A–C). This
abnormality was only vaguely sug-
gested on initial 5-mm-thick sec-
tions (not shown). The ability to
retrospectively reconstruct high-res-
olution images through suspicious
airways without the need to rescan
a patient is one of the major advan-
tages of initially acquiring volu-
C, D metric data with 1-mm collimation.
images frequently discloses otherwise unsuspected airway in patients with diffuse airway inflammation, in particular
disease, especially involving the peripheral airways. Second, (Fig. 5-2).
in cases in which additional anatomic resolution is needed, These advantages must be weighed against a number
the ability to retrospectively reconstruct select images of potential disadvantages. These include added recon-
through regions of interest without the need to rescan struction and postprocessing times, as well as storage
patients frequently improves diagnostic certainty (Fig. 5-1). requirements for potentially several hundred additional
In this regard, it is also now possible to obtain high-quality images. Of greater concern is the potential of added risk
multiplanar and/or volumetric reconstructions through resulting from increased radiation exposure. This problem
the whole lung, providing a more global assessment of the can be minimized by routinely using reduced milli-
extent and severity of disease—techniques of potential value amperage whenever feasible, especially in younger patients,
Chapter 5: Airways 455
Figure 5-2 Volumetric data acquisition: advanced image processing. A: 1-mm

section through the lower lobes imaged with lung windows shows ill-defined
nodular densities throughout the right lower lobe. There is evidence of
bronchiectasis in the middle lobe and heterogeneous density in the left lower
lobe. Based on this image, precise anatomic localization of disease is limited.
B, C: Coronal and sagittal maximum intensity projections (MIPs), respectively,
show to better advantage the finding of multiple clusters of small nodules with a
tree-in-bud configuration consistent with small airway infection, in this case due
C to active tuberculosis with endobronchial spread of infection.
women, and patients for whom multiple follow-up studies Another approach of potential value, specifically in
are planned. Diagnostic images can almost always be patients with suspected bronchiectasis, is to utilize 3-mm
obtained using between 80 and 100 mAs and frequently as rather than 1-mm sections. As shown by Remy-Jardin et al.
low as 40 mAs in thin patients (16,17). As shown by Yi (20), no difference in accuracy or estimation of disease
et al. (18), no significant loss of diagnostic accuracy can be extent or severity could be identified when 1-mm and
identified when evaluating patients with bronchiectasis 3-mm sections were compared, leading these authors to
with tube current setting as low as 70 mA. conclude that the use of scans generated by 4 2.5 mm col-
Although volumetric high-resolution CT (HRCT) has limation in place of 4 1 mm collimation could allow as
been reported to result in a considerably greater radiation great as 20% decrease in radiation dose without sacrificing
dose (19), use of low-dose scanning coupled with dose mod- diagnostic accuracy. It should be emphasized, however, that
ulation techniques available on newer CT scanners can be sections thicker than 3 mm are suboptimal for evaluating
employed to minimize this problem. One solution is to ini- bronchiectasis and should never be relied on to make this
tially perform volumetric data acquisition, with subsequent diagnosis (Fig. 5-3) (21).
follow-up examinations performed with dedicated high- Scans are routinely performed in deep inspiration
resolution images only using low-dose technique, whenever with the patient supine. Expiratory images may prove
feasible, to minimize subsequent radiation exposure. useful to either detect or confirm a suspicion of air
Figure 5-3 Bronchiectasis: evalua-

tion with thick versus thin sections.
A, B: Magnified 5- and 1-mm sections,
respectively, through the right upper
lobe bronchus clearly demonstrate
the advantage of high-resolution thin
sections for diagnosing bronchiectasis
A, B (compare arrows in A and B).
trapping (Fig. 5-4) (22–25). Expiratory HRCT scans may choice of window levels and widths (Fig. 5-5). In one study
be obtained either during suspended respiration after using inflation-fixed lungs imaged with 1.5 mm high-
forced exhalation (static expiratory CT) or during forced resolution technique, optimal evaluation of peripheral
exhalation (dynamic expiratory CT) (26,27). Expiratory bronchial walls requires the use of window centers between
scans may be obtained at selected levels, typically 3 to 5, 250 and 700 HU, with corresponding window widths
or through the entire lung using volumetric technique. between 1,000 and 1,400 HU, respectively (28).
Although assessment of the central airways may be ac- Although visualization of the airways per se rarely
complished using a wide range of window settings, visuali- requires administration of intravenous contrast material,
zation of smaller peripheral airways, as well as identification in cases in which atelectasis is identified on correspon-
of subtle focal air trapping, is dependent on the appropriate ding radiographs, the use of a bolus of contrast media is
A B
Figure 5-4 Expiratory imaging. A, B: Images obtained at the same level first in deep inspiration (A) and then following forced exhalation
(B) demonstrate the use of expiratory images for detecting otherwise occult air trapping. Note that the central airways markedly narrow
on expiration, consistent with bronchomalacia. In this case, these findings were due to clinically documented postinfectious obliterative
(constrictive) bronchiolitis.
A B
Figure 5-5 Imaging airways: effect of window level and width selection. A, B: Identical sections through the upper lobes imaged with
window center and widths of 700 HU and 1,400 HU (A) and 450 HU and 850 HU (B), respectively. Note that when window levels or
widths are inappropriately low, there is considerable distortion of the airways with resulting pseudobronchial wall thickening and luminal
narrowing (compare arrows in A and B). At the same time, narrowing windows and levels has the beneficial effect of accentuating the extent
of mosaic attenuation due to small airway disease with diffuse air trapping. In this case, there is evidence of extensive central bronchiectasis
due to allergic bronchopulmonary aspergillosis.
recommended when there is a suspicion of mediastinal mucosal disease; (b) limited accuracy in identifying sub-
disease or to facilitate differentiation between central mucosal disease, especially central extension of neoplasm;
tumor and peripheral consolidated or atelectatic lung (c) limited accuracy in the estimation of the length of
(Fig. 5-6). tracheal and bronchial stenoses; and (d) limited assess-
ment of obliquely coursing airways. Endobronchial lung
cancers, even those located in the central airways, can be
Advanced Reconstruction Techniques
missed on conventional CT, particularly when thick 7- to
Although routine axial images remain the standard for 10-mm sections only are acquired.
identifying airway disease, this method has important lim- With the introduction of MDCT, various high-quality
itations, including (a) inability in most cases to identify reconstruction techniques have become routinely available
Figure 5-6 Neoplastic airway

obstruction. A, B: Sequential con-
trast-enhanced sections through
the left mainstem and distal left
upper and lower lobe bronchi, res-
pectively, show a well-defined intra-
luminal mass obstructing the distal
left mainstem bronchus. Following
intravenous contrast administration,
note that the tumor enhances
differentially from the peripheral col-
lapsed left upper and lower lobes,
allowing precise definition of the
true extent of disease (arrows in A
and B). A moderate-sized left pleural
effusion is also evident. Intravenous
contrast optimally should be used to
evaluate all patients presenting with
radiographic evidence of collapse.
(From Naidich DP, Webb WR,
Grenier PA, et al. Imaging of the
airways: functional and radiologic
correlations. Philadelphia: Lippincott
Williams & Wilkins; 2005, with
A, B permission.)
(29–31). These include multiplanar reconstructions MPRs (32–34), especially useful for assessing airway
(MPRs), maximum intensity projections (MIPs) and mini- stenoses, particularly those after prolonged tracheal intuba-
mum intensity projections (MinIPs), external rendering tion or lung transplantation (35). Although less commonly
with either three-dimensional (3D) shaded surface displays employed, MIPs have proved of value for assessing bron-
(SSDs) or volumetric rendering, and internal rendering, or chiolar disease, especially for patients with so-called
so-called virtual bronchoscopy (VB) (Fig. 5-7). Of these tree-in-bud opacities (Fig. 5-2). Least often used in our expe-
various techniques, the most useful, to date, have been rience, although of considerable promise, is VB.
A B
C D
Figure 5-7 Image reconstruction: Kartagener syndrome. A–D: Coronal multiplanar reconstruction (MPR) (A), minimum intensity projec-
tion image (MinIP) (B), maximum intensity projection image (MIP) (C), and volumetric rendered image (D) through the carina in a patient with
situs inversus. These images were all reconstructed using five contiguous 1-mm sections obtained from a 16-detector CT scanner using
1-mm collimators. Although MinIPs accentuate the tracheobronchial tree, effectively erasing the pulmonary vasculature, and MIPs enhance
visualization of the pulmonary vessels at the expense of optimal visualization of the airways, both are 2D representations and therefore lose
any perception of depth. In distinction, volumetric rendering preserves depth perception by including all pixels within the volume of inter-
est. MPRs strike an effective balance, equally delineating both vessels and airways, and are most commonly used for obtaining additional
views of the airways and lungs. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations.
Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
To date, numerous reports have shown that VB is a tromagnetic catheter navigation (12). This technique re-
reliable means for displaying the central airways, compara- quires correlating anatomic reference points identified on
ble to routine fiberoptic bronchoscopy (FB) in detecting ob- virtual bronchoscopic images with those obtained at the
structing endobronchial lesions or focal airway stenoses time of endoscopy. Although not widely available at this
(29,30,36–42). However, despite these studies, VB has rarely time, it may be anticipated that as this technique gains vali-
proved of real diagnostic value. One exception is the use of dation it will emerge as a standard method for diagnosing
VB in guiding ultrathin bronchoscopy (Fig. 5-8) (43,44). An especially peripheral lung lesions otherwise too difficult to
additional new application for which VB is essential is elec- biopsy using standard bronchoscopic technique.
A, B C, D
E, F G, H
I, J K, L
M, N O, P
Figure 5-8 Virtual bronchoscopic CT-guided ultrathin bronchoscopy. See Color Figure 5-8I–P. A–H: Select magnified axial (A–D) and cor-
responding sagittal (E–H) images beginning at the secondary carina and proceeding caudally in a patient with a left lower lobe nodule. I–L:
Virtual bronchoscopic views corresponding to the axial and sagittal views shown in A to H. M–P: Bronchoscopic views obtained using ultra-
thin bronchoscopy. These correspond precisely with the virtual bronchoscopic views at the same levels shown in I to L. Q: CT scanogram ob-
tained at the time of bronchoscopy showing the extremely peripheral position of the ultrathin bronchoscope. R: Magnified view of the left
lower lobe showing the tip of the ultrathin bronchoscope within the left lower lobe nodule (arrow). These images confirm that virtual bron-
choscopy may be indispensable by providing a road map to guide ultrathin bronchoscopes into regions of the lung previously difficult to ac-
cess by bronchoscopists. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations.
Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.) (continued)
Q R
NORMAL AIRWAY ANATOMY extrathoracic and intrathoracic portions; the trachea can be
considered intrathoracic at the point at which it passes
posterior to the manubrium. In a CT study of 50 anatomi-
Trachea and Mainstem Bronchi cally normal persons, the extrathoracic trachea measured
2 to 4 cm in length, whereas the intrathoracic trachea
Anatomy
measured 6 to 9 cm in length (mean, 7.5 0.8 cm).
The trachea is a cartilaginous and fibromuscular tube The trachea contains 16 to 22 cartilaginous “rings,”
extending from the inferior aspect of the cricoid cartilage which are in fact horseshoe shaped and incomplete poste-
(at the level of the sixth cervical vertebra) to the carina (at riorly (Fig. 5-9). These help to support the tracheal wall
the level of the fifth thoracic vertebra). It measures from and to maintain an adequate tracheal lumen during forced
10 to 12 cm in length (45,46). The trachea is divided into expiration. The posterior portion of the tracheal wall, lying
A B
Figure 5-9 Normal tracheal anatomy. See Color Figure 5-9B. A: Cross-sectional drawing shows typical appearance of a horseshoe-shaped
tracheal cartilage within the anterior and lateral tracheal wall, which is markedly thicker than the posterior tracheal membrane. Although
labeled, tracheal mucosa and submucosa are rarely identifiable with CT. B: Histopathologic section shows characteristic horseshoe-shaped
cartilage as dark. The posterior tracheal membrane is bowed anteriorly. (From Webb EM, Elicker BM, Webb WR. Using CT to diagnose
nonneoplastic tracheal abnormalities: appearance of the tracheal wall. AJR Am J Roentgenol. 2000;174:1315–1321, with permission.)
between the open ends of the tracheal cartilages, is a thin oval, with a horseshoe-shaped trachea with a flat posterior
fibromuscular membrane termed the posterior tracheal membrane noted in nearly 25% (45).
membrane (47) (Fig. 5-9). The blood supply to the trachea Tracheal diameter can vary widely in anatomically nor-
is bilateral and tenuous (48). Superiorly, blood is supplied mal persons. In men, tracheal diameter averages 19.5 mm
from branches of the inferior thyroid arteries and the and ranges from 13 to 25 mm (mean, 3 standard
extreme right intercostal artery, whereas inferiorly, bran- deviations) in the coronal plane and 13 to 27 mm in the
ches of the bronchial arteries and the third intercostal sagittal plane (49). In women, tracheal diameter is
artery supply the trachea. slightly less, averaging 17.5 mm and ranging from 10 to
There is marked variability in the cross-sectional appe- 21 mm in the coronal plane and 10 to 23 mm in the
arance of the trachea, which may appear rounded, oval, or sagittal plane (49). In children, tracheal diameter
horseshoe shaped (Fig. 5-10) (49). The posterior tracheal increases as the child grows, and tracheal cross-sectional
membrane may appear convex posteriorly, flat, or convex area correlates most closely with height (50,51). In chil-
anteriorly (35,38). In one study of anatomically normal dren, variation in tracheal dimensions at different levels is
individuals, the trachea most often appeared round or as much as 20% (50,51).
A, B C
D, E F
Figure 5-10 A–F: Normal trachea: high-resolution CT sections through the trachea, obtained at 1-cm intervals. The tracheal wall is well
defined internally by air in the tracheal lumen and is well defined externally by mediastinal fat. Tracheal cartilage within the anterior and
lateral tracheal walls appears a few millimeters thick; calcification (black arrows) is often present in older patients, more commonly in
women. The posterior tracheal wall represents the posterior tracheal membrane (white arrow) and is relatively thin. In this patient, it is
sharply outlined by lung. The posterior tracheal membrane appears flattened, whereas the cartilage anteriorly gives the trachea a rounded
or horseshoe shape. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations.
Figure 5-11 Tracheobronchomalacia.

Paired inspiratory-expiratory imaging.
A, B: Sections at the level of the carina
and proximal right upper lobe bronchus
acquired in inspiration (A) and following
forced expiration (B) show marked nar-
rowing of the airways consistent with
tracheobronchomalacia. Note mosaic
attenuation throughout the right lung in
B, consistent with peripheral airway
obstruction with resultant air trapping.
Anterior bowing of the central airways
is an important sign that the image was
obtained in expiration, and when
marked it correlates with bronchoscopic
evidence of tracheobronchomalacia.
(From Naidich DP, Webb WR, Grenier
PA, et al. Imaging of the airways:
functional and radiologic correlations.
Philadelphia: Lippincott Williams &
A, B Wilkins; 2005, with permission.)
The shape of the intrathoracic trachea (as well as the Airways divide by dichotomous branching, with approx-
mainstem and lobar bronchi) can change dramatically with imately 23 generations of branches from the trachea to
expiration or dynamic respiratory maneuvers (Figs. 5-4 and the alveoli. The wall thickness of conducting bronchi
5-11). Most commonly, invagination of the posterior and bronchioles is approximately proportional to their
membrane occurs with forced expiration, resulting in a diameter, at least for bronchi distal to the segmental level.
crescent-moon or horseshoe-shaped lumen (27). Marked In general, the thickness of the wall of a bronchus or bron-
anterior bowing of the posterior tracheal membrane can be chiole less than 5 mm in diameter should measure 1/6 to
used on expiratory CT as a marker of forced expiration. It 1/10 of its diameter (Table 5-1) (52). Because a bronchus
may also be of value when assessing the effectiveness of a is usually recognized only when its walls are visible, one
patient’s expiratory effort when attempting to diagnose cannot see bronchi with a wall thickness of less than
tracheobronchomalacia or peripheral air trapping. Slight 300 mm on CT or HRCT. As a consequence, normal bronchi
side-to-side narrowing of the trachea also occurs during ex- less than 2 mm in diameter or closer than 2 cm from pleu-
piration, a finding accentuated in the presence of tracheo- ral surfaces equivalent to seventh to ninth order airways
malacia. In a study of dynamic CT during forced expiration, are generally below the resolution even of high-resolution
the sagittal tracheal diameter decreased an average of 32%, CT unless they happen to course in precisely the same
from 19.6 2.3 to 13.3 3.5 mm. In the coronal plane, plane as the CT section (Table 5-1) (53,54).
mean tracheal diameter decreased 13%, from 19.4 2.7 Accurate evaluation of the airways necessitates a detailed
to 16.9 2.6 mm during forced expiration (27). The knowledge of both normal anatomy and variants (Fig. 5-12
extrathoracic trachea can increase in diameter slightly and Table 5-2). Identification of individual airways is a
during forced expiration or during a Valsalva maneuver. function of their size and orientation, as well as recognition
of the spurs that demarcate their origins. Spurs appear
either as triangular wedges or as a linear septum dividing
Lobar and Segmental Bronchi
airways when seen longitudinally. Alternatively, spurs may
Similar to the trachea, the bronchi are composed of both appear as only faint curvilinear densities when sectioned
cartilaginous and fibromuscular elements. However, the horizontally. Recognizing the normal appearance of spurs
distinction between those parts of the bronchial lumen not only allows confident identification of individual
that are supported by cartilage and those that are not is less airways, but also, as significantly, serves as an important
clear cut than in the trachea. For a short distance, the main indication of pathology when abnormally thickened (55).
bronchi contain horseshoe-shaped cartilage plates, as does
the trachea, but the cartilage plates become less regularly Bronchial Nomenclature
shaped at more peripheral levels. On dynamic expiratory Nomenclature for describing segmental lung anatomy has
scans, some posterior invagination of the walls of the undergone several changes, but that described by Jackson
central bronchi is often visible on CT, as in the trachea, and Huber in 1943 (56) has been generally adopted.
reflecting this anatomy. Bronchi are also designated using a numeric system
TABLE 5-1
AIRWAY DIMENSIONS
Airway Diameter Wall Thickness
Generation Division (mm) (mm) Number
Segmental 2–4 5–8 1.5 30

Bronchi 6–8 1.5–3 0.3 500
11 1.0 0.15 2,000
Terminal bronchiole 13 0.7 0.01 8,000
Respiratory bronchiole 15 0.5 0.05 30,000
Intra-acinar 16–23 0.3 — —
Modified from Weibel ER. High resolution computed tomography of the pulmonary parenchyma: anatomi-
cal background. Presented at the Fleischner Society Symposium on Chest Disease, Scottsdale, Arizona,
1990, with permission.
popularized by Boyden (57), in which segments are often is the result of variation in the relationship of bronchi
designated by B followed by a number (e.g., B1) and sub- to the plane of the scan rather than true anatomic variation.
segments are indicated by the segmental number followed It is also easy to distort the appearance of a normal airway if
by a small letter (e.g., B1a) (Fig. 5-12). The numbering of the section thickness exceeds the width of the airway, hence
segments is roughly in their order of origin from the air- the rationale for being able to retrospectively reconstruct
way as one proceeds distally from the carina. Standard high-resolution sections through select airways as needed.
nomenclature and 1961 revisions in letter-number codes Finally, there is considerable variation observed in the
are used in this book (Table 5-2) (58,59). branching patterns of subsegmental bronchi, an additional
cause of potential confusion unless care is taken to trace
these airways centrally toward the carina (Table 5-2).
Right-Sided Bronchial Anatomy
On the right side there are six characteristic sections
The right and left bronchial trees are considered independ- that serve as convenient, easily recognizable landmarks
ently. Emphasis is placed on recognition of characteristic (Fig. 5-13). These include (a) the distal trachea/carina
axial sections (Fig. 5-13). Some variability in the cross- and apical segment bronchus; (b) the right upper lobe
sectional appearance of the airways is to be anticipated; this bronchus including the anterior and posterior segmental
bronchi; (c) the bronchus intermedius; (d) the origins of
the middle and right lower lobe bronchi, often including
the origin and proximal portion of the superior segmental
bronchus; (e) the trunk of the proximal basilar bronchi,
the “truncus basalis”; and (f ) the peripheral basilar lower
lobe bronchi. The right upper lobe bronchus is termed
eparterial because it arises above the right main pul-
monary artery. It has a horizontal course usually originat-
ing just below the carina. Air characteristically marginates
the posterior wall of the right upper lobe bronchus as well
as the posterior wall of the bronchus intermedius. Less
commonly, an aberrant segmental vein can be identified
posterior to the bronchus intermedius draining into the
right superior pulmonary vein, most often originating
from the posterior segment of the right upper lobe and less
frequently from the superior segment of the right lower
lobe (60). Unlike the right upper lobe anterior and poste-
rior segmental bronchi that lie in the same plane as the CT
section, the apical segmental bronchus courses superiorly
and can be identified as a lucency superimposed on the
distal end of the right upper lobe bronchus.
Unlike the right upper lobe bronchus, the middle lobe
bronchus has an oblique course as it originates at the distal
Figure 5-12 Anatomic drawing illustrates terminology of the
central airways. (From Prakash UBS. Bronchoscopy. New York: end of the bronchus intermedius (Fig. 5-13). It is easily dis-
Raven Press; 1994, with permission.) tinguished from the origin of the right lower lobe bronchus
TABLE 5-2
SEGMENTAL AND SUBSEGMENTAL BRONCHI: NOMENCLATURE AND VARIANTS
Bronchi and Segments Subsegments Most Common Variants/Commentsa
Right lung
Right upper lobe
B1, apical a, apical B1 origin from B3 or less likely B2
b, anterior
B2, posterior a, apical B2 arising cephalad to B3 (56%), at same level as B2 or B2b (44%)
b, posterior
B3, anterior a, lateral Axillary subsegment formed by B2a or B3b (or both)
b, anterior
Middle lobe
B4, lateral a, lateral B4 and B5 equivalent in size (44%)
b, medial Trifurcation B4a, B4b, B5 (13%)
B5, medial a, superior B5 larger than B4 (27%)
b, inferior
Right lower lobe
B6, superior segmental a, medial B6c, B6a b bifurcation (60%)
bronchus b, superior
c, lateral
B*, subsuperior bronchus Variable, arises between B6 and B7 (30%)
Truncus basalis Bronchial trunk caudal to B6 giving rise to basal segments
B7, medial basilar bronchus a, anterior B7a b anterior to the right inferior pulmonary vein (60%)
b, medial
B8, anterior a, lateral Rudimentary B8 (10%)
b, basilar
B9, lateral a, lateral B8, B9 10 bifurcation (60%)
b, basilar B8, B9, B10 trifurcation (15%)
B10, posterior basilar bronchi a, posterior
b, lateral
c, basilar
Left lung
Left upper lobe Bifurcation into superior (B1 2, B3) and inferior trunks (B4, B5) (75%)
Trifurcation of left upper lobe into B1 2, B3, B4 5 (16%–25%)
B1 2, apicoposterior a, apical
b, posterior
c, lateral Variation in origin of B1 2c common
B3, anterior a, lateral B3 may arise anywhere between B1 2 and B4 5
b, medial
c, superior B3 poorly defined (25%)
B4, superior lingular bronchus a, lateral Well-developed B4a (40%); size varies depending on distribution of B3
b, anterior
B5, inferior lingular bronchus a, superior
b, inferior
Left lower lobe
B6, superior bronchus a, medial B6a, B6b c bifurcation (45%)
b, superior
c, lateral
B*, subsuperior bronchus Variable, arises between B6 and B7, a single stem (30%)
Truncus basalis Bronchial trunk giving rise to basal segments, longer than on right
B7 8, anteromedial bronchus B7a, medial Separate origin of B7 (5%)
B7b, lateral B7 8, B9–10 bifurcation (45%)
B8a, lateral B7 8, 9, 10 trifurcation (15%)
B8b, basilar
B9, lateral basilar bronchus a, lateral
b, basilar
B10, posterior basilar bronchus a, lateral
b, basilar
aPercentages shown are approximate.
From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations. Philadelphia: Lippincott Williams &
Wilkins; 2005, with permission.
A B
C D
Figure 5-13 Characteristic cross-sectional anatomy of the airways. A, B: Diagrammatic illustrations of the right and left bronchial trees,
respectively, each showing six characteristic axial sections illustrated in C to L. C–L: Axial sections corresponding to the diagrammatic rep-
resentations in A and B (see Table 5-2 for correlation). Tr, trachea; Ap Seg (B1), right upper lobe apical segmental bronchus; Ap P (B1 2), left
upper lobe apical-posterior segmental bronchus; Ant Seg (B3), right upper lobe anterior segmental bronchus; RULB, right upper lobe
bronchus; Post Seg (B2), right upper lobe posterior segmental bronchus; BI, bronchus intermedius; LMB, left mainstem bronchus; RMB, right
mainstem bronchus; Ant Seg Br (B3), left upper lobe anterior segmental bronchus; Ap-Post (B1 2), left upper lobe apical posterior
segmental bronchus; LULB, left upper lobe bronchus; LB, lingual bronchus; LLLB, left lower lobe bronchus; MLB, middle lobe bronchus;
RLLB, right lower lobe bronchus; Inf LB (B5), inferior lingular bronchus; LLLB, left lower lobe bronchus; Sup LB (B4), superior lingular
bronchus; Sup Seg (B6), left lower lobe superior segmental bronchus; Sup Seg Br (B6), right lower lobe superior segmental bronchus; Lat
Seg Br (B4), middle lobe lateral segmental bronchus; Med Seg Br (B5), middle lobe medial segmental bronchus; TB-RLL, truncus basalis,
right lower lobe; TB-LLL, truncus basialis, left lower lobe; AB-RLL (B8), right lower lobe anterobasilar segmental bronchus; MB-RLL (B7), right
lower lobe medial basilar segmental bronchus; LB-RLL (B9), right lower lobe lateral basilar segmental bronchus; PB-RLL (B10), right lower
lobe posterobasilar segmental bronchus; AB-LLL (B8), left lower lobe anterobasilar segmental bronchus; MB-LLL (B7), left lower lobe medial
basilar segmental bronchus; LB-LLL (B9), left lower lobe lateral basilar segmental bronchus; PB-LLL (B10), left lower lobe posterobasilar
segmental bronchus. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations.
Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.) (continued)
E F
G H
I J
by identifying the lobe spur. As it courses obliquely, the the right lower lobe also can be identified at this level origi-
distal aspect of the middle lobe bronchus appears to taper, nating from the posterior aspect of the right lower lobe
giving it a narrowed, triangular appearance. This should bronchus. Similar to the right upper lobe bronchus, the
not be misinterpreted as abnormal. The medial and lateral superior segmental bronchus also courses horizontally and
segmental bronchi arise from the distal middle lobe is generally well seen along its entire length.
bronchus and are generally easily recognized on sequential Identification of the basilar segmental airways is facili-
thin sections. Typically, the superior segmental bronchus of tated by following their course on sequential axial images.
K L
A reliable landmark is the finding of the medial basilar bronchus. At this point, the origin of the apical-posterior
segmental bronchus anterior to the right inferior pul- segmental bronchus can be identified as a circular lucency
monary vein. superimposed on the distal aspect of the upper portion of
the left upper lobe bronchus.
In distinction, a section obtained through the lower
Left-Sided Bronchial Anatomy portion of the left upper lobe bronchus will also include
the secondary carina marking the origin of the left lower
Similar to right-sided bronchial anatomy, there are also
lobe bronchus, identifiable as a triangular wedge of soft
six characteristic axial sections that can routinely be iden-
tissue located posterolaterally. At this level the left interlo-
tified through the left-sided airways (Fig. 5-13). These in-
bar pulmonary artery lies lateral to both the left upper and
clude (a) the distal trachea/carina and apical-posterior
lower lobe bronchi and typically has a smooth, rounded
segmental bronchus of the left upper lobe, (b) the
contour laterally. A circular lucency at the end of the left
left mainstem bronchus and proximal apical-posterior
upper lobe bronchus at this level indicates the origin of
segmental bronchus, (c) the distal left mainstem
the lingular bronchus. The superior segmental bronchus is
bronchus/proximal left upper lobe bronchus, (d) the dis-
also often seen coursing posteriorly at or near this level.
tal left upper lobe bronchus and proximal lingular
Unlike the superior segmental bronchus, however, the
bronchus, (e) the trunk of the proximal left lower lobe
anterior segmental bronchus to the left upper lobe is more
bronchus—the “truncus basalis,” and (f) the proximal
unpredictable: most often arising as a trifurcation from the
basilar segmental bronchi.
distal end of the left upper lobe bronchus, this same air-
Among these characteristic sections, the most complex
way may also arise from the apical-posterior segmental
anatomically are those through the left upper lobe
bronchus or even the proximal lingular bronchus. Despite
bronchus. Because of its complexity, it is possible to obtain
this, the anterior segmental bronchus is usually easily
multiple sections through this same airway. As a conse-
identified as it characteristically courses horizontally in
quence, the appearance of the left upper lobe bronchus
the same plane as the CT scan. Regarding the basilar
and of its branches is best described at two specific levels
airways, the only difference between the right and left sides
(corresponding to the third and fourth characteristic
is that on the left, the medial and anterobasilar segmental
sections described previously) (Fig. 5-13): through the
bronchi most often arise from a common trunk. Other-
upper portion of the left upper lobe bronchus, which
wise, the right and left lower lobe bronchi tend to be
includes the origin of the apical-posterior segmental
mirror images of each other.
bronchus, and through the lower portion of the left upper
lobe bronchus, which includes the lingula bronchus,
respectively.
Anatomic Variants
The key to identifying a section through the upper
aspect of the left upper lobe bronchus is to note that the In adults the most commonly encountered lung bud
posterior wall of the bronchus at this level is smoothly anomalies are either anomalous or supernumerary
concave and is further marked by a band of hazy increased bronchi (61). Additional anomalies, including agenesis,
density coursing across the left upper lobe bronchus congenital bronchial atresia (Fig. 5-14), congenital lobar
caused by a partial volume effect from the left main pul- emphysema, congenital cystic adenomatoid malforma-
monary artery as it passes over the left upper lobe tion, and bronchogenic cysts, may also be seen, although
A B
C D
Figure 5-14 Bronchial atresia. A, B: Axial images through the upper lobes show characteristic branching pattern of mucoid impaction.
Note that the peripheral portion of the left upper lobe is emphysematous. C, D: Volumetrically rendered coronal section and sagittal MIP,
respectively, show to better advantage the characteristic branching pattern of a blind-ending, mucoid-impacted left upper lobe bronchus as
well as the extent of peripheral emphysema and, in this case, apical consolidation. Involvement of the left upper lobe is characteristic of
bronchial atresia. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations. Philadelphia:
Lippincott Williams & Wilkins; 2005, with permission.)
less commonly. By definition, a displaced or anomalous (63), the following slightly modified classification has
bronchus is one that arises at a lower level than normal in proved especially clinically useful.
the bronchial tree. If this same bronchus also supplies the
same lung segment, it is more aptly labeled supernumer- 1. Anomalies arising from normal higher order bronchial
ary. Although several hypotheses have been advanced to divisions. This includes accessory superior segmental
account for these anomalies (62), as proposed by Wu et al. bronchi, for example, identifiable as two parallel
bronchi both supplying the superior segment of the the bronchus intermedius before the origin of the
right lower lobe, and axillary bronchi, identifiable as a superior segmental bronchus to the right lower lobe
supernumerary segmental bronchus supplying the and usually before the origin of the middle lobe
lateral aspect of the right upper lobe. bronchus (Fig. 5-19). The bronchus is then directed
2. Anomalies arising from sites typically lacking branches. caudally toward the mediastinum, hence the “car-
Included in this group are tracheal bronchi, accessory diac” appellation. The length of the cardiac bronchus
cardiac bronchi, and bridging bronchi (Figs. 5-15 and is variable, ranging from a short, blind-ending diver-
5-16). The most common of these is the tracheal or pig ticulum to a longer, branching structure. The short
bronchus (Fig. 5-15) (64–67). Present in up to 1% to type is usually a simple bronchial stump without
2% of normal individuals, tracheal bronchi exclusively associated alveolar tissue, whereas the longer subtype
arise on the right side and may involve displacement has been noted both with and without associated
of the entire right upper lobe, the apical segmental rudimentary alveolar tissue (73). Investigators have
bronchus only, or additional supernumerary lobar or speculated that this anomaly, in fact, represents a
separate apical segmental bronchi (68). Although this rudimentary accessory lobe. Rarely, this blind-ending
anomaly may superficially mimic the rarer entity of airway may serve as a potential reservoir for infec-
tracheal diverticula endoscopically, CT readily distin- tious material, leading to chronic inflammation and
guishes between these entities (Fig. 5-17) (69). hypervascularity that result in recurrent episodes of
Tracheal bronchi also need to be distinguished from aspiration pneumonitis or hemoptysis (73,74,76).
tracheoceles (Fig. 5-18). Resulting from foci of muscu- 3. Anomalies associated with abnormalities of situs. In this
lar weakness within the trachealis muscle, tracheoceles category, simple reversal of the right- and left-sided
have been divided into two groups: those with wide airways is the most commonly encountered anomaly
mouths, likely acquired, and those with narrow (Fig. 5-7). Less frequent are anomalies due to
mouths, likely congenital in origin. These lesions may bronchial isomerism, in which the pattern of airway
grow to be capacious and thus may serve as reservoirs branching and pulmonary lobation is identical in the
for debris, leading to dumping of their contents into two lungs (63,68). This includes bilateral left-sided
the tracheobronchial tree (70,71). airway anatomy, either as an isolated finding or asso-
Another rare congenital anomaly of the tracheo- ciated either with the venolobar syndrome or less
bronchial tree, the accessory cardiac bronchus is the commonly polysplenia, and bilateral right-sided air-
only true supernumerary anomalous bronchus (72), way anatomy. Right-sided airway isomerism is usually
with an incidence of less than 0.5% in the general associated with asplenia and severe congenital heart
population (73–75). It arises from the medial wall of disease and is only rarely seen in adults.
A, B C, D
Figure 5-15 Tracheal bronchus. Four images A–D: at contiguous levels in a patient with a tracheal bronchus, obtained with spiral
technique 7-mm collimation and reconstruction at 7-mm intervals. The tracheal bronchus (arrow, B) arises from the lateral tracheal wall. It
supplies the apical segment of the right upper lobe (B1), seen in cross section in A. The right upper lobe bronchus (RUL) gives rise to the
anterior (B3) and posterior (B2) segmental bronchi. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and
radiologic correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
A B
Figure 5-16 Bridging bronchus. A, B: Sequential axial images at

the level of the carina show an unusual configuration of the right
upper lobe bronchus originating from the distal trachea. C:
Coronal reconstruction shows that the right upper lobe bronchus
arises separately from the distal trachea, whereas the bronchus
intermedius arises from the proximal left mainstem bronchus. This
configuration is typical of a so-called bridging bronchus. (From
Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways:
functional and radiologic correlations. Philadelphia: Lippincott
C Williams & Wilkins; 2005, with permission.)
A B
Figure 5-17 Tracheal diverticulum. A, B: Axial and coronal multiplanar reconstructions (MPRs), respectively, show a well defined blind-
ending tracheal diverticulum originating from the lateral aspect of the mid trachea. This entity should be distinguished from both anomalous
bronchi, which typically occur nearer to the carina (see Fig. 5-15), and tracheoceles, which typically originate at a higher level (see Fig. 5-18).
(From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations. Philadelphia: Lippincott Williams
& Wilkins; 2005, with permission.)
A B
Figure 5-18 Tracheocele. A, B: Sequential axial images show characteristic outpouching of the posterolateral wall of the intrathoracic
trachea at the level of the thoracic inlet. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic
correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
CENTRAL AIRWAYS has the advantage of correlating with the approximate

range of visualization of the airways by routine FB.
Along with rapid advances in CT and magnetic reso-
Computed Tomography/Bronchoscopic
nance (MR) technology, the past decade has seen remark-
Correlations
able advances in the field of bronchoscopy, both in the
For purposes of analysis, the airways can be divided into development of new devices such as ultrathin broncho-
central airways, extending from the trachea to the segmen- scopes or endobronchial ultrasound (EBUS) as well as
tal bronchi, and peripheral airways, extending from the sophisticated interventional bronchoscopic techniques
subsegmental bronchi to the bronchioles. This approach (77,78). Accurate assessment of the relative value of
A B
Figure 5-19 Cardiac bronchus. A, B: Axial and coronal multiplanar reconstruction (MPR) images, respectively, show characteristic
appearance of an anomalous bronchus originating from the medial wall of the bronchus intermedius, coursing inferiorly. This may serve as a
reservoir for debris, resulting in repeated episodes of aspiration or hemoptysis. In this case note the presence of ill-defined infiltrate in the
superior segment of the right lower lobe resulting from aspiration. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways:
functional and radiologic correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
these various techniques necessitates a thorough under- patients with central airway obstruction, CT often allows
standing of CT-FB correlations (79). In this chapter, atten- distinction between central tumor and peripheral atelectasis
tion is focused on CT-FB correlations in evaluating and (Figs. 5-6 and 5-22) (87,88). In patients with evidence of
diagnosing central airway lesions; correlation between CT direct invasion of adjacent mediastinal organs, CT can
and FB in diagnosing peripheral lesions is discussed in identify those with unresectable tumor (Figs. 5-6, 5-22,
Chapter 6. 5-26, 5-27). CT also allows visualization of distal airways
that otherwise cannot be evaluated bronchoscopically
because of proximal airway obstruction (Fig. 5-28). Addi-
Principles of Interpretation
tionally, in patients with central obstructing tumors, CT can
Compared with CT, bronchoscopic evaluation of the be used to predict which patients will benefit from interven-
airways has numerous advantages. Most important, bron- tional bronchoscopic techniques, including selecting opti-
choscopy allows direct visualization of airway lumina to mal sites for transbronchial needle aspiration (TBNA),
the fifth generation, enabling identification of subtle identifying candidates for the placement of endobronchial
mucosal, submucosal, and endobronchial abnormalities. stents (Figs. 5-28 and 5-29), and defining the extent of
Bronchoscopy also allows acquisition of bacteriologic, extraluminal disease in candidates for potential laser
cytologic, and histologic material from endobronchial, therapy, for example (89,90).
peribronchial, and parenchymal sites with minimal risk Despite these advantages, important limitations with
(80). In select cases, bronchoscopy may provide localiza- regard to bronchoscopy have also been noted. CT is of lim-
tion of bleeding sites before surgery. In addition, an in- ited value, for example, in the detection of endobronchial
creasing number of therapeutic interventions may now be lesions smaller than 2 to 3 mm (Fig. 5-30). In a prospective
performed endoscopically, including traditional proce- study of 105 lesions in 98 patients with radiographically
dures, such as removal of mucus plugs or foreign bodies, occult squamous cell carcinoma (SCC) identified on the
as well as a growing number of more sophisticated appli- basis of positive sputum cytology, for example, Usuda et al.
cations, including the use of various stents to provide (91) showed that although 55 (53%) of these lesions were
airway patency. Despite these advantages, FB has impor- larger than 2 mm (including 22 characterized as either poly-
tant limitations, especially for those patients without poid or nodular, with the remainder appearing flat or as
evidence of endobronchial disease in whom a specific his- focal areas of bronchial irregularity), 27 lesions (26%)
tologic diagnosis is sought (80,81). proved to be smaller than 2 mm, and 23 lesions (22%) were
In comparing CT with FB, terminology appropriate for endoscopically invisible, being identified by bronchial
bronchoscopic evaluation may not be appropriate for CT. brushings alone. From these data, it is unlikely that CT, even
Descriptive terms such as mucosal or submucosal, in par- optimally performed, would have identified more than 20%
ticular, may be misleading (Fig. 5-20) (82,83). CT evidence of these lesions, rendering CT of limited value in routine
of smooth airway narrowing, for example, may be associ- screening of the central airways in patients with positive
ated endoscopically with endobronchial, submucosal, or sputum cytology. It should be emphasized that, although FB
peribronchial disease (Fig. 5-21). In fact, as a general rule, can detect subtle endobronchial disease not identifiable
CT is imprecise in characterizing abnormalities identified with CT, it should not be assumed that all such lesions are
bronchoscopically (82). Exceptions include the finding of a significant. Even in patients presenting with hemoptysis,
discrete polypoid endoluminal lesion (Figs. 5-22 and 5-23) investigators have shown that many such lesions are non-
or, less commonly, calcific endobronchial filling defects, as specific. In one prospective study evaluating both CT and FB
may occur, for example, in patients with aspirated foreign in 57 consecutive patients presenting with hemoptysis (3),
bodies (Fig. 5-24) or broncholiths. Also diagnostic is the focal endobronchial lesions could be identified broncho-
finding of a fat-containing intraluminal lesion characteris- scopically in a total of 18 (32%) cases. Although 6 of these
tic of an endobronchial hamartoma (Fig. 5-25). In the case lesions proved to be due to bronchogenic carcinoma,
of broncholithiasis, CT may prove more sensitive than FB all prospectively identified by CT, in the remaining 12 cases,
by identifying both intraluminal and peribronchial calcifi- transbronchial biopsies proved diagnostic in only 1, a pati-
cations distal to inflamed and narrowed airways rendered ent with mucosal Kaposi sarcoma.
inaccessible to bronchoscopy (84–86). As documented by
Conces et al. (84) in a study of 15 patients with proved
Transbronchial Needle Aspiration
broncholithiasis, whereas CT correctly localized 6 of 10
endobronchial lesions and 4 of 5 calcified peribronchial The ability of CT to provide lymph node “mapping” to
nodes, the diagnosis of broncholithiasis was established assist in the planning and performance of TBNA serves
endoscopically in only 5 cases. as a paradigm for the integration and correlation of
CT has proved of greatest value prior to bronchoscopy by CT and bronchoscopy (92). TBNA allows access to medi-
defining the precise location and extent of disease within astinal and hilar tissue through the bronchoscope, poten-
and surrounding the central airways in patients with known tially obviating more invasive procedures such as
or suspected lung cancer (Figs. 5-6, 5-22, 5-23, 5-26). In mediastinoscopy or thoracotomy. Although CT may be
Figure 5-20 CT classification of airway pathology. A–H: Diagrammatic representations of the appearance of various airway abnormalities
as seen on CT sections. In each case, abnormalities are represented as they would appear in axial, oblique (A, B), and sagittal planes. Note
that this classification makes no attempt to correlate CT findings with a diagnosis of mucosal or submucosal disease. With rare exception,
this distinction is typically impossible to establish. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and
radiologic correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
A B
C D
Figure 5-21 CT-bronchoscopic correlation: diffuse squamous cell carcinoma. See Color Figure 5-21D. A–C: Sequential axial CT images
through the carina, right upper lobe bronchus, and lower lobe bronchi, respectively, imaged with lung windows in a patient presenting with
a chronic cough show diffuse thickening of the airway walls, associated with extensive mediastinal and hilar adenopathy identifiable even
with wide windows, and small bilateral effusions. This appearance is nonspecific and compatible with a diffuse infiltrative or inflammatory
process involving the entire tracheobronchial tree. D: Flexible bronchoscopic images in this same case demonstrate diffuse mucosal thick-
ening and nodular irregularity covering most of the surface of the airways. Forceps biopsy revealed diffusely infiltrative squamous cell carci-
noma. In this case, CT and bronchoscopy proved complementary by disclosing different aspects of this patient’s clinically unsuspected unre-
sectable tumor. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations. Philadelphia:
indispensable by identifying optimal biopsy sites, CT To date, TBNA has most often been used to evaluate
may also be of value by identifying cases for which TBNA patients with suspected intrathoracic neoplasia, with most
is not indicated, as it has been shown that the yield of series reporting a diagnostic yield of between 60% and
TBNA in patients with negative CT scans is sufficiently 80% sensitivity for diagnosing and staging lung cancer
low to obviate this procedure (92a). CT can also help to (94–96,96a). In one multicenter trial, TBNA resulted in a
avoid potentially false-positive TBNA by identifying le- positive diagnosis in 48% of 279 patients with non–small
sions that abut the central airways, mimicking mediasti- cell lung cancer (NSCLC), and 62% of 81 patients with
nal or hilar adenopathy, thus making the procedure unre- small cell carcinoma. Importantly, TBNA also provided
liable as a method for staging. the only bronchoscopic specimen diagnostic for lung
A B
Figure 5-22 Central airway tumor—CT evaluation. A, B: CT sections through the lower trachea imaged with wide and narrow windows,
respectively, show evidence of a well-defined polypoid mass causing narrowing of the trachea lumen. In this case, even without intravenous
contrast, tumor is clearly identifiable (arrows in B), easily separable from peripheral atelectasis. In this case CT is complementary to bron-
choscopy by clearly delineating both the intra- and extraluminal extent of disease.
A B
C D
Figure 5-23 Central tumor: CT-bronchoscopic correlations. See Color Figure 5-23C,D. A, B: Sequential axial images through the right
mainstem bronchus show a well-defined intraluminal mass in this patient status post prior left pneumonectomy. C: Endoscopic view con-
firming the presence of a well-defined endoluminal tumor consistent with recurrent non–small cell lung cancer. Note the close correlation
between CT and bronchoscopy in this case. D: Endoscopic view following cauterization showing near complete removal of the tumor.
475
D
Figure 5-24 CT-bronchoscopic correlation: endobronchial foreign body. See Color Figure 5-24C,D. A, B: Magnified CT sections through
the right mainstem–right upper lobe bronchus imaged with wide and narrow windows, respectively, show an unusual-appearing intraluminal
filling defect in a patient presenting with bacterial pneumonia and empyema. C: Endoscopic view revealing a foreign body coated with se-
cretions partially obstructing the right main bronchus. This was successfully removed by forceps. D: Image of the same foreign body follow-
ing removal revealing this to be an aluminum foil packet. The patient subsequently admitted accidentally aspirating the packet containing
crack cocaine that he had hidden in his mouth while fleeing the police a month earlier. In this case, the unchanging nature of the endo-
bronchial density on sequential CT scans several days apart prompted the suspicion of a foreign body rather than the initial impression of
retained secretions. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations.
Figure 5-25 Endobronchial hamartoma. Section through the

bronchus intermedius shows a well-defined endobronchial mass
within which foci of both calcification and fat (arrow) can be identi-
fied. Histologically proved endobronchial hamartoma. (From
Williams & Wilkins; 2005, with permission.)
A B
C D
Figure 5-26 Central tumor: CT-bronchoscopic correlations. See Color Figure 5-26I. A–D: Sequential axial images through the left hilum
imaged with lung windows show a large mass causing marked narrowing/obstruction of the left lower lobe bronchus just distal to the origin
of the superior segmental bronchus (arrows in B and C) as well as the superior lingula segment (curved arrows in B and C). Note that this
mass crosses the left major fissure to involve both the left upper and lower lobes. E–H: Identical sections as shown in A–D, imaged with me-
diastinal windows. The precise margins of the tumor are well defined following a bolus of intravenous contrast. The subcarinal nodes are
slightly enlarged. I: Endoscopic view of the proximal left lower lobe bronchus confirms the presence of a well-defined intraluminal mass just
below the origin of the superior segmental bronchus (arrow). This case again illustrates the complementary nature of CT by delineating the
true extraluminal extent of tumor (compare with Figs. 5-6 and 5–21) (continued).
477
E F
G H
I
A B
C
Figure 5-27 CT-bronchoscopic correlation: tumor localization—adenocarcinoma of the lung. See Color Figure 5-27C. A, B: Sequential
contrast-enhanced CT images through the subcarinal space and aortic root, respectively, imaged with narrow windows, in a 48-year-old ex-
smoker presenting with two episodes of scant hemoptysis, shows extensive tumor throughout the mediastinum. Note that the carina is
splayed, and there is slight deformity of the medial and posterior aspect of both the right and left mainstem bronchi (arrows in A) as well as
marked narrowing of the middle lobe bronchus (arrowhead in B) and apparent obstruction of the proximal right lower lobe bronchus (arrow
in B). This appearance is nonspecific but most often indicative of diffuse extrinsic compression of the airways by extensive malignant
adenopathy. Note, as well, marked irregularity of the posterior wall of the left atrium (curved arrow in B). Although these findings are all
suggestive of unresectable tumor, evaluation requires more definitive confirmation. C: Sequential images obtained at the time of bron-
choscopy in this same patient reveal extrinsic compression of the bronchus intermedius resulting from enlarged adjacent nodes. In this case
there is also evidence of eruption of malignant subcarinal lymph node through the wall of the left main bronchus (arrows in C), resulting in
endobronchial tumor. Forceps biopsy in the left main bronchus revealed adenocarcinoma. Although the presence of malignant subcarinal
lymph nodes staged this patient as at least IIIA (based on an N2 node), confirmation of involvement of the (contralateral) left main bronchus
confirmed that there was T4 unresectable tumor. This case points out both the complementary and the competing nature of CT and bron-
choscopy for assessing the true extent of tumor. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radi-
ologic correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
cancer in 18% of patients (93). Similarly, in a study of Less well appreciated is the fact that CT also may be of
123 patients with documented malignancy evaluated by value as a guide to TBNA in patients with benign mediasti-
TBNA, Sharafkhaneh et al. (94) reported a sensitivity of nal or hilar disease (96–98). Morales et al. (97), in a study
69%, with significant correlations noted between a posi- of 51 consecutive patients with suspected sarcoidosis who
tive TBNA and cell type (p 0.001) and nodal size underwent both transbronchial lung biopsy (TBBx) and
(p 0.05). Comparable with other reported series, TBNA flexible TBNA of mediastinal or hilar nodes, showed that
proved most successful in diagnosing small cell carci- 23% of patients with stage 1 disease and 10% of patients
noma, less often successful in diagnosing non–small cell with stage 2 disease had their diagnoses established only
carcinoma, and least successful in diagnosing lymphoma with TBNA. Overall, the combined use of TBBx and TBNA
(95,96). allowed a histologic diagnosis in 83% of these cases. In
A B
C D
Figure 5-28 Central tumor—airway and esophageal invasion. A: Magnified image through the carina following a bolus of intravenous
contrast imaged with narrow windows shows extensive tumor infiltrating the mediastinum. There is marked narrowing and irregularity of
both the right and left mainstem bronchi (arrows in A) as well as narrowing and displacement of the esophagus (curved arrow). B: Image
obtained just below A showing apparent occlusion of the right mainstem bronchus with residual patency of the left mainstem bronchus. The
esophagus is barely discernible. C: Magnified view of the right lung several centimeters below A and B shows marked irregularity of the
distal bronchus intermedius and superior segmental bronchus indicative of extensive tumor infiltration. Endoscopic visualization of airways
distal to the right mainstem bronchus could not be obtained. D: Magnified view just below the carina obtained following a bolus of intra-
venous contrast shows stents in both the left mainstem bronchus and the esophagus placed after endoscopic confirmation of tumor
invasion of the carina and esophagus due to non–small cell lung cancer. This case illustrates the use of CT to define airways distal to those
identified at bronchoscopy, as well as determining the need for both airway and esophageal stents.
patients for whom TBNA is contemplated, CT may also of diagnosing sarcoid in 7 patients who underwent
prove cost-effective. In the same study, 9 of 51 patients who both TBNA and TBBx, improving the diagnostic rate by
would otherwise have required surgical assessment to 47% (96).
obtain a definitive diagnosis were successfully diagnosed CT has also proved of value in patients with acquired
by TBNA; thus, in the judgment of these investigators, the immunodeficiency syndrome (AIDS). In immunocompro-
results more than financially justify the routine prebron- mised patients, the finding of low-density mediastinal or
choscopic use of HRCT (97). Similar findings have been hilar lymph nodes, in particular, strongly correlates with
reported by Cetinkaya et al. (96) in a study of 60 consecu- active mycobacterial infection (100). In these patients,
tive patients with mediastinal or hilar adenopathy in which TBNA may represent the least invasive means of obtaining
TBNA with a 22-gauge Wang cytology needle successfully diagnostic material, especially when sputum smears prove
diagnosed tuberculosis (TB) in 13 (65%) of 20 cases, pro- nondiagnostic. As documented by Harkin et al. (98), TBNA
viding the sole means of diagnosis in 8 (40%). Similar proved diagnostic in 23 (51%) of 45 procedures performed
findings have been reported by Trisolini et al. (99), who in 42 human immunodeficiency virus (HIV–positive)
reported that TBNA revealed non-necrotizing granulomas patients. This included identification of 21 of 23 docu-
in 23 (72%) of 32 patients with subsequently documented mented cases of mycobacterial infection, of which TBNA
sarcoidosis. Importantly, TBNA provided the sole means provided the only diagnostic specimen in 13 (57%) (98).
A B
Figure 5-29 Squamous cell carcinoma of the lung: endobronchial stenting. A, B: Axial CT sections through the distal carina and proximal
mainstem bronchi, respectively, show extensive mediastinal tumor causing eccentric narrowing of the distal trachea, near complete
obstruction of the right mainstem bronchus, and apparent extrinsic compression of the medial aspect of the left mainstem bronchus. Note
that there is also slight nodular irregularity of the anterior wall of the left mainstem bronchus, suggesting possible tumor infiltration. C, D:
Axial and coronal images, respectively, following successful insertion of a stent in the distal trachea and left mainstem bronchus. In this non-
surgical case, palliation was best achieved by insuring patency of the left mainstem bronchus. (From Naidich DP, Webb WR, Grenier PA,
et al. Imaging of the airways: functional and radiologic correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
(continued)
C D
A B
Figure 5-30 CT-bronchoscopic correlation: sarcoidosis. See Color Figure 5-30C,D. A, B: CT sections through the right and left mainstem
bronchi, respectively, imaged with wide windows show well-defined focal lesions (arrows) protruding into the proximal right main (A) and
proximal and distal left main (B) bronchi, respectively. Note the presence of diffuse nodular infiltrates in both lungs with a predominantly
perilymphatic distribution. C, D: Bronchoscopic images obtained the next day revealed no evidence of focal endobronchial lesions in the
right (C) or left (D) main bronchi but did show focal areas of fine mucosal nodularity unsuspected from the CT (not shown). Presumably the
large “lesions” shown by CT represented secretions that were no longer present by the time of bronchoscopy, although the mucosal nodu-
larity was too subtle to be detected by CT. In this case, although CT disclosed findings characteristic of parenchymal sarcoidosis, apparent
endobronchial lesions proved to represent retained secretions. In the setting of focal endobronchial lesions, performing a repeat CT follow-
ing strenuous coughing may be of value by minimizing the number of false-positive examinations. (From Naidich DP, Webb WR, Grenier PA,
et al. Imaging of the airways: functional and radiologic correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
C D
Alternate Methods for Performing Transbronchial Computed Tomography Fluoroscopy

Needle Aspiration and Virtual Bronchoscopy
Despite the proved efficacy of TBNA as a diagnostic tool, One approach has been to perform TBNA under direct CT
unfortunately, this technique remains underutilized. In a guidance (Fig. 5-31). CT fluoroscopy, in particular, has the
survey of 871 North American bronchoscopists, only 12% advantage of real-time cine visualization of the tip of the
routinely used TBNA to diagnose malignancy and only 2% needle. Rong and Cui (102), for example, compared their
to diagnose benign disease (101). Lack of utilization prima- diagnostic yield first performing routine TBNA without
rily reflects the widely disparate results reported by most and then under direct CT guidance and found this tech-
bronchoscopists. As a consequence, a variety of techniques nique improved their yield from 20% to 60%. Using CT
have been recently developed in an attempt to improve the fluoroscopy it has been demonstrated that 41% of 116
diagnostic yield of TBNA. needle passes were improperly positioned, and subsequent
Figure 5-31 CT-guided trans-

bronchial needle aspiration (TBNA).
A: Scanogram obtained during CT-
guided bronchoscopy shows the tip
of the TBNA needle extending into
the subcarinal space. B: Magnified
CT section obtained in the same
patient following repositioning of
the bronchoscope confirms that the
tip of the needle is within the center
of an enlarged right paratracheal
node. Despite CT guidance, in this
case TBNA proved nondiagnostic.
Subsequent mediastinoscopy con-
A, B firmed non–small cell lung cancer.
passes established diagnoses in enlarged lymph nodes in all differentiate the various layers of the bronchial wall.
12 patients studied (103). All patients in this group had pre- Specifically, as many as seven layers within the wall of large
viously undergone a nondiagnostic TBNA by the conven- airways have been identified, including two of the inner
tional method. Similar results have been reported by White mucosa, three of cartilage, and two of the outer adventitia
et al. (104), who reported a yield for CT fluoroscopy–guided (Fig. 5-32). Vessels can be reliably identified by their
TBNA of 83% in 12 patients, all of whom had also under- pulsation and low internal echo density and are easily
gone a nondiagnostic TBNA by the conventional method. differentiated from lymph nodes as small as 2 mm that can
Additional alternate approaches have made use of VB to be identified by their high echo density. In distinction,
plan for optimal biopsy sites (105–107). VB as a method hypoechoic areas and invasion of the cartilage layer have
for preplanning TBNA, for example, has been shown to been shown to characterize malignant tissue (112).
improve the yield of TBNA (108,109). In one study (110), As a result, EBUS provides a unique method for assessing
volumetric rendering was used to first highlight a total of pathologic changes both in and adjacent to the airways.
36 sites in 14 patients in whom enlarged nodes could be Takahishi et al. (123), for example, in a study of 22 patients
identified on routine 3-mm axial images. Although VB did with radiographically occult lesions, have shown that
not alter the approach to subcarinal or aorticopulmonary EBUS can evaluate the depth of tumor invasion of airway
window nodes, virtual bronchoscopic imaging with nodal walls differentiating tumors that remained separate from
highlighting increased the successful choice of biopsy sites those extending to the cartilaginous layer with a sensitiv-
for hilar nodes (from 37% to 83%; p 0.001) and for pre- ity of 86%, a specificity of 67%, and an accuracy of
tracheal nodes (from 59% to 85%; p 0.07), respectively. 80%. Despite its overall accuracy for detecting bronchial
pathology, CT is incapable of such fine distinctions. Herth
Endobronchial Ultrasound et al. (124), for example, comparing the ability of CT versus
Over the past decade, considerable interest has focused on EBUS to distinguish extrinsic airway compression from
the potential of EBUS to diagnose both central (111–120) tumor invasion in 105 patients, confirmed the superior
and peripheral lesions (120a,121,122). In comparison to sensitivity (89% vs. 75%), specificity (100% vs. 28%), and
all other imaging modalities, including direct endoluminal accuracy (94% vs. 51%) of EBUS. This compares favorably
visualization, EBUS has the great advantage of being able to with a previous report of 93% accuracy for endobronchial
A B
Figure 5-32 A–E: Endobronchial ultrasound (EBUS). See Color Figure 5-32C,D. A: In vitro endobronchial ultrasound appearance of the
normal trachea showing five identifiable portions of the airway wall. From inside out these include one hyperechoic layer representing mu-
cosa and submucosa, a second anechoic layer representing cartilage, and a third echogenic layer representing the adventitia. As illustrated
in this figure, these layers measure 1, 1.1, and 0.9 mm, respectively. Note that the endobronchial probe lies eccentrically within the tracheal
lumen and that the definition of the tracheal wall is nonuniform, resulting from lack of uniform contact between the probe and the mucosa.
(From Herth F, et al. Endobronchial ultrasound improves classification of suspicious lesions detected by autofluorescence bronchoscopy.
J Bronchol. 2003;10:249–252, with permission.) B: Endoscopic ultrasound image shows evidence of asymmetric wall thickening measuring
3 mm (vs. normal 1.4 mm), in this case due to tumor confined to the bronchial wall without evidence of adenopathy. C, D: Endoscopic view
of a bulky adenoid cystic carcinoma prior to and following laser resection. In this case, laser resection appears to have eradicated the tumor.
E: Follow-up endobronchial ultrasound examination shows that there is residual tumor eccentrically located in the anterolateral tracheal wall
(arrow). These cases illustrate the advantage of EBUS for exquisite delineation of tracheal and bronchial wall anatomy. (B–E courtesy of
Heinrich Becker, MD, University of Heidelberg, Germany.)
C D
E Figure 5-32 (continued)
ultrasonographic evaluation of tracheobronchial invasion well as limitations due to the necessity for endoscopic
by tumor in 15 surgically resected patients (111). probes to be specially configured to provide consistent
Not surprisingly, given the ability of EBUS to accurately positioning adjacent to bronchial walls. Although a dual-
identify nodes as small as 2 mm adjacent to the central channel bronchoscope has been developed, allowing direct
airways, attention has focused on the use of this technique visualization of the TBNA needle entering lymph nodes
for guiding TBNA. To date, numerous reports have docu- directly, this scope is currently limited to this one use only.
mented the efficacy of EBUS for staging lung cancer, in
particular (113,114,116–120,125–129), as well as in the Therapeutic Correlations
diagnosis of benign disease (130,131). Herth et al. (117), In addition to the use of CT for detecting bronchial lesions
for example, using a dedicated EBUS probe, reported and guiding diagnostic procedures, CT also plays an essen-
successful sampling by EBUS-guided TBNA of lymph nodes tial role in treatment planning. This is especially important
ranging from 8 to 43 mm in 207 (86%) of 242 patients given the wide range of potential interventional broncho-
with malignancy confirmed in 72% (117). In a follow-up scopic techniques now available for treating both benign
randomized trial, EBUS-guided TBNA was found to be and malignant airway disease, making interventional bron-
significantly better than conventional TBNA for all nodal choscopy an essential option for treating endobronchial
stations combined except those in the subcarinal space disease (9). In this category are a number of different
(84% vs. 58%) (132). Currently, the major limitations of and competing techniques, including neodymium:
EBUS-guided TBNA are the expense of this procedure, as yttrrium-aluminum-garnet (Nd:YAG) laser phototherapy,
A, B C
Figure 5-33 CT-bronchoscopic correlation: use in therapeutic planning—stent evaluation. See Color Figure 5-33B,C. A: CT of a patient
with esophageal carcinoma who had previously required endobronchial stent placement in the left main bronchus resulting from extrinsic
compressive obstruction of the bronchus by the primary tumor. CT was obtained to evaluate dyspnea 4 weeks after an uncovered self-
expanding metal stent had been placed within the left mainstem bronchus. Note that in addition to extensive posterior mediastinal tumor,
there is soft tissue density filling the lumen of the stent (arrow). B: Flexible bronchoscopy confirmed the presence of tissue obstructing the
stent. Forceps biopsy revealed invasive esophageal carcinoma. C: Flexible bronchoscopic appearance of the left main bronchus after re-
moval of the obstructing tumor by argon plasma coagulation via the flexible bronchoscope. Patent distal airways are seen (arrow). Because
of the electroconductive nature of metal stents, this technique is safer than electrocautery within the lumen of a stent and is much less likely
to damage the stent than is the Nd:YAG laser. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radio-
logic correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
photodynamic therapy, cryotherapy, electrocautery, brachy-

therapy, and argon plasma coagulation (77). In these
TABLE 5-3
cases, the main role of CT is to delineate the true extralumi-
CLASSIFICATION OF DISEASES OF THE
nal extent of tumor (Fig. 5-23). CT also plays an especially
important role in the use of airway stents by predetermin- TRACHEA AND MAINSTEM BRONCHI
ing their optimal diameter and length (78). As shown by Focal disease
Miyazawa et al. (132a), in conjunction with spirometry, Tracheal strictures
ultrathin bronchoscopy, and EBUS, multiplanar CT images Congenital
are of use in determining the choke point of flow limita- Acquired
tion in malignant endobronchial lesions prior to and Postintubation
following stent placement. CT is also of proven value for Infection (tuberculosis, fungal disease)
follow-up evaluation to identify stent migration or obstruc- Benign neoplasms
tion resulting from excessive granulation tissue or tumor Commonest
recurrence (Fig. 5-33) (37). Squamous cell papilloma (solitary)
Laryngotracheobronchial papillomatosis
Primary malignant neoplasms
Abnormalities Involving the Trachea Commonest
and Mainstem Bronchi Squamous cell carcinoma
Adenoid-cystic carcinoma
Identification of tracheal disease is notoriously difficult. Carcinoid tumors
Clinically, it has been estimated that lesions need to
Secondary malignant neoplasms
occlude more than 75% of the lumen before the develop-
ment of symptoms related to airway obstruction, and even Diffuse disease
then these are frequently mistaken as being due to asthma. Increased diameter
Tracheobronchomegaly
Early diagnosis is rare and usually occurs as a result of Decreased diameter
hemoptysis. Radiographically, the trachea has been Relapsing polychondritis, amyloidosis, sarcoidosis, Wegener
described as the blind spot in the chest, with lesions granulomatosis, tracheopathia osteochondrolytica, ulcerative
characteristically identified only when large (133). For colitis, saber-sheath trachea, infection
descriptive purposes, this section addresses first focal and Adapted from Kwong JS, Muller NL, Miller RR. Disease of the trachea
then diffuse tracheal abnormalities, with an acknowledg- and main-stem bronchi: correlation of CT with pathologic findings.
ment that this distinction may be arbitrary (Table 5-3). Radiographics. 1992;12:645–657, with permission.
Focal Tracheal Abnormalities commonly, in shearing of the mainstem bronchi from the
carina (Fig. 5-34). Although the entire length of the tra-
Tracheal Trauma chea is vulnerable following blunt trauma, more than
It is unusual for patients suffering tracheal trauma to live 80% of cases occur within 2.5 cm of the carina (134). The
long enough to be evaluated: not surprisingly, therefore, left mainstem bronchus is less frequently disrupted, likely
most series are limited in number (134). Trauma may be because of its longer course through the mediastinum,
either penetrating or more often blunt (Fig. 5-34). which offers some extra measure of protection. Additional
Although penetrating tracheal injuries typically involve causes of tracheal rupture include overdistension of endo-
the neck, blunt traumatic injuries usually involve the tracheal tubes and traumatic intubation (136).
intrathoracic portions of the trachea, often extending to Chest radiographic findings most often include nonspe-
the mainstem bronchi (135). Blunt trauma may also cific pneumomediastinum and/or deep cervical emphy-
result in pulmonary contusion or laceration, thoracic sema (137). Pneumomediastinum most commonly occurs
spine injuries, and esophageal disruption, as well as secondary to air leaks originating in the pulmonary intersti-
multiple fractures. Extrathoracic organ injury is also tium, the so-called Macklin effect (138). Pneumothoraces
frequently encountered (135a). Most often the result of are also common, presumably the result of disruption of
deceleration or so-called steering wheel injuries, proposed the mediastinal pleura. In those cases in which the airway
mechanisms of injury include compression of the airways is partly occluded by blood clots or compression by an
between the sternum and thoracic spine with lateral adjacent mediastinal hematoma, the first sign of tracheal
expansion of the thorax, shearing at fixation points espe- rupture may be a persistent air leak following chest tube
cially at the carina, and elevated intratracheal pressure, insertion. In more severe cases, bronchial avulsion may
especially against a closed glottis (135b). These result lead to a falling away of the involved lung from the hilum
either in transverse lacerations, in disruption of the due to disruption of central attachments, the so-called
posterior, membranous portion of the trachea or, less fallen lung sign (137).
Figure 5-34 Blunt trachea trauma:

imaging techniques. See Color
Figure 5-34D. A, B: Multiplanar
A, B (MPR) and minimum projection
(MinIP) coronal reconstructions, re-
spectively, through the trachea show
extensive mediastinal and deep cer-
vical emphysema. Note the absence
of a pneumothorax. C: Axial image
through the mid trachea shows a
focal defect in the posterior membra-
nous trachea with resultant exten-
sive pneumomediastinum. D: Virtual
bronchoscopic image pointing from
within the mid trachea toward the ca-
rina shows a deep groove in the pos-
terior membranous trachea (arrow)
extending toward the origin of the
right mainstem bronchus. Sub-
sequent bronchoscopic evaluation
confirmed the presence of a mem-
branous tear corresponding anatomi-
cally to the virtual bronchoscopic
image. (Case courtesy of James
C, D Gruden, MD, Phoenix, Arizona.)
CT findings in patients with blunt trauma have been In addition to postintubation stenoses, tracheobronchial
described (138–142). In one retrospective study of 14 strictures also occur as a result of infection, in particular, TB
patients with tracheal rupture either due to trauma and histoplasmosis, and after lung transplantation. After
or following intubation, CT proved diagnostic in 85% of lung transplantation, complications commonly occur at the
cases, providing direct visualization of tracheal tears in level of the bronchial anastomosis and include dehiscence,
10 (71%). Direct signs of tracheal rupture included focal necrosis, and stenosis from malacia, fibrous stricture,
wall defects in more than 50% of cases (Fig. 5-34) and, less prominent granulation tissue formation, or a combination
commonly, a deformed tracheal lumen. In this same of these conditions (35,146–153).
series, overdistension of the tube balloon was noted to Regardless of the etiology, CT, especially when per-
be particularly common, occurring in 5 of 7 intubated formed using volumetric techniques, has proved a reliable
patients. Of particular value was the finding of herniation method for assessing tracheobronchial strictures (Fig. 5-35)
of the endotracheal tube balloon outside the confines of (1,154–158). In all cases, accurate evaluation requires that
the tracheal wall. MPR or VB images should be obtained in the precise length and degree of stenosis be assessed as well
all cases in which tracheal injuries are suspected, as these as associated peritracheal abnormalities. As documented by
may disclose subtle abnormalities not readily identified in Whyte et al. (158), in their study of 25 patients with known
cross section (Fig. 5-34). or suspected stenosis, CT with MPRs demonstrated the site
Despite the use of advanced imaging techniques, defini- and degree of tracheal or main bronchial stenoses with a
tive diagnosis and therapy nearly always require bron- sensitivity of 93% and a specificity of 100%. As discussed in
choscopy. Rare exceptions include patients in whom the the section on trauma, in select cases the degree of stenoses
diagnosis can be made presumptively owing either to may also be accurately judged by use of VB (40).
direct observation or to evidence of a massive air leak Although endoscopic management of tracheal stenoses
following chest tube placement (143). Although treatment has been proposed, especially the use of lasers or stents,
typically involves surgical repair, conservative therapy has these typically provide only temporary benefits. Primary tra-
been advocated for individuals with postintubation mem- cheal and/or laryngotracheal resection and reconstruction
branous tracheal ruptures if these are less than 2 cm in remain the optimal method for treating non-neoplastic
length (136,144). Late manifestations of tracheal rupture stenoses, with resection up to 4 cm in length feasible, the
include wheezing secondary to tracheal stricture, poten- equivalent of eight cartilaginous rings (159). In one series of
tially mistakenly diagnosed as asthma. 58 cases of postintubation or idiopathic stenoses, major
complications occurred in 12%, including anastomotic
Tracheal Stenosis dehiscence and vocal cord paralysis (159).
Benign airway strictures may be either congenital or
acquired. In either case, accurate evaluation requires that Tracheal Neoplasia
the precise length and degree of stenosis be assessed along Etiology. Although many different tracheal tumors, both
with associated peritracheal or bronchial abnormalities. benign and malignant, have been reported, primary
Among acquired strictures, the most common cause is tracheal neoplasia is rare (160). These tumors include both
iatrogenic injury resulting from intubation. Although primary epithelial and mesenchymal neoplasms and sec-
investigators have estimated that complications occur in ondary neoplasia due to either metastases or, more com-
up to 10% to 15% of patients after intubation and in up monly, direct tracheal invasion by adjacent mediastinal
to 40% of patients after tracheotomies, the incidence of neoplasms, especially carcinomas arising from the lung,
serious complications after intubation has decreased thyroid, and esophagus (161). In fact, only a few lesions
dramatically since the introduction of high-volume/low- occur with any frequency, with SCCs and adenoid cystic
pressure tubes (145). The two principal sites of postintu- carcinomas (AACs) together accounting for up to 86%
bation stenosis are at the stoma or at the level of the of cases (162). Other, relatively rarer malignant lesions
endotracheal tube or tracheostomy tube balloon (48). In encountered include carcinoid and mucoepidermoid
both cases, strictures result from pressure necrosis, which tumors, whereas hamartomas and bronchial papillomas are
causes ischemia and subsequent scarring. Stenosis also the most common benign tumors encountered. Although
may be due to inflammation with subsequent thinning clearly distinct histologically, most of these entities clini-
and weakening of the tracheal wall (tracheomalacia) cally present with nonspecific symptoms including cough,
in the segments between the stoma and the cuff; narrow- wheezing, stridor, and hemoptysis, and radiographic and
ing may also be caused by granulation tissue resulting CT appearances of both benign and malignant tumors
from direct tracheal injury by the endotracheal or tra- similarly tend to overlap.
cheostomy tube tip (48). Other complications of intuba-
tion include tracheal rupture and tracheo-brachiocephalic Squamous Cell Carcinoma. SCC most often occurs in
artery or tracheoesophageal fistulas, the latter usually middle-aged male smokers, and not surprisingly it has
occurring in the presence of combined tracheal and eso- been found to be associated with other malignant diseases
phageal intubation. of the respiratory tract in as many as one third of patients
A, B
C
Figure 5-35 Tracheal stenosis following intubation. A: Axial section at the

level of the thoracic inlet shows markedly increased soft tissue density internal
to the tracheal cartilage causing irregular narrowing of the tracheal lumen. B, C:
Multiplanar (MPR) and 3D surface rendered coronal reconstructions, respec-
tively, through the trachea show to good advantage the true extent of the
stricture, information essential for optimal surgical planning. D: Comparison of
the appearance of tracheal stenosis caused by granulation tissue and a normal
trachea. Soft tissue internal to the tracheal cartilage is irregularly thickened.
(A–C Courtesy of Phillip Boiselle, MD, Boston, Massachusetts; D from Webb
EM, Elicker BM, Webb WR. Using CT to diagnose nonneoplastic tracheal abnor-
malities: appearance of the tracheal wall. AJR Am J Roentgenol. 2000;174:
D 1315–1321, with permission.)
(Fig. 5-36) (162). In approximately 10% of cases, these

tumors prove to be multifocal and are frequently found to
extend into the mainstem bronchi or adjacent esophagus,
resulting in malignant esophageal-bronchial fistulas (162).
Sialadenoid Tumors. Sialadenoid tumors are pulmonary

analogs of the salivary glands arising from tracheo-
bronchial mucous glands and most often prove to be AACs.
The second most frequent lesions in this group are
mucoepidermoid cancers, making up approximately 5% of
the total. Less commonly encountered are mucous gland
adenomas, pleomorphic adenomas, and acinic cell carcino-
mas, and still more rare are pulmonary oncocytomas and
myoepitheliomas (163–165). AACs are low-grade lesions
that typically arise from the tracheobronchial mucous
glands, most frequently on the posterolateral wall of the
trachea (Figs. 5-29 and 5-37). Along with SCC, ACC shows Figure 5-36 Squamous cell carcinoma. Axial section shows soft
a marked propensity for local, submucosal, and perineural tissue mass involving the carina, resulting in marked narrowing of
both the right and especially the left mainstem bronchi. Biopsy
lymphatic invasion. Metachronous hematogenous metas- proved squamous cell carcinoma. (Case courtesy of Dr. Moulay
tases have been reported to be common, occurring in Meziane, Cleveland, Ohio.)
A, B
C
Figure 5-37 Adenoid cystic carcinoma. A, B: Axial CT sections through the mid trachea in two different patients, both with focal adenoid
cystic carcinomas. Note that in each case there is an eccentric soft tissue mass involving only a portion of the tracheal wall with variable de-
grees of tracheal narrowing and extension into the adjacent soft tissues of the mediastinum. In both cases, CT clearly shows limited
extraluminal invasion, confirming that these lesions are resectable. C: Multiplanar reconstruction in a different patient than shown in either A
or B shows a lesion marginating the left side of the trachea, causing slight irregularity and narrowing of the tracheal lumen. (From Naidich
DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations. Philadelphia: Lippincott Williams & Wilkins;
2005, with permission.)
44% of cases in one series (166). Unlike SCCs, but in com- 5 years and usually is restricted to the larynx, with sponta-
mon with mucoepidermoid tumors and carcinoids, no neous resolution noted in most cases. Furthermore, papillo-
known association exists between ACCs and cigarette matosis has been shown to be etiologically linked to
smoking and there is no predilection for either sex. ACC infection with human papillomavirus, either contracted at
typically occurs in slightly younger patients with a mean birth or acquired through sexual transmission. Papillo-
age of between 40 and 50 years. matosis has been noted to occur in patients with AIDS.
Widespread disease occurs in a few patients. As shown by
Carcinoid Tumors. Carcinoid tumors are neuroendocrine Kramer et al. (172) in a literature review of 532 patients,
neoplasms that arise from bronchial or bronchiolar epithe- tracheal and bronchial involvement occurs in only approxi-
lium. They are presumed to arise either from existing mately 5% of cases, whereas pulmonary parenchymal
Kulchitsky cells, neuroepithelial bodies, or pluripotential disease occurs in less than 1%. Unfortunately, in this
bronchial epithelial stem cells. They present a spectrum of population, spontaneous remission is unusual, with malig-
histologic subtypes ranging from low- to high-grade neo- nant transformation to SCC reported in adults. The origin
plasms and are classified as typical carcinoids, atypical carci- of parenchymal disease is controversial. Because tracheo-
noids, and large cell neuroendocrine carcinomas, with the bronchial and parenchymal disease is rarely encountered
highest grades merging with small cell carcinoma (167). In without coexistent laryngeal disease, investigators have pos-
addition to findings related to bronchial obstruction, carci- tulated that spread is by endobronchial dissemination. An
noid tumors cause hemoptysis in up to 50% of cases, association between distal spread and prior tracheostomy
reflecting their rich hypervascular stroma. In addition, carci- further supports this cause. Typically, lesions appear first
noid tumors are known to produce a variety of hormones as well-defined nodules, occasionally identifiable as cen-
and neuroamines, including adrenocorticotropic hormone trilobular, that subsequently undergo central necrosis
(ACTH), serotonin, and somatostatin, among others. resulting in multiple thin-walled cavities. Bronchiectasis,
Cushing syndrome has been reported, although rarely. with or without parenchymal consolidation and atelectasis,
Carcinoid syndrome may also occur, although only in is often identified in association with nodular densities, pre-
patients with known liver metastases (167). sumably the result of bronchial occlusion and subsequent
Most endobronchial carcinoid tumors are central in inflammation or infection.
location. Easily identified bronchoscopically, they typically Another lesion rarely encountered is an endobronchial
appear as smooth, cherry-red lesions with a propensity to hamartoma. Commonly presenting as a solitary pul-
bleed on biopsy (Fig. 5-38). Although massive hemoptysis monary nodule (Fig. 5-25), endobronchial hamartomas
has been reported, in fact, these lesions are usually safely are rare, estimated to represent between 10% and 20%
biopsied. They tend to exhibit variable growth patterns. of intrathoracic hamartomas (174). Hamartomas are mes-
Most often they extend extraluminally, resulting in so- enchymal lesions composed of a mixture of cartilage, fat,
called iceberg lesions. With more peripheral lesions, airway and fibrous tissue and most often occur in segmental
obstruction may result in focal mucoid impaction. bronchi; tracheal involvement is distinctly unusual. In a
On CT, endobronchial carcinoid tumors most often report of 43 patients with documented endobronchial
present either as hilar masses or discrete intraluminal hamartomas, the most common presenting symptoms
lesions (Fig. 5-38). A characteristic triad of features has were recurrent respiratory tract infections and hemoptysis,
been described, including the finding of a well-defined occurring in 44% and 33% of cases, respectively (174). On
round or slightly lobulated tumor causing narrowing of CT, although infrequently reported, the presence of fat is
the adjacent airway, associated with eccentric calcifications considered diagnostic of this entity (Fig. 5-15) (175,176).
(168). These latter are commonly seen, with calcifications
identifiable in up to a third of cases (169,170). Carcinoid Tracheal and Mainstem Bronchial Neoplasms: CT
tumors are also well known to markedly enhance follow- Findings. Not surprisingly, the CT appearances of tracheal
ing administration of intravenous contrast media. neoplasms overlap. Indeed, regardless of cell type, these
tumors have in common the finding of an endobronchial
Benign Tracheal Neoplasms. The most common benign tra- soft tissue mass or asymmetric tracheal or bronchial wall
cheal neoplasm is squamous cell papilloma. Although, thickening causing eccentric narrowing of the tracheal or
like SCCs, these most often occur in middle-aged male carinal lumen (Figs. 5-29 and 5-36 to 5-39) (177).
smokers, squamous cell papillomas typically involve the Asymmetric tracheal or bronchial wall thickening may be
larynx or bronchi, with tracheal involvement uncommon: exceedingly subtle, especially in patients with benign
More important, involvement, when present, is limited to lesions such as papillomas. Because retained secretions
the tracheal wall (161). may also have this appearance, the examiner often needs
Laryngotracheobronchial papillomatosis refers to a to acquire additional images after requesting patients to
condition associated with multiple papillomas (Fig. 5-39) cough (Fig. 5-40). Although the presence of calcification
(171–173). Histologically similar to solitary papillomas, should raise the suspicion of a mesenchymal tumor,
this condition most often occurs in children younger than especially chondromatous lesions, definitive diagnosis still
A B
C D
Figure 5-38 A: Axial image through the distal left upper lobe bronchus shows a well-defined, discrete intrabronchial lesion (arrow). B, C:
Coronal and sagittal multiplanar reconstructions (MPRs), respectively, and (D) virtual bronchoscopic image show this same lesion to good
advantage (arrow in B and C), again confirming its endobronchial location. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the air-
ways: functional and radiologic correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
A B
C D
Figure 5-39 Laryngotracheal papillomatosis in a patient with acquired immunodeficiency syndrome (AIDS). A, B: Axial CT images
through the upper and mid trachea, respectively, show marked nodular irregularity of the tracheal lumen resulting from multiple polypoid
lesions. C, D: Axial CT sections show typical appearance of diffuse laryngotracheal papillomatosis resulting in scattered nodules, some of
which are cavitary. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations.
Figure 5-40 Retained secretions.

A, B: Magnified views of the carina
imaged with wide and narrow
windows, respectively, show a well-
defined filling defect in the left
mainstem bronchus mimicking the
appearance of an endobronchial
tumor (arrow) (compare with Fig.
5-38). C, D: Magnified views at the
same levels as shown in A and B
following cough. On occasion
retained secretions and/or endolu-
minal hematomas may simulate
A, B neoplasia (continued ).
C, D Figure 5-40 (continued)
requires histologic evaluation in cases of focal tracheal smoothly narrowed airways without evidence of either
disease. However, whereas endoscopic biopsy can reliably wall thickening or active peribronchial inflammation
establish the diagnosis, adequate preoperative evaluation (178,181–183). Unlike active infection, which involves
requires accurate definition of the extent of extraluminal both mainstem bronchi equally, fibrotic TB usually in-
involvement if present. This is especially important volves the left main bronchus, perhaps because of its
because as much as one half of the length of the trachea greater length. CT findings in patients with central airway
may be successfully resected in the absence of mediastinal TB are rarely specific. As noted by Moon et al. (183), find-
invasion (48). ings that help to differentiate tuberculous disease from
bronchogenic carcinoma include longer circumferential
involvement and the absence of an intraluminal mass.
Tuberculosis of the Trachea
Notwithstanding these distinctions, accurate diagnosis
and Mainstem Bronchi usually requires histologic evaluation.
TB typically involves the distal trachea and proximal main-
stem bronchi; isolated tracheal disease is rare (178–184).
Diffuse Tracheal Disease
Active inflammation causes irregular circumferential
bronchial wall thickening, resulting in narrowed or even In addition to tracheal tumors, many different inflamma-
obstructed airways. Enhancement of the tracheal wall has tory abnormalities, both infectious and noninfectious, dif-
been reported in patients with acute inflammation after fusely affect the trachea and mainstem bronchi (Table 5-3)
the administration of intravenous contrast material (183). (185). These include both primary tracheal abnormalities
In most cases, active infection results from extension from and those resulting from spread of adjacent mediastinal
peribronchial lymphatic disease, although endobronchial inflammation, and they may manifest as either diffuse
implantation from contaminated sputum has also narrowing or dilatation.
been postulated as a potential mechanism (Fig. 5-41)
(181,183). Evidence of associated mediastinal inflamma- Diffuse Tracheal Widening/
tion, in the form of diffuse increased density of mediasti- Tracheobronchomegaly
nal fat or enlarged mediastinal or pericarinal nodes, may Also referred to as Mounier-Kuhn syndrome, the term
also be identified (178,181–183). If this condition is tracheobronchomegaly refers to a heterogeneous group of
untreated, ulceration with resulting fistulas between nodes patients who have marked dilatation of the trachea and
and adjacent airways may develop. With appropriate ther- mainstem bronchi, frequently in association with tracheal
apy, these changes are usually, although not invariably, diverticulosis, recurrent lower respiratory tract infections,
reversible. Unlike active infection, fibrotic TB results in and bronchiectasis (186–188). The etiology of this disorder
A B
Figure 5-41 Tuberculous tracheitis. See Color Figure 5-41C. A, B:

Contrast-enhanced axial images through the mid trachea, imaged with wide
and narrow windows, respectively, show extensive low-density, rim-enhancing
right paratracheal nodes causing eccentric narrowing of the posterolateral
tracheal wall. Note apparent small nodule in the anterior tracheal wall in A
(arrow). C: Bronchoscopic image corresponding to the CT section in A shows
extrinsic compression of the posterolateral tracheal wall on the right side (as-
terisk), associated with diffuse, irregular nodularity of the tracheal mucosa.
Suspected polypoid lesion seen on CT is seen to better effect endoscopically.
Endoscopically proved tuberculosis. (From Naidich DP, Webb WR, Grenier PA,
et al. Imaging of the airways: functional and radiologic correlations.
C Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
is controversial. Findings in favor of a congenital etiology CT findings in patients with tracheobronchomegaly

include histopathologic evidence of deficiency of tracheo- have been extensively described (188,191–193). Using a
bronchial muscle fibers and absence of the myenteric tracheal diameter of greater than 3 cm measured 2 cm
plexus as well as an association with other congenital or above the aortic arch, and diameters of 2.4 and 2.3 cm
connective tissue disorders, including ankylosing spondyli- for the right and left main bronchi, respectively, the
tis, Marfan syndrome, cystic fibrosis (CF), Ehlers-Danlos diagnosis of tracheobronchomegaly is relatively straight-
syndrome, and cutis laxa in children (189,190). In distinc- forward (191). Additional findings include tracheal scal-
tion, findings in favor of an acquired etiology include the loping or diverticula, the latter especially common along
fact that tracheobronchomegaly most often is diagnosed in the posterior tracheal wall. Also common is the finding
men in their third and fourth decades without an of marked tracheal flaccidity, identifiable as a marked
antecedent history of respiratory tract infection, often in decrease in the diameter of the trachea on expiration,
association with chronic cigarette smoking (189,190). An even to the point of airway occlusion, indicative of
association between tracheomegaly and diffuse pulmonary tracheomalacia (27). Optimal evaluation of this disorder
fibrosis has also been reported, presumably the result of requires volumetric data acquisition as a minimum
increased traction on the tracheal wall due to increased through the length of the trachea, ideally through the
elastic recoil pressure in both lungs exerting opposing force entire thorax (193a). In addition to allowing the recon-
(189,190). struction of high-quality MPRs or 3D volumetric images,
volumetric data acquisition also provides invaluable Diffuse Tracheal Narrowing

information regarding peripheral air trapping, a frequent Included in the group of patients with diffuse narrowing
ancillary finding. Although precise quantitative estima- are those with infectious disorders, including bacterial,
tion of the severity of tracheomalacia has also been re- viral, and fungal infections either primary or secondarily
ported using electron beam CT (EBCT), this technique re- resulting from adjacent mediastinitis, as well as patients
mains generally less available (193b). with a variety of noninfectious disorders, including tra-
The importance of tracheomegaly lies in its association cheopathia osteochondrolytica, relapsing polychondritis,
with distal airway inflammation. In a study of 75 con- Wegener granulomatosis (WG), amyloidosis, sarcoidosis,
secutive patients referred for CT evaluation of possible and ulcerative colitis (185,186,196). Although many of
bronchiectasis, Roditi and Weir (189) found that, overall, these have similar appearances in cross section, some are
12% of their patients proved to have dilated tracheas, sufficiently distinctive to allow a definitive diagnosis with
including 7 (17%) of 42 patients with CT evidence of appropriate clinical correlation (197).
bronchiectasis as well as 3 (6%) of 32 without. These data
suggest that tracheomegaly may play a causative role in Saber-Sheath Trachea. Saber-sheath trachea is a com-
the development of bronchiectasis as a result of a predis- mon variation in tracheal morphology (Fig. 5-42) in
position to infection from abnormal mucus clearance in which the internal side-to-side diameter of the trachea is
patients with inefficient cough and stagnant mucus. decreased to half or less the corresponding sagittal diame-
Tracheomalacia may also be seen in association with a ter. Although this condition is classically described as a
number of disorders in addition to tracheobronchomeg- static deformity, further narrowing of the tracheal lumen
aly, including most causes of diffuse tracheal inflamma- can be documented when patients are examined using CT
tion such as relapsing polychondritis, as well as following obtained during forced expiration or a Valsalva maneuver
trauma or especially in patients with chronic bronchitis, (196,198). Although it has long been appreciated that
among others. Symptoms are typically nonspecific and saber-sheath tracheas are always associated with chronic
include chronic cough, dyspnea, and occasionally hemop- obstructive pulmonary disease (199), only recently has it
tysis (194). Tracheomalacia has also been associated with been shown that there is a significant relationship between
spontaneous tracheal rupture (144). In cases for which tracheal dimensions and the severity of emphysema (200).
this diagnosis is suspected, additional expiratory scans In a study evaluating tracheal morphology in patients
may be essential. Aquino et al. (195) showed that the before and after lung volume reduction surgery, the length
probability of tracheomalacia varied between 89% and of the trachea decreased while the width increased in size.
100% when the change in the sagittal diameter of the Unfortunately, these changes proved to be poor predictors
trachea exceeded 29%. of postoperative lung function (200).
Figure 5-42 Saber-sheath trachea.

See Color Figure 5-42B. A: Axial sec-
tion at the level of the aortic arch
shows that the sagittal length of the
trachea is more than twice the width
of the trachea—a finding characteris-
tic of a saber-sheath trachea. B:
Bronchoscopic image in the same pa-
tient shows similar appearance of
marked narrowing of the width of the
trachea. Note that, although the
lumen of the trachea appears smooth
on CT, accurate evaluation of the tra-
cheal mucosa requires endoscopy.
(From Naidich DP, Webb WR, Grenier
PA, et al. Imaging of the airways: func-
tional and radiologic correlations.
Philadelphia: Lippincott Williams &
A, B Wilkins; 2005, with permission.)
Infectious Tracheobronchitis. A number of infections, maxillary sinuses, this infection may also involve the tra-
both acute and more often chronic, may affect the trachea chea and proximal bronchi in up to 10% of cases (204).
and proximal bronchi, resulting in both focal and diffuse Although RS is endemic in tropical and subtropical areas,
airway disease. Of particular interest is the rare presenta- with more than 80% of cases reported from Central and
tion of acute bacterial tracheitis and rhinoscleroma. South America, Africa, India, Indonesia, and Eastern and
Acute tracheitis is almost always seen in children, only Central Europe (205), the prevalence of this disease is
rarely presenting in adults, usually in association with rising in the United States because of increasing ease of
immunodeficiency syndromes or AIDS (201). It is charac- international travel. The diagnosis is generally established
terized by acute upper airway obstruction resulting in either by biopsy or positive cultures, present in approxi-
stridor. Endoscopically there is evidence of purulent mately 50% of cases.
debris, mucosal ulcerations, and edema (Fig. 5-21). On CT there is evidence of nonspecific diffuse nodular
Rhinoscleroma (RS) is a slowly progressive granuloma- thickening of the trachea and proximal bronchi causing
tous infection caused by the bacterium Klebsiella rhinoscle- airway narrowing (Fig. 5-43). There may also be evidence
romatis, a capsulated gram-negative bacterium (202–204). of mediastinal or hilar adenopathy. With airway ob-
Characterized by involvement of the upper respiratory struction, patchy parenchymal consolidation may also
tract, including the nose, upper lip, hard palate, and be identified. Although antibiotics generally result in
A B
C D
Figure 5-43 Rhinoscleroma. See Color Figure 5-43D. A–C: Axial CT sections at the level of the thoracic inlet, distal trachea, and proximal
lower lobe bronchi, respectively, in a 26-year-old immigrant complaining of persistent cough and chest pain. These images show a combina-
tion of nodular thickening of the tracheal and bronchial walls, resulting in marked luminal narrowing associated with mediastinal and hilar
adenopathy. D: Endoscopic biopsy of the tracheal wall shows squamous mucosa with acute inflammation, characterized by dense infiltration
with plasma cells, lymphocytes, and histiocytes. Subsequent evaluation revealed intracellular bacilli consistent with rhinoscleroma. (From
Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations. Philadelphia: Lippincott Williams &
Wilkins; 2005, with permission.)
immediate improvement, RS is difficult to eradicate, with Tracheobronchopathia Osteochondroplastica

the infection frequently relapsing. Tracheobronchopathia osteochondroplastica is a rare dis-
order involving the trachea and main bronchi in which
Fungal Tracheobronchitis/Aspergillosis. Airway disease multiple submucosal nodules representing bone, cartilage,
caused by aspergillosis takes many forms, including sapro- or calcified acellular protein matrix may be identified,
phytic colonization, tracheobronchitis, ulcerative tracheo- resulting in diffuse narrowing of the airway lumen
bronchitis with or without pseudomembrane formation, (Fig. 5-44) (211–213). Characteristically, these nodules
and necrotizing (invasive) aspergillosis (206), the latter spare the membranous posterior wall of the trachea. Often
further subdivided into acute and chronic airway invasive asymptomatic, tracheopathia osteochondroplastica has
forms (207). Acute tracheobronchitis typically occurs only been reported to cause hemoptysis in as many as 25% of
in severely immunocompromised patients, including those cases (214). Other symptoms, including cough, dyspnea,
with underlying malignancies or AIDS or following bone hoarseness, stridor, and recurrent lower respiratory tract in-
marrow, lung, or heart transplants (208). Histologically, fections, have also been noted. Typically occurring in mid-
there is evidence of respiratory epithelial ulceration and dle-aged men, the course of this disease is generally be-
submucosal inflammation typically restricted to the central nign, and in most cases the diagnosis is made only at
airways. CT examination reveals nonspecific multifocal or autopsy.
diffuse tracheobronchial wall thickening, causing either The origin of this disease is unknown. Thought by
smooth and/or nodular narrowing of the airway lumen Virchow to represent enchondroses and exostoses arising
(206,207,209). Unfortunately, antifungal therapy generally from the perichondrium of cartilage rings, calcified nod-
proves unsuccessful, especially in severely immunocompro- ules alternatively have been attributed to cartilaginous and
mised patients. Acute obstruction of the right mainstem bony metaplasia involving elastic tissue (215). Although
bronchus with subsequent bronchial wall necrosis and investigators have postulated that tracheopathia osteo-
rupture of the adjacent pulmonary artery due to necrotizing chondroplastica represents an end stage of tracheal amy-
tracheobronchitis in an immunocompromised patient has loid, to date the presence of amyloid has been reported in
been reported as a cause of mortality (210). only a handful of cases.
A B
Figure 5-44 CT-bronchoscopic correlations: tracheobronchopathia osteochondroplastica (TO). See Color Figure 5-44B. A: Magnified view
of the mid trachea imaged with narrow windows shows characteristic appearance of calcified nodules resulting in mild narrowing of the
tracheal lumen. Note the absence of involvement of the posterior membranous tracheal wall. This proved to be an incidental finding.
B: Corresponding flexible bronchoscopic evaluation demonstrates precise CT-endoscopic correlation with characteristic nodular excrescences
protruding into airway, pathognomonic of this disease. Spirometry and flow-volume loop revealed minimal obstructive dysfunction. (From
Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations. Philadelphia: Lippincott Williams &
Wilkins; 2005, with permission.)
Radiologically, the diagnosis is difficult unless the patient been reported (218). Histologically, one sees evidence
has marked nodularity and narrowing of the trachea. of chondritis with perichondral inflammation, loss of
Bronchoscopically, these lesions appear as multiple yellow- basophilic staining of cartilage matrix, and ultimately
white, hard papillary nodules that may be difficult to fibrous replacement of involved cartilage.
sample for biopsy because of their bony nature. In distinc- Characterized by episodic inflammation, relapsing
tion, numerous reports have documented a characteristic polychondritis results in diffuse tracheal and bronchial
CT appearance of calcified nodules measuring between narrowing in approximately 50% of cases, secondary to
3 and 8 mm in diameter, resulting in irregular narrowing a combination of edema, granulation tissue, cartilage
of the lower trachea and mainstem bronchi with sparing of destruction, and ultimately fibrosis of the tracheal wall
the posterior tracheal wall (173,177). (219). When severe, this disease may result in airway
obstruction and recurrent episodes of atelectasis and pneu-
Relapsing Polychondritis monia (220). CT findings in this disease include tracheal
Relapsing polychondritis is a rare disorder that results in narrowing resulting from thickening of the anterior and
destruction of cartilage within the pinna of the ear, nose, lateral tracheal walls with characteristic sparing of the
joints, and upper airways, including the larynx and posterior tracheal wall (Fig. 5-45) (173,217,218,221).
subglottic trachea (216–219). This condition is thought Because this disorder is often indistinguishable from other
to be related to abnormal acid mucopolysaccharide causes of diffuse tracheal narrowing, the diagnosis rests on
metabolism and an association has also been noted in a combination of pathologic, clinical, and radiologic find-
approximately 25% of cases of autoimmune vasculitis ings, especially the presence of abnormal calcifications in
(217). Although nonerosive polyarteritis, nasal defor- the pinna of the ear (218). As documented by Im et al.
mity, and auricular chondritis are characteristic, they (217), CT may be of particular value in documenting a
need not be present in all cases. Arteritis resulting in aor- response to steroid therapy by demonstrating decreased
titis and aortic insufficiency, mitral insufficiency, and tracheal wall thickening, potentially obviating more inva-
aneurysms involving medium-sized arteries have also sive follow-up procedures.
A B
Figure 5-45 Relapsing polychondritis. A–C: Axial images

through the upper, mid, and lower trachea/carina, respectively,
show typical appearance of diffuse, uniform tracheal and mainstem
bronchial wall thickening with sparing of the posterior membra-
nous wall of the trachea. Note the absence of mediastinal adeno-
pathy. (A from Naidich DP, Webb WR, Grenier PA, et al. Imaging of
the airways: functional and radiologic correlations. Philadelphia:
Lippincott Williams & Wilkins; 2005, with permission; B and C from
Webb EM, Elicker BM, Webb WR. Using CT to diagnose nonneo-
plastic tracheal abnormalities: appearance of the tracheal wall. AJR
C Am J Roentgenol. 2000;174:1315–1321, with permission.)
Wegener Granulomatosis Lesions may be either focal or diffuse and may involve
WG is a systemic necrotizing vasculitis typically involving any portion of the airways from the hypopharynx to the
middle-aged adults (222). Although characteristically lobar bronchi (Figs. 5-46 and 5-47) (225–228). Involve-
described by the classic triad of upper respiratory tract, ment of segmental bronchi may result in atelectasis or
lung, and renal involvement with glomerulonephritis, the infection (224). Typically, the disease causes diffuse wall
disease also often involves skin and joints as well as the thickening resulting in circumferential narrowing of the
middle ear, eye, and nervous system. In 90% of cases of airway lumen. Although it is uncommon, mediastinal or
generalized WG, serum antineutrophil cytoplasmic anti- hilar adenopathy may also be identified.
bodies characterized by a diffuse granular cytoplasmic CT findings have been extensively reported
immunofluorescent staining pattern (C-ANCA) may be (223–226,227). In one study of 30 patients with docu-
detected. These are thought to be highly specific for the mented WG, 29 of 30 (97%) (229,230) had abnormal CT
disease, although false-positive test results have been studies at the time of presentation (226). Nodules and
noted in patients with a variety of both inflammatory and masses were most commonly observed, occurring in 90%
neoplastic diseases (222). of patients (Figs. 5-46 and 5-47). Most often bilateral,
Within the lungs, WG usually manifests either as multi- these characteristically proved to be either subpleural
ple parenchymal nodules or as areas of focal consolida- (simulating pulmonary infarction) or peribronchovascular
tion, frequently with evidence of cavitation. Involvement in distribution (226). Interestingly, feeding vessels were a
of the central and peripheral airways is an important man- frequent observation, consistent with underlying vasculi-
ifestation of this disease. Cordier et al. (223), in a study of tis. Central cavitation, both irregular and thick walled, and
77 patients with documented WG, found that 55% in smooth and thin walled, was also common, typically
whom FB was performed proved to have airway involve- occurring only in lesions greater than 2 cm in size. In this
ment, including inflammatory lesions with and without study, patchy foci of consolidation or ground-glass attenu-
bronchial stenosis, ulcerations and pseudotumors, and ation were less frequently noted, occurring in approxi-
isolated hemorrhagic foci. Similarly, Daum et al. (224), in mately 25% of cases (226).
their study of 51 patients, all evaluated bronchoscopically, Whereas airway narrowing is an infrequent presenting
found endobronchial abnormalities in 59%, including manifestation of disease, airway lesions often occur in
subglottic stenosis in 17%, ulcerating tracheobronchitis in the course of therapy, possibly because of prolonged
60%, tracheal or bronchial stenoses in 13%, and hemor- survival with use of cytotoxic agents. In one series of
rhage without an identifiable source in another 4% of 19 patients, airway lesions occurred in 5 patients in the
patients. In this same study, there was no correlation noted course of eight relapses and always proved symptomatic,
between C-ANCA levels and the presence of tracheo- manifesting as hoarseness, cough, or stridor (229). In
bronchial inflammation. Furthermore, a specific histologic selected cases, CT may prove invaluable by demonstrating
diagnosis of WG could only be made bronchoscopically in optimal sites for tracheostomy as well as by defining the
approximately 17% of cases. true extent of disease when bronchial narrowing precludes
A B
Figure 5-46 Wegener granulomatosis. A: Axial image through the mid trachea showing irregular nodular thickening of the tracheal wall
associated with increased soft tissue density in the right paratracheal space. B: Axial section shows marked narrowing of the origin of the
middle lobe bronchus associated with hilar adenopathy. In this case, the lung parenchyma appeared normal. (From Naidich DP, Webb
WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with
permission.)
A B
C D
Figure 5-47 Wegener granulomatosis. A, B: Axial section through the upper trachea imaged with wide and narrow windows, respec-
tively, shows a tiny, well-defined apparently calcified nodule in the posterolateral wall of the trachea. C: Coronal reconstruction at the level
of the proximal right mainstem bronchus shows an additional well-defined nodule. D: Virtual bronchoscopic image shows diffuse nodularity
throughout the trachea, with the largest nodule corresponding to the nodule identified in C. In this case, the true extent of disease is best
seen on the virtual endoscopic images. Findings were subsequently confirmed bronchoscopically. (From Grenier PA, et al. New frontiers in
CT imaging of the airways. Eur Radiol. 2002;12:1022–1044, with permission.)
complete bronchoscopic evaluation. The use of VB has Although amyloidosis is divided into two broad cate-
also been recommended to evaluate the trachea prior to gories, primary or secondary, as defined by the World
bronchoscopy (231). Health Organization, precise classification is dependent
on recognition of specific fibrillar proteins (232). Primary
Amyloidosis amyloidosis is defined by the presence of a fragment of the
Amyloidosis entails a number of entities characterized by variable immunoglobulin (Ig) light chain (AL), and thus
the deposition of an autologous fibrillar protein and asso- varies from person to person. In distinction, secondary
ciated protein derivatives in extracellular tissues. Together amyloidosis involves the amino acid terminus of the acute
these elements accumulate either in single organs or phase protein serum amyloid A (SAA) and is identical in
diffusely throughout the body, resulting in a wide range of all patients. SAA is subsequently cleaved by macrophages
clinical symptoms. Histochemically, amyloid binds with to form amyloid A (AA) (233). Secondary amyloidosis
Congo red, showing green birefringence in polarized light. usually develops as a response to a number of different
chronic inflammatory diseases, including CF, TB, and than tracheobronchial amyloidosis and is usually charac-
bronchiectasis, among others. terized by multiple peripheral or subpleural nodules that
Primary systemic amyloidosis results from clonal vary from a few millimeters to several centimeters in size.
expansion of plasma cells within the bone marrow and Calcifications are identified in up to 50% of lesions.
results in multiorgan accumulation of amyloid. It is this Patients are rarely symptomatic (237). Diffuse parenchy-
form of the disease that is frequently associated with either mal amyloidosis, although typically identified in patients
multiple myeloma or Waldenström macroglobulinemia with primary systemic amyloidosis, is the least common
and tends to have an inexorable downhill course. In the form of localized amyloidosis. It is characterized by diffuse
lungs, primary systemic amyloidosis results in diffuse reticulonodular opacities with prominent interlobular and
interstitial infiltration. intralobular septae (237).
In distinction, localized amyloidosis typically involves a The CT appearance of tracheobronchial amyloid has
single organ only. Although this form of the disease is also been frequently reported (Fig. 5-48) (233,236–240). In its
characterized by deposition of monoclonal Ig light chains, nodular form, amyloid appears as a focal mass partially or
there is no evidence of marrow plasmacytosis, nor is there completely occluding airways with infiltration of the
evidence of proteinuria as is usually seen in patients with adjacent peritracheal or bronchial tissues, often with foci
primary systemic amyloidosis. Within the lung, three types of calcification. In this form, it may mimic the appearance
of localized amyloidosis have been identified: tracheo- of both benign and malignant tracheal neoplasms. In
bronchial, nodular, and diffuse alveolar septal amyloidosis its more diffuse form, tracheal amyloid leads to concen-
(234,235). tric nodular thickening of the tracheal wall due to submu-
Although tracheobronchial amyloidosis is the most cosal deposition of nodules or plaques, often resulting
common form of respiratory system involvement in in luminal narrowing (Fig. 5-48). Mural calcifications are
patients with localized amyloidosis, to date, only several prominent features, although calcifications are rarely
hundred cases have been reported (236). The disease is identified bronchoscopically. In patients with recurrent
characterized by the presence either of multifocal or infiltrates, bronchiectasis may also be identified. Differ-
diffuse submucosal plaques or masses. Endobronchial ential diagnosis includes relapsing polychondritis, tracheo-
biopsies are diagnostic. Clinical manifestations reflect the bronchopathia osteochondroplastica, and granulomatous
portion of the tracheobronchial tree affected (235): tracheitis.
patients with proximal subglottic or laryngeal involve- In patients with severe narrowing, therapy usually
ment typically present with hoarseness or stridor. In involves repeated attempts at debulking lesions either using
distinction, involvement of the distal trachea or mainstem, forceps or Nd:YAG laser resection. Despite initial response,
lobar, or proximal segmental bronchi results in cough, disease progression often necessitates repeated resections
wheezing, dyspnea, or hemoptysis. In cases in which (236). In patients with subglottic involvement, tra-
bronchial obstruction occurs, patients may present with cheostomies may be required. In these cases, unfortunately,
fever secondary to obstructive pneumonitis (237). an association has been noted between proximal death and
Nodular parenchymal amyloidosis is less commonly seen early death due to respiratory insufficiency (235).
A B
Figure 5-48 Amyloidosis. A, B: Axial images through the carina and subcarinal space, respectively, show extensive nodular calcifications
throughout the bronchial tree associated with mild bronchial wall thickening. This appearance should not be confused with the cartilaginous
calcifications typically seen in elderly individuals. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radi-
ologic correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
Inflammatory Bowel Disease/Ulcerative Colitis LOBAR ATELECTASIS

Although this manifestation is rare, changes involving
both large and small airways have been reported to occur Evaluation of patients with radiographic evidence of lobar
in patients with ulcerative colitis. Within small airways, or segmental atelectasis represents one of the most impor-
findings consistent with constrictive bronchiolitis have tant indications for the use of CT. The radiographic
been described reminiscent of sclerosing cholangitis, patterns of lobar atelectasis have been extensively reviewed
although the two diseases rarely coexist (241). Similar (88). A wide range of abnormalities has been described
findings of submucosal fibrosis and airway narrowing characterizing the appearance of the atelectatic lobe as well
have also been described within the larger airways, includ- as secondary, compensatory changes, including shifts of
ing the trachea, often in association with bronchiectasis. mediastinal structures, the hila, the hemidiaphragms, and
To date, no apparent relationship has been established the fissures.
between disease activity within the colon and airway Lobar atelectasis results from certain variably categorized
abnormalities: in fact, airway disease may occur even years mechanisms (242). These include the following: (a) resorp-
after colectomy. On CT, the tracheobronchial walls appear tion or obstructive atelectasis, resulting from endobronchial
thickened, resulting in irregular narrowing indistinguish- obstruction (Figs. 5-6, 5-49 to 5-53); (b) cicatrization
able from that due to the other causes of diffuse airway atelectasis, resulting from parenchymal fibrosis (Figs. 5-54
narrowing described earlier (173,185). and 5-55); (c) compression or passive atelectasis, which is
A B
C D
Figure 5-49 Right upper lobe atelectasis—central airway obstruction. A–C: Sequential images through the middle and lower trachea
and right upper lobe bronchus, respectively. The collapsed right upper lobe can be identified as a wedge extending along the medi-
astinum to the anterior chest wall. Hyperinflation of the middle and lower lobes can be clearly separated by identification of the lateral
margin of the oblique fissure (straight arrows in A and B). The middle lobe is insinuated between the collapsed upper lobe medially and
the lateral chest wall (curved arrows in A and B). Note the abrupt obstruction of the distal portion of the right upper lobe bronchus (arrow
in C). D–F: Identical images as in A to C, imaged with narrow windows. After intravenous contrast administration, one can differentiate the
central tumor mass (curved arrows in D; asterisk in F) from the peripherally collapsed right upper lobe (curved arrows in E and F) because
of subtle differences in attenuation between tumor tissue and atelectatic lung. Note the presence of markedly enlarged mediastinal nodes
(open arrow in E). G, H: Coronal reconstructions through the carina and more posteriorly through the descending thoracic aorta convey to
better advantage the three dimensional appearance of right upper lobe collapse, in particular the location of the major and minor fissures
(continued ).
E F
G H
A B
Figure 5-50 Right upper lobe atelectasis: value of intravenous contrast administration. A: Section at the level of the right upper lobe
bronchus after intravenous contrast administration shows evidence of tumor replacing most of the right upper lobe (arrows) in addition to
obstructing the right upper lobe bronchus; only a small volume of atelectatic lung is still identifiable (asterisk). A small right pleural effusion
is also present. B: Section through the right upper lobe in a different patient from that in A. In this case, tumor is located centrally (arrows)
and is surrounded by lung within which normal branching vessels can still be identified. Posterior bulging of the major fissure is noted. These
findings are consistent with so-called “drowned lung” secondary to central bronchial obstruction (compare with A and Fig. 5-49). These
cases illustrate the importance of administering intravenous contrast media to identify the presence and extent of tumor accurately.
A B
C D
Figure 5-51 Combined right upper and middle lobe atelectasis. A, B: Posteroanterior and lateral radiographs, respectively, show an ill-
defined hazy density obscuring the right hilum on the posteroanterior radiograph with apparent displacement of the entire length of the
major fissure (arrows in B), a pattern more consistent with left rather than right upper lobe collapse. C, D: Contrast-enhanced CT sections
through the right upper and middle lobes, respectively, show that both these airways are obstructed by tumor (arrows), with resultant com-
bined upper and middle lobe atelectasis. In this case, despite the use of contrast, differentiation between tumor and atelectatic lung is diffi-
cult. At endoscopy, obstructing endobronchial lesions with identical histologic features were identified in both the right upper and middle
lobe bronchi. Combined right upper and middle lobe collapse accounts for the apparent left upper lobe pattern of collapse seen on the cor-
responding posteroanterior and lateral chest radiographs shown in A and B.
collapse caused by extrinsic compression, such as from of lobar atelectasis, as occurs, for example, in patients with
pleural fluid or air, or the presence of any space-occupying combined middle and lower lobe atelectasis or even right
intrathoracic lesion resulting in extrinsic compression of upper and lower lobe atelectasis (246). In the absence of
adjacent parenchyma (Fig. 5-56); and (d) adhesive atelec- bronchial obstruction, characteristic alterations in the
tasis, which is collapse resulting from loss of surfactant normal appearance of the airways consequent to volume
(Fig. 5-57). loss with resultant changes in the caliber and position of
CT has proved to be an essential adjunct to routine radi- airways may also be identified far more accurately with CT
ography in the evaluation of atelectasis (87,243,244). This than on corresponding chest radiographs.
is primarily because CT can accurately identify and localize Differentiation between the various causes of lobar
the presence of an obstructing lesion (79,82,245). CT is atelectasis requires not only precise visualization of the
especially valuable in patients with unusual combinations airways but also detailed examination of the atelectatic
A B
Figure 5-52 Left upper lobe atelectasis: value of intravenous contrast administration. A: Contrast-enhanced section shows obstruction of
the left upper lobe bronchus. A large mass can be identified by the subtle difference in attenuation between the mass and peripheral
atelectatic lung (arrows). This tumor extends into the left hilum and displaces the left interlobar pulmonary artery (curved arrow), as well as
medially, causing partial obstruction of the left superior pulmonary vein. B: Contrast-enhanced section through the lower trachea in a differ-
ent patient from the one in A shows a large tumor mass occupying most of the left upper lobe at this level (arrows), within which subtle
dystrophic calcification can be identified. This patient had bronchoscopically confirmed metastatic mucinous carcinoma of the colon.
A B
C D
Figure 5-53 Left lower lobe atelectasis. A–D: Select contrast-enhanced images in a patient with left lower lobe atelectasis. The marked
narrowing of the left mainstem (arrow in A) and left upper lobe bronchi (arrow in B) is consistent with submucosal and peribronchial tumor
infiltration; the left lower lobe bronchus is obstructed at its origin. Tumor also infiltrates the left hilum and narrows the left interlobar pul-
monary artery (curved arrows in A and B). A large, necrotic tumor mass is identifiable distinct from atelectatic lung (arrows in C). Fluid-filled
bronchi are present in the lower lobe (arrows in D). There is marked posterior and medial displacement of the major fissure (open arrows in
A–C). A small quantity of pleural fluid is also present. This patient had bronchoscopically confirmed non–small cell lung cancer.
A B
Figure 5-54 Cicatrization atelectasis: middle lobe (ML). A–C:

Sequential 5-mm sections show typical appearance of cicatrization
atelectasis of the ML with a patent middle lobe bronchus (arrow in
A) leading to a markedly shrunken middle lobe, within which di-
lated, bronchiectatic airways are apparent (arrow in B). The major
fissure is displaced anteromedially (curved arrow in C); in distinc-
tion, the minor fissure is more difficult to identify owing to partial
C volume averaging.
A B
Figure 5-55 Cicatrization atelectasis: right lower lobe (RLL). A, B: Selected 5-mm sections imaged with wide windows show a shrunken
right lower lobe outlined laterally by the markedly displaced major fissure (arrows in A and B). Note the patency of the right lower lobe
bronchus (curved arrow in A) and posterior displacement of the middle lobe bronchus, associated with extensive bronchiectasis. In this case,
one sees evidence of a marked difference in lung density between the right upper lobe (asterisk) and lingula versus the right middle lobe
and left lower lobe, likely the result of associated small airway disease in the upper lobes. C: Identical section as B, imaged with narrow
windows, shows that bronchiectatic airways fill the shrunken right lower lobe (arrow) (continued).
C Figure 5-57 Adhesive atelectasis. A 5-mm section shows typical

pattern of paramediastinal ground-glass attenuation and consoli-
Figure 5-55 (continued) dation approximately 6 months after radiation therapy for
Hodgkin disease. The mild thickening of the secondary lobular
septa anteriorly is compatible with mild lymphatic obstruction
lung itself, as well as evaluation of the hilum and medi- (arrows).
astinum. In most cases, this requires use of a bolus of
intravenous contrast. In addition to allowing differentia-
tion between central tumor and peripherally collapsed delineation of accompanying pleural fluid, especially
lung, contrast administration facilitates more precise when loculated.
staging of central tumors by allowing more accurate
assessment of nodal disease as well as direct mediastinal
invasion. Additionally, administration of contrast often Obstructive (Resorption) Atelectasis
provides detailed evaluation of parenchymal changes
within the atelectatic lung itself, including the presence of Mechanisms and Causes
generalized or focal infection, as well as more precise
Most commonly the result of primary bronchogenic carci-
noma, otherwise indistinguishable endobronchial masses
have been reported to represent a wide range of benign
and malignant lesions (Table 5-4).
Endobronchial Tumors
In addition to metastases, most commonly of breast,
colon, kidney, and thyroid carcinoma and of melanoma,
but also including those of uterine, prostatic, testicular,
and adrenal malignancies (247), other neoplastic lesions
to be considered include carcinoid tumors (170) and lym-
phoma (248,249). Malignant lesions are generally more
likely to be associated with extraluminal soft tissue masses,
usually identifiable after administration of intravenous
contrast medium. Benign tumors involving the tracheo-
bronchial tree are less common than malignant lesions
and, with the exception of amyloidomas (which are of
unknown origin), originate from a large number of diverse
Figure 5-56 Compression atelectasis. Contrast-enhanced CT
section following a bolus of intravenous contrast shows typical cell lines, including papillomas (epithelial gland origin),
appearance of compression atelectasis of the right lower lobe due myoblastomas (neural cell origin), and, most commonly,
to a large right pleural effusion. In this case there is also evidence hamartomas and chondromas (connective tissue origin)
of a moderate-sized pericardial effusion also causing compression
of the adjacent lung. Note that there is uniform enhancement of (250). Rarely, specific features such as the presence of
the periphery of the left lower lobe and that the lobe is markedly calcium, as occurs in patients with central carcinoid
shrunken. These two signs are consistent with a lack of significant tumors, or fat with or without calcification (Fig. 5-25), as
pathology within the lobe, with its normal vascular supply intact
and a lack of infection, infarction, or tumor. Proximal patency of may be seen in patients with endobronchial hamartomas,
the left lower lobe airways is also noted. allow a specific diagnosis.
cases. Atelectasis secondary to bronchial strictures also

may mimic the appearance of discrete central endo-
TABLE 5-4
bronchial lesions, especially as a result of TB (181,183). In
RESORPTION ATELECTASIS: CAUSES
these cases, the use of MPRs and of 3D reconstructions
Endobronchial neoplasms may be of value by demonstrating the true extent and con-
Common figuration of the lesions (33,253). Often, these lesions are
Bronchogenic carcinoma associated with extraluminal nodal calcifications, thereby
Bronchial carcinoid/adenoid cystic carcinoma
Metastases (kidney, breast, colon, melanoma)
simplifying the differential diagnosis.
Lymphoma The accuracy of CT for detecting endobronchial
Kaposi sarcoma obstructing lesions nears 100%. In a retrospective analysis
Rare of 50 patients with segmental or lobar atelectasis,
Benign neoplasms Woodring (254) compared the accuracy of chest radi-
Hamartoma, papilloma, granular cell myoblastoma
ographs with CT in identifying patients with central
Non-neoplastic endobronchial lesions obstructing tumors. Using the findings of bronchial
Common
obstruction and a central mass causing a contour abnor-
Mucus plugs
Aspirated foreign bodies mality, unlike chest radiographs, which were diagnostic in
Inflammatory strictures 24 (89%) of 27 patients, CT proved 100% sensitive, cor-
Tuberculous/fungal infection rectly identifying all 27 obstructing carcinomas. In 3 cases
Broncholithiasis (10%), CT findings led to false-positive diagnoses: in each,
Rare
benign causes of bronchial obstruction led to the false
Bronchial laceration
Bronchial torsion assumption of central tumor. Importantly, in no case in
Wegener granulomatosis which the proximal airways were shown to be patent was
Amyloidosis tumor ultimately found. The finding of peripheral air
Sarcoidosis bronchograms in itself does not exclude central obstruc-
Extrinsic bronchial obstruction tion, especially in patients with more peripheral, segmen-
Common tal occlusion. CT air bronchograms may be seen within
Enlarged hilar nodes (tuberculosis, sarcoidosis) otherwise obstructed segments or lobes because central
Aortic aneurysm
obstruction is of recent onset, because it is incomplete, or
Adapted from Woodring JH, Reed JC. Types and mechanisms of pul- because the effects are mitigated by collateral air drift. As
monary atelectasis. J Thorac Imaging. 1996;11:92–108 with permission. documented by Woodring (254), in 32 cases in which
peripheral air bronchograms could be identified with CT, a
central obstructing tumor was present in 11 (34%).
As noted previously, most malignant endobronchial
Nontumoral Obstruction lesions are associated with extraluminal soft tissue masses
Endobronchial obstruction may also result from certain (Figs. 5-6 and 5-49 to 5-53). In many cases these lesions
non-neoplastic causes, including the following: retained may be outlined as distinct from postobstructive pneu-
secretions; blood clots; inflammatory strictures, in particu- monitis and atelectasis by administration of a bolus of
lar those secondary to granulomatous infections, with or intravenous contrast medium. Differentiation is based on
without associated broncholithiasis; aspirated foreign numerous factors, most importantly including tumor vas-
bodies; and bronchial trauma. cularity, the presence and extent of tumor necrosis,
Obstruction commonly results from retained secre- evidence of pulmonary artery obstruction, and the state of
tions. Although this disorder may cause confusion with the atelectatic lung, including the presence or absence of
central endobronchial masses, in general, inspissated infection. As documented by Burke and Fraser (255). in
secretions may be identified either because of demonstra- their study of 50 consecutive patients undergoing pul-
ble fluid density (including the presence of an air-fluid monary resection for bronchogenic carcinomas, a wide
level) or because of the finding of a clear separation spectrum of histologic changes can be identified within
between irregular intraluminal densities and the adjacent atelectatic lung when atelectasis occurs over a prolonged
bronchial wall (especially when viewed with narrow, time interval. Initially, histologic changes reflect non-
mediastinal windows) (Fig. 5-58). This appearance should infectious processes, including mucoid impaction and
be differentiated, of course, from the finding of fluid- dilatation of the airways, frequently associated with
filled, distended bronchi within the peripheral portions of lymphocytic infiltration of bronchial walls, and the
lobes obstructed by central tumors (251). The finding of presence of edema fluid and eosinophilic proteinaceous
intraluminal blood, on the other hand, may indeed mimic material within airspaces, representing, at least in part,
the appearance of an endoluminal soft tissue mass (252). retained surfactant. Subsequently, aggregates of lipid-laden
This finding, however, in our experience, is sufficiently rare macrophages accumulate, and progressive lymphocytic
in the absence of central tumor to obviate concern in most infiltration and collagen deposition occur within the
A B
Figure 5-58 Inspissated secretions. A: Axial section thought the mid lung imaged with lung windows shows extensive retained secretions
throughout the length of the left upper lobe bronchus as well as in the bronchus intermedius associated with extensive consolidation and
volume loss in the left lower lobe. Secretions are usually easily identifiable when they result in mottled intraluminal densities (compare with
Fig. 5-40). B: Coronal reconstruction showing to better advantage the extent of retained secretions bilaterally.
pulmonary interstitium (so-called lipid pneumonia).

Finally, with sufficient time, interstitial fibrosis supervenes.
TABLE 5-5
Surprisingly, in the foregoing series, infection proved
NONOBSTRUCTIVE ATELECTASIS: CAUSES
relatively rare and, when present, appeared either as focal
bronchitis or as bronchiolitis, with only minimal or absent Cicatrization atelectasis (localized)
parenchymal involvement. Common
Tuberculosis or fungal infection
Bronchiectasis
Radiation fibrosis
Nonobstructive Atelectasis
Compressive (passive) atelectasis
Mechanisms of nonobstructive atelectasis have been Common
described and include cicatrization atelectasis, passive Pleural effusion
atelectasis, and adhesive atelectasis. Table 5-5 presents Malignant
Benign (especially empyema)
their various causes.
Pneumothorax
Diaphragm dysfunction or elevation
Aortic aneurysm (especially left lower lobe)
Cicatrization (Cicatricial) Atelectasis Rare
Pleural tumors
Cicatrization atelectasis results from scarring and fibrosis Emphysematous bullae
consequent to long-standing inflammatory disease. Often
Adhesive atelectasis (localized)
the result of chronic granulomatous infection, similar Common
changes may be identified in a nonanatomic distribution Acute radiation pneumonitis
in patients with radiation fibrosis and more diffusely in Rare
patients with interstitial pulmonary fibrosis (242). Pulmonary embolism
Cicatrization atelectasis may be differentiated from other Adapted from Woodring JH, Reed JC. Types and mechanisms of pul-
forms of lobar atelectasis, especially obstructive causes, by monary atelectasis. J Thorac Imaging. 1996;11:92–108, with permission.
recognition of the following criteria: (a) endobronchial patent central airways, with the sole exception of a case
obstruction is typically absent; with careful scan technique, with bronchoscopically identified broncholithiasis.
patency of the bronchial tree subtending the involved
segment or lobe can be confirmed (Figs. 5-54 and 5-55);
Compressive (Passive) Atelectasis
(b) inflammatory changes in the lobar bronchi often result
in airway irregularity, narrowing, or distortion; bronchiecta- Compressive atelectasis denotes volume loss secondary to
tic changes are frequently present within the involved lobe extrinsic compression of adjacent lung. Although a distinc-
peripherally and further help to identify collapse as the tion may be made between compressive and passive atelec-
result of chronic inflammation and fibrosis; (c) lobar tasis, defined as loss of volume resulting from interruption
bronchi are displaced to a far greater degree than typically of the normal balance between elastic recoil of the lung
seen in patients with obstructive forms of atelectasis; in the and the opposing outward pull of the chest wall and
case of the right upper lobe, the right upper lobe bronchus diaphragm, from the standpoint of CT interpretation, this
is characteristically displaced posteriorly, with resultant pos- distinction is arbitrary and of no clinical significance. Most
terior orientation of the right main bronchus and bronchus often, both compressive atelectasis and passive atelectasis
intermedius; in the extreme form, lobar airways may appear result from the presence of fluid (benign or malignant) or
to lie in a paramediastinal location; (d) the degree of air (with or without tension pneumothorax) within the
volume loss, in general, is more marked in cicatrization pleural space (Fig. 5-56). Less frequently, compressive
atelectasis than in collapse secondary to endobronchial atelectasis is caused by space-occupying lesions arising
obstruction, presumably as a reflection of the longer time from the pleura, chest wall, or mediastinum, including, as
course over which cicatrization atelectasis occurs; and reported by Gierada et al. (258), atelectasis resulting from
(e) especially in patients with chronic TB, calcified hilar or severe bullous emphysema.
mediastinal nodes with or without punctate parenchymal Fluid within the pleural space is easily identified, distinct
calcifications are frequently identifiable; rarely, an entire from underlying lung, especially after the administration of
lobe or even lung may calcify. The term “auto-lobectomy” or intravenous contrast media (Figs. 5-6 and 5-55). Pleural
“auto-pneumonectomy” seems appropriate for these cases fluid, when sufficiently massive, distorts the usual configu-
of extreme volume loss, with or without diffuse calcifica- ration of collapsed lobes seen in patients with endo-
tion, analogous to the changes seen in patients with long- bronchial lesions. Although investigators have reported that
standing renal TB. exclusion of tumor in patients with atelectatic lobes may be
The chronic middle lobe syndrome best typifies cica- difficult (254), if the lobar bronchi are patent and air
trization atelectasis (Fig. 5-54). As detailed by Kwon et al. bronchograms can be defined in every portion of the
(256) in a clinicopathologic study of 21 patients, middle atelectatic lobe, tumor can be confidently excluded. Also
lobe syndrome can result from a wide range of both strongly suggestive of a lack of tumor within collapsed lung
neoplastic and non-neoplastic causes. These include is the finding of a markedly shrunken lobe surrounded
extrinsic causes of bronchial obstruction, usually from by or floating within pleural fluid (Fig. 5-55). In our experi-
enlarged peribronchial nodes and often the result of ence, this degree of collapse excludes the possibility of an
prior granulomatous infection; intrinsic bronchial infiltrative tumor because only a relatively spongy lobe not
obstruction, as may occur in chronic foreign body aspira- fixed or infiltrated by tumor can collapse to this degree.
tion or broncholithiasis; and nonobstructing conditions, In patients with pleural effusions, CT is also of value in
in particular bronchiectasis (256). Pathologically, a com- identifying malignant changes along pleural surfaces. As
bination of bronchiectasis and bronchiolitis is most discussed in greater detail in Chapter 9, the diagnosis of a
commonly seen, frequently in association with secondary malignant effusion depends on recognition of an abnormal
findings of organizing pneumonia. The finding of granu- parietal pleura, identifiable either as nodular foci of soft
lomas histologically is also common and should suggest tissue density or as an enveloping rind of thickened nodular
the possibility of atypical mycobacterial infection, espe- pleura, usually in association with loculated areas of pleural
cially when it is associated with small, ill-defined fluid (243).
nodules on HRCT. Pathologic changes in patients with Unusual configurations of collapse may result when air,
middle lobe syndrome have been attributed both to lack fluid, or both are present in the pleural space. The magni-
of normal collateral ventilation in these relatively iso- tude of pressure exerted on lung by adjacent pleural fluid
lated lung segments and to poor drainage because of the or air may be great; the result may be initially confusing
oblique orientation of these smaller and shorter lobar radiographically. Of particular interest is the phenomenon
airways. of lung torsion (259). Torsion has been noted to occur
Despite the reported association between middle lobe spontaneously, usually in association with some other pul-
syndrome and bronchial obstruction, in most cases, classic monary abnormality, such as pneumonia or tumor, after
middle lobe atelectasis develops in the presence of a traumatic pneumothorax, and especially as a complication
widely patent lobar bronchus (257). In the study reported of surgery. Radiographically, the hallmark of this entity is
by Kwon et al. (256), for example, 20 of 21 patients with the finding of a collapsed lobe, almost always either the
documented middle lobe syndrome were found to have right or the left upper lobe, less commonly the middle
lobe, and only rarely the lower lobes, occupying an PERIPHERAL AIRWAYS
unusual position. Marked change in position of the
collapsed lobe on sequential radiographs is considered
characteristic. Although the condition is rare, the diagnosis Bronchiectasis
is significant, especially in postoperative patients, because Bronchiectasis is defined as localized, irreversible dilatation
unrecognized torsion is associated with a high mortality of the bronchial tree. Although many different disorders
from resultant pulmonary infarction. have been associated with bronchiectasis, it most
commonly results from acute, chronic, or recurrent infec-
tion (Table 5-6) (261–263). Clinically, the diagnosis of
Adhesive Atelectasis
bronchiectasis is usually made only in patients with severe
Adhesive atelectasis refers to atelectasis resulting from an disease (264,265). Patients can present with a history of
absence of surfactant and may be thought of as a form recurrent pulmonary infections, persistent cough produc-
of resorption atelectasis occurring without bronchial tive of large or small amounts of sputum, or hemoptysis.
obstruction. The prototype for this form of atelectasis is Hemoptysis is frequent, occurring in up to 50% of cases,
acute radiation pneumonitis (Fig. 5-57). The changes and it may be the only clinical finding (2,266).
seen in the thorax consequent to radiation therapy have Bronchiectasis most often results from necrotizing
been described (260). In patients with acute radiation infections such as those caused by Staphylococcus, Klebsiella,
pneumonitis, the predominant finding is ground-glass at- or Bordetella pertussis (261). Granulomatous infections,
tenuation, typically conforming to known radiation including those caused by Mycobacterium tuberculosis
ports. With time, fibrosis ensues with resultant cicatricial (267), atypical mycobacteria, especially Mycobacterium
atelectasis. avium-intracellulare complex (MAC) (268–273), and fungal
TABLE 5-6
CAUSES OF BRONCHIECTASIS AND POSSIBLE MECHANISMS
Condition Mechanisms
Infection (bacteria, mycobacteria, Impaired mucociliary clearance, disruption of respiratory epithelium, microbial toxins, host-mediated
fungus, virus) inflammation, fibrosis, bronchostenosis
Immunodeficiency Genetic or acquired predisposition to recurrent infection
Bronchial obstruction (tumor, Impaired mucociliary clearance, recurrent infection, mucus plugging
foreign body, congenital
abnormalities)
Cystic fibrosis Abnormal airway epithelial chloride transport, impaired mucus clearance, recurrent infection
Asthma Airway inflammation, mucus plugging
Allergic bronchopulmonary Type I and Type III immune responses to fungus in airway lumen, mucus plugging
aspergillosis (ABPA)
Dyskinetic cilia syndrome Genetic defect, absent or dyskinetic cilia, impaired mucus clearance, recurrent infection
(Kartagener syndrome)
Young syndrome Abnormal mucociliary clearance
(obstructive azoospermia)
Yellow nails and lymphedema Unknown, lymphatic hypoplasia and sometimes immune deficiency, predisposition to
syndrome recurrent infection
Alpha1-antitrypsin deficiency Proteinase–antiproteinase imbalance
Tracheobronchomegaly Congenital deficiency of membranous and cartilaginous parts of tracheal and bronchial walls,
(Mounier-Kuhn syndrome) impaired mucus clearance, recurrent infection
Williams-Campbell syndrome Congenital deficiency of bronchial cartilage, obstruction, impaired mucus clearance,
recurrent infection
Systemic diseases (e.g., collagen Various, including inflammation, infection, fibrosis
vascular diseases, inflammatory
bowel disease)
Chronic fibrosis Traction bronchiectasis
Bronchiolitis obliterans Bronchial wall inflammation, epithelial damage, recurrent infection in some cases
(postinfectious, lung
transplant, etc.)
Aspiration, toxic fume inhalation Inflammation
Modified from Davis AL, Salzman SH, eds. Bronchiectasis. Philadelphia: WB Saunders; 1991, with permission.
organisms such as Histoplasma, are also associated with secondary to peribronchial fibrosis and volume loss, evi-
bronchiectasis. Bronchiectasis is increasingly frequently dence of bronchial wall thickening, and, in more severely
identified in HIV-positive patients (274,275). Bronch- affected patients, the presence of discrete cystic masses
iectasis is also a frequent manifestation of common vari- occasionally containing air-fluid levels. Most of these find-
able immunodeficiency syndrome. This heterogeneous ings are nonspecific; a definitive diagnosis of bronchiectasis
entity is a primary immunodeficiency disease character- can seldom be made radiographically (301).
ized by hypogammaglobulinemia and recurrent bacterial
infections in the absence of any known genetic abnormal-
Computed Tomography Evaluation:
ity (276). In one study of 47 documented cases retro-
spectively evaluated, bronchiectasis was diagnosed in 32
Diagnostic Criteria
(68%) (276). There is also a known association between Bronchiectasis traditionally has been classified as cylindric,
this entity and autoimmune diseases, including autoim- varicose, or cystic (264). Although differentiation among
mune hemolytic anemia and autoimmune thrombocy- these categories is useful in assessing disease severity and
topenia, as well as a higher than expected incidence of occasionally may be of value in the differential diagnosis,
lymphoma (277). these distinctions are less important clinically than defin-
Bronchiectasis may occur in association with a variety ing the presence and extent of disease. Evaluating the extent
of genetic abnormalities, generally associated with of bronchiectasis is particularly important, because surgery
abnormal mucociliary clearance, immunodeficiency, or is only rarely performed in patients with involvement of
structural abnormalities of the bronchial wall. This multiple lung segments (296–298).
group includes patients with CF, as well as those with Various descriptive terms have been used to identify
the dyskinetic cilia syndrome and Young syndrome. bronchiectasis. These include “tram tracks” to describe dila-
Bronchiectasis is also frequently identified in patients ted airways coursing within the scan plane, “signet rings”
with a1-antitrypsin deficiency (278). King et al. (279), in to describe vertically coursing dilated airways adjacent to
a study of 14 patients with a1-antitrypsin deficiency, eccentrically located pulmonary artery branches, and both
found evidence of bronchiectasis in 6 (43%). This corre- “string of pearls” and “cluster of grapes” to describe the
lates well with the finding that approximately 50% of appearance of saccular airways seen either along their length
patients with this deficiency manifest symptoms of or in atelectatic lung segments, respectively (299). Of inter-
airway disease, in particular chronic sputum production. est as purely descriptive terms, they are of only limited value
Surprisingly, a frequent association between bronchiecta- for CT interpretation.
sis and other forms of emphysema has not been well HRCT is now the gold standard imaging modality for
documented (279). Patients with Williams-Campbell syn- evaluating bronchiectasis (300). The CT diagnosis of
drome (congenital deficiency of bronchial cartilage) have bronchiectasis is based on recognition of a variety of find-
defective cartilage in the fourth- to sixth-order bronchi ings, both direct and indirect (Table 5-7). Direct findings
often in association with distal air trapping (280,281). include bronchial dilatation (Fig. 5-59), lack of normal
Bronchiectasis is also associated with Mounier-Kuhn bronchial tapering (Figs. 5-3 and 5-60), and visibility of
syndrome (congenital tracheobronchomegaly); with airways in the peripheral 1 cm of lung. Indirect signs include
immunodeficiency syndromes, including Bruton hypo- bronchial wall thickening and irregularity (Figs. 5-1,
gammaglobulinemia and immunoglobulin A (IgA) and 5-61, 5-62), as well as the presence of mucoid impaction
combined IgA-IgG subclass deficiencies; and with the (Figs. 5-1 and 5-63). Ancillary signs have also been described
yellow nail syndrome (yellow nails, lymphedema, and and include tracheomegaly, focal air trapping (as a manifes-
pleural effusions). tation of small airway disease identifiable on expiration
Other diseases that result in airway inflammation, scans) (Fig. 5-64), and emphysema. Needless to say, with
mucus plugging, and bronchiectasis include allergic increasing severity, combinations of these findings may be
bronchopulmonary aspergillosis (ABPA) (282,283) and identified in most cases (Fig. 5-65).
asthma (284–289). Bronchiectasis also has been noted to
occur in patients with obliterative bronchiolitis, espe- Bronchial Dilatation
cially in patients with chronic rejection after heart-lung Recognition of bronchial dilatation is the key to the CT
or lung transplantation (290–294) or bone marrow diagnosis of bronchiectasis. Unfortunately, to date, no
transplantation, as a result of rejection or chronic graft- consistent CT criteria have been determined (Table 5-8)
versus-host disease (295). An increased incidence of (286,301–304).
bronchiectasis has also been noted in patients with coal As initially defined (299), bronchiectasis is most often
workers’ pneumoconiosis presumably secondary to coal diagnosed when the internal diameter of a bronchus
dust exposure (327). appears greater than the diameter of the adjacent pul-
The radiographic manifestations of bronchiectasis have monary artery branch (Figs. 5-59, 5-65, 5-66) (305–309). In
been well described. They include a loss of definition of patients with bronchiectasis, the bronchial diameter is often
vascular markings in specific lung segments, presumably much larger than the pulmonary artery diameter, a finding
TABLE 5-7
HIGH-RESOLUTION COMPUTED TOMOGRAPHIC FINDINGS
IN BRONCHIECTASIS
Direct Signs Indirect Signs
1. Bronchial dilatation 1. Bronchial wall thickening

Increased bronchoarterial ratio 0.5 times the diameter of an adjacent pulmonary
Signet ring sign (vertically artery (vertically oriented bronchi)
oriented bronchi) 2. Mucoid impaction or fluid-filled bronchi; tubular or
Contour abnormalities, Y-shaped structures; branching or rounded opacities
cylindrical, varicose, or cystic in cross section
bronchiectasis Air-fluid levels
2. Lack of tapering 2 cm distal 3. Centrilobular nodules or tree-in-bud opacities
to bifurcation 4. Mosaic perfusion
3. Visibility of peripheral airways 5. Air trapping on expiratory scan
within 1 cm of the costal pleura 6. Bronchial artery hypertrophy
7. Atelectasis or emphysema
From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations.
Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.
that reflects not only the presence of bronchial dilatation, subjects) evaluated had abnormal BA ratios. Similarly, Kim
but also some reduction in pulmonary artery size, likely a et al. (303) also documented the unreliability of abnormal
consequence of hypoxia in regions of bronchiectasis result- BA ratios in the diagnosis of bronchiectasis, especially for
ing in decreased lung perfusion and a corresponding persons living at high altitudes. In this study, 9 of 17 (53%)
decrease in pulmonary artery size (303). of healthy persons living at an altitude of 1,600 m had evi-
The definition of bronchoarterial (BA) ratio varies widely dence of at least one bronchus equal to or greater in size
among authors. Typically defined as the internal or luminal than the adjacent pulmonary artery; of greater import, these
diameter divided by the diameter of the adjacent pul- same authors found that 2 of 16 (12.5%) persons living at
monary artery (Fig. 5-65 and 5-66), the reliability of this sea level similarly showed evidence of at least one abnor-
sign is limited (301,303,310). Lynch et al. (311), in an assess- mally dilated airway. Furthermore, and surprisingly, evalua-
ment of 27 healthy subjects living in Denver, Colorado, tion of the distribution of these abnormal airways failed
found that 37 (26%) of 142 bronchi (or 59% of study to identify any significant difference in the likelihood of
A B
Figure 5-59 Bronchiectasis: direct sign—increased bronchoarterial (BA) ratio. A, B: 1-mm sections through the lower lung zones.
respectively, show typical appearance of dilated bronchi resulting in increased BA ratios (so-called signet ring sign (arrows in A and B).
Identifying increased BA ratios is easiest when bronchi are sectioned at right angles.
Figure 5-61 Measurement of bronchial wall thickness using

the T/D ratio. This ratio is defined as wall thickness (T) divided
by the total diameter of the bronchus (D). In normal persons, it
averages about 0.2 or 20%. (From Naidich DP, Webb WR, Grenier
PA, et al. Imaging of the airways: functional and radiologic
correlations. Philadelphia: Lippincott Williams & Wilkins; 2005,
with permission.)
diameters, important differences may be also attributable

to the use of visual inspection versus objective measure-
ments of bronchial and arterial dimensions. As empha-
sized by Kim et al. (303), visual inspection alone may
Figure 5-60 Bronchiectasis: direct sign—lack of peripheral lead to overestimation of BA ratios because of a subtle
tapering. CT section through the carina shows lack of tapering of optical illusion in which the diameter of hollow circles
the upper lobe bronchi as they extend peripherally (arrows). Note appears larger than that of solid circles despite their
that the bronchial walls also appear slightly nodular.
identical size.
Despite potential variability in normal bronchial size,
abnormal BA ratios in airways either by lobe or by antero- CT measurements of bronchial diameter have proved
posterior location within the lungs (303). A similar lack of remarkably consistent among and between readers (312,
variation by segments, lobes, or lungs has also been 313). Desai et al. (313) evaluated both interobserver and
reported by Kim et al. (310). intraobserver variation in CT measurements of bronchial
As noted, the definition of an abnormal BA ratio varies wall circumference in 61 subsegmental bronchi and found
widely among reported series (301–304,310). In addition the reproducibility of these measurements sufficiently
to variations in size criteria, with a range as great as one to significant to be clinically useful in demonstrating the pro-
two times the diameter of the adjacent pulmonary artery, gression of bronchiectasis. However, despite these data,
as well as the use of internal versus external bronchial time-consuming measurements of this type are unlikely to
Figure 5-62 Bronchiectasis: indi-

rect sign—bronchial wall thickening.
A–C: Target reconstructed views
through the right lower lobe show-
ing focal bronchial wall thickening
(arrows). This appearance is equally
typical of chronic airway inflamma-
tion. Rarely, when focal, early carci-
noma may also result in a similar
appearance. A follow-up low-dose
CT study obtained within a few
months is usually adequate to differ-
A, B, C entiate these entities.
A B
Figure 5-63 Bronchiectasis: indirect sign—mucoid impaction. A: Target reconstructed 1-mm section through the base of the middle lobe
shows multiple nodules, some with a subtle branching configuration in the base of the middle lobe. Note the presence of dilated air-filled
bronchi as well. B: Follow-up magnified 5-mm section at the same level as shown in A following antibiotic therapy now shows that the nodu-
lar opacities seen in A represent mucus-filled peripheral airways.
A B
Figure 5-64 Chronic airway inflammation; bronchitis—bronchial wall thickening and focal air trapping. A: Magnified 1-mm section
through the right lung base shows typical appearance of bronchial wall thickening consistent with chronic airway inflammation. There is het-
erogeneous density noted in the right lower lobe and diffuse decreased lung density in the base of the middle lobe. B: Identical image as
shown in A imaged with narrower windows. Although this shows to better advantage the extent of mosaic attenuation due to diffuse air
trapping, note that bronchial wall thickening appears artifactually worse. In this case a combination of windowing is most advantageous to
display the true extent of disease (compare with Fig. 5-5).
A B
Figure 5-65 CT findings in severe bronchiectasis: cystic fibrosis. A, B: 1-mm sections through the mid and lower lung fields in a patient
with diffuse bronchiectasis show typical appearance of all findings characteristic of bronchiectasis, including increased bronchoarterial
ratios, lack of peripheral tapering, bronchial wall thickening and luminal sacculations, and multiple foci of mucoid impaction (arrows in B).
be of routine clinical use. As emphasized by Diederich

et al. (304), visual inspection remains the mainstay for
TABLE 5-8
assessing bronchial dilatation because obtaining objective
BRONCHIECTASIS: COMPUTED TOMOGRAPHIC
measurements (with or without the use of calipers) is time
consuming and often clinically impracticable. In this CRITERIA
regard, visual inspection has been shown to have accept- Bronchial dilatation
able interobserver variability. Using visual inspection only, ID bronchus adjacent PA (214)
Diederich et al. (304) found close agreement among three OD bronchus adjacent PA (222)
readers in both the detection (k 0.78) and the assess- Minimum ID 110% (211,213)
Minimum ID 150% (210)
ment of the severity (k 0.68) of bronchiectasis. Grade 1, 2 times diameter adjacent PA; grade 2, 3 times;
Recently, attempts have been made to provide automatic grade 3, 3 times (223)
segmentation of bronchi to generate CT bronchograms
Lack of tapering
(Fig. 5-67). These have the potential to more accurately as- Visual inspection (194,209)
sess true luminal diameters and bronchial wall thickening Maintains same diameter 2 cm after branching (221)
by ensuring that true cross-sectional images are obtained, Peripheral bronchial dilatation
especially through obliquely coursing airways (314). Peripheral 1/2 of lung (223)
Peripheral 1/3 of lung (194,211,336)
Within 1 cm of costal/mediastinal pleura (210)
Abutting mediastinal pleura (210)
Bronchial wall thickening
Subjective (191,194,210)
Bronchial wall 2 times thickness normal adjacent bronchi (230)
Minimum diameter adjacent PA (229)
Grade 1, 0.5 diameter adjacent PA; grade 2, 0.5–1.0; grade 3
Figure 5-66 Normal and abnormal bronchoarterial (BA) ratios. 1 times (223)
A normal BA ratio averages 0.65 to 0.7 in young or middle-aged ID 80% total outside diameter (213)
patients. A BA ratio of 1 or more may be seen in some normals
older than 65 years or in patients living at high altitude. A BA ratio Mucoid impaction
of 1.5 is typical of bronchiectasis and usually reflects increased Tubular or Y-shaped structures branching or rounded
bronchial diameter and decrease in size of the artery. This appear- opacities in cross section differentiated from blood
ance mimics a signet ring and is termed the signet ring sign. (From vessels by position relative to adjacent patent
Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: bronchi (229)
Williams & Wilkins; 2005, with permission.) ID, internal diameter; OD, outer diameter; PA, pulmonary artery.
A B
Figure 5-67 Airway segmentation. A: 3D externally rendered view of the segmented tracheobronchial tree providing a CT broncho-
graphic view of the airways. B: Segmented airway tree in a different individual than in A, within which the central axes or center lines are
delineated. These may serve either as a template for developing automated virtual bronchoscopic images or as a means for obtaining true
cross-sectional images through user-selected airways, providing precise measurements of luminal area or wall thickness. (A Courtesy of AP
Kiraly, Siemens Corporate Research, Princeton, New Jersey; B from Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: func-
tional and radiologic correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
Lack of Bronchial Tapering. Lack of bronchial tapering indicative of bronchial dilatation, these authors found that
has come to be increasingly recognized as a critical means visualization of bronchi within 1 cm of the pleura could
for diagnosing bronchiectasis, in particular, subtle cylindric be identified in 81% and 53% of patients with clinical or
bronchiectasis (Figs. 5-60 and 5-65). First emphasized by pathologic evidence of cylindric bronchiectasis, respec-
Lynch et al. (315) as a necessary finding, lack of bronchial tively. Most important, although bronchi were identified
tapering has been reported by some investigators to be the within 1 cm of the mediastinal pleura in 40% of anatomi-
most sensitive means for diagnosing bronchiectasis. Kang cally normal persons, in no case was a bronchus identified
et al. (301), for example, in an assessment of 47 lobes with within the same distance from the costal pleura in any
pathologically proved bronchiectasis, found lack of taper- anatomically normal person (316). Although bronchi nor-
ing of bronchial lumina in 37 (79%) cases as compared mally may be seen within 1 cm of the mediastinal pleura,
with abnormal BA ratios seen in only 28. Accurate assess- bronchi abutting the mediastinal pleura were always
ment of this finding is difficult in the absence of contigu- abnormal.
ous sections, especially for vertically or obliquely oriented
airways. Despite frequent citation, the value of this sign in Bronchial Wall Thickening
studies in which noncontiguous, high-resolution 1- to In distinction to bronchial dilatation, the finding of
3-mm sections only are obtained is doubtful. Use of con- bronchial wall thickening is not specific for bronchiectasis,
tiguous images is especially critical when identification re- being frequently seen in patients with nonspecific chronic
quires that the diameter of airways remain unchanged for bronchitis, for example. Not surprisingly, however, bronchial
at least 2 cm after branching (303). wall abnormalities are commonly found in patients with
bronchiectasis, thus making identification of this finding im-
Visualization of Peripheral Airways. Another frequently portant (Figs. 5-1, 5-62, 5-64, 5-65). Similar to variations in
cited manifestation of bronchiectasis is identification of the definition of BA ratio, there are variations in suggested
airways in the lung periphery (Fig. 5-60 and 5-65). methods for assessing bronchial wall thickness, as well.
Variously defined as the ability to identify airways in the As reported by Weibel (317), 2nd- to 4th-generation
peripheral one half to one third of the lung, this finding segmental airways have mean diameters of between 5 and
has proved useful because normally, even with HRCT tech- 8 mm, with walls measuring approximately 1.5 mm;
nique, airways are rarely identifiable in the outer portions 6th- to 8th-generation airways have mean diameters meas-
of the lung parenchyma. Kim et al. (316) more precisely uring between 1.5 and 3 mm, with walls of approximately
characterized the value of this sign by differentiating 0.3 mm; and 11th- to 13th-generation airways have diame-
between airways near the costal pleura and those near the ters measuring 0.7 to 1 mm, with walls of 0.1 to 0.15 mm
mediastinal pleura. Using 1 cm from the pleural surface as (Table 5-1). Because a bronchus is usually recognized only
when its walls are visualized, one cannot resolve bronchi internal diameter of airways as being either between 50%
with a wall thickness of less than 300 mm. This corre- and 80% or more than 50% of the diameter of the adja-
sponds to bronchi of about 1.5 to 2 mm in diameter, cent pulmonary artery (indicative of mild or severe
equivalent to the 7th to 9th generations of airways. Rarely, bronchial wall thickening, respectively) is an acceptable
smaller bronchi, approximately 1 mm in diameter, can alternative (304).
be recognized coursing perpendicularly with the plane of a
1-mm-thick section by virtue of their lower density and
proximity to an adjacent pulmonary artery, even though Mucoid Impaction
their walls cannot be resolved (318,319). The appearance of fluid- or mucus-filled airways depends
To date, identification of thickened bronchial walls has on both their size and orientation relative to the CT scan
been largely subjective. Attempts to define bronchial wall plane. Larger fluid-filled airways result in abnormal lobular
thickening more precisely have largely involved visually or branching structures when they lie in the same plane as
assessing or measuring ratios between bronchial walls and the CT scan (Fig. 5-63). Although confusion between fluid-
adjacent pulmonary arteries or nearby normal airways filled branching airways and abnormally dilated vessels is
(20,304,316,320). In difficult cases, because bronchiectasis possible, especially in studies performed without intra-
and bronchial wall thickening are often multifocal rather venous contrast administration, in nearly all cases, recogni-
than diffuse and uniform, a comparison of one lung tion of mucoid-impacted airways is straightforward. The
region with another can be helpful in making this diagnosis is usually simplified by the identification of
diagnosis. As a general rule, normal airways have a T/D other foci of bronchiectasis (Fig. 5-65). In problematic
ratio—defined as wall thickness divided by total bronchial cases, the distinction between larger fluid-filled bronchi
diameter—of 0.2 or 20% (Fig. 5-61). Alternatively, the and dilated blood vessels is easily made by repeat scanning
relationship between bronchial wall thickness and of patients after a bolus of intravenous contrast medium
bronchial diameter may be assessed using the bronchial (Fig. 5-68).
lumen ratio (BLR), defined as the inner diameter of the The finding of dilated, mucus-filled airways, especially
bronchus divided by its outer diameter (310). At the sub- when central or predominantly segmental or lobar in
segmental level, the BLR in normal individuals averages distribution, should alert one to the possibility of central
0.66 0.06 with a range of 0.51 to 0.86. Despite the endobronchial obstruction, resulting either from tumor
reported accuracy and reproducibility of these measure- (Fig. 5-69) or from foreign body aspiration (Fig. 5-24). As
ments, in clinical practice, using a visual estimation of the previously discussed, the use of intravenous contrast
Figure 5-68 Mucoid impaction—

evaluation with contrast enhanced
CT. A: Magnified view of the right
lower lobe imaged with lung win-
dows shows a tubular structure
in the posterobasilar segment. B:
Same image as A imaged with nar-
row windows. Lack of enhancement
following intravenous contrast is
diagnostic of mucoid impaction. In
this case these findings proved
to be secondary to long-standing
bronchiectasis. This appearance, how-
ever, has also been described in pa-
tients with focal bronchial obstruc-
tion due to both central and periph-
eral tumors, including carcinoids and
rarely bronchoalveolar cell carcino-
mas (see Fig. 5-69). In indeterminate
cases, follow-up low-dose CT studies
and/or bronchoscopy may be neces-
A, B sary for definitive diagnosis.
Figure 5-69 Endobronchial metas-

tasis. A: Axial section shows a filling
defect obstructing the bronchus
intermedius (arrow). B: Axial image
obtained below A shows extensive
mucoid impaction of the bronchi
within the superior segment second-
ary to proximal endobronchial
obstruction. Biopsy proved metasta-
tic colon cancer. (From Naidich DP,
Webb WR, Grenier PA, et al. Imaging
of the airways: functional and ra-
diologic correlations. Philadelphia:
Lippincott Williams & Wilkins; 2005,
A, B with permission.)
media in this setting may allow differentiation between in 85%. Results included 6 lobes with obliterative bronchi-
central tumor and fluid-filled peripheral airways (Fig. 5-6 olitis, 18 with inflammatory or suppurative bronchi-
and 5-70). olitis, and 16 with both obliterative and inflammatory
Mucoid impaction also occurs as a component of the bronchiolitis. Of these, CT findings consistent with bron-
sequestration spectrum, either as an isolated phenome- chiolitis were identified in 30 (75%) of 47 lobes, including
non, usually affecting the left upper lobe in patients with a pattern of mosaic attenuation (n 21), bronchiolectasis
bronchial atresia (Fig. 5-14), or typically as a component (n 17), centrilobular nodular or branching opacities
of intralobar sequestration. (n 10), and consolidation (n 8) (301).
Fluid-filled small airways in the peripheral lung are The finding of mosaic attenuation or focal air trapping
usually identifiable either as branching structures within on expiratory scans, in particular, may prove of special
the center of secondary lobules, aptly described as having interest as an early manifestation of bronchiectasis (Fig. 5-
a tree-in-bud appearance (267), or as ill-defined cen- 64). In one study of 70 patients (324) with CT evidence of
trilobular nodules (321). These are frequently identified bronchiectasis visible in 52% of lobes evaluated, areas of
in association with bronchiectasis (Fig. 5-2) and are dis- decreased attenuation (mosaic perfusion) were visible on
cussed in greater detail later. inspiratory scans in 20% of lobes and on expiration (air
One unique manifestation of mucoid impaction is trapping) in 34%. Although areas of decreased attenuation
the finding of high-attenuation mucus (Fig. 5-71). This on expiratory scans were more prevalent in lobes with se-
appearance is highly suggestive of ABPA and is frequently vere (59%) or localized (28%) bronchiectasis, in 17% of
also present in the nasal sinuses as well (322,323). The lobes, air trapping could be identified in the absence of as-
finding of high-density mucus presumably is due to sociated bronchiectasis. This finding has led to speculation
the deposition of calcium within chronically inspissated that evidence of bronchiolar disease may, in fact, precede
secretions and has been reported to be seen in nearly one and even lead to the development of bronchiectasis (324).
quarter of cases (323). In this same study, the presence of decreased attenuation on
expiratory scans was also associated with mucoid im-
Ancillary Signs: Bronchiolar Disease paction, seen in 73% of lobes with large mucus plugs and in
As CT has become established as the modality of choice 58% of those with centrilobular mucus plugs. These same
for assessing bronchiectasis, investigators have become authors noted a correlation (r 0.40, p 0.001) between
increasingly aware that, in addition to signs of large airway the total extent and severity of bronchiectasis and the extent
inflammation, most patients also have evidence of small of decreased attenuation shown on expiratory CT. Not sur-
airway disease (Fig. 5-64). Kang et al. (301), for example, prisingly, in 55 patients who had pulmonary function tests,
in their study of 47 resected lobes with documented the extent of expiratory attenuation abnormalities proved
bronchiectasis, found pathologic evidence of bronchiolitis inversely related to measures of airway obstruction such as
forced expiratory volume in 1 second and the ratio of that sensus exists on the best method for estimating extent and
parameter to forced vital capacity (FEV1/ FVC) (234). severity of disease in patients with bronchiectasis (279,
302,316,320,324–329). Although of only limited value in
Extent and Severity distinguishing among the various causes of bronchiectasis
Analogous to the lack of standardization for estimating (316), assessing disease extent and severity is important,
both bronchial dilatation and wall thickening, little con- especially in patients for whom sequential CT studies are
B–E
Figure 5-70 Mucoid impaction: CT-bronchoscopic correlations. See Color Figure 5-70H,I. A: 5-mm section through the mid lung shows a
tubular structure in the periphery of the left upper lobe suggestive of focal mucoid impaction. B–E: Magnified views of this lesion following
a bolus of intravenous contrast shows no evidence of enhancement. F, G: Magnified views at the same level as A–E obtained at the time of
CT-guided ultrathin bronchoscopy confirming that the tip of the bronchoscope is at the medial end of this lesion. H: View through the tip of
the ultrathin bronchoscope shows purulent secretions obscuring the orifice of a seventh- or eight-order bronchus. I: View obtained several
seconds after H following aspiration of secretions showing a normal patent bronchus. In this case, initial prebronchoscopic planning was ac-
complished using virtual bronchoscopy (not shown) (continued).
F, G H
mucus plugging, peribronchial thickening, generations of

bronchial divisions involved, the number of bullae, and
the presence of emphysema, in addition to a three-point
severity scale to calculate a global CT score based on
25 points to correspond with prior radiographic scoring
systems. Based on this approach, these authors found CT
to be a valuable tool for objectively evaluating the extent
I
to be used to evaluate response to therapy and progression

of disease (320,325–329).
In general, attempts have focused on the use of grading Figure 5-71 Mucoid impaction: high-density mucus. Non–
systems designed to assess the number of either segments contrast-enhanced CT section in a patient with right upper lobe
or lobes, with corresponding correlation with disease collapse shows multiple high-density nodular opacities within
the collapsed lobe. This appearance on non–contrast-enhanced
severity. Bhalla et al. (320), in the most detailed method scans is consistent with central bronchiectasis due to allergic
to date, used nine separate variables, including extent of bronchopulmonary aspergillosis.
and severity of bronchiectasis in patients with CF. Unfor- evaluation using a modification of the scoring system
tunately, this system is extremely time consuming and is recommended by Bhalla et al., replaced the assessment of
not of value for routine clinical assessment. segments (which are often difficult to identify) with that
A number of modifications to this approach have been of lobes using a five-point scale based on a visual assess-
suggested (302,330–335). For the most part, these have ment of the number of abnormal bronchi as less than
emphasized differences in the definition of bronchial 25%, 25% to 49%, 50% to 74%, or more than 75%,
dilatation, bronchial wall thickening, and the extent of dis- respectively. Although this approach has the benefit of
eases as measured by either segments or lobes. Other relative simplicity, it has not proved extremely reliable.
differences include methods for describing the extent of Using this same system, Diederich et al. (304) found only
disease on axial images, with some investigators assessing moderate interobserver agreement among three readers
the number of generations of abnormal airways (320), (k 0.58) for assessing disease severity in a study of
nodules, and mosaic attenuation—findings indicative of 88 patients with suspected bronchiectasis. Although the
small airway involvement (331,332,334,335). use of CT scoring systems provides precise assessment of
Allowing for differences in approach, these scoring disease severity, the implications of these data are that with
systems have in common consistent good correlation with the exception of specific clinical indications like monitor-
more traditional radiographic, clinical, and functional ing response to therapy in children and young adults with
criteria (302,320,331,334,335). Although most of these CF, general acceptance of the need for routine CT scoring
studies have focused on the use of CT in patients with CF, for bronchiectasis will have to await an approach based on
similar close correlation between the degree of bronchial automated computer assisted diagnosis.
wall thickening, the severity and extent of bronchiec-
tasis, and the extent of mosaic attenuation or air trap-
ping with functional indices of airway obstruction have Pitfalls in Diagnosis
been shown in patients with causes of chronic airway Despite the overall accuracy of CT for diagnosing bronchiec-
inflammation and infection other than CF (326,336–338). tasis, certain potential pitfalls have been described (309).
Lynch et al. (337), for example, in a study of 261 patients Particularly frequent are those due to motion artifacts, both
with symptomatic bronchiectasis, excluding patients with respiratory and cardiac (339). Respiratory artifacts, in partic-
CF, ABPA, and fungal and mycobacterial infections, found a ular, result in ghosting that can closely mimic the appear-
weak but significant correlation between the degree of mor- ance of tram tracks.
phologic abnormalities on CT and the extent of Bronchiectasis is especially difficult to diagnose in
physiologic impairment. Specifically, CT scores for the patients with either parenchymal consolidation or atelecta-
severity and extent of bronchiectasis correlated with FEV1 sis because CT often discloses dilated peripheral airways
(r 0.362, p 0.0001) and forced vital capacity (FVC) that will revert to normal after resolution of the lung
(r 0.362, p 0.0001), and CT scores for bronchial wall disease, so-called reversible bronchiectasis. Another poten-
thickening correlated with the FEV1 (r 0.367, tial pitfall related to consolidation is that, as discussed
p 0.0001) and forced vital capacity (FVC) (r 0.239, previously, consolidation may obscure vascular anatomy
p 0.0001). and thus may render interpretation of BA ratios difficult or
Similar correlations have been reported by others (336). impossible (301).
Roberts et al. (338), for example, showed that the extent Rarely, the appearance of cavitary nodules in patients
and severity of bronchiectasis and bronchial wall thicken- with widespread bronchoalveolar cell carcinoma, cavitary
ing correlated with the severity of airflow obstruction, in metastases, or even cystic Pneumocystis carinii (now called
particular with decreased attenuation of the lung on expira- Pneumocystis jiroveci) pneumonia or Langerhans cell histi-
tory scans having the closest correlation (r 0.55, ocytosis may result in “pseudobronchiectasis” (301). In
p 0.00005). Edwards et al. (336), in a comparison of CT patients with Langerhans cell histiocytosis, bizarrely
findings in children with and without CF, showed that, shaped cysts are often seen, especially in the upper lobes.
although there was no correlation between CT findings Because these cysts may seem to branch, their appearance
and sputum production, strongest relationships could may be suggestive of bronchiectasis. In fact, pathologi-
be shown between the extent of bronchiectasis, bronchial cally, some of these cystic abnormalities do indeed repre-
wall thickening, and air trapping on CT with FEV1 and sent abnormally dilated bronchi, ostensibly the result of
FEF25%–75%. peribronchiolar inflammation.
Despite these data, the use of systematic scoring systems Most important, bronchial dilatation may occur as a
has not translated to routine clinical practice. In part, this component of diffuse fibrotic lung disease. The result is
reflects the cumbersome nature of the task of evaluating a so-called traction bronchiectasis (Fig. 5-72). In most cases,
large number of variables in increasingly large datasets. In this condition is easily identified because peripheral
addition, not surprisingly, there are problems related to bronchi appear irregularly thick walled or corkscrewed and
both interobserver and intraobserver variability. Smith are invariably found in association with diffuse reticular
et al. (302,320), for example, in an attempt to simplify changes or with honeycombing (340).
A, B
Figure 5-72 Traction bronchiecta-

sis—two examples. A, B: Magnified
high-resolution images through the
right lung show dilated peripheral
airways associated with findings
consistent with early idiopathic
interstitial pneumonitis, including
fine peripheral, subpleural reticula-
tion and ground-glass attenuation.
C, D: Magnified views in a patient
with sarcoidosis, different from the
patient in A and B, show the typical
appearance of dilated airways
associated with focal scarring and
architectural distortion. Bronchial
dilatation in these cases should
not be mistaken for bronchiectasis
C, D resulting from infection.
Specific Etiologies
between the various causes of bronchiectasis. However, a
The CT appearances of several diseases associated with few conditions associated with bronchiectasis have been
bronchiectasis have been described. As discussed later, reported to have distinctive HRCT appearances that can aid
although CT is of value for identifying morphologic in their diagnosis. The most important of these diseases in
features in patients with both acute and chronic airway our experience are mycobacterial disease, including atypical
infection, CT is of considerably less value for differentiating mycobacterial infection, CF, and ABPA.
Mycobacterial Infection
Nontuberculous (Atypical) Mycobacterial Infection.
Bronchiectasis in association with ill-defined clusters of
centrilobular lung nodules on CT is characteristic of nontu-
berculous mycobacterial infection resulting from nontuber-
culous Mycobacterium (NTMB), in particular by MAC and
Mycobacterium kansasii (271–273,341). NTMBs are obliga-
tory aerobic gram-positive rods with variable acid-fast
staining that are ubiquitous organisms widely distributed
throughout the environment, resulting in widespread
human exposure. NTMBs are typically low-grade pathogens
that infrequently result in clinical infection (341).
Given the indolent nature of infection in most cases,
the diagnosis is often difficult to establish. To distinguish
definite infection from colonization, therefore, the follow-
ing diagnostic criteria have been proposed: (a) isolation
of NTMB from lung biopsy specimens; (b) identification Figure 5-73 Atypical mycobacterial (MAC) infection. Section
of granulomas or acid-fast bacilli on transbronchial biopsy through the mid lung shows characteristic appearance of varicose
specimens in association with a positive culture from and cystic bronchiectasis within the middle lobe and lingula with
relative sparing of the lower lobes. Note the presence of periph-
other respiratory secretions; (c) four or more sputum eral bronchiolectasis in both lower lobes.
cultures demonstrating heavy growth of NTMB; (d) bron-
choscopic washings demonstrating M. kansasii, because
this organism is only rarely present as a contaminant; and CT findings in patients with NTMB infection often are
(e) positive cultures obtained from bronchoscopic wash- indistinguishable from those seen in patients with TB
ings in association with positive blood or marrow cultures (Fig. 5-74). However, as noted previously, a distinct pat-
in AIDS patients only (342). Lesser degrees of diagnostic tern of involvement predominantly affecting the middle
certainty result either in patients with bronchoscopic lobe and lingula has been identified in elderly women
brushings demonstrating acid-fast bacilli with positive (271–273,341). As documented by Tanaka et al. (343) in a
cultures obtained from bronchoalveolar lavage or in prospective study of 26 patients evaluated over a 4-year
patients with two or three positive sputum cultures in period with findings on CT suggestive of MAC pulmonary
association with radiographic evidence of pulmonary disease, including clusters of small nodules in the lung
infiltrates or nodules (341,342). Definite diagnosis is re- periphery and bronchiectasis, 13 (50%) proved to have
quisite because patients usually require a prolonged course positive MAC cultures from bronchial washings. Of these
of multidrug therapy. patients, epithelioid granulomas were demonstrated in
Three groups of patients have been identified to be at risk 8 of 13. These findings are nearly identical to those
for pulmonary infection. Most commonly, NTMB infection reported by Swensen et al. (272), who also reported an
affects elderly white men in their 60s who have pre-existing approximately 50% rate of positive cultures in patients
lung disease, usually emphysema. Other predisposing with bronchiectasis and nodules on CT. These nodules
factors include diabetes, alcoholism, and nonpulmonary have a distinct appearance, identifiable as clustered
malignant diseases. Infection in these patients results in a around peripheral vessels and airways frequently having a
spectrum of radiographic findings indistinguishable from tree-in-bud configuration. Histologically, this pattern is
those seen in patients with TB, including linear and nodular due either to impaction of peripheral bronchioles or to
densities most often involving the upper lobes and superior peribronchiolar granulomas (272). The finding of granu-
segments of the lower lobes, with or without evidence of lomas is especially significant in this population because
cavitation and endobronchial spread. The second group of it constitutes evidence that bronchiectasis, instead of
patients at risk comprises elderly women, typically without being a precursor, is likely the result of chronic infection
underlying predisposing factors. Radiographic findings in (273,343).
this group include scattered nodular densities associated
with bronchiectasis, preferentially involving the lingular Cystic Fibrosis
and middle lobes. This presentation is especially striking on CF is a multisystem autosomal recessive disease in which a
CT (Fig. 5-73). Finally, also at risk are immunocompro- structural defect of the CF transmembrane regulator pro-
mised patients, especially those with AIDS. Unlike x-ray tein leads to abnormal chloride transport across epithelial
studies in the preceding two groups, radiographs are typi- membranes. This phenomenon leads to abnormally low
cally normal in these patients. To date, MAC infection has water content of airway mucus, resulting in decreased
been identified in up to 20% of AIDS patients, with esti- mucus clearance and mucus plugging, and eventually to
mates as high as 50% of patients infected at autopsy (341). an increased incidence of bacterial infections, in particular
Figure 5-74 Tuberculosis. A, B:

Magnified views of sections through
the right upper lobe show a combi-
nation of severe bronchiectasis
associated with extensive mucoid
impaction and consolidation. In this
case, the remainder of the lungs was
clear (not shown). This pattern of
extensive upper lobe disease should
suggest the possibility of active
A, B tuberculosis.
those due to Pseudomonas. Bronchial wall inflammation, These areas are best appreciated on expiratory scans, which
progressing to widespread and severe bronchiectasis, is should be obtained in all patients with suspected CF (25).
universal in patients with long-standing disease, and these In addition to findings of bronchiectasis and bronchi-
features are commonly visible even on routine chest radi- olectasis, cystic or bullous lung lesions can also be identified
ographs (344). CF is the commonest cause of pulmonary predominantly in the subpleural regions of the upper lobes.
insufficiency in the first three decades of life (345,346). Hilar or mediastinal lymph node enlargement and pleural
CT findings in patients with CF have been well abnormalities can also be seen, largely reflecting chronic
described (Fig. 5-65). The hallmark of disease is the find- infection. Pulmonary artery dilatation resulting from pul-
ing of diffuse bronchiectasis limited to the central airways monary hypertension can also be noted in patients with
in up to one third of cases (25,320,324,331,347,348). long-standing disease.
These findings mimic those seen in patients with ABPA, a HRCT can demonstrate morphologic abnormalities in
frequent complication in patients with CF. Although all patients with early CF, even those whose pulmonary func-
lobes are involved in advanced cases, early in the course tion tests and radiographs are normal. Santis et al. (349),
of disease, findings may be restricted to the upper lobes, in a study of 38 patients with mild CF, showed that chest
with right upper lobe predominance frequently noted radiographs were normal in nearly one half; in distinction,
(324,331,332,334,335,346,349,350). Less frequently, HRCT disclosed evidence of bronchiectasis in 77% of these
disease has been noted to exclusively involve the lower same patients and in 65% of those with normal chest radi-
lobes (351). ographs (349). Similar findings have been reported by
Bronchial wall or peribronchial interstitial thickening is Lynch et al. (350), including HRCT evidence of bronchial
also commonly present in patients with CF, especially early wall thickening, bronchiectasis, centrilobular small airway
in the course of disease. Thickening of the wall of proximal abnormalities, and lobular or segmental inhomogeneities
right upper lobe bronchi was the earliest abnormal feature representing mosaic perfusion or air trapping all in pa-
visible on HRCT in one study of patients with mild CF tients with normal chest radiographs. Helbich et al. (334),
(349). Mucus plugging is also common, frequently visible in a study of HRCT findings in 117 patients with CF, fur-
in all lobes (348). Other parenchymal findings include ther confirmed the utility of CT to identify morphologic
consolidation or atelectasis, with volume loss identified in changes in the lungs of CF patients by identifying bron-
up to 20% of lobes in patients with advanced disease. Also chiectasis in 80.3%, peribronchial wall thickening in
frequently noted, especially early in the course of disease, 76.1%, mosaic perfusion in 63.9% and mucus plugging
are findings indicative of acute infectious bronchiolitis in 51.3%.
resulting in branching, tree-in-bud, or nodular centrilobu- Even in patients with more extensive disease, CT may
lar opacities as well as focal geographic areas of decreased be of value by disclosing otherwise unsuspected sites of
lung attenuation resulting from regional air trapping (350). disease. As reported by Bhalla et al. (320), in a study of
14 patients with CF, of a total of 162 segments assessed, cise parameters and findings at CT, leading the study’s
bronchiectasis was detected in 124 segments using authors to speculate that comprehensive evaluation of these
HRCT, compared with only 71 segments identified radi- patients requires that all three tests be obtained (329).
ographically. Similar findings have been reported by Although CT is of value for assessing morphologic
Hansell and Strickland (324). changes in patients with CF, its role in the assessment of
patients prior to lung transplantation, especially to rule
CT-Clinical Correlation. Given the ability of HRCT to out occult lung tumors, has been questioned (356).
identify the entire spectrum of morphologic changes that
occur in the lungs of patients with CF, it is not surprising Allergic Bronchopulmonary Aspergillosis
that CT has also been evaluated as a possible means for ABPA is a hypersensitivity reaction to Aspergillus species
clinical assessment of disease severity as well as a means commonly found in the soil. It occurs almost exclusively
for monitoring disease progression and response to in patients with either asthma or CF—in 1% to 2% of
therapy (Fig. 5-65). HRCT has been shown to identify patients with asthma and in about 10% of patients with
morphologic changes in asymptomatic CF patients with CF (282,357). The disease is the result of immunologic
normal chest radiographs and pulmonary function tests responses to the endoluminal growth of fungal species.
(349,350). In one study of 38 patients with mild CF who The diagnosis of ABPA is based on a constellation of both
had normal pulmonary function and chest radiographs, primary and secondary diagnostic clinical, laboratory,
for example, 17 (45%) proved to have abnormal HRCT and radiologic criteria. The acronym ARTEPICS has been
studies (349). suggested for the primary criteria, which include
The ability of CT compared with pulmonary function A (asthma), R (radiologic evidence of pulmonary infil-
tests to provide useful outcome measures has been trates), T (positive skin test for Aspergillus fumigatus),
reviewed (325,335,352). As previously discussed in depth, E (eosinophilia), P (precipitating antibodies to A. fumiga-
to date, a number of scoring systems have been suggested tus), I (elevated IgE), C (central bronchiectasis), and
(353–355). Based on an assessment of the degree and S (elevated A. fumigatus serum-specific IgE and IgG) (357).
extent of bronchiectasis, bronchial wall thickening, mucus A diagnosis of ABPA is nearly certain when six of these
plugging, atelectasis, and emphysema, Bhalla et al. (320) eight criteria are fulfilled. Secondary criteria include the
showed a statistically significant correlation between presence of A. fumigatus in sputum, a history of expectora-
CT scores and the percent ratios of FEV1/FVC (r 0.69, tion of mucus plugs, and delayed cutaneous reactivity to
p 0.006). In another study based on assessment of Aspergillus antigen (282,357).
bronchiectasis and mucus plugging (355), CT scores corre- Many of these findings are nonspecific: up to 10% of
lated highly with clinical (r 0.88, p 0.0001) and chest asthmatic patients, for example, have serum precipitating
radiographic (r 0.93, p 0.0001) scores, as well as pul- antibodies to A. fumigatus, and up to 25% demonstrate
monary function tests. More recently, de Jong et al. (325) cutaneous hypersensitivity (358). Similarly, although
retrospectively evaluated sequential CT scans in a cohort of peripheral blood eosinophilia and elevated levels of
116 patients aged 5 to 52 years and documented that, serum IgE are suggestive of the infection, they are hardly
although bronchiectasis worsened in 68 (74%) of 92 pathognomonic. Central bronchiectasis, thought essential
patients (with only single follow-up CT available), FEV1 to the diagnosis in most series, likely represents an
worsened in only 54 (59%) of 92 cases. Furthermore, CT advanced form of the disease resulting from chronic
evidence of worsening bronchiectasis was identified in inhalation and growth of A. fumigatus spores within the
27 patients with stable or improving pulmonary function, bronchial tree (358).
leading these investigators to speculate that CT and pul- Investigators have suggested that disease progression
monary function tests measure different aspects of CF lung be divided into five separate phases. These are as fol-
disease (325). Similar findings have been reported by lows: (a) an acute phase; (b) resolution, during which
Brody et al. (353), who showed that changes identified on time pulmonary infiltrates clear and serum IgE declines;
sequential HRCT studies correlated with clinical improve- (c) remission, when all diagnostic criteria recur; (d) a
ment, including improvement in peribronchial thickening phase of dependence on corticosteroids; and (e), finally, in
and mucus plugging. some cases, diffuse pulmonary fibrosis (357).
Although HRCT offers a reliable method for initially CT findings in patients with ABPA have been exten-
assessing disease as well as monitoring progression, not all sively reported (Figs. 5-71 and 5-75) (265,283,358–365).
studies have found CT useful in patients with more Neeld et al. (358), in a comparative study of 8 patients
advanced disease (354). It has also been suggested that fulfilling clinical criteria of ABPA and 7 asthmatic patients
HRCT cannot serve as a surrogate for exercise limitations. In evaluated with CT, identified bronchial dilatation in 19
one study of children with CF, CT findings were correlated (41%) of 46 lobes in the patients with ABPA, as compared
with functional measurements, including heart rate, oxygen with 15% in the patients with asthma. Moreover, in this
consumption, and time of exercise; no consistent relation- same study, patients with APBA tended to have more
ships could be established between lung function or exer- severe forms of bronchiectasis in comparison with
A B
Figure 5-75 Allergic bronchopulmonary aspergillosis (ABPA).

A, B: High-resolution sections through the mid lungs show
predominantly central bronchiectasis with relative sparing of the
lung periphery. This pattern, although nonspecific, is characteristic
of ABPA. C: Section in a different patient than in A and B shows
evidence of severe cystic bronchiectasis primarily involving the
lingula. In this case, discrete filling defects are identifiable within
cystic airways, a finding consistent with fungal superinfection
C (arrows in C).
asthmatic patients, in whom cylindric bronchiectasis only evidence of pleural abnormalities, especially focal pleural
could be identified. In this regard, CT may be valuable in thickening.
the early identification of lung damage in patients with The significance of central bronchiectasis as a specific
ABPA and thus may help in planning treatment (359). marker for ABPA is disputed (330,363). As noted by Neeld
Small airway abnormalities, with dilatation due to mucus et al. (358), central bronchiectasis is less likely a specific
or fluid-filled centrilobular bronchioles, can also be seen marker for ABPA as much as it is indicative of long-standing,
on HRCT, resulting in a tree-in-bud appearance (366). severe bronchial inflammation. Supportive data have been
Abnormalities of lung attenuation reflecting mosaic reported by Reiff et al. (330), who also found that although
perfusion and air trapping on expiratory CT scans can also central bronchiectasis was more widespread and severe than
be seen. seen in patients with idiopathic bronchiectasis, the finding
In addition to central bronchiectasis, APBA also results of central bronchiectasis by itself proved to have a sensitivity
in a variety of parenchymal and pleural abnormalities. As of only 37%. More specific is the finding of high-attenua-
reported by Panchal et al. (362), in their assessment of tion mucoid impaction (Fig. 5-71) (322, 367–369). First
23 patients with ABPA, in addition to central bronchiecta- described in association with chronic fungal sinusitis, high-
sis identified in all patients, parenchymal abnormalities density mucus presumably represents the presence of
could be identified in 43% of cases and included consoli- calcium or metallic ions within viscous mucus (370). The
dation, collapse, cavitation, bullae, and parenchymal prevalence of this finding has been noted to be as high as
scarring, all of which were most pronounced in the 28% in one series, and when it is present, it should be
upper lobes. An identical percentage of cases also had considered characteristic (367).
ABPA, a hypersensitivity reaction in which organisms sis of the etiology was made by three experienced chest
remain within the airway lumen, should be differentiated radiologists in only 45% of cases, whereas a high confi-
from the distinct entity of invasive aspergillosis (369). This dence level of diagnostic certainty was reached in only
latter entity is defined as either angioinvasive or airway 9% of cases, of which the correct diagnosis was made in
invasive and almost always occurs in immunosuppressed only 35%. Unfortunately, as in most studies assessing
persons. In the airway-invasive form of the disease, organ- diagnostic accuracy with CT, scans were interpreted with-
isms can be identified deep to the airway basement mem- out correlative clinical data (Fig. 5-76).
brane. Clinically, in distinction to the angioinvasive form More recently, Cartier et al. (333), in a retrospective
of disease, airway-invasive aspergillosis may occur even evaluation of 82 consecutive patients with bronchiectasis
when the degree of immunosuppression is mild. Radio- with an established etiology, reported that a correct diag-
logically, airway-invasive aspergillosis most often manifests nosis was established by two independent observers in
as patchy areas of parenchymal consolidation. On CT 61% of cases, including 68% of patients with CF, 67%
scans, most patients have been reported to have a distinct of patients with TB, and 56% of patients with ABPA.
peribronchial or peribronchiolar distribution of disease, Although these results represent an improvement over
ranging from patchy areas of consolidation to poorly those reported by others, the overall diagnostic sensitivity
defined centrilobular nodules (369). In one study, sequen- is not sufficiently high to warrant confident predictions in
tial CT studies showed that airway-invasive aspergillosis all save a small percentage of cases, specifically in this
resulted in bronchiectasis in patients without prior airway study those with an asymmetric upper lobe distribution
dilatation (369). suggestive of TB and a lower lobe distribution suggestive of
In addition to immunosuppression related to leuke- a history of childhood viral infections.
mia, bone marrow or renal transplantation, or other
immunosuppressive therapies, approximately 10% of
patients with AIDS may show evidence of aspergillosis of
the airways (274). Various descriptive terms have been
Small Airway Disease/Bronchiolitis
used, including invasive aspergillosis, necrotizing tra-
cheitis, obstructing bronchopulmonary aspergillosis, and
Clinicopathologic Classification
chronic cavitary parenchymal aspergillosis. Involvement By definition, bronchiolitis is a generic term used to
of the airways leads to certain abnormalities ranging from describe bronchiolar inflammation of various causes.
subtle nodular irregularity of airway walls to complete To date, several clinicopathologic classifications have been
airway obstruction. In the latter case, obstructing proposed, but unfortunately, none has gained widespread
bronchial aspergillosis has been reported to represent a acceptance (372,373). In part, this confusion is due to a
stage before frank tissue invasion and is characterized by bewildering array of redundant or even misleading
an acute onset of symptoms including fever, dyspnea, and descriptive terms that have been applied to this group of
cough associated with expectoration of bronchial casts diseases, in particular, bronchiolitis obliterans (BO).
laden with fungi (274). Further complicating this subject is the concept of small
airway disease. First proposed by Hogg et al. (374) to
describe inflammatory changes in peripheral airways of
Computed Tomographic Differentiation
smokers, resulting in moderate to severe airflow obstruc-
of Causes of Bronchiectasis
tion, this term was subsequently refined by Macklem et al.
The reliability of CT in distinguishing between various (375) to indicate an idiopathic syndrome of chronic
etiologies of bronchiectasis is controversial (330,333,371). airflow obstruction in patients without evidence of under-
In one study evaluating HRCT findings in 168 patients lying emphysema or chronic bronchitis. As such, the con-
with chronic sputum production, with the exception of cept of small airway disease is essentially physiologic,
lower lobe predominance in patients with syndromes with resulting from abnormalities involving airways between
impaired mucociliary clearance, no significant differences 2 and 3 mm in size. Although this category includes air-
could be identified in lobar distribution between cases of ways larger than bronchioles, the concept of small airway
idiopathic bronchiectasis and those with known etiologies disease has now acquired the status of broad usage as
(330). Although central bronchiectasis was more common a general synonym for bronchiolar disease.
in patients with ABPA, the diagnostic sensitivity of this Histologically, bronchiolitis is an inflammatory disor-
finding proved to be only 38% (330), as previously noted. der involving the bronchiolar wall, occurring as a reaction
Nor in this same study were estimates of the extent and to injury that is either focal or multifocal and affecting
severity of disease of diagnostic value. from a few to nearly all bronchioles (373). The injury can
Similar findings have been reported by Lee et al. (363), be due to a number of known causes, or can be of un-
who found in a study of 100 patients with a variety of known origin (Table 5-9). To date, classification of small
causes of bronchiectasis that a correct first-choice diagno- airway disease has focused either on clinical correlations
A B
C D
Figure 5-76 Bronchiectasis—differential diagnosis. A, B: Sections through the mid and lower lung fields show diffuse bronchiectasis (ar-
rows). Without clinical correlation, this appearance is nonspecific. In this case, bronchiectasis is due to immunosuppression in a patient with
AIDS. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations. Philadelphia: Lippincott
Williams & Wilkins; 2005, with permission.) C: Section through the lower lung fields in a different patient than shown in A and B. There is
evidence of diffuse bronchiectasis as well as findings consistent with small airway inflammation with bronchiolectasis and tree-in-bud
opacities in the lung periphery. D: Contrast-enhanced CT in the same patient as shown in C, imaged with narrow windows, shows marked
dilatation of the esophagus, in this case due to achalasia. In this case bronchiectasis was secondary to atypical mycobacterial infection, a
common complication in patients with achalasia.
or pathologic findings. As discussed, although both these Depending on the etiologic agent, there is either healing or
approaches have merit, it is also possible to construct a resorption and residual scarring. Chronic bronchiolitis is
radiologic-CT classification with considerably greater prac- typically associated with more prolonged injury and is
tical use. characterized by bronchiolar infiltration by mononuclear
Etiologic classifications of small airway disease typically cells followed by the development of a localized peribron-
also differentiate between acute or chronic bronchiolitis. chiolar fibrotic process (373). In addition, the small air-
Acute bronchiolitis characteristically results from processes ways may be involved indirectly by processes adjacent to
that cause bronchiolar injury over a short period of time bronchioles such as diseases characterized by diffuse
such as viral infection or more rarely inhalation of toxic lung fibrosis or peribronchiolar inflammation, such as
gases. Lesions mainly involve the bronchiolar epithelium, sarcoidosis, which can cause distortion of the bronchiole
resulting in necrosis and desquamation followed by and narrowing of the lumen (376).
exudation, fibrin deposition, and inflammatory cell infil- The current pathologic classification of bronchiolitis is
tration occasionally causing a granulomatous reaction. based on three main histologic patterns: cellular bronchi-
reaction in the more distal airspaces. In such cases,

bronchiolitis occurs in association with pneumonitis—
TABLE 5-9
hence the term bronchiolitis obliterans with organizing
ETIOLOGIES AND CLINICAL CONDITIONS
pneumonia (BOOP) (378). Reflecting not only the presence
ASSOCIATED WITH BRONCHIOLITIS of intraluminal fibrotic buds but also more extensive
Inhalation of gases, fumes, and dusts inflammatory changes involving alveolar ducts and alveoli,
Infection (e.g., viruses, Mycoplasma pneumoniae, Chlamydia BOOP characteristically results in predominantly restrictive
species, Aspergillus fumigatus) lung disease manifested radiographically as ill-defined areas
Irradiation
of parenchymal consolidation (379).
Aspiration
Drugs and chemicals Most often idiopathic, BOOP also may be associated
Organ transplantation with infection, toxic fume inhalation, or clinical syn-
Bone marrow dromes resulting in a wide range of diffuse interstitial lung
Heart-lung diseases, including idiopathic pulmonary fibrosis, connec-
Lung
tive tissue diseases, and vasculitides (372).
Connective tissue disease
Rheumatoid disease
Sjögren syndrome Obliterative (or Constrictive) Bronchiolitis
Systemic lupus erythematosus BO is characterized by the development of irreversible
Dermatomyositis circumferential submucosal fibrosis resulting in bronchiolar
Progressive systemic sclerosis
narrowing or obliteration in the absence of intraluminal
Others
Hypersensitivity pneumonitis granulation tissue polyps or surrounding parenchymal
Autoimmune diseases inflammation (Fig. 5-77D) (380). Fibrosis extends predom-
Chronic eosinophilic pneumonia inantly between the epithelium and the muscular mucosae
Neoplasia (carcinoid tumor, neuroendocrine cell hyperplasia) and along the long axis of the airway, impairing collateral
From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: ventilation and leading to airflow obstruction. The epithe-
functional and radiologic correlations. Philadelphia: Lippincott Williams & lium overlying the abnormal fibrous tissue, although
flattened or metaplastic, usually remains intact. Clinically,
constrictive bronchiolitis is associated with marked airflow
obstruction usually not responsive to steroid therapy (381)
olitis, BO with intraluminal polyps, and obliterative or Radiographically, little, if any, abnormality apart from mild
constrictive bronchiolitis (Fig. 5-77) (373,377). hyperinflation is apparent. Conditions associated with con-
strictive bronchiolitis include, most importantly, chronic
Cellular Bronchiolitis allograft rejection associated with heart-lung or lung trans-
Cellular bronchiolitis is characterized by inflammatory plantation and graft-versus-host disease associated with
cellular infiltrates involving both the bronchiolar lumen bone marrow transplantation. Other associated conditions
and walls, usually associated with some degree of fibrosis include collagen vascular diseases, drug toxicity, and child-
(Fig. 5-77B). According to the clinical presentation, acute hood infections (Swyer-James or Macleod syndrome).
or chronic, and the predominant cell type, classification Although histologic characterization clearly represents the
includes a number of specific etiologies, including: infec- gold standard for diagnosis, in most cases accurate diagnosis
tious bronchiolitis; respiratory bronchiolitis (RB); requires open (surgical) lung biopsy; transbronchial biopsies
follicular bronchiolitis; panbronchiolitis; aspiration result in insufficient tissue to allow specific diagnoses to be
bronchiolitis; bronchiolitis associated with hypersensitiv- made in the vast majority of cases. Worse, limited histologic
ity pneumonitis; and asthma. Cellular bronchiolitis may sampling may lead to erroneous diagnoses.
regress with treatment with antibiotic or anti-inflamma- With the introduction of HRCT, a method is now avail-
tory therapy or progress to one of the two following types able that allows routine classification of the various histo-
of fibrotic reaction. logic patterns of bronchiolitis (Table 5-10). The impact of
CT on assessing small airway disease is clearly one of the
Bronchiolitis Obliterans with Intraluminal Polyps more significant contributions of CT over the past decade
More appropriately referred to as cryptogenic organizing (293,376,382–388). As illustrated, this approach relies
pneumonia (COP) or proliferative bronchiolitis, this entity heavily on familiarity with pertinent secondary lobular
histologically is characterized by the presence of granulation anatomy (321).
tissue polyps or plugs of fibroblastic tissue extending from Anatomically, bronchioles are airways that lack cartilage
areas of epithelial damage into the lumens of respiratory (Fig. 5-77A). These include membranous bronchioles,
bronchioles, resulting in partial or occasionally complete which are purely air conducting, and respiratory bronchi-
obstruction (Fig. 5-77C). Although this pattern may be the oles, which contain alveoli within their walls. Along with
only abnormality present in the lungs, in most cases, it is their accompanying pulmonary artery branches, bronchi-
associated with similar epithelial injury and fibroblastic oles lie within the center of secondary pulmonary lobules
A B
C D
Figure 5-77 Bronchiolitis: histologic classification. See Color Figure 5-77A–D. A: High-power view of a normal bronchiole showing normal
wall thickness and normal surrounding alveoli. Note the absence of inflammatory cells or intraluminal abnormalities. B: High-power view in a
patient with cellular bronchiolitis. Inflammatory cells are noted infiltrating the bronchiolar wall, which is markedly thickened, as well as within
the bronchiolar lumen. The adjacent alveolar walls also appear slightly thickened, and the adjacent alveoli appear distended. The accompa-
nying arteriole (upper left corner) appears normal. C: High-power view in a patient with organizing pneumonia shows extensive granulation
tissue polyps within bronchioles (arrow) associated with organizing pneumonia (curved arrow). D: High-power view in a patient with obliter-
ative (constrictive) bronchiolitis shows extensive fibrosis surrounding the bronchiolar lumen, which is markedly narrowed. The adjacent alve-
oli are distended. Only scant inflammation is identified.
(Fig. 5-78). Although normal intralobular structures cannot trilobular lucencies, respectively (321,389). Indirect signs
be identified, direct and indirect signs of bronchiolar have also been described, the most important of which is
disease have been described (293,376,382–388). Direct the finding of focal air trapping, manifested either as areas
signs result from the presence of bronchiolar secretions, of mosaic attenuation on scans obtained in inspiration or
peribronchiolar inflammation, or, less commonly, bronchi- as areas of decreased attenuation, focal or global, on scans
olar wall thickening. Most characteristically, these changes obtained at end-expiration (see Fig. 5-86) (390).
result in branching or Y-shaped linear densities, as well as Using these signs, one can classify bronchiolitis into one
poorly defined centrilobular nodules or focal small cen- of four basic patterns (Table 5-10). These are as follows:
TABLE 5-10
BRONCHIOLAR DISEASE: COMPUTED TOMOGRAPHIC CLASSIFICATION
Bronchiolar diseases with tree-in-bud pattern BOOP associated with toxic fume inhalation
Common Collagen vascular diseases
Mycobacterium tuberculosis infection Prior infection
Atypical mycobacterial infections (MAC) Radiation therapy
Bacterial infections (especially in patients with cystic fibrosis, Uncommon
and HIV-positive and AIDS patients) BOOP associated with unrelated pathologic processes,
Asian panbronchiolitis including neoplasms, infectious granulomas, and vasculitides,
Uncommon or as a component of other diseases, including hypersensi-
Viral and fungal infections, including PCP and CMV tivity pneumonitis and Langerhans’ cell histiocytosis
Bronchiolar diseases with poorly defined centrilobular nodules Bronchiolar disease associated with decreased lung attenuation
Common Common
Subacute hypersensitivity pneumonitis Constrictive bronchiolitis after heart-lung and lung and/or
RB-ILD bone marrow transplantation
Uncommon Uncommon
LIP in AIDS patients Idiopathic
Follicular bronchiolitis After childhood infections
Mineral dust–induced bronchiolitis Associated with collagen vascular diseases
Sarcoidosis Toxic fume inhalation
After consumption of Sauropus androgynus
Bronchiolar disease associated with ground-glass attenuation
Constrictive bronchiolitis associated with neuroendocrine
and/or consolidation
hyperplasia
Common
Sarcoidosis
Idiopathic BOOP
MAC, Mycobaterium avium-intracellulare complex; HIV, human immunodeficiency virus; AIDS, acquired immunodeficiency syndrome; PCP, Pneumo-
cystis carinii pneumonia; CMV, cytomegaloviral infection; RB-ILD, respiratory bronchiolitis with interstitial lung disease; LIP, lymphocytic interstitial
pneumonitis; BOOP, bronchiolitis obliterans organizing pneumonia.
(a) acute cellular bronchiolitis almost invariably caused by associated with intraluminal polyps and adjacent pneu-
infection, resulting in a tree-in-bud appearance (Figs. 5-2, monitis, resulting in diffuse, ground-glass attenuation or
5-3, 5-76, 5-79, 5-80); (b) subacute and chronic cellular consolidation (Figs. 5-83 and 5-84); and (d) bronchiolar
bronchiolitis, resulting in poorly defined centrilobular diseases associated with submucosal fibrosis, resulting in
opacities (Figs. 5-81 and 5-82); (c) bronchiolar diseases a pattern of diffuse mosaic attenuation (Figs. 5-4, 5-85 to
5-87). This classification has the virtue of being com-
plementary to the pathologic classification outlined
previously and allows easy differentiation between the
most important types of bronchiolitis most often clinically
encountered (Fig. 5-77). It is not exceptional for the correct
diagnosis in these cases to first be suggested on the basis of
CT findings.
Computed Tomography Classification
Bronchiolar Diseases Associated with

a Tree-in-Bud Pattern
The hallmark of this group of diseases is the finding of
dilated, mucus-filled bronchioles resulting in a pattern of
centrilobular nodular, branching, or Y-shaped densities that
has been aptly referred to as simulating a tree in bud. First
described by Akira et al. (391) in association with diffuse
Figure 5-78 Normal bronchogram of the right bronchial tree panbronchiolitis (Fig. 5-88) and by Im et al. (267) in
targeted on the right upper lobe. The arrows show the normal patients with endobronchial spread of TB (Figs. 5-2, 5-73,
appearance of lobular and terminal bronchioles. (From Naidich DP, 5-74, 5-80), this pattern, in fact, is almost always the result
Webb WR, Grenier PA, et al. Imaging of the airways: functional and
radiologic correlations. Philadelphia: Lippincott Williams & Wilkins; of acute or subacute infectious bronchiolitis (195). This
2005, with permission.) includes bacterial, viral, and fungal causes and pertains
Figure 5-79 Acute cellular bron-

chiolitis—tree-in-bud opacities. A,
B: Target reconstructed images
through the right lung show charac-
teristic appearance of clusters of
small ill-defined nodules adjacent to
peripheral vessels, some with a
distinctive branching configuration
(arrows). Note that these opacities
rarely extend to the pleural surface,
indicative of their centrilobular
A, B distribution.
to both immunocompetent and immunocompromised the lung periphery. Less commonly, the “buds” reflect the
patients. presence of peribronchiolar granulomas, a finding espe-
As documented by Im et al. (267), tree-in-bud opacities cially common in patients with chronic infection due to
on CT correlate pathologically with the presence of inspis- mycobacterial infections (195). Other manifestations of
sated secretions within dilated terminal and respiratory acute infection, including areas of ground-glass attenuation
bronchioles. These bronchioles characteristically are cen- or consolidation, may also be present. Similarly, one can
trilobular in distribution and are most easily identified in also usually identify poorly defined centrilobular nodules
Figure 5-80 Acute cellular bron-

chiolitis—tuberculosis. A, B: Target
reconstructed 1-mm images through
the right upper lobe show evidence
of typical tree-in-bud opacities
(arrow in A), as well as mild bronchial
dilatation and mucoid impaction of
larger peripheral airways (arrow in B)
in this patient with documented spu-
tum-positive tuberculosis (compare
A, B with Figs. 5-73 and 5-74).
A B
Figure 5-81 Chronic cellular bronchiolitis—hypersensitivity pneumonitis. A: High-resolution 1-mm image through the upper lobes in a
patient with subacute hypersensitivity pneumonitis shows typical pattern of poorly defined centrilobular nodules diffusely present through-
out both lungs. B: Target reconstructed view in the same patient shows poorly defined centrilobular nodules to better advantage.
without evidence of branching, the result of sections secondary lobular emphysema may be identified, presum-
through distended centrilobular bronchioles seen in cross ably the result of bronchial obstruction.
section (Fig. 5-89). These findings, however, are almost In Asia, especially in Japan and Korea, a form of diffuse
always ancillary to the main finding of a tree-in-bud pattern. bronchiolitis termed diffuse panbronchiolitis is common
With healing, a distinct tree-in-bud pattern associated with (Fig. 5-88) (391–394). This disorder has been defined as
Figure 5-82 Chronic cellular

bronchiolitis— follicular bronchiolitis.
A, B: Target reconstructions through
the left lung show multiple poorly
defined centrilobular nodules, most
associated with dilated peripheral
airways (arrows). This diagnosis was
subsequently confirmed by open
A, B lung biopsy (not shown).
A B
Figure 5-83 Airspace consolidation—cryptogenic organizing pneumonia. Bilateral areas of airspace consolidation having both peripheral
and peribronchovascular predominant distribution. A: Thin-section CT at the level of the upper lobes and superior segments of the lower
lobes. B: Thin-section CT scan at the level of the lower lobes. (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: func-
tional and radiologic correlations. Philadelphia: Lippincott Williams & Wilkins; 2005, with permission.)
A B
C
Figure 5-84 Bronchiolar disease associated with intraluminal polyps—bronchiolitis obliterans organizing pneumonia (BOOP)/cryptogenic
organizing pneumonia (COP). See Color Figure 5-84B,C. A: Section through the lower lobes shows evidence of ill-defined foci of nodular
consolidation and patchy ground-glass attenuation. This appearance is nonspecific and almost always requires open lung biopsy for defini-
tive diagnosis. B, C: High-power hematoxylin and eosin and trichrome sections, respectively, obtained from an open lung biopsy show char-
acteristic patchy appearance of intraluminal polyps composed of immature fibroblastic tissue (so-called Masson bodies) typical of COP.
A B
Figure 5-85 Mosaic perfusion pattern at multidetector computed tomography (MDCT). Postinfectious bronchiolitis. Thin-collimation
MDCT acquisition. A: Thin-section CT at the level of the lower part of the lungs showing patchy areas of hypoattenuation in the lower lobes
and the right middle lobe. B: Lateral reformation after thin-collimation multidetector row CT acquisition using the multiplanar volume refor-
mation and minimum-intensity projection techniques in combination. Left (sagittal reformation of the right lung) and right (sagittal reforma-
tion of the left lung) images show well-demarcated areas of hypoattenuation and decreased perfusion, reflecting the territories where the
lesions of obliterative bronchiolitis are situated. Notice the presence of bronchiectasis in the apicoposterior segment of the left upper lobe
(arrow). (From Naidich DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional and radiologic correlations. Philadelphia:
A B
Figure 5-86 Constrictive bronchiolitis. A, B: 1-mm sections through the upper and lower lobes, respectively, in a 25-year-old woman
presenting with increasingly severe dyspnea and a normal chest radiograph (not shown). Evidence of mosaic attenuation is identifiable as
geometric areas of apparent ground-glass attenuation admixed with areas of relatively low parenchymal attenuation. In this case, the size of
the vessels does not vary from region to region; in addition, there is evidence of diffuse bronchiectasis. Open lung biopsy revealed only
scant peribronchiolar inflammation without evidence of intraluminal granulation tissue or surrounding airspace or interstitial inflammation.
These findings are characteristic of constrictive bronchiolitis.
Figure 5-87 Obliterative (constrictive) bronchiolitis—Swyer-

James syndrome. High-resolution 1-mm image through the middle
lung zone shows hyperinflation of the lingula within which visualized
vessels appear attenuated. There is also evidence of both central
and peripheral bronchiectasis with a few tree-in-bud opacities also
seen. By history this patient suffered from a childhood, presumably,
viral pneumonia. Although the Swyer-James syndrome ostensibly
causes unilateral hyperinflation, on CT there is almost always evi-
dence of diffuse, although asymmetric airway disease.
537
survival rates reported to be in the range of 60% and 30%,

respectively (393).
Not surprisingly, the tree-in-bud pattern is frequently
found in association with other signs of proximal airway
infection. As documented by Acquino et al. (195), in a
prospective study of 27 cases with a tree-in-bud pattern,
26 had evidence of associated bronchiectasis or proximal
airway wall thickening (Fig. 5-78). Diagnoses in this study
included CF, postinfectious bronchiectasis, ABPA, TB, P.
carinii pneumonia, and bronchopneumonia, among oth-
ers (195). Most important, when compared with a control
population of 141 patients with other documented airway
diseases, including emphysema, RB, constrictive bronchi-
Figure 5-88 Diffuse panbronchiolitis. A 1-mm section shows olitis, BOOP, and extrinsic allergic alveolitis, in none of
diffuse ill-defined nodular and branching or Y-shaped densities
throughout both lower lobes in this patient with documented these cases could a tree-in-bud pattern be identified. In
diffuse panbronchiolitis. distinction, in this same control population, 25% of
patients with bronchiectasis and 17% of patients with
acute infections bronchitis or pneumonia proved to have
a clinicopathologic entity characterized by symptoms of this pattern (195). These findings reinforce the concept
chronic cough, sputum, and dyspnea associated with that a tree-in-bud pattern should be interpreted as a mani-
abnormal pulmonary function tests usually indicating festation of infectious bronchiolitis.
mild to moderate airway obstruction and radiographic
evidence of ill-defined nodular infiltrates typically basilar Bronchiolar Diseases Associated with Poorly
in distribution (393). Histologically, diffuse panbronchi- Defined Centrilobular Nodules
olitis is characterized by chronic inflammation with In distinction to infectious bronchiolitis, the hallmark of
mononuclear cell proliferation and foamy macrophages this group of diseases is the finding of ill-defined centrilob-
predominantly involving the walls of respiratory bronchi- ular nodules in the absence of a tree-in-bud appearance.
oles, adjacent alveolar ducts, and alveoli, constituting the Pathologically, the finding of centrilobular nodules
so-called unit lesion of PB (393). Inclusion of this entity in correlates with the finding of peribronchiolar inflamma-
the category of infectious causes of bronchiolitis seems tion without evidence of bronchiolar dilatation or retained
justified because nearly all patients develop superinfection secretions (321,389). Evidence of proximal bronchial
with Pseudomonas. On CT, a characteristic diffuse tree- inflammation is also lacking. In our opinion, this pattern is
in-bud pattern is invariably seen, predominantly affecting sufficiently distinctive to warrant separate classification.
the lung bases, that has been shown to respond to treat- Within this category are a diverse set of diseases that
ment with erythromycin (395). Unfortunately, despite have in common the finding of peribronchiolar inflam-
initial improvement, diffuse panbronchiolitis is usually mation in the absence of widespread parenchymal
considered a progressive disease, with 5- and 10-year consolidation (Table 5-10). This group includes subacute
hypersensitivity pneumonitis (Fig. 5-81) (396,397),
lymphocytic interstitial pneumonitis (LIP) (340),
follicular bronchiolitis (Fig. 5-82) (398), respiratory
bronchiolitis–interstitial lung disease (RB-ILD) (Fig. 5-
83) (399–401), and mineral dust–induced bronchiolitis,
among others. In most cases, differential diagnosis in
this group is simplified by detailed clinical correlation,
including careful occupational and environmental expo-
sure histories.
Characteristic of HRCT findings in this diverse group are
those seen in patients with subacute hypersensitivity
pneumonitis (Fig. 5-81). Hypersensitivity pneumonitis is
associated with a chronic, nonspecific inflammation that,
early in its course, is primarily distributed around
Figure 5-89 Acute tuberculosis. Thin-section CT at the level of respiratory bronchioles with relative sparing of intervening
the upper part of the lungs showing bilateral patchy areas of small lung (396,397,402,403). This pattern may persist even later
centrilobular and linear branching opacities (tree-in-bud sign). (From in the course of disease. Additionally, nearly two thirds of
functional and radiologic correlations. Philadelphia: Lippincott patients also have evidence of non-necrotizing granulomas,
Williams & Wilkins; 2005, with permission.) again typically localized to the peribronchiolar interstitium.
Foci of BOOP may also be identified, in some cases, with accumulation of pigmented foamy macrophages pre-
characteristic findings of intraluminal fibrous plugs. dominantly within and around the walls of respiratory
Together, these findings result in a pattern of diffuse, poorly bronchioles and alveolar ducts, with or without intralu-
defined centrilobular nodular opacities, especially in the minal mucus plugs (399–401,407). When RB is the only
subacute phase of disease. cause of disease in patients with clinical evidence of
Similar CT findings have been identified in patients interstitial lung disease, it is then referred to as RB-
with LIP, especially in AIDS patients. Characterized by an associated interstitial lung disease. Radiographically, up
interstitial infiltrate of mature lymphocytes, LIP is part of to three fourths of patients show evidence of reticular or
the spectrum of hyperplasia of bronchus-associated reticulonodular infiltrates (399–401,407). CT findings in
lymphoid tissue that also includes follicular bronchiolitis smokers with RB have been described and include
(404). In this latter disease, lymphoid aggregates contain- parenchymal and subpleural micronodules, predomi-
ing reactive germinal centers can be identified within the nantly within the upper lobes, associated with areas of
walls of bronchioles representing a localized form of ground-glass attenuation resulting from filling of
lymphoid hyperplasia. Although characterized by more airspaces with pigmented macrophages and bronchial
diffuse interstitial involvement, in fact, histologic differen- dilatation (Fig. 5-90) (399–401,407). The finding of a
tiation between follicular bronchiolitis and LIP may be tree-in-bud pattern is distinctly unusual (195).
difficult, a finding raising the possibility that the poorly
defined centrilobular nodules identified on CT in most Bronchiolar Diseases Associated with Focal
cases of LIP diagnosed by transbronchial biopsy actually Ground-Glass Attenuation or Consolidation
represent follicular bronchiolitis (Fig. 5-82). Follicular This pattern of presentation is characteristic of BOOP
bronchiolitis has also been described in association with (340,388,408–410). As previously noted, BOOP is charac-
rheumatoid arthritis (RA) (405). In one report in which terized histologically by the presence of granulation tissue
CT findings showed evidence both of branching or polyps within respiratory bronchioles and alveolar ducts
Y-shaped densities and poorly defined centrilobular (Masson bodies) associated with patchy organizing pneu-
densities in patients with RA with documented follicular monia (Fig. 5-77). Investigators have suggested that the
bronchiolitis, in all cases the disease was either stabilized term COP is preferable to BOOP because the predominant
or reversed after treatment with erythromycin, suggesting abnormality clinically, functionally, and radiologically is
that the finding of a tree-in-bud pattern may reflect super- the organizing pneumonia and because involvement of
imposed infection (406). the bronchioles may be absent in up to 30% of cases
Also representative of this CT category is RB–RB-ILD (295). However, BOOP is too well established in the litera-
(Fig. 5-90). A form of reversible bronchiolitis most often ture to be replaced. For the remainder of this chapter, the
incidentally identified in young smokers, this entity is two terms will be used interchangeably.
now classified as part of the spectrum of smoking related Two distinct patterns of intraluminal fibrosis or Masson
disorders, including—in order of worsening disease—RB, bodies have been described (411). Type 1 Masson bodies
RB-ILD, and desquamative interstitial pneumonitis are those that show abundant myxoid matrix, sparse
(DIP). RB is characterized histologically by the patchy fibrosis, and little or no fibrin and presumably represent
A B
Figure 5-90 Respiratory bronchiolitis. A, B: 1-mm sections through the trachea and carina, respectively, show poorly defined
centrilobular nodules (arrows) non-uniformly distributed throughout both lungs. Note the absence of reticular densities, architectural distor-
tion, or tubular or branching structures (compare with Fig. 5-88).
immature mesenchymal cells. In distinction, type 2 Masson nonsegmental, unilateral or bilateral foci of airspace
bodies contain fibrin and have histochemical properties consolidation (385); however, in a smaller number of
of myofibroblasts. This distinction may be important cases, both focal nodules as well as irregular predomi-
because, at least in one study, patients with type 1 disease nantly basilar reticular densities have been described.
showed good response to steroid therapy, whereas Infiltrates may be migratory. Honeycombing is rare,
those with type 2 disease generally failed to respond to although as noted by King (419), this rarity may reflect
treatment (411). the relative insensitivity of radiographs versus CT scans for
Although most cases of BOOP are idiopathic, similar detecting subtle areas of parenchymal involvement.
findings may be seen in association with many different Although most patients with ill-defined areas of patchy
clinical presentations (Fig. 5-84). As described by airspace consolidation respond to treatment with corti-
Katzenstein (222), BOOP may be classified into three sep- costeroids, the response in patients with interstitial
arate categories. In the first, BOOP is the primary cause of infiltrates has been noted to be worse.
respiratory illness. This category includes idiopathic BOOP Several studies have reviewed the CT and HRCT findings
as well as BOOP resulting from RA (412), toxic inhalants, in patients with BOOP (340,388,409,410,420,421). The
drug toxicity (413,414), and the entire spectrum of colla- most common abnormality consists of patchy bilateral con-
gen vascular disease (415), as well as BOOP in response to solidation, seen in approximately 80% of cases, frequently
prior infection, both viral and bacterial (Fig. 5-91). Also with a predominantly peribronchial and subpleural
included in this category is BOOP in association with distribution (Figs. 5-83 and 5-84). Bronchial wall thicken-
acute radiation pneumonitis (416). In the second category ing and dilatation are commonly present in the areas with
are cases in which BOOP is found as a nonspecific reaction consolidation. Although small (1 to 10 mm), ill-defined,
along the periphery of unrelated pathologic processes, predominantly peribronchial or peribronchiolar nodules
including neoplasms, infectious granulomas, vasculitides, are seen in 30% to 50% of cases, the finding of a tree-in-bud
and even infarcts. Finally, BOOP may also be identified pattern is distinctly unusual (384). Irregular linear opacities
as a minor component of other diseases including hyper- are present in approximately 10% of cases.
sensitivity pneumonitis, nonspecific interstitial pneumoni- Areas of ground-glass attenuation may be seen in up to
tis, and Langerhans cell histiocytosis (222). For this 60% of immunocompetent patients with BOOP, but they
reason, the finding of features consistent with BOOP on are seldom the predominant abnormality in these patients
transbronchial biopsy is of only limited value in the (Fig. 5-84). Areas of ground-glass attenuation are seen
absence of detailed clinical and radiologic correlation. more commonly in immunocompromised patients with
Clinically, patients with idiopathic BOOP usually pre- BOOP and may be the predominant or only abnormality
sent with a 1- to 3-month history of nonproductive seen in these patients (421). Small nodules are seen more
cough, low-grade fever, and increasing shortness of breath commonly in immunocompromised patients; nodules
(409,410,417). Various radiographic patterns have been 1 to 10 mm in diameter are observed in 6 of 11 (55%) of
described (418). Most often, radiographs show patchy, immunocompromised patients as compared with 7 of
32 (22%) of immunocompetent patients with idiopathic
BOOP reported by Lee et al. (421).
In a few cases, BOOP may present either as multiple
large (1 to 5 cm) nodules or masses (422) or less com-
monly as a solitary, poorly defined peripheral lesion. In the
latter case, definitive diagnosis usually requires open lung
biopsy (Fig. 5-91) (423). In a report by Akira et al. (422),
12 (20%) of 50 patients with BOOP presented with multi-
ple nodules or masses as the predominant finding. Of a
total of 60 lesions, 88% proved to have irregular margins,
whereas 45% were associated with air bronchograms, and
38% had pleural tags. The finding of an irregular masslike
area of consolidation adjacent to pleural surfaces, in partic-
ular, proved suggestive. Similar findings have also been
reported by Bouchardy et al. (424), who also noted nodu-
lar or masslike opacities to be a frequent finding, occurring
in 5 (42%) of 12 patients. Recently, it has been noted that
Figure 5-91 Postinfectious bronchiolitis obliterans organizing
pneumonia (BOOP). Thin-section CT showing multiple bilateral BOOP may result in a distinctive curvilinear ringlike band
rounded areas of airspace consolidation having either a subpleural of airspace consolidation. Dubbed the “reverse halo” or
or a peribronchovascular distribution. (From Naidich DP, Webb “atoll” sign, this finding typically involves the subpleural
WR, Grenier PA, et al. Imaging of the airways: functional and radio-
logic correlations. Philadelphia: Lippincott Williams & Wilkins; portions of the lower lobes and, when present, is highly
2005, with permission.) suggestive of the diagnosis (Fig. 5-92).
A, B
Figure 5-92 Reverse halo sign—bronchiolitis obliterans organizing pneumonia (BOOP). A, B: High-resolution sections through the lung
bases show characteristic curvilinear subpleural airspace consolidation within which hazy ground-glass density can be identified, resulting in
the so-called reverse halo sign.
Bronchiolar Diseases Associated with Decreased Also known as the Swyer-James syndrome, constrictive
Lung Attenuation bronchiolitis frequently occurs as the sequela of childhood
In this category are patients with obliterative or constric- respiratory tract infections (Fig. 5-87) (431), most often
tive bronchiolitis. Constrictive bronchiolitis is defined viral, although similar findings may also be seen in children
histologically by the presence of concentric fibrosis involv- after infection with Mycoplasma. Obliterative bronchiolitis
ing the submucosal and peribronchial tissues of terminal has also been linked to consumption of Sauropus androgynus
and respiratory bronchioles exclusively, with resulting (432,433), a small, lowland shrub found in Asia and con-
bronchial narrowing or obliteration (Fig. 5-77). This pro- sumed in the form of an uncooked juice as a means of
cess is typically nonuniform, and because the surrounding weight reduction, especially in Taiwan. A CT pattern of
parenchyma is normal, the condition may be difficult to mosaic attenuation has also been reported in association
identify even on open lung biopsy (379,417). Clinically, with neuroendocrine hyperplasia, especially in patients
although these patients may be relatively asymptomatic, in with carcinoid tumors (434,435).
most cases progressive airway obstruction results in severe
respiratory compromise that is usually unresponsive to Obliterative Bronchiolitis Syndrome. Of all causes of con-
steroid therapy. strictive bronchiolitis, most important is the association
Unlike BOOP, which usually proves to be idiopathic, between constrictive bronchiolitis and heart-lung or lung
obliterative bronchiolitis is only rarely idiopathic. Condi- transplants (Fig. 5-93) (291,294,436–438). In this setting,
tions associated with constrictive bronchiolitis include constrictive bronchiolitis results from lymphocyte-mediated
heart-lung or lung transplantations, chronic allograft rejec- chronic rejection and is a major cause of both morbidity and
tion, allogeneic bone marrow transplantation with chronic mortality. The primary risk factor for post-transplantation
graft-versus-host disease, and collagen vascular diseases, constrictive bronchiolitis appears to be the frequency and
especially RA (290,292,294,425,426). severity of acute cellular rejection that nearly always
The association between RA and, to a lesser degree, occurs in patients in the early postoperative setting (439).
Sjögren syndrome and bronchiolitis warrants special men- Investigators have estimated that as many as 50% of patients
tion. Obliterative bronchiolitis is frequently identified in receiving heart-lung or lung transplants develop constrictive
patients with RA, occasionally exhibiting rapid progression bronchiolitis, typically within 6 months to a year after organ
(427,428). In fact, RA is associated with a wide range of transplantation, and between 25% and 40% of these pati-
abnormalities affecting both large and small airways. ents will die as a direct result. Although the term bronchi-
As shown by Perez et al. (429) in a prospective evaluation olitis obliterans syndrome has been proposed to describe
of 50 patients with documented RA, HRCT demonstrated this occurrence, as previously discussed, in our opinion, it is
bronchial or parenchymal abnormalities in 70% of cases, preferable to refer to this entity simply as posttransplanta-
including air trapping in 32%, cylindrical bronchiectasis tion constrictive bronchiolitis (439).
in 30%, and centrilobular opacities in 6%. Although less In distinction to the recipients of heart-lung and lung
common, obliterative bronchiolitis can also be identified transplants, constrictive bronchiolitis is far less common in
in patients with Sjögren syndrome, in one study occurring patients after bone marrow transplantation, although it is
in 54% of 35 patients (430). still problematic. In this setting, constrictive bronchiolitis is
characteristic HRCT findings consist of mosaic attenua-

tion, air trapping, and bronchial dilatation (Figs. 5-93
and 5-94) (385,387). Mosaic attenuation results from
decreased perfusion of areas distal to bronchiolar obstruc-
tion with blood flow redistribution to noninvolved lung.
Air trapping resulting from partial airway obstruction is
best seen on expiratory HRCT scans (293,384–386,441).
To date, because of the low yield of transbronchial
biopsy in the diagnosis of constrictive bronchiolitis, the
diagnosis of posttransplantation constrictive bronchiolitis
has been based on a constellation of clinical, radiologic,
and primarily physiologic findings, as follows: a decline of
20% in forced expiratory flow at midlung volumes
(FEV25–75) from peak values obtained after transplantation
and in the absence of definable pathogens on bron-
Figure 5-93 Constrictive bronchiolitis associated with lung
transplantation. A 1-mm section in a 48-year-old man with bronchi- choalveolar lavage, in association with a normal radi-
olitis obliterans after double lung transplantation demonstrates ographic appearance of the transplanted lung (442).
markedly dilated bronchi in both lower lobes and in the right Although mosaic perfusion is commonly seen in patients
middle lobe. Also note the patchy areas of decreased attenuation
and perfusion. with obliterative bronchiolitis, unfortunately, this feature
is of limited value in the diagnosis because it is also com-
mon in healthy control subjects and in posttransplant
a manifestation of chronic graft-versus-host disease patients without evidence of constrictive bronchiolitis
(425,440) and it has been reported in 2% to 13% of (436). The two most helpful findings in the diagnosis are
patients who underwent allogeneic bone marrow trans- the presence of bronchial dilatation, particularly in the
plantation. Clinically, it may develop any time after the lower lobes, and the presence of air trapping on expiratory
third month after bone marrow transplantation. HRCT. Worthy and Flower (436) reviewed the HRCT scans
The chest radiograph in patients with constrictive bron- in 15 patients with pathologically proven BO after lung
chiolitis may be normal, or it may show nonspecific transplantation and in 18 control subjects. Findings seen
abnormalities, including variable degrees of hyperinfla- in patients with constrictive bronchiolitis included
tion, peripheral attenuation of the vascular markings, and bronchial dilatation in 80%, mosaic perfusion in 40%,
evidence of central airway dilatation (385,387). The bronchial wall thickening in 27%, and air trapping in 80%
A B
Figure 5-94 Bronchial dehiscence following lung transplantation. A, B: Sections through the right upper lobe imaged with lung and
mediastinal windows, respectively, following right lung transplantation show focal ulceration extending from the anterior aspect of the
proximal right upper lobe bronchus (arrow in B). In this case, there is also evidence of a small right-sided pneumothorax. Findings were con-
firmed bronchoscopically. (Case courtesy of Dr. Timothy Harkin, Mt. Sinai Medical Center, New York, New York.)
of cases, compared with the control subjects, who had

bronchial dilatation in 22% of cases, mosaic perfusion in
22% of cases, bronchial wall thickening in 0% of cases,
and air trapping in 6% of cases (436).
The role of repeated HRCT scans in monitoring the
development of BO syndrome after lung transplantation
was also assessed by Ikonen et al. (149). In a study of
13 lung transplant recipients who underwent a total of 126
HRCT scans during a mean follow-up period of 23 months,
8 of the 13 patients developed BO syndrome. The authors of
this study demonstrated that the HRCT findings coincided
with the development of the BO syndrome. The overall sen-
sitivity of HRCT for the diagnosis of BO was 93%, and the
specificity was 92% (149).
Figure 5-95 Severe persistent asthma. Axial thin-section CT

ASTHMA scan at full inspiration. Diffuse decreased lung attenuation and
bronchial wall thickening are noted in both lungs. (From Naidich
Asthma is a chronic inflammatory condition characterized DP, Webb WR, Grenier PA, et al. Imaging of the airways: functional
and radiologic correlations. Philadelphia: Lippincott Williams &
by airway inflammation due to a variety of stimuli resulting Wilkins; 2005 with permission.)
in bronchial hyperresponsiveness (442a). Chronic inflam-
mation results in structural changes, including smooth
muscle thickening, neovascularity, and fibrosis, leading to
irreversible airway narrowing (443). Bronchiectasis may HEMOPTYSIS
also result from chronic inflammation likely associated
with recurrent airway infection. Hemoptysis represents the ideal subject for evaluating
The clinical indications for CT in patients with asthma correlations between bronchoscopy and CT. Hemoptysis is
are generally limited and include identification of subtle a common manifestation of both pulmonary and extra-
radiographically occult airspace infection or treatable causes pulmonary diseases (Table 5-11). Reported causes of
of airway inflammation, including ABPA and hypersensi- hemoptysis have varied during the past 50 years, depend-
tivity pneumonitis, for example (444). CT is also of use in ing on changing disease prevalence and on the addition of
selected cases for documenting emphysema in smokers new diagnostic techniques such as FB and chest CT. Before
with asthma (445). CT may also rarely be of value for iden- 1970, TB, bronchiectasis, and bronchogenic carcinoma
tifying structural causes of localized wheezing, as may occur were the most common causes of hemoptysis (450–452).
with focal tracheal or mainstem bronchial lesions. High- With the advent of FB in 1970, most studies concluded
resolution CT findings in patients with asthma have been that chronic bronchitis and bronchogenic carcinoma had
described and cover nearly the entire gamut of large and emerged as the predominant causes of hemoptysis
small airway abnormalities (315,446–448). As reported by (453,454). Indeed, by the late 1980s, TB and bronchiecta-
Park et al. (447), however, only three findings are identified sis were considered unlikely causes of routine hemoptysis,
with greater frequency in asthmatic patients than in normal although they were still significant in the differential
individuals: bronchial wall thickening, bronchial dilatation, diagnosis of massive hemoptysis.
and expiratory air trapping. Reports of the prevalence of In the late 1980s, HRCT of the chest became the
bronchial wall thickening vary widely, ranging between standard radiologic technique for detecting airway disease,
16% and 92% of cases (315). Bronchial dilatation is gener- particularly bronchiectasis (455,456). Use of HRCT during
ally mild and likely reflects in part a reduction in the diame- this period demonstrated that bronchiectasis again was
ter of adjacent pulmonary arteries due to hypoxia. Similar responsible for a significant portion of cases of otherwise
to bronchial wall thickening, the frequency with which unexplained cases of hemoptysis (2,3,266,457). At pres-
bronchial dilatation is identified varies widely, ranging from ent, numerous studies have validated the usefulness of CT
15% to 77% of cases, likely due to differences in CT criteria in evaluating patients with hemoptysis (Table 5-12)
employed (315,447–449). Abnormal expiratory air trapping (2,3,79,266,457–461). In particular, CT has proved of
has been observed in 50% of asthmatic patients over the value in establishing the etiology in patients in whom no
course of the disease (447). Focal and diffuse areas of de- prior diagnosis could be established by FB, in particular,
creased lung attenuation seen in 20% to 30% of asthmatic patients with radiographically occult peripheral cancers
patients are likely due to a combination of air trapping and and bronchiectasis. Millar et al. (266) demonstrated that
pulmonary oligemia secondary to alveolar hypoventilation routine CT provided a diagnosis in 20 of 40 patients
(Fig. 5-95) (315,448). (50%) with hemoptysis in whom both chest radiographs
TABLE 5-11
CONDITIONS ASSOCIATED WITH HEMOPTYSIS
Common Uncommon Rare
Neoplasia Alveolar hemorrhage syndromes Cardiac disease (mitral stenosis/heart failure)

Bronchogenic carcinoma, carcinoid tumors Goodpasture syndrome Arteriovenous malformations
Acute infection Vasculitis (Wegener granulomatosis) Hematologic disorders
Active tuberculosis Broncholithiasis
Bacterial and fungal pneumonia Foreign body aspiration
Lung abscess
Chronic airway infection and inflammation
Bronchitis
Bronchiectasis
Cystic fibrosis
Mycetomas in pre-existing cavities
Trauma
Pulmonary infarction
Iatrogenic
Lung biopsy (transbronchial/percutaneous)
Anticoagulation
Pulmonary artery catheterization
Sarcoidosis
Cryptogenic
From Naidich DP, Webb WR, Grenier PA et al. Imaging of the airways: functional and radiologic correlations. Philadelphia: Lippincott Williams &
and FB failed to yield a diagnosis. Set et al. (457) com- with hemoptysis evaluated both with CT and FB, found
pared the utility of FB and HRCT in 91 patients and found that CT identified all cancers. Furthermore, the overall
that CT detected all 27 tumors seen bronchoscopically diagnostic yield of bronchoscopy was documented to be
as well as an additional 7, 5 of which were beyond bron- less than that of CT (47% compared with 61%, respec-
choscopic range. Similarly, McGuinness et al. (3), in a tively). It is worth noting that in addition to identifying
prospective study of 57 consecutive patients presenting the cause and often location of bleeding, CT also plays a
TABLE 5-12
HEMOPTYSIS: COMPUTED TOMOGRAPHY–FLEXIBLE BRONCHOSCOPY CORRELATIONS
Hirshberg McGuinness Set Naidich Millar Haponik
et al.a,b (1997) et al. (1994) et al. (1993) et al. (1990) et al. (1992) et al. (1987)
(n 208) (n 57) (n 91) (n 58) (n 40) (n 32)
Bronchiectasis (%) 20 25 15 17 18 0
Tuberculosis (%) 1.4 16 0 2 5 0
Lung cancer (%) 19 12 37 33 5 0
Fungal infection (%) 1.4 12 0 2 0 6
Cryptogenic (%) 8 19 34 33 50 37
Bronchitis (%) 18 7 5 0 0 10
Miscellaneous (%) 33 5 9 10 22 3
Multiple causes (%) 0 4 —c 0 —d 0
aReferences: Hirshberg B, Biran I, Glazer M, Kramer MR. Hemoptysis: etiology, evaluation, and outcome in a tertiary referral hospital. Chest. 1997;
112:440–444; McGuinness G, Beacher JR, Harkin TJ, et al. Hemoptysis: prospective high-resolution CT/bronchoscopic correlation. Chest. 1994;105:
1155–1162; Set PAK, Flower CDR, Smith IE, et al. Hemoptysis: comparative study of the role of CT and fiberoptic bronchoscopy. Radiology. 1993;
189:677–680; Naidich DP, Funt S, Ettenger NA, Arranda C. Hemoptysis: CT-bronchoscopic correlations in 58 cases. Radiology. 1990;177:357–362;
Millar A, Boothroyd A, Edwards D, Hetzel M. The role of computed tomography (CT) in the investigation of unexplained hemoptysis. Respir Med.
1992;86:39–44; Haponik EF, et al. Computed tomography in the evaluation of hemoptysis: impact on diagnosis and treatment. Chest. 1987;91:80–85.
Gudjberg CE. Roentgenologic diagnosis of bronchiectasis: an analysis of 112 cases. Acta Radiol. 1955;43:209–226.
bCT performed in 129 (62%) of 208 cases.
c Two cases of lung cancer and bronchiectasis.
dOne case of peripheral neoplasm and bronchiectasis.
Adapted from Hirshberg B, Biran I, Glazer M, Kramer MR. Hemoptysis: etiology, evaluation, and outcome in a tertiary referral hospital. Chest.
1997;112:440–444, with permission.
potentially important role by serving as a roadmap for 81% of 53 patients with hemoptysis. Although the site of
bronchoscopists prior to performing TBNA or other inter- bleeding was localized in 5 patients, no specific diagnoses
ventional bronchoscopic procedures. were made, and carcinoma was not detected in any patient
Although most patients are currently evaluated with during the procedure or at follow-up.
some combination of CT and bronchoscopy, especially The value of bronchoscopy for establishing a diagnosis
those with recurrent hemoptysis, the sequence with which of lung cancer in patients with normal or nonlocalizing
these studies are performed remains somewhat controver- chest radiographs, in particular, is controversial (467,
sial (462–464). Tak et al. (464), for example, in a recent 470,472). Adelman et al. (473), evaluating the clinical sig-
study of 50 patients presenting with hemoptysis and a nificance of a negative endoscopic examination in patients
normal or nonlocalizing radiograph reported that, at risk for lung cancer, reviewed the clinical outcome of
although a definitive diagnosis could be established in 67 patients over a 3-year period with cryptogenic hemopty-
17 (34% of cases), the etiology was established by HRCT sis after nondiagnostic bronchoscopy. Although 9 patients
in 30% of cases compared with only 10% by bron- subsequently died of causes unrelated to thoracic disease,
choscopy. Importantly, CT identified all patients with only one patient was subsequently diagnosed with lung
focal airway pathology established by FB. Similar findings cancer 20 months after bronchoscopy, without recurrence
have been reported, even in cases of severe or massive of hemoptysis; the remaining 85% remained well.
hemoptysis (461). It should be noted that a slightly higher incidence of
Although these findings have led these investigators to lung cancer developing in patients initially assessed as
suggest that CT should be the initial method for evaluating normal has been reported. In a retrospective evaluation
patients with hemoptysis, others disagree. Lenner et al. of follow-up data on 115 patients with hemoptysis of
(462), for example, in a recent review of the diagnosis and unknown origin initially evaluated with both chest
management of patients presenting with hemoptysis, has radiography and bronchoscopy, lung cancer developed in
concluded that following an initial chest radiograph, bron- 7 (6%); although all occurred in smokers older than the
choscopy should remain the next step in the diagnostic age of 40 years, unfortunately, all proved unresectable at
algorithm—this despite acknowledging that the overall the time of presentation (474). Although these data might
diagnostic accuracy of CT exceeds bronchoscopy and that, suggest a potential role for CT surveillance, to date, there is
furthermore, a preliminary CT might be of value by direct- little evidence to suggest that CT is of value for early
ing the bronchoscopist to areas of abnormality for the detection of endobronchial lesions in particular.
purposes of both diagnosis and staging. More recently, in an assessment of the cost-effectiveness
In fact, the role of FB in the evaluation of patients of thoracic CT and bronchoscopy in patients referred for
presenting with hemoptysis is controversial (465). FB is of evaluation of hemoptysis, Law et al. (463) evaluated
proven efficacy in evaluating patients with central endo- 56 patients, of whom 39 had CT examinations and 42
bronchial disease, with definite diagnoses made in more were evaluated by FB (with the others excluded for a vari-
than 95% of endoscopically visible primary malignant ety of reasons, including obvious diagnoses by history and
diseases (466). In addition to identifying central endo- physical findings, poor patient condition, and lack of
bronchial disease, bronchoscopy has also proven valuable patient compliance) and found that in none of these cases
in the diagnosis of localizing bleeding sites before surgery, did either procedure result in a diagnosis of lung cancer.
removal of blood clots that may cause obstruction, and Furthermore, in comparing the incidence of carcinoma
removal of foreign bodies. Despite this efficacy, the overall prior to and following 1992, a time when widespread use
diagnostic accuracy of FB in patients presenting with of bronchoscopy and CT combined became available, the
hemoptysis in most reported series is surprisingly low, incidence of lung cancer was noted to decrease, leading
especially in patients presenting with normal or nonlocal- these investigators to conclude that both bronchoscopy
izing chest radiographs (467–470). and CT are overutilized.
In one review of 48 consecutive patients presenting Although the value of CT for assessing patients with
with hemoptysis, FB resulted in a diagnosis other than hemoptysis has been established, it remains unclear if
endobronchial inflammation in only 4 patients, only 2 of there is any role for CT in the management of patients
whom proved to have lung cancer (469). Similarly, Peters presenting with massive hemoptysis. Traditionally, patient
et al. (468), in a retrospective evaluation of 113 patients management has been guided by the quantity of bleeding,
with hemoptysis, found that a specific cause was identified with massive hemoptysis usually defined as 600 mL per
in none of the 26 patients with normal radiographs, 24 hours, reflecting the observation that oxygen transfer is
compared with 35 of 87 patients (40%) in whom impaired when approximately 400 mL of blood fills the
corresponding chest radiographs proved localizing. Poe alveolar space (4). Bleeding may originate from several
et al. (471), in a study of 196 patients with normal or non- sources within the lungs, including the bronchial and
localizing chest radiographs, found that in only 33 cases pulmonary arteries as well as the pulmonary capillary
(17%) were specific causes established bronchoscopically. bed. In 90% of cases, massive hemoptysis originates from
Sharma et al. (470) reported normal FB examinations in high-pressure bronchial arteries; however, as noted by
Corder, this accounts for only 1.5% of all patients 3. McGuinness G, Beacher JR, Harkin TJ. Hemoptysis: prospective
high-resolution CT/bronchoscopic correlation. Chest. 1994;105:
presenting with hemoptysis.
1155–1162.
Recently, the role of CT in assessing patients with mas- 4. Corder (4), R. Hemoptysis. Emerg Med Clin North Am. 2003;21:
sive hemoptysis has been re-evaluated. In one retrospective 421–435.
5. McLoud TC. Imaging techniques for diagnosis and staging of
study comparing the results of CT versus bronchoscopy in
lung cancer. Clin Chest Med. 2002;23:123.
patients with severe hemoptysis, CT proved superior to FB 6. Molina PL, Hiken JN, Glazer HS. Imaging evaluation of obstruc-
both for identifying the site of bleeding (72% vs. 39%, tive atelectasis. J Thorac Imaging. 1996;11:176–186.
7. Hennebicque AS, et al. CT findings in severe thoracic sarcoido-
respectively; p 0.005) and the etiology (92% vs. 32%
sis. Eur Radiol. 2005;15:23–30.
respectively; p 0.0001) (461). Similarly, Revel et al. 8. Walser EM. Stent placement for tracheobronchial disease. Eur
(475), in a retrospective study of 80 patients admitted to J Radiol. 2005;55:321–330.
9. Beamis JF. Interventional pulmonology techniques for treating
the intensive care unit with massive hemoptysis, noted
malignant large airway obstruction: an update. Curr Opin
that CT proved more effective than bronchoscopy for iden- Pulmon Med. 2005;11:292–295.
tifying the cause of bleeding (77% vs. 8%, respectively). In 10. Noppen M, et al. Interventional bronchoscopy for treatment of
tracheal obstruction secondary to benign or malignant thyroid
distinction, CT and FB proved equally effective in localiz-
disease. Chest. 2004;125:723–730.
ing the site of bleeding (70% vs. 73%, respectively). 11. Falcone F, Fois F, Grosso D. Endobronchial ultrasound.
Clearly, in patients in whom bleeding constitutes a Respiration. 2003;70:179–194.
12. Hautmann H, et al. Electromagnetic catheter navigation during
medical emergency, immediate control of hemorrhage via
bronchoscopy—validation of a novel method by conventional
rigid bronchoscopy remains the procedure of choice. In fluoroscopy. Chest. 2005;128:382–387.
patients for whom arterial embolization is an option, the 13. Boiselle PM. Multislice helical CT of the central airways. Radiol
Clin North Am. 2003;41:561.
use of CT angiographic techniques, in particular, both MIPs
14. Grenier PA, et al. Multidetector-row CT of the airways. Semin
and 3D volume-rendered images, may prove invaluable Roentgenol. 2003;38:146–157.
in select cases by depicting the nonbronchial systemic 15. Colice GL. Chest CT for known or suspected lung cancer. Chest.
1994;106:1538–1550.
origin of hemoptysis as well as identifying the ectopic
16. Zhang JB, et al. Dynamic expiratory volumetric CT imaging of
origin of bronchial arteries, facilitating their successful the central airways: comparison of standard-dose and low-dose
catheterization (476). techniques. Acad Radiol. 2003;10:719–724.
17. Boiselle PM, et al. Multiplanar and 3D imaging of the central
In summary, in the setting of hemoptysis, CT and
airways: comparison of image quality and radiation dose of sin-
bronchoscopy should be considered complementary pro- gle-detector row CT and multi-detector row CT at differing tube
cedures. HRCT cannot replace bronchoscopy, especially currents in dogs. Radiology. 2003;228:107–111.
18. Yi CA, et al. Multidetector CT of bronchiectasis: effect of radiation
in its ability to provide biopsy specimens for histologic or
dose on image quality. AJR Am J Roentgenol. 2003;181:501–505.
cytologic examination. In turn, bronchoscopy is limited 19. Leswick DA, et al. Radiation cost of helical high-resolution chest
in detecting peribronchial abnormalities, as well as in CT. AJR Am J Roentgenol. 2005;184:742–745.
20. Remy-Jardin M, et al. Diagnosis of bronchiectasis with multi-
diagnosing parenchymal disease, especially bronchiecta-
slice spiral CT: accuracy of 3-mm-thick structured sections. Eur
sis. In our experience, CT should be obtained in most Radiol. 2003;13:1165–1171.
patients prior to bronchoscopy. In patients with normal 21. Giron J, Skaff F, Maubon A. The value of thin-section CT scans
in the diagnosis and staging of bronchiectasis: comparison with
or nonlocalizing chest radiographs, CT is of value by
bronchography in a series of fifty-four patients. Ann Radiol.
detecting otherwise radiographically occult tumors, in- 1988;31:25–33.
cluding both subtle endobronchial lesions as well as oc- 22. Vikgren J, et al. Value of air trapping in detection of small air-
ways disease in smokers. Acta Radiol. 2003;44:517–524.
cult peripheral lung tumors. CT is especially valuable in
23. Zhang J, et al. Frequency and severity of air trapping at dynamic
patients with peripheral tumors by providing a roadmap expiratory CT in patients with tracheobronchomalacia. AJR Am
to lesions beyond the limit of bronchoscopic visualiza- J Roentgenol. 2004;182:81–85.
24. Tanaka N, et al. Air trapping at CT: High prevalence in asympto-
tion. In addition, in those cases in which TBNA may be of
matic subjects with normal pulmonary function. Radiology.
value, CT may play an indispensable role by assisting in 2003;227:776–785.
the selection of optimal biopsy sites. Similarly, in pa- 25. Dorlochter L, et al. High resolution CT in cystic fibrosis—the
contribution of expiratory scans. Eur J Radiol. 2003;47:193–198.
tients in whom therapeutic intervention may be indi-
26. Stern EJ, Webb WR, Gamsu G. Dynamic quantitative computed
cated, CT can assist in the selection of appropriate candi- tomography: a predictor of pulmonary function in obstructive
dates as well as optimizing technique. Finally, CT is also lung diseases. Invest Radiol. 1994;29:564–569.
27. Stern EJ, et al. Normal trachea during forced expiration:
indispensable for diagnosing bronchiectasis as well as
dynamic CT measurements. Radiology. 1993;187:27–31.
other causes of diffuse lung disease associated with he- 28. Bankier AA, et al. Bronchial wall thickness: appropriate window
moptysis, including sarcoidosis. settings for thin-section CT and radiologic-anatomic correla-
tion. Radiology. 1996;199:831–836.
29. De Wever W, Bogaert J, Verschakelen JA. Virtual bronchoscopy:
accuracy and usefulness—an overview. Semin Ultrasound CT
REFERENCES MRI. 2005;26:364–373.
30. Finkelstein SE, et al. Virtual bronchoscopy for evaluation of
malignant tumors of the thorax. J Thorac Cardiovasc Surg.
1. Boiselle PM, Ernst A. State-of-the-art imaging of the central air-
2002;123:967–972.
ways. Respiration. 2003;70:383–394.
31. Ravenel JG, McAdams HP. Multiplanar and three-dimensional
2. Naidich DP, et al. Hemoptysis: CT-bronchoscopic correlations
imaging of the thorax. Radiol Clin North Am. 2003;41:475.
in 58 cases. Radiology. 1990;177:357–362.
32. Ferretti GR, Bricault I, Coulomb M. Virtual tools for imaging of 60. Kim JS, Choi D, Lee KS. CT of the bronchus intermedius: fre-
the thorax. Eur Respir J. 2001;18:381–392. quency and cause of a nodule in the posterior wall on normal
33. Quint LE, et al. Stenosis of the central airways: evaluation by scans. AJR Am J Roentgenol. 1995;165:1349–1352.
using helical CT with multiplanar reconstructions. Radiology. 61. Zylak CJ, et al. Developmental lung anomalies in the adult:
1995;194:871–877. Radiologic-pathologic correlation. Radiographics. 2002;22:
34. Salvolini L, et al. Clinical applications of 2D and 3D CT imag- S25–S43.
ing of the airways—a review. Eur J Radiol. 2000;34:9–25. 62. Ghaye B, et al. Congenital bronchial anomalies. Radiographics.
35. McAdams HP, et al. Bronchial anastomotic complications in 2001;21:105–119.
lung transplant recipients: virtual bronchoscopy for noninva- 63. Wu JW, et al. Variant bronchial anatomy: CT appearance and
sive assessment. Radiology. 1998;209:689–695. classification. AJR Am J Roentgenol. 1999;172:741–744.
36. De Wever W, et al. Multidetector CT-generated virtual bron- 64. Morrison SC. Demonstration of a tracheal bronchus by com-
choscopy: an illustrated review of the potential clinical indica- puted tomography. Clin Radiol. 1988;39:208–209.
tions. Eur Respir J. 2004;23:776–782. 65. Freeman SJ, Harvey JE, Goddard PR. Demonstration of super-
37. Ferretti GR, et al. Follow-up after stent insertion in the tracheo- numerary tracheal bronchus by computed tomographic
bronchial tree: role of helical computed tomography in scanning and magnetic resonance imaging. Thorax. 1995;50:
comparison with fiberoptic bronchoscopy. Eur Radiol. 2003;13: 426–427.
1172–1178. 66. O’Sullivan BP, Frassica JJ, Rayder SM. Tracheal bronchus: a cause
38. Finkelstein SE, et al. Comparative evaluation of super high- of prolonged atelectasis in intubated children. Chest. 1998;113:
resolution CT scan and virtual bronchoscopy for the detec- 537–540.
tion of tracheobronchial malignancies. Chest. 2003;124: 67. Shipley RT, et al. Computed tomography of the tracheal
1834–1840. bronchus. J Comput Assist Tomogr. 1985;9:53–55.
39. Haliloglu M, et al. CT virtual bronchoscopy in the evaluation of 68. Landing BH, Dixon LG. Congenital malformation and genetic
children with suspected foreign body aspiration. Eur J Radiol. disorders of the respiratory tract. Am Rev Respir Dis. 1979;120:
2003;48:188–192. 151–185.
40. Hoppe H, et al. Multidetector CT virtual bronchoscopy to grade 69. Karcaaltincaba M, Haliloglu M, Ekinci S. Partial tracheal dupli-
tracheobronchial stenosis. AJR Am J Roentgenol. 2002;178: cation: MDCT bronchoscopic diagnosis. AJR Am J Roentgenol.
1195–1195. 2003;183:290–292.
41. Hoppe H, et al. Grading airway stenosis down to the segmental 70. Moller GM, ten Berge EJFM, Stassen CM. Tracheocele: a rare
level using virtual bronchoscopy. Chest. 2004;125:704–711. cause of difficult endotracheal intubation and subsequent
42. Lacasse Y, et al. Accuracy of virtual bronchoscopy to detect pneumomediastinum. Eur Respir J. 1994;7:1376–1377.
endobronchial lesions. Ann Thorac Surg. 2004;77:1774–1780. 71. Tanaka H, et al. Paratracheal air cysts communicating with the
43. Kiraly AP, et al. Three-dimensional path planning for virtual trachea: CT findings. J Thorac Imaging. 1997;12:38–40.
bronchoscopy. IEEE Trans Med Imaging. 2004;23:1365–1379. 72. Mangiula VG, Razvzn VS. The accessory cardiac bronchus; bron-
44. Shinagawa N, et al. CT-guided transbronchial biopsy using an chologic aspect and review of the literature. Chest. 1968;54:
ultrathin bronchoscope with virtual bronchoscopic navigation. 35–38.
Chest. 2004;125:1138–1143. 73. McGuinness G, et al. Accessory cardiac bronchus: computed
45. Gamsu G, Webb WR. Computed tomography of the trachea: nor- tomographic features and clinical significance. Radiology.
mal and abnormal. AJR Am J Roentgenol. 1982;139:3321–326. 1993;189:563–566.
46. Gamsu G, Webb WR. Computed tomography of the trachea and 74. Keane MP, et al. Accessory cardiac bronchus presenting with
mainstem bronchi. Semin Roentgenol. 1983;18:51–60. haemoptysis. Thorax. 1997;52:490–491.
47. Webb EM, Elicker BM, Webb WR. Using CT to diagnose non- 75. Ghaye B, Kos X, Dondelinger RF. Accessory cardiac bronchus:
neoplastic tracheal abnormalities: appearance of the tracheal 3D CT demonstration in nine cases. Eur Radiol. 1999;9:45–48.
wall. AJR Am J Roentgenol. 2000;174:1315–1321. 76. Bentala M, et al. Cardiac bronchus: a rare cause of hemoptysis.
48. Zwischenberger JB, Sankar AB. Surgery of the thoracic trachea. Case report. Eur J Cardiothorac Surg. 2002;22:643–645.
J Thorac Imaging. 1995;10:199–205. 77. Chan AL, et al. Advances in the management of endobronchial
49. Breatnach E, Abbott GC, Fraser RC. Dimensions of the normal lung malignancies. Curr Opin Pulmon Med. 2003;9:301–308.
human trachea. AJR Am J Roentgenol. 1984;141:903–906. 78. Ernst A, et al. Central airway obstruction. State-of-the-art. Am
50. Griscom NT. Computed tomographic determination of tracheal J Respir Crit Care Med. 2004;169:1278–1297.
dimensions in children and adolescents. Radiology. 1982;145: 79. Naidich DP, Harkin TJ. Airways and lung: correlation of CT with
361–364. fiberoptic bronchoscopy. Radiology. 1995;197:1–12.
51. Effman EL, et al. Tracheal cross-sectional area in children: CT 80. Shure D, Fedullo PF. Transbronchial needle aspiration of
determination. Radiology. 1983;149:137–140. peripheral masses. Am Rev Respir Dis. 1983;128:1090–1092.
52. Weibel ER, Taylor CR. Design and structure of the human lung. 81. Minami H, et al. Interbronchoscopist variability in the diagno-
In Fishman AP, ed. Pulmonary diseases and disorders. New York: sis of lung cancer by flexible bronchoscopy. Chest. 1994;105:
McGraw-Hill; 1988:11–60. 1658–1662.
53. Murata K, et al. Optimization of computed tomography tech- 82. Naidich SP, et al. Comparison of CT and fiberoptic bron-
nique to demonstrate the fine structure of the lung. Invest choscopy in the evaluation of bronchial disease. AJR Am
Radiol. 1988;23:170–175. J Roentgenol. 1987;148:1–7.
54. Webb WR, et al. Normal and diseased isolated lungs: high- 83. Colice G, et al. Comparison of computed tomography with
resolution CT. Radiology. 1988;166:81–87. fiberoptic bronchoscopy in identifying endobronchial abnor-
55. Foster WL, et al. Localized peribronchial thickening: A CT sign malities in patients with suspected lung cancer. Am Rev Respir
of occult bronchogenic carcinoma. AJR Am J Roentgenol. Dis. 1985;131:397–400.
1985;144:906–908. 84. Conces DJ, Tarver RD, Vix VA. Broncholithiasis: CT features in
56. Jackson CL, Huber JF. Correlated applied anatomy of the 15 patients. AJR Am J Roentgenol. 1991;157:249–253.
bronchial tree and lungs with a system of nomenclature. Dis 85. Seo JB, et al. Broncholithiasis: review of the causes with radio-
Chest. 1943;9:319–326. logic-pathologic correlation. Radiographics. 2002;22:S199–S213.
57. Boyden EA. The nomenclature of the bronchopulmonary 86. Kim TS, et al. Endobronchial actinomycosis associated with
segments and their blood supply. Dis Chest. 1961;39:1–6. broncholithiasis: CT findings for nine patients. Am J Roentgenol.
58. Ferry RM, Boyden EA. Variations in the bronchovascular 2005;185:347–353.
patterns of the right lower lobe in fifty lungs. J Thorac Cardiovasc 87. Glazer HS, Anderson DJ, Sagel SS. Bronchial impaction in lobar
Surg. 1951;22:188–201. collapse: CT demonstration and pathologic correlation. AJR Am
59. Pitel M, Boyden EA. Variations in the bronchovascular patterns J Roentgenol. 1989;153:485–488.
of the left lower lobe in fifty lungs. J Thorac Cardiovasc Surg. 88. Woodring JH, Reed JC. Radiographic manifestations of lobar
1953;26:633–653. atelectasis. J Thorac Imaging. 1996;11:109–144.
89. Gelb AF, et al. Computed tomography and bronchoscopy in 112. Baba M, et al. Correlation between endobronchial ultrasonog-
chest radiographically occult mainstem neoplasm: diagnosis raphy (EBUS) images and histologic findings in normal and
and Nd-YAG laser treatment in 8 patients. West J Med. 1990;153: tumor-invaded bronchial wall. Lung Cancer. 2002;35:65–71.
385–389. 113. Larsen SS, et al. Endoscopic ultrasound guided biopsy of medi-
90. George PJM, et al. Bronchography in the assessment of patients astinal lesions has a major impact on patient management.
with lung collapse for endoscopic laser therapy. Thorax. 1990; Thorax. 2002;57:98–103.
45:503–508. 114. Miyazu Y, et al. Endobronchial ultrasonography in the assess-
91. Usuda K, et al. Relation between bronchoscopic findings ment of centrally located early-stage lung cancer before photo-
and tumor size of roentgenographically occult bronchogenic dynamic therapy. Am J Respir Crit Care Med. 2002;165:832–837.
squamous cell carcinoma. J Thorac Cardiovasc Surg. 1993;106: 115. Penman I, Fergusson RJ. Endoscopic ultrasound: a useful tool
1098–1103. to assess the mediastinum in patients with lung cancer? Thorax.
92. Naidich DP, et al. Bronchoscopic evaluation of the airways with 2002;57:95–96.
CT correlation. In: Imaging of the airways. Philadelphia: Lippincott, 116. Fritscher Ravens A. Endoscopic ultrasound evaluation in the
Williams & Wilkins; 2005:29–70. diagnosis and staging of lung cancer. Lung Cancer. 2003;41:
92a. Schenk DA, Bower JH, Bryan CL. Transbronchial needle aspira- 259–267.
tion staging of bronchogenic carcinoma. Am Rev Respir Dis. 117. Herth FJ, Becker HD, Ernst A. Ultrasound-guided trans-
1986;134:146–148. bronchial needle aspiration—an experience in 242 patients.
93. Harrow EM, et al. The utility of transbronchial needle aspiration Chest. 2003;123:604–607.
in the staging of bronchogenic carcinoma. Am J Respir Crit Care 118. Herth FJF. Mediastinal staging—the rote of endobronchial and
Med. 2000;161:601–607. endo-oesophageal sonographic guided needle aspiration. Lung
94. Sharafkhaneh A, et al. Yield of transbronchial needle aspiration Cancer. 2004;45:S63–S67.
in diagnosis of mediastinal lesions. Chest. 2003;124:2131–2135. 119. Caddy G, et al. The accuracy of EUS-FNA in assessing mediasti-
95. Khoo KL, et al. Transbronchial needle aspiration: initial experi- nal lymphadenopathy and staging patients with NSCLC. Eur
ence in routine diagnostic bronchoscopy. Respir Med. 2003;97: Respir J. 2005;25:410–415.
1200–1204. 120. Eloubeidi MA, et al. Endoscopic ultrasound-guided fine needle
96. Cetinkaya E, et al. Diagnostic value of transbronchial needle aspiration of mediastinal lymph node in patients with sus-
aspiration by Wang 22-gauge cytology needle in intrathoracic pected lung cancer after positron emission tomography and
lymphadenopathy. Chest. 2004;125:527–531. computed tomography scans. Ann Thorac Surg. 2005;79:
96a. Hermens FHW, et al. Diagnostic yield of transbronchial histol- 263–268.
ogy needle aspiration in patients with mediastinal lymph node 120a. Herth FJF, Ernst A, Becker HD. Endobronchial ultrasound-guided
enlargement. Respiration. 2003;70:631–635. transbronchial lung biopsy in solitary pulmonary nodules and
97. Morales CF, et al. Flexible transbronchial needle aspiration in peripheral lesions. Eur Respir J. 2002;20:972–974.
the diagnosis of sarcoidosis. Chest. 1994;106:709–711. 121. Kikuchi E, et al. Endobronchial ultrasonography with guide-
98. Harkin TJ, et al. Transbronchial needle aspiration in the diagno- sheath for peripheral pulmonary lesions. Eur Respir J. 2004;24:
sis of mediastinal mycobacterial infection [Abstract]. Am Rev 533–537.
Respir Dis. 1993;147:A801. 122. Paone G, et al. Endobronchial ultrasound-driven biopsy in the
99. Trisolini R, et al. The value of flexible transbronchial needle diagnosis of peripheral lung lesions. Chest. 2005;128:3551–3557.
aspiration in the diagnosis of stage I sarcoidosis. Chest. 123. Takahashi H, et al. A prospective evaluation of transbronchial
2003;124:2126–2130. ultrasonography for assessment of depth of invasion in early
100. Pastore SM, et al. Intrathoracic adenopathy associated with bronchogenic squamous cell carcinoma. Lung Cancer. 2003;42:
pulmonary tuberculosis in patients with human immunodefi- 43–49.
ciency virus infection. Chest. 1993;103:1433–1437. 124. Herth F, et al. Endobronchial ultrasound reliably differentiates
101. Prakash UBS, Offord KP, Stubbs SE. Bronchoscopy in North between airway infiltration and compression by tumor. Chest.
America: the ACCP survey. Chest. 1991;100:1668–1675. 2003;123:458–462.
102. Rong F, Cui B. CT scan directed transbronchial needle aspiration 125. Becker H, Herth F. Is endobronchial ultrasound indispensable
biopsy for mediastinal nodes. Chest. 1998;114:36–39. in clinical practice? Pro: endobronchial ultrasound. J Bronchol.
103. Goldberg SN, et al. Mediastinal lymph node biopsy: diagnostic 2002;9:145–151.
yield of transbronchial mediastinal lymph node biopsy with CT 126. LeBlanc JK, et al. Endoscopic ultrasound in non-small cell lung
fluoroscopic guidance—initial experience. Radiology. 2000;216: cancer and negative mediastinum on computed tomography.
764–767. Am J Respir Crit Care Med. 2005;171:177–182.
104. White CS, et al. Transbronchial needle aspiration—guidance 127. Pass HI. Mediastinal staging 2005: pictures, scopes, and
with CT fluoroscopy. Chest. 2000;118:1630–1638. scalpels. Semin Oncol. 2005;32:269–278.
105. Bricault I, Ferretti G, Cinquin P. Registration of real and CT- 128. Rintoul RC, et al. Endobronchial and endoscopic ultrasound-
derived virtual bronchoscopic images to assist transbronchial guided real-time fine-needle aspiration for mediastinal staging.
biopsy. IEEE Trans Med Imaging. 1998;17:703–714. Eur Respir J. 2005;25:416–421.
106. Sherbondy AJ, et al. Virtual bronchoscopic approach for com- 129. Yasufuku K, et al. Endobronchial ultrasound guided trans-
bining 3-D CT and endoscopic video. In: Medical Imaging bronchial needle aspiration for staging of lung cancer. Lung
2000: Physiology and Function from Multidimensional Images. Proc. Cancer. 2005;50:347–354.
SPIE. 2000;3978:104–116. 130. Miyazu Y, et al. Endobronchial ultrasonography in the diagnosis
107. Mori K, Deguchi D, Sugiyama Y, et al. Tracking of a bron- and treatment of relapsing polychondritis with tracheo-
choscope using epipolar geometry analysis and intensity- bronchial malacia. Chest. 2003;124:2393–2395.
based image registration of real and virtual endoscopic images. 131. Annema JT, Veselic M, Rabe KF. Endoscopic ultrasound-guided
Med Image Anal. 2002;6:321–326. fine-needle aspiration for the diagnosis of sarcoidosis. Eur
108. Crawford SW, Colt HG. Virtual reality and written assessments Respir J. 2005;25:405–409.
are of potential value to determine knowledge and skill in flexi- 132. Herth F, Beck R, Ernst A. Conventional vs. endobronchial ultra-
ble bronchoscopy. Respiration. 2003;71:269–275. sound-guided transbronchial needle aspiration: a randomized
109. Moorthy K, et al. Evaluation of virtual reality bronchoscopy as a trial. Chest. 2004;125:322–325.
learning and assessment tool. Respiration. 2003;70:195–199. 132a. Miyazawa T, et al. Stenting at the flow-limiting segment in tra-
110. Hopper KD, et al. Transbronchial biopsy with virtual CT cheobronchial stenosis due to lung cancer. Am J Respir Crit Care
bronchoscopy and nodal highlighting. Radiology. 2001;221: Med. 2004;169:1096–1102.
531–536. 133. Berkmen YM. The trachea: the blind spot in the chest. Radiol
111. Tanaka F, et al. Evaluation of tracheo-bronchial wall invasion Clin North Am. 1984;22:539–562.
using transbronchial ultrasonography (EBUS). Eur J Cardio 134. Karmy-Jones R, Jurkovich GJ. Blunt chest trauma. Curr Prob
Thorac Surg. 2000;17:570–574. Surg. 2004;41:223–380.
135. Rossbach MM, et al. Management of major tracheobronchial in- 160. Hazama K, et al. Clinicopathological investigation of 20 cases
juries: a 28-year experience. Ann Thorac Surg. 1998;65:182–186. of primary tracheal cancer. Eur J Cardio Thor Surg. 2003;
135a. Baumgartner F, et al. Tracheal and main bronchial disruptions 23:1–5.
after blunt chest trauma: presentation and management. Ann 161. McCarthy MJ, Rosado-de-Christenson ML. Tumors of the tra-
Thorac Surg. 1990;50:569–574. chea. J Thorac Imaging. 1995;10:1810–198.
135b. Kirsh MM, et al. Management of tracheobronchial disruption 162. Grillo HC, Mathisen DJ, Wain JC. Management of tumors of the
secondary to nonpenetrating trauma. Ann Thorac Surg. trachea. Oncology. 1992;6:61–67.
1976;22:93–101. 163. Kwon JW, et al. Mucous gland adenoma of the bronchus: CT
136. Carbognani P, et al. Management of postintubation membra- findings in two patients. J Comput Assist Tomogr. 1999;23:
nous tracheal rupture. Ann Thorac Surg. 2004;77:406–409. 758–760.
137. Unger JM, et al. Tears of the trachea and main bronchi caused 164. Kim TS, et al. Sialadenoid tumors of the respiratory tract:
by blunt trauma: radiologic findings. AJR Am J Roentgenol. radiologic-pathologic correlation. AJR Am J Roentgenol. 2001;
1989;153:1175–1180. 177: 1145–1150.
138. Chen JD, et al. Using CT to diagnose tracheal rupture. AJR Am 165. Maziak DE, et al. Adenoid cystic carcinoma of the airway: thirty-
J Roentgenol. 2001;176:1273–1280. two-year experience. J Thorac Cardiovasc Surg. 1996;112:
139. Wintermark M, Schnyder P, Wicky S. Blunt traumatic rupture of 1522–1531.
a mainstem bronchus: spiral CT demonstration of the “fallen 166. Shepard JO. The bronchi: imaging perspective. J Thorac Imaging.
lung” sign. Eur Radiol. 2001;11:409–411. 1995;10:236–254.
140. Rollins RJ, Tocino I. Early radiographic signs of tracheal rup- 167. Rosado de Christenson ML, et al. Thoracic carcinoids: radiologic-
ture. AJR Am J Roentgenol. 1987;148:695–698. pathologic correlation. Radiographics. 1999;19:707–736.
141. Palder SB, Shandling B, Manson D. Rupture of the thoracic tra- 168. Naidich DP. CT/MR correlation in the evaluation of tracheo-
chea following blunt trauma: diagnosis by CAT scan. J Pediatr bronchial neoplasia. Radiol Clin North Am. 1990;28:555–571.
Surg. 1991;26:1320–1322. 169. Jeung M-Y, et al. Bronchial carcinoid tumors of the thorax: spec-
142. Lupetin AR. Computed tomographic evaluation of laryngotra- trum of radiologic findings. Radiographics. 2002;22:351–365.
cheal trauma. Curr Probl Diagn Radiol. 1997;26:185–206. 170. Zwiebel BR, Austin JHM, Grines MM. Bronchial carcinoid
143. Balci AE, et al. Surgical treatment of post-traumatic tracheo- tumors: assessment with CT of location and intratumoral calci-
bronchial injuries: 14-year experience. Eur J Cardio Thorac Surg. fication in 31 patients. Radiology. 1991;179:483–486.
2002;22:984–989. 171. Gruden JF, Webb WR, Sides DM. Adult-onset disseminated
144. Tsunezuka Y, et al. Spontaneous tracheal rupture associated tracheobronchial papillomatosis: CT features. J Comput Assist
with acquired tracheobronchomalacia. Ann Thorac Cardiovasc Tomogr. 1994;18:640–642.
Surg. 2003;9:394–396. 172. Kramer SS, et al. Pulmonary manifestations of juvenile laryngo-
145. Stark P. Imaging of tracheobronchial injuries. J Thorac Imaging. tracheal papillomatosis. AJR Am J Roentgenol. 1985;144:
1995;10:206–219. 687–694.
146. Devine N, et al. Bronchial anastomotic complications in lung 173. Prince JS, et al. Nonneoplastic lesions of the tracheobronchial
transplantation: CT assessment. J Radiol. 1996;77:477–481. wall: Radiologic findings with bronchoscopic correlation.
147. Fullerton DA, Campbell DN. Airway problems in lung trans- Radiographics. 2002;22:S215–S230.
plantation. Semin Respir Crit Care Med. 1996;17:187–196. 174. Cosio BG, et al. Endobronchial hamartoma. Chest. 2002;122:
148. Herman S, Rappaport DC, Weisbrod G, et al. Single-lung trans- 202–205.
plantation: imaging features. Radiology. 1989;170:89–93. 175. Siegelman SS, Khouri NF, Scott WW. Pulmonary hamartoma:
149. Ikonen T, et al. Lung and heart-lung transplantation: value of CT findings. Radiology. 1986;160:313–317.
high-resolution computed tomography in routine evaluation 176. Reittner P, Muller NL. Tracheal hamartoma: CT findings in two
of lung transplantation recipients during development of bron- patients. J Comput Assist Tomogr. 1999;23:957–958.
chiolitis obliterans syndrome. J Heart Lung Transplant. 1996;15: 177. Marom EM, Goodman PC, McAdams HP. Pictorial essay—focal
587–595. abnormalities of the trachea and main bronchi. AJR Am
150. McAdams HP, et al. Telescoping bronchial anastomoses for J Roentgenol. 2001;176:707–711.
unilateral or bilateral sequential lung transplantation: CT 178. Beigelman C, Sellami D, Brauner M. CT of parenchymal and
appearance. Radiology. 1997;203:202–206. bronchial tuberculosis. Eur Radiol. 2000;10:699–709.
151. Murray JG, et al. Pictorial essay. Complications of lung trans- 179. Fukuoka K, et al. Endobronchial lesions involved in Myco-
plantation: radiologic findings. AJR Am J Roentgenol. 1996;166: bacterium avium infection. Respir Med. 2003;97:1261–1264.
1405–1411. 180. Hoheisel G, et al. Endobronchial tuberculosis: diagnostic fea-
152. Paulson EC, et al. Bronchial sleeve resection for posttransplant tures and therapeutic outcome. Respir Med. 1994;88:593–597.
stricture. Ann Thorac Surg. 2003;76:2075–2076. 181. Kim Y, et al. Tuberculosis of the trachea and main bronchi: CT
153. Schafers HJ, et al. Surgical treatment of airway complications findings in 17 patients. AJR Am J Roentgenol. 1997;168:1051–1056.
after lung transplantation. J Thorac Cardiovasc Surg. 1994;107: 182. Lee KS, et al. Endobronchial tuberculosis: CT features. J Comput
1476–1480. Assist Tomogr. 1991;15:424–428.
154. Gotway MB, et al. Benign tracheobronchial stenoses: Changes 183. Moon WK, et al. Tuberculosis of the central airways: CT findings
in short-term and long-term pulmonary function testing after of active and fibrotic disease. AJR Am J Roentgenol. 1997;169:
expandable metallic stent placement. J Comput Assist Tomogr. 649–653.
2002;26:564–572. 184. VanDyck P, et al. Imaging of pulmonary tuberculosis. Eur Radiol.
155. Marel M, et al. Management of benign stenoses of the large air- 2003;13:1771–1785.
ways in the University Hospital in Prague, Czech Republic, in 185. Marom EM, Goodman PC, McAdams HP. Pictorial essay—
1998–2003. Respiration. 2005;72:622–628. diffuse abnormalities of the trachea and main bronchi. AJR Am
156. Shitrit D, et al. Accuracy of virtual bronchoscopy for grading J Roentgenol. 2001;176:713–717.
tracheobronchial stenosis—correlation with pulmonary func- 186. Choplin RH, Wehunt WD, Theros EG. Diffuse lesions of the
tion test and fiberoptic bronchoscopy. Chest. 2005;128: trachea. Semin Roentgenol. 1983;18:38–50.
3545–3550. 187. Schwartz M, Rossoff L. Tracheobronchomegaly. Chest. 1994;
157. Sorantin E, et al. Spiral CT based assessment of tracheal steno- 106:1589–1590.
sis using 3-D skeletonization. IEEE Trans Med Imaging. 2002;21: 188. Dunne MG, Reiner B. CT features of tracheobronchomegaly.
263–273. J Comput Assist Tomogr. 1988;2:388–391.
158. Whyte RI, et al. Helical computed tomography for the evalua- 189. Roditi GH, Weir J. The association of tracheomegaly and
tion of tracheal stenosis. Ann Thorac Surg. 1995;60:27–31. bronchiectasis. Clin Radiol. 1994;49:608–611.
159. Rea F, et al. Benign tracheal and laryngotracheal stenosis: surgi- 190. Woodring JH, et al. Acquired tracheomegaly in adults as a com-
cal treatment and results. Eur J Cardio Thorac Surg. 2002;22: plication of diffuse pulmonary fibrosis. AJR Am J Roentgenol.
352–356. 1989;152:743–747.
191. Boiselle PM, et al. Tracheobronchomalacia: evolving role of 216. Behar JV, et al. Relapsing polychondritis affecting the lower
dynamic multislice helical CT. Radiol Clin North Am. 2003;41: respiratory tract. AJR Am J Roentgenol. 2002;178:173–177.
627. 217. Im J-G, et al. CT manifestations of tracheobronchial involve-
192. Hasegawa I, et al. Tracheomalacia incidentally detected on CT ment in relapsing polychondritis. J Comput Assist Tomogr.
pulmonary angiography of patients with suspected pulmonary 1988;12:792–793.
embolism. AJR Am J Roentgenol. 2003;181:1505–1509. 218. Mendelson DS, et al. Relapsing polychondritis studied by com-
193. Hubbard M, Masters IB, Chang AB. Rapidly progressing case of puted tomography. Radiology. 1985;157:489–490.
Mounier-Kuhn syndrome in early childhood. Pediatr Pulmonol. 219. Trentham DE, Le CH. Relapsing polychondritis. Ann Intern Med.
2003;36:353–356. 1998;129:114–122.
193a. Nishino M, et al. Value of volumetric data acquisition in expira- 220. Sarodia BD, Dasgupta A, Mehta AC. Management of airway
tory high-resolution computed tomography of the lung. manifestations of relapsing polychondritis. Case reports and
J Comput Assist Tomogr. 2004;28:209–214. review of literature. Chest. 1999;116:1669–1675.
193b. Hein E, et al. Dynamic and quantitative assessment of tracheo- 221. Booth A, et al. The radiologic manifestations of relapsing poly-
malacia by electron beam tomography: correlation with clini- chondritis. Clin Radiol. 1989;40:147–149.
cal symptoms and bronchoscopy. J Comput Assist Tomogr. 222. Katzenstein A-LA. Major problems in pathology. In: Katzenstien
2000;24:247–252. and Askin’s surgical pathology of non-neoplastic lung disease.
194. Aquino SL, et al. Acquired tracheomalacia: detection by expira- Philadelphia: WB Saunders Co; 1997.
tory CT scan. J Comput Assist Tomogr. 2001;25:394–399. 223. Cordier JF, et al. Pulmonary Wegener’s granulomatosis. A clini-
195. Aquino SL, et al. Tree-in-bud pattern: frequency and significance cal and imaging study of 77 cases. Chest. 1990;97:906–912.
on thin section CT. J Comput Assist Tomogr. 1996;20:594–599. 224. Daum TE, et al. Tracheobronchial involvement in Wegener’s
196. Kwong JS, Muller NL, Miller RR. Disease of the trachea and granulomatosis. Am J Respir Crit Care Med. 1995;151:522–526.
main-stem bronchi: correlation of CT with pathologic findings. 225. Screaton NJ, et al. Tracheal involvement in Wegener’s gran-
Radiographics. 1992;12:645–657. ulomatosis: Evaluation using spiral CT. Clin Radiol. 1998;53:
197. Naidich DP, et al. Trachea and central bronchi. In: Imaging of 809–815.
the airways. Philadelphia: Lippincott Williams & Wilkins; 226. Lee KS, et al. Thoracic manifestation of Wegener’s granulomato-
2005:70–106. sis: CT findings in 30 patients. Eur Radiol. 2003;13:43–51.
198. Shepard JO, McLoud TC. Imaging the airways. Clin Chest Med. 227. Papiris SA, et al. Imaging of thoracic Wegener’s granulomatosis:
1991;12:151–168. the computed tomographic appearance. Am J Med. 1992;93:
199. Greene R. “Saber sheath” trachea: relation to chronic ob- 529–536.
structive pulmonary disease. AJR Am J Roentgenol. 1978;130: 228. Lohrmann C, et al. Pulmonary manifestations of Wegener
441–445. granulomatosis: CT findings in 57 patients and a review of the
200. Leader JK, et al. Size and morphology of the trachea before and literature. Eur J Radiol. 2005;53:471–477.
after lung volume reduction surgery. AJR Am J Roentgenol. 229. Aberle DR, Gamsu G, Lynch D. Thoracic manifestations of
2004;183:315–321. Wegener granulomatosis: diagnosis and course. Radiology.
201. Valor RR, et al. Bacterial tracheitis with upper airway obstruction 1990;174:703–709.
in a patient with the acquired immunodeficiency syndrome. Am 230. Deremee RA, Homburger HA, Specks U. Lesions of the respira-
Rev Respir Dis. 1992;146:1598–1599. tory tract associated with the finding of anti-neutrophil
202. Yigla M, et al. Laryngotracheobronchial involvement in a cytoplasmic autoantibodies with a perinuclear staining pattern.
patient with nonendemic rhinoscleroma. Chest. 2000;117: Mayo Clin Proc. 1994;69:819–824.
1795–1798. 231. Summers RM, et al. CT virtual bronchoscopy of the central
203. Sedano HO, Carlos R, Koutlas IG. Respiratory scleroma: a clini- airways in patients with Wegener’s granulomatosis. Chest. 2002;
copathologic and ultrastructural study. Oral Surg Oral Med Oral 121:242–250.
Pathol Oral Radiol Endod. 1996;81:665–671. 232. WHO-IUIS Nomenclature Sub-Committee. Nomenclature of amy-
204. Amoils CP, Shindo ML. Laryngotracheal manifestations of loid and amyloidosis. Bull World Health Org. 1993;71:105–112.
rhinoscleroma. Ann Otol Rhinol Laryngol. 1996;105:336–340. 233. Georgiades CS, et al. Amyloidosis: review and CT manifesta-
205. Hart CA, Rao SK. Rhinoscleroma [Editorial]. J Med Microbiol. tions. Radiographics. 2004;24:405–416.
2000;49:395–396. 234. Utz JP, Swensen SJ, Gertz MA. Pulmonary amyloidosis. The
206. Franquet T, et al. Necrotizing aspergillosis of large airways. CT Mayo Clinic experience from 1980 to 1993. Ann Intern Med.
findings in eight patients. J Comput Assist Tomogr. 2002;26: 1996;124:407–413.
342–345. 235. O’Regan A, et al. Tracheobronchial amyloidosis. The Boston
207. Buckingham SJ, Hansell DM. Aspergillus in the lung: diverse University experience from 1984 to 1999. Medicine. 2000;79:
and coincident forms. Eur Radiol. 2003;13:1786–1800. 69–79.
208. Miller WTJ, et al. Pulmonary aspergillosis in patients with 236. Mares DC, et al. Tracheobronchial amyloidosis: a review of clin-
AIDS. Clinical and radiographic correlations. Chest. ical and radiographic characteristics, bronchoscopic diagnosis,
1994;105:37–44. and management. J Bronchol. 1998;5:147–155.
209. Franquet T, et al. Aspergillus infection of the airways—computed 237. Kim HY, et al. Localized amyloidosis of the respiratory system:
tomography and pathologic findings. J Comput Assist Tomogr. CT features. J Comput Assist Tomogr. 1999;23:627–631.
2004;28:10–16. 238. Urban BA, et al. CT evaluation of amyloidosis: spectrum of
210. Putnam JB, et al. Acute airway obstruction and necrotizing disease. Radiographics. 1993;13:1295–1308.
tracheobronchitis from invasive mycosis. Chest. 1994;106: 239. Kirchner J, et al. CT findings in extensive tracheobronchial
1265–1267. amyloidosis. Eur Radiol. 1998;8:352–354.
211. Bioque JC, et al. Tracheobronchopathia osteochondroplastica. 240. Howard ME, Ireton J, Daniels F. Pulmonary presentations of
Clinical study and follow-up in nine cases. J Bronchol. 2001;8: amyloidosis. Respirology. 2001;6:61–64.
78–83. 241. Wilcox P, et al. Airway involvement in ulcerative colitis. Chest.
212. Manning JE, et al. Tracheobronchopathia osteochondroplas- 1987;92:18–22.
tica. Clin Radiol. 1998;53:302–304. 242. Woodring JH, Reed JC. Types and mechanisms of pulmonary
213. Meyer CN, Dossing M, Broholm H. Tracheobronchopathia atelectasis. J Thorac Imaging. 1996;11:92–108.
osteochondroplastica. Respir Med. 1997;91:499–502. 243. Naidich DP, et al. CT of lobar collapse. Semin Roentgenol.
214. Clee MD, Anderson JM, Johnston RN. Clinical aspects of 1984;19:222–235.
tracheobronchopathia osteochondroplastica. Br J Dis Chest. 244. Mintzer RA, Sakowicz BA, Blonder JA. Lobar collapse. Usual and
1983;77:308–314. unusual forms. Chest. 1988;94:615–620.
215. Mariotta S, et al. Spiral CT and endoscopic findings in a case of 245. Naidich DP, Lee JJ, Garay SM. Comparison of CT and fiberoptic
tracheobronchopathia osteochondroplastica. J Comput Assist bronchoscopy in the evaluation of bronchial disease. AJR Am
Tomogr. 1997;21:418–420. J Roentgenol. 1987;148:1–7.
246. Gurney JW. Atypical manifestations of pulmonary atelectasis. 274. McGuinness G, et al. AIDS-related airway disease. Am J Roentgenol.
J Thorac Imaging. 1996;11:165–175. 1997;168:67–77.
247. Heitmiller RF, et al. Endobronchial metastasis. J Thorac 275. McGuinness G, et al. AIDS-associated bronchiectasis: CT
Cardiovasc Surg. 1993;106:537–542. features. J Comput Assist Tomogr. 1993;17:260–266.
248. Kiani B, Magro CM, Ross P. Endobronchial presentation of 276. Thicket KM, et al. Common variable immune deficiency: respi-
Hodgkin lymphoma: A review of the literature. Ann Thorac Surg. ratory manifestations, pulmonary function and high-resolution
2003;76:967–972. CT scan findings. Q J Med. 2002;95:655–662.
249. Kiani B et al. Polypoid endobronchial Hodgkin lymphoma with 277. Aghamohammadi A, Farhoudi A, Moin M, et al. Clinical and
an initial response to photodynamic therapy. Ann Thorac Surg. immunological features of 65 Iranian patients with common
2003;76:940–942. variable immunodeficiency. Clin Diagn Lab Immunol. 2005;12:
250. Shah H, et al. Benign tumors of the tracheobronchial tree. 825–832.
Endoscopic characteristics and role of laser resection. Chest. 278. McMahon MA, et al. Alpha-1 antitrypsin deficiency and com-
1995;107:1744–1751. puted tomography findings. J Comput Assist Tomogr. 2005;29:
251. Woodring JH. The computed tomography mucous bron- 549–553.
chogram sign. J Comput Tomogr. 1988;12:165–168. 279. King MA, et al. Alpha(1)-antitrypsin deficiency: Evaluation of
252. Eisenhuber E, Brunner C, Bankier AA. Blood clots mimicking bronchiectasis with CT. Radiology. 1996;199:137–141.
peripheral intrabronchial tumors in patients with hemoptysis: 280. Watanabe Y, et al. Congenital bronchiectasis due to cartilage
CT and bronchoscopic findings. J Comput Assist Tomogr. 2000; deficiency: CT demonstration. J Comput Assist Tomogr. 1987;
24:47–51. 11:701–703.
253. Quint LE, et al. Three dimensional CT of central lung tumors. 281. Kaneko K, et al. Computed tomography findings in Williams-
Clin Imaging. 1990;14:323–329. Campbell syndrome. J Thorac Imaging. 1991;6:11–13.
254. Woodring JH. Determining the cause of pulmonary atelectasis: 282. Wark P. Pathogenesis of allergic bronchopulmonary aspergillo-
a comparison of plain radiography and CT. AJR Am J Roentgenol. sis and an evidence-based review of azoles in treatment. Respir
1988;150:757–763. Med. 2004;98:915–923.
255. Burke M, Fraser R. Obstructive pneumonitis: a pathologic and 283. Kullnig P, et al. Computerized tomography in the diagnosis of
pathogenetic reappraisal. Radiology. 1988;166:699–704. allergic bronchopulmonary aspergillosis. Radiology. 1989;29:
256. Kwon KY, et al. Middle lobe syndrome: A clinicopathological 228–231.
study of 21 patients. Hum Pathol. 1995;26:302–307. 284. Kasahara K, et al. Correlation between the bronchial subepithe-
257. Rosenbloom SA, Ravin CE, Putman CE, et al. Peripheral middle lial layer and whole airway wall thickness in patients with
lobe syndrome. Radiology. 1983;149:17–21. asthma. Thorax. 2002;57:242–246.
258. Gierada DS, Glazer HS, Slone RM. Pseudomass due to atelecta- 285. Gono H, et al. Evaluation of airway wall thickness and air
sis in patients with severe bullous emphysema. AJR Am trapping by HRCT in asymptomatic asthma. Eur Respir J. 2003;
J Roentgenol. 1997;168:85–92. 22:965–971.
259. Pinstein ML, et al. Middle lobe torsion following right upper 286. Matsuoka S, et al. Bronchoarterial ratio and bronchial wall thick-
lobectomy. Radiology. 1985;155:580. ness on high-resolution CT in asymptomatic subjects: correlation
260. Mah K, et al. Assessment of acute radiation-induced pulmonary with age and smoking. AJR Am J Roentgenol. 2003;180:513–518.
changes using computed tomography. J Comput Assist Tomogr. 287. Lee YM, et al. High-resolution CT findings in patients with near-
1986;10:736–743. fatal asthma—comparison of patients with mild-to-severe
261. Davis AL, Salzman SH, eds. Bronchiectasis. In: Cherniack asthma and normal control subjects and changes in airway
NS, ed. Chronic Obstructive Pulmonary Disease. Philadelphia: WB abnormalities following steroid treatment. Chest. 2004;126:
Saunders; 1991:316–338. 1840–1848.
262. Barker AF, Bardana EJ. Bronchiectasis: update on an orphan 288. Vignola AM, et al. Airway remodelling assessed by sputum and
disease. Am Rev Respir Dis. 1988;137:969–978. high-resolution computed tomography in asthma and COPD.
263. Stanford W, Galvin JR. The diagnosis of bronchiectasis. Clin Eur Respir J. 2004;24:910–917.
Chest Med. 1988;9:691–699. 289. Park SW, et al. Differences in radiological/HRCT findings in
264. Kumar NA, Nguyen B, Maki D. Bronchiectasis: current clinical eosinophilic bronchitis and asthma: implication for bronchial
and imaging concepts. Semin Roentgenol. 2001;36:41–50. responsiveness. Thorax. 2006;61:41–47.
265. Currie DC, et al. Comparison of narrow section computed 290. Lee ES, et al. Early bronchiolitis obliterans following lung trans-
tomography and plain chest radiography in chronic allergic plantation: accuracy of expiratory thin-section CT for diagnosis.
bronchopulmonary aspergillosis. Clin Radiol. 1987;38:593–596. Radiology. 2000;216:472–477.
266. Millar A, et al. The role of computed tomography (CT) in the 291. Reynaud-Gaubert M, et al. Early detection of airway involve-
investigation of unexplained hemoptysis. Respir Med. 1992;86: ment in obliterative bronchiolitis after lung transplantation—
39–44. functional and bronchoalveolar lavage cell findings. Am J Respir
267. Im JG, et al. Pulmonary tuberculosis: CT findings—early active Crit Care Med. 2000;161:1924–1929.
disease and sequential change with antituberculous therapy. 292. Boehler A, Estenne M. Post-transplant bronchiolitis obliterans.
Radiology. 1993;186:653–660. Eur Respir J. 2003;22:1007–1018.
268. Han DH, et al. Radiographic and CT findings of nontuberculous 293. Sibtain NA, Padley SPG. HRCT in small and large airways
mycobacterial pulmonary infection caused by Mycobacterium diseases. Eur Radiol. 2004;14:L31–L43.
abscessus. AJR Am J Roentgenol. 2003;181:513–517. 294. Shah RM, Miller W. Pulmonary complications of transplanta-
269. Fujita J, et al. Pathological findings of bronchiectases caused by tion: radiographic considerations. Clin Chest Med. 2005;26:545.
Mycobacterium avium intracellulare complex. Respir Med. 2003; 295. Crawford SW, Clark JG. Bronchiolitis associated with bone mar-
97:933–938. row transplantation. Clin Chest Med. 1993;14:741–749.
270. Jeong YJ, et al. Nontuberculous mycobacterial pulmonary 296. Agasthian T, et al. Surgical management of bronchiectasis. Ann
infection in immunocompetent patients: Comparison of thin- Thorac Surg. 1996;62:976–978.
section CT and histopathologic findings. Radiology. 2004;231: 297. Mazieres J, et al. Limited operation for severe multisegmental
880–886. bilateral bronchiectasis. Ann Thorac Surg. 2003;75:382–387.
271. Hartman TE, Swense SJ, Williams DE. Mycobacterium avium- 298. Ayed AK. Resection of the right middle lobe and lingula in
intracellulare complex: evaluation with CT. Radiology. 1993;187: children for middle lobe/lingula syndrome. Chest. 2004;125:
23–26. 38–42.
272. Swenson SJ, Hartman TE, Williams DE. Computed tomography 299. Naidich DP. Computed tomography of bronchiectasis. J Comput
in diagnosis of Mycobacterium avium-intracellulare complex in Assist Tomogr. 1982;6:437–444.
patients with bronchiectasis. Chest. 1994;105:49–52. 300. Naidich DP, Harkin TJ. Airways and lung: CT versus bronchogra-
273. Moore EH. Atypical mycobacterial infection in the lung: CT phy through the fiberoptic bronchoscope. Radiology. 1996;200:
appearance. Radiology. 1993;187:777–782. 613–614.
301. Kang EY, Miller RR, Müller NL. Bronchiectasis: comparison of 327. Altin R, et al. Presence and HRCT quantification of bronchiecta-
preoperative thin-section CT and pathologic findings in resected sis in coal workers. Eur J Radiol. 2004;52:157–163.
specimens. Radiology. 1995;195:649–654. 328. Morice AH, et al. The diagnosis and management of chronic
302. Smith IE, et al. Chronic sputum production: correlations between cough. Eur Respir J. 2004;24:481–492.
clinical features and findings on high resolution computed tomo- 329. Edwards EA, et al. HRCT lung abnormalities are not a surrogate
graphic scanning of the chest. Thorax. 1996;51:914–918. for exercise limitation in bronchiectasis. Eur Respir J. 2004;24:
303. Kim JS, et al. Bronchoarterial ratio on thin section CT: compari- 538–544.
son between high altitude and sea level. J Comput Assist Tomogr. 330. Reiff DB, et al. CT findings in bronchiectasis: limited value in
1997;21:306–311. distinguishing between idiopathic and specific types. AJR Am
304. Diederich S, Jurriaans E, Flower CDR. Interobserver variation in J Roentgenol. 1995;165:261–267.
the diagnosis of bronchiectasis on high-resolution computed 331. Shah RM, et al. High-resolution CT in the acute exacerbation of
tomography. Eur Radiol. 1996;6:801–806. cystic fibrosis: evaluation of acute findings, reversibility of those
305. Joharjy IA, Bashi SA, Abdullah AK. Value of medium-thickness findings, and clinical correlation. AJR Am J Roentgenol. 1997;
CT in the diagnosis of bronchiectasis. AJR Am J Roentgenol. 169:375–380.
1987;149:1133–1137. 332. Maffessanti M, et al. Cystic fibrosis in children: HRCT findings
306. Cooke JC, Currie DC, Morgan MP. Role of computed tomography and distribution of disease. J Thorac Imaging. 1996;11:27–38.
in the diagnosis of bronchiectasis. Thorax. 1987;42:272–277. 333. Cartier Y, et al. Bronchiectasis: accuracy of high-resolution CT in
307. Silverman PM, Godwin JD. CT/bronchographic correlations in the differentiation of specific diseases. AJR Am J Roentgenol.
bronchiectasis. J Comput Assist Tomogr. 1987;11:52–56. 1999;173:47–52.
308. Young K, Aspestrand F, Kolbenstvedt A. High resolution CT and 334. Helbich TH, et al. Cystic fibrosis: CT assessment of lung involve-
bronchography in the assessment of bronchiectasis. Acta Radiol. ment in children and adults. Radiology. 1999;213:537–544.
1991;32:439–441. 335. Helbich TH, et al. Evolution of CT findings in patients with
309. McGuinness G, Naidich DP, McAuley DI. Bronchiectasis: CT cystic fibrosis. AJR Am J Roentgenol. 1999;173:81–88.
evaluation (pictorial essay). AJR Am J Roentgenol. 1993;160: 336. Edwards EA, Metcalfe R, Milne DG, et al. Retrospective review of
253–259. children presenting with non cystic fibrosis bronchiectasis:
310. Kim SJ, Im JG, Kim IO. Normal bronchial and pulmonary arte- HRCT findings and clinical relationships. Pediatr Pulmonol.
rial diameters measured by thin section CT. J Comput Assist 2003;36:87–93.
Tomogr. 1995;19:365–369. 337. Lynch DA, et al. Correlation of CT findings with clinical evalua-
311. Lynch DA, et al. Frequency of bronchial dilatation at high- tions in 261 patients with symptomatic bronchiectasis. AJR Am
resolution CT in asthmatics and control subjects. Radiology. J Roentgenol. 1999;173:53–58.
1991;181(P):250. 338. Roberts HR, et al. Airflow obstruction in bronchiectasis: correla-
312. Seneterre E, et al. Measurement of the internal size of bronchi tion between computed tomography features and pulmonary
using high resolution computed tomography (HRCT). Eur function tests. Thorax. 2000;55:198–204.
Respir J. 1994;7:596–600. 339. Tarver RD, Conces DJ, Godwin JD. Motion artifacts on CT sim-
313. Desai SR, et al. The reproducibility of bronchial circumference ulate bronchiectasis. AJR Am J Roentgenol. 1988;151:1117–1119.
measurements using computed tomography. Br J Radiol. 340. Lynch DA, et al. Idiopathic interstitial pneumonias: CT features.
1994;67:257–262. Radiology. 2005;236:10–21.
314. Kiraly AP, et al. Three-dimensional human airway segmentation 341. Patz EF, Swensen SJ, Erasmus J. Pulmonary manifestations of
methods for clinical virtual bronchoscopy. Acad Radiol. 2002; nontuberculous mycobacterium. Radiol Clin North Am. 1995;33:
9:1153–1168. 719–729.
315. Lynch DA, et al. Uncomplicated asthma in adults: comparison 342. Albelda SM, Kern JA, Martinelli DL. Expanding spectrum of
of CT appearance of the lungs in asthmatic and healthy subjects. pulmonary disease caused by nontuberculous mycobacteria.
316. Kim JS, et al. Cylindrical bronchiectasis: diagnostic findings on 343. Tanaka H, et al. Yield of computed tomography and bron-
thin-section CT. AJR Am J Roentgenol. 1997;168:751–764. choscopy for the diagnosis of Mycobacterium avium complex
317. Weibel ER. High resolution computed tomography of the pulmonary pulmonary disease. Am J Respir Crit Care Med. 1997;155:
parenchyma: anatomical background. Presented at the Fleischner 2041–2046.
Society Symposium on Chest Disease, Scottsdale, Arizona, 344. Boucher RC. New concepts of the pathogenesis of cystic fibrosis
1990. lung disease. Eur Respir J. 2004;23:146–158.
318. Murata K, Itoh H, Todo G. Centrilobular lesions of the lung: 345. Wood BP. Cystic fibrosis: 1997. Radiology. 1997;204:1–10.
demonstration by high-resolution CT and pathologic correla- 346. Stern RC. The diagnosis of cystic fibrosis. N Engl J Med.
tion. Radiology. 1986;161:641–645. 1997;336:487–491.
319. Webb WR, et al. CT of a bronchial phantom: factors affecting 347. Demirkazik FB, et al. High resolution CT in children with cystic
appearance and size measurements. Invest Radiol. 1984;19: fibrosis: correlation with pulmonary functions and radi-
394–398. ographic scores. Eur J Radiol. 2001;37:54–59.
320. Bhalla M, et al. Cystic fibrosis: scoring system with thin-section 348. Brody AS, et al. Computed tomography in the evaluation of cys-
CT. Radiology. 1991;179:783–788. tic fibrosis lung disease. Am J Respir Crit Care Med. 2005;172:
321. Webb WR. Thin-section CT of the secondary pulmonary lobule: 1246–1252.
anatomy and the image—the 2004 Fleischner Lecture. Radiology. 349. Santis G, Hodson ME, Strickland B. High resolution computed
2006;239:322–339. tomography in adult cystic fibrosis patients with mild lung
322. Goyal R, et al. Case report. High attenuation mucous plugs in disease. Clin Radiol. 1991;44:20–22.
allergic bronchopulmonary aspergillosis: CT appearance. J Comput 350. Lynch DA, et al. Pediatric pulmonary disease: assessment with
Assist Tomogr. 1992;16:649–650. high-resolution ultrafast CT. Radiology. 1990;176:243–248.
323. Logan PM, Muller NL. High-resolution computed tomography 351. Marchant JM, Masel JP, Dickenson FL, et al. Application of chest
and pathologic findings in pulmonary aspergillosis: a pictorial high-resolution computer tomography in young children with
essay. Can Assoc Radiol J. 1996;47:444–452. cystic fibrosis. Pediatr Pulmonol. 2001;31:24–29.
324. Hansell DM, Strickland B. High-resolution computed tomogra- 352. Robinson TE. High-resolution CT scanning: potential outcome
phy in pulmonary cystic fibrosis. Br J Radiol. 1989;62:1–5. measure. Curr Opin Pulmon Med. 2004;10:537–541.
325. de Jong PA, et al. Progression of lung disease on computed 353. Brody AS, et al. High-resolution computed tomography of the
tomography and pulmonary function tests in children and chest in children with cystic fibrosis: support for use as an out-
adults with cystic fibrosis. Thorax. 2006;61:80–85. come surrogate. Pediatr Radiol. 1999;29:731–735.
326. Ooi GC, et al. High-resolution CT quantification of bronchiec- 354. Sanatmaria F, et al. Cystic fibrosis: when should high-resolution
tasis: Clinical and functional correlation. Radiology. 2002;225: computed tomography of the chest be obtained? Pediatrics.
663–672. 1998;101:908–913.
355. Nathanson I, et al. Ultrafast computerized tomography of the 383. Garg K, et al. Proliferative and constrictive bronchiolitis: classifi-
chest in cystic fibrosis: a new scoring system. Pediatr Pulmonol. cation and radiologic features. AJR Am J Roentgenol. 1994;162:
1991;11:81–86. 803–808.
356. Marom EM, et al. Cystic fibrosis: usefulness of thoracic CT in 384. Hartman TE, et al. CT of bronchial and bronchiolar diseases.
the examination of patients before lung transplantation. Radiographics. 1994;14:991–1003.
Radiology. 1999;213:283–288. 385. Müller NL, Miller RR. Diseases of the bronchioles: CT and
357. Zhaomking W, Lockey RF. A review of allergic bronchopulmonary histopathologic findings. Radiology. 1995;196:3–12.
aspergillosis. J Invest Allergol Clin Immunol. 1996;6:144–151. 386. Teel GS, et al. Imaging of small airways disease. Radiographics.
358. Neeld DA, et al. Computerized tomography in the evaluation of 1996;16:27–41.
allergic bronchopulmonary aspergillosis. Am Rev Respir Dis. 387. Sharma V, et al. The radiological spectrum of small-airway dis-
1990;142:1200–1206. eases. Semin Ultrasound CT MRI. 2002;23:339–351.
359. Shah A, et al. CT in childhood allergic bronchopulmonary 388. Ryu JH, Myers J, Swenson SJ. Bronchiolar disorders. Am J Respir
aspergillosis. Pediatr Radiol. 1992;22:227–228. Crit Care Med. 2003;168:1277–1292.
360. Sandhu M, Mukhopadhyay S, Sharma SK. Allergic bronchopul- 389. Gruden JF, Webb WR, Warnock M. Pictorial essay.
monary aspergillosis: a comparative evaluation of computed Centrilobular opacities in the lung on high-resolution CT:
tomography with plain chest radiography. Australas Radiol. diagnostic considerations and pathologic correlation. AJR Am J
1994;38:288–293. Roentgenol. 1994;162:569–574.
361. Angus RM, et al. Computed tomographic scanning of the lung 390. Desai SR, Hansell DM. Small airways disease: expiratory com-
in patients with allergic bronchopulmonary aspergillosis and in puted tomography comes of age. Clin Radiol. 1997;52:
asthmatic patients with a positive skin test to Aspergillus fumi- 332–337.
gatus. Thorax. 1994;49:586–589. 391. Akira M, et al. Diffuse panbronchiolitis: evaluation with high-
362. Panchal N, et al. Central bronchiectasis in allergic bronchopul- resolution CT. Radiology. 1988;168:433–438.
monary aspergillosis—comparative evaluation of computed 392. Akira M, et al. Diffuse panbronchiolitis: follow-up CT examina-
tomography of the thorax with bronchography. Eur Respir J. tion. Radiology. 1993;189:559–562.
1994;7:1290–1293. 393. Iwata M, Colby TV, Kitaichi M. Diffuse panbronchiolitis: diag-
363. Lee PH, et al. Accuracy of CT in predicting the cause of nosis and distinction from various pulmonary diseases with
bronchiectasis. Clin Radiol. 1995;50:839–841. centrilobular interstitial foam cell accumulations. Hum Pathol.
364. Logan PM, Muller NL. CT manifestations of pulmonary 1994;25:357–363.
aspergillosis. Crit Rev Diagn Imaging. 1996;37:1–37. 394. Nishimura K, et al. Diffuse panbronchiolitis: correlation of high-
365. Ward S, et al. Accuracy of CT in the diagnosis of allergic bron- resolution CT and pathologic findings. Radiology. 1992;184:
chopulmonary aspergillosis in asthmatic patients. AJR Am 779–785.
J Roentgenol. 1999;173:937–942. 395. Ichikawa Y, et al. Reversible airway lesions in diffuse panbron-
366. Webb WR. High-resolution computed tomography of obstruc- chiolitis. Detection by high-resolution computed tomography.
tive lung disease. Radiol Clin North Am. 1994;32:745–757. Chest. 1995;107:120–125.
367. Logan PM, Muller NL. High-attenuation mucous plugging in 396. Hansell DM, et al. Hypersensitivity pneumonitis: correlation of
allergic bronchopulmonary aspergillosis. Can Assoc Radiol J. individual CT patterns with functional abnormalities. Radiology.
1996;47:374–377. 1996;199:123–128.
368. Roberts CM, Citron KM, Strickland B. Intrathoracic asper- 397. Moysan P, et al. Hypersensitivity pneumonitis: CT findings.
gilloma: role of CT in diagnosis and treatment. Radiology. 1987; J Radiol. 1995;76:435–440.
165:123. 398. Romero S, et al. Follicular bronchiolitis: Clinical and pathologic
369. Logan PM, et al. Invasive aspergillosis of the airways: radi- findings in six patients. Lung. 2003;181:309–319.
ographic, CT, and pathologic findings. Radiology. 1994;193: 399. Davies G, Wells AU, duBois RM. Respiratory bronchiolitis
383–388. associated with interstitial lung disease and desquamative inter-
370. Zeinreich SJ, Kennedy SW, Malat M, et. al. Fungal sinusitis: diag- stitial pneumonia. Clin Chest Med. 2004;25:717.
nosis with CT and MR imaging. Radiology. 1988;169:439–444. 400. Ryu JH, et al. Desquamative interstitial pneumonia and respira-
371. Primack SL, et al. High-resolution CT of the lung: Pitfalls in the tory bronchiolitis-associated interstitial lung disease. Chest.
diagnosis of infiltrative lung disease. AJR Am J Roentgenol. 2005;127:178–184.
1996;167:413–418. 401. Sadikot RT, et al. Respiratory bronchiolitis associated with
372. King TE. Overview of bronchiolitis. Clin Chest Med. 1993;14: severe dyspnea, exertional hypoxemia, and clubbing. Chest.
607–610. 2000;117:282–285.
373. Colby TV. Bronchiolitis: pathologic considerations. Am J Clin 402. Lynch DA, et al. Hypersensitivity pneumonitis: sensitivity of
Pathol. 1998;109:101–109. high-resolution CT in a population-based study. AJR Am
374. Hogg JC, Macklem PT, Thurlbeck WM. Site and nature of airway J Roentgenol. 1992;159:469–472.
obstruction in chronic obstructive lung disease. N Engl J Med. 403. Akira M, et al. Summer-type hypersensitivity pneumonitis: com-
1968;268:1355–1360. parison of high-resolution CT and plain radiographic findings.
375. Macklem P, Mead J. Resistance of central and peripheral airways AJR Am J Roentgenol. 1992;158:1223–1228.
measured by a retrograde catheter. J Appl Physiol. 1967;22: 404. McGuinness G, Scholes JV, Jagirdar JS, et al. Unusual lympho-
395–401. proliferative disorders in nine adults with HIV or AIDS: CT and
376. Worthy SA, Muller NL. Small airway diseases. Radiol Clin North pathologic findings. Radiology. 1995;197:59–65.
Am. 1998;36:163. 405. Kinoshita M, et al. Follicular bronchiolitis associated with
377. Myers JL, Colby TV. Pathologic manifestations of bronchiolitis, rheumatoid arthritis. Intern Med. 1992;31:674–677.
constrictive bronchiolitis, cryptogenic organizing pneumonia, 406. Hayakawa H, et al. Bronchiolar disease in rheumatoid arthritis.
and diffuse panbronchiolitis. Clin Chest Med. 1993;14:611–622. Am Rev Respir Crit Care Med. 1996;154:1531–1536.
378. Epler GR, Colby TV. The spectrum of bronchiolitis obliterans. 407. Heyneman LE, et al. Respiratory bronchiolitis, respiratory bron-
Chest. 1983;83:161–162. chiolitis-associated interstitial lung disease, and desquamative
379. Myers JL, Colby TV. Pathologic manifestations of bronchiolitis, interstitial pneumonia: Different entities or part of the spectrum
constrictive bronchiolitis, cryptogenic organizing pneumonia of the same disease process? AJR Am J Roentgenol. 1999;173:
and diffuse panbronchiolitis. Clin Chest Med. 1993;14:611–622. 1617–1622.
380. Epler GR, et al. Idiopathic bronchiolitis obliterans with organiz- 408. Ujita M, et al. Organizing pneumonia: perilobular pattern at
ing pneumonia. N Engl J Med. 1985;312:152–159. thin-section CT. Radiology. 2004;232:757–761.
381. Meyers JL. Respiratory bronchiolitis with interstitial lung 409. Cordier JF. Cryptogenic organizing pneumonia. Clin Chest Med.
disease. Semin Respir Med. 1992;13:134–139. 2004;25:727.
382. Lynch DA. Imaging of small airways diseases. Clin Chest Med. 410. Cordier JF. Bronchiolitis obliterans organizing pneumonia.
1993;14:623–634. Semin Respir Crit Care Med. 2000;21:135–146.
411. Yoshinouchi T, et al. Clinicopathological study on two types of 438. Leung AN, et al. Bronchiolitis obliterans after lung transplan-
cryptogenic organizing pneumonitis. Respir Med. 1995;89: tation: detection using expiratory HRCT. Chest. 1998;113:
271–278. 365–370.
412. Triggiani M, et al. Lung involvement in rheumatoid arthritis. 439. Dauber JH. Posttransplant bronchiolitis obliterans syndrome.
Sarcoidosis Vasculitis Diffuse Lung Dis. 2003;20:171–179. Where have we been and where are we going? Editorial. Chest.
413. Camus P, et al. Interstitial lung disease induced by drugs and 1996;109:857–859.
radiation. Respiration. 2004;71:301–326. 440. Soubani AO, Miller KB, Hassoun PM. Pulmonary complications
414. Epler GR. Drug-induced bronchiolitis obliterans organizing of bone marrow transplantation. Chest. 1996;109:1066–1077.
pneumonia. Clin Chest Med. 2004;25:89. 441. Pipvath SJ, et al. Radiologic and pathologic features of bronchi-
415. Wells AU, duBois RM. Bronchiolitis in association with connec- olitis. AJR Am J Roentgenol. 2005;185:354–363.
tive tissue diseases. Clin Chest Med. 1993;14:655–666. 442. Keller CA, et al. Bronchiolitis obliterans in recipients of single,
416. Crestani B, et al. Migratory bronchiolitis obliterans organizing double, and heart-lung transplantation. Chest. 1995;107:
pneumonia after unilateral radiation therapy for breast carci- 973–980.
noma. Eur Respir J. 1995;8:318–321. 442a. American Thoracic Society. Standards for the diagnosis and care
417. Colby TV, Myers JL. The clinical and histologic spectrum of patients with chronic obstructive pulmonary disease (COPD)
of bronchiolitis obliterans including bronchiolitis obliterans and asthma. Am J Respir Crit Care Med. 1995;152:S77–S121.
organizing pneumonia (BOOP). Semin Respir Dis. 1992;13: 443. Jeffery PK. Comparative morphology of the airways in asthma
119–133. and chronic obstructive pulmonary disease. Am J Respir Crit
418. Cordier JF, Loire R, Brune J. Idiopathic bronchiolitis obliterans Care Med. 1994;150(5 suppl):S6–S13.
organizing pneumonia. Chest. 1989;96:999–1005. 444. Lynch DA. Imaging of asthma and allergic bronchopulmonary
419. King TE. BOOP: an important cause of migratory pulmonary mycosis. Radiol Clin North Am. 1998;36:129.
infiltrates? Eur Respir J. 1995;8:193–195. 445. Khondoh Y, Taniguchi H, Yokoyama S, et al. Emphysematous
420. Pipavath S, Godwin JD. Imaging of the chest: idiopathic intersti- change in chronic asthma in relation to cigarette smoking:
tial pneumonia. Clin Chest Med. 2004;25:651. assessment by computed tomography. Chest. 1990;97:
421. Lee KS, et al. Cryptogenic organizing pneumonia: CT findings 845–849.
in 43 patients. AJR Am J Roentgenol. 1994;162:543–546. 446. Newman K, Lynch DA, Newman LS, et al. Quantitative com-
422. Akira M, Yamamoto S, Sakatani M. Bronchiolitis obliterans puted tomography detects air trapping due to asthma. Chest.
organizing pneumonia manifesting as multiple large nodules of 1994;106:105–109.
masses. AJR Am J Roentgenol. 1998;170:291–295. 447. Park CS, et al. Airway obstruction in asthmatic and healthy in-
423. Watanabe K, et al. Organizing pneumonia presenting as a dividuals: Inspiratory and expiratory thin-section CT findings.
solitary nodular shadow on a chest radiograph. Respiration. Radiology. 1997;203:361–367.
2003;70:507–514. 448. Grenier P, et al. Abnormalities of the airways and lung
424. Bouchardy LM, et al. CT findings in bronchiolitis obliterans parenchyma in asthmatics: CT observations in 50 patients and
organizing pneumonia (BOOP) with radiographic, clinical, and inter- and intraobserver variability. Eur Radiol. 1996;6:199–206.
histologic correlation. J Comput Assist Tomogr. 1993;17:352–357. 449. Paganin F, Seneterr E, Chanez P, et al. Computed tomography
425. Dodd JD, Muller NL. Bronchiolitis obliterans organizing pneu- of the lungs in asthma: influence of disease severity and etiol-
monia after bone marrow transplantation—high-resolution ogy. Am J Respir Crit Care Med. 1996;152:110–114.
computed tomography findings in 4 patients. J Comput Assist 450. Moersch HJ. Clinical significance of hemoptysis. JAMA. 1952;
Tomogr. 2005;29:540–543. 148:1461–1465.
426. Strange C, Highland KB. Interstitial lung disease in the patient 451. Pursel SE, Lindskog GE. Hemoptysis: a clinical evaluation
who has connective tissue disease. Clin Chest Med. 2004;25:549. of 105 patients examined consecutively on a thoracic surgical
427. McDonagh J, et al. High resolution computed tomography of service. Am Rev Respir Dis. 1961;84:329–336.
the lungs in patients with rheumatoid arthritis and interstitial 452. Abbott GA. The clinical significance of pulmonary hemorrhage:
lung disease. Br J Rheumatol. 1994;33:118–122. a study of 1316 patients with chest disease. Dis Chest. 1948;14:
428. Remy-Jardin M, et al. Lung changes in rheumatoid arthritis: CT 824–842.
findings. Radiology. 1994;193:375–382. 453. Weaver LJ, Sollioday N, Cugell DW. Selection of patients with
429. Perez T, Remyjardin M, Cortet B. Airways involvement in hemoptysis for fiberoptic bronchoscopy. Chest. 1979;76:7–10.
rheumatoid arthritis: clinical, functional, and HRCT findings. 454. Johnson H, Reisz G. Changing spectrum of hemoptysis: under-
Am J Respir Crit Care Med. 1998;157:1658–1665. lying causes in 148 patients undergoing diagnostic flexible
430. Taouli B, Brauner MW, Mourey Lea. Thin-section chest CT bronchoscopy. Arch Intern Med. 1989;149:1666–1668.
findings of primary Sjögren’s syndrome: correlation with 455. Munro NC, et al. Comparison of thin section computed to-
pulmonary function. Eur Radiol. 2002;12:1504–1511. mography with bronchography for identifying bronchiectatic
431. Zhang LJ, et al. High-resolution computed tomography in pedi- segments in patients with chronic sputum production. Thorax.
atric patients with postinfectious bronchiolitis obliterans. 1990;45:135–139.
J Thorac Imaging. 1999;14:85–89. 456. Grenier P, Lenoir S, Brauner M. Computed tomographic assess-
432. Chang H, et al. Histopathological study of Sauropus androgy- ment of bronchiectasis. Semin Ultrasound CT MR. 1990;11:
nus-associated constrictive bronchiolitis obliterans. Am J Surg 430–441.
Pathol. 1997;21:35–42. 457. Set PAK, et al. Hemoptysis: comparative study of the role of CT
433. Yang CF, et al. Correlation of high-resolution CT and and fiberoptic bronchoscopy. Radiology. 1993;189:677–680.
pulmonary function in bronchiolitis obliterans: A study based 458. Worthy SA, Flower CDR. Computed tomography of the air-
on 24 patients associated with consumption of Sauropus ways. Eur Radiol. 1996;6:717–729.
androgynus. AJR Am J Roentgenol. 1997;168:1045–1050. 459. Tasker AD, Flower CDR. Imaging the airways—hemoptysis,
434. Miller RR, Müller NL. Neuroendocrine cell hyperplasia and bronchiectasis, and small airways disease. Clin Chest Med.
obliterative bronchiolitis in patients with peripheral carcinoid 1999; 20:761.
tumors. Am J Surg Pathol. 1995;19:653–658. 460. Revel MP, et al. Can CT replace bronchoscopy in the det-
435. Brown MJ, English J, Müller NL. Case report. Bronchiolitis ection of the site and cause of bleeding in patients with
obliterans due to neuroendocrine hyperplasia: high-resolution large or massive hemoptysis? AJR Am J Roentgenol. 2002;179:
CT-pathologic correlation. AJR Am J Roentgenol. 1997;168: 1217–1224.
1561–1562. 461. Vernhet H, et al. Value of thoracic CT in the management of
436. Worthy SA, et al. Bronchiolitis obliterans after lung transplanta- severe hemoptysis. J Radiologie. 2003;84:685–691.
tion: high resolution CT findings in 15 patients. AJR Am 462. Lenner R, Schilero GJ, Lesser M. Hemoptysis: diagnosis and
J Roentgenol. 1997;169:673–677. management. Comp Ther. 2002;28:7–14.
437. Shepard JAO. Imaging of lung transplantation. Clin Chest Med. 463. Law GTS, et al. Cost-effectiveness of CT thorax and bron-
1999;20:827. choscopy in haemoptysis with normal CXR for exclusion of
lung cancer. Australas Radiol. 2002;46:381–383.
464. Tak S, et al. Haemoptysis in patients with a normal chest radi- 471. Poe RH, et al. Utility of fiberoptic bronchoscopy in patients
ograph: bronchoscopy—CT correlation. Australasian Radiol. with hemoptysis and a nonlocalizing chest roentgenogram.
1999;43:451–455. Chest. 1988;92:70–75.
465. Rohwedder JJ. Editorial. Enticements for fruitless bronchoscopy. 472. Lederle FA, Nichol KL, Parenti CM. Bronchoscopy to evaluate
Chest. 1989;96:708–710. hemoptysis in older men with nonsuspicious chest roentgeno-
466. Dierkesman R. The diagnostic yield of bronchoscopy. Cardiovasc grams. Chest. 1989;95:1043–1047.
Intervent Radiol. 1991;14:24–28. 473. Adelman M, et al. Cryptogenic hemoptysis: clinical features,
467. Santiago S, Tobias J, Williams AJ. A reappraisal of the causes of bronchoscopic findings, and natural history in 67 patients. Ann
hemoptysis. Arch Intern Med. 1991;151:2449–2451. Intern Med. 1985;102:829–834.
468. Peters J, McClung HC, Teague RB. Evaluation of hemoptysis in 474. Herth F, Ernst A, Becker HD. Long-term outcome and lung
patients with a normal chest roentgenogram. West J Med. cancer incidence in patients with hemoptysis of unknown
1984;141:624–626. origin. Chest. 2001;120:1592–1594.
469. Jackson CV, Savage PJ, Quinn DL. Role of fiberoptic bron- 475. Aziz ZA, et al. Functional impairment in emphysema:
choscopy in patients with hemoptysis and normal chest Contribution of airway abnormalities and distribution of
radiographs. Chest. 1985;87:142–144. parenchymal disease. AJR Am J Roentgenol. 2005;185:1509–1515.
470. Sharma SK, et al. Fiberoptic bronchoscopy in patients with 476. Remy-Jardin M, et al. Bronchial and nonbronchial systemic
haemoptysis and normal chest roentgenograms. Indian J Chest arteries at multi-detector row CT angiography: comparison with
Dis Allied Sci. 1991;33:15–18. conventional angiography. Radiology. 2004;233:741–749.
Focal Lung Disease 6
DETECTION 557 diagnosis. Finally, paralleling these developments has been

Magnetic Resonance Imaging 560 the growing availability of advanced imaging workstations
allowing routine use of increasingly sophisticated postpro-
CHARACTERIZATION 561 cessing techniques such as automated volumetric assess-
Definition 561 ment as well as a potential role of computer-aided diagnosis
Clinical Evaluation 561 (CAD) and characterization (CADx) to further refine and
Morphologic Evaluation 562 enlarge our diagnostic armamentarium.
Contrast-Enhancement Computed Tomography Despite these technologic innovations, however, and
Evaluation 589 considerable advances in biopsy techniques, controversy
Magnetic Resonance Imaging 593 persists concerning optimal evaluation and management
Positron Emission Tomography 596 of these cases, with ongoing controversy regarding the
Growth Characteristics 601 definition of standard of care. For the purposes of this
chapter, evaluation of pulmonary nodules is considered
MANAGEMENT 604 under the umbrella of focal lung disease, as many of the
Factors Affecting Clinical Management 604 features pertinent to lung nodule evaluation also pertain
Follow-up Evaluation of Indeterminate to this more inclusive category. The topic of lung cancer
Lung Nodules 605 screening is dealt with separately in Chapter 7. In an
Workup of Indeterminate Nodules attempt to organize this vast topic, our approach is to
Less Than 10 mm 605 consider three separate, albeit overlapping, categories:
Workup of Indeterminate Nodules 1 to 3 cm 606 detection, characterization, and management.
Biopsy 607
FUTURE DIRECTIONS: COMPUTER-AIDED DETECTION

DIAGNOSIS 610
Failure to recognize lung cancer on routine chest radi-
ographs (CXRs) is one of the most frequent causes of
Although assessment of patients with focal lung abnormali- missed diagnosis in radiology, frequently resulting in a
ties continues to represent an important problem in pul- significant delay in diagnosis (4). Quekel et al. (5) looked
monary diagnosis (1–3), the past decade has seen important at CXRs retrospectively in 259 patients with proven
developments both in hardware and in software applica- non–small cell lung cancer (NSCLC) and found a 19%
tions that have transformed our approach to both the identi- incidence of missed diagnosis. Those patients with missed
fication and especially the characterization of lung nodules. lesions had significantly smaller nodules (median diame-
The introduction in particular of multidetector computed ter 16 mm), with indistinct borders frequently obscured
tomography (MDCT) scanners capable of high-resolution by overlying osseous structures. Most important, missed
imaging of the entire thorax in a single breath-hold has lesions led to a significant delay in diagnosis when com-
revolutionized our evaluation of lung nodules. Of equal pared with lesions initially correctly diagnosed (472 vs.
importance has been the development and continued clin- 29 days, respectively), resulting in 43% of lesions being
ical validation of 2-[fluorine-18]-fluoro-2-deoxy-D-glucose upstaged from TI to T2 lesions during the delay period.
positron emission tomography (FDG-PET) scanning for From the outset, it has been appreciated that CT leads
providing uniquely physiologic data, further refining to the earlier identification of lung nodules, further
improved with the introduction of spiral CT (6–9). With CT remain a frequent occurrence (Fig. 6-2) (15–20). Of
CT, the ability to detect lung nodules is contingent on particular concern is the potential for overlooking small
a number of factors. These include variations between lung cancers. Hazelrigg et al. (21), for example, in a study
scanners and scan techniques, variations in nodule char- of 281 patients undergoing lung reduction surgery for
acteristics, and interobserver variability. Advantages first emphysema, found that of 17 (22%) of 78 lung
identified resulting from single-detector (SD) CT scanners nodules subsequently proved to be malignant surgically,
included reduction in the time of examination as well 14 retrospectively proved to have false-negative CT studies.
as the ability to obtain overlapping thin CT sections It is worth noting that the causes of missed nodules on
without the need to obtain additional scans (10,11). CT are quite distinct from those identified as causes for
As spiral CT technology has evolved from SD to 4-, 16-, missed nodules on routine CXRs. In a recent review of
and 64-detector rows, the number of nodules that are 40 NSCLCs seen only retrospectively between 1993
routinely identified has significantly increased (Fig. 6-1). and 2001, the most common causes for missed cancers
Although nodules have been reported to occur in up to 45% included upper lobe distribution (with right-sided lesions
of individuals screened for lung cancer using single-slice low- occurring in 45% compared with 28% on the left side),
dose computed tomography (LDCT) with 5-mm collimation location in the apical and posterior segments or sub-
(12), more recent reports using MDCT show that nodules segments (60%), and obscuration by an overlying clavicle
may be routinely identified in greater than 60% of individu- (22%) (22). The mean diameter of missed lesions in this
als (13,14). Similarly, McWilliams et al. (13), in a lung can- study proved to be 1.9 cm, considerably larger than
cer CT screening study that compared the results of a total of on CT.
332 baseline CT scans performed with an SD CT scanner Reasons for missing nodules on CT include both detec-
using 7-mm collimation with 229 studies performed either tion errors and misinterpretation. Causes of detection
with a 4 or 8 MDCT scanner using 1.25-mm collimation, errors include factors related to scan technique (use of thin
found that abnormal scans were reported twice as frequently vs. thick sections) (23–25), contiguous versus overlapping
with MDCT (60 vs. 36%, respectively, p 0.001). sections (9,11), or data compression (26), for example;
Despite well-documented improvements in nodule factors related to nodule characteristics (e.g., nodule
detection with MDCT, paradoxically, missed nodules on density, location, relation to adjacent structures, and edge
A, B C
Figure 6-1 Improved nodule detection with multidetector computed tomography (MDCT). A: Target reconstructed 1-mm image through
the left mid lung shows a subtle central lung nodule (arrow). This lesion would easily have been overlooked with routine thick-section axial
images. B, C: Target reconstructed images at the same level as shown in A; several months later the presence of a malignant cavitary nodule
is confirmed (arrows in B and C), in this case from squamous cell carcinoma of the head and neck.
Chapter 6: Focal Lung Disease 559
A, B C, D
Figure 6-2 Missed lung cancer. A, B: Magnified contiguous views through the medial aspect of the right upper lobe shows a tiny, poorly
marginated opacity barely discernible, not initially identified. C, D: Magnified views at the same level as A and B obtained approximately 3
months later show the presence of a rapidly growing lung nodule (arrow in C). Pathologically confirmed stage 1A lung cancer.
characteristics) (27); and factors related to interobserver nodules measuring 3 mm, 5 mm, and 7 mm of 37%, 62%,
variability (20). In this latter category are “missed” nodules and 80%, respectively.
due to interpretative errors that occur because of differences In addition to limitations related to technique and nod-
among individual readers’ thresholds for dismissing abnor- ule characteristics, problems with interobserver variability
malities as “trivial” (19). have also long been noted. In one early study, for example,
This wide variety of causes is reflected in numerous pub- the value of CT in staging 202 patients with Wilms’ tumors
lished reports. Swensen et al. (15), in an evaluation of and normal CXRs was assessed by three independent
annual incidence screening CT scans, found that, in 26% of observers (28). Although the three observers identified a
patients, missed nodules could be retrospectively identified total of 78 nodules, only 19 (24%) of these nodules were
(Fig. 6-2). Similar findings also have been reported by identified by all three readers. Of still greater concern, only
Heinschke et al. (18) for missed nodules on low-dose 14 of the 202 patients actually developed documented
screening CT studies. Kakinuma et al. (19), in an evaluation pulmonary disease. Of these 14, only 5 had nodules iden-
of missed nodules in a screened population, demonstrated tified by all three readers; furthermore, in five other cases,
seven overlooked nodules that proved to lie adjacent to CT studies were interpreted as normal by all three readers.
regions of the lung scarred by prior inflammation (n 2), Although differences in the level of reader expertise have
to have ground-glass attenuation (n 3), or to be located often been cited to account for interobserver variability, this
adjacent to pulmonary vessels (n 2). White et al. (17), in factor alone cannot account for the wide differences in
a retrospective review, demonstrated that nodules were readers’ interpretations. In a retrospective evaluation of 116
often missed either when endobronchial or when located nodules comparing contiguous versus overlapping 8-mm
in the lower lobes. In the most detailed evaluation of the sections using four reviewers of varying interpretative skill,
causes of missed nodules to date, Rusinek et al. (27) for example, Buckley et al. (10) showed that, although over-
demonstrated reduced sensitivity for detecting simulated lapping 8-mm sections significantly improved sensitivity
nodules on LDCT scans when central versus peripheral (p 0.055) and decreased the number of false-positive
(58% vs. 74%, for nodules with and without vessels attach- interpretations, nonetheless a total of 58 false-positive
ments, respectively). Not surprisingly, smaller nodules studies were still recorded by even the most experienced
also proved more difficult to identify, with sensitivities for interpreters.
More recently, in a comparison of the sensitivity of MIPs with the number identified by routine contiguous axial
single versus double readings of standard (SDCT) and images, Diederich et al. (29) showed that MIPs using a slab
LDCT MDCT studies, of 390 nodules identified in nine thickness of 15 mm outperformed 5-mm axial image evalua-
patients with documented pulmonary metastases, the tion as well as improved readers’ confidence in identifying
average sensitivity for detection of nodules significantly “definite” nodules. As documented by Gruden et al. (30),
increased with double readings (63% and 64% vs. 74% the addition of 8-mm axial MIP slabs reconstructed from
and 79% for standard-dose CT and LDCT studies, respec- 3.75-mm axial CT sections reconstructed at 3-mm intervals
tively) (20). As important, only 47% and 52% of nodules led to a significant improvement in the detection of central
identified with LDCT and standard-dose CT, respectively, nodules, in particular, for all readers regardless of their level
were concordantly identified by all readers. of expertise, and the detection of even peripheral nodules by
In an attempt to improve nodule detection, interest less experienced readers. These data support the idea that
has focused on the use of a number of alternative image pro- newer methods of interrogating large datasets will be requi-
cessing techniques. The best documented of these makes use site to make optimal use of the volume of information now
of contiguous or sliding maximum intensity projection routinely available with MDCT. In this regard, as discussed
(MIP) images reconstructed from sequential 1- to 3-mm in greater detail later in this chapter, considerable interest has
source images (Fig. 6-3) (29,30). Unlike routine cine view- focused on the use of CAD in detecting lung nodules (34).
ing (31,32), sliding MIPs take advantage of high-resolution
data without requiring readers to scroll through hundreds of
thin-section axial images as may routinely be reconstructed,
especially with newer MDCT scanners (33). In a study The potential for magnetic resonance (MR) to detect
comparing the number of pulmonary nodules identified by pulmonary nodules is limited because of inferior spatial
Figure 6-3 Advanced image pro-

cessing: maximum intensity projec-
tion (MIP) images and volumetric
renderings. A, B: Magnified 5-mm
MIPs and volumetrically rendered
coronal images through the right mid
lung, respectively, show a nodule in
the right lower lobe to good advan-
tage. Compared with routine axial
images, MIPs and volumetric ren-
dered images allow rapid interro-
gation of large datasets routinely
A, B acquired with MDCT scanners.
resolution as compared with CT. Nonetheless, MR does Clinical Evaluation

present some advantages that in select cases may allow
identification of lesions with more certainty than CT. As In the evaluation of pulmonary nodules, a number of
noted by Müller et al. (35), in an early study of 25 patients clinical and historical data should be considered (43). In
evaluated with both CT and MR using T2-weighted addition to age and smoking history, these include occu-
sequences, whereas CT allowed more precise identification pational exposure, especially to asbestos, as well as the
of small nodules, MR was superior for detecting lesions presence of coexisting parenchymal disease, for example,
adjacent to blood vessels. Feuerstein et al. (36), in a idiopathic pulmonary fibrosis (IPF); a known prior
prospective study comparing CT with MRI in the evaluation malignancy; geographic considerations, including travel
of 12 cases of suspected pulmonary metastases with history; and results of diagnostic tests, especially skin
surgical correlation, found that MR and CT were equivalent testing (1–3).
for classifying patients as having positive or negative status Of these, probably the most reliable is assessment of
for metastatic disease: surprisingly, this also proved true for age. SPNs are almost invariably benign in patients younger
individual nodules as well. Given the greater contrast than 30 years. Taylor et al. (44) resected 81 nodules in
resolution of MR, it is perhaps not surprising that MR patients aged 20 to 29 years and found 1 bronchogenic
would perform so well, especially in the detection of carcinoma, 1 “bronchial adenoma,” and 79 benign lesions,
central or perihilar nodules (37). As reported by Doppman including 67 granulomas. Thus, unless there is a known
et al. (38) in an early study, MR was shown to be of primary malignancy, solitary lesions in patients younger
value in detecting small, centrally located adrenocorti- than 30 years may be presumed to be benign, although
cotrophic hormone–producing bronchial carcinoid tumors. definitive diagnosis may still be requisite. In distinction,
More recently, Schroderer et al. (39), comparing a two- pulmonary malignancies should not be considered rare
dimensional (2D) half-Fourier single shot turbo spin-echo in the 30- to 39-year-old age group. Among 39 lesions
(HASTE) MR sequence with MDCT as a gold standard to resected in this age group, Davis et al. (45), found 9 bron-
evaluate 30 patients with known pulmonary metastases, chogenic carcinomas. At age 40 and beyond, there is greater
reported MR sensitivity of 73% for lesions less than 3 mm, probability of malignancy in any newly discovered SPN.
86.3% for lesions between 3 and 5 mm, and 95.7% for Given that nearly 90% of these lesions may prove
lesions between 6 and 10 mm. Despite these findings, given resectable, early diagnosis remains essential.
the advantages of greater spatial resolution and the ability SPNs are a common finding on routine CXRs in adults.
to detect calcification, it is likely that MDCT will remain Although traditionally the vast majority of such inciden-
the method of choice for detecting pulmonary nodules tally detected lesions are benign (46,47), lung cancer,
for the foreseeable future. particularly early treatable lesions, often manifests as a
peripheral pulmonary nodule, especially in asymptomatic
patients (41). In an extensive review of 1,267 lung cancers
CHARACTERIZATION by Theros (48), 507, or 40% of cases, initially presented as
peripheral lung masses. Not surprisingly, considerably
Although differentiating benign from malignant nodules higher prevalence of malignant nodules has been reported
remains problematic, recent improvements in morphologic in clinical studies of patients specifically referred for nod-
classification, contrast-enhanced densitometry, techniques ule evaluation. In this setting, estimates of the prevalence
for more accurate assessment of nodule growth, and of cancer may be as high as 70%, presumably reflecting
especially use of FDG-PET scanning have redefined our differences due to referral bias (49).
approach to the problem of the indeterminate lung nodule. Nearly as important as age as a predictor of malignancy
is smoking history. Stitik and Tockman (50), in an early
surveillance study of male smokers, found that 72% of
Definition
the cancers detected in asymptomatic subjects were
Various authors have applied a wide range of restrictions peripheral nodules. It should be noted that these data do
on what constitutes a solitary nodule. However, any not take into account the recently documented increase
patient whose CXR shows a single rounded or ovoid lesion in the prevalence of adenocarcinoma, in particular, devel-
in the lung parenchyma less than 3 cm, which is not asso- oping in women as a consequence of changed cigarette
ciated with atelectasis or pneumonia, must be considered smoking demographics (51). The use of chest radiogra-
to have a solitary pulmonary nodule (SPN). The term coin phy for evaluating asymptomatic smokers has been most
lesion was suggested by O’Brien et al. (40) to describe recently assessed in a large-scale screening study—the
a focal mass. The term enjoyed a short-lived popularity but Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer
has since been disparaged because the description implies Screening Trial (52). Altogether, a total of 77,465 individ-
a flattened object instead of a spherical mass. SPN should uals aged 55 to 74 years without prior history of malig-
be considered the most appropriate designated name for nancy (including smokers, prior smokers, and never
such lesions (41,42). smokers) were evaluated in this study with a single-view
posteroanterior (PA) CXR. Of these, 5,991 (8.9%) were Density

interpreted as suspicious for cancer (9.6% for men; 8.2%
for women), with the highest rates recorded for older Optimal evaluation of nodules requires determination of
individuals and smokers (52). Of these, 206 (3.4%) sub- nodule density, including differentiating between subsolid
sequently underwent biopsy, with 126 (61%) shown to and solid nodules, as well as identifying the presence and
have lung cancer, 44% of which were stage 1 NSCLCs. pattern of calcifications, fat, air, and fluid (Table 6-1).
Although, overall, 1.9 lung cancers were detected per
1,000 screens, the rate clearly reflected smoking status, Subsolid Versus Solid Lung Nodules
with current smokers, former smokers, and nonsmokers Largely the result of early lung cancer screening trials cou-
having rates of 6.3, 4.9, and 0.4 lung cancers detected per pled with the widespread availability of MDCT scanners
1,000 screens, respectively. In all, nonsmokers accounted capable of providing contiguous high-resolution images
for 11% of lung cancers detected, including 12 adenocar- throughout the thorax in a single breath-hold period, it is
cinomas and 2 carcinoid tumors. All save 2 of these were now appreciated that noncalcified pulmonary nodules are
identified in women. As discussed in depth in Chapter 7, best characterized as either subsolid or solid in appearance
it should be noted that, although most lesions in (55). Subsolid nodules may be further divided into pure
this study proved to be potentially resectable, the value ground-glass nodules (focal densities in which underlying
of screening (both radiographic and CT) remains to be lung morphology is preserved) and mixed solid/ground-
determined. glass nodules.
Patients with a history of extrathoracic malignancies To date, numerous reports have shown that there is a
clearly pose a particular diagnostic problem, especially close but imperfect correlation between pure ground-glass
given the prevalence of small, less than 10-mm, nodules in lesions and either atypical adenomatous hyperplasia (AAH)
the general public. These are difficult to perform biopsy on
percutaneously and may be difficult to palpate at surgery,
especially when central. In this context, clinical correlation
can be of value. The propensity of certain tumors to metas- TABLE 6-1
tasize to the lungs, in particular, varies and should be COMPUTED TOMOGRAPHY DENSITOMETRY
taken into account. Gilbert and Kagan (55), for example, IN EVALUATION OF FOCAL LUNG DISEASE
have studied the estimated incidence of lung metastases
for various primary neoplasms at presentation and autopsy. Calcification
In their study, those tumors most likely to metastasize to Common
Calcified granulomas (central or diffuse)
the lung included choriocarcinoma, renal cell carcinoma, Hamartomas (central, diffuse, popcorn)
and bone and soft tissue sarcomas, in particular. A similar Carcinoid tumors (punctate, eccentric)
propensity for sarcomas, as compared with carcinomas, to Uncommon
metastasize to the lungs has been documented by Peuchot Bronchogenic carcinomas (punctate, eccentric)
and Libshitz (54). Metastases (diffuse, punctate)
Rare
Mucoid impaction (branching)
Morphologic Evaluation Sclerosing hemangiomas (punctate, eccentric)
Fat
By virtue of unobstructed visualization of the lungs and Common
improved contrast resolution compared with routine Hamartomas ( ) calcification
CXRs, CT not only allows exquisite definition of the den- Uncommon
sity, shape, and edge characteristics of focal lung lesions Exogenous lipoid pneumonia
but also enables assessment of the relationship between Contrast Enhancement
lesions and surrounding structures, including airways, ves- Common
sels, and pleural surfaces. As a consequence, CT has Bronchogenic carcinomas
Carcinoid tumors
become the standard method for initial characterization Arteriovenous malformations
of lesions either suspected or seen on corresponding
Fluid
CXRs. Acknowledging that optimal evaluation of the
Common
malignant versus benign potential of focal lesions requires Lung abscesses (bacterial and fungal)
assessing all pertinent imaging features in concert, for the Mucoid impaction
purposes of this chapter the following features pertaining Rare
to nodule morphology are discussed independently: nod- Intrapulmonary bronchogenic cyst
ule density, edge characteristics, size, and location. Subse- Iodine
quent sections focus on the use of both CT and MR Amiodarone toxicity
contrast enhancement, FDG-PET imaging, and techniques From Naidich DP, Garay SM. Radiographic evaluation of focal lung
for measuring nodule growth. disease. Clin Chest Med. 1991;12:77–95, with permission.
Premalignant AAH
A localized BAC Slow

growth Ground-
B localized BAC with VDT glass
foci of structural >1 yr
collapse
C localized BAC, Inter- Mixed
active fibroblastic mediate
proliferation Figure 6-4 Adenocarcinoma: CT-pathologic cor-
relations. This figure shows the relationship between
D,E,F poorly differentiated Fastest Predom- the Noguchi classification of primary adenocarcino-
tubular, papillary mas of the lung and the corresponding CT appear-
VDT inantly
adenocarcinoma ances of these lesions. AAH, atypical adenomatous
<1 yr
solid hyperplasia; BAC, bronchoalveolar cell carcinoma.
Note that as lesions become more invasive, their
doubling times become shorter, and their CT appea-
rance changes from pure ground-glass opacities to
mixed ground-glass/solid lesions and predominantly
Pure BAC — 100% solid lesions, respectively. As noted in the figure,
Mixed BAC/Invasive — 75% pure BAC has a 100% 5-year survival, decreasing to
Pure Invasive — 52% 52% for purely invasive adenocarcinomas.
or true bronchoalveolar cell carcinoma (BAC), correspond- Mirtcheva et al. (61), correlating CT with pathologic
ing to Noguchi classification types A and B (Fig. 6-4) findings in 29 patients (15 with Noguchi type A and B
(56–64). As discussed in greater detail in Chapter 7, tumors, 14 with type C tumors), also reported close corre-
Noguchi et al. (65) proposed a new classification of adeno- lation between the appearance of lesions on CT and their
carcinomas based on a review of pathologic features of 236 pathologic appearance. Specifically using a classification of
cases of adenocarcinomas less than 2 cm. Briefly, types ground-glass opacities suggested by Gaeta et al. (69), these
A and B correspond to true BACs characterized by pure authors noted the finding of lesions appearing as pure
lepidic growth in the absence of active fibroblastic prolifera- ground-glass opacities, with or without air bronchograms
tion, whereas type C corresponds to a mixed-type adenocar- or with surrounding ground-glass “halos,” were predictive
cinoma [as defined by the new World Health Organization of type A and B BACs, whereas mixed solid/ground-glass
(WHO) classification] (66), defined as having bronchoalve- lesions (Fig. 6-7) or those with superimposed reticulation
olar features as well as active fibroblastic proliferation. As (Fig. 6-8) were highly predictive of invasive (type C) ade-
originally defined, types A, B, and C together were referred nocarcinomas (61).
to as “replacement tumors” and ostensibly represented Not surprisingly, these authors and others have also
progression from truly localized tumors to a more invasive reported good correlation between the extent of ground-
form. Types D to F, in distinction, represent progressively glass attenuation and prognosis (Figs. 6-6 to 6-10)
more invasive, poorly differentiated adenocarcinomas with- (70–72). Takashima et al. (72), for example, in a study
out BAC features and likely arose de novo (65,67). Using this correlating lesion size, percentage of ground-glass opacity,
classification, Noguchi found significant differences in and air bronchograms with clinical data and pathologic
prognosis, with types A and B having a 100% 5-year sur- findings in 52 consecutive patients with adenocarcinoma
vival; type C tumors, with a higher percentage of nodal and with BAC components less than 3 cm, documented a
distal metastases, having a 75% 5-year survival; and types D significant difference in prognosis based on the extent
to F having a still worse prognosis, with 5-year survivals of of ground-glass attenuation (p 0.043) as well as the
approximately 50%. presence of an air bronchogram within the tumor
To date, numerous studies have positively correlated CT (p 0.003).
findings with Noguchi’s classification. Kim et al. (67a), for Although more controversial, several reports to date
example, in a study of 224 patients (132 with BAC, 92 have shown that serial CT studies are of value by confirm-
with adenocarcinoma), showed that the extent of ground- ing that there is progression of disease from pure ground-
glass opacity identified by CT was greater in BAC than in glass opacities to more complex mixed solid/ground-glass
other adenocarcinomas (p 0.001). Small ground-glass lesions (Fig. 6-9) (61,73). In one study evaluating
lesions, in particular, are likely to represent either foci of serial CT findings in 73 patients with a spectrum of
AAH or pure BACs (Figs. 6-5 and 6-6). Nakata et al. (68) adenocarcinomas, a significant linear trend was noted
showed that only 7% of tumors less than 1 cm with a pure confirming progression from AAH, through type A and B,
ground-glass pattern proved to be mixed adenocarcinomas; to type C adenocarcinomas, with lesions first presenting
the remaining were all BACs. as focal ground-glass opacities, subsequently increasing
A, B
Figure 6-5 Adenocarcinoma: CT-pathologic correlations. A, B: Magnified 1-mm sections through the right upper lobe in two different
patients show characteristic appearance of a pure ground-glass opacity. The former is presumed to be a focus of atypical adenomatous hyper-
plasia (AAH) (arrows in A), whereas the latter is one of many lesions identified in a patient with documented multifocal bronchoalveolar cell
carcinoma (BAC) (arrows in B). Not surprisingly, these lesions are generally only identifiable when contiguous high-resolution 0.75- to 1-mm
images are acquired. Distinction between AAH and pure BAC based solely on their CT appearances is problematic and is even difficult when
evaluated histologically (Fig. 6-6). In general, lesions less than 5 mm are assumed to represent AAH, whereas lesions greater than 1 cm almost
always represent BAC. Lesions intermediate in size are indeterminate and require close monitoring, frequently over several years.
Figure 6-6 The spectrum of ade-

nocarcinomas of the lung: pathol-
ogic appearances. See Color Figure
6-6A–F. Low- and high-powered mag-
nified views show characteristic ap-
pearance of atypical adenomatous
hyperplasia (AAH) (A and B), muci-
nous bronchoalveolar cell carcinoma
(BAC) (C and D), and nonmucinous
BAC (E and F), respectively. It is
apparent why accurate identification
of these lesions requires the entire
specimen to be histologically evalu-
ated, making both transbronchial
and transthoracic needle aspiration
or biopsy of only limited value.
(Courtesy of William Travis, MD,
Memorial Sloan Kettering Hospital,
A, B New York, New York.)
C D
E F
Figure 6-7 Adenocarcinoma: CT-

pathologic correlations. (A and B)
Shown are 5-mm and 1-mm magni-
fied views of two different patients,
respectively, illustrating the charac-
teristic appearance of mixed
ground-glass/solid lesions (arrows).
This appearance typically correl-
ates well with Noguchi type C bron-
choalveolar cell carcinoma (BAC)
or World Health Organization
(WHO) classification invasive mixed
A, B adenocarcinoma.
A, B C
D, E F
Figure 6-8 Adenocarcinoma: CT-pathologic correlations. A–F: Sequential magnified 1-mm sections through the medial aspect of the
superior segment of the right lower lobe show characteristic appearance of a mixed ground-glass/solid lesion with so-called reticular
features (arrows in F), appearing as a fine linear mesh within which vessels and airways can still be identified. This appearance typically cor-
relates with at least a Noguchi type C bronchoalveolar carcinoma.
Figure 6-9 Adenocarcinoma: dis-

ease progression. A: Magnified
1-mm section through the medial
aspect of the right lung shows two
discrete ground-glass opacities in
this patient with documented multi-
focal bronchoalveolar cell carci-
noma. B: Magnified view at the
same level as shown in A, 6 months
later, shows clear progression of
disease with both lesions enlarging,
with evidence of a solid compo-
nent now apparent in the more
posterior lesion. Progression from
pure ground-glass lesions to mixed
ground-glass/solid lesions is consis-
tent with evolution from pure bron-
choalveolar cell carcinoma (BAC) to
A, B more invasive adenocarcinoma.
in size and density with the appearance of solid Calcification

components (74). It is well documented that the radiographic presence of cal-
It should be emphasized that, although pure ground- cification within a pulmonary lesion is generally a reliable
glass opacities are likely to represent either AAH or BAC, sign that a lesion is benign (46,47). The best documenta-
not all such lesions prove malignant (Fig. 6-11). In one tion of the nature of calcification within pulmonary
study of 10 pure ground-glass opacities followed by serial nodules was carried out at the Mayo Clinic and reported in
CT examinations for a median period of 32 months a series of articles in the 1950s (75–78). In addition to con-
and shown to enlarge, 3 proved to be due to lymphopro- firming a strong correlation between radiographic and/or
liferative disease and 1 proved to be due to focal fibrosis tomographic evidence of calcification and benign disease,
(60). Similarly, the finding of a mixed part solid–part these authors also defined characteristic patterns of calcifi-
ground-glass or solid lesion, although more likely to rep- cation. Based on specimen radiography obtained on a total
resent a mixed adenocarcinoma, may in fact represent of 207 solitary pulmonary masses, four different patterns of
virtually any type of lung cancer, including small cell lung benign calcification were identified (Fig. 6-12): (a) a lami-
cancer. nated pattern with calcium deposited in concentric layers,
A B
Figure 6-10 Adenocarcinoma: correlation of CT appearance and prognosis. A, B: Magnified view through the medial aspect of the left
upper lobe shows a predominantly solid lesion with only a minimal ground-glass component. This appearance correlates with a poorer prog-
nosis when compared with adenocarcinomas appearing either as pure ground-glass or mixed ground-glass/solid lesions, with a greater
propensity for nodal and distant metastases to be present at the time of diagnosis.
A B
Figure 6-11 Focal inflammation

simulating adenocarcinoma. A, B:
Sequential magnified 1-mm images
through the right upper lobe show a
characteristic-appearing ground-
glass nodule initially diagnosed as
consistent with probable bron-
choalveolar cell carcinoma (BAC).
C, D: Sequential magnified 1-mm
images through the right upper
lobe at the same level as shown in
A and B, obtained 3 months later,
show near complete resolution of
this lesion, now presumed to repre-
sent focal nonspecific inflammation.
Although there is extremely close
correlation between subsolid nod-
ules and adenocarcinoma (Figs. 6-6
to 6-10), the diagnosis of malignancy
should not be assumed unless le-
sions remain stable for at least 3 to 6
C, D months.
between rings of fibrous tissue; (b) a dense central calcified be emphasized that for small nodules, particularly those
nidus; (c) diffuse and/or irregular or nodular (so-called that are 1 cm or less, the commonest CT finding is diffuse
popcorn) calcifications; and (d) eccentric, single or multi- homogeneous calcification.
ple punctuate calcifications. With minor variation, these The concept of using CT to identify otherwise radi-
guidelines for pattern recognition remain in effect. Al- ographically occult calcifications within pulmonary nodules
though the four patterns cited adequately categorize the was first published by Siegelman et al., in 1980 (79). In
range of heterogeneous calcifications encountered, it should this initial series, 91 patients without radiographic or
requires that contiguous images through nodules be ob-

Calcifications tained with slice collimation less than half the diameter of
the nodule to eliminate partial volume averaging, as even
with the use of MDCT only a few sections will truly be rep-
resentative. For small lesions, attenuation values will be
overestimated by an edge-enhancing (sharp) algorithm
and underestimated by a smoothing algorithm (80).
Benign Selection of a high-spatial-frequency reconstruction algo-
rithm, in particular, may artificially create the erroneous
impression of calcification within lesions, especially along
their edge (82). As discussed in the section on contrast
enhancement, this same logic applies to the timed evalua-
tion of nodules following administration of intravenous
contrast media. This technical consideration should not
obscure the fact that 1-mm sections frequently disclose
the presence of calcification when thicker 7- to 10-mm
Malignant sections do not.
Primary lung malignancies may contain areas of calcifi-
cation or ossification (Fig. 6-14). Theros (48), for example,
Figure 6-12 Patterns of calcification. Central, diffuse, laminated found 7 examples of radiologically detected calcification
or curvilinear, and popcornlike calcification patterns are charac- in 1,267 lung cancers, a prevalence of approximately 1%.
teristic of benign nodules, whereas eccentric or multiple coarse
calcifications are more often seen in malignant nodules. O’Keefe et al. (78), using radiographs of surgical lung speci-
mens, found calcification in 8 (15%) of 52 resected primary
lung cancers. With the addition of CT, the likelihood of
tomographic evidence of calcification were evaluated with detecting calcification within tumors has increased (83,84).
thin-section CT and assigned a representative CT number As documented in one study, 53 (10.6%) of 500 consecutive
based on an average of continuous pixels measured within patients with provisional diagnoses of lung cancer were
the estimated center of the lesion. These authors reported found to have calcification on CT (83). Calcification within
that although all 45 primary neoplasms and 13 metastases tumors may result for a number of reasons. These include
had representative CT numbers less than 164 Hounsfield (a) tumor engulfing previous calcification, as may occur
units (HU), 20 of 33 benign lesions proved to have repre- with a preexisting granuloma; (b) dystrophic calcification,
sentative CT numbers greater than 164 HU. If the correct di- occurring in areas of tumor necrosis; and (c) primary tumor
agnosis of a benign lesion is considered a true positive and calcification, as may be seen in particular within mucinous
the designation of a malignancy as benign is considered a adenocarcinomas. Radiographically, calcification within
false positive, then the overall sensitivity of CT in this initial lung neoplasms typically appears stippled or eccentric and
study proved to be 59%, with a sensitivity of 73% for lesions usually occurs in large central lesions (78). In their study of
1 cm or less in diameter and a specificity of 100% (79). 53 patients with histologically verified calcification com-
The density of a lesion, as measured by the degree to pared to a control group of 15 patients, Grewal and Austin
which it attenuates a collimated beam of x-rays, is an (84a) recorded a mean diameter of 6.2 3 cm for calcified
innate property of its tissue content. However, the determi- neoplasms versus 4.4 2.3 cm for noncalcified tumors,
nation of a reliable measurement in HU is subject to with 33 (85%) of the calcified tumors measuring more than
extensive variation. Measurements vary as a function of 3 cm. In a smaller study, Mahoney et al. (84) also noted
partial volume averaging, the reconstruction algorithm, a propensity for calcification to occur in larger lesions: of
true slice thickness, the size of the lesion, kilovoltage, and 20 lesions, 15 (75%) were 5 cm or greater, whereas only
beam-hardening artifacts (80). The use of CT to evaluate 3 were 2 cm or smaller. Although a number of different pat-
nodules is now well established, as problems related to terns of calcification have been noted in these studies,
standardization of tissue attenuation measurements have including peripheral, central, diffuse, and amorphous,
been notably lessened, in particular, by algorithms to cor- to date no study has shown any consistent correlation
rect beam-hardening artifacts (81). It is generally held that between patterns of calcification and histologic subtype.
in the majority of cases, calcified lesions are readily recog- Indeed, virtually all primary lung cancers, including small
nized by “obvious” high-attenuation values without the cell carcinoma, may have areas of calcification identified on
use of specific HU thresholds or the need for a reference CT scans (83–87).
phantom. As a consequence, it is mandatory that studies In addition to bronchogenic carcinoma, calcification
be performed with close attention to scan technique has been identified in other primary lung tumors (88). By
(Fig. 6-13). Although thin-section CT suffices in the major- far the commonest of these are central carcinoid tumors
ity of cases, in questionable instances, an adequate study (Fig. 6-15) (83,87,89,90). Classified as part of the spectrum
A, B
C, D
E, F Figure 6-13 CT densitometry: 1-mm versus

7-mm section. A–C: Sequential 7-mm helical
sections through the left upper lobe recon-
structed every 6 mm show a small nodule in
the left upper lobe. D, Identical section as
in B, imaged with narrow windows, shows
no obvious calcifications within the nodule.
E–G: Sequential 1-mm helical images
through the same portion of the left upper
lobe as in A–C again show a small peripheral
lung nodule. H: Identical image as in F,
imaged with narrow windows, clearly shows
this lesion to be densely calcified. This case
illustrates that 1-mm sections are often nec-
essary to perform adequate CT densitome-
try. Helical acquisition is especially helpful for
evaluating very small nodules, as it markedly
simplifies acquisition of sections through the
G, H center of lesions.
Figure 6-14 Tumoral calcification.

A, B: Magnified views through the
right upper lobe imaged with wide
and narrow windows, respectively,
show a spiculated irregular peri-
pheral lesion within which punctate
calcifications can be identified. Note
that in this case there is also evi-
dence of enlarged mediastinal
nodes. The presence of calcifica-
tions per se is not a reliable sign of
A, B benign disease.
A B
C D
Figure 6-15 Carcinoid tumor. A–D: Sequential 1-mm sections show a well-defined central mass in the lower portion of the right hilum
causing marked compression of the basilar segmental airways (arrows in A, B, and D). E–H: Identical sections as shown in A–D, imaged with
mediastinal windows, show the presence of peripheral calcification. This constellation of findings is characteristic of a carcinoid tumor,
subsequently verified surgically (continued ).
E F
G H
of neuroendocrine lung neoplasms, carcinoids are further central lesions but was seen in only 1 (8%) peripheral
characterized as either typical or atypical (91). These lesion. In a similar study of 12 pulmonary carcinoids,
tumors tend to exhibit variable growth patterns. Most arise Magid et al. (92) also noted a propensity for central tumors
as endobronchial obstructing lesions, resulting on occasion to preferentially calcify. In our judgment, the triad of a well-
in an appearance of focal mucoid impaction; alternatively, defined central lesion abutting or narrowing an adjacent
they may extend extraluminally, resulting in so-called ice- bronchus with eccentric areas of calcification is all but
berg lesions (Fig. 6-16). Unlike typical carcinoids, atypical pathognomonic of this diagnosis (93). The one exception
carcinoids show a marked propensity to metastasize early, is the rare case of a central, endobronchial hamartoma (see
especially to regional lymph nodes. A typical triad of fea- Fig. 5-25).
tures has been described for central carcinoid tumors and Despite the fact that most calcified parenchymal
includes a well-defined round or slightly lobulated lesion neoplasms are identifiable as such with CT, it should be
that abuts or narrows a central airway in which eccentric emphasized that malignant lesions may still mimic the
calcifications can be identified (Fig. 6-16). In addition, appearance of benign calcified granulomas. In addition to
these lesions are characteristically extremely vascular and the well-described finding of calcified metastases from
therefore show marked contrast enhancement following osteogenic sarcomas and chondrosarcomas, adenocarcino-
intravenous contrast administration. Although accurate mas, both primary and metastatic, not infrequently are
preoperative evaluation requires precise definition of both calcified. Review of previous published studies confirms that
the intraluminal and the extraluminal components of these primary and metastatic adenocarcinoma remains the lesion
tumors, recognition of the vascular nature of these lesions most likely to be misinterpreted as benign (85–87,94).
is also of obvious value, especially prior to bronchoscopy, Although routine axial CT images are now the accepted
given the propensity of these lesions to hemorrhage, some- gold standard for evaluating the presence of calcifications
times massively, when biopsy is performed. within nodules, it is worth noting that two alternative
In a retrospective study of 31 patients with documented approaches have also been advocated. The first involves
carcinoids (27 typical, 4 atypical), Zweibel et al. (89) found the use of dual kVp to analyze nodules as an alternate
that 18 (58%) were central, whereas 13 (42%) were periph- method for detecting calcium (95). Although the necessity
eral in location. Calcification was identified in 7 (39%) for images to be obtained at precisely the same level for
comparison purposes has, to date, proved a major limita- reported series with data on histology, 37 of 68 (54%) of
tion to this technique (96), the recent introduction of a the lesions contain fat (101–103). Calcification is more
dual source CT scanner raises the distinct possibility that common in large lesions, occurring in less than 10% of
this approach may supersede traditional CT analysis. The lesions smaller than 2 cm, one third of tumors 3 to 4 cm
second involves the use of dual-energy digital subtraction in diameter, and 75% of lesions larger than 5 cm (104).
radiography (97). Briefly, dual-subtraction chest radiogra- Cartilage usually predominates, but occasionally fat or
phy makes use of the fact that structures containing myxomatous connective tissue may be most prominent.
calcium selectively attenuate lower energy photons in the Hamartomas may be diagnosed when they are manifest
x-ray beam, allowing identification of calcified nodules, on CT as lesions that contain fat, or fat plus calcium, and
otherwise indistinguishable from noncalcified nodules have a smooth contour and sharp outline (87,98). In a
by routine radiographic techniques. Although the use of report of 47 lesions studied by CT, 30 (63.8%) contained
this technique is not currently widespread, commercially fat only (n 18), calcium and fat (n 10), or calcium
available digital radiographic systems are now in use, alone (n 2) (98). The remaining lesions were diagnosed
making this a potentially attractive alternative to CT. as indeterminate and managed with surgery or lung biopsy
(98). In our experience, lung cancers never manifest as
Fat smoothly marginated lesions containing fat. Use of CT to
The presence of fat within the parenchyma is indicative definitively diagnose pulmonary hamartomas requires the
of one of two diagnoses: hamartoma (Figs. 6-17 to 6-19) same attention to scan technique as discussed previously
or exogenous lipoid pneumonia (Figs. 6-20 and 6-21). pertaining to CT identification of calcifications: namely,
Hamartomas, the third commonest cause of a SPN, are that contiguous (and/or overlapping) thin sections less
considered benign neoplasms that originate in fibrous con- than half the diameter of the nodule be acquired recon-
nective tissue beneath the mucous membrane of the structed with a soft tissue algorithm to obviate potential
bronchial wall (98,99). These lesions contain mixtures of misdiagnoses due to partial volume averaging either from
myxomatous connective tissue, cartilage, epithelial-lined adjacent lung parenchyma or from subtle cavitation result-
clefts, and variable amounts of fat, smooth muscle, marrow, ing in the spurious appearance of HU units in the range
and bone (100). Fat is a key element; if one combines three of fat.
A B
Figure 6-16 Hilar carcinoid: evaluation with contrast enhancement. A, B: Enlargement of 3-mm section acquired helically through the left
hilum, imaged with wide and narrow windows, respectively, 1 minute after the administration of intravenous contrast media shows a left
hilar mass causing slight compression of the adjacent bronchus (arrow in A). Note that with mediastinal windowing this mass has the same
density as the adjacent left pulmonary artery and aorta, measuring 163 HU, indicative of an extremely vascular lesion. This degree of
enhancement within a central mass abutting a bronchus is characteristic of a carcinoid tumor, subsequently verified at surgery. C, D:
Magnified images from a different patient than shown in A and B, first without (C) and then following a bolus of intravenous contrast media
(D), again demonstrate that carcinoid tumors are highly vascular and enhance following intravenous contrast administration. (A, B from
Costello P, Naidich DP. Protocols for helical CT of the chest. In: Silverman PM, ed. Helical (spiral) computed tomography. A practical
approach to clinical protocols. Philadelphia: Lippincott–Raven; 1998:65–101, with permission.) (continued )
C D
It is worth emphasizing that, despite their benign nature, sema (Fig. 6-23). Less commonly, central endobronchial
hamartomas grow, albeit slowly. Bateson and Abbott (104), tumors may present with focal mucoid impaction. In these
for example, noted slow growth in 40 of 45 patients with cases, the use of a bolus of intravenous contrast may be of
available serial chest roentgenograms. In one study of 11 help in differentiating the central enhancing tumor from
cases that exhibited enlargement, recognizable changes were peripheral fluid-filled airways. Care should be taken to
present only in those studied for at least 3 years (105). ensure that tumor has not invaded the airways directly
by measuring attenuation values within these areas both
Fluid prior to and following the injection of contrast media.
A number of focal parenchymal lesions are distinguished Alternatively, differentiation of tumor from fluid-filled
by the presence of fluid. Included in this group are lung airways may be obtained with T2-weighted MR images
abscesses (Fig. 6-22), mucoid impaction of the airways (Fig. 6-25) (107,108).
(Fig. 6-23), and, rarely, intrapulmonary bronchogenic cysts
(106) (Table 6-1) (Fig. 6-24). Even when surrounded by ex- Air
tensive consolidation or within atelectatic lung, abscesses CT is exquisitely sensitive to the presence of air, and it is
characteristically appear well defined and smooth walled not surprising that subtle cavitation within nodules is
prior to the development of bronchial or pleural communi- easily identified when otherwise radiographically occult
cations following the administration of intravenous con- (Figs. 6-26 to 6-28). In addition to assessing the relation-
trast media and consequently are generally easily differenti- ship between cavities and airways and surrounding
ated from tumors with CT. Similarly, mucoid impaction of parenchyma, CT allows precise evaluation of the thickness
the airways is easily identifiable by the presence of charac- of cavity walls as well as the presence of intracavitary
teristic focal fluid-filled, nonenhancing branching struc- filling defects and/or air-fluid levels. As reported by
tures, frequently associated with peripheral areas of emphy- Woodring et al. (109), the most important of these is
A B
Figure 6-17 Hamartoma. A, B: Sequential magnified 1-mm sections through a well-defined nodule in the left lower lobe show both dense
laminated calcifications and a small quantity of fat. This combination is diagnostic of a pulmonary hamartoma.
A B
Figure 6-18 Hamartoma. A, B: Magnified images through the right upper lobe imaged with narrow and wide windows, respectively,
show a well-defined subpleural nodule within which calcification and a small quantity of fat can be identified, consistent with the diagnosis
of a pulmonary hamartoma (compare with Fig. 6-17).
Figure 6-19 Hamartoma. Contrast-enhanced CT section shows

a giant fat-containing hamartoma within the left upper lobe. In this
case partial volume averaging is not a significant concern. Figure 6-21 Lipoid pneumonia: left lower lobe mass in an
elderly patient using oily drops for nasal dryness. Note focal
infiltrate with large areas of low attenuation in the range of fatty
tissues typical of this entity. (Case courtesy of Dr. Norman
assessment of cavity wall thickness. Of 65 solitary lung Ettenger, Bronx VA Hospital, New York, New York.)
cavities initially assessed by these authors using plain radi-
ographs, all lesions in which the thickest part of the cavity
wall was 1 mm were benign. In distinction, of those with
greatest wall thickness measuring between 5 and 15 mm, lesions. Gaeta et al. (111), in a review of thin-section CT in
51% were benign; of those over 15 mm, 95% were malig- 11 patients with nodular BAC, reported 2 lesions with
nant. In a follow-up prospective study of 61 additional serpentine radiolucencies that subsequently were shown
patients reported by these same authors, 19 (95%) of histologically to represent air-containing glandular spaces
20 cavities with a maximum wall thickness of 4 mm again within the tumor (Fig. 6-8). Kuhlman et al. (112), in a
proved benign; whereas 16 (84.2%) of 19 cavities with study of 30 patients with documented focal BACs, similarly
maximum wall thickness of 16 mm proved malignant described a pattern they interpreted as suggestive of
(110). In our experience, although thin-walled cavities are pseudocavitation in 18 cases. Defined as small oval areas of
almost always benign, thick-walled cavities are usually of lucency appearing in or around pulmonary masses, this
indeterminate etiology (Fig. 6-26). appearance was interpreted as consistent with residual
As important as assessing the thickness of the wall of dilated bronchioles and expanded air spaces, the result
cavities is the finding of bubblelike lucencies within focal of the distinctive propensity of these tumors to extend
along alveolar walls without distortion of the underlying
architecture. As noted by Zwirewich et al. (113), bubblelike
lucencies are seen more commonly in BAC (50%) than in
acinar adenocarcinoma (31%) or other lung tumors (11%),
making this sign highly suggestive of BAC (113).
Contour and Edge Characteristics

Malignant nodules are much more likely to have an irregu-
lar contour or spiculated margins (Figs. 6-10, 6-14, and
6-29) (113,114). In one study using high-resolution images
to evaluate 104 consecutive patients with focal lung lesions,
not only was spiculation per se a significant predictor of
malignancy but so, too, was the finding of spicules extend-
ing to the pleural surface as well as the greater length of
spicules (114). Benign lesions, on the other hand, much
more commonly exhibit a round or oval contour and a
sharply defined edge. This has proved true both for primary
and metastatic nodules. In patients with known neoplasms
Figure 6-20 Lipoid pneumonia. CT section through the midpor- and multiple pulmonary nodules, for example, Gross et al.
tions of both lungs shows bilateral ill-defined areas of parenchymal
consolidation within which there is unmistakable evidence of fat (115) found that metastatic lesions on CT are more likely to
density (arrows). be spherical or ovoid, whereas linear, triangular, irregularly
A B
Figure 6-22 Lung abscess. A: Contrast-enhanced CT section at the

level of the carina shows a well-defined fluid-filled lesion within which
a small quantity of air can be identified in an immunocompromised
patient. Note the presence of a similar-appearing smaller lesion in the
left upper lobe. In the appropriate clinical setting, the finding of a
well-defined, thin-walled fluid-filled lesion is characteristic of a lung
abscess. B: Contrast-enhanced CT section in a different patient than
shown in A shows a focal area of peripheral consolidation in the pos-
terobasilar segment of the right lower lobe within which there is a
sharply defined rounded area of lucency measuring in the range of
fluid (HU not shown). The finding of a sharply defined fluid collection
distinguishes a lung abscess from necrotic tumor, the margins of
which are never as sharply defined. This image also emphasizes the
need to evaluate complex parenchymal disease following intravenous
contrast administration. (C) Section at the same level as B, following a
6-week course of antibiotic therapy, shows that the previous abscess
has almost completely resolved, leaving behind only an ill-defined lin-
C ear scar to mark the spot.
Figure 6-23 Mucoid impaction.

A, B: Magnified images through the
left hilum following intravenous con-
trast administration show extensive
tumor obstructing the origin of the
left lower lobe bronchus. Note exten-
sive mucoid impaction in the proximal
portions of the superior segmental
bronchi, identifiable as linear branch-
ing structures extending posteriorly
from the left hilum (arrow). (See also
A, B Fig. 5-70.)
577
A B
Figure 6-24 Intrapulmonary bronchogenic cyst. A, B: Identical sections imaged with wide and narrow windows, respectively, show a well-
defined peripheral nodule in the superior segment of the right lower lobe in close proximity to a subsegmental bronchus. Note that one
minute following a bolus of intravenous contrast media, this lesion is uniform in appearance and measures 12 HU, indicating that it is within
the range of fluid. Subsequent images obtained at 2, 3, and 4 minutes failed to show any significant enhancement, suggesting the possibil-
ity of an intraparenchymal cyst or abscess. At transthoracic needle biopsy, approximately 2 mL of clear sterile fluid was aspirated. At thora-
cotomy this lesion proved to have cartilage in the wall, confirming the diagnosis of an intrapulmonary bronchogenic cyst. (From Costello P,
Naidich DP. Protocols for helical CT of the chest. In: Silverman PM, ed. Helical (spiral) computed tomography. A practical approach to clinical
protocols. Philadelphia: Lippincott–Raven; 1998:65–101, with permission.)
shaped, or multiple ill-defined lesions, especially when Nonetheless, although the contour and shape of
centrally located, are less likely to represent metastases. lesions may provide some indication as to the likelihood
These authors also found that if noncalcified, round lesions of malignancy, it is not unusual both for primary
larger than 2.5 cm were present or if more than 10 lesions lung neoplasms and, in particular, a solitary metastasis to
were visualized in any one patient, the likelihood of present with smooth, sharply marginated contours.
metastatic disease was very high. Similarly, although less common, benign lesions may
Figure 6-25 Hilar carcinoid: differentiation between tumor and

fluid-filled airways. A: Contrast-enhanced CT section at the level of
the inferior pulmonary veins shows a discrete mass in the right
hilum, easily differentiated from adjacent obstructed airways. B–E:
Sequential gadolinium-enhanced MR images corresponding to
the same approximate level as shown in A also show the presence
of a central tumor causing obstruction of the adjacent fluid-filled
airways. In this case, identification of central tumor is equally well
A established using either contrast-enhanced CT or MRI.
B C
D E
A, B
Figure 6-26 Thick-walled cavities: differen-

tial diagnosis. A–D: Sequential high resolution
sections through the right upper lobe show a
thick-walled cavity associated with adjacent
nodules with a tree-in-bud configuration (arrow
in B). This constellation of findings is strongly
suggestive of active tuberculosis, subsequently
verified at bronchoscopy. E: CT section in a dif-
ferent patient than in A–D also shows bilateral
irregular thick-walled cavities. In this case, the
lesion in the right upper lobe proved to be a
cavitary lung cancer, whereas the lesion in the
left upper lobe proved to be caused by tuber-
culosis (biopsy proved). As shown in these
cases, the presence of a thick-walled cavity in
C, D itself is nonspecific (continued).
E
appear spiculated (Fig. 6-29). Hirakata et al. (116), in a had poorly defined smooth margins and 30% had poorly
study of 87 metastatic lesions identified at autopsy, defined, irregular margins. Well-defined lesions correlated
including 43 nodules less than 5 mm, found 38% of nod- with the presence of tumor displacing surrounding nor-
ules to be well defined with smooth margins, using high- mal structures by uniform expansion, whereas irregular
resolution computed tomography (HRCT), whereas 16% margins correlated with proliferation of tumor into the
A B
C
Figure 6-27 Cavitation: benign disease. A: High-resolution CT section shows a multiseptate, thin-walled cavity in the posterobasilar seg-
ment of the right lower lobe, associated with minimal eccentric consolidation and ground-glass attenuation. The findings were consistent
with clinically documented cryptococcal infection. B: High-resolution CT section shows a sharply defined, thin-walled cavity in the left lower
lobe in an acquired immunodeficiency syndrome (AIDS) patient, subsequently documented to have atypical mycobacterial infection (MAI).
Note as well the presence of a focal area of mucoid impaction in the right lower lobe. The finding of thin-walled cavities, with or without
septations, is a reliable sign of benignity. C: CT section through the distal trachea shows multiple thick-walled cavities in this patient with
prior treated tuberculosis. Within the cavities are nodular filling defects (arrows), findings consistent with superimposed fungal infection, in
this case caused by Aspergillus species.
A B
C D
Figure 6-28 Cavitary disease: septic emboli. A–D: CT sections through the lungs, beginning at the level of the aortic arch and extending
to the middle lobe bronchus, show numerous predominantly peripheral/subpleural nodules in varying stages of cavitation, characteristic
of septic pulmonary emboli (arrows in A–C). Note that some of these lesions lie adjacent to pulmonary vessels, suggestive evidence of their
hematologic origin.
adjacent interstitium. Less well appreciated is the fact that matic SPNs, noted a benign-to-malignant ratio of 13:1 in a
benign lesions may also have spiculated margins, further group of 112 lesions 1 cm or less versus a ratio of 1:4 in
limiting the usefulness of this sign. 144 lesions greater than 3.0 cm in diameter (41).
Surprisingly, given the importance usually attributed to The size of a lung nodule identified with CT is also a
the edges of lesions, significant interobserver variability in good indicator of the likelihood of malignancy, especially
the assessment of this sign has also been reported. In a when interpreted in association with lesion density. This is
study characterizing nodules in 66 patients, the largest especially true for small nodules (less than 10 mm) identi-
discrepancy between four board-certified readers was in fied on CT. As documented in a number of recent reports,
classification of nodule contour (94). This suggests that the likelihood of a nodule less than 4 mm incidentally
accurate evaluation of this sign may require a more precise identified on CT being malignant is exceedingly small,
definition or perhaps the use of a more quantitative assess- regardless of whether or not it is subsolid or solid in
ment (117,118). appearance (119–121). Only rarely will such lesions prove
malignant (Figs. 6-1 and 6-2). In a recent retrospective
review of 2,987 prevalence LDCT screening studies, for
Size
example, 0 of 378 nodules less than 5 mm proved to be
The likelihood of malignancy in an indeterminate pul- malignant on yearly follow-up studies, whereas in this
monary nodule detected by chest roentgenography is a same study 13 of 328 nodules between 5 and 9 mm
direct function of the size of the lesion. This principle was proved malignant on subsequent examination (122). In
well established in the years preceding the advent of CT. distinction, nodules between 4 and 10 mm incidentally
Thus, Steele et al. (41), in a study of 756 resected asympto- identified with CT should be considered indeterminate,
Figure 6-29 Nodule characterization: spic-

ulation. A: Magnified view at the level of the
posterior aspect of the aortic arch shows an ir-
regular, spiculated lesion, the appearance of
which is highly suggestive of a primary lung
neoplasm, subsequently surgically verified as
non–small cell lung cancer. B: Magnified sec-
tion through the right upper lobe in a different
patient than A shows a similar-appearing spic-
ulated nodule seemingly partially obstructing
the medial branch of the anterior segmental
bronchus, consistent with a presumed prior
lung neoplasm. C: Follow-up CT study several
weeks later, without therapy, obtained prior to
diagnostic bronchoscopy shows that the previ-
ous lesion is no longer identifiable, consistent
with a benign, presumably inflammatory, etiol-
ogy. Although spiculation is highly suggestive
of a malignant etiology, in any individual case
it remains an unreliable finding in itself.
A
B C
whereas lesions larger than 10 mm are best considered cially in symptomatic patients. In one retrospective study
malignant until proven otherwise (123). The incidence of of 64 patients with nodules less than 1 cm subsequently
malignancy in a lung lesion larger than 3 cm is so great evaluated by video-assisted thoracoscopy (VATS), reported
that all these lesions should be surgically resected unless by Munden et al. (124), 38 (57%) nodules proved to be
medically contraindicated. neoplastic, including 37 nodules in patients without prior
Although size has proved to be a statistically significant history of malignancy. Strikingly, in 8 cases with nodules
discriminator between benign and malignant lesions, less than 5 mm (so-called ditzels), 6 of 8 proved malig-
especially in a screened population, the potential for tiny nant, whereas 2 proved to be due to intrapulmonary
nodules to be malignant cannot be entirely ignored, espe- lymph nodes.
It should be emphasized that the likelihood of small Select Morphologic Features

(1 cm) nodules being malignant will also be influenced
by the presence of additional lesions. Yuan et al. (125), in a In addition to the morphologic features discussed previ-
retrospective review of preoperative CT scans in 223 pa- ously, additional select morphologic signs have been
tients, evaluated the probability of malignancy occurring in described that are worth describing separately owing to
small nodules less than 1 cm coexisting with potentially their widespread use. These include the “feeding vessel
operable lung cancers. Areas of evaluation included nodule sign,” evidence of a presumed hematogenous origin or
size and location and the clinical stage of the primary evidence of an arteriovenous malformation (AVM); a
tumors. A total of 71 coexisting nodules were identified in “positive bronchus sign,” defined as a dilated bronchus
58 (26%) patients, of which 18 (25%) proved malignant. leading into or through a nodule; and the “CT halo sign,”
Although the probability of malignancy was greater in identifiable as the presence of ground-glass attenuation
nodules located within primary tumor lobes, four (57%) surrounding a solid nodular core.
coexisting synchronous cancers were identified in lobes
other than those containing the dominant lesion. The Feeding Vessel Sign
likelihood of malignancy also proved to be less in patients Focal disease, in which identification of the relationship
with clinical stage 1A disease. between nodules and adjacent or feeding vessels has
Although there is little dispute regarding the relation- been described, includes metastases, infarcts, and AVMs
ship between size and the likelihood of malignancy, there (Figs. 6-28, 6-30, and 6-31). Particularly intriguing has
is controversy regarding the significance of size as an been the observations made by some investigators of a
indicator of cancer stage. Although it has been argued that connection between pulmonary metastases and adjacent
size is not an accurate reflection of likely tumor stage pulmonary arteries, the so-called feeding vessel sign
(126), recent data strongly suggest that a stage-size relation- (Fig. 6-30). Meziane et al. (131), in a radiologic-
ship exists, in particular in a screened population. Among pathologic correlative study, noted frequent identification
464 diagnosed lung cancers in 28,689 screened individuals, of pulmonary vessels leading to the center of metastatic
including 84% surgically staged and the remainder evalu- lesions. These results have received further confirmation
ated with follow-up imaging studies, the percentage of using microangiographic techniques (132). However, in a
cases with no metastases (N0M0) was 91% for lesions more recent study comparing the thin-section CT appear-
15 mm or less, 83% for lesions 16 to 25 mm, and 68% for ance of metastases with histopathologic findings,
lesions between 26 and 35 mm (18,127). Hirakata et al. (116) studied 87 metastatic lesions identi-
fied at autopsy, including 43 nodules less than 5 mm.
Location
It has long been established that primary lung cancers are
more commonly found in the upper lobes, especially on
the right side. This remains true to the present. In a recent
review of 40 solitary nodules identified in an outpatient
department, SPNs were identified in the right lung in 70%
of cases and in the upper lobes in 65% (128). Not surpris-
ingly, radiographically SPNs were most often identified in
the lung periphery in 60% of cases, compared with the
medial third and middle third of the lung in 10% and
30%, respectively (128). Metastatic lesions tend to be sub-
pleural in location. In a study of serial sections of resected
lungs containing metastatic nodules, Scholten and Kreel
(129) found that 60% of the lesions were in an immediate
pleural or subpleural location and an additional 25% were
located in the outer third of the lung. Furthermore, two
thirds of metastatic lesions occur in the lower lobes.
Unfortunately, a number of other lesions also have a
predilection for subpleural portions of the lungs. As
documented by Bankoff et al. (130), in a study including
96 patients with well-circumscribed peripheral pulmonary
nodules undergoing minithoracotomies, 17 (18%) proved
to have intrapulmonary lymph nodes. It should be Figure 6-30 Magnified CT section through the right upper lobe
shows an irregular nodule toward which a peripheral pulmonary
noted that most of these proved to be solitary, with 12 arterial branch is clearly coursing (arrow). Documented metastatic
located within the lower lobes. colon cancer.
A, B, C D, E
Figure 6-31 Feeding vessel sign:

arteriovenous malformation (AVM).
A–E: Sequential magnified contrast-
enhanced sections through the
medial aspect of the left lower lobe
show a markedly enhancing periph-
eral nodule associated with enlarged
vessels leading from the hilum
toward the nodule, consistent with
an AVM. F, G: 3D renderings of the
lesion clearly demonstrate the
relationship between feeding arteries
and draining veins to better advan-
tage than sequential axial images,
facilitating therapeutic/interventional
F, G planning.
Contrary to previous reports, these investigators found possibility of partial volume effects making a lesion appear
that in the majority of lesions no discernible relationship connected to an adjacent but unrelated vessel, especially
could be identified between metastases and adjacent pul- one coursing obliquely through the scan plane. To obvi-
monary vessels. It is worth noting that this latter study ate this pitfall, it should be emphasized that thin sections
was based on careful analysis of HRCT images. It is likely and/or MPRs are necessary to properly assess this sign.
that the higher prevalence of the feeding vessel sign in Other indirect signs suggesting a hematogenous or
previous investigations was in part artifactual because of vascular etiology have been described. One is the finding of a
volume averaging of thicker CT sections. zone of hypodensity distal to nodules. Presumably, this rep-
Although a vascular connection is not commonly seen, resents hypoperfusion distal to a pulmonary artery occluded
when present it is helpful for distinguishing hematogenous by metastases. Even after successful chemotherapy in pa-
metastases from either primary carcinomas or granuloma- tients with documented metastatic lung disease, a mass-ves-
tous lesions (Fig. 6-30). Although the finding of a feeding sel connection can still be seen. Other indirect but suggestive
vessel sign has been cited as reliable in diagnosing patients findings indicative of vascular disease have been described in
with septic emboli, recently Dodd et al. (132a) have shown patients with documented pulmonary infarction (133).
that in fact most often vessels actually course around nod-
ules in these patients when assessed with MPRs (132a). Arteriovenous Malformations. Not surprisingly, the most
Similar considerations likely apply to the finding of a feed- compelling evidence substantiating the value of the feeding
ing vessel in patients with pulmonary vasculitis (Fig. 6-28). vessel sign comes from patients with known or suspected
A major potential pitfall in using this mass-vessel sign is the AVMs (Figs. 6-31 and 6-32). The presence of pulmonary
A B
Figure 6-32 Feeding vessel sign:

complex arteriovenous malforma-
tion (AVM). A, B: Magnified con-
trast-enhanced CT sections through
the left lower lobe and just above
the left hemidiaphragm, respec-
tively, show a complex tangle of ab-
normal, seemingly separate vessels.
C, D: Axial and off-axis sagittal volu-
metric renderings allow visualization
of the complex interrelationship be-
tween these vessels, otherwise
nearly impossible to reconstruct using
axial images alone. (Case courtesy of
Elliot Fishman, MD, Johns Hopkins
C, D Hospital, Baltimore, Maryland.)
vascular malformations is often suspected based on a clinical Although nonspecific, pulmonary lesions that directly
history and findings on plain chest radiography. A high per- abut, displace, or narrow a visible bronchial lumen should
centage of patients with lung lesions are affected with Osler- not be assessed as benign (137–140). In a retrospective
Weber-Rendu disease. Chest roentgenography often shows a study of 132 patients with solitary nodules evaluated with
complex lesion consisting of a lobulated or serpiginous mass CT, Kui et al. (140) identified a total of 34 lesions with air
associated with tortuous feeding arteries and draining veins. bronchograms; altogether, 33 (28.7%) of 115 proved lung
However, for smaller lesions the vasculature may be subtle or cancers but only 1 (5.9%) of 17 benign lesions had air
inapparent and the lesion may be manifest as an SPN; asso- bronchograms identified. When assessed morphologically,
ciated vasculature may be visualized only on CT. airways seen in relation to lung cancers were invariably
Remy et al. (134,135) have demonstrated that spiral abnormal, being tortuous, ectatic, or cut-off, confirming
CT allows accurate assessment of the presence and num- earlier descriptions by Gaeta et al. (141).
ber of arteriovenous malformations, being superior to
angiography for demonstrating small lesions. As impor- Computed Tomography Halo Sign
tant, unenhanced three-dimensional (3D) CT can pro- The halo sign by definition is a discrete nodule sur-
vide a reliable analysis of the angioarchitecture of lesions rounded by a circular margin of ground-glass attenuation
and is therefore useful in follow-up both of treated and (Fig. 6-35). Usually representing hemorrhage surrounding
untreated patients. a focal infarct, this sign has proved useful in the proper
clinical situation as an early indication of invasive pul-
Positive Bronchus Sign monary aspergillosis (IPA) (142–146). When identified in
The presence of a positive bronchus sign or air bron- immunocompromised patients, particularly those with
chogram is an important and common finding with thin- acute leukemia, following aggressive therapeutic marrow
section CT (Figs. 6-33 and 6-34). Gaeta et al. reported a suppression accompanied by severe leukopenia, detection
prevalence of a positive bronchus sign in 66.7% of SPNs of the halo sign has proved sufficiently characteristic
evaluated by thin-section CT, whereas Kuriyama et al. of fungal infection to warrant empirical therapy (147).
(136,137) reported a prevalence of 35.0% (14 out of The halo sign is most useful early in the course of infec-
40 cases) in patients with documented malignant nodules. tion. As noted by Blum et al. (146), in their prospective
A, B Figure 6-33 Positive bronchus sign. A–D:

Sequential 1-mm sections obtained volumet-
rically through the left mid lung show the
presence of a poorly defined lesion in the left
lower lobe. Note that in this case the anter-
obasilar bronchus can be traced on sequen-
tial images into this lesion (arrow in C). This
appearance correlates with a greater likeli-
hood of obtaining a histologic diagnosis on
transbronchial biopsy. E: Histologic section
obtained by transbronchial biopsy identifies
this lesion as a non–small cell lung cancer.
Note the close proximity of tumor cells to
the adjacent bronchial mucosa (arrow). (From
Naidich DP. Volumetric CT in the evaluation
of focal pulmonary disease. In: Remy-Jardin
M, Remy J, eds. Spiral CT of the chest.
Berlin: Springer-Verlag; 1997:129–151, with
C, D permission.)
E
A B
C D
Figure 6-34 Positive bronchus sign. A–H: Sequential 1-mm helical CT sections from above-downward in a patient with a spiculated right
upper lobe nodule first through the right upper lobe bronchus (A–D) and then the bronchus intermedius (E–H) allow central airway branches
to be traced sequentially. Note that in this case, a lateral subsubsegmental branch of the anterior segmental bronchus can be traced to this
lesion (arrows in E–G). The ability to identify a relationship between airways and nodules has clear implications for the likely success of trans-
bronchial biopsy in establishing a definitive histologic diagnosis. Bronchoscopically verified non–small cell lung cancer (continued).
E F
G H
A B
Figure 6-35 A: Early invasive aspergillosis. Patient in leukopenic phase following bone marrow transplant. Note halo of intermediate density
in the transition zone between the nodule and the normal lung. This sign is suggestive of infection by angiotropic organisms, the most common
of which is Aspergillus fumigatus. B: One-millimeter target reconstructed image through the right upper lobe shows multiple ill-defined nodules
all associated with adjacent areas of ground-glass attenuation. In this case of documented pulmonary Kaposi sarcoma, areas of ground-glass at-
tenuation represent associated lung hemorrhage. Based on these cases, it may be concluded that the finding of a halo sign per se is nonspecific.
A similar appearance may also be seen in patients with bronchoalveolar cell carcinoma and/or adenocarcinoma (Figs. 6-7, 6-9, and 6-10).
study of 38 neutropenic patients presenting with radio- spiculated nodules may still be due to chronic inflamma-
graphic evidence of nodular lung disease, a halo sign was tory lesions. As a consequence, most lesions remain
noted in 16 of 22 patients evaluated within the first 7 days indeterminate, requiring the use of additional imaging
with documented IPA, with no false-positive studies. In techniques for diagnosis.
distinction, a halo sign could be identified in only 3 of 13
patients with proved invasive aspergillosis evaluated only
Contrast-Enhancement Computed
after 10 days. As most of these patients also have ex-
tremely low platelet counts, making biopsy unfeasible,
Tomography Evaluation
CT findings play a critical role in the management The phenomenon of enhancement of pulmonary malig-
of these patients (144). It is only later following successful nancies by intravenous injection of contrast media was first
therapy when granulocytes begin to return to the blood- documented by Littleton et al. (151), using tri-spiral tomog-
stream and patients begin to recover that the characteristic raphy of lung masses. These authors showed that there is a
air crescent sign traditionally identified on plain radi- fundamental difference in vascularity between focal malig-
ographs as diagnostic of fungal infection is seen as nant and benign lesions: malignant lesions contain neovas-
the central infarcted portion of the lesion becomes cularity and hence are much more apt to enhance following
demarcated. injections of contrast media. Numerous publications to
Mori et al. (147), in a study of 33 febrile bone marrow date have shown that this distinctive feature may also be
recipients, found that CT showed nodules in 20 of 21 successfully exploited using angiography, contrast-enhanced
episodes of documented fungal infection but none in 9 conventional tomography, FDG-PET, gadolinium-enhanced
bacteremic episodes. Based on these data, these authors magnetic resonance imaging, Doppler ultrasonography,
concluded that CT studies documenting the presence of and CT (152).
nodules in this population could be taken as presumptive In an initial study of 163 patients, Swenson et al.
evidence of fungal infection warranting empiric therapy (153) found the median enhancement of malignant
without the need for bronchoscopy. It should be empha- nodules (n 111) to be 40 HU (range 20 to 108 HU)
sized, however, that the halo sign is nonspecific and has compared with 12 HU (range 4 to 58 HU) for benign
been reported in a variety of other lesions in which both lesions (n 52). Using 20 HU as a threshold separating
inflammatory and neoplastic processes compromise malignant from benign lesions, these authors reported
blood vessels, resulting in nodules surrounded by hemor- sensitivity of 100%, specificity of 76.9%, and accuracy of
rhage. Primack et al. (148) have described this sign in 92.6% (Fig. 6-36). In a follow-up to their original report,
patients with other infections, including candidiasis, again using 20 HU as a threshold to evaluate 107
cytomegalovirus infection, and herpes pneumonia, as patients, these same authors again noted that the median
well as in patients with Kaposi sarcoma and metastatic enhancement in malignant lesions (n 52) proved
angiosarcoma (Fig. 6-35B). The halo sign has also been significantly higher than that in benign lesions with a
reported to occur in patients with tuberculomas (149). sensitivity of 98%, specificity of 73%, and accuracy of
85% and positive and negative predictive values measur-
ing 77% and 98%, respectively (154). Only one false-
Morphologic Features: Overview
negative study was reported (Figs. 6-37 and 6-38). Eight
Although individual morphologic features are of value in nodules were identified that measured between 16 and
differentiating benign from malignant nodules, especially 24 HU: four of these proved malignant, leading the
when interpreted in concert, unfortunately, even together authors to define this range as inconclusive (154).
they are of only limited use in individual cases (Fig. 6-29B The value of contrast enhancement has been further
and C) (150). In one study evaluating the role of HRCT in validated by several other investigators (155–158). In the
characterizing 104 consecutive cases of pathologically largest series published to date, Swensen et al. (159)
proved solitary nodules, statistically useful findings to dif- prospectively evaluated 356 solid nodules between 5 and
ferentiate benign from malignant disease included (a) 40 mm. Using 15 HU as a threshold for a positive test
spiculation, especially when prominent, extending to the result, malignant lesions enhanced significantly more than
adjacent visceral pleura or associated with pleural retrac- granulomas and benign neoplasms (median 38.1 HU vs.
tion and/or focal pleural thickening; (b) a positive 10 HU, p 0.01), with a sensitivity of 98%, specificity of
bronchus sign; (c) the presence of subsolid nodules with 58%, and accuracy of 77% for predicting malignancy.
both solid and ground-glass components; and (d) the den- As initially described, contrast-enhanced CT nodule
sity of lesions (fat vs. fluid vs. calcification) (114). Using studies were first performed prior to the introduction of
these features combined, CT proved to have a sensitivity of SD CT scanners and performed as follows: after initial
91.4% for identifying malignant nodules, with a corre- unenhanced scans a total of 420 mg I/kg, 300 mg I/mL
sponding specificity of 56.5% and an overall accuracy of (100 mL for a 70-kg subject) was injected at a rate of
83.7%. As noted by these authors, well-defined round 2 mL/second. Three-millimeter-thick sections were then
metastases may mimic benign disease, whereas irregular obtained using a standard reconstruction algorithm
A, B C
D, E F
Figure 6-36 Nodule characterization: contrast-enhanced CT. A, B: Magnified 3-mm sections through the anterior aspect of the left upper
lobe imaged with wide and narrow windows, respectively, show a discrete, otherwise indeterminate 5-mm nodule. C–F: Sequential images
at the same level as shown in A and B obtained at 1, 2, 3, and 4 minutes following a bolus of 100 mL of nonionic contrast media
administered at a rate of 2 mL/second. Note that the maximum increase in density within this lesion from a baseline of 94 HU (B) occurs at
3 minutes (E) and equals exactly 20 HU, consistent with a presumed benign etiology. This lesion subsequently proved to be without change
in appearance over several years. In fact, this size lesion in general cannot be reliably evaluated with this technique owing to the likelihood
of significant noise resulting from the need to use thin sections to avoid partial volume averaging.
employing a 19-cm field-of-view, with images obtained at as well as patterns of contrast enhancement. Defining
1-, 2-, 3-, and 4-minute intervals using an identical tech- nodules as either malignant (n 42), benign (n 16), or
nique. Image interpretation was performed by identifying inflammatory—defined as foci of active inflammation
the one axial image most clearly obtained through the (n 7)—these authors showed that significant differences
center of the nodule selected for each separate timed could be measured in the peak height of enhancement
acquisition, with subsequent HU analysis using a region- within malignant (41.9 HU) and inflammatory nodules
of-interest (ROI) occupying 60% of the total diameter of (13.4 HU) compared with benign nodules (13.4 HU)
the nodule (154). (p 0.001). Similarly, significant differences could be
Modifications of this technique have been suggested. identified between these groups, calculating both the ratio
Zhang and Kono (155), for example, following rapid con- of peak enhancement within nodules to the aorta as well
trast infusion (4 mm/second) used single-level dynamic as nodule perfusion (155). Using an absolute peak en-
SD CT to acquire clusters of images at 15 and 65 seconds, hancement of 20 HU (in place of an incremental change
allowing these authors to assess a number of parameters, in density of 20 HU as initially proposed by Swensen et
including peak levels of nodule enhancement (including al.), 2 of 42 malignant nodules were interpreted as benign.
the ratio of peak enhancement within nodules versus the Interestingly, significant differences were also identified
aorta), time-attenuation curves, and nodule perfusion in this study when patterns of nodule enhancement were
assessed. In addition to absence of enhancement, images, volumetric datasets through nodules were acquired
specific patterns evaluated included homogeneous versus at 30, 60, 90, and 120 seconds and then at 4, 5, 9, 12, and
heterogeneous enhancement and central versus peripheral 15 minutes following the injection of 120 mL of contrast
or rim enhancement. Significantly, although benign nod- medium. Using the combination of wash-in of 25 HU or
ules characteristically showed either no enhancement greater coupled with washout of 5 to 31 HU, these authors
(56%) or less commonly peripheral enhancement (31%), reported a sensitivity, specificity, and accuracy of 94%,
there were no cases of heterogeneous enhancement within 90%, and 92%, respectively, for diagnosing malignancy. In
this group. In distinction, malignant nodules most com- distinction, benign lesions were characterized by a number
monly showed homogeneous enhancement (55%) with of patterns, most frequently less than 25 HU wash-in
heterogeneous enhancement next most commonly seen (160). Although suggestive, this technique clearly raises
(30%); in no case did a malignant nodule show no meas- issues regarding radiation dose.
urable enhancement. In an attempt to further delineate the relationship
More recently, using MDCT, Jeong et al. (160) evaluated between contrast enhancement and malignancy, Yi et al.
the use of combined wash-in and washout data to more (157) measured peak and net enhancement of 54 nodules
precisely characterize a total of 107 pulmonary nodules following 120 mL of contrast medium and showed signifi-
(49 malignant, 58 benign). Following initial unenhanced cant correlation between peak enhancement of nodules
A B
C D
Figure 6-37 Nonenhancing lung cancer related to a scar. A–D: Three-millimeter section obtained at the same level 1, 2, 3, and 4 minutes
following a bolus of intravenous contrast and imaged with lung windows shows the presence of a slightly irregular, oblong nodule adjacent
to an area of focal pulmonary emphysema anteriorly. E–H: Identical images as shown in A–D, imaged with narrow windows. In this case the
mean density within this lesion (shown on the bottom of each image) measured 8.2, 16.6, 9.6, and 20.7 HU, respectively, never exceeding
15 HU above the baseline noncontrast section (not shown). Histologic section through this lesion confirmed the presence of foci of adeno-
carcinoma. By visual inspection, greater than 90% of this lesion proved to be fibrotic tissue, accounting for the lack of contrast enhancement
noted in D–G. Extensive fibrosis within pulmonary nodules, whether or not this constitutes a scar carcinoma, represents a potential pitfall in
the use of contrast enhancement to identify benign pulmonary nodules. I: Histologic section through this lesion confirming the presence of
foci of adenocarcinoma (arrows). By visual inspection, greater than 90% of this lesion proved to be fibrotic tissue, accounting for the lack of
contrast enhancement noted in D–G (continued).
E F
G H
I
and the extent of microvascular density (MVD) (p 0.006) negative predictive value of 97%, and an accuracy of 78%
and vascular endothelial growth factor (VEGF) staining for predicting malignancy.
(p 0.042). Using 30 HU net enhancement as a cut-off, Despite apparently compelling data, numerous ques-
these investigators further reported a sensitivity of 99%, a tions remain regarding the role of contrast-enhanced
specificity of 54%, a positive predictive value of 71%, a nodule studies for evaluating pulmonary nodules. These
A, B C
Figure 6-38 Nodule characterization: contrast-enhanced CT. A: Magnified view through the base of the middle lobe imaged with wide
windows shows a slightly irregular, ovoid lesion suggestive of focal mucoid impaction (compare with Fig. 6-23). B, C: Images at the same
level as shown in A, obtained 1 and 3 minutes following intravenous contrast administration, show lack of significant enhancement, with
density measurements of 16 and 17 HUs, respectively. Despite lack of enhancement, this lesion was resected and proved to be a mucinous
bronchoalveolar cell carcinoma. Lack of enhancement in this case reflects the fact that most of this lesion represented a mucus-filled periph-
eral airway. (Case courtesy of Jeffrey Klein, MD, University of Vermont Medical Center, Burlington, Vermont.)
include optimal methods for performing and interpreting Given a striking negative predictive value, it is perhaps
contrast-enhanced studies, including an optimal thresh- surprising that this technique has not gained greater
old, if any, for differentiating benign from malignant acceptance. No doubt this is due to a number of factors, not
lesions; optimal methods for reliably measuring contrast least the fact that this approach requires meticulous atten-
enhancement; determination of the lower limit of size for tion to scan technique. In this regard, it should be noted that
accurate CT assessment; and guidelines for evaluating one limitation—namely, the need to individually identify
nodules with heterogeneous density (Fig. 6-39). Also an optimal section through the mid portion of nodules
unclear is the optimal number of data acquisitions to from each timed acquisition and subsequently select identi-
properly assess nodules while minimizing radiation dose. cal ROIs from each series—can be obviated by use of auto-
As noted by Swensen et al. (161), in their initial study of mated nodule volume segmentation providing a single
163 patients, 50 (31%) had peak nodule enhancement at measure of density throughout the entire nodule (Fig. 6-40).
either 3 or 4 minutes; in six of these studies, failure to Although it may seem that the introduction of FDG-PET
obtain delayed images would have resulted in false- (discussed later) would diminish the need for contrast-
negative diagnoses. enhanced nodule studies, it is of interest to note that in at
Allowing for these limitations, contrast-enhanced CT least one report from Australia in which four separate diag-
clearly has a potentially important role to play as an nostic strategies for evaluating nodules were assessed,
additional means to reach an informed decision on the including conventional CT, conventional CT followed by a
management of select patients with pulmonary nodules. contrast-enhanced CT nodule study (QECT), conventional
Undoubtedly, the greatest potential use is its role in pro- CT followed by FDG-PET, and a combination of all three, a
viding support for conservative follow-up of noncalcified strategy using a combination of contrast-enhanced CT and
lesions, which are considered likely benign clinically and PET proved most cost-effective, assuming a prevalence of
morphologically. The use of contrast enhancement malignancy of approximately 50% (162).
clearly will be less helpful in cases with features sugges-
tive of malignancy such as large nodules or those with
spiculated borders. Such cases usually demand tissue
diagnosis. Enhancement of a lesion, per se, is of less The potential use of gadolinium (Gd)-enhanced MRI to
value, given the significant number of false-positive characterize lung nodules has been long appreciated
examinations (154). (163,164). Compared with contrast-enhanced CT, dynamic
Faster acquisition times allow the evaluation of a

number of dynamic MR indices based on measuring SIs,
including the maximum intensity enhancement ratio
(MER) (defined as the maximum SI following contrast
enhancement minus the SI prior to contrast enhancement),
the slope of enhancement, and the time to maximum
enhancement (Tmax), among others (Figs. 6-41 and 6-42).
It should be emphasized that when applied to tumors, SI
indices reflect a combination of factors, including lesion
perfusion, relative blood volume, capillary surface area,
vascular permeability, and the volume of extravascular fluid
(165). In fact, good correlation between MRI findings and
the extent of tumor vascularity has been reported. Using
a T1-weighted spin-echo sequence, Fujimoto et al. (166), in
a retrospective study of 94 resected lung nodules, showed
that the MER and the slope of the time-SI curve were posi-
tively correlated and Tmax was negatively correlated with
measurements of MVD and VEGF assessed immunohisto-
chemically. Also observed was good correlation between
prolonged SI washout (calculated as the percentage
decrease in SI between Tmax and SI 3 minutes later) and
the density grade of elastic and collagen fibers within the
tumor interstitium.
Most important, MR has been shown to be of value for
differentiating those nodules requiring subsequent evalua-
tion or treatment from those for which no further workup
is necessary. Ohno et al. (166), employing a 3D radio-
frequency spoiled gradient-echo sequence, showed that
patients with malignant nodules and those with active
Figure 6-39 Nodule characterization: limitations of contrast- infection could be differentiated from patients with inac-
enhanced CT. Magnified view of a 1-mm section through a nodule in
the right lower lobe imaged 1 minute following a bolus of 100 mL of tive benign nodules using dynamic MR indices with a
nonionic contrast media administered at a rate of 2 mL/second. sensitivity, specificity, and accuracy of 100%, 70%, and
Note that, although the density within this lesion measures 1.6 HU, 95%, respectively. This included 8 of 10 nodules in which
the standard deviation (arrow) measures 26 HU, clearly limiting the
value of this technique. Noise resulting from use of thin sections to interventional studies obtained to differentiate benign
evaluate small lesions is easily identified, visually resulting in marked from malignant disease revealed active infections for
heterogeneity in nodule density. With rare exception (see Fig. 6-38), which subsequent treatment was required.
contrast-enhanced CT optimally should be reserved for lesions larger
than 8 mm to ensure the reliability of density measurements. Although a number of studies to date have shown that
MR is effective in identifying malignant nodules, the role of
Gd-enhanced MR for differentiating between various types
of tumors is less clear. Fujimoto et al. (166), in their study
MRI using fast imaging techniques including gradient-echo of Gd-enhanced MRI, found no statistical differences in
(GRE) or echo-planar imaging, in particular, has the consid- dynamic MR indices between adenocarcinomas (n 67)
erable advantage of superior temporal resolution while and squamous cell carcinomas (n 27). In distinction,
avoiding ionizing radiation (165). In this regard, numerous Ohno et al. (167) have reported that use of dynamic
studies have emphasized the need to consider more than MR imaging is of value for differentiating between
just peak enhancement when assessing pulmonary nodules. BACs and peripheral adenocarcinomas. Presumably reflect-
In an early prospective study of 28 SPNs (20 malignant, ing a greater number of intact arterioles assessed histologi-
8 benign) between 10 and 30 mm, for example, although cally, both the maximum relative enhancement ratio and
no statistical difference between malignant and benign nod- mean slope of enhancement of BACs proved significantly
ules was identified using percent change in signal intensity greater than in patients with either mixed BACs or invasive
(SI) before and after enhancement, in this same population, adenocarcinomas.
significantly higher SIs were demonstrated in malignant Despite these findings, a number of limitations inhibit
lesions when enhancement curves (% SI/second) were effective assessment of the role of Gd-enhanced MRI for
evaluated during the first transit of a bolus of contrast evaluating pulmonary nodules. In addition to poorer
media using dynamic snapshot sagittal GRE images spatial resolution compared with CT as well as numerous
(p 0.0001%) (164). artifacts, these include use of different image acquisition
A, B, C D, E, F
G H
Figure 6-40 Nodule characterization: 3D evaluation. See Color Figure 6-40I,J. A–F: Magnified views through a well-defined nodule in the
right upper lobe evaluated on a 16-detector CT scanner with images obtained prior to (A) and then at 35 (B), 65 (C), 95 (D), 155 (E), and 190
(F) seconds following a bolus of 100 mL of nonionic contrast media administered at a rate of 2 mL/second. Note that in this case, maximum
contrast enhancement occurred at 35 seconds, measuring 97.4 HU, increasing by 43.4 HU as measured from a region of interest manually
placed in the center of the lesion on one axial image through the center of the nodule, findings consistent with possible neoplasm. G, H:
Magnified views obtained in the same patient as shown in A through F at the same time, imaged with wide and narrow windows prior to the
injection of IV contrast media. I, J: Images showing the result of automatic segmentation of both lesions illustrated above, respectively.
Note that this provides automated evaluation of the length of these lesions in the x, y, and z plane, as well as a global measurement of the
HU density of these lesions. K, L: Graphic display of 3D volume densitometry in distinction to the method for density measurements derived
for both lesions using representative HU density measurements from individual axial images versus 3D volume densitometry. Note that as
the information generated by these two techniques is approximately equal, an automated 3D approach would be preferable both to mini-
mize interobserver variability as well as to simplify the task of obtaining these measurements. It should also be noted that rapid acquisition
of data potentially not only allows evaluation of global patterns of contrast enhancement but additionally offers the potential to measure tis-
sue perfusion and permeability. At surgery, the lesion in the right upper lobe proved to be an adenocarcinoma; in distinction, the lesion in
the left upper lobe has been known to be stable for several years, and is consistent with a presumed hamartoma (continued).
I J
45 RUL RUL 3D
LUL 45
LUL 3D
Enhancement
Enhancement
30
30
(HU)
(HU)
15
15
0 0
0 50 100 150 200 0 50 100 150 200
K Time (s) Time (s) L

techniques (e.g., fast SE vs. 3D gradient-echo sequences); Positron Emission Tomography

lack of standardization of techniques for intravenous con-
trast administration (including rate, volume, or lack of Initially introduced in the early 1990s (169–174), over the
correlation with body mass and height); a variety of meth- past few years PET with FDG is now widely accepted as
ods for obtaining pathologic correlation; differences in routinely indicated in the assessment of pulmonary
populations studied, especially given the small sample nodules. In this section, the main focus is on the use of
sizes reported in most series; and use of a variety of meth- PET for specifically evaluating pulmonary nodules. A more
ods for deriving dynamic MR indices, including use of detailed assessment of the role of PET imaging in lung can-
retrospective derived thresholds. Although many of these cer staging is presented in Chapter 7.
considerations also apply to CT, until MR becomes more Briefly, FDG is a D-glucose analog radiopharmaceutical
accessible and less expensive, it is unlikely that MR will labeled with a positron emitter (18F) that is transported
compete with either CT or FDG-PET (see later) in the fore- through the cell membrane and phosphorylated using
seeable future. In one retrospective study comparing normal glycolytic pathways. Increased uptake and accumu-
dynamic MRI and contrast-enhanced CT, for example, in lation of FDG has been shown to occur in tumor cells
23 cases in which both modalities were obtained there was owing to increased expression of glucose transporter
no statistically significant difference between MR and CT messenger RNA and increased transporter protein, al-
for differentiating benign from malignant disease (168). though the effect is nonspecific and may be seen as well in
This was so despite the fact that in 2 discordant cases MR foci of inflammation. Once inside tumor cells, FDG is only
correctly differentiated the benign from the malignant slowly metabolized, allowing identification with PET
lesions, including 1 tuberculoma and 1 adenocarcinoma. (Figs. 6-43 and 6-44) (171).
A,B C
D, E F
Figure 6-41 Solitary nodule: 3T MR evaluation. A–C: Dynamic MR images acquired in a 71-year-old man with primary lung cancer (adeno-
carcinoma) in the left lower lobe, with a 3D radiofrequency spoiled gradient-echo sequence (TR 2.7 ms/TE 0.6 ms/flip angle 20, 28 96
matrix, 256 192 reconstructed matrix, rectangular field of view 400 280 mm, slab thickness 110 mm, 11 partitions) using an overcon-
tiguous slice in the coronal plane. Dynamic MR images (A) t 9.9 s, (B) t 11.0 s, and (C) t 18.7 s demonstrate the enhancement
effect within a solitary pulmonary nodule measuring 25 mm (arrow). Note that t indicates the time after bolus injection of contrast media.
The maximum relative enhancement ratio was 0.68, whereas the slope of enhancement in this case was 0.077/s. D–F: Dynamic MR
images in a 56-year-old man with a documented healed tuberculoma in the right upper lobe, acquired with exactly the same parameters
as shown in A–C. Dynamic MR images—in this case (D) t 0 s; (E) t 12.1 s; and (F) t 23.1 s)—demonstrate the effect of
enhancement on this 11-mm solitary nodule (arrow). Distinct from the adenocarcinoma illustrated in A–C, in this case of a benign fibrotic
nodule the maximum relative enhancement ratio was only 0.04, whereas the slope of enhancement was 0.0036/s. (Cases courtesy of
Dr. Ohno, Boston, Massachusetts.)
A, B C
Figure 6-42 Solitary nodule: 3T MR evaluation. A–C: Dynamic MR images in a 67-year-old woman with an 18-mm nodule in the left lower
lobe, acquired with exactly the same imaging parameters as shown in Figure 6-41 (arrow). Dynamic MR images—in this case (A) t 5.5 s,
(B) t 7.7 s, and (C) t 13.2 s—demonstrate a different pattern than shown in the case illustrated in Figure 6-41, with the maximum rela-
tive enhancement ratio of 0.95 and a slope of enhancement of 0.172/s. This case subsequently proved to be active infection due to atypical
mycobacteria. (Case courtesy of Dr. Ohno, Boston, Massachusetts.)
A B
D C
Figure 6-43 Nodule characterization: CT–positron emission tomography (PET) evaluation. See Color Figure 6-43D. A: Nonenhanced 5-
mm CT section obtained in quiet respiration as part of a combined PET-CT examination shows a small, slightly irregular nodule in the right
lower lobe. B, C: Corrected (top right) and uncorrected (bottom right) images, respectively, obtained from the 2-[fluorine-18]-fluoro-2-
deoxy-D-glucose (FDG)-PET scan show increased activity in the medial aspect of the right lower lobe when compared visually with back-
ground mediastinal activity (arrow in B). D: CT-PET fusion image (bottom left) shows that the increased activity in the PET images superim-
poses on the nodule identified on the corresponding CT scan shown in A. Use of combined CT-PET scanning enhances diagnostic accuracy
by simultaneously providing physiologic and anatomic data from the same study, allowing more accurate staging. Histologically verified
stage 1A non–small cell lung cancer.
B, C D
Figure 6-44 Nodule characterization: CT–positron emission tomography (PET) staging. A: CT section through the lower neck in a patient
with a lesion in the right upper lobe (not shown) following intravenous contrast administration shows no apparent adenopathy. B–D:
Select coronal, axial, and sagittal images, respectively, from a corresponding 2-[fluorine-18]-fluoro-2-deoxy-D-glucose (FDG)-PET scan show
marked activity in a lesion in the right upper lobe (B) as well as focal increased activity in a supraclavicular node on the right side (arrows in
B–D). Note the presence of markedly enhancing mediastinal and right hilar nodes. Despite apparent lack of adenopathy on the CT study, an
excisional biopsy was performed on the supraclavicular node, obviating more invasive surgical procedures, establishing this as a stage
3B lung cancer.
From the start, it has been apparent that FDG-PET is of value of delayed imaging also has been evaluated
value in differentiating benign from malignant nodules, (188,189). In one report, SUVs increased by 20% (p 0.01)
with initial studies reporting sensitivities ranging between when images obtained at 60 and 90 minutes were com-
93% and 100% and specificities between 82% and 88% pared, resulting in a sensitivity of 100% and a specificity of
(171,173,175,176). Meta-analyses similarly confirm that 89% for detecting malignant nodules (188). Despite these
PET is of value for staging lung cancer, especially when data, use of delayed scanning has not gained widespread
compared with CT, albeit with considerably wider variabil- acceptance, likely, in part, due to difficulties in scheduling.
ity in both sensitivity and specificity (177–179). Gould In another attempt to improve diagnostic accuracy, use
et al. (179), for example, in a meta-analysis of more than of SUV contrast ratios has been advocated. Nomori et al.
1,400 published cases, reported a mean sensitivity and (184), in a study of 161 nodules between 1 and 3 cm
specificity of 96% and 73.5%, respectively, for the use of (108 malignant and 53 benign), compared visual assess-
PET to stage lung cancer. Birim et al. (177), in a later meta- ment of lesions (classified into 3 grades, including defini-
analysis using receiver operating characteristic curves, tively positive, faintly positive, and negative) and SUVs,
documented somewhat less enthusiastic results, with an respectively, with contrast ratios between nodules and the
overall sensitivity of 83% and a specificity of 92%. contralateral lung. Using receiver operator curves (ROCs)
A number of issues regarding the use of PET for evaluat- to determine cut-off values for each method, these authors
ing pulmonary nodules remain, including the accuracy of noted that there was no significant difference between
FDG-PET for characterizing nodules less than 1 cm as well these techniques for lesions visually rated as definitely
as optimal methods for assessing FDG uptake [visual assess- positive or negative; however, of 17 lesions visually rated
ment vs. use of standardized uptake values (SUVs)]. These as faintly positive, whereas the SUV failed to detect any
factors and others have been shown to influence both the additional true positive cases, use of contrast ratios proved
sensitivity and specificity of FDG-PET for evaluating lung significantly more sensitive, allowing positive identifica-
nodules. tion of an additional 9 lesions (p 0.001). Although of
Although some have reported evaluation of lesions as interest, this approach also has yet to gain widespread
small as 5 mm (180,181), most investigators have deter- acceptance.
mined that PET is less accurate for nodules smaller than In addition to technical factors, it should be empha-
8 to 10 mm (180,182,183). This is because with most sized that the efficacy of PET for evaluating malignant
current PET scanners, lesion contrast decreases when the nodules is primarily a function of tumor cell volume,
diameter is smaller than twice the spatial resolution of the density, and avidity for FDG. In this regard, a number of
scanner. In one study of 87 patients evaluated surgically, studies have documented that the accuracy of FDG has
PET proved inaccurate in differentiating benign from proved both less sensitive and less specific for diagnosing
malignant lesions in 27% of nodules less than 1 cm, and adenocarcinomas (especially when well differentiated,
10% of lesions between 1 and 2 cm (182). Similar results including bronchoalveolar cell tumors) when compared
documenting the relative inaccuracy of PET for staging with other types of lung cancers, especially squamous
lesions less than 1 cm have been reported in an evaluation cell carcinomas (Fig. 6-45) (180,183,185). Other lesions
of 64 patients with clinical stage T1–2 tumors (183). that have been reported to yield a high percentage of false-
There is also little consensus in the literature regarding negative studies include carcinoid tumors, in particular.
the optimal method for assessing FDG uptake. Many studies Despite the occurrence of both false-positive and false-
make use of semiquantitative SUVs to evaluate PET scans, negative studies (Figs. 6-45 and 6-46), the potential for PET
especially when the SUV is greater than 2.5 (175,183–185). to provide prognostic information has been documented.
In distinction, others have argued that visual inspection In one study of 44 stage T1N0M0 adenocarcinomas, lesions
using background mediastinal activity as a control is at least with an SUV contrast ratio of less than 0.5 proved to have
equivalent or even superior to SUVs, as SUVs suffer a num- statistically less lymphatic, vascular, and pleural invol-
ber of limitations, including decreased sensitivity due to vement and were more likely to be well differentiated (186).
hyperglycemia, which has the effect of decreasing both the With a slightly different approach, similar results have been
blood clearance and accumulation of FDG in tumors, and reported in a separate retrospective review of 315 patients
variations resulting from the reliance on body weight to in whom nodules with high maximum SUVs (SUV 10)
estimate SUV without allowances made for individual dif- were statistically more likely to have poorly differentiated
ferences in the percentage of body fat (186). Indeed, the tumors presenting in an advanced stage, with patients having
mean SUV of malignant nodules has been reported to vary stage IB and II disease with mean maximum uptake values
from 5.5 to as high as 10.1 (186). greater than the median value compared with those in
In an attempt to improve diagnostic accuracy, a number the same stage, in particular, having a lower disease-free
of novel approaches have been investigated. It has been survival (185). Correlation between negative PET findings
suggested, for example, that the accuracy of PET can be and early-stage disease in patients with newly diagnosed
improved by correcting for the size of lesions, although to NSCLCs have led some investigators to recommend that
date this technique is not widely used (187). The added these patients may be safely managed conservatively with
Figure 6-45 Nodule character-

ization: false-negative CT–positron
emission tomography (PET) evalua-
tion. A: Magnified view through
the left upper lobe shows typical
appearance of a predominantly
ground-glass nodule within which
bubblelike lucencies representing
dilated bronchioles are apparent,
findings characteristic of non–small
cell lung cancer, likely bronchoalveo-
lar cell carcinoma. B: Coronal view
from a corresponding 2-[fluorine-
18]-fluoro-2-deoxy-D-glucose (FDG)-
PET scan shows no evidence of
activity in the lungs. Although bron-
choalveolar cell carcinoma (BAC) is
a well-established cause of a false-
negative FDG-PET study, when
associated with characteristic CT
appearances, it has been reported
to correspond to improved survival.
A, B Biopsy documented BAC.
A B
D C
Figure 6-46 Nodule characterization: CT–positron emission tomography (PET) correlations in inflammatory disease. See Color Figure
6-46D. A: Nonenhanced 5-mm CT section through the right upper lobe shows an irregular nodule causing apparent narrowing of a subseg-
mental bronchus, the same case as illustrated in Figure 6-11. B, C: Corrected (top right) and uncorrected (bottom right) images from a 2-[flu-
orine-18]-fluoro-2-deoxy-D-glucose (FDG)-PET study show increased activity in the medial aspect of the right upper lobe (arrow in B). D: CT-
PET fusion image (bottom left) shows the increased activity in the PET images, corresponding precisely with the nodule identified on the
corresponding CT study. In this case, as shown in Figure 6-11, subsequent follow-up showed complete resolution of these findings, despite
the initial clinical impression of neoplasm. E: Coronal section from an FDG-PET scan in a different patient than shown in A–D shows multiple
foci of increased activity in the right upper lobe, in this case due to active tuberculosis. These cases illustrate that, despite high sensitivity for
detecting neoplasia, FDG-PET is less accurate for differentiating active inflammation from tumor.
Nodule Doubling Time
V = 4⁄3(d⁄2)3
Double in Volume
d increase by 1⁄4
Figure 6-47 Estimates of nodule doubling time. Schematic
representation indicating that for a nodule to double in volume, its
diameter need only increase by 25%.
this degree of accuracy even with use of electronic calipers

(Fig. 6-49) (195). This becomes even more problematic
for slower growing malignancies with doubling times in
the range of 800 days, as typically occurs with small local-
ized BACs (Fig. 6-50) (196,197).
For these reasons, recent attention has focused on
obtaining automated segmentation and volume assessment
of nodules (Figs. 6-51 and 6-52) (194,198–202). Although
E accurate and reproducible assessment of the volume of
Figure 6-46 (continued) small nodules is dependent on methodology (199), this
approach represents a clear improvement over standard
serial CT follow-up studies to monitor for signs of growth bidimensional cross-sectional measurements. Measurement
(190). Clearly, such an approach will require further of volume doubling time (VDT) assumes that nodules are
prospective validation. essentially spherical: once nodule diameters are converted
to volumes, VDT can then be calculated if the time differ-
ence (t), initial volume (V0), and volume at time t (Vt) are
Growth Characteristics
In the setting of a nonspecific morphologic appearance,
differentiation between benign and malignant nodules Detecting Volume Change:
frequently requires assessing the presence of growth. As Small Nodules
a consequence of the large number of small, otherwise
volume volume
nonspecific, lesions less than 8 to 10 mm for whom further
doubling doubling
evaluation with contrast-enhanced CT, PET, or biopsy
remains problematic, considerable attention has focused
on developing reliable methods for assessing subtle 4 mm 5 mm 6.2 mm
changes in nodule size (191). Unfortunately, accurate
assessment of growth rates in small lung nodules,
especially those less than 5 mm, is often problematic,
particularly when reliance is placed solely on the use
of bidimensional perpendicular measurements (192).
Numerous studies, to date, have documented wide inter-
and intraobserver variability in obtaining accurate nodule
diameters (193,194). This is because a nodule doubles in 3 cm 3.75 cm 4.70 cm
volume when its diameter increases by only one fourth Figure 6-48 Schematic representation of progressive volume
(Figs. 6-47 and 6-48). A 3-mm nodule, therefore, doubles doublings of a hypothetical solid 4-mm nodule progressing to
4.7 cm. Note that each progressive volume doubling is repre-
in volume when it reaches 3.8 mm in diameter. Not sur- sented by an increase in the diameter of the nodule by one
prisingly, in routine clinical practice it is difficult to reach quarter (see Fig. 6-47).
Figure 6-49 Estimates of nodule

growth: use of electronic calipers.
A, B: Identical magnified images
through the left lower lobe show a
tiny solid nodule in the lung periph-
ery. The only difference between
these images is the apparent size
of the lesion as measured with elec-
tronic calipers. Note that, although
the nodule appears to measure
4 mm in A, the same nodule meas-
ures 3 mm in B, representing a
potential volume doubling, if real.
As exemplified in this figure, elec-
tronic calipers are of only limited use
for obtaining precise measurements
A, B of small pulmonary nodules.
known (201). Typically, VDTs for lung cancer range between sectional diameter and area measurements, volumetric
20 and 400 days, in distinction to benign nodules with measurements suffer from some important limitations.
VDTs either less than 20 days or greater than 400 days, as In addition to the fact that automated volume assessment
occurs in patients with hamartomas or granulomas. usually requires use of a standalone workstation, limita-
Unfortunately, despite obvious improvement over cross- tions have been noted in the reproducibility of these

growth: ground-glass opacities. A, B:
Magnified sequential 1-mm images
through the middle lobe show an ill-
A, B defined pure ground-glass lesion ap-
parently measuring 13.58 mm in
length using electronic calipers. C, D:
Magnified sequential 1-mm images
at the same exact levels as shown in
A and B, respectively, obtained 6
months later. The lesion now meas-
ures 17.99 mm in length. That there
has been unequivocal interval growth
is best ascertained not by use of the
electronic calipers, which are inexact
when attempting to precisely meas-
ure poorly defined ground-glass lesi-
ons, but by noting the lesion has
clearly increased in size when aligned
with adjacent anatomic reference
points, in this case adjacent periph-
C, D eral pulmonary vessels.

size: automated 3D segmentation.
A: Magnified view of a tiny central
nodule imaged with wide windows.
Accurate measurement of this lesion
would be extremely difficult using
electronic calipers. B: Image depicts
3D segmentation of the nodule
shown in A. Note that this technique
provides automated estimates of the
volume of the nodule, as well as its
cross-sectional dimensions in the x,
y, and z planes, as well as minimum
and maximum diameters measured
in off-axis planes. 3D segmentation
is more precise for analyzing small
nodules and is less likely to result in
interobserver variability, especially
important when accurate longitudi-
A, B nal measurements are required.
measurements for tiny (3 mm) nodules, as well as those It should be emphasized that opinions vary on whether
adjacent to blood vessels, airways, and pleural surfaces. growth rate is a sufficiently reliable parameter to be useful
Measurements are also affected by variations in acquisition in the management of patients. Although absence of growth
parameters (203) and motion artifacts, making use of identified on serial radiographs remains a useful indication
temporal comparisons over time especially problematic. of benignity (150), recent data—largely derived from lung
A, B
Figure 6-52 Estimates of nodule size: effect of
respiratory variations. A, B: Magnified views of a
tiny nodule in the right lung imaged only minutes
apart during the same examination in expiration
and inspiration. Note that there is a substantial
change in the measured volume of this nodule of
approximately 14% [10.3 vs. 12.0 mm in expira-
tion (A) and inspiration (B), respectively] when as-
sessed using automated 3D segmentation. (Image
courtesy of Carol Novak, PhD, Siemens Corporate
Research, Princeton, New Jersey.)
cancer screening trials—have called this assumption into and 33 of 102 patients (32%) undergoing major resections
question, especially for incidentally identified small non- (lobectomies or pneumonectomies) proved to have benign
solid (pure ground-glass) lung nodules. Although it is disease. Although more recent data have shown a larger
generally accepted that most lung cancers have doubling proportion of malignant nodules (113), the potential of
times in the range of 30 to 490 days, there is significant unnecessary surgical interventions remains problematic
variability in the growth rates of most cancers, especially (217,218).
adenocarcinomas of the lung (204). Hasegawa et al. (205), It should be emphasized that these data predate the era
for example, in a seminal study of 82 screen-detected of MDCT. Unfortunately, the use of MDCT has led to an
cancers, noted that although 61 of these had an average VDT explosion of finding small indeterminate lung nodules not
of between 52 and 1733 days, subsolid (ground-glass) nod- only in individuals undergoing LDCT screening but in vir-
ules had an average doubling time of 813 days. tually all patients undergoing CT examinations, regardless
of the clinical indication. Given these data, it is more criti-
cal than ever that newer approaches be developed at a
The Dormant Scar Carcinoma
minimum to ensure that unnecessary interventions—from
It has long been noted that cancers frequently occur at the the number of repeat imaging procedures to the number
periphery of old scars. Bennett et al. (206), for example, of unnecessary biopsies—be reduced to the greatest extent
found that almost half of adenocarcinomas of the lung in possible. In this regard, it is clear that the previous unitary
men were associated with pre-existent pulmonary scars. approach to the management of lung nodules is no longer
The cause of premalignant fibrosis is usually attributable tenable. Instead, recognition of the need to differentiate
to prior tuberculosis, old infarcts, or nonspecific granulo- between management options in patients with small nod-
mas. Active scars, presumably with increased metabolic ules of less than 8 to 10 mm and management options in
activity, may display peripheral areas of epithelial hyper- patients with larger nodules of between 1 and 3 cm is now
plasia, atypical metaplasia, and finally adenocarcinoma essential. Parenthetically, it is worth noting that lesions
(206). However, despite the fact that focal fibrosis and car- larger than 3 cm are appropriately designated masses and,
cinoma are frequently associated histologically, a causal especially in patients older than the age of 30 years, are
relationship between a scar and carcinoma is difficult to overwhelmingly likely to be malignant (90%). In these
prove. Furthermore, it is well known that carcinoma can cases, especially when the lesions are otherwise deemed
lead to active production of collagen and focal desmoplas- potentially resectable, management should proceed
tic reaction (207,208). Therefore, although focal areas of directly to biopsy and, if indicated, resection. Immediate
fibrosis may indeed be pathogenetically important in the surgery may also be applicable in patients for whom reli-
development of bronchogenic carcinomas in some able follow-up assessment many not be possible.
patients, the incidence of true scar carcinoma has probably Not surprisingly, such an approach will necessitate a
been overestimated in the past. An increase in the inci- reconsideration of the use of currently available methods
dence of bronchogenic carcinoma has been clearly shown, for evaluating pulmonary nodules, including clinical
however, especially in patients who have IPF or fibrosis assessment of the likelihood of malignancy (including the
secondary to asbestos exposure or, less commonly, pro- potential use of statistical methodologies such as bayesian
gressive systemic sclerosis, rheumatoid disease, dermato- analysis); imaging techniques used to characterize indeter-
myositis, or sarcoidosis (209–213). minate nodules; and indications for proceeding to lung
biopsy, via fiberoptic bronchoscopy (FB), transthoracic
needle biopsy (TTNB), or surgery. All things being equal,
MANAGEMENT the proper approach to the evaluation of a focal pulmo-
nary lesion should be conservative. Patients with a pre-
It is estimated that nearly 130,000 individuals or 52 of sumptive diagnosis of a benign nodule should nonetheless
every 100,000 people present each year with radiographic be followed with serial chest roentgenograms to confirm
evidence of pulmonary nodules (171). Unfortunately, to a lack of growth. Similarly, no lesion with a questionable
date, there is still little consensus concerning optimal or disturbing morphologic feature should be certified as
management of these patients (1,93,214). Patients with benign. In a difficult situation, in our opinion it is prefer-
indeterminate nodules are particularly problematic, as able to designate a benign mass as indeterminate (false
nearly 60% of nodules in most surgical series subsequently positive), with a goal of never diagnosing a malignant
prove benign. In an evaluation of pulmonary resections lesion as benign (false negative).
performed over a 20-year period in patients with solitary
lung nodules, for example, Ray et al. (215) found that
Factors Affecting Clinical Management
only 27 of 179 lesions (15%) were malignant. Only slightly
better results have been reported by Keagy et al. (216), who As previously discussed, clinical correlation continues to
found that 79 of 122 patients (65%) undergoing minor play a critical role in the assessment of patients with pul-
resections (biopsy, wedge resections, or segmentectomies) monary nodules. Characteristics such as older age, a history
of cigarette smoking, and a previous history of cancer all smaller than 5 mm (193,194,203); concerns regarding
increase the probability that an SPN is malignant (43). For radiation dose, especially resulting from cumulative expo-
example, a lung nodule in a patient older than the age of sures due to numerous follow-up studies (227); limita-
40 years, especially a smoker, is much more likely to be a tions resulting from inter- and intraobserver variability
bronchogenic carcinoma than a nodule seen in a younger (228); and cost (229).
patient or an immunocompromised patient presenting
with fever (219,220). In this regard a number of sophisti-
Workup of Indeterminate Nodules
cated methods for obtaining more precise estimates of
malignancy versus benignity for radiographically identified
Less Than 10 mm
nodules have been developed, including Bayes theorem, To date, a number of reports have addressed the specific
logistic regression models, and neural networks (221–223). issue of the appropriate time for CT follow-up of small
As reported by Gurney et al. (221), for example, using neu- (1 cm) indeterminate lung nodules. The reports include
ral networks trained and tested on 318 pulmonary nodules, those primarily focused on LDCT lung cancer screening
bayesian analysis led to a total of only 21 false-positive and (119,122,205), as well as serial observations of nodules
6 false-negative predictions for malignancy. identified on routine CT examination (121,225,226).
Similar attempts have also been made using PET data. Although specifics vary, there is an emerging consensus that
Dewan et al. (224) compared the utility of PET scanning initial recommendations for obtaining follow-up CT scans
with standard criteria using bayesian analysis in a retro- at 3-, 6-, 12-, and 24-month intervals are no longer appro-
spective study of 52 patients also undergoing CT evalua- priate for this group of lesions (230).
tion and found that the likelihood ratios for malignancy in Based on 5 years’ experience with LDCT screening of
SPN with abnormal FDG-PET scans was 7.11 (indicative of 1,520 individuals, Swensen et al. (119) recommend that
a very high probability of malignancy for positive studies) for lesions less than 4 mm and those between 4 and 7 mm
and 0.06 for benignity (indicative of a very high probabil- follow-up CT studies be performed at 12 months and
ity of benignity for negative studies). FDG-PET proved to 3 months, respectively, whereas those for nodules larger
be a better predictor of malignancy as a stand-alone test than 8 mm include an immediate diagnostic HRCT study.
than standard radiographic and clinical criteria either These recommendations correspond, at least in part, to
alone or, surprisingly, even in conjunction with FDG-PET those followed both by the PLCO study (52) and the
results (224). American College of Radiology Imaging Network (231)
Nonetheless, despite such promising results, as pub- regarding the need to differentiate lesions smaller than
lished in the American College of Chest Physician (ACCP) 10 mm as a discrete subgroup of pulmonary nodules and
Guidelines for management of patients with SPNs, and the need to obtain only yearly CT follow-up studies for
although attempts at statistical modeling are “laudable,” nodules of 4 mm. The need for only limited follow-up CT
as a rule they have proven to be “cumbersome” and of examinations for small lung nodules identified at LDCT
“little practical use” to clinicians (150). screening has also been documented by Henschke et al.
(122), who found that of 378 nodules smaller than 5 mm
in their screened population, none developed a cancer on
Follow-up Evaluation of Indeterminate
initial yearly follow-up incidence screenings.
Lung Nodules
Similar results have also been reported in nonscreened
As emphasized previously, based largely on data derived populations evaluated by CT. Benjamin et al. (120), in a ret-
from LDCT screening studies, a unitary diagnostic approach rospective review of 87 nonscreened cases in which lung
to the evaluation of lung nodules is currently no longer nodules smaller than 10 mm were identified (from among
tenable. Although the evaluation of nodules larger than 3,446 consecutive thoracic CT studies), for which definitive
1 cm in most cases can be reliably done with contrast- 2-year follow-up was available, found that, although 10
enhanced CT studies, PET, PET/CT (see previous sections), (11%) of these were malignant, 9 proved to be metastases in
or biopsy, an optimal approach to the management of small a patient with known extrathoracic malignancies. More
lung nodules (1 cm) remains problematic (25). recently, Piyavisetpat et al. (121), in a retrospective review of
Recommendations regarding the number and time inpatients identified from a total of 1,826 patients undergoing
tervals for follow-up of small indeterminate lung nodules CT for a noncalcified lung nodule and who had no history
by CT need to reflect the clinical likelihood of identifying of neoplasm, infection, fibrosis, or immunodeficiency,
incidental benign versus malignant nodules (120,226). A reported that no nodule less than 4 mm grew on follow-up
number of important additional considerations must also CT within a 12-month period.
be addressed, including the rate of nodule growth as a Based on these data, guidelines for the management of
reflection of both size and especially CT morphology small pulmonary nodules detected on CT scans have been
(solid vs. subsolid lesions) (58,60,205); the accuracy recommended in a consensus statement from the Fleischner
of available techniques for establishing cross-sectional Society (Table 6-2) (230). These call for differentiating “low-
and/or volumetric measurements, especially for nodules risk” from “high-risk” individuals based on smoking history
proper care (1,214). It is important to emphasize that the

choice of strategy for managing patients with pulmonary
TABLE 6-2
nodules entails considerations of the various levels of
FLEISCHNER TABLE
expertise (including those of the interventional radiolo-
Nodule Size Low-Risk Patient High-Risk Patient gist, bronchoscopist, cytologist, and surgeon) as well as
patient factors, in particular the effect of the patient’s risk-
4 mm No follow-up f/u 12 mo; no —stop
taking attitude toward testing strategies (232).
4–6 mm f/u 12 mo; no —stop f/u 6–12 mo; no —f/u In general, one of three broad strategies is usually
18–24 mo employed for nodules greater than 1 cm. These include
6–8 mm f/u 6–12 mo; no f/u 3–6 mo; no —f/u (a) additional imaging tests, followed by biopsy if interval
—f/u 18–24 mo 18–24 mo change is noted; (b) bronchoscopic and/or transthoracic
8 mm CT f/u 3, 9, 24 mo needle aspiration (TTNA) or biopsy, followed by contin-
or CT-PET, or biopsy ued observation or surgery as indicated; and (c) immediate
(TTNB/VATS) surgery. As noted by Cummings et al. (233), the choice
f/u, follow-up; no , no change; CT, computed tomography; PET, between these strategies may indeed be a close call. Using
positron emission tomography; TTNB, transthoracic needle biopsy; decision analysis, these authors showed that immediate
VATS, video-assisted thoracoscopy.
surgery produced a slightly longer life expectancy, pro-
vided the probability of cancer clinically was high, whereas
or other known risk factors (including known immuno- observation produced a slightly longer life expectancy in
deficiency, among others) as follows. patients with a low probability of malignancy (233). These
In “low-risk” individuals, nodules smaller than 4 mm differences, however, were extremely small. Unfortunately,
need not be followed up; those larger than 4 to 6 mm many studies purporting to evaluate optimal strategies
should be followed up yearly, without the need for follow- based on either decision analysis or cost-effectiveness have
up if unchanged; those larger than 6 to 8 mm should be been reported prior to the introduction and widespread
followed up in 6 to 12 months and then at 18 to 24 months availability of PET scanning or video-assisted surgery,
if unchanged; and those larger than 8 mm should be in particular, limiting their value (233,234).
followed up more aggressively at 3, 9, and 24 months with Given the many considerations outlined previously, our
dynamic contrast-enhanced CT, PET, or biopsy as options. general recommendations for nodules between 1 and 3 cm
In distinction, in “high-risk” individuals, nodules smaller are as follows:
than 4 mm need to be followed up only once at 12 months,
provided there is no change in size; those larger than 4 to 1. Nodules that are unchanged in size for 2 years or that
6 mm should be followed up initially at 6 to 12 months show benign patterns of calcification or fat may
and then in 18 to 24 months if unchanged; those larger be safely observed (Figs. 6-13, 6-17, and 6-18). It should
than 6 to 8 mm should be followed up initially at 3 to be emphasized, however, given numerous reports
6 months, then subsequently at 9 to 12 and 24 months if documenting the propensity for some adenocarcinomas,
unchanged; and those larger than 8 mm to be followed up especially slow-growing BACs with predominant
in the same manner as the low-risk group. ground-glass attenuation, to grow extremely slowly, that
It should be noted that controversy remains regarding continued yearly radiologic or LDCT surveillance exami-
the duration of follow-up for both part solid and especially nations should be obtained as clinically indicated.
pure subsolid (ground-glass) nodules (58,60,205). As a 2. In patients with indeterminate CT studies, either a dedi-
consequence, long-term follow-up extending over years cated contrast-enhanced CT nodule study (Figs. 6-36
should be requisite in the appropriate clinical setting, espe- and 6-40) or preferentially FDG-PET imaging (Figs. 6-43
cially if there is an antecedent history of lung cancer. In all and 6-44), when available, should be performed.
cases, whenever possible and regardless of the indication, Although both studies suffer important limitations both
follow-up studies should be performed with the lowest in sensitivity and in specificity, these examinations
possible radiation dose (ideally between 40 and 80 mA may prove invaluable by (a) increasing the degree of
whenever possible) to minimize cumulative radiation certainty that a lesion is benign, especially in patients in
exposure in individuals for whom multiple follow-up CT whom the suspicion of malignancy is low, as, for exam-
examination may be anticipated. ple, patients below the age of 40, and (b) in the case of
FDG-PET, adding invaluable additional findings perti-
nent to staging malignant nodules.
Workup of Indeterminate Nodules 1 to 3 cm
3. In patients for whom tissue diagnosis is deemed neces-
It is apparent that despite considerable technical advances, sary, FB should be reserved only for those patients
both radiographic and surgical, optimal management of in whom lesions either are clearly endobronchial on
patients with solitary nodules still requires considerable CT (Figs. 6-15, 6-33 and 6-34), or for whom TBNA
clinical acumen. No single algorithm applies in all cases, or biopsy is deemed appropriate for definitive staging
making clinical judgment an absolute prerequisite for of mediastinal nodes.
4. TBNA or biopsy should be performed preferentially in entails considerably greater cost, given the need for gen-
patients with indeterminate nodules, especially those eral anesthesia in most cases, coupled with the fact
deemed likely to be benign, or in patients with known that 30% to 40% of cases require conversion to thora-
extrathoracic malignancies in whom metastatic dis- cotomy either for definitive resection of lung cancer or
ease is likely (Fig. 6-53). In these cases, core needle because of inability to localize disease. Therapeutic
biopsies should be performed if the initial cytology indications yet to be validated include video-assisted
proves nonspecific or when more definitive characteri- lobectomies and metastatectomies. The use of either
zation of malignancy is indicated, especially in CT-guided wire or endoscopic ultrasound localization
patients with suspected lymphoma. Although core or localization by presurgical injection of radioisotope
needle biopsies result in greater morbidity, this is markers may be of value in select cases, especially when
more than offset by the higher yield of true benign nodules are central in location (235).
diagnoses. CT guidance is recommended particularly
for those cases in which nodules are small, central in
Biopsy
location, or difficult to identify on biplane fluoroscopy.
5. VATS should be considered primarily a diagnostic Because of the limitations of currently available noninvasive
modality and reserved for those cases in which alterna- methods for assessing nodules, biopsy is often required for
tive methods prove nondiagnostic and/or for cases in a definitive diagnosis. It is also not surprising that as with
which routine surgical resection is considered either other aspects of SPN evaluation there are differences of
too risky or contraindicated. Routine use of VATS opinion regarding the best method for performing nodule
biopsies. In addition to sputum cytology, currently available
methods include FB, TTNB, and surgery, including VATS.
Sputum Cytology
The accuracy of sputum cytology is dependent on a number
of factors, including the number of specimens collected and
lesion location, among others. In a systematic review of
available data collected as background for the Clinical
Practice Guidelines on Lung Cancer, Scheiber and McCrory
(236) reported a pooled sensitivity and specificity of
0.66 and 0.99, respectively, with sensitivity higher for central
versus peripheral lesions. In our experience, even with a pos-
itive sputum cytology, additional invasive procedures are al-
most always necessary to further assess and stage lesions.
Fiberoptic Bronchoscopy
Routine FB is of limited value in the assessment of lung
nodules (237–239). Although extremely accurate for assess-
ing focal endobronchial lesions, including those within
lobar and segmental airways, the yield of FB for peripheral
nodules is considerably poorer. In one review of 30 studies
with endobronchial lesions evaluated bronchoscopically,
the overall diagnostic sensitivity was 0.88 when results from
endobronchial biopsy, cytobrushing, and washing were
combined (236). In distinction, the overall diagnostic sensi-
tivity for peripheral lesions proved considerably worse, with
an average reported sensitivity of only 0.69 when cytobrush-
ing, transbronchial biopsy, and bronchoalveolar lavage
(BAL) or washing are combined (236). Although a number
of reasons account for this lower sensitivity, including
whether or not bronchoscopy is performed under fluoro-
Figure 6-53 Transthoracic needle biopsy (TTNB). Magnified scopic guidance, as well as variations in bronchoscopists’
view from a CT-guided TTNB confirming that the tip of the needle experience, not surprisingly the most important determi-
is in the lesion. This proved to be a non–small cell lung cancer. The nant is the size of lesions. Evaluating the yield of cytobrush-
overall sensitivity of TTNB is greater than 90% for diagnosing
peripheral lung lesions and is especially appropriate for lesions ing, transbronchial biopsy, and BAL together, although
less than 2 cm. overall reported sensitivity is 0.69, the sensitivity for lesions
smaller than 2 cm was 0.33 versus 0.62 for lesions larger nodules, a number of newer imaging techniques have
than 2 cm (236). Although the yield of FB increases when been developed. These include the use of multiplanar
only those cases in which nodules are associated with air- reconstructions to enhance visualization of peripheral
ways are selected when evaluated by CT, even in this setting, airways (241); ultrathin bronchoscopy, with (248,249)
the sensitivity of FB increases only to approximately two or without virtual bronchoscopic guidance (250–253);
thirds of cases (137,141,240,241). FB has also proved of endoscopic ultrasonography (EBUS) (254,255); and elec-
only limited value in diagnosing benign disease, even in tromagnetic navigation (256,257). Although promising,
situations in which the prevalence of benign disease is high, these techniques will require further validation prior to
as in select populations with suspected tuberculosis (242). gaining widespread acceptance.
An important consideration for the use of FB is the case
in which a parenchymal nodule(s) is associated with either
Transthoracic Needle Biopsy
hilar or more importantly mediastinal adenopathy. In these
cases, the ability to perform transbronchial needle aspira- The use of TTNA and/or TTNB for assessing solitary
tions and biopsies (TBNA or Wang needle biopsies) using nodules is well established. To date, numerous studies have
CT as a road map, in the appropriate hands, may allow de- shown that the accuracy of needle aspiration for diagnosing
finitive staging of lung cancer in a significant proportion of malignancy is uniformly greater than 90%, although a
cases (243–247). TBNA allows cytologic and/or histologic trend toward lower sensitivity for lesions smaller than 2 cm
sampling in select cases, obviating more invasive surgical recently has been noted (Fig. 6-53) (236). Advances in this
procedures, including mediastinoscopy and thoracotomy. technique, including the use of CT guidance and/or CT
In most series, TBNA has resulted in sensitivities between fluoroscopy, immediate cytopathology, the introduction of
60% and 80% for diagnosing and staging lung cancer. core biopsy techniques, and postbiopsy positional restric-
In an attempt to improve the diagnostic accuracy tions to limit the incidence of pneumothoraces, have made
of transbronchial biopsy for assessing peripheral lung this option still more attractive (Fig. 6-54).
A B
Figure 6-54 Transthoracic needle biopsy: complications. A: CT

section through the right upper lobe shows the presence of a 21-
gauge aspiration needle in close proximity to a right upper lobe
nodule. B: Section obtained at approximately the same level as
A following withdrawal of the needle shows ill-defined airspace
consolidation appearing adjacent to the nodule, resulting from
hemorrhage. Despite this finding the patient reported no undue
discomfort, nor was there evidence of subsequent hemoptysis.
Cytologically verified non–small cell lung cancer. C: CT section
through the middle lobe in a different patient than seen in A and B
shows a 21-gauge cytology needle anteriorly with the tip within a
right upper lobe nodule. In this case, transthoracic biopsy resulted
in a small pneumothorax (arrow), incidentally confirming that the
lesion is separate from the adjacent parietal pleura. Despite this
complication, aspiration was successful, confirming this lesion as a
C non–small cell lung cancer, subsequently resected.
In our judgment, in most cases the true utility of TTNB be noted that in this series, originating from the American
is in diagnosing benign disease (258). Unfortunately, Southwest, 30% of benign lesions were caused by coc-
despite some reports suggesting that a benign diagnosis cidioidomycosis, an unusual distribution compared with
could be established in as many as 64% of lesions (259), most other series. The use of an automated spring-loaded
in most series the accuracy of TTNA for definitive diagno- gun allows better penetration of firm lesions less
sis of benign disease has proved limited, typically below amenable to fine aspirating needles. Another advantage of
50% (260,261). Calhoun et al. (262), in an early study core needle biopsy is the ability to more precisely charac-
using fine needle aspiration under fluoroscopic guidance, terize malignant lesions, including pulmonary metastases
for example, were able to definitively diagnose only and especially parenchymal lymphoma. Unfortunately,
16 (12%) of 132 cases interpreted as showing no evidence the incidence of pneumothoraces does increase to as high
of malignancy on initial aspiration biopsy, and 1 of as 54% with use of a core needle biopsy gun (258). In dis-
these 16 subsequently proved malignant. Furthermore, in tinction, although the incidence of pneumothoraces
the same group of 132 patients, 38 lesions (29%) subse- resulting from TTNA has been reported to be as high
quently proved malignant. as 50%, in most series, the incidence averages around
These data emphasize the need to clearly distinguish 20% to 25%, with the need for chest tube placement
between a true benign versus a nonspecific diagnosis at occurring in less than 5% (264). As noted by Kazerooni et
TTNA and biopsy. Included in the former category are al. (268), the risk of complications does increase with
hamartomas, granulomas, infectious etiologies in which smaller lesions, especially when centrally located. Given
microorganisms can be demonstrated, including lung these data, the use of core needle biopsy is advocated
abscesses, and cores of fibrous tissue; included in the in patients in whom the suspicion of malignancy is
latter category are the findings of blood, histiocytes, low or for whom initial cytologic evaluation has proved
necrosis, inflammatory cells, alveolar lining cells, and nonspecific.
bronchial epithelium (262,263). A number of reasons It should be emphasized that although the accuracy of
have been cited to explain the low yield of TTNB for estab- TTNA for peripheral lesions overall is 90%, TTNA is of
lishing a benign diagnosis. These include an insufficient limited use for evaluating pure subsolid, ground-glass
number of biopsies, lack of on-site cytology, technical lesions, especially when these are less than 1 cm. Given
inexperience, and small lesion size, among others. Use of the small sample size, differentiation between AAH, BAC,
repeated biopsies has been shown to increase the likeli- and invasive adenocarcinoma remains problematic, and,
hood of a benign diagnosis by biopsies as much as in these cases, surgical biopsy is indicated for definitive
30% (264). Similarly, the presence of on-site cytologic diagnosis.
evaluation has been shown to reduce the number of false-
negative studies in at least some studies. The effect of
Video-Assisted Thoracoscopy
small lesion size (1.5 cm) is somewhat more controver-
sial (263,265,266). Li et al. (265) have reported a signifi- Advances in video technology have led to the use of VATS
cant difference in the diagnostic accuracy of TTNB for for a number of indications, including assessment of inde-
small and large lesions (74% vs. 96%, respectively). More terminate pulmonary nodules (213,269,270). VATS is per-
recently, however, Wescott et al. (263), in a study of formed using a double-lumen endobronchial tube to allow
74 biopsies of 64 small (less than 1.5 cm) lesions, includ- ventilation of the contralateral lung, usually under general
ing 21 benign lesions and 43 malignant lesions, reported anesthesia. Following collapse of the ipsilateral lung, three
a sensitivity of 93% and a specificity of 100% with an incisions are typically made with introduction of a telescope
accuracy of 95%. As noted by these authors, discrepancies coupled to a video camera; if necessary, these incisions also
in the diagnostic accuracy of TTNB for small lesions likely allow insertion of a palpating finger in those cases in which
reflect a combination of experience and the number of peripheral abnormalities are not apparent by inspection.
attempted biopsies. With these issues in mind, in most Wedge and, less commonly, lobar resections are performed
centers, cases that meet truly benign criteria can now either using an endoscopic stapler or, less frequently,
safely be managed conservatively. neodymium:yttrium-aluminum-garnet (YAG) laser resec-
The ability to establish a definitive benign diagnosis tion (271). Following the procedure, a chest tube is always
may be improved significantly with coaxial core needle inserted.
biopsy using an automated cutting biopsy needle (267). Initial results using video-assisted thoracoscopic exci-
The advantage of this approach is inherent in the ability sional biopsy have confirmed that this is a viable approach
to acquire histologic sections. Klein et al. (258), in a retro- for diagnosing peripheral pulmonary nodules. Mack et al.
spective study comparing the accuracy of routine cytology (272), in a study of 242 indeterminate pulmonary nodules,
versus core needle biopsy, found that although no differ- were able to establish definitive diagnoses in all patients,
ence could be seen for diagnosing malignancy, fine needle including benign diagnoses in 127 patients who otherwise
aspiration yielded a benign diagnosis in only 44% of may have required thoracotomies (52%) (272). Viewed
cases as compared with 100% using core biopsy. It should from a different perspective, of course, it is unclear how
many of these lesions may have been successfully diagnosed FUTURE DIRECTIONS: COMPUTER-
with modern TTNB techniques. Of the 127 patients with AIDED DIAGNOSIS
benign diagnoses, 12 had pulmonary hamartomas and
34 had granulomas, diagnoses that may have been estab- CAD is best defined as a method of assisting radiologic inter-
lished definitively using TTNB for which resection is unnec- pretation by means of computer image analysis (279,280).
essary (272). Furthermore, in the 115 patients diagnosed Ideally, the result is improved decision making and perform-
with malignant nodules, 64 (56%) of the nodules proved to ance due to enhanced detection and evaluation of complex
be metastatic in origin, a diagnosis again that preferentially imaging features, decreased interobserver variability, and
should be made prior to thoracoscopy. elimination of otherwise repetitive or tedious tasks.
In most large series initially reported, as many as 25% CAD can be used to include any computer-aided tool
of cases required conversion to thoracotomy, because of or postprocessing algorithm applicable to routine axial
the need for more extensive or definitive resection in pa- images. In addition to automated segmentation techniques
tients with resectable lung cancers, inability to detect and detection algorithms, this includes such seemingly
pathology, prior adhesions, bleeding, and/or equipment mundane applications as use of gray level image manipula-
failure. As reported by Hazelrigg et al. (270), in a coopera- tion, electronic calipers, and a variety of postprocessing
tive study of 1,820 cases, 865 of which were performed for image algorithms such as MIP and minimum intensity pro-
indeterminate pulmonary nodules, there were 65 cases jection (MinIP) images, rotating multiplanar reconstruc-
(7.5%) in which nodules could not be detected. Especially tions (MPRs), and virtual bronchoscopy. As suggested by
problematic then and now are central lesions variously de- Summers (281), CAD has the potential to serve as a
fined as greater than 1 to 3 cm from adjacent visceral method for global assessment of the thorax, including eval-
pleura or adjacent to segmental or subsegmental pul- uating the presence and severity of atherosclerosis (calcium
monary vasculature (273). In this setting, the diagnostic scoring and aneurysm detection) as well as bony pathology
accuracy of VATS may be as low as 67%. As a result, a num- as measured by bone densitometry and vertebral height.
ber of investigators have suggested CT-guided wire place- Presently, the major impetus for the development of CAD
ment using a variety of needles with or without injection is detecting lung nodules. This is largely the result of the
of methylene blue dye prior to VATS (274–276). Using a widespread availability of MDCT scanners capable of pro-
Kopans wire, Shah et al. (274) were able to successfully viding as many as 250 to 350 images per case, representing
place hook wires in 16 of 17 nodules prior to VATS with a clear challenge to routine daily clinical interpretation.
only one of these dislodging. Similarly, Shepard et al. Major indications that have been proposed for lung nodule
(275), also using a Kopans hook wire, reported success- evaluation with CAD include improved sensitivity or detec-
fully positioning 8 of 10 wires with only 2 dislodging prior tion, the ability to evaluate complex imaging features for
to surgery. Asamura et al. (277) have suggested use of CT- nodule characterization, elimination of repetitive tasks
guided injection of a metallic coil with subsequent using automatic nodule registration, and decreased inter-
thoracoscopic resection under fluoroscopic guidance as observer variability.
a method to improve thoracoscopic localization. More CAD devices can serve in two principal manners: to
recently, additional aids to improve localization include enable easier identification of findings or to function auto-
intraoperative ultrasonography and use of radiotracer lo- matically to help identify and/or characterize lesions.
calization techniques (235). This latter approach involves Automated CAD offers the potential not only to decrease
the preoperative percutaneous injection of gamma-emit- detection and recognition errors as a second reader but also
ting radioisotopes, such as technetium-labeled macroag- to reduce mistakes related to misinterpretation (282–288).
gregated albumin typically used for lung perfusion studies, Published reports to date have demonstrated consider-
that can subsequently be localized at surgery using a radio- able variability in CAD sensitivity for detecting pulmonary
probe. Preliminary experience using commercially avail- nodules, ranging between 38% and 95% (282,284–287,
able vascular embolization microcoils presurgically 289–293). Unfortunately, almost all studies reported to
delivered percutaneously via a 22-gauge needle to mark this time have been performed without pathologic docu-
the location of small peripheral nodules has also been re- mentation, limiting validation. In addition to a lack of a
ported (278). “gold standard,” numerous other limitations can be iden-
Despite these reports, in most cases, prethoracoscopic tified, leading both to suboptimal CAD performance and
nodule localization may not be necessary, provided sur- to difficulties in comparing CAD devices. These include,
geons make adequate use of finger palpation at the time of among others, differences among various CAD algorithms,
thoracoscopy (213). As pointed out by Westcott (263), it is leading to inaccurate segmentation of nodules, especially
difficult to follow the logic of performing wire localiza- those adjacent to the pleura, fissures, and blood vessels, or
tion—or, more recently, radionuclide injection—without those predominantly subsolid ground-glass in density;
first attempting needle aspiration or biopsy, given that the partial volume averaging due to both respiratory and
procedure requires identical skills, equipment, and expense cardiac motion; and use of thick (5 to 10 mm) CT sections
as routine CT-guided TTNB. for both routine CT and LDCT studies. Additional
confounding factors include measuring the impact of CAD and equaled CAD performance in interpreting 2-mm sec-
on reader vigilance as well as assessing variations in case tions, whereas CAD outperformed readers when contigu-
selection. Further confounding analysis is the fact that ous 0.075-mm images were reconstructed (297).
considerable variability has been documented between In one study of 68 nodules greater than or equal to
readers, even among “experts,” not only for identifying 5 mm, Wormanns et al. (292), using a CAD system specifi-
nodules but also for the definition of “actionability.” cally designed to detect nodules greater than or equal to
Nodule characteristics, in particular, affect CAD per- 5 mm, found that although radiologists had a sensitivity of
formance. In a study of 20 cases with 135 nodules inter- 85% compared with 38% for CAD, 15% of nodules greater
preted by two readers, CAD proved superior for nodules than 5 mm would have been missed without CAD. Given
smaller than 5 mm (p 0.5) and for those without vascu- that most CAD studies to date have reported sensitivities in
lar attachments (p 0.006), whereas readers outperformed the range of 70% to 80%, these data suggest that CAD may
CAD for nodules with vascular attachments (p 0.001) play an important role as a second reader, especially for
(294). Overall, CAD had a sensitivity of 76.3% compared studies performed with MDCT scanners. This includes cases
with 52.6% for both readers combined. As important, read- initially interpreted as normal, as well as cases for which
ers detected a greater number of false-positive findings per the primary indication is cancer detection. In one study of
case, ranging between 0.15 and 0.25 per case compared 100 normal CT examinations referred for a variety of clini-
with 0.55 for CAD. Not surprisingly, slice thickness also has cal indications in which nodules were identified as being of
been well documented to affect the utility of CAD. In addi- high (10 mm), medium (5 to 9 mm), or low (5 mm)
tion to an unacceptable number of false-positive cases per significance, Peldschus et al. (298) reported that a total of
study (often per slice) (290,295,296), use of thick sections 53 lesions in 33% of patients were identified, of which
adds little to overall diagnostic accuracy. In one study com- 5 (9.4%) were rated high and 21 (36.9%) were rated inter-
paring CAD to reader performance in the evaluation of mediate in significance.
20 patients with nodules less than 10 mm, for example, Of equal importance is the potential for CAD to identify
readers outperformed CAD in interpreting 4-mm sections otherwise overlooked lung cancers (Fig. 6-55). Armato et al.
A B
Figure 6-55 Missed lung cancer: computer-assisted diagnosis (CAD). See Color Figure 6-55E–H. A, B: Magnified CT images through
the left lung imaged with wide and narrow windows, respectively, in a patient who had had a prior right upper lobe resection for
non–small cell lung cancer. These images were interpreted as showing no significant abnormalities. C, D: Magnified CT images obtained
4 months later as part of routine CT surveillance now show a central lobular nodule (arrow in C) and enlarged mediastinal nodes consis-
tent with recurrent tumor. E–H: Retrospective analysis of the initial dataset evaluated on a standalone workstation using specialized soft-
ware (LungCare, Siemens Medical Solutions, Inc, Malvern, PA), including a user-initiated CAD program, shows that the enlarging nodule
seen in C was identified as a potential candidate nodule on the initial postoperative study. The size of this lesion is most accurately as-
sessed employing 3D segmentation (F). The potential for CAD to be used as a second reader, especially in cases with a high clinical sus-
picion for malignancy, is apparent (continued ).
C D
E, F
G, H
(34), for example, in a study of 38 LDCT screening studies (19 of 37 benign nodules), and an accuracy of 77.6% (83
performed with 10-mm collimation, identified a total or of 107 nodules).
38 missed cancers, including 23 due to detection errors Although it is apparent that CAD is still clearly in a
(20 of 23 10 mm, 12 5 mm) and 15 due to interpretive development stage, it can be recognized that CAD has the
errors; of these, CAD was able to successfully identify 32 potential to profoundly alter current practice in a number
(84%). Using a more sophisticated approach, Li et al. (280), of ways. This assumption underlies the rationale for the
in a study of 17 missed peripheral adenocarcinomas multi-institutional consortium—the Lung Image Database
matched with 10 control studies, employed a CAD scheme Consortium—organized with the explicit intent of creating
utilizing “difference images” (subtracting nodule-sup- a Web-based database for evaluating various computer-
pressed images from nodule-enhanced images) to compare aided diagnostic techniques (http://cip.cancer.gov). Inten-
CAD performance with 14 readers who evaluated cases first ded to provide a standardized set of images to compare
without and then with CAD. The average sensitivity for different CAD algorithms, this approach will require
detecting the 17 cancers improved from an average of 52% developing a consensus on a wide variety of technical and
to 68% (p 0.001). clinical issues. In addition to answering basic questions,
In addition to detection, there are numerous potential such as defining optimal techniques of scan acquisition
applications for CAD as a means to characterize nodules. and reconstruction, as well as ensuring accurate nodule
These include automatic 3D segmentation to obtain lung measurements, of still greater importance is the question
nodule volumes as well as automated contrast-enhanced of how such a database will be compiled in the absence of
densitometry, diagnostic feature extraction, and, perhaps a pathologic gold standard. This will necessarily entail
in the future, data mining correlating imaging features addressing such issues as variations in the definition of
with clinical and physiologic parameters and peripheral nodules. Also of concern will be questions related to
biomarkers. matching cases and readers across modalities as well as
Nodule characterization using automated, quantitative issues related to location uncertainty and the reader.
methods offers the potential to better interpret the rela- Although advanced statistical methods are being devel-
tionship of specific characteristics to the likelihood of oped to address many of these issues, ultimately a distinc-
malignancy or benignity. In clinical practice, morphologic tion will have to be made between measurements of CAD
assessment has primarily focused on visual assessment of utility on the one hand and assessment of clinical utility
a nodule’s borders (299), but with computer techniques, on the other. These caveats aside, however, it is to be antic-
internal architecture and attenuation assessments are now ipated that the creation of a robust database accessible to
aspects that can be documented in an easily reproducible all with an interest in developing CAD methodologies
manner. As most CAD algorithms designed for nodule represents an important advance in gaining clinical accept-
detection extract nodule characteristics to classify candidate ance for CAD.
ROIs, these same CAD algorithms can also be applied to
analyzing nodule morphology.
Using contiguous CT sections less than or equal to REFERENCES
3 mm and SPNs in 31 patients, McNitt-Gray et al. (300),
for example, applied a stepwise discriminant analysis to 1. Webb WR. Radiologic evaluation of the solitary pulmonary
identify combinations of quantitative nodule features nodule. AJR Am J Roentgenol. 1990;154:701–708.
2. Shaham D, Guralnik L. The solitary pulmonary nodule: radio-
that best characterized nodules as benign or malignant. logic considerations. Semin Ultrasound CT MRI. 2000;21:97–115.
Employing quantitative measures, including nodule tex- 3. Hartman TE. Radiologic evaluation of the solitary pulmonary
ture, 90.3% of 31 cases were correctly classified as either nodule. Radiol Clin North Am. 2005;43:459–465.
4. Potchen EI, Bisesi MA. When is it malpractice to miss lung
benign or malignant using linear discriminant analysis. cancer on chest radiographs? Radiology. 1990;175:29–32.
Using a different approach, Kawata et al. (301) proposed 5. Quekel LGBA, Kessels AGH, Goei R, vanEngelshoven JMA. Miss
the use of a previously established 3D database of proved rate of lung cancer on the chest radiograph in clinical practice.
Chest. 1999;115:720–724.
lung nodules—both benign and malignant—that can 6. Kalender WA, Seissler W, Klotz E, Vock P. Spiral volumetric CT
be automatically retrieved to exhibit morphologic and with single-breath-hold technique, continuous transport, and
density characteristics consistent with indeterminate nod- continuous scanner rotation. Radiology. 1990;176:181–183.
7. Costello P, Anderson W, Blume D. Pulmonary nodule: evalua-
ules identified. Using a combination of image-based tion with spiral volumetric CT. Radiology. 1991;179:875–876.
elements (including shape, curvature indices, and vector 8. Vock P, Soucek M, Daepp M, Kalender WA. Lung: spiral volumet-
fields as well as features of the surrounding lung ric CT with single-breath-hold technique. Radiology. 1990;176:
864–867.
parenchyma) derived from histologically confirmed 9. Kalender WA, Polacin A, Suss C. A comparison of conventional
benign and malignant lesions, as well as text-based fea- and spiral CT: an experimental study on the detection of spheri-
tures including age, sex, and smoking history, these cal lesions. J Comput Assist Tomogr. 1994;18:167–176.
10. Buckley JA, Scott WW, Siegelman SS, et al. Pulmonary nodules:
authors were able to compute a similarity coefficient and effect of increased data sampling on detection with spiral
likelihood of malignancy to record a sensitivity of 91.4% CT and confidence in diagnosis. Radiology. 1995;196:
(64 of 70 malignant nodules), a specificity of 51.4% 395–400.
11. Diederich S, Lentschig MG, Winter F, et al. Detection of pul- 35. Müller NL, Gamsu G, Webb WR. Pulmonary nodules: detection
monary nodules with overlapping vs non-overlapping image using magnetic resonance and computed tomography. Radiology.
reconstruction at spiral CT. Eur Radiol. 1999;9:281–286. 1985;155:687–690.
12. Diederich S, Wormanns D, Semik M, et al. Screening for lung 36. Feuerstein IM, Jicha DL, Pass HI, et al. Pulmonary metastases:
cancer with low-dose spiral CT: prevalence in 817 asymptomatic MR imaging with surgical correlation—a prospective study [see
smokers. Radiology. 2002;222:773–781. comments]. Radiology. 1992;182:123–129.
13. McWilliams A, Mayo J, MacDonald S, et al. Lung cancer screen- 37. Panicek DM. Editorial. MR imaging for pulmonary metastases?
ing—a different paradigm. Am J Respir Crit Care Med. 2003;168: Radiology. 1992;182:10–11.
1167–1173. 38. Doppman JL, Pass HI, Nieman LK, et al. Detection of ACTH-
14. Swensen SJ, Jett JR, Hartman TE, et al. Lung cancer screening producing bronchial carcinoid tumors: MR imaging vs CT. AJR
with CT: Mayo Clinic experience. Radiology. 2003;226:756–761. Am J Roentgenol. 1991;156:39–43.
15. Swensen SJ, Jett JR, Sloan JA, et al. Screening for lung cancer 39. Schroeder T, Ruehm SG, Debatin JF, et al. Detection of pul-
with low-dose spiral computed tomography. Am J Respir Crit monary nodules using a 2D HASTE MR sequence: Comparison
Care Med. 2002;165:508–513. with MDCT. AJR Am J Roentgenol. 2005;185:979–984.
16. Gurney JW. Missed lung cancer at CT: imaging findings in nine 40. O’Brien EJ, Tuttle WM, Ferkaney JE. The management of
patients. Radiology. 1996;199:117–122. the pulmonary “coin” lesion. Surg Clin North Am. 1948;28:
17. White CS, Romney BM, Mason AC, et al. Primary carcinoma of 1313–1322.
the lung overlooked at CT: analysis of findings in 14 patients. 41. Steele JD. The solitary pulmonary nodule. Report of a coopera-
Radiology. 1996;199:109–115. tive study of resected asymptomatic solitary pulmonary nodules
18. Henschke CI, Naidich DP, Yankelevitz DF, et al. Early lung cancer in males. J Thorac Cardiovasc Surg. 1963;46:21–39.
action project: initial findings on repeat screenings. Cancer. 42. Walske BR. The solitary pulmonary nodule. Dis Chest. 1966;49:
2001;1:153–159. 302–304.
19. Kakinuma R, Ohmatsu H, Kaneko M, et al. Detection failures in 43. Swensen SJ, Silverstein MD, Ilstrup DM, et al. The probability of
spiral CT screening for lung cancer: analysis of CT findings. malignancy in solitary pulmonary nodules. Application to
Radiology. 1999;212:61–6. small radiologically indeterminate nodules. Arch Intern Med.
20. Wormanns D, Ludwig K, Beyer F, et al. Detection of pulmonary 1997;157:849–855.
nodules at multirow-detector CT: effectiveness of double read- 44. Taylor RR, Rivkin LN, Salyer JM. The solitary pulmonary nodule.
ing to improve sensitivity at standard-dose and low-dose chest Ann Surg. 1958;147:197–202.
CT. Eur Radiol. 2005;15:14–22. 45. Davis EW, Peabody JE, Katz S. The solitary pulmonary nodule.
21. Hazelrigg SR, Boley TM, Weber D, et al. Incidence of lung nod- J Thoracic Surg. 1956;32:728–770.
ules found in patients undergoing lung volume reduction. Ann 46. Holin SM, Dwork RE, Glaser S, et al. Solitary pulmonary nodules
Thorac Surg. 1997;64:303–306. found in a community-wide chest roentgenographic survey: a
22. Shah PK, Austin JHM, White CS, et al. Missed non-small cell five year follow-up study. Am Rev Tuberc. 1959;79:427–439.
lung cancer: radiographic findings of potentially resectable 47. McClure CD, Boucot KE, Shipman GA, et al. The solitary pul-
lesions evident only in retrospect. Radiology. 2003;226:235–241. monary nodule and primary lung malignancy. Arch Environ
23. Remy-Jardin M, Remy J, Giraud F, Marquette CH. Pulmonary Health. 1961;3:127–139.
nodules: detection with thick-section spiral CT versus conven- 48. Theros EG. Varying manifestations of peripheral pulmonary
tional CT. Radiology. 1993;187:513–520. neoplasms: a radiologic-pathologic correlative study. AJR Am
24. Fischbach F, Knollmann F, Griesshaber V, et al. Detection of J Roentgenol. 1977;128:893.
pulmonary nodules by multislice computed tomography: 49. Gurney JW. Determining the likelihood of malignancy in soli-
improved detection rate with reduced slice thickness. Eur Radiol. tary pulmonary nodules with Bayesian analysis. Part 1. Theory.
2003;13:2378–2383. Radiology. 1993;186:405–413.
25. Zhao BS, Schwartz LH, Moskowitz CS, et al. Pulmonary metas- 50. Stitik FP, Tockman MS. Radiographic screening in the early
tases: effect of CT section thickness on measurement—initial detection of lung cancer. Radiol Clin North Am. 1978;16:347–366.
experience. Radiology. 2005;234:934–939. 51. Patel JD, Bach PB, Kris MG. Lung cancer in US women: a con-
26. Ko JP, Rusinek H, Naidich DP, et al. Wavelet compression of temporary epidemic. JAMA. 2004;291:1763–1768.
low-dose chest CT data: effect on lung nodule detection. 52. Oken MM, Marcus PM, Hu P, et al. Baseline chest radiograph for
Radiology. 2003;228:70–75. lung cancer detection in the randomized prostate, lung, colorec-
27. Rusinek H, Naidich DP, Mcguinness G, et al. Pulmonary nodule tal and ovarian cancer screening trial. J Natl Cancer Inst. 2005;
detection: low-dose versus conventional CT. Radiology. 1998;209: 97:1832–9.
243–249. 53. Gilbert HA, Kagan AR. Metastases: incidence, detection and evalua-
28. Wilimas JA, Kaste SC, Kauffman WM, et al. Use of chest com- tion without histologic confirmation. Amsterdam: North-Holland
puted tomography in the staging of pediatric Wilms’ tumor: Publishing; 1976.
interobserver variability and prognostic significance. J Clin 54. Peuchot M, Libshitz HI. Pulmonary metastatic disease: radio-
Oncol. 1997;15:2631–2635. logic-surgical correlation. Radiology. 1987;164:719–722.
29. Diederich S, Lentschig MG, Wormanns D, Heindel W. Detection 55. Henschke CI, Yankelevitz DF, Mirtcheva RM, et al. CT screening
of pulmonary nodules at spiral CT: comparison of maximum for lung cancer: frequency and significance of part-solid and
intensity projection sliding slabs and single-image reporting. nonsolid nodules. AJR Am J Roentgenol. 2002;178:1053–1057.
Eur Radiol. 2001;11:1345–1350. 56. Aoki T, Tomoda Y, Watanabe H, et al. Peripheral lung adenocar-
30. Gruden JF, Ouanounou S, Tigges S, et al. Incremental benefit of cinoma: correlation of thin-section CT findings with histologic
maximum-intensity-projection images on observer detection prognostic factors and survival. Radiology. 2001;220:803–809.
of small pulmonary nodules revealed by multidetector CT. AJR 57. Ikeda N, Maeda J, Yashima K, et al. A clinicopathological study
Am J Roentgenol. 2002;179:149–157. of resected adenocarcinoma 2 cm or less in diameter. Ann
31. Seltzer SE, Judy PF, Feldman U, et al. Influence of CT image size Thorac Surg. 2004;78:1011–1016.
and format on accuracy of lung nodule detection. Radiology. 58. Kakinuma R, Ohmatsu H, Kaneko M, et al. Progression of focal
1998;206:617–622. pure ground-glass opacity detected by low-dose helical com-
32. Tillich M, Kammerhuber F, Reittner P, et al. Detection of pul- puted tomography screening for lung cancer. J Comput Assist
monary nodules with helical CT: comparison of cine and film- Tomogr. 2004;28:17–23.
based viewing. AJR Am J Roentgenol. 1997;169:1611–1614. 59. Kishi K, Homma S, Kurosaki A, et al. Small lung tumors
33. Napel S, Rubin GD, Jeffrey RB. Technical note. STS-MIP: a new with the size of 1 cm or less in diameter: clinical, radiological,
reconstruction technique for CT of the chest. J Comput Assist and histopathological characteristics. Lung Cancer. 2004;44:
Tomogr. 1993;17:832–838. 43–51.
34. Armato SG, Li F, Giger ML, MacMahon H, et al. Lung cancer: per- 60. Kodama K, Higashiyama M, Yokouchi H, et al. Natural history
formance of automated lung nodule detection applied to cancers of pure ground-glass opacity after long-term follow-up of more
missed in a CT screening program. Radiology. 2002; 225:685–692. than 2 years. Ann Thorac Surg. 2002;73:386–392.
61. Mirtcheva RM, Vasquez M, Yankelevitz DF, Henschke CI. 84a. Grewal RG, Austin JH. CT demonstration of calcification in car-
Bronchioloalveolar carcinoma and adenocarcinoma with bron- cinoma of the lung. J Comput Assist Tomogr. 1994;18:867–871.
chioloalveolar features presenting as ground-glass opacities on 85. Zerhouni EA, Stitik FP, Siegelman SS, et al. CT of the pulmonary
CT. Clin Imaging. 2002;26:95–100. nodule: a cooperative study. Radiology. 1986;160:319–327.
62. Nakajima R, Yokose T, Kakinuma R, et al. Localized pure 86. Jones FA, Wiedemann HP, O’Donovan PB, Stoller JK. Com-
ground-glass opacity on high-resolution CT: histologic charac- puterized tomographic densitometry of the solitary pulmonary
teristics. J Comput Assist Tomogr. 2002;26:323–329. nodule using a nodule phantom. Chest. 1989;96:779–783.
63. Nakamura H, Saji H, Ogata A, et al. Lung cancer patients showing 87. Khan A, Heerman PG, Vorwerk PS, et al. Solitary pulmonary
pure ground-glass opacity on computed tomography are good nodules: comparison of classification with standard, thin-section,
candidates for wedge resection. Lung Cancer. 2004;44:61–68. and reference phantom CT. Radiology. 1991;179:477–481.
64. Nomori H, Ohtsuka T, Naruke T, Suemasu K. Differentiating 88. Im J-G, Kin WH, Han MC, et al. Sclerosing hemangiomas of the
between atypical adenomatous hyperplasia and bronchioloalve- lung and interlobar fissures: CT findings. J Comput Assist Tomogr.
olar carcinoma using the computed tomography number 1994;18:34–38.
histogram. Ann Thorac Surg. 2003;76:867–871. 89. Zweibel BR, Austin JHH. Carcinoid tumors: CT evaluation.
65. Noguchi M, Morikawa A, Kawasaki M, et al. Small adenocarci- Radiology. 1991;178:473–486.
noma of the lung: histologic characteristics and prognosis. 90. Naidich DP, McCauley DI, Siegelman SS. Computed tomogra-
Cancer. 1995;75:2844–2852. phy of bronchial adenomas. J Comput Assist Tomogr. 1982;6:
66. Travis WD, Colby TV, Corrin B, et al. Histologic typing of lung 725–732.
and pleural tumors. In: WHO World Health Organization 91. Forster BB, Müller NL, Miller RR, et al. Neuroendocrine carcino-
Pathology Panel e, ed. WHO international histological classifica- mas of the lung: clinical, radiologic, and pathologic correlation.
tion of tumors, 3rd ed. New York: Springer Verlag; 1999:46–55. Radiology. 1989;170:441–445.
67. Noguchi M, Shimosato Y. The development and progression 92. Magid D, Siegelman SS, Eggleston JC. Pulmonary carcinoid
of adenocarcinoma of the lung. Cancer Treat Res. 1995;72: tumors: CT assessment. J Comput Assist Tomogr. 1989;13:
131–142. 244–247.
67a. Leek S, Kim Y, Han J, et al. Bronchioalveolar carcinoma; clinical, 93. Naidich DP, Garay SM. Radiographic evaluation of focal lung
histopathologic, and radiologic findings. Radiographics. 1997;17: disease. Clin Chest Med. 1991;12:77–95.
1345–1347. 94. Gurney JW, Lyddon DM, McKay JA. Determining the likelihood
68. Nakata M, Saeki H, Takata I, et al. Focal ground-glass opacity de- of malignancy in solitary pulmonary nodules with Bayesian
tected by low-dose helical CT. Chest. 2002;121:1464–1467. analysis: part II. Application. Radiology. 1993;186:415–422.
69. Gaeta M, Caruso R, Barone M, et al. Ground-glass attenuation 95. Bhalla M. Use of dual kVp to assess solitary pulmonary nodules.
in nodular bronchioloalveolar carcinoma: CT patterns and J Comput Assist Tomogr. 1995;19:44–47.
prognostic value. J Comput Assist Tomogr. 1998;22:215–219. 96. Swensen SJ, Yamashita K, McCollough CH, et al. Lung nodules:
70. Higashiyama M, Kodama K, Yokouchi H, et al. Prognostic value dual-kilovolt peak analysis with CT—multicenter study. Radiology.
of bronchiolo-alveolar carcinoma component of small lung 2000;214:81–85.
adenocarcinoma. Ann Thorac Surg. 1999;68:2069–2073. 97. MacMahon H. Improvement in detection of pulmonary nod-
71. Asamura H, Suzuki K, Watanabe S, et al. A clinicopathological ules: digital image processing and computer-aided diagnosis.
study of resected subcentimeter lung cancers: A favorable prog- Radiographics. 2000;20:1169–1177.
nosis for ground glass opacity lesions. Ann Thorac Surg. 2003; 98. Siegelman SS, Khouri NF, Scott WW, et al. Pulmonary hamar-
76:1016–1022. toma: CT findings. Radiology. 1986;160:313–317.
72. Takashima S, Maruyama Y, Hasegawa M, et al. High-resolution 99. Bateson EM. So-called hamartoma of the lung—a true neoplasm
CT features: prognostic significance in peripheral lung adeno- of fibrous connective-tissue of the bronchi. Cancer. 1973;31:
carcinoma with bronchioloalveolar carcinoma components. 1458–1467.
Respiration. 2003;70:36–42. 100. Bateson EM. Relationship between intrapulmonary and endo-
73. Takashima S, Sone S, Li Fea. Small solitary pulmonary nodules bronchial cartilage-containing tumors (so-called hamartomata).
( or equal to 1 cm) detected at population-based CT screening Thorax. 1965;20:447.
for lung cancer: reliable high-resolution CT features of benign 101. Hickey PM, Simpson WM. Primary chondroma of the lung. Acta
lesions. AJR Am J Roentgenol. 2003;180:955–964. Radiol. 1926;5:475–500.
74. Takashima S, Maruyama Y, Hasegawa M, et al. CT findings and 102. McDonald JR, Harrington SW, Clagett OT. Hamartoma (often
progression of small peripheral lung neoplasms having a called chondroma) of the lung. J Thoracic Surg. 1945;14:128–143.
replacement growth pattern. Am J Roentgenol. 2003;180:817–826. 103. Benninghoven CD, Pierce CB. Primary chondroma of the lung.
75. McClure CD, Boucot KE, Shipman GA, et al. The solitary pul- AJR Am J Roentgenol. 1933;29:805–812.
monary nodule and primary lung malignancy. Arch Environ 104. Bateson EM, Abbott EK. Mixed tumors of the lung, or hamarto-
Health. 1961;3:127–139. chondromas: a review of the radiological appearances of cases
76. Good CA, Hood RT Jr, McDonald JE. Significance of a solitary published in the literature and a report of fifteen new cases. Clin
mass in the lung. AJR Am J Roentgenol. 1953;70:543–554. Radiol. 1960;11:232–247.
77. Good CA, Wilson TW. The solitary circumscribed pulmonary 105. Jensen KG, Schiodt T. Growth conditions of hamartoma of the
nodule. JAMA. 1958;166:210–215. lung. Thorax. 1958;13:233–237.
78. O’Keefe ME Jr, Good CA, McDonald JE. Calcification in solitary 106. Ribet ME, Copin M-C, Gosselin BH. Bronchogenic cysts of the
nodules in the lung. AJR Am J Roentgenol. 1957;77:1023–1033. lung. Ann Thorac Surg. 1996;61:1636–1640.
79. Siegelman SS, Zerhouni EA, Leo FP, et al. CT of the solitary 107. Naidich DP, Rumanick WM, Ettenger NA. Congenital anomalies
pulmonary nodule. AJR Am J Roentgenol. 1980;135:1–13. of the lung in adults: MR diagnosis. AJR Am J Roentgenol.
80. Zerhouni EA, Spivey JF, Morgan RH, et al. Factors influencing 1988;151:13–19.
quantitative CT measurements of solitary pulmonary nodules. 108. Weinreb JC, Naidich SP. Thoracic magnetic resonance imaging.
J Comput Assist Tomogr. 1982;6:1075–1087. Clin Chest Med. 1991;12:33–54.
81. Im J-G, Gamsu G, Gordon D. CT densitometry of pulmonary 109. Woodring HH, Fried AM, Chung VP. Solitary cavities of the
nodules in a frozen human phantom. AJR Am J Roentgenol. 1988; lung: diagnostic implications of cavity wall thickness. AJR Am
150:61–66. J Roentgenol. 1980;135:1269–1271.
82. Swensen SJ, Morin RL, Aughenbaugh GL, Leimer DW. CT re- 110. Woodring JH, Fried AM. Significance of wall thickness in soli-
construction algorithm selection in the evaluation of solitary tary cavities of the lung: a follow-up study. AJR Am J Roentgenol.
pulmonary nodules. J Comput Assist Tomogr. 1995;19:932–935. 1983;140:473–474.
83. Grewal RG, Austin JHM. CT demonstration of calcification in 111. Gaeta M, Barone M, Caruso R, et al. CT-pathologic correlation
carcinoma of the lung. J Comput Assist Tomogr. 1994;18:867–871. in nodular bronchoalveolar carcinoma. J Comput Assist Tomogr.
84. Mahoney MC, Shipley RT, Cocoran HL, Dickson BA. CT demon- 1994;18:229–232.
stration of calcification in carcinoma of the lung. AJR Am 112. Kuhlman JE, Fishman EK, Kuhajda FP, et al. Solitary bronchoalve-
J Roentgenol. 1990;154:255–258. olar carcinoma: CT criteria. Radiology. 1988;167:379–382.
113. Zwirewich CV, Vedal S, Miller RR, Müller NL. Solitary pul- 136. Kuriyama K, Tateishi R, Doi O. Prevalence of air bronchograms
monary nodule: high-resolution CT and radiologic-pathologic in small peripheral carcinomas of the lung on thin-section CT:
correlation. Radiology. 1991;179:469–476. comparison with benign tumors. AJR Am J Roentgenol. 1991;
114. Seemann MD, Staebler A, Beinert T, et al. Usefulness of mor- 156:921–924.
phological characteristics for the differentiation of benign from 137. Gaeta M, Pandolfo I, Volta S, et al. Bronchus sign on CT in
malignant solitary pulmonary lesions using HRCT. Eur Radiol. peripheral carcinoma of the lung: value in predicting results
1999;9:409–417. of transbronchial biopsy. AJR Am J Roentgenol. 1991;157:
115. Gross BH, Glazer GM, Bookstein FL. Multiple pulmonary 1181–1185.
nodules detected by computed tomography: diagnostic impli- 138. Siegelman SS, Khouri NF, Leo FP, et al. Solitary pulmonary nod-
cations. J Comput Assist Tomogr. 1985;9:880–885. ules: CT assessment. Radiology. 1986;160:307–312.
116. Hirakata K, Nakata H, Haratake J. Appearance of pulmonary 139. Kuriyama K, Tateishi R, Doi O, et al. CT-pathologic correlation
metastases on high-resolution CT scans: comparison with in small peripheral lung cancers. AJR Am J Roentgenol. 1987;
histopathologic findings from autopsy specimens. AJR Am 149:1139–1143.
J Roentgenol. 1993;161:37–43. 140. Kui M, Templeton PA, White CS, et al. Evaluation of the air bron-
117. Kido S, Kuriyama K, Kumatani Y, et al. Fractal analysis of inter- chogram sign on CT in solitary pulmonary lesions. J Comput Assist
nal texture of small peripheral pulmonary adenocarcinomas Tomogr. 1996;20:983–986.
in thin-section computed tomography: radiologic-pathologic 141. Gaeta M, Barone M, Russi EG, et al. Carcinomatous solitary pul-
correlation. Radiology. 1997;205:396. monary nodules: evaluation of the tumor-bronchi relationship
118. McNitt-Gray MF, Hart EM, Wycoff N, et al. Characterization of with thin-section CT. Radiology. 1993;187:535–539.
solitary pulmonary nodules using measures extracted from 142. Oren I, Goldstein N. Invasive pulmonary aspergillosis. Curr
high resolution CT images. Radiology. 1997;205:395. Opin Pulmon Med. 2002;8:195–200.
119. Swensen SJ, Jett JR, Hartman TE, et al. CT screening for lung 143. Kuhlman JE, Fishman EK, Burch PA, et al. Invasive pulmonary
cancer: five-year prospective experience. Radiology. 2005;235: aspergillosis in acute leukemia: the contribution of CT to early
259–265. diagnosis and aggressive management. Chest. 1987;92:95–99.
120. Benjamin MS, Drucker EA, McLoud TC, Shepard JAO. Small 144. Kuhlman JE, Fishman EK, Burch PA, et al. CT of invasive pul-
pulmonary nodules: detection at chest CT and outcome. monary aspergillosis. AJR Am J Roentgenol. 1988;150:1015–1020.
Radiology. 2003;226:489–493. 145. Barloon TJ, Galvin JR, Mori M, et al. High-resolution ultrafast
121. Piyavisetpat N, Aquino SL, Hahn PE, et al. Small incidental chest CT in the clinical management of febrile bone marrow
pulmonary nodules—how useful is short-term interval CT transplant patients with normal or nonspecific chest roentgen-
follow-up? J Thorac Imaging. 2005;20:5–9. ograms [see comments]. Chest. 1991;99:928–933.
122. Henschke CI, Yankelevitz DF, Naidich DP, et al. CT screening 146. Blum U, Windfuhr M, Buitrago-Tellez C, et al. Invasive pul-
for lung cancer: suspiciousness of nodules according to size on monary aspergillosis. MRI, CT and plain radiographic findings
baseline scans. Radiology. 2004;231:164–168. and their contribution for early diagnosis. Chest. 1994;106:
123. Mayne ST, Buenconsejo J, Janerich D. Familial cancer history 1156–1161.
and lung cancer risk in United States nonsmoking men and 147. Mori M, Galvin JR, Barloon TJ, et al. Fungal pulmonary infec-
women. Cancer Epidemiol Biomarkers Prev. 1999;8:1065–1069. tions after bone marrow transplantation: evaluation with
124. Munden RF, Pugatch RD, Liptay MJ, et al. Small pulmonary radiography and CT. Radiology. 1991;178:721–726.
lesions detected at CT: clinical importance. Radiology. 1997; 148. Primack SL, Hartman TE, Lee KS, Müller NL. Pulmonary nod-
202:105–110. ules and the CT halo sign. Radiology. 1994;190:513–515.
125. Yuan Y, Matsumoto T, Hiyama A, et al. The probability of ma- 149. Gaeta M, Volta S, Stroscio S, et al. CT “halo sign” in pulmonary
lignancy in small pulmonary nodules coexisting with poten- tuberculoma. J Comput Assist Tomogr. 1992;16:827–828.
tially operable lung cancer detected by CT. Eur Radiol. 2003;13: 150. Tan BB, Flaherty KR, Kazerooni EA, Iannettoni MD. The solitary
2447–2453. pulmonary nodule. Chest. 2003;123:89S–96S.
126. Patz Jr EF, Swensen SJ, Herndon JE 2nd. Estimate of lung cancer 151. Littleton JT, Durizch ML, Moeller G, Herbert DE. Pulmonary
mortality from low-dose spiral computed tomography screen- masses: contrast enhancement. Radiology. 1990;177:861–871.
ing trials: implications for current mass screening recommen- 152. Swensen SJ. Functional CT: lung nodule evaluation. Radio-
dations. J Clin Oncol. 2004;22:2202–2206. graphics. 2000;20:1178–1181.
127. Investigators I-E-TIELCAP. Computed tomographic screening 153. Swensen SJ, Brown LR, Colby TV, Weaver AL. Pulmonary
for lung cancer. Arch Intern Med. 2006;166:321–325. nodules: CT evaluation of enhancement with iodinated contrast
128. Libby DM, Henschke CI, Yankelevitz DF. The solitary pul- material. Radiology. 1995;194:393–398.
monary nodules: update 1995. Am J Med. 1995;99:491–496. 154. Swensen SJ, Brown LR, Colby TV, et al. Lung nodule enhance-
129. Schholten ET, Kreel L. Distribution of lung metastases in the ment at CT: prospective findings. Radiology. 1996;201:447–455.
axial plane: a radiological-pathological study. Radiol Clin 155. Zhang M, Kono M. Solitary pulmonary nodules: evaluation
(Basel). 1977;46:248–265. of blood flow patterns with dynamic CT. Radiology. 1997;205:
130. Bankoff MS, McEniff NJ, Bhadelia RA, et al. Prevalence of 471–478.
pathologically proven intrapulmonary lymph nodes and their 156. Yamashita K, Matsunobe S, Tsuda T, et al. Solitary pulmonary
appearance on CT. AJR Am J Roentgenol. 1996;167:629–630. nodule: preliminary study of evaluation with incremental
131. Meziane MA, Hruban RH, Zerhouni EA, et al. High resolution dynamic CT. Radiology. 1995;194:399–405.
CT of the lung parenchyma with pathologic correlation. Radio- 157. Yi CA, Lee KS, Kim EA, et al. Solitary pulmonary nodules:
graphics. 1988;8:27–54. dynamic enhanced multi-detector row CT study and compari-
132. Milne ENC, Zerhouni EA. Blood supply of pulmonary metas- son with vascular endothelial growth factor and microvessel
tases. J Thorac Imag. 1987;2:15–23. density. Radiology. 2004;233:191–199.
132a. Dodd JD, Souza CA, Müller NL. High resolution MDCT of pul- 158. Loftus WK, Kew J. Lung nodule enhancement at CT. Radiology.
monary septic embolism: evaluation of the feeding vessel sign. 1997;204:283.
AJR AM J Roentgenol. 2006;187:623–629. 159. Swensen SJ, Viggiano RW, Midthun DE, et al. Lung nodule
133. Balakrishnan J, Meziane MA, Siegelman SS. Pulmonary infarc- enhancement at CT: multicenter study. Radiology. 2000;214:73–80.
tion: CT appearance with pathologic correlation. J Comput 160. Jeong YJ, Lee KS, Jeong KS, et al. Solitary pulmonary nodule:
Assist Tomogr. 1989;13:941. characterization with combined wash-in and washout features at
134. Remy J, Remy-Jardin M, Wattinne L, Deffontaines C. Pulmonary dynamic multi-detector row CT. Radiology. 2005;237:675–683.
arteriovenous malformations: evaluation with CT of the chest 161. Swensen SJ, Brown LR, Colby TV, Weaver AL. CT evaluation
before and after treatment. Radiology. 1992;182:808–816. of pulmonary nodule enhancement—response. Radiology. 1996;
135. Remy J, Remy-Jardin M, Giraud F, Wattinne L. Angio- 200:586.
architecture of pulmonary arteriovenous malformations: clini- 162. Comber LA, Keith CJ, Griffiths M, Miles KA. Solitary pulmonary
cal utility of three-dimensional helical CT. Radiology. 1994; nodules: impact of quantitative contrast-enhanced CT on the
191:657–664. cost-effectiveness of FDG-PET. Clin Radiol. 2003;58:706–711.
163. Kono M, Adachi S, Kusumoto M, Sakai E. Clinical utility of 185. Cerfolio RJ, Bryant AS, Ohja B, Bartolucci AA. The maximum
Gd-DTPA-enhanced magnetic resonance imaging in lung standardized uptake values on positron emission tomography
cancer. J Thorac Imag. 1993;8:1:18–26. of a non-small cell lung cancer predict stage, recurrence, and
164. Guckel C, Schnabel K, Deimling M, Steinbrich W. Solitary survival. J Thorac Cardiovasc Surg. 2005;130:151–159.
pulmonary nodules: MR evaluation of enhancement patterns 186. Nomori H, Watanabe K, Ohtsuka T, et al. Fluorine 18-tagged
with contrast-enhanced dynamic snapshot gradient-echo imag- fluorodeoxyglucose positron emission tomographic scanning to
ing. Radiology. 1996;200:681–686. predict lymph node metastasis, invasiveness, or both, in clinical
165. Ohno M, Hatabu H, Takenaka D, et al. Solitary pulmonary nod- T1 NO MO lung adenocarcinoma. J Thorac Cardiovasc Surg.
ules: potential role of dynamic MR imaging in management— 2004;128:396–401.
initial experience. Radiology. 2002;224:503–511. 187. Hickeson M, Yun M, Matthies A, et al. Use of a corrected stan-
166. Fujimoto K, Abe T, Muller NL, et al. Small peripheral pul- dardized uptake value based on the lesion size on CT permits
monary carcinomas evaluated with dynamic MR imaging: accurate characterization of lung nodules on FDG-PET. Eur
Correlation with tumor vascularity and prognosis. Radiology. J Nucl Med. 2002;29:1639–1647.
2003;227:786–793. 188. Matthies A, Hickeson M, Cychiara A, Alavi A. Dual time point
167. Ohno Y, Hatabu H, Takenaka D, et al. Dynamic MR imaging: F-18-FDG PET for the evaluation of pulmonary nodules. J Nucl
value of differentiating subtypes of peripheral small adenocarci- Med. 2002;43:871–875.
noma of the lung. Eur J Radiol. 2004;52:144–150. 189. Kubota K, Itoh M, Ozaki K, et al. Advantage of delayed whole-
168. Kim JH, Kim H-J, Lee K-H, et al. Solitary pulmonary nodules. body FDG-PET imaging for tumor detection. Eur J Nucl Med.
A comparative study evaluated with contrast-enhanced dynamic 2001;28:696–703.
MR imaging and CT. J Comput Assist Tomogr. 2004;28:766–775. 190. Cheran SK, Nielsen ND, Patz Jr EF. False-negative findings for
169. Gambhir SS, Hoh EK, Phelps ME, et al. Decision tree sensitivity primary lung tumors on FDG positron emission tomography:
analysis for cost-effectiveness of FDG-PET in the staging and staging and prognostic implications. AJR Am J Roentgenol. 2004;
management of non-small-cell lung carcinoma. J Nucl Med. 182:1129–1132.
1996;37:1428–1436. 191. Yankelevitz DF, Reeves AP, Kostis WJ. Small pulmonary nodules:
170. Gupta NC, Frank AR, Dewan NA, et al. Solitary pulmonary volumetrically determined growth rates based on CT evaluation.
nodules: detection of malignancy with PET with 2-[F-18]- Radiology. 2000;217:251–256.
fluoro-2-deoxy-D-glucose. Radiology. 1992;184:441–444. 192. Therrasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to
171. Gupta NC, Maloof J, Gunel E. Probability of malignancy in soli- evaluate the response to treatment in solid tumors. European
tary pulmonary nodules using fluorine-18-FDG and PET. J Nucl Organization for Research and Treatment of Cancer, National
Med. 1996;37:943–948. Cancer Institute of the United States, National Cancer Institute
172. Cook GJR, Maisey MN. Review. The current status of clinical of Canada. J Natl Cancer Inst. 2000;92:205–216.
PET imaging. Clin Radiol. 1996;51:603–613. 193. Revel MP, Bissery A, Bienvenu M, et al. Are two-dimensional CT
173. Duhaylongsod FG, Lowe VJ, Patz EF, et al. Detection of primary measurements of small noncalcified pulmonary nodules reli-
and recurrent lung cancer by means of F-18 fluorodeoxyglucose able? Radiology. 2004;231:453–458.
positron emission tomography (FDG PET). J Thorac Cardiovas 194. Jennings SG, Winer-Muram HT, Tarver RD, Farber MO. Lung
Surg. 1995;110:130–140. tumor growth: assessment with CT—comparison of diameter
174. Patz EF Jr, Lowe VJ, Hoffman JM. Focal pulmonary abnormali- and cross-sectional area with volume measurements. Radiology.
ties; evaluation with F-18 fluorodeoxyglucose PET scanning. 2004;231:866–871.
Radiology. 1993;188:487–490. 195. Wormanns D, Diederich S, Lentschig MG, et al. Spiral CT of pul-
175. Hubner KF, Buonocore E, Gould HR, et al. Differentiating benign monary nodules: interobserver variation in assessment of lesion
from malignant lung lesions using “quantitative” parameters of size. Eur Radiol. 2000;10:710–713.
FDG PET images. Clin Nucl Med. 1996;21:941–949. 196. Aoki T, Nakaata H, Watanabe H, et al. Evolution of peripheral
176. Bury T, Dowlati A, Paulus P, et al. Evaluation of the solitary pul- lung adenocarcinomas: CT findings correlated with histology and
monary nodule by positron emission tomography imaging. Eur tumor doubling times. AJR Am J Roentgenol. 2000;174:763–768.
Resp J. 1996;9:410–414. 197. Yankelevitz DF, Henschke CI. Does 2-year stability imply that
177. Birim O, Kappetein AP, Stijnen T, Bogers AJJC. Meta-analysis pulmonary nodules are benign? AJR Am J Roentgenol. 1997;
of positron emission tomographic and computed tomogra- 168:325–328.
phic imaging in detecting mediastinal lymph node metastases 198. Wormanns D, Kohl G, Klotz E, et al. Volumetric measurements
in nonsmall cell lung cancer. Ann Thorac Surg. 2005;79: of pulmonary nodules at multi-row detector CT: in vivo repro-
375–382. ducibility. Eur Radiol. 2004;14:86–92.
178. Dwamena BA, Sonnad SS, Angobaldo JO, Wahl RL. Metastases 199. Ko JP, Rusinek H, Jacobs EL, et al. Small pulmonary nodules:
from non-small cell lung cancer: mediastinal staging in the Volume measurement at chest CT—phantom study. Radiology.
1990’s—meta-analytic comparison of PET and CT. Radiology. 2003;228:864–870.
1999;213:530–536. 200. Kostis WJ, Reeves AP, Yankelevitz DF, Henschke CI. Three-
179. Gould MK, Maclean CC, Kuschner WG, et al. Accuracy of dimensional segmentation and growth-rate estimation of small
positron emission tomography for diagnosis of pulmonary nod- pulmonary nodules in helical CT images. IEEE Trans Med
ules and mass lesions: a meta-analysis. JAMA. 2001;285:914–924. Imaging. 2003;22:1259–1274.
180. Marom EM, Sarvis S, Herndon II JE, Patz EF Jr. T1 lung cancers: 201. Yankelevitz DF, Reeves AP, Kostis WJ. Small pulmonary nodules:
sensitivity of diagnosis with fluorodeoxyglucose PET. Radiology. volumetrically determined growth rates based on CT evaluation.
2002;223:453–549. Radiology. 2000;217:251–256.
181. Herder GJ, Golding RP, Hoekstra OS, et al. The performance of 202. Tateishi U, Uno H, Yonemori K, et al. Prediction of lung adeno-
F-18-fluorodeoxyglucose positron emission tomography in carcinoma without vessel invasion—a CT scan volumetric
small solitary pulmonary nodules. Eur J Nucl Med Mol Imaging. analysis. Chest. 2005;128:3276–3283.
2004;31:1231–126. 203. Winer-Muram HT, Jennings SG, Meyer CA, et al. Effect of varying
182. Imdahl A, Jenkner S, Brink I, et al. Validation of FDG positron CT section width on volumetric measurement of lung tumors
emission tomography for differentiation of unknown pulmonary and application of compensatory equations. Radiology. 2003;
lesions. Eur J CardioThoracic Surg. 2001;20:324–329. 229:184–194.
183. Port JL, Andrade RS, Levin MA, et al. Positron emission tomo- 204. Winer-Muram HT, Jennings SG, Tarver RD, et al. Volumetric
graphic scanning in the diagnosis and staging of non-small cell growth rate of stage I lung cancer prior to treatment: serial CT
lung cancer 2 cm in size or less. J Thorac Cardiovasc Surg. 2005; scanning. Radiology. 2002;223:798–805.
130:1611–1615. 205. Hasegawa M, Sone S, Takashima S, et al. Growth rate of small
184. Nomori H, Watanabe K, Ohtsuka T, et al. Visual and semi- lung cancers detected on mass CT screening. Br J Radiol.
quantitative analyses for F-18 fluorodeoxyglucose PET scanning 2000;73: 1252–1259.
in pulmonary nodules 1 cm to 3 cm in size. Ann Thorac Surg. 206. Bennett DE, Sasser WF, Ferguson TB. Adenocarcinoma of the
2005;79:984–989. lung in men. Cancer. 1969;23:431–439.
207. El-Torky M, Giltman LI, Dabbous M. Collagens in scar carci- 235. Daniel TM. A proposed diagnostic approach to the patients with
noma of the lung. Am J Pathol. 1985;121:322–326. the subcentimeter pulmonary nodules: techniques that facilitate
208. Barsky SH, Huang SJ, Bhuta S. The extracellular matrix of video-assisted thoracic surgery excision. Semin Thorac Cardiovasc
pulmonary scar carcinoma is suggestive of a desmoplastic Surg. 2005;17:115–122.
origin. Am J Pathol. 1986;124:412–419. 236. Schreiber G, McCrory DC. Performance characteristics of differ-
209. Kawai T, Yakumaaru K, Suzuki M, et al. Diffuse interstitial pul- ent modalities for diagnosis of suspected lung cancer—summary
monary fibrosis and lung cancer. Acta Pathol Jpn. 1987;37:11–19. of published evidence. Chest. 2003;123:115S–128S.
210. Hughes JM, Weil H. Asbestosis as a precursor of asbestos related 237. Gasparini S, Ferretti M, Secchi EB, et al. Integration of trans-
lung cancer: results of a prospective mortality study. Br J Ind bronchial and percutaneous approach in the diagnosis of
Med. 1991;48:229–233. peripheral pulmonary nodules or masses: experience with 1,027
211. Mellemkjaer L, Linet MS, Gridley G, et al. Rheumatoid arthritis consecutive cases. Chest. 1995;108:131–137.
and cancer risk. Eur J Cancer. 1996;32:1752–1757. 238. Baaklini WA, Reinoso MA, Gorin AB, et al. Diagnostic yield of
212. Roumm AD, Medsger TA Jr. Cancer and systemic sclerosis: an fiberoptic bronchoscopy in evaluating solitary pulmonary nod-
epidemiologic study. Arthritis Rheum. 1985;28:1336–1340. ules. Chest. 2000;117:1049–1054.
213. Kaiser LR, Shrager JB. Video-assisted thoracic surgery: the current 239. Reichenberger F, Weber J, Tamm M, et al. The value of trans-
state of the art. AJR Am J Roentgenol. 1995;165:1111–1117. bronchial needle aspiration in the diagnosis of peripheral
214. Swensen SJ, Jett JR, Payne WS, et al. An integrated approach to pulmonary lesions. Chest. 1999;116:704–708.
evaluation of the solitary pulmonary nodule. Mayo Clin Proc. 240. Naidich DP, Sussman R, Kutcher WL, et al. Solitary pulmonary
1990;65:173–86. nodules: CT-bronchoscopic correlation. Chest. 1988;93:595–598.
215. Ray JF, Lawton BR, Magnin GE, et al. The coin lesion story: update 241. Bandoh S, Fujita J, Tojo Y, et al. Diagnostic accuracy and safety
1976. Twenty years experience with early thoracotomy for 179 of flexible bronchoscopy with multiplanar reconstruction
suspected malignant coin lesions. Chest. 1976;70:332–336. images and ultrafast Papanicolaou stain—evaluating solitary
216. Keagy BA, Starek PJK, Murray GF. Major pulmonary resection for pulmonary nodules. Chest. 2003;124:1985–1992.
suspected but unconfirmed malignancy. Ann Thorac Surg. 242. Lai RS, Lee SSJ, Ting YM, et al. Diagnostic value of trans-
1984;38:314–316. bronchial lung biopsy under fluoroscopic guidance in solitary
217. Romano PS, Mark DH. Patient and hospital characteristics pulmonary nodule in an endemic area of tuberculosis. Respir
related to in-hospital mortality after lung cancer resection. Med. 1996;90:139–143.
Chest. 1992;101:1332–1337. 243. Trisolini R, Agli LL, Cancellieri A, et al. The value of flexible
218. Bach PB, Cramer LD, Schrag D, et al. The influence of hospital transbronchial needle aspiration in the diagnosis of stage I
volume on survival after resection for lung cancer. N Engl J Med. sarcoidosis. Chest. 2003;124:2126–2130.
2001;345:181–188. 244. Sharafkhaneh A, Baaklini W, Gorin AB, Green L. Yield of trans-
219. Scott WW, Kuhlman JE. Focal pulmonary lesions in patients bronchial needle aspiration in diagnosis of mediastinal lesions.
with AIDS: percutaneous transthoracic needle biopsy. Radiology. Chest. 2003;124:2131–2135.
1991;180:419–421. 245. Hermens FHW, VanEngelenburg TCA, Visser FJ, et al. Diagnostic
220. End A, Helbich T, Wisser W, Dekan G, Klepetko W. The pul- yield of transbronchial histology needle aspiration in patients
monary nodule after lung transplantation: cause and outcome. with mediastinal lymph node enlargement. Respiration. 2003;70:
Chest. 1995;107:1317–1322. 631–635.
221. Gurney JW, Swensen SJ. Solitary pulmonary nodules: determin- 246. Khoo KL, Chua GSW, Mukhopadhyay A, Lim TK. Transbron-
ing the likelihood of malignancy with neural network analysis. chial needle aspiration: initial experience in routine diagnostic
Radiology. 1995;196:823–829. bronchoscopy. Respir Med. 2003;97:1200–1204.
222. Henschke CI, Yankelevitz DF, Mateescu I, et al. Neural networks 247. Cetinkaya E, Yildiz P, Altin S, Yilmaz V. Diagnostic value of trans-
for the analysis of small pulmonary nodules. Clin Imaging. 1997; bronchial needle aspiration by Wang 22-gauge cytology needle
21:390–399. in intrathoracic lymphadenopathy. Chest. 2004;125: 527–531.
223. Nakamura K, Yaoshida H, Engelmann R, et al. Computerized 248. Asano F, Matsuno Y, Komaki C, et al. CT-guided transbronchial
analysis of the likelihood of malignancy in solitary pulmonary diagnosis using ultrathin bronchoscope for small peripheral
nodules with use of artificial neural networks. Radiology. 2000; pulmonary lesions [English abstract]. Nihon Kokyuki Gakkai
214:823–830. Zasshi. 2002;40:11–16.
224. Dewan NA, Shehan CJ, Reeb SD, et al. Likelihood of malig- 249. Rooney CP, Wolf K, McLennan G. Ultrathin bronchoscopy as an
nancy in a solitary pulmonary nodule: comparison of Bayesian adjunct to standard bronchoscopy in the diagnosis of peripheral
analysis and results of FDG-PET scan. Chest. 1997;112:416–422. lung lesions: a preliminary report. Respiration. 2002;69:63–68.
225. Ost D, Fein AM, Feinsilver SH. The solitary pulmonary nodule. 250. Moriya H, Hashimoto N, Honjo H, et al. Visibility of peripheral
N Engl J Med. 2003;348:2535–2542. bronchi in virtual bronchoscopy with 0.5 mm collimation mul-
226. Libby DM, Smith JP, Altorki NK, et al. Managing the small pul- tidetector CT [in Japanese]. Jpn J Clin Radiol. 2001;46:1–7.
monary nodule discovered by CT. Chest. 2004;125:1522–1529. 251. Higgins WE, Ferguson JS, Thomas K, et al. Progress towards
227. Mayo JR, Aldrich J, Müller NL. Radiation exposure at chest CT: a virtual-bronchoscopic guidance of peripheral nodule biopsy.
statement of the Fleischner Society. Radiology. 2003;228:15–21. Am J Respir Crit Care Med. 2003;167:A535.
228. Leader JK, Warfel TE, Fuhrman CR, et al. Pulmonary nodule 252. Shinagawa N, Yamazaki K, Onodera Y, et al. CT-guided trans-
detection with low-dose CT of the lung: agreement among radi- bronchial biopsy using an ultrathin bronchoscope with virtual
ologists. AJR Am J Roentgenol. 2005;185:973–978. bronchoscopic navigation. Chest. 2004;125:1138–1143.
229. Gould MK, Sanders GD, Barnett PG, et al. Cost-effectiveness of 253. Suda Y, Tanaka A, Shindoh Y, Iijima H. Practical methodology
alternative management strategies for patients with solitary for bronchial route analysis using virtual bronchoscopy. Am
pulmonary nodules. Ann Intern Med. 2003;138:724–735. J Respir Crit Care Med. 2003;167:A527.
230. MacMahon H, Austin JHM, Gamsu G, et al. Guidelines for manage- 254. Schwarz Y, Mehta AC, Ernst A, et al. Electromagnetic navigation
ment of small pulmonary nodules detected on CT scans: a state- during flexible bronchoscopy. Respiration. 2003;70:516–522.
ment from the Fleischner Society. Radiology. 2005; 237:395–400. 255. Solomon SB, White P, Wiener CM, et al. Three-dimensional CT-
231. www.acrin.org/current_protocols.html guided bronchoscopy with a real-time electromagnetic position
232. Raab SS, Hornberger J. The effect of a patient’s risk taking atti- sensor: a comparison of two image registration methods. Chest.
tude on the cost effectiveness of testing strategies in the evalua- 2000;118:1783–1787.
tion of pulmonary lesions. Chest. 1997;111:1583–1590. 256. Goldberg BB, Steiner RM, Liu JB. US-assisted bronchoscopy with
233. Cummings SR, Lillington GA, Richard RJ. Estimating the proba- use of miniature transducer-containing catheters. Radiology.
bility of malignancy in solitary pulmonary nodules: a Bayesian 1994;190: 233–237.
approach. Am Rev Respir Dis. 1986;134:449–452. 257. Kurimoto N, Murayama M, Yoshioka S, et al. Assessment of use-
234. Goldberg-Kahn B, Healy JC, Bishop JW. The cost of diagnosis: a fulness of endobronchial ultrasonography in determination
comparison of four different strategies in the workup of solitary of depth of tracheobronchial tumor invasion. Chest. 1999;
radiographic lung lesions. Chest. 1997;111:870–876. 115:1500–1506.
258. Klein JS, Salomon G, Stewart EA. Transthoracic needle biopsy 281. Summers RM. Roadmaps for advancement of radiologic computer-
with a coaxially placed 20-gauge automated cutting needle: aided diagnosis in the 21st century. Radiology. 2003;229: 11–13.
results in 122 patients. Radiology. 1996;198:715–720. 282. Armato SG, Giger ML, MacMahon H. Automated detection of
259. Khouri NF, Stitik FP, Erozan YS, et al. Transthoracic needle aspi- lung nodules in CT scans: preliminary results. Med Phys. 2001;28:
ration biopsy of benign and malignant lung lesions. AJR Am J 1552–1561.
Roentgenol. 1985;144:281–288. 283. Ko JP, Naidich DP. Lung nodule detection and characterization
260. Stanley JH, Fish GD, Andiole JG, et al. Lung lesions: cytologic with multislice CT. Radiol Clin North Am. 2003;41(3):575–597.
diagnosis by fine-needle biopsy. Radiology. 1987;162: 389–391. 284. Brown MS, Goldin JG, Rogers S, et al. Computer-aided lung
261. Fraser RS. Transthoracic needle aspiration. The benign diagno- nodule detection in CT: Results of large-scale observer test. Acad
sis. Arch Pathol Lab Med. 1991;115:751–761. Radiol. 2005;12:681–686.
262. Calhoun P, Feldman PS, Armstrong P, et al. The clinical out- 285. Brown MS, Goldin JG, Suh RD, et al. Lung micronodules:
come of needle aspirations of the lung when cancer is not diag- automated method for detection at thin-section CT—initial
nosed. Ann Thorac Surg. 1986;41:592–596. experience. Radiology. 2003;226:256–262.
263. Westcott JL, Rao N, Colley DP. Transthoracic needle biopsy of 286. Gurcan MN, Sahiner B, Petrick N, et al. Lung nodule detection
small pulmonary nodules. Radiology. 1997;202:97–103. on thoracic computed tomography images: preliminary evalua-
264. Westcott JL. Diagnostic approach to solitary pulmonary nodule. tion of a computer-aided diagnosis system. Med Phys. 2002;29:
J Bronchol. 1996;3:316–323. 2552–2558.
265. Li HQ, Boiselle PM, Shepard JAO, et al. Diagnostic accuracy and 287. Zhao B, Gamsu G, Ginsberg MS, et al. Automatic detection of
safety of CT-guided percutaneous needle aspiration biopsy of small lung nodules on CT utilizing a local density maximum
the lung: comparison of small and large pulmonary nodules. algorithm. J Appl Clin Med Phys. 2003;4:248–260.
AJR Am J Roentgenol. 1996;167: 105–109. 288. Armato SG 3rd, Altman MB, La Riviere PJ. Automated detection
266. Yankelevitz DF, Henschke CI, Koizumi JH, et al. CT-guided of lung nodules in CT scans: effect of image reconstruction algo-
transthoracic needle biopsy of small solitary pulmonary nod- rithm. Med Phys. 2003;30:461–472.
ules. Clin Imaging. 1997;21:107–110. 289. Qian J, Fan L, Wei G, et al. Knowledge-based automatic
267. Moulton JS, Moore PT. Coaxial percutaneous biopsy tech- detection of multi-type lung nodules from multidetector
nique with automated biopsy devices: value in improving CT studies. In: Sonka M, Fitzpatrick J, eds. Medical Imaging 2002:
accuracy and negative predictive value. Radiology. 1993;186: Image Processing; 2002: Proceedings of SPIE. 2002: 689–697.
515–522. 290. Armato SG, Giger ML, Moran CJ, et al. Computerized detection
268. Kazerooni EA, Mikhail A, Martinez FJ. Risk of pneumothorax in of pulmonary nodules on CT scans. Radiographics. 1999;19:
CT-guided transthoracic needle aspiration biopsy of the lung. 1303–1311.
Radiology. 1996;198:371–375. 291. Brown MS, McNitt-Gray MF, Goldin JG, et al. Patient-specific
269. Bernard A, Azorin J, Bellenot F, et al. Resection of pulmonary models for lung nodule detection and surveillance in CT
nodules using video-assisted thoracic surgery. Ann Thorac Surg. images. IEEE Trans Med Imaging. 2001;20:1242–1250.
1996;61:202–204. 292. Wormanns D, Fiebich M, Saidi M, et al. Automatic detection of
270. Hazelrigg SR, Nunchuck SK, LoCicero J III, et al. Video Assisted pulmonary nodules at spiral CT: clinical application of a com-
Thoracic Surgery Study Group Data. Ann Thorac Surg. 1993; puter-aided diagnosis system. Eur Radiol. 2002;12: 1052–1057.
56:1039–1044. 293. Novak CL, Shen H, Odry B, et al. Performance of an automatic
271. Kirby TJ, Mack MJ, Landreneau RJ, Rice TW. Initial experience system for nodule correspondence in follow-up CT studies of
with video-assisted thoracoscopic lobectomy. Ann Thorac Surg. the lung. Radiology. 2002;225:476.
1993;56:1248–1253. 294. Marten K, Engelke C, Seyfarth T, et al. Computer-aided detec-
272. Mack MJ, Haellrigg SR, Landreneau RJ, Acuff TE. Thoracoscopy tion of pulmonary nodules: influence of nodule characteristics
for the diagnosis of the indeterminate solitary pulmonary nod- on detection performance. Clin Radiol. 2005;60:196–206.
ule. Ann Thorac Surg. 1993;56:825–832. 295. Giger ML, Bae KT, Macmahon H. Computerized detection of
273. Schwarz CD, Lengliner F, Eckmayr J, et al. VATS (video-assisted pulmonary nodules in computed tomography images. Invest
thoracic surgery) of undefined pulmonary nodules. Preop- Radiol. 1994;29:459–465.
erative evaluation of videoendoscopic resectability. Chest. 296. Kanazawa S, Kawata Y, Niki N, et al. Computer-aided diagnosis
1994;106:1570–1574. for pulmonary nodules based on helical CT images. Comput
274. Shah RM, Spirn PW, Salazar AM, et al. Localization of periph- Med Imaging Graphics. 1998;22:157–167.
eral pulmonary nodules for thoracoscopic excision: value of 297. Marten K, Grillhosl A, Seyfarth T, et al. Computer-assisted detec-
CT–guided wire placement. AJR Am J Roentgenol. 1993;161: tion of pulmonary nodules: evaluation of diagnostic perform-
279–283. ance using an expert knowledge-based detection system with
275. Shepard J-AO, Mathisen DJ, Muse VV, et al. Needle localization variable reconstruction slice thickness settings. Eur Radiol. 2005;
of peripheral lung nodules for video-assisted thoracoscopic sur- 15:203–212.
gery. Chest. 1994;105:1559–1563. 298. Peldschus K, Herzog P, Wood SA, et al. Computer-aided diagno-
276. Shennib H. Intraoperative loculation techniques for pulmonary sis as a second reader—spectrum of findings in CT studies of the
nodules. Ann Thorac Surg. 1993;56:745–748. chest interpreted as normal. Chest. 2005;128:1517–1523.
277. Asamura H, Kondo H, Narike T, et al. Computed tomography- 299. Furuya K, Murayama S, Soeda H, et al. New classification of
guided coil injection and thoracoscopic pulmonary resection small pulmonary nodules by margin characteristics on high
under roentgenographic fluoroscopy. Ann Thorac Surg. 1994;58: resolution CT. Acta Radiol. 1999;40:496–504.
1542–1544. 300. McNitt-Gray MF, Hart EM, Wyckoff N, et al. A pattern classifica-
278. Finley RJ, Clifton J, Mayo J. Image-guided video-assisted thora- tion approach to characterizing solitary pulmonary nodules
coscopic resection of small peripheral lung nodules. Adv Surg. imaged on high-resolution CT: preliminary results. Med Phys.
2005;39:263–284. 1999;26: 880–888.
279. Ko J, Naidich DP. Computer-aided diagnosis in the evaluation 301. Kawata Y, Niki N, Ohmatsu H, Moriyama N. Example-based
of lung disease. J Thorac Imaging. 2004;19:136–155. assisting approach for pulmonary nodule classification in three-
280. Li Q, Suzuki K, Shiraishi J, et al. Computer-aided diagnosis in dimensional thoracic computed tomography images. Acad
thoracic CT. Semin Ultrasound CT MR. 2005;26:357–363. Radiol. 2003;10:1402–1415.
Lung Cancer 7
INTRODUCTION/EPIDEMIOLOGY 621 2004 that 173,700 Americans will be diagnosed with lung
cancer, and 164,440 will have died of the disease (2). Lung
PATHOLOGY 622 cancer currently accounts for 33% of deaths from cancer in
Preinvasive Lesions 622 men, compared with carcinoma of the prostate, the second
Adenocarcinoma 623 most common cause, which accounts for 14%. Similarly,
Squamous Cell Carcinoma 623 lung cancer accounts for 24% of cancer deaths in women
Large Cell Carcinoma 623 in the United States compared with 18% caused by breast
Tumors with Neuroendocrine Morphology 624 carcinoma, now the second most common cause of death
Miscellaneous Lung Tumors 626 from cancer in women (3). The 5-year survival of lung can-
cer remains only 10% to 15%, despite intensive efforts
NON–SMALL CELL LUNG CANCER 626 over the past several decades, because most bronchogenic
TNM Classification 626 carcinomas are still unresectable when first diagnosed.
Stage Groupings 631 Complete surgical excision of carcinoma remains the most
effective form of therapy and offers the patient the only
MULTIDISCIPLINARY IMAGING IN THE reasonable hope for cure.
EVALUATION AND STAGING OF LUNG Over the past several years, there has been a growing
CANCER 637 awareness of important changes in the epidemiology of
Computed Tomography Evaluation 637 lung cancer (4). For example, it is now established that
Magnetic Resonance Imaging Evaluation 648 adenocarcinomas have supplanted squamous cell carci-
Positron Emission Tomography (FDG-PET) noma (SCC) as the most frequent form of the disease,
Evaluation 652 across both genders and all races. Lung cancer remains
predominantly a disease caused by tobacco exposure.
SMALL CELL LUNG CANCER 658 Despite efforts to eliminate smoking, 50 million Americans
Therapeutic Considerations 658 still smoke; as important, even among those who do quit,
Imaging Evaluation 658 the risk of developing lung cancer remains significantly
higher than in never smokers for as long as 30 years (5).
SURGICAL EVALUATION 659 Unfortunately, although the rate of smoking in younger
Superior Sulcus (Pancoast) Tumors 659 individuals has declined, it remains problematic. In addi-
tion to tobacco exposure, numerous other causes have been
RADIOFREQUENCY ABLATION identified and include exposure to asbestos, radon, and
SCREENING 660 a host of chemical agents, including nickel, polycyclic
aromatic hydrocarbons, and chromates, among others. In
the developing world, concern has also focused on the role
of exposure to cooking oils (6).
INTRODUCTION/EPIDEMIOLOGY There is also increasing appreciation of gender and race
differences in lung cancer (6). Especially concerning is the
At the outset of the 21st century, lung cancer remains one epidemic of lung cancer appearing in women, including
of the world’s leading causes of preventable death (1). In nonsmokers particularly in Asian populations (7). As
the United States alone, it was estimated as recently as reported by Toh et al. (8) in a retrospective study from
Singapore, whereas women comprised 33.8% of all lung Preinvasive Lesions

cancer patients, they accounted for 73.9% of cancer occur-
ring in nonsmokers. This raises the important question of In addition to the four basic subtypes of lung cancer, a
whether a difference exists in the types and prognoses of number of preinvasive lesions have also been identified,
cancer in smokers versus nonsmokers. Women have been including atypical adenomatous hyperplasia (AAH) and
reported to have longer survival times, especially after sur- diffuse idiopathic neuroendocrine cell hyperplasia (DIP-
gical resection of early-stage adenocarcinomas (9). In fact, NECH). As described extensively in Chapter 6, AAH is
the risk of developing lung cancer is multifactorial, the defined as a localized proliferation of mildly to moder-
result of synergistic interactions between environmental ately atypical cells lining the alveoli in the absence of
causes and genetics (4). Although it is outside the scope of underlying inflammation or fibrosis (14). Based on stud-
this chapter, awareness of the role of genetic susceptibility ies showing a variety of molecular alterations, including
to developing cancer represents one of the most important the frequent finding of k-ras and p53 mutations, AAH is
developments over the past decade in our understanding now considered by most to represent a premalignant
of the epidemiology of this multifaceted disease (10). lesion, part of the spectrum of adenocarcinoma of the
Understanding these factors is already essential to the lung. These lesions are frequently identified incidentally in
development of molecular imaging techniques (11–13). lobes resected for primary adenocarcinomas and typically
appear as poorly defined ground-glass opacities smaller
than 5 mm (Fig. 7-1 and Figs. 6-4 through 6-6).
PATHOLOGY Whereas AAH proves to be a common finding, espe-
cially on multidetector CT screening studies, DIP-NECH is
As most recently specified by the 2004 World Health an exceptionally rare abnormality that is defined by a
Organization (WHO), primary lung tumors are classified purely intraepithelial proliferation of neuroendocrine cells,
by their light-microscopic appearance into four broad
categories, including adenocarcinoma, SCC, large cell
carcinoma, and small cell carcinoma (Table 7-1) (14).
Multiple subdivisions of these types have been defined,
some of which are based on molecular biologic characteris-
tics. Although these are not currently considered of primary
importance for classification, in selected cases, knowledge
of immunohistologic subtyping is important for differen-
tial diagnosis and may ultimately prove of prognostic and
therapeutic value (15). From an imaging standpoint, some
distinctions between the various subtypes of non–small
cell lung cancer are worth identifying, in particular, be-
tween adenocarcinoma and squamous cell carcinoma.
TABLE 7-1
HISTOLOGIC AND CYTOLOGIC DIFFERENCES
BETWEEN MAJOR HISTOLOGIC CLASSES
OF LUNG CANCER
Proportions by Histologic Type
Histologic Type/(Subtype) Total
Squamous carcinoma 29%

Small cell lung carcinoma 20%
Large cell lung carcinoma 9%
Large cell neuroendocrine 2%
Adenocarcinoma 32% Figure 7-1 Mixed adenocarcinoma with bronchoalveolar cell
(bronchioloalveolar) 3% features. Magnified CT section through the right upper lobe
Others 12% shows a part solid, part ground-glass nodule. Although this lesion
appears to be predominantly ground-glass, the finding of a subtle
From Pass HL, Carbone DP, Johnson DH, et al., eds. Lung cancer: prin- solid component makes this more likely to represent a mixed
ciples and practice, 3rd ed. Philadelphia: Lippincott Williams & Wilkins, adenocarcinoma. Accurate histologic diagnosis of these lesions
2005, with permission. requires that the entire nodule be resected.
Chapter 7: Lung Cancer 623
identifiable on CT either as scattered tiny micronodules or, Although adenocarcinomas with BAC components are
more characteristically, as a cause of mosaic lung attenua- properly classified as adenocarcinomas, mixed subtype, it
tion due to small-airway involvement resulting in diffuse has been shown that those with prominent BAC features
air trapping (16–18). This entity is frequently associated may be further subdivided by additional histologic
with multiple pulmonary tumorlets because of a nodular features, including the size of scars (5 mm vs. 5 to
proliferation of airway neuroendocrine cells that extend 15 mm), percentage of lepidic growth, and the presence of
beyond the epithelium into the adjacent lung. Although vascular invasion, among others (19). These findings are
usually identified alone, evidence of bronchiectasis may be of importance, as a close correlation appears to exist be-
found as well as an association with carcinoid tumors. tween these features and those identified on CT scans—in
particular, the extent of ground-glass attenuation—
with important implications for prognosis (Figs.6-1 and
Adenocarcinoma 6-7 to 6-10); specifically, the greater the degree of ground-
Currently, adenocarcinoma is the most frequent type of glass attenuation, the better outcome, with 5-year survival
lung cancer identified in both men and women, represent- of patients with surgically resected pure BAC defined
ing approximately 30% to 35% of all primary lung histologically and pure ground-glass lesions identified by
neoplasms. These tumors most often are initially seen as pe- CT as 100% (see Chapter 6). Despite these correlations, at
ripheral lung nodules ranging in density from pure ground- present no consensus exists regarding a definition of “min-
glass to solid lesions, less frequently as central imally invasive” adenocarcinoma (19). Although most
endobronchial lesions. As noted in the 2004 WHO classifi- commonly first seen either as solitary or multiple nodules,
cation, adenocarcinomas are most often heteroge- adenocarcinoma, especially BACs, less commonly appears
neous, containing more than one subtype in a majority of as diffuse parenchymal consolidation, frequently lobar in
cases, leading to the newer designation of “adenocarci- distribution (Fig. 7-2) (21). The differential in these cases
noma, mixed type” as the most common form of adenocar- should include, in particular, primary parenchymal lym-
cinoma of the lung (Table 7-2) (19). From an imaging phoma. In selected cases in which differentiation between
standpoint, the most important distinction among the primary and metastatic adenocarcinoma proves problem-
various subtypes of adenocarcinoma is bronchoalveolar cell atic, diagnosis may be facilitated by immunohistochemical
carcinoma (BAC). Incorporating initial seminal observa- evaluation. Specifically, it has been shown that thyroid
tions by Noguchi et al. (20), we define BAC strictly transcription factor (TTF-1) is a highly specific marker for
as a tumor showing only pure lepidic growth without evi- primary lung adenocarcinomas distinct from metastatic
dence of interstitial or stromal invasion. Although BAC is adenocarcinoma, positive in approximately 85% of cases
further subdivided into nonmucinous, mucinous, and (14) (Table 7-3). Histochemical staining may also be of
mixed types, these distinctions have no apparent radiologic value in differentiating between diffuse adenocarcinoma
counterparts. Most important, from a diagnostic stand- involving the pleura and mesothelioma.
point, definitive diagnosis requires that the entire lesion be
evaluated histologically; the diagnosis should not be made Squamous Cell Carcinoma
solely on the basis of transbronchial and/or transthoracic
needle biopsy, as it is possible that focal areas of invasive In distinction to adenocarcinoma, SCC more often is
adenocarcinoma may otherwise not be identified. initially seen as central lesions resulting in airway obstruc-
tion and peripheral atelectasis. Histologically, SCC may be
differentiated from adenocarcinoma in particular by the
presence of keratinization and intercellular bridges. Only
TABLE 7-2 approximately 25% of SCCs appear as peripheral lesions,
ADENOCARCINOMA: 2004 WORLD HEALTH frequently associated with central cavitation (Fig. 7-3). As a
ORGANIZATION CLASSIFICATION result of involvement of the central airways, patients with
■ Adenocarcinoma—mixed subtype
SCC often are seen earlier than are those with most
■ Acinar adenocarcinoma cases of adenocarcinoma with cough, fever, or hemoptysis.
■ Papillary adenocarcinoma Unfortunately, as documented by numerous early lung can-
■ Bronchioloalveolar carcinoma— cer CT screening trials, unlike for peripheral lung nodules
nonmucinous/mucinous/mixed or indeterminate that may be identified when they are as small as 3 to 4 mm,
■ Solid adenocarcinoma—mucin production CT has proved of little value for early identification of focal
■ Fetal adenocarcinoma airway lesions.
■ Mucinous (colloid) adenocarcinoma
■ Mucinous cystadenocarcinoma
■ Signet-ring adenocarcinoma Large Cell Carcinoma
■ Clear-cell adenocarcinoma
Large cell lung cancers represent approximately 10% of all
From Travis WD, et al. Evolving concepts in the pathology and computed
tomography imaging of lung adenocarcinoma and bronchioloalveolar primary lung neoplasms and are characterized by lacking
carcinoma. J Clin Oncol. 2005;23:3279–3287, with permission. features of either adenocarcinomas or SCCs. They
C B
Figure 7-2 Bronchoalveolar cell carcinoma (BAC): lobar distribution. See Color Figure 7-2B. A: Left mid lung shows evidence of dense
consolidation in the periphery of the left upper lobe, marginated posteriorly by the left major fissure, associated with a diffuse nodular infil-
trate throughout the remainder of the left upper lobe. Many of these nodules appear to be centrilobular in distribution, some with a tree-in-
bud configuration. Note that despite extensive disease, no evidence exists of architectural distortion; the anterior segmental bronchus is
patent. Although nonspecific, when chronic, these findings are consistent with diffuse tumor manifesting at least in part a pattern of lepidic
growth. B: Fluorodeoxyglucose–positron emission tomography (FDG-PET) correlation. Coronal maximum intensity projection image shows
intense activity in the lateral aspect of the left upper lobe, corresponding to changes identified in A. No other abnormalities are identified.
Although FDG-PET is often negative in cases with solid nodules due to BAC, in this case, PET is of value by further confirming that the
disease is restricted to a single lobe and therefore is potentially resectable. C: High-resolution CT section, in a different patient than shown
in A and B, demonstrates the appearance of diffuse consolidative BAC. Innumerable ill-defined ground-glass nodules can be identified in
portions of the lung less extensively involved (arrows).
typically are first seen as peripheral lung tumors similar to carcinoid tumors are considered low-grade tumors with few
adenocarcinomas, although statistically are more likely to mitoses per high-power field and are currently not consid-
be larger than 3 cm at presentation. As is discussed later, ered as a part of a continuum with either LCNEC or SCLC,
they may also show neuroendocrine features, in which despite sharing neuroendocrine features. They characteristi-
case they are designated large cell neuroendocrine cancer. cally appear as well-defined lesions, either peripheral or
more commonly central in location, the latter appearing
as well-defined endobronchial lesions, approximately one
Tumors with Neuroendocrine Morphology
third of which have evidence of calcification and/or ossi-
Included in this category are lesions that share common fication (see Figs. 6-15 and 6-16). Carcinoids are also
morphologic and immunohistochemical characteristics and extremely vascular, a feature that is easily identified when
include both typical and atypical carcinoid tumors as well scans are performed after a bolus of intravenous contrast
as large cell neuroendocrine cancer (LCNEC) and small cell media (23).
lung cancer (SCLC) (22). Together these lesions account Both SCC and LCNEC are high-grade tumors. SCC
for 25% of primary lung neoplasms. Typical and atypical represents approximately 20% of lung cancers and
A B
C
Figure 7-3 Squamous cell carcinoma (SCC). A: Non–contrast-enhanced CT image through the left upper lobe shows a thick-walled irreg-
ular cavity within which a discrete air-fluid level is identified. Note that this lesion marginates the mediastinum anteriorly and the pleura and
chest wall anterolaterally, without definite evidence of either mediastinal or chest-wall invasion. Note the absence of obvious mediastinal
adenopathy or effusion. Although this appearance is consistent with a cavitary SCC, especially in the absence of signs of infection or vas-
culitis, neither mediastinal nor pleural/chest wall invasion can be excluded, pointing out the limitations of CT for differentiating between T3
and T4 lesions. The presence of an air-fluid level is also nonspecific, consistent either with superimposed infection or hemorrhage. B:
Fluorodeoxyglucose–positron emission tomography (FDG-PET) correlation. Corresponding axial image demonstrates similar findings as
shown in A. Although PET images are of limited use for establishing mediastinal or pleural/chest-wall invasion, in this case, PET is of value by
excluding significant mediastinal or hilar adenopathy, or remote pleural malignancy. C: Contrast-enhanced CT study in a different patient
than shown in A and B also shows a thick-walled cavitary mass occupying most of the left lower lobe. In this case, direct communication is
found between the tumor and the left lower lobe bronchus. Histologically verified SCC.
TABLE 7-3
FREQUENCIES OF POSITIVE IMMUNOHISTOCHEMICAL RESULTS FOR LUNG CARCINOMAS
OF VARIOUS HISTOLOGIC TYPES BY MARKER
Chromo- CD-117
Ki-67 EGFR CK-7 CK20 TTF-1 dranin A (c-KIT) CEA Calretinin
SqCa

BAC
AdCa
LCLC
LCNEC
Carcinoid
SCLC
Meso
Met
SqCa, squamous carcinoma; BAC, bronchioloalveolar carcinoma; AdCa, adenocarcinoma; LCLC, large cell lung carcinoma; LCNEC, large cell
neuroendocarcinoma; SCLC, small cell lung cancer; Meso, mesothelioma; Met, met gene.
Modified from Pass HL, Carbone DP, Johnson DH, et al., eds. Lung cancer: principles and practice, 3rd ed. Philadelphia: Lippincott Williams & Wilkins,
A B
Figure 7-4 Small cell carcinoma. A, B: Contrast-enhanced CT images at the level of the aortic arch and carina show the characteristic
appearance of bulky adenopathy involving all visualized mediastinal compartments. Whereas some nodes appear discrete, others appear
coalescent, a finding consistent with extranodal spread. Note that in this case, a spiculated nodule appears in the medial aspect of the left
upper lobe in A, likely the primary tumor.
characteristically is first seen with extensive bulky central type, including both mucoepidermoid and adenoid cystic
tumor (Fig. 7-4), frequently causing obstruction of the carcinomas (14). Of these, the most distinctive radio-
superior vena cava, often without a definable primary graphically are the salivary gland tumors, as they typically
lesion. SCC is almost always associated with heavy ciga- involve the trachea and proximal mainstem bronchi.
rette smoking; only rarely is SCC seen as a solitary lung
nodule, occurring in less than 5% of cases, and when
found, has no distinguishing imaging features. NON–SMALL CELL LUNG CANCER
Large cell neuroendocrine carcinomas, although classi-
fied as large cell tumors, show features of neuroendocrine
TNM Classification
differentiation, falling between atypical carcinoids and SCC
in aggressiveness. The diagnosis of these tumors requires Accurate staging is critical in the selection of treatment and
demonstration of specific neuroendocrine markers, includ- in the evaluation of prognosis in patients with bron-
ing, for example, chromogranin A, synaptophysin, or neural chogenic carcinoma. The responsibility for accurate staging
cell adhesion molecule (14,23). LCNEC accounts for is shared by the surgeon, pulmonologist, radiologist, and
approximately 3% of resected lung cancers and 19% of pathologist (24). A uniform staging system enables every
neuroendocrine tumors (23). Although both SCC and participant in the management of the patient to communi-
LCNEC are aggressive neuroendocrine lesions with a poor cate results unambiguously. Familiarity with the current
prognosis, they differ importantly in that SCC typically staging system for lung cancer as proposed by the
responds to chemotherapy in distinction to LCNEC, for International System for Staging Lung Cancer and Regional
which surgery is the primary method of therapy. Lymph Nodes is essential to an understanding of the role
It should be noted that neuroendocrine markers may of imaging in the assessment of these patients (Table 7-4)
be identified in as many as 20% of non–small cell lung (see refs. 29,82,83).
cancers (NSCLCs), which are accordingly designated non– The staging system devised for lung cancer has evolved
small cell lung cancers with neuroendocrine differentiation over the years, after the establishment of the American
(NSCLC-NDSCC). Furthermore, both SCC and LCNEC Joint Committee on Cancer and End Results Reporting
may occur in combination with NSCLC, as well as with (AJCC) in 1959. In 1970, the task force on lung cancer of
each other, so-called “combined SCC/LCNEC” tumors. the AJCC adopted the TNM system originally proposed
by Pierre Denoix (25). Lung cancer stage is determined
by the extent of the primary tumor (T), the presence
Miscellaneous Lung Tumors
of intrapulmonary, hilar, or mediastinal lymph node
In addition to the main variants described earlier, a num- metastases (N), and the presence of extrathoracic metas-
ber of additional lesions may be encountered, including tases (M). Small cell carcinoma is still staged as either
adenosquamous cell cancer, pleomorphic carcinomas with limited or extensive disease and consequently is consid-
sarcomatous features, and carcinomas of salivary gland ered separately (26).
TABLE 7-4
DEFINITION OF PRIMARY TUMOR, NODAL INVOLVEMENT, AND DISTANT METASTASES (TNM)
CHARACTERISTICS IN LUNG CANCER
Primary Tumor (T) T4 Tumor of any size that invades any of the following:
TX Primary tumor cannot be assessed, or tumor proven by mediastinum, heart, great vessels, trachea, esophagus,
the presence of malignant cells in sputum or bronchial vertebral body, carina; or tumor with a malignant pleural
washings but not visualized by imaging or bronchoscopy or pericardial effusion,b or with satellite tumor nodule(s)
T0 No evidence of primary tumor within the ipsilateral primary-tumor lobe of the lung
Tis Carcinoma in situ
Regional Lymph Nodes (N)
T1 Tumor 3 cm in greatest dimension, surrounded by lung or
NX Regional lymph nodes cannot be assessed
visceral pleura, without bronchoscopic evidence of
N0 No regional lymph node metastasis
invasion more proximal than the lobar bronchusa
N1 Metastasis to ipsilateral peribronchial and/or ipsilateral hilar
(i.e., not in the main bronchus)
lymph nodes, and intrapulmonary nodes involved by
T2 Tumor with any of the following features of size or extent:
direct extension of the primary tumor
3 cm in greatest dimension
N2 Metastasis to ipsilateral mediastinal and/or subcarinal lymph
Involves main bronchus, 2 cm distal to the carina
node(s)
Invades the visceral pleura
N3 Metastasis to contralateral mediastinal, contralateral hilar,
Associated with atelectasis or obstructive pneumonitis
ipsilateral or contralateral scalene, or supraclavicular
that extends to the hilar region but does not involve the
lymph node(s)
entire lung
T3 Tumor of any size that directly invades any of the following: Distant Metastasis (M)
chest wall (including superior sulcus tumors), diaphragm, MX Presence of distant metastasis cannot be assessed
mediastinal pleura, parietal pericardium; or tumor in the M0 No distant metastasis
main bronchus 2 cm distal to the carina, but without M1 Distant metastasis presentc
involvement of the carina; or associated atelectasis or
obstructive pneumonitis of the entire lung
aThe uncommon superficial tumor of any size with its invasive component limited to the bronchial wall, which may extend proximal to the main
bronchus, is also classified T1.
bMost pleural effusions associated with lung cancer are due to tumor. However, there are a few patients in whom multiple cytopathologic examina-
tions of pleural fluid show no tumor. In these cases, the fluid is not bloody and is not an exudate. When these elements and clinical judgment dictate
that the effusion is not related to the tumor, the effusion should be excluded as a staging element, and the patient’s disease should be staged T1, T2,
or T3. Pericardial effusion is classified according to the same rules.
cSeparate metastatic tumor nodule(s) in the ipsilateral nonprimary-tumor lobe(s) of the lung also are classified M1.
From Mountain CF. Revisions in the International System for Staging Lung Cancer. Chest. 1997;111:1710–1717, with permission.
The initial classification scheme was published in 1973, T3 lung cancer is defined as a tumor of any size with
and a modified version appeared in 1979 (27,28). At the direct extension into the chest wall (including superior
World Conference on Lung Cancer in 1985, a new uniform sulcus tumors), diaphragm, or the mediastinal pleura or
international staging system was adopted (see ref. 43). This pericardium without involving the heart, great vessels, tra-
system, based largely on the prior AJCC formulation, has chea, esophagus, or vertebral body, or a tumor in the main
successively been modified to reflect changes in the thera- bronchus within 2 cm of the tracheal carina without
peutic approach to lung cancer. The present International involvement of the carina (Table 7-4). This may be associ-
Staging System for Lung Cancer (ISSLC), revised in 1997, ated with either atelectasis or obstructive pneumonitis
again has two major components: anatomic extent of involving the entire lung.
disease (TNM) and cell type (29; see refs. 82,83). In distinction to T3, T4 tumors may be of any size
when evidence exists of invasion of the heart or mediasti-
nal organs, including the great vessels, trachea, esophagus,
Primary Tumor Extent
or carina. Additional qualifiers include direct invasion of
T is the descriptor given to the primary tumor and its local vertebral bodies or evidence of malignant pleural or peri-
extent. The precise definitions are given in Table 7-4. As cardial effusions. Importantly, in comparison with prior
pertains to radiologic interpretation, T staging is subdi- staging systems, the current ISSLC classification also
vided in four groups: T1 to T4. T1 lung cancer is a tumor includes lesions with satellite tumor nodule(s) within the
that is 3 cm or less in greatest dimension, is surrounded by ipsilateral primary-tumor lobe of the lung (29).
lung or visceral pleura, and has no evidence of invasion
proximal to a lobar bronchus at bronchoscopy.
Nodal Disease
T2 lung cancer is a tumor more than 3.0 cm in greatest
diameter, or a tumor of any size that invades the visceral The descriptor N refers to the presence or absence of
pleura or has associated atelectasis or obstructive pneu- regional lymph node metastases (Fig. 7-5). The definitions
monitis extending to the hilar region. proposed in the 1997 ISSLC revision are listed in Table 7-1
A B
C D
Figure 7-5 Nodal mapping. See Color Figure 7-5 A–D. A: Diagram depicting intrathoracic nodal groups as classified by the International
Staging System. B–D: Color-coded cross-sectional images illustrating the location of high paratracheal nodes (B), aortopulmonary window
nodes (C), and pericarinal nodes (D), respectively. (From Ko JP, Drucker EA, Shepard JO, et al. CT Depiction of regional nodal stations for
lung cancer staging. Am J Roentgenol. 2000;174:775–782, with permission.)
and include anatomic landmarks for 14 hilar, intrapul- These nodes are all within the lung and are completely
monary, and mediastinal lymph node stations (Table 7-5). covered by pleura. Although the presence of N1 disease
It should be emphasized that although prognosis is affected adversely affects prognosis, it usually does not affect resec-
by the number and size of involved nodes, as well as the tability. However, N1 status may necessitate a pneumonec-
levels of involvement, and intracapsular versus extracapsu- tomy (rather than a lobectomy) for total removal of disease.
lar, or microscopic versus macroscopic nodal disease, these N2 status includes all patients with ipsilateral mediastinal
characteristics are not considered in the current TNM lymph node or subcarinal lymph node involvement. Sub-
staging system. carinal nodes remain N2 despite the fact that in a few series,
N0 indicates no demonstrable metastases to regional when careful nodal analysis was undertaken, metastases
lymph nodes. Only 10% to 15% of patients have N0 status at to the subcarinal lymph nodes carried a poor prognosis
diagnosis. N1 status is characterized by direct tumor exten- (30–32). Some disagreement exists about the significance of
sion or metastases to ipsilateral peribronchial or hilar nodes. uppermost ipsilateral mediastinal lymph node involvement,
TABLE 7-5
LYMPH NODE MAP DEFINITIONS
Nodal station Anatomic landmarks
N2 nodes—All N2 nodes lie within the mediastinal pleural envelope
1 Highest mediastinal nodes Nodes lying above a horizontal line at the upper rim of the brachiocephalic (left innominate) vein
where it ascends to the left, crossing in front of the trachea at its midline
2 Upper paratracheal nodes Nodes lying above a horizontal line drawn tangential to the upper margin of the aortic arch and
below the inferior boundary of N1 nodes
3 Prevascular and Prevascular and retrotracheal nodes may be designated 3A and 3P; midline nodes are considered
retrotracheal nodes to be ipsilateral
4 Lower paratracheal nodes The lower paratracheal nodes on the right lie to the right of the midline of the trachea between a
horizontal line drawn tangential to the upper margin of the aortic arch and a line extending across
the right main bronchus at the upper margin of the upper lobe bronchus, and contained within the
mediastinal pleural envelope; the lower paratracheal nodes on the left lie to the left of the midline
of the trachea between a horizontal line drawn tangential to the upper margin of the aortic arch
and a line extending across the left main bronchus at the level of the upper margin of the left
upper lobe bronchus, medial to the ligamentum arteriosum and contained within the
mediastinal pleural envelope
Researchers may wish to designate the lower paratracheal nodes as N4s (superior) and N4i (inferior)
subsets for study purposes; the N4s nodes may be defined by a horizontal line extending across
the trachea and drawn tangential to the cephalic border of the azygos vein; the N4i nodes may be
defined by the lower boundary of N4s and the lower boundary of N4, as described above.
5 Subaortic nodes (aortopulmonary Subaortic nodes are lateral to the ligamentum arteriosum or the aorta or left pulmonary artery and
window) proximal to the first branch of the left pulmonary artery and lie within the mediastinal pleural
envelope
6 Para-aortic nodes Nodes lying anterior and lateral to the ascending aorta and the aortic arch or the innominate
(ascending aorta or artery, beneath a line tangential to the upper margin of the aortic arch
phrenic)
7 Subcarinal nodes Nodes lying caudal to the carina of the trachea, but not associated with the lower lobe bronchi or
arteries within the lung
8 Paraesophageal nodes Nodes lying adjacent to the wall of the esophagus and to the right or left of the midline, excluding
(below carina) subcarinal nodes
9 Pulmonary ligament nodes Nodes lying within the pulmonary ligament, including those in the posterior wall and the lower part
of the inferior pulmonary vein
N1 nodes—All N1 nodes lie distal to the mediastinal pleural reflection and within the visceral pleura
10 Hilar nodes The proximal lobar nodes, distal to the mediastinal pleural reflection and the nodes adjacent to the
bronchus intermedius on the right; radiographically, the hilar shadow may be created by
enlargement of both hilar and interlobar nodes
11 Interlobar nodes Nodes lying between the lobar bronchi
12 Lobar nodes Nodes adjacent to the distal lobar bronchi
13 Segmental nodes Nodes adjacent to the segmental bronchi
14 Subsegmental nodes Nodes around the subsegmental bronchi
From Mountain CF, Dresler CM. Regional lymph node classification for lung cancer staging. Chest. 1997;111:1718–1723, with permission.
which is listed as a contraindication for surgical resection by sequently, metastases to the scalene and supraclavicular
some authors (33,34) but not by others (35). nodes are classified as N3 rather than M1.
N3 status includes metastases to contralateral mediasti-
nal, contralateral hilar, and both ipsilateral and contralateral
Distant Metastases
supraclavicular nodes. Although, as is discussed later, some
controversy surrounds the interpretation of N2 disease, there The descriptor M relates to the presence of distant metas-
is general agreement that N3 status implies nonresectability. tases (M1) or their absence (M0) (Table 7-4). Although sca-
It should be emphasized that in selected cases, N3 status lene, cervical, and contralateral hilar nodes are not
may be conferred as the result of clinical sequelae, as in considered evidence of distant metastases, one of the
patients with signs of recurrent laryngeal nerve involvement. most important changes introduced in the 1997 ISSLC
It should also be noted that in the current staging system, classification was altering M1 status to include patients with
both ipsilateral and contralateral supraclavicular and scalene separate metastatic tumor nodules in the ipsilateral non–
nodes are designated “regional” nodes because radiothera- primary-tumor lobe(s) of the lung (29). Extrathoracic
pists can include them in the radiation therapy portal. Con- metastases are present in approximately 40% of patients
A B
C
Figure 7-6 Squamous cell carcinoma: metastatic disease. A: Enlarged CT section through the left lower lobe shows a cavitary thick-walled
lesion consistent with squamous cell carcinoma. Note an apparent filling defect within the cavity, in this case consistent with intracavitary
hemorrhage. B: Magnified CT image shows the characteristic appearance of a heterogeneous adrenal metastasis on the left side. C:
Gadolinium-enhanced MRI showing a rim-enhancing left frontal lesion associated with extensive edema, causing considerable mass effect.
with newly diagnosed bronchogenic carcinoma (36) and reflect differences in survival between these two groups; 67%
carry a very poor prognosis (Fig. 7-6). Unfortunately, of patients with surgical–pathologic stage IA disease survive
clinically silent sites of distant metastatic disease are 5 years compared with 57% with stage IB disease (29).
frequently discovered post mortem. Mathews et al. (37) Similarly, the new classification calls for subdividing
studied 200 patients who died within 30 days of presumed previous stage II disease into IIA (T1N1M0) and IIB
curative surgery. At autopsy, 48 (23.4%) of these patients (including both T2N1M0 and T3N0M0), reflecting the
had metastases, 18 of them to the adrenal gland and 16 to prognostic implications of hilar nodal status, tumor size,
the liver. In a comparable population, a 26% incidence of and tumor location (including extrapulmonary extension
extrathoracic metastases was found at autopsy (38). In a in the absence of adenopathy); 55% of patients with stage
study of 95 patients with bronchogenic carcinoma in whom IIA survive 5 years compared with 38% and 39% of
preoperative CT scan showed a solitary lung mass without patients with T2N1M0 and T3N0M0 tumors, respectively
evidence of hilar or mediastinal lymphadenopathy, 25% (29). It should be noted that as part of this revision,
had extrathoracic metastases (39). Interestingly, in this same T3N0M0, previously classified as stage IIIA, is now consid-
study, the majority of these patients (67%) had adenocarci- ered stage IIB (Table 7-6 and Fig. 7-7).
noma. More recently, Finke et al. (40) reviewed autopsy Modifications in the stage III classifications include
results performed within 30 days of surgery in a total of reclassifying T3N0M0 as stage IIB, as noted earlier to
32 patients who had undergone curative resection of reflect the generally better prognosis of peripheral lesions
NSCLC and identified extrathoracic metastases in 5 (16%) with limited focal chest-wall invasion. Otherwise, stage III
patients, including in the liver in two patients and the lung, remains unchanged, albeit heterogeneous, subdivided
epicardium, adrenal gland, and kidney in one patient each. into two subgroups. Stage IIIA (T3N1M0 or T1-T3N2M0)
Brain metastases are the most frequent site of distant represents patients with advanced disease in whom exten-
metastatic disease in patients without lymph node disease sive surgery (including resection of all ipsilateral and
(Fig. 7-6). Although more often occurring in patients with subcarinal lymph nodes) may be of value. Stage IIIB in-
SCLC, overall, in approximately 15% of patients with cludes all cases of more advanced disease without distant
NSCLC, brain metastases will develop within 5 years after metastases for which surgery is generally contraindicated.
resection (41). The likelihood of developing brain metas- Stage IV represents any patient with evidence of distant
tases increases with increasing tumor stage (42). Although metastases.
brain lesions clearly fall into the category of distant metas-
tases, interestingly, no differentiation is made between
Current Controversies
solitary and multiple brain metastases, although isolated
brain metastases may be an indication for surgical resec- Although the current ISSLC proposed in 1997 has proved a
tion. Whereas resection of solitary brain metastases is consistent and reliable method for assessing prognosis, a
indicated, unfortunately, as documented by one recent number of questions and controversies regarding this system
study, the 5-year survival rate in patients with operable have arisen.
brain metastases remains identical, regardless of initial In one study, for example, assessing the reliability of the
stage, equivalent to survival in patients with stage IIIA new staging system when applied to an unselected lung can-
disease at 23% (41). cer population, each stage and substage accurately reflected
differences in survival (with the sole exception of stage IIA,
for which insufficient numbers of cases were available for
Stage Groupings
statistically meaningful evaluation). However, comparing
Based on comparable extent of disease, anticipated treat- the 1986 and the 1997 classifications, no substantial differ-
ment, and prognosis, TNM subsets have been defined ences were observed in a cohort of 1,296 consecutive
corresponding to four stages of disease, not including patients (p 0.0001), leading these authors to speculate
carcinoma in situ (Fig. 7-7 and Table 7-6). Initially formu- that simple reliance on TNM descriptors alone might repre-
lated on the basis of results of careful studies by Mountain sent an improvement over the current staging system (44).
(43) of the prognostic factors and treatment outcomes in Although a complete review of potential limitations of
3,753 patients treated in tertiary care facilities (43), modifi- the current staging system is outside the scope of the pres-
cations have been adopted most recently in 1997 (29). The ent review, a number of issues seem especially likely to
most important of these include dividing previous stage 1 lead to revisions when the newest recommendations are
and stage II disease into stage IA, IB, IIA, and IIB categories; reported in 2007 (45). These include, among others,
the assignment of T3N0M0 to stage IIB; the designation of substaging stage 1 to reflect current data regarding differ-
satellite lesions in the same lobe as the primary tumor as ences in prognosis among T1 lesions of different sizes;
T4; and the assignment of primary lesions with synchro- re-evaluation of the significance of satellite lesions, espe-
nous satellite tumor within any other lobe as M1 disease. cially those identified in the same lobe as the primary
As noted, in the latest classification, stage I is now subdi- tumor; and the significance of small pleural effusions
vided into IA (T1N0M0) and IB (T2N0M0) categories to identified only by CT.
A B
C D
Figure 7-7 A–E: Diagrammatic representations depicting stage

IA through IV non–small cell lung cancer, respectively. See Color Figure
7-7 A–E. (Images courtesy of Thomas W. Rice, MD. Reprinted with per-
E mission of The Cleveland Clinic Foundation.)
Substaging Stage I Lung Cancer T1 lesions have the same prognosis, raising the question of
One question that has received considerable attention is whether substaging of stage I lesions should be performed
whether the size of T1 lesions matters. At present, T1 lesi- (46). In a review of 7,620 stage I patients with NSCLCs
ons are defined as smaller than 3 cm. However, growing undergoing curative resections, by using data compiled
evidence—most recently derived from data obtained from from the Surveillance, Epidemiology, and End Results
low-dose lung cancer screening trials—indicates that not all (SEER) 2003 registry, for example, Wisnivesky et al. (47)
TABLE 7-6
TNM STAGING OF LUNG CANCER
Mediastinal
Scalene (ipsi-/contralateral)
Peribronchial (ipsilateral)
Hilar
Lymph Node (N)

Supraclavicular
Subcarinal
(contralateral)
(contralateral)
Node (N)
(ipsilateral)
(ipsilateral)
Lymph
Stage IV
M1 (any T or N)
N3 Stage IIIB
/ / / N3 N2 Stage IIIA
MO
&/ N2 N1 Stage IIA Stage IIB
N0 Stage IA Stage IB Stage IIB
*
&/ N1
T1 T2 T3 T4
N0
Primary
T1 T2 T3 T4
Stage 0 Tumor (T)
(Tis, N0, M0)
a&b&c any of a,b,c,d (a&c)/b/d (a&c)/d Criteria
3 cm 3 cm any any a. Size
No invasion Main bronchus Main bronchus b. Endo-
Metastases (M)
proximal to the ( 2 cm distal ( 2 cm distal – bronchial
M0: Absent
lobar bronchus to the carina) to the carina) location
M1: Present
Mediastinum/
Separate metastatic tumor nodule(s) in Chest wall**/
trachea/heart/
the ipsilateral non–primary-tumor lobe(s) Surrounded by diaphragm/
great vessels/ c. Local
of the lung also are classified M1 lung or visceral Visceral pleura mediastinal/
esophagus/ invasion
pleura pleura/parietal
vertebral body/
pericardium
carina
Tis: Carcinoma in situ Malignant
Atelectasis/
Staging is not relevant for occult pleural/peri-
obstructive
carcinoma (Tx, N0, M0) cardial effusion
pneumonltis Atelectasis/
or
that extends to obstructive
– satellite tumor d. Other
* Including direct extension to the hilar region pneumonltis of
nodule(s) within
intrapulmonary nodes but doesn’t the entire lung
the ipsilateral
** Including superior sulcus tumor involve the
primary-tumor
entire lung
lobe of the lung
(&: and) (/: or) (&/ : and/or)
Reprinted from Mason RJ, Broaddus UC, Murray JF, et al., eds. Murray and Nadel’s textbook of respiratory medicine, 4th ed. Philadelphia: Elsevier;
divided lesions into those between 5 and 15 mm, 16 and 130 consecutive resections for pathologic stage IA NSCLCs,
25 mm, 26 and 35 mm, 36 and 45 mm, and larger than in which lesions were classified as 0 to 20 mm versus 21 to
45 mm in size, respectively, and showed statistically higher 30 mm in size, with a median follow-up of 6.8 years,
cure rates for all categories of tumors (p 0.05) except reported by Birim (49), both the 5-year disease-free interval
those between 26 and 35, and 36 and 45 mm (47). Gajra and overall survival proved superior for smaller lesions
et al. (48), in a retrospective review of 246 consecutive (68% vs. 48% and 69% vs. 51%, respectively, for lesions
patients with pathologically proved stage IA cancers, smaller than versus larger than 2 cm).
divided cases into those with lesions 1.5 cm or larger It should be noted that not all reports have concluded
(n 86) and those with lesions measuring 1.6 to 3.0 cm that subdividing T1 lesions will necessarily lead to altered
(n 160) and showed a significantly improved overall prognosis (Fig. 7-8). In one retrospective study of 510
5-year survival in the group with smaller lesions (85.5% vs. patients with pT1N0MO lesions who were identified from a
78.6%, respectively; p 0.05). In a similar study evaluating tumor registry and were undergoing resection, no statistically
A B
Figure 7-8 Peripheral adenocarcinoma, stage IIIA disease. A: Magnified view through the right upper lobe imaged with lung windows
shows an approximate 1.5-cm well-defined peripheral nodule. B: Magnified view of the mediastinum at the level of the aortic arch shows an
enlarged right paratracheal node (4R). Density along the left lateral edge of the trachea is consistent with retained secretions. This case
illustrates that even small lesions (2 cm) may still be biologically aggressive. Metastatic nodal disease was biopsy confirmed.
significant relationship could be established between including 436 NSCLCs (376 prevalence cancers, 88 inci-
survival and tumor size for lesions smaller than 3 cm dence cancers) further categorized by both size and consis-
(p 0.701) (50). Although these data contradict the notion tency (53a). Of these, surgical staging was obtained in 368
of the need to substage stage I lesions, it should be noted (84%) of 426 patients. Stage IA tumors were identified in
that reported observed survival in this study was based on 91% of solid lesions smaller than 15 mm, 83% of solid
deaths from all causes rather than disease-free survival lesions between 16 and 25 mm, 68% of solid lesions
or overall survival. Furthermore, although only 26 (5%) between 26 and 35 mm, and 55% of solid lesions measur-
patients had lesions smaller than 1 cm, 80% of lesions meas- ing greater than 36 mm.
ured between 2 and 3 cm, 90% of which proved resectable, Although reasons exist to consider subdividing stage IA
suggesting that advanced-stage cancers were commonly not lesions into T1a and T1ab, there is no consensus as to the
registered as nonsurgical candidates. A question remains appropriate size cut-off. Further confounding interpreta-
whether this particular study had sufficient power, as only tion, most studies, to date, have failed to take into account
62 deaths were reported among the 510 patients studied the morphology of lesions as established by CT—specifi-
(51). It should be emphasized that not all lesions smaller cally, whether lesions are subsolid ground-glass opacities,
than 2 cm are necessarily earlier-stage lesions. In fact, 21% of mixed solid–subsolid, or solid lesions.
clinical stage IA lesions prove to have mediastinal adenopa- It should also be noted that a similar rationale has been
thy at the time of diagnosis, whereas in 24%, evidence exists offered to distinguish between T1 and T2 lesions. In one
of distant metastases, 13% at the time of diagnosis and 11% study, for example, using 5 cm as a cut-off compared with
within 1 year (52). Nonetheless, the likelihood of adenopa- lesions measuring between 3 and 5 cm, a significant differ-
thy in particular does appear to be related to T status. In one ence in survival was noted (46% vs. 61%) (46), leading to
retrospective study of 503 patients with completely resected the suggestion that lesions larger than 5 cm should be
NSCLCs 3 cm or smaller, lesions larger than 2 cm proved reclassified as T3 (54).
twice as likely to have nodal metastases as did those 2 cm or
smaller (53). Furthermore, by using multivariate analysis, Synchronous Satellite Lesions
only tumor size proved to correlate with nodal status (53). Also controversial is the significance of synchronous satel-
To date, the most compelling data supporting the no- lite lesions. It has been suggested, for example, that classifi-
tion that smaller size conveys an earlier stage have been cation of synchronous pulmonary nodules in the same
obtained from low-dose (LD) CT screening trials. In one lobe as the primary tumor as T4 stage IIIB lesions is unwar-
study, of a total of 28,689 asymptomatic individuals ranted (Figs. 7-9 and 7-10). To date, several studies have
screened with LDCT, 464 lung cancers were diagnosed, shown that the finding of synchronous satellite lesions in
A B
C D
Figure 7-9 Tumor staging: T4 satellite lesions. A: CT section

at the level of the carina imaged with lung windows shows a
2.6 2.3–cm irregular nodule in the medial aspect of the left
upper lobe. B: Contrast-enhanced CT section at approximately the
same level as in A shows an enlarged ipsilateral anterior mediasti-
nal para-aortic (station 6) node. Based on these images, this repre-
sents a stage II (T2N2M0) lesion. C: Section several centimeters
above A shows an approximate 9 9-mm satellite lesion in the
same lobe. This finding technically results in this tumor being
restaged as stage IIIB (T4). D: Follow-up contrast-enhanced CT
section at the same level as shown in B obtained 3 months after
adjuvant chemotherapy, showing an apparent complete response.
E: Section obtained at the same time as D shows that whereas
the primary lesions has nearly completely disappeared, this satel-
lite lesion has only marginally decreased in size. This case
illustrates the difficulty in accurately assessing the significance
of satellite lesions in the absence of definitive histologic staging as
well as the limitations of CT in predicting response to therapy. This
E patient is currently awaiting definitive surgical resection.
the same lobe as the primary tumor does not necessarily significant improvement in survival was noted at 2 years
imply a worse prognosis, especially when these are dis- (41.5% vs. 20%) when compared with synchronous lesions
covered only incidentally at the time of surgical resection identified in different lobes. This difference in prognosis
(55,56). As reported by Shimizu et al. (57), for example, may reflect differences in the manner in which disease is
in a study of 42 patients with satellite lesions in the spread, with satellite lesions in the same lobe as the primary
same lobe as primary tumors not detected before surgery, a lesions resulting from local pulmonary arterial invasion and
tics to differentiate between these subsets of patients

presumably must await further prospective evaluation.
Pleural Effusions Identified Only

by Computed Tomography
As noted in the 1997 revision, tumors associated with
malignant effusions are currently defined as T4 lesions. As
cytology is frequently negative despite malignant pleural
disease, it has been proposed that the finding of a hemor-
rhagic and/or exudative effusion in itself could also be
considered malignant (29). Some have argued that the find-
ing of a malignant effusion should more properly be consid-
ered M1 disease, as the prognosis in these cases is similar to
the finding of extrathoracic disease. Interpreting the signifi-
cance of pleural effusions has become more problematic, as
it is not infrequent for small quantities of pleural fluid to be
identified by CT, otherwise impossible to identify on rou-
tine chest radiographs and too small to aspirate easily.
Figure 7-10 Tumor staging: synchronous cancers vs. metastatic Although not directly addressed by the current staging
disease. CT section shows two discrete nodules of approximately
the same size in the superior segment of the right lower lobe and
system, it is likely that many of these will prove to be be-
the middle lobe, respectively, in the absence of obvious hilar or nign, especially in the absence of CT evidence of pleural
mediastinal adenopathy (not shown). Although subtle differences thickening or loculation. Although the potential for fluo-
are seen in their morphology, it is not possible, on the basis of this
image, to differentiate two synchronous primary lesions from one
rodeoxyglucose–positron emission tomography (FDG-PET)
dominant lesion and a metastasis involving a different lobe. In the to differentiate more definitively between benign and malig-
absence of mediastinal adenopathy or evidence of extrathoracic nant pleural and/or pericardial effusions is currently under
disease, both lesions should be considered resectable. At surgery,
both proved to be adenocarcinomas. Lack of lymphatic involve-
evaluation, definitive guidelines regarding the use of FDG-
ment is consistent with presumed synchronous lesions. PET for this purpose remain to be established (60–62).
subsequent embolization rather than true hematogenous Additional Staging Issues

metastatic disease (45). Determining whether a synchro- In addition to reconsidering the need to substage stage I
nous satellite lesion represents a true second primary versus lesions, re-evaluate the staging of satellite lesions, and deter-
a metastatic lesion remains problematic, especially when mine the significance of small effusions identified only by
they have similar histologic appearances. Similar difficulty CT, a number of relatively rarer presentations of lung cancer
is encountered when differentiating metachronous pri- occur that differ from predicted outcomes by the current
maries from recurrent metastatic disease. As suggested by staging system. Patients with skip mediastinal lymph node
Martini and Melamed (58), synchronous tumors are metastases, for example, have a slightly better prognosis
defined either as having different histology, or if the same than do patients in whom both hilar and mediastinal nodes
histology is present, defined as synchronous if they are iden- are identified (63). Similarly, better outcomes have been
tified in different segments, lobes, or the contralateral lung, described in patients with so-called “early hilar lung cancer
provided no evidence of tumor is found in lymphatics (EHLC).” As defined by the Japan Lung Cancer Society,
common to both lesions, and no evidence of metastatic EHLC is a T1N0M0 cancer that is located in a segmental,
disease is seen. Identical guidelines have been proposed as lobar, or main bronchus with invasion limited to the
well for differentiating metachronous primaries from recur- bronchial wall. Terzi et al. (64), in a study of 29 patients
rent metastatic lesions. with squamous cell tumors conforming to this definition,
Although of proven value, these guidelines were pro- not surprisingly reported a 5-year cumulative survival rate
posed based on routine radiographic evaluation before the after resection of 96%. A number of interventional bron-
widespread availability of CT scanners. Numerous studies choscopic approaches to these focal endobronchial lesions
have subsequently established that subcentimeter lesions have been proposed, including brachytherapy and electro-
as small as only a few millimeters, impossible to identify cautery, among others, with varying results (65).
with chest radiography, may be identified in as many as The importance of histologic evidence of blood vessel
60% of asymptomatic individuals with multidetector CT, invasion has also been questioned (66), as has the signifi-
the vast majority of which prove to be benign (59). This cance of microscopic nodal disease, with implications
suggests that the finding of a subcentimeter nodule—one for the use of routine immunohistochemical staining
unlikely to be identified on a routine chest radiograph— (15). It has been shown, for example, that in patients with
should not necessarily be construed as evidence of T4 histologic evidence of N0 disease, immunohistochemical
status and in our opinion should not in itself preclude analysis may reveal otherwise unsuspected micrometas-
surgery (45). Determination of acceptable CT characteris- tases in as many as 40% of cases, leading to statistically
poorer survival (15). A similar impact on survival has been disease based on routine chest radiographs as well as an
described in cases in which evidence is found of blood estimate of the patient’s clinical status. For example, in
vessel invasion by the primary tumor. In one study of patients with suspected stage IA disease in whom a diag-
72 patients with T1–3N0 disease undergoing resection, nosis of malignancy has not yet been established, initial
although the overall 5-year survival equaled 62.5%, in scans should be obtained before the administration of in-
cases with documented blood vessel invasion, the survival travenous contrast, using thin (1.5 to 2 mm) sections to
rate was only 23.5% (66). Although these issues have clear measure tissue density accurately. In selected cases, further
therapeutic implications, it is unlikely that they will result evaluation using contrast enhancement to evaluate tissue
in specific stage modifications. vascularity may be indicated, especially for well-defined
rounded lesions (see Figs. 6-36 to 6-39). In distinction, in
cases in which tumor appears primarily central on corre-
MULTIDISCIPLINARY IMAGING sponding chest radiographs, especially when associated
IN THE EVALUATION AND STAGING with peripheral atelectasis, images should be acquired
OF LUNG CANCER after a bolus administration of intravenous contrast
media to allow optimal differentiation between central
lesions and postobstructive pneumonitis (Fig. 7-11).
Attention should also focus on the need to ascertain the
There is no optimal CT technique to stage lung cancer. relation between tumor and the carina, frequently necessi-
Even in the era of multidetector CT scanners, each study tating retrospective reconstruction of overlapping thin
should reflect an initial estimate of the probable stage of (1 mm) sections.
A B
Figure 7-11 Central endobronchial tumor: use of contrast

enhancement. A–C: Sequential contrast-enhanced CT images
beginning at the level of the distal trachea (A) and extending infe-
riorly to the proximal ascending aorta (C) show extensive tumor
infiltrating the right hilum, encasing the proximal right interlobar
pulmonary artery, narrowing the right mainstem bronchus, and
obstructing the remaining right-sided airways, with evidence of
tumor infiltration of the subcarinal space (arrow in B). A subtle but
distinct difference is seen in the density of tumor vs. peripheral
atelectatic lung (arrows in C). Also noted are markedly enlarged
mediastinal nodes, as well as marked compression of the postero-
lateral wall of the right atrium, likely causing obstruction of the
right inferior pulmonary vein. A large right-sided effusion is also
present. The use of intravenous contrast enhancement is clearly of
value in cases in which complex disease involves multiple anatomic
C structures.
Computed Tomography Evaluation with sensitivities and specificities ranging from 63% to
of Tumor Extent 90% and specificities of 84% to 86% (24).
As will be discussed, one notable exception for the role
Chest-wall Invasion of MR is the evaluation of superior sulcus tumors. A num-
The main role of CT is in differentiating T3 from T4 ber of other CT findings have also been assessed but are
lesions. Unfortunately, from the outset, it has been appar- considerably less accurate. Glazer et al. (72), in a study of
ent that CT is limited in assessing chest-wall invasion, with 47 patients with peripheral tumors contacting the chest
reported sensitivities and specificities ranging from 38% to wall or pleural surface, found that the presence of an
87% and from 40% and 89%, respectively (67). The only obtuse angle between the mass and pleura, pleural contact
CT findings that allow a reliable diagnosis of chest-wall over a distance greater than 3 cm, and associated pleural
invasion are the presence of rib destruction (68) or an thickening yielded a sensitivity of 87% for chest-wall
obvious chest-wall mass (Fig. 7-12) (69). These findings, invasion, with a specificity of only 59% and a diagnostic
however, have a low sensitivity, being present in only 20% accuracy of 68%. In this study, the presence of focal chest
to 40% of patients with surgically proven chest-wall inva- pain, although not as sensitive as CT, was more specific
sion (70,71). Despite superior tissue contrast, MR has also (94%) and accurate (85%) in making the diagnosis of
proved of limited value for identifying chest-wall invasion, chest-wall invasion (72). Thus tumors abutting the pleura,
B C
Figure 7-12 CT evaluation of tumor extent: chest-wall invasion. A: Magnified CT section through the right apex shows a superior sulcus
tumor with subtle extension into the adjacent chest wall, resulting in lack of definition of chest-wall fascial planes (arrow in A). B: Coned-
down posteroanterior chest radiograph from a different patient than in A shows extensive soft tissue density obliterating the left apex
associated with destruction of the adjacent second and third ribs. C: Magnified CT section through the left apex, in the same patient as in B,
demonstrates extensive tumor infiltration into the adjacent chest wall associated with lytic destruction of the second rib. In the absence of
the signs illustrated in these two cases, definitive evaluation of chest-wall infiltration with CT is extremely limited.
even when associated with local pleural thickening, may invasive because associated desmoplastic reaction and
not be invasive (70–72). inflammation can simulate tumor extension into an adja-
Two studies have suggested that the accuracy of CT in cent structure (70). Likewise, microscopic invasion may be
the detection of parietal pleural and chest-wall invasion can completely missed. The low reliability of CT in this context
be improved by assessing the movement of lesions along has been documented by several authors (77–79). Overall,
the chest wall during expiration as evidence of a lack the sensitivity, specificity, and accuracy of CT for identifying
of pleural fixation or invasion (73,74). Although based on a mediastinal invasion have been reported to vary, respec-
limited number of cases, it appears that the finding of respi- tively, from 40% to 84%, 57% to 94%, and 56% to 89%
ratory phase shifts in the relation between tumor and (24). Although additional criteria have been proposed
the adjacent pleural surfaces is evidence of a lack of parietal to identify patients with indeterminate mediastinal or
pleural and/or chest-wall invasion. Despite this, use of pleural invasion whose lesions are likely to be technically
paired inspiratory–expiratory images to evaluate tumor– resectable—including (a) less than 3 cm of contact between
pleural relations has not become standard practice. Sim- the tumor and the adjacent mediastinum; (b) less than 90
ilarly, whereas it has been reported that the use of induced degrees of circumferential contact between the tumor and
pneumothorax results in a sensitivity of 100% in the detec- the aorta, and the presence of mediastinal fat between the
tion of chest-wall and mediastinal pleural invasion on CT, mass and adjacent mediastinal structures—these findings
with specificities ranging from 57% to 80% (75,76), this have proven unreliable as a means for assessing individual
technique has not gained widespread acceptance and is cases (Fig. 7-13) (77).
unlikely ever to be used in routine clinical practice. With the advent of bronchial reconstruction procedures,
selected patients whose tumors are within 2 cm of the tra-
Mediastinal Invasion cheal carina (T3) can also undergo curative resection
Mediastinal invasion that may be technically resectable (T3) (80,81). Determination of the proximity of lesions to the
includes limited invasion of the pericardium, mediastinal carina, however, may be problematic. This is especially true
pleura or fat, and vagus and phrenic nerves (74,77). A lesion of left-sided SCCs in which the oblique orientation of the
that abuts or makes only limited contact with the left mainstem bronchus makes assessment of submucosal
mediastinum, even if adjacent to areas of pleural or pericar- extension of tumor to the carina particularly difficult to
dial thickening, cannot confidently be diagnosed as being diagnose. As tumors more than 2 cm from the carina may
A B
Figure 7-13 CT evaluation of tumor extent: mediastinal invasion. A: Magnified CT section at the level of the great vessels after intravenous
contrast administration shows a lobular mass (arrow) in the medial aspect of the right upper lobe abutting the lateral and posterior wall of the
trachea. Although a biopsy of this lesion can be performed via transbronchial needle aspiration, the results should not be overinterpreted to
suggest either mediastinal invasion and/or nodal disease. B: Contrast-enhanced CT in a different patient than shown in A through the lower
lobes shows a slightly irregular mass abutting the lateral wall of the mid descending aorta. Although it has been suggested that less than 3 cm
of contact or less than 90 degrees of circumferential contact between the tumor and the aorta indicates resectability, only near-circumferential
encasement with tumor extending into the adjacent mediastinal fat should be taken as signs of unresectability (see Fig. 7-14). In both cases,
surgical resection proved successful.
A B
Figure 7-14 CT evaluation of tumor extent: mediastinal inva-

sion. A: Contrast-enhanced CT at the level of the carina shows
extensive tumor infiltration into the left hilum encasing the left
main pulmonary artery to its origin. Note that tumor also extends
to the lateral aspect of the left mainstem bronchus and extends
abutting a considerable portion of the anterolateral aortic wall. In
addition to the finding of direct mediastinal invasion, evidence of
tumor extending to the carina in itself precludes surgical resection.
B, C: CT sections in a different patient than shown in A reveal
an irregular cavitary mass in the right upper lobe (B) associated
with extensive tumor, causing marked narrowing and irregular
nodular thickening of the right and left mainstem bronchi. In these
cases, CT findings are sufficiently definitive to obviate FDG-PET
C staging.
be resectable by using a “sleeve lobectomy” in place of a

pneumonectomy, accurate assessment of the true extent of
disease requires bronchoscopic correlation (81a).
As previously noted, T4 status includes tumors of any
size with invasion of the mediastinum or involving the
heart, great vessels, trachea, esophagus, vertebral bodies,
or tracheal carina or associated with a malignant pleural
effusion (Figs. 7-14 to 7-16). As recently modified, T4 also
now includes evidence of satellite tumor nodule(s) within
the ipsilateral primary-tumor lobe of the lung (Fig. 7-9).
Technically, patients with T4 tumors are considered to
have surgically unresectable disease with overall 5-year
survival of less than 10% (29,82,83). As discussed earlier,
the significance of a satellite lesion in the same lobe as a
primary tumor is controversial and does not necessarily
preclude surgical resection.
The most important role for CT in assessing tumor extent
is determining whether a T4 classification is warranted on
the basis of unequivocal CT findings (Figs. 7-11, 7-12, and
7-14 to 7-16). In selected cases, CT can reliably determine Figure 7-15 CT evaluation of tumor extent: vertebral body inva-
sion. Contrast-enhanced CT section shows a large tumor (T4) in the
the presence of involvement of major mediastinal struc- right lower lobe directly invading the adjacent vertebral body
tures, such as the heart, great vessels, trachea, esophagus, (arrow).
main pulmonary artery, right or left pulmonary artery

beyond the mediastinal pleural reflection, or a major
branch of any vessel within the mediastinal fat.
3. Erosion of a vertebral body or involvement of the
brachial plexus.
4. Malignant effusions. Although a pleural effusion per se
is nonspecific, the finding of associated discrete soft
tissue nodules involving the visceral or parietal pleura
may be taken as definitive signs of a malignant effusion.

of Nodal Disease
Identification of enlarged nodes and assignment of these
to particular stations requires detailed knowledge of cross-
sectional anatomy. The current system for interpreting
nodal disease, as defined by the American Thoracic
Figure 7-16 CT evaluation of tumor extent: stage IIIB mediasti- Society, is based on the concept that thoracic nodes are
nal, pleural, and chest-wall invasion. Contrast-enhanced CT section better defined if their relation to fixed anatomic structures
at the level of the right main pulmonary artery shows extensive (such as the azygos vein, bronchi, or aortic arch) can be
tumor (T4) infiltrating the mediastinum (black arrow) as well as
the chest wall (curved arrow). Note that the pleura on the right side
described (Fig. 7-5) (82). These anatomic structures are
is circumferentially thickened, nodular, and irregular because of ex- easily recognizable at surgery, mediastinoscopy, and for
tensive infiltration by tumor (white arrows). No pleural fluid could the most part on CT scans. In this regard, a number of
be identified despite extensive pleural and chest wall involvement.
pictorial reviews have been published clarifying these rela-
tions (84,85), as well as delineating the source and direc-
and vertebral bodies. However, definite involvement should
tion of thoracic lymphatics (86).
be considered only if clear-cut evidence of destruction exists
N1 nodes include all nodes within the pleural reflec-
or if CT evidence of encasement or interdigitation of tumor
tion at surgery (Fig. 7-17). This definition is somewhat
around a given structure can be demonstrated. As a general
arbitrary because the visceral pleura cannot always be
guide, CT diagnosis of T4 status can be confidently diag-
visualized and some nodes may be partially covered by vis-
nosed if the following conditions are met:
ceral pleura, even though they are anatomically within the
1. Tumors involving the trachea or narrowing the carina, hilum. In this regard, separation between these groups
usually involving both sides of the carina. with current imaging methods—both CT and MR—is
2. Tumors surrounding, encasing, or abutting for more problematic, especially when attempting to differentiate
than 180 degrees the superior vena cava, the aorta, 4R (lower paratracheal) from10R (right hilar) nodes. To
A B
Figure 7-17 Nodal staging: N1 disease. A, B: Contrast-enhanced CT sections in two different patients showing evidence of bulky right
hilar (A) and left hilar (B) nodes. Note lack of obvious subcarinal adenopathy in both cases.
facilitate interpretation, it has been proposed that nodes

lying between the cephalic border of the azygos vein and a
line extending across the right main bronchus at the upper
margin of the right upper lobe bronchus are designated
4i (for inferior) and should be interpreted as lying within
the mediastinal pleural envelope (82).
N2 nodes include ipsilateral mediastinal and subcarinal
nodes, whereas N2 nodes include contralateral hilar and
mediastinal nodes as well as both ipsilateral and contralat-
eral scalene nodes and supraclavicular nodes (Figs. 7-4, 7-5,
7-8, and 7-17 to 7-19). Although these distinctions are
generally easily made, distinction between N2 ipsilateral
and N3 contralateral mediastinal nodes in particular may
Figure 7-18 Nodal staging: N3 disease. Non–contrast-enhanced prove problematic, especially when nodes lie directly ante-
section through the lower neck shows an enlarged right supraclavic- rior to the carina (Fig. 7-20). Similar problems arise in
ular node (arrow). Identification of this accessible node facilitates
histologic evaluation. interpreting subcarinal nodes. Although these technically
A B
Figure 7-19 Nodal staging: N3 disease. A: Contrast-enhanced CT section shows a subpleural lesion in the left upper lobe. B: Slightly
lower section through the proximal carina shows evidence of contralateral low right paratracheal (R4) mediastinal nodes. Although this
appearance suggests that this lesion is unresectable, confirmation is always requisite, given the lack of CT specificity.
A B
Figure 7-20 Nodal staging: N2 versus N3 disease. See Color Figure 7-20B. A: Contrast-enhanced CT section through the carina shows
a heterogeneous mass in the posterior segment of the right upper lobe. Enlarged mediastinal nodes are present, including one that lies
immediately anterior to the trachea. In this case, differentiation between N2 and N3 disease is problematic. In practical terms, these lesions
should be assumed to represent N2 disease in the absence of other findings to suggest unresectability. B: Endobronchial map of trans-
bronchial needle-aspiration biopsy sites for mediastinal and hilar lymph nodes as visualized bronchoscopically. By convention, the airways
are oriented from above-downward, with right-sided airways on the right side. It should be noted that although widely accepted, the num-
bering system used in this map is different from that proposed by the American Thoracic Society.
are designated N2 nodes, some consider the finding of sub- pulmonary ligament and adjacent to the esophagus (91). A
carinal nodes adjacent to the medial wall of the contralat- group of lymph nodes located between the left upper lobe
eral mainstem bronchus more properly to be considered bronchus and the left superior segmental bronchus drains
N3 disease (87). both the left upper and left lower lobes. In addition, the
Critical for interpretation of CT studies in patients left upper lobe drains to nodes in the aorticopulmonary
being assessed for lung cancer staging is familiarity with window, the anterior mediastinum, the left paratracheal
the accessibility of nodes with mediastinoscopy as well as area, and the subcarinal nodes. As on the right side, the
transbronchial needle aspiration and biopsy (TBNA) (88) left lower lobe exhibits preferential drainage to the sub-
(Fig. 7-20). As a rule, mediastinoscopy should allow evalu- carinal nodes, paraesophageal nodes, and nodes of the
ation of nodes in the upper paratracheal (2R, 2L), lower pulmonary ligament.
paratracheal (4R, 4L), right tracheobronchial (10R), left As noted previously, it is not unusual for some proximal
peribronchial (10L), and subcarinal (7) nodal stations. nodes to be spared while more distal nodes are involved by
Aortopulmonary window and anterior mediastinal nodes disease. In some series, as many as 30% of patients with
(stations 5 and 6) are typically not considered accessible to lung cancer had “skipped” metastases to mediastinal lymph
the mediastinoscope. nodes with no involvement of lobar or hilar nodes (35,92).
As documented by Riquet et al. (63), in a retrospective
Patterns of Lymphatic Spread study comparing 209 patients with pathologically proved
Optimal assessment of nodal disease is facilitated by N1-negative, N2-positive disease [(N1-)N2] with 522 pa-
awareness of the pattern of tumor spread to regional tients with N1-positive, N2-positive disease (N1)N2,
lymph nodes. These patterns have been recently illustrated showed that patients with (N1-)N2 disease were more likely
(86). Knowledge of the patterns of metastatic spread to to have upper lobe tumors with metastases involving a sin-
regional lymph nodes is valuable when interpreting CT gle nodule station only, resulting in a greater number of
findings (89). Typically, lymph flows into intrapulmonary lobectomies than in patients with (N1)N2 disease.
nodes around segmental bronchi, then to lobar or interlo- Furthermore, these individuals had a better prognosis, with
bar nodes, and then to lymph nodes at the hilar areas. The 5-year survival rates of 34.4% versus 18.5%, respectively
spread of lung cancer generally follows the same pathway. (p 0.00006). Thus it should not be assumed that absence
The lymphatic pathways from the individual lobes to the of hilar nodes indicates the absence of disease in the medi-
mediastinum are constant. astinal nodes. Likewise, it is important to understand differ-
In 1952, Borrie (90) described more precisely the mech- ences in the patterns of tumor spread throughout the
anisms of lymph drainage. In the right lung, a collection of various pulmonary lobes. Although, as a rule, bronchogenic
intrapulmonary lymph nodes that lie between the upper carcinoma tends to spread to ipsilateral mediastinal lymph
lobe bronchus and the superior segmental or middle lobe nodes, significant exceptions occur.
bronchi appeared to be the common drainage pathway for Although right lung cancers initially spread almost exclu-
all three lobes. In addition, the right upper lobe drains to sively to ipsilateral mediastinal nodes, left lung tumors have
nodes in the region of the azygos vein, also called the right a higher propensity for contralateral spread (Fig. 7-21). In
tracheobronchial nodes, contiguous with nodes in the one study, contralateral spread was observed in 3.6% of
lower paratracheal area. The middle and lower lobes patients with right upper lobe tumors, in 3.7% of those with
also drain to the subcarinal nodes and to nodes in the right lower lobe tumors, in 11% of patients with left upper
A B
Figure 7-21 Patterns of spread: N3 disease. A: CT section shows tumor in the left lower lobe in the absence of obvious hilar adenopathy.
B: CT section through the mediastinum in the same patient shows evidence of enlarged contralateral right paratracheal (N3) nodes. These
were subsequently confirmed to be malignant by mediastinoscopy. Left lower lobe tumors in particular have a tendency to involve con-
tralateral right paratracheal nodes.
lobe tumors, and in 25% of patients with left lower lobe short axis used as a cutoff for assessing subcarinal, precari-
tumors (93). Right lung cancers tend to involve mainly sta- nal, and tracheobronchial lymph nodes. By using these cri-
tions 10R and 4R and extend to the higher paratracheal teria, Ikezoe et al. (97) noted improved specificity when
node stations (2R) and finally to the scalene or inferior compared with routine 10-mm short-axis measurements
deep cervical nodes. Tumors in the right lower and middle (94% vs. 77%, respectively), albeit at the cost of some loss
lobe frequently spread to subcarinal (station 7), tracheo- of sensitivity (69% vs. 74%, respectively). Alternative ap-
bronchial (10R), and paratracheal nodal stations (4R and proaches that have been suggested include the use of a
2R). It should be noted that subcarinal nodes are midline minimum difference of 5 mm between adjacent nodal sta-
nodes connected to both right and left lymphatic systems tions before interpreting nodes as enlarged (98). Neither
and from which contralateral spread can occur. of these approaches has gained widespread acceptance.
Contralateral lymphatic spread is more common with As a consequence, the most important role for CT in
tumors originating on the left side, especially those in the patients with enlarged mediastinal and hilar nodes is as a
left lower lobe (Fig. 7-21). Regional lymph node stations tool for determining the optimal method for obtaining his-
4L and 10L are commonly involved. Left upper lobe tumors tologic verification. Currently, in addition to routine trans-
often metastasize to subcarinal (station 7), the aorticopul- cervical mediastinoscopy and standard TBNA, a number of
monary window (station 5), and anterior mediastinal innovative approaches are currently under evaluation, in-
nodes (station 6). The left lower lobe lesions are those most cluding CT-guided transbronchial needle aspiration with
likely to spread contralaterally via the subcarinal node real-time CT fluoroscopy or virtual bronchoscopic guid-
(station 7), to the right paratracheal regions (4R), and to the ance, endoscopic ultrasound-guided bronchoscopy (EUB)
pulmonary ligament and paraesophageal stations. and, most recently, electromagnetic-guided bronchoscopy
(99–102). Choice among these options depends largely on
Computed Tomography Sensitivity, both experience and access to state-of-the-art technology.
Specificity, and Accuracy Although it is generally accepted that enlarged medi-
It is well established that CT is limited in its ability to stage astinal nodes are nonspecific, requiring histologic evalua-
nodal disease accurately, with reported sensitivities, speci- tion for definitive staging, controversy remains over the
ficities, and accuracies ranging between 46% and 87%, need to perform mediastinoscopy when CT is negative. In
69% and 89%, and 65% and 84%, respectively (67). most centers, a normal CT study obviates mediastinoscopy
Identification of malignant mediastinal and hilar nodes on the assumption that malignant nodes will be identified
with CT is limited because of a number of considerations. surgically in only 5% to 15% of cases (103). However, this
The detectability of mediastinal nodes correlates directly is by no means universally accepted. Tahara et al. (104), in
with the amount of mediastinal fat that provides a natural a retrospective evaluation of 29 patients with normal CT
contrasting background. In individuals with a paucity of fat, examinations in whom mediastinoscopy was performed,
it is often difficult to distinguish lymph nodes from medi- found that of 27 patients with documented cancer with
astinal structures, even after administration of intravenous otherwise clinically resectable disease, 11% had positive
contrast media. Another problem is related to the partial- nodes at mediastinoscopy, obviating surgical resection.
volume effect in anatomic areas oriented obliquely relative In this same study, in an additional 10% of cases, metasta-
to the plane of scanning. Comparison of lymph node size tic nodes were identified at surgery. A nearly identical
and number in cadavers imaged by CT and followed percentage of unsuspected metastatic nodes identified at
by lymph node dissection has shown good correlation mediastinoscopy was previously reported by Seely et al.
between CT and pathology for right-sided mediastinal (105), who also identified nodal metastases in 21% of
lymph nodes (94,95). However, CT has been shown to be 104 patients with otherwise resectable T1 lesions. As is
less accurate on the left side, failing to demonstrate sev- discussed later, reliance on the use of CT as a method for
eral normal nodes in the left tracheobronchial region (10L), determining the need for histologic sampling of medi-
aorticopulmonary window (station 5), and subcarinal astinal nodes has been lessened by the introduction of
region (station 7) (95,96). Therefore, a risk exists of under- FDG-PET scanning.
staging lymph nodes on the left side of the mediastinum.
The most important limitation of CT, however, is that it
is restricted to morphologic evaluation only. Lymph nodes
are typically considered enlarged when they measure
of Metastatic Disease
greater than 10 mm in short axis. However, nodal enlarge- The most important contribution of CT to the evaluation of
ment, per se, is of little value in distinguishing between distant metastases is identification of adrenal lesions, as
benign hyperplastic and malignant nodes, necessitating these are easily incorporated into every routine CT examina-
histologic verification in all cases for which accurate stag- tion of the thorax (Figs. 7-6 and 7-22 to 7-24). The adrenal
ing is required. In an attempt to improve the accuracy glands are a common site of pathology, with lesions esti-
of CT, it has been suggested that different size criteria be mated to be present in nearly 9% of the general population
applied to different nodal stations, with 13 mm in the and identified on approximately 5% of routine abdominal
A B
Figure 7-22 Benign adrenal disease: CT evaluation. A: Magnified unenhanced CT section shows a well-defined right-sided adrenal lesion
measuring less than 10 Hounsfield units (HU). B: Magnified CT section from a different patient also shows a well-defined low-density adrenal
lesion measuring less than 10 HU after intravenous contrast administration. In both cases, the findings of either 10 HU on unenhanced
sections or lack of enhancement after contrast administration are reliable signs of benign lipid-filled adrenal adenomas for which histologic
correlation is unnecessary.
C
Figure 7-23 Benign adrenal disease: CT–FDG-PET correlations.
See Color Figure 7-23C. A: CT section shows a 3.9-cm mixed solid
ground-glass lesion in the left upper lobe, subsequently diagnosed as
an adenocarcinoma. B: Contrast-enhanced section through the upper
abdomen in the same patient as shown in A demonstrates slightly
enlarged nodular adrenals bilaterally. This appearance is nonspecific,
requiring additional confirmation with either follow-up MR or prefer-
entially FDG-PET. C: Coronal images from a corresponding FDG-PET
scan show intense activity in the left upper lobe tumor, but no evi-
dence of increased activity in either adrenal. This combination of CT
with PET findings is consistent with benign bilateral adrenal
hyperplasia. Incidentally noted, however, are multiple osseous metas-
B tases (arrows), making this a stage IV unresectable lesion.
be reliably obtained in most cases based solely on CT. This

is because a fair proportion of cases of nodular adrenal
enlargement will be found to represent benign adrenal
adenomas or adrenal hyperplasia (107).
Differentiation of adrenal adenomas from metastatic
bronchogenic carcinoma on CT is facilitated by measure-
ment of attenuation values in scans performed either
before injection (108,109) or after delayed imaging
(Figs. 7-22 and 7-24) (110). As documented by Blake
et al. (111), in a study of 122 adrenal masses including
17 malignant lesions, precontrast images proved most use-
ful, with lesions measuring less than 10 HU indicative of
benign lesions and lesions measuring greater than 43 HU
indicative of malignant lesions. In this same study, incor-
porating baseline non–contrast-enhanced measurements
with an assessment of absolute percentage washout of le-
Figure 7-24 Adrenal metastases: CT evaluation. Contrast- sions (defined as the initial density measured 75 seconds
enhanced CT section shows an enlarged heterogeneous mass in after contrast administration minus the lesion density
the left adrenal gland, with evidence of rim enhancement. Note measured at 10 minutes divided by the initial density
that the right adrenal gland is normal in appearance. This appear-
ance is consistent with an adrenal metastasis (compare with
minus the precontrast density determined following
Fig. 7-22B). a bolus of 120 mm of nonionic contrast media at a rate of
2.5 mm/sec), CT proved 100% sensitive for detecting
CT examinations (106,107). Although a wide range of malignant adrenal masses, with a specificity of 98% (111).
pathologic entities may result in adrenal masses, including, The use of delayed contrast-enhanced imaging is recom-
among others, pheochromoctyomas, cortical carcinomas, mended, as approximately 30% of adenomas prove to be
neuroblastomas, lymphomas, and myelolipomas, as well as lipid poor and are thus difficult to characterize on unen-
simple cysts, in the setting of known or suspected lung hanced scans only. As shown by Macari et al. (109), initial
cancer, the major differential remains between incidental noncontrast scans through the adrenals are easily obtained
adrenal adenomas and adrenal metastases. Although histo- in those cases in which a heightened suspicion exists for
logic confirmation of adrenal lesions may be required, a the presence of disease. This approach and others using
confident diagnosis of a benign adrenal adenoma may now both MR and FDG-PET are summarized in Figure 7-25.
Adrenal Staging
CT/MR Evaluation
< 10 HUs (+) Contrast CT Cyst
Lipid rich
adenoma
N Y
Lipid poor
adenoma
Y
1 2 3 Y
C + < 60%
washout
Relative % Intracellular Pt. return:

washout < 60% fat on C - study
N
of + contrast CT chemical N HU<10? N
or < 30 HU at shift MR
10 min
N biopsy
Y Y Y
Figure 7-25 Proposed algorithm for evaluating adrenal lesions. With this approach, lesions less than 10 HU on unenhanced CT sections
can be considered diagnostic of lipid-rich adrenal adenomas. For those lesions that enhance after contrast administration, any of the follow-
ing steps may be taken: either follow-up CT images showing less than 60% washout compared with the initial CT image or less than 30 HU
on 10-minute delayed images; chemical-shift MR images; follow-up repeated unenhanced CT; or FDG-PET imaging. (Courtesy of Jane Ko,
MD, New York, New York.)
A B
Despite the ability of CT to detect distant metastases, its

use as a routine method for evaluating extrathoracic
disease in asymptomatic patients has been questioned
(112). Even before the widespread use of FDG-PET, the
utility of routine imaging tests including CT for evaluating
distant metastases has been questioned. In one study of
755 patients with clinical T1–2 N0 NSCLC evaluated with
CT scans of the abdomen and pelvis, radionuclide bone
scans, and either CT or MR brain scans, whereas a total of
27 (4%) of 754 patients proved to have distant metastases,
occult disease was identified in only 0.5% of T1N0 cases
and 0.9% of T2N0 cases, respectively (113).
Despite these data, still no consensus has been reached
on the need to acquire presurgical MR or CT studies espe-
cially of the brain. This is because FDG-PET has proved
of little use for identifying brain metastases because of
normal uptake of the tracer by the brain. In a more recent
report than cited earlier using CT to evaluate 108 con-
secutive neurologically intact patients with potentially
resectable lung cancers, brain metastases were identified in
4.8% of cases (114).
C
Figure 7-26 Follow-up CT evaluation: local recurrence. A: Post-Therapeutic Computed
Magnified unenhanced CT section 3 months after right upper lobe Tomography Evaluation
resection and mediastinal lymph node dissection. A line of surgical
clips is present anterior to the right interlobar pulmonary artery. B: At present, follow-up CT imaging is obtained in all
Follow-up contrast-enhanced magnified CT image at the same
level as shown in A, approximately 12 months later. Note the soft patients undergoing chemotherapy or radiotherapy, as
tissue density lateral to the surgical clips, not present on the initial well as in patients after surgery (Figs. 7-26 and 7-27). No
postoperative study. C: CT-guided transthoracic needle biopsy compelling data support the value of postoperative screen-
confirmed the presence of locally recurrent tumor. Although CT is
able to identify locally recurrent disease before the development ing to detect either local or distant recurrences (115). As
of symptoms, no evidence indicates that this prolongs survival. noted by Walsh et al. (115), in a study of 358 consecutive
In patients being treated with radiotherapy, the role of CT

initially is to provide the best possible definition of tumor
margins to assist in treatment planning. Although postradia-
tion changes have been well defined on CT, no current indi-
cations exist for its use to monitor response to radiotherapy.
This is because distinguishing postradiation change from
locally recurrent tumor is especially problematic with CT,
even after a bolus of intravenous contrast media.
For patients being treated with chemotherapy, the major
indication for CT is also to monitor the response to therapy.
In addition to identifying new or enlarging mediastinal or
hilar nodes, or identifying new parenchymal lesions, most
often this involves the use of either WHO bidimensional
measurements or RECIST unidimensional measurements to
evaluate typically three to five previously identified and
annotated nodules or masses (24). By definition, the diam-
eter of measurable lesions must be greater than twice the
Figure 7-27 Follow-up CT evaluation: evaluation of the post- slice thickness to be accurately assessed. The strengths and
pneumonectomy space. Contrast-enhanced CT section at the level
of the aorticopulmonary window shows evidence of local disease limitations of these are reviewed in depth in Chapter 6 and
recurrence after left-sided pneumonectomy (arrow). illustrated in Figure 6-49. Unfortunately, considerable inter-
and intraobserver variability has been reported, even among
dedicated chest radiologists, especially when lesions are
small or poorly defined, even by using electronic calipers
patients who had undergone complete resections for (118,119). It should also be emphasized that it is not un-
NSCLCs, recurrences occurred in 135 individuals (local common for some lesions to enlarge while others shrink, re-
and/or local and distant in 45, distant only in 90), only 33 inforcing the concept that CT is limited to anatomic and not
of whom were asymptomatic. By using multivariate analy- physiologic assessment of disease.
sis, these authors showed that whereas detection of asymp-
tomatic recurrences led to earlier diagnosis by as long as
Magnetic Resonance Imaging Evaluation
8 to 10 months, resulting choice of treatment and overall
survival were not affected, leading these investigators to MRI offers inherent advantages that may be of use in lung
doubt the utility of routine CT surveillance in previously cancer staging. First, unlike CT, MRI does not require the use
treated asymptomatic patients. Similar findings have been of ionizing radiation. In addition, MRI demonstrates the
reported more recently for patients with tumors deemed relationship between tumor and major mediastinal vascular
unresectable but treated with curative intent with structures without the need for administration of intra-
chemotherapy and/or radiotherapy (116). venous contrast media. Third, the increased soft tissue dif-
These caveats notwithstanding, follow-up CT studies are ferentiation achievable with MRI compared with CT may be
now routinely acquired in most centers to monitor disease of value in selected cases by differentiating tumor from adja-
progression and/or the presence of local recurrences after cent structures such as the chest wall. Similarly, MRI can
surgery. In distinction, CT is only rarely used to detect help by differentiating obstructing carcinoma from sur-
asymptomatic extrathoracic recurrences, with the excep- rounding atelectatic lung, especially on T2-weighted or
tion of adrenal lesions identifiable on all standard thoracic gadolinium-enhanced images, as consolidated lung typi-
CT examinations. cally exhibits a higher signal intensity than does the central
In addition to recurrent disease, CT also has been hilar mass (120–122). Although initially cited as an impor-
suggested as a method for detecting second primary tant advantage, acquisition of high-quality sagittal and coro-
lesions. The risk of a second metachronous primary has nal images is no longer an advantage, given the widespread
been estimated to occur in 1% to 2% of patients per year availability of multidetector CT (MDCT) scanners providing
(117). Although CT is capable of detecting metachronous equally high-quality off-axis reconstructions.
lesions, the value of early detection of second primaries is MRI also exhibits significant disadvantages relative to CT.
unclear. In the study by Walsh et al. (115), for example, It is costly and remains less accessible. More important,
whereas new primary tumors were identified in 35 of 358 spatial resolution is less than that achievable with CT. As
postoperative patients, no definite survival advantage a consequence, CT is superior to MRI in determining the
could be documented by early diagnosis. The value of CT presence or extent of endobronchial tumor, for example,
for detecting new primaries approximates that expected or involvement of central airways or the presence of
for low-dose CT screening and consequently remains tumor-traversing fissures (120). Artifacts and blurring of
controversial. structures related to respiratory and cardiac motion are also
a potential problem, although this has been greatly reduced ence, we recommend performing MRI of the brain in these
by the advent of motion-compensation techniques and the patients, even when they are asymptomatic.
systematic use of cardiac gating. Perhaps most important, MR is also a reliable method for differentiating among
MR is insensitive to the presence of calcium, rendering iden- the various causes of adrenal masses (Figs. 7-29 and 7-30).
tification of calcified nodes less accurate than with CT. As a Most helpful diagnostically is the use of chemical-shift
consequence, in the vast majority of cases, CT remains the imaging with in-phase and out-of-phase spoiled gradient-
imaging modality of choice for assessing the presence of recalled echo (GRE) sequences to differentiate between
intrathoracic tumor (103). adrenal adenomas and malignant adrenal masses (124).
Characteristically, tumors show increased signal inten- Adrenal adenomas contain intracellular fat and are there-
sity on both T1- and T2-weighted sequences and typically fore recognizable due to signal loss on out-of-phase
enhance after intravenous administration of gadolinium images, especially when this measures more than 20%
(Fig. 7-28). As discussed in Chapter 6, the use of MR for (125). They are also typically smaller than 3 cm. In distinc-
evaluating peripheral lung lesions, although currently tion, adrenal metastases demonstrate no change in sig-
under evaluation, is limited. Currently, the strongest indi- nal intensity on out-of-phase images. Additionally, they
cations for the use of MR include imaging brain metastases, typically demonstrate low signal intensity on T1-weighted
evaluating patients with Pancoast tumors, and evaluating images and high signal intensity on T2-weighted sequences
the adrenal glands (Figs. 7-29 through 7-32). and show progressive enhancement after administration of
Although both CT and MRI have been advocated for intravenous contrast material (124).
detecting brain metastases, in most series, gadolinium- In selected cases, chest-wall invasion may be better
enhanced MRI has proved far superior to CT in the detec- demonstrated by MRI than by CT (Figs. 7-31 and 7-32).
tion of cerebral metastases (Fig. 7-6) (123). It is likely that This is because MRI provides better tissue contrast, particu-
the prevalence of asymptomatic brain metastases in larly on T2-weighted images, between tumor, chest wall,
asymptomatic patients is higher than the approximately fat, and muscles (126–128). A thin layer of extrapleural
5% suggested on meta-analysis of CT data (112). Because fat separating the tumor mass from the chest wall can
the prevalence of metastases is higher in patients with almost always be seen on high-quality MR images. This
T3 lesions and those with adenocarcinoma, in our experi- thin layer of extrapleural fat, located beneath the parietal
A B
Figure 7-28 Malignant lung mass and nodule: MR assessment. A: Breath-hold 5-mm fat-suppressed T2-weighted sequence shows a large
soft tissue mass in the right upper lobe with heterogeneous, intermediate signal intensity (thick arrow) differentiated from adjacent
parenchymal atelectasis (thin arrow) by the higher T2 signal of the latter, secondary to retained secretions. Foci of higher T2 signal in the
tumor may reflect areas of liquefaction or necrosis. A second smaller nodular lesion is noted in the left lower lobe (curved arrow) with
slightly higher signal intensity noted within. B: A 1-mm section through the same plane as A, with a fat-suppressed volumetric T1 spoiled
gradient-echo sequence (VIBE). The mass and nodule demonstrate homogeneously low signal intensity. Benign lung nodules and masses
cannot reliably be differentiated by their T1 and T2 signal characteristics alone, which are often nonspecific, usually low to intermediate on
T1-weighted sequences and intermediate to high on T2-weighted sequences. C–E: The 5-mm axial reconstructions from three consecutive
20-second breathhold VIBE sequences (T1 fat saturated) obtained in the first 80 seconds after IV injection of gadolinium at approximately
the same level as in A and B. These images demonstrate early enhancement of the compressed vascular parenchyma and progressive
enhancement of the tumor mass. This differential enhancement may be very useful in delineating the presence and size of underlying masses
from adjacent underlying parenchyma (see Fig. 9-143, as well). Note that in this case, considerably greater enhancement appears in the
smaller lesion. Whereas the absence of enhancement strongly favors a benign etiology similar to lack of contrast enhancement on CT,
patterns of enhancement with MR are nonspecific. In this case, differential enhancement in the smaller lesion is compatible with both
metastatic disease and a second primary carcinoma (continued ).
C D
E
A B
Figure 7-29 Adrenal adenoma: MR evaluation. A, B: Magnified in-phase (A) and out-of-phase (B) images show marked decrease in signal
intensity on the out-of-phase image, a finding consistent with lipid-rich adenoma (arrows in A and B). MR, CT, and PET are comparable in
their ability to differentiate between benign and malignant adrenal lesions (see Figs. 7-23 and 7-25).
A B
Figure 7-30 Adrenal metastasis: MR evaluation. A, B: T1- and T2-weighted MR images, respectively, at the same level as in A. The right
adrenal mass appears moderately hyperintense on the T2-weighted image relative to the liver, with slight heterogeneity, features consistent
with metastatic disease (histologically confirmed adrenal metastases).
A B
C D
Figure 7-31 Chest-wall invasion: MR evaluation. Pancoast tumor. A, B: Sequential MR images through the thoracic inlet from below-
upward show a lobular mass in the left lung apex (arrow in A). Note that the medial border of the tumor is well delineated by fat. A large
supraclavicular lymph node is present on the left, as well, just posterior to the left jugular vein and lateral to the left carotid artery (arrow in
B). C, D: Coronal and sagittal images through the lung apex, respectively, show sharp demarcation between the tumor and adjacent medi-
astinal and extrapleural fat along most of the course of the tumor–pleural interface (arrows). No contact is seen with the spine. Although
these findings suggested a lack of gross chest-wall invasion, at surgery, microscopic invasion was documented with minimal tumor extension
into the extrapleural fat. The supraclavicular node shown in A proved to be positive.
A B
C
Figure 7-32 Chest-wall and mediastinal invasion: MR evaluation. A, B: T1- and T2-weighted axial MR images through the left lung apex
show a poorly marginated tumor mass extending laterally into the chest wall (arrows). Note that medially, the mediastinal fat is well pre-
served. C: Coronal MR image again shows that the mass is invading the chest wall laterally. Although the mass clearly parallels a consider-
able portion of the mediastinum, the mediastinal fat remains intact (arrows). At surgery, chest-wall invasion was confirmed; the mediastinum
proved to be normal.
pleura, may be effaced in the presence of early inva- Services (CMS) for the staging and restaging of NSCLC. In
sion (Fig. 7-33). With conventional CT, this fat layer is brief, FDG is a D-glucose analogue radiopharmaceutical
more difficult to see (129–132). Padovani et al. (128), in labeled with a positron emitter (18F) that is facilitatively
34 patients’ MRIs, reported a sensitivity of 90% and a transported through the cell membrane and phospho-
specificity of 86% in the detection of chest-wall invasion. rylated by using normal glycolytic pathways, especially by
To date, MR has proved most useful for evaluating superior primitive fermenting cells with a high anaerobic/aerobic
sulcus tumors, where improved tissue differentiation is of metabolic ratio (142). This includes neoplastic cells, irre-
greatest need for identifying the brachial plexus (133,134). spective of cell type, as well as inflammatory cells. Failure
of further metabolism results in intracellular trapping of
the tracer. (18F)FDG normally accumulates in a number of
Positron Emission Tomography
sites, including especially the brain, heart, renal collecting
(FDG-PET) Evaluation
system, and bladder. Tracer also accumulates in so-called
“brown fat,” potentially mimicking the appearance of
Initial Staging of Non–Small Cell Lung Cancer
mediastinal adenopathy. Moderate activity is also com-
Since its introduction in the early 1990s, PET with monly noted in the liver, spleen, bone marrow, breast, and
2-(fluorine-18)-fluoro-2-deoxy-D-glucose (FDG–PET) has gastrointestinal tract. Within the thorax, (18F)FDG-PET
become an essential component for lung cancer staging also accumulates in both residual thymic tissue as well as
(135–141). This is especially true after approval for reim- in thyroid tissue, including nodular goiters, sites easily
bursement by the Centers for Medicare and Medicaid mistaken for mediastinal metastases (143). It should also
A B
Figure 7-33 Stage IIIA disease: ipsilateral mediastinal adenopathy/chest-wall invasion. A: T1-weighted MRI scan shows a right hilar mass
with apparent extension into the mediastinum, causing compression of the superior vena cava (straight arrow). Laterally, a suggestion exists
that tumor has infiltrated extrapleural fat, identifiable as focal areas of decreased signal intensity (curved arrows). B: T2-weighted MRI scan,
at a level slightly higher than in A. Tumor clearly extends into the chest wall (arrows). At surgery, the vena cava was found not to be invaded.
Hilar and mediastinal nodes proved positive, without evidence of mediastinal invasion. The tumor was successfully resected despite chest-
wall invasion.
be noted that 18FDG-PET may be of limited value in dia- standard imaging techniques proved to have futile thora-
betics; this is because elevations of blood glucose (or cotomies compared with only 21% in the group addition-
insulin levels) lead to increased accumulation of FDG in ally evaluated with FDG-PET, irrespective of clinical stage
muscle and decreased accumulation of uptake within (147). This included 8% with distant metastatic disease
malignant tissue (144). documented solely by PET.
In addition to its role in evaluating pulmonary nodules Reed et al. (146), in a multi-institutional study of 303
(see discussion in Chapter 6), to date, numerous reports patients with otherwise resectable disease as determined
similarly confirm that FDG-PET is of value for staging by routine staging techniques, also documented the superi-
lung cancer, especially when compared with CT (145–148). ority of FDG-PET compared with CT for identifying both
Pieterman et al. (145), for example, in a widely cited pro- N1 disease (42% vs. 13%; p 0.177) and N2–3 disease
spective study comparing a standard approach to lung (58% vs. 32%; p 0.004), respectively, with a negative pre-
cancer staging (including CT, ultrasonography, and bone dictive value of FDG-PET for mediastinal nodal disease of
scanning) with another including FDG-PET in 102 patients, 87%. In this same study, unsuspected distant metastases
reported a sensitivity and specificity of FDG-PET for detect- and/or second primary malignancies were also identified in
ing mediastinal nodes of 91% (95% CI, 81% to 100%) and 18 (6.3%) of 287 patients, leading to the avoidance of
86% (95% CI, 78% to 94%), respectively, compared with unnecessary thoracotomies in 20% of patients (146). It is
75% (95% CI, 60% to 90%) and 66% (95% CI, 55% to worth noting that not all studies to date have concluded
77%) for CT. As important, in this same study, FDG-PET that PET leads to a significant decrease in the number of
identified otherwise unsuspected distant metastases in 11 of thoracotomies performed when compared with standard
102 patients. Combining mediastinal nodal evaluation with staging techniques. In one randomized controlled trial
assessment of distant metastases, overall, FDG-PET had a of 1,854 patients with stage I or II NSCLC, whereas PET
sensitivity of 95% (95% CI, 88% to 100%) and a specificity allowed more appropriate stage-specific therapy in patients
of 83% (95% CI, 74% to 92%), leading to an alteration of with stage I and II disease, no statistical difference was
staging in 62 patients including upstaging in 42 when com- noted in the number of thoracotomies performed between
pared with standard staging techniques (145). the two groups (149).
Similar findings have been reported from the PLUS The impact of FDG-PET on clinical practice is most
(PET in Lung Cancer Staging) multicenter randomized apparent when assessed from the perspective of resul-
trial (147). In this study of 188 patients from nine centers ting alterations in patient management. In one study of
who had tumors deemed clinically resectable, patients 198 patients, including 37 patients specifically evaluated
were randomized either for routine evaluation or routine for staging, overall, PET proved to have a significant impact
evaluation plus FDG-PET. For purposes of analysis, the on patient management, affecting both diagnostic and
primary outcome measure for this study was futile tho- therapeutic decisions alike in more than 70% of cases,
racotomy, defined as either surgical evidence of benign including altering stage in 22.2% and providing otherwise
disease, pathologic stage IIIA-N2 or IIIB disease, or unexpected diagnoses in another 4% (148).
postoperative recurrence or death within 12 months of In support of these data, a number of meta-analyses
randomization. Overall, 41% of patients evaluated with have similarly supported the benefit of FDG-PET for
staging NSCLC. Gould et al. (150), for example, in a meta- same patients were accurately staged in 83% of cases
analysis of more than 1,400 published cases, reported a while overstaging 10% and understaging 7%. Interestingly,
mean sensitivity and specificity of 96% and 73.5%, respec- in this same study, PET-CT proved superior to the use of
tively, for the use of PET to stage lung cancer. Dwamena fusion software to co-register separately acquired PET and
et al. (151,152), in a meta-analysis pooling the results of CT images. Comparing integrated PET-CT imaging with
FDG-PET in 14 studies including 514 patients and/or CT in separately co-registered CT and PET images in 25 patients,
another 29 studies including 2,226 patients, also docu- integrated PET-CT overall proved statistically more accurate
mented a significant advantage of FDG-PET for detecting in staging patients (p 0.05), with attempts to perform
nodal metastases, with a mean sensitivity and specificity of co-registration proving unsuccessful in 32% of cases (158).
79% and 91% for FDG-PET versus 60% and 77% for CT, Clearly, the introduction of integrated PET-CT studies repre-
respectively (p 0.001) (151,152). sents an important technical improvement that will likely
More recently, in a meta-analysis of 23 studies using replace dedicated PET scanning over the next several years.
summary receiver operating characteristic curves, similar, Less well established is the use of FDG-PET to evaluate
but somewhat less compelling, results have been reported pleural and/or pericardial effusions. A number of reports
by Birim et al. (153), although still clearly superior to CT. have shown that FDG-PET is more sensitive than CT for
As assessed in this study, the sensitivity of FDG-PET for differentiating benign from malignant pleural effusions,
detecting mediastinal nodes ranged from 66% to 100%, with malignant effusions showing increased PET activity
with an overall sensitivity of 83% (95% CI, 77% to 87%), (60–62). In one retrospective study in which CT and FDG-
and corresponding specificities ranging from 81% to PET first alone and then in combination were evaluated in
100%, with an overall specificity of 92% (95% CI, 89% to 92 patients with newly diagnosed NSCLC with pleural
95%). In comparison, overall sensitivity and specificity of abnormalities, FDG-PET proved superior to CT, with a com-
CT proved to be 59% (95% CI, 50% to 67%) and 78% bined sensitivity, specificity, negative predictive value (NPV),
(95% CI, 70% to 84%), respectively (153). positive predictive value (PPV), and accuracy of 100%,
It is worth emphasizing that these data do not include 76%, 67%, 100%, and 84%, respectively. Similar results
an assessment of the role of recently introduced integrated have also been reported in a prospective study of 98 con-
PET-CT scanners compared with dedicated PET scanners secutive patients with pleural abnormalities, with FDG-
for staging lung cancer. Logically, the two imaging modali- PET increased activity in 61 (96.8%) of 63 patients with
ties should be considered complementary, with CT provid- documented malignant effusions and absent activity in 31
ing greater anatomic detail and with PET providing physi- (88.5%) of 35 patients with benign pleural disease (62).
ologic assessment (154,155). FDG-PET, for example, is of Although taken together, these data support the use of
little value for assessing local tumor extent (T disease), FDG-PET for routine staging of NSCLC, it should be
especially for evaluating invasion of contiguous structures emphasized that not all patients require FDG-PET scans
such as the chest wall, bronchi, or mediastinal structures for accurate staging (Fig. 7-34) (116). In particular, cases
(156). Additionally, as shown by Cerfolio et al. (154), in a in which clear CT evidence of mediastinal invasion with
prospective blinded study comparing dedicated PET scans encasement of mediastinal organs is found, for example,
with integrated PET-CT studies in the same 129 patients, or evidence of direct or metastatic bony metastases or CNS
integrated PET-CT proved statistically superior for both metastases, do not require the addition of FDG-PET for
stage I disease (52% vs. 33%; p 0.03) and stage II disease assessing additional diagnostic or therapeutic options.
(70% vs. 36%; p 0.04), respectively. Integrated PET-CT Given currently established advantages and limitations
also proved superior for assessing both T and N status, for the use of FDG-PET as an initial method for staging
proving more sensitive and specific with a higher positive lung cancer, in our judgment, the following guidelines are
predictive value for both N1 and N2 disease (p 0.05), in suggested (Table 7-6):
particular, for 4R, 5, 7, 10L, and 11 nodal stations. Overall,
PET-CT proved to have an accuracy of 96% and 90% for as- 1. FDG-PET is superior to both CT and MRI for evaluating
sessing N1 and N2 nodes, respectively. nodal disease and, with the exception of brain metas-
Similar data favorably comparing integrated PET-CT with tases, distant metastatic disease (Fig. 7-35). Its utility is
dedicated PET scanning have been reported by others less clear in patients with unequivocal evidence of unre-
(157,158). In one study comparing integrated PET-CT with sectable disease, as established either clinically or with
either modality alone, PET-CT proved more accurate in 27% alternate imaging tests.
of cases, positively affecting management in 12.5% (157). 2. Given lack of specificity, the finding of positive nodes or
Halpern et al. (158), in another study of 36 patients with evidence of extrathoracic disease on FDG-PET still neces-
NSCLC, compared dedicated PET studies with integrated sitates either confirmation with additional imaging tests
PET-CT examinations and found that PET-CT outperformed or preferably histologic confirmation, whenever possi-
dedicated PET scanning, with dedicated PET accurately ble (Fig. 7-36). In this regard, FDG-PET may be of con-
staging 57% of cases while overstaging 20% and understag- siderable value by identifying optimal sites for definitive
ing 23%, compared with integrated PET-CT, in which these histologic staging.
A B
Figure 7-34 CT–PET correlations: stage II lung cancer. See Color Figure 7-34B. A: Contrast-enhanced CT section shows a larger (3 cm)
lesion abutting the posterior pleura and the posterior mediastinum without evidence of enlarged nodes or clear evidence of pleural, chest-
wall, or mediastinal invasion. No effusion was present (not shown). B: Fused PET–CT axial image shows increased activity localized to the
right upper lobe lesion, shown in A. Again note lack of obvious pleura, chest-wall, or mediastinal invasion. Even with fused CT–PET imaging,
evaluation of tumor extent frequently remains problematic.
A B
Figure 7-35 CT–PET correlations: stage IV lung cancer. See Color Figure 7-35B. A: Unenhanced CT section obtained as part of a com-
bined PET–CT study shows enlarged high right paratracheal (2R) nodes medial to the brachiocephalic artery, causing obscuration of the
mediastinal fat on the right side in a patient with a right upper lobe tumor (not shown). B: Corresponding fused PET–CT image shows focal
increased activity both in the right paratracheal nodes as well as in the anterior chest wall posterior to the pectoralis muscles, identified only
in retrospect on the CT section. Excisional biopsy of this node confirmed stage IV non–small cell lung cancer. One of the most important
advantages of FDG-PET versus CT is the ability to identify more accurately diffuse metastatic disease (compare with Figs. 6-23B and 6-44).
A B
Figure 7-36 CT–PET correlations: stage IIIB lung cancer. See Color Figure 7-36B. A: Unenhanced CT section obtained as part of a
combined PET–CT study shows enlarged precarinal (4R and 4L) nodes. A small quantity of pleural fluid can be identified on the right side.
B: Corresponding fused PET–CT axial image shows markedly increased activity in the same nodes. Despite both CT- and PET-positive
studies, histologic correlation is still necessary for definitive diagnosis. Mediastinoscopy confirmed stage IIIB lung cancer.
3. A negative FDG-PET scan coupled with a negative

Additional Indications for PET Evaluation
CT scan may be sufficiently specific to obviate medi-
astinoscopy, although this remains controversial Additional potential indications for which FDG-PET is
(Fig. 7-37). currently being evaluated besides initial staging of NSCLC
4. Although the use of FDG-PET for differentiating between include predicting prognosis, monitoring response to treat-
benign and malignant effusions is promising, analogous ment, and determining the need and appropriate timing for
to the significance of PET-positive nodes, definitive eval- restaging as well as detecting recurrences in asymptomatic
uation still requires cytologic and/or histologic confirma- patients (24,159–161). It is worth noting that FDG-PET
tion whenever possible. is currently approved for restaging lymphoma, melanoma,
A B
C D
E F
Figure 7-37 CT–PET correlations: stage IA lung cancer. See Color Figure 7-37B. A: CT section imaged with lung windows shows an irreg-
ular, slightly spiculated subpleural lesion (arrow) in the periphery of the right upper lobe. B: Fused PET–CT coronal image shows increased
activity localized to the right upper lobe lesion shown in A (arrow), without evidence of nodal or distal metastases. In our experience, the
combination of negative CT and PET is sufficiently specific to warrant obviating mediastinoscopy. Surgically proved T1N0M0 non–small cell
lung cancer. C–F: Selected contrast-enhanced CT images in the same patient show no evidence of mediastinal, hilar, or adrenal lesions.
and cervical cancer, as well as assessing response to therapy tumor were 43% and 100%, respectively, although the
in patients with breast cancer (160). PPV of FDG-PET for assessing N2 disease, specifically,
proved to be less than 20% (167).
Assessing Prognosis at Diagnosis, The role of PET for monitoring response to treatment
and Response to Initial Therapy has also been evaluated in the expectation that PET imag-
A number of reports to date have suggested a role for ing might allow tailoring of treatment by identifying
FDG-PET in assessing prognosis both at the time of diag- early responders for whom continued therapy would be
nosis and after induction of chemotherapy (162–165). indicated (161). As reported by Weber et al. (163) in a
Ideally, FDG-PET would provide information allowing study of 57 patients with stage III and IV NSCLC evalu-
therapy to be appropriately tailored to individual pa- ated before and after a single course of platinum-based
tients, providing an incentive to continue therapy in therapy, evidence of a 20% reduction in SUV correlated
responders and discontinue therapy early in its course with subsequent response as measured on CT by using
among nonresponders. RECIST criteria with a sensitivity of 96% and a speci-
Cerfolio et al. (166), for example, in an assessment of ficity of 84%. Although suggestive, these data remain
PET as an independent predictor of the biologic aggressive- preliminary.
ness of pulmonary nodules, correlated standard uptake
values (SUVs) in 315 patients with the degree of tumor Post-treatment Evaluation and Restaging
differentiation, as well as the likelihood of complete surgi- Similar to data suggesting a role for PET in predicting
cal resection, and reported that the maximum SUV proved prognosis to initial therapy, it has also been suggested
a reliable predictor of both disease-free survival and that FDG-PET is an accurate means for predicting survival
survival. Specifically, when patients with SUVs above ver- (161). Several studies have shown that patients with
sus below the median SUVs assessed for individual stage either NSCLC (168) or SCLC (169) who have positive
groups as a whole were separately compared, the actual PET scans after therapy have a worse prognosis than do
4-year survival was 80% versus 66% for stage IB disease those with negative studies. In one study of 113 patients
(p 0.48), 64% versus 32% for stage II disease (p 0.28), with NSCLC, for example, median survival for those
and 64% versus 16% for stage IIIA disease (p 0.12), with a positive PET scan after therapy was 12 months
respectively. compared with those with negative studies (11 of 13
In a related study, Port et al. (167), evaluated the use of patients), in whom a median of 34 months of follow-up
PET for predicting the response to induction che- was recorded (168). Data supporting the potential for
motherapy in 25 patients before and after neoadjuvant FDG-PET to identify local recurrence also have been
therapy. Defining a significant response as a 50% or reported (Fig. 7-38) (170), including for patients in
greater decrease in the SUV, PET results were correlated whom newly identified nodules or masses are identified
with pathologic response, considered major if no disease by CT as well as for evaluation of patients after radiation
or only microscopic disease remained in the primary therapy. Of particular importance is choosing appropri-
tumor. With these definitions, the PPV and NPV of FDG- ate time intervals for reevaluation. It is current practice
PET for predicting response to therapy in the primary that restaging be performed 1 to 2 months after surgery
A B
Figure 7-38 CT–PET correlations: tumor restaging. See Color Figure 7-38B. A: Contrast-enhanced CT section obtained as part of a
combined PET–CT study in a patient after left upper lobe resection shows no evidence of local recurrence. B: Fused PET–CT image at the
same level as A shows markedly increased activity in an otherwise normal-sized left hilar node. Subsequent follow-up examinations con-
firmed local disease recurrence. Although FDG-PET scans are more sensitive than CT for detecting both local and distant recurrence, the
clinical value of early detection of tumor recurrence remains to be established.
and 2 to 6 months after completion of chemo- and/or refers to any tumor burden more than limited disease with
radiotherapy. As recently summarized, the main contri- obvious metastatic disease and is present in 60% to 80%
bution of PET in this setting is differentiation between of patients with newly diagnosed SCLCs.
persistent or recurrent tumor and residual fibrosis or
scarring, especially after radiation therapy (160). Follow-
up PET scans may also be of value by identifying optimal Therapeutic Considerations
sites for biopsy, analogous to its role in initial staging.
Untreated patients with small cell carcinoma have a
Despite these data, a number of issues remain to be
median survival of only 6 to 17 weeks, but even with
resolved regarding potential uses of FDG-PET for
treatment, the 5-year survival is usually less than 10%
post-treatment evaluation. In addition to the need for
(174). Limited-stage disease is typically treated with
standardization of technique and timing, additional
combined chemotherapy and radiation therapy, with
prospective, ideally, randomized studies are still required
a median survival time of approximately 18 months,
to establish a definitive role for PET for predicting prog-
with cure occurring in approximately 20% of cases.
nosis and as a means for follow-up evaluation in treated
Recent evidence suggests that improved 5-year survival
patients (160,161).
rates are obtained with concurrent chemotherapy and
Given presently available data, the following guidelines
radiotherapy (175). In patients with small cell carcinoma
are suggested for the use of FDG-PET as a means for evalu-
first seen with a single lung nodule or mass (T1–2
ating potential response to therapy and/or follow-up of
lesions) without associated lymphadenopathy (stage I
patients with treated lung cancer:
disease), surgical resection followed by chemotherapy
1. FDG-PET is of proven value for restaging tumors. results in a 5-year survival of 30% to 50% (176–178).
Optimal restaging is best accomplished 1 to 2 months Review of long-term survivors with small cell carcinoma
after surgery and 2 to 6 months after completion of demonstrated that only patients who have undergone
chemoradiation therapy for differentiating between per- surgery for treatment of solitary pulmonary nodules may
sistent or recurrent tumor and fibrosis in cases with be expected to survive beyond 2 years (179). Although no
residual radiographic or CT abnormalities. evidence exists to suggest a role for maintenance ther-
2. FDG-PET is also of value for determining optimal sites apy, it has been suggested that in patients achieving
for biospy in patients with suspected recurrences. complete remission, prophylactic cranial irradiation be
3. Although suggestive, the efficacy of FDG-PET as a means undertaken (171).
for predicting response to therapy, including induction In distinction, extensive-stage disease is primarily
therapy before potential resection, as well as recur- treated with platinum-based chemotherapy alone. Whereas
rent disease, both local and metastatic, remains to be initial response rates range between 60% and 70%, with
established. complete response rates between 20% and 30%, median
survival remains only approximately 9 months for patients
with extensive-stage disease.
SMALL CELL LUNG CANCER
Imaging Evaluation
SCLC is a tumor with a very high biologic malignancy
that tends to be widespread at the time of diagnosis. Presently, CT remains the mainstay of imaging for patients
Fortunately, the number of cases of SCLC has diminished with SCLC. Indications for the use of FDG-PET remain
over the past decade, accounting for 13.8% of cases unclear, although a growing body of data suggests that PET
of bronchogenic carcinoma, as of 1998 (171). Unlike scanning may prove more appropriate in select cases
patients with NSCLC, following guidelines developed by (180–182). FDG-PET has been shown in small series to be
the Veterans Administration Lung Cancer Study Group, more accurate than CT in staging patients with extensive-
patients with small cell carcinoma are classified as having stage disease, but it is also of value in assessing prognosis.
either limited or extensive-stage disease (172). Limited dis- Pandit et al. (169), for example, in a retrospective study
ease refers to tumor restricted to a single hemithorax that of 62 scans in 46 patients (including 8 untreated and
can be treated within a single tolerable radiation port 38 treated patients) showed that patients with SCLC with
(172,173). This includes regional hilar, ipsilateral, and PET-positive studies had significantly worse outcomes than
contralateral mediastinal nodes, ipsilateral and contralat- PET-negative patients. In all cases, adequate pretherapeutic
eral scalene nodes, and ipsilateral pleural effusions evaluation should include either contrast-enhanced CT or
independent of whether the pleural fluid is benign or preferentially MRI of the brain, including in patients with-
malignant (174), with the proviso that patients with con- out neurologic symptoms. This reflects the high prevalence
tralateral hilar and/or supraclavicular nodes or those with of brain metastases that occur in patients with SCLC. As
malignant effusions, either pleural or pericardial, are previously discussed, in one prospective study of 134 pa-
excluded, except from clinical trials. Extensive disease tients evaluated with both contrast-enhanced CT and
double-dose gadolinium-enhanced MRI, MRI identified for surgery. Advocates have noted that in comparison with
brain lesions in a total of 19 patients, whereas CT identified standard lobectomies, for example, VATS has comparable
lesions in only 12 (123). survival statistics while minimizing pain, decreasing the
need for blood transfusions, and substantially decreasing
the length of hospital stay (190,191). As pointed out by
SURGICAL EVALUATION others, however, it is not infrequent for conversion to a
thoracotomy to occur, although the frequency with which
Surgery remains the only definitive method for curing lung this occurs will depend on surgical experience. It has also
cancer and is indicated in all cases of potentially operable been noted in at least one series that advantages obtained
stage IA tumors. In distinction, surgery is not indicated from preferentially performing VATS are lost by 2 weeks
in patients with stage IIIB (183), and in most with stage IV after surgery (192).
disease (184). Although indications for surgical resection Although VATS is frequently used to evaluate small, sub-
are less clearly defined for stage IIB–IIIA lesions (185,186), centimeter pulmonary nodules, as reviewed by Daniel
assessing resectability remains the most important function (193), at present no standard algorithm exists for the surgi-
of imaging. Part of this assessment calls for an understand- cal approach to these lesions. Factors that influence choice
ing of the limitations of surgery. of technique include nodule morphology, nodule accessi-
In an attempt to address disparities in reported results bility (including assessment of the relationship between
from surgical resections, the International Association for nodules and chest-wall structures), as well as a thorough
the Study of Lung Cancer (IASLC) (187) proposed that knowledge of techniques that facilitate VATS excisional
a formal definition of a “compete resection” meet the biopsies, such as use of preoperative coils, radiotracer local-
following criteria: that the margins of the resection be ization techniques, and intraoperative ultrasonography.
microscopically free of tumor; that systematic nodal
dissection be undertaken; that no evidence is found of
Superior Sulcus (Pancoast) Tumors
extracapsular tumor; and that the highest mediastinal
nodes resected be free of tumor as well. In those cases in Superior sulcus or Pancoast tumors constitute a separate
which systematic nodal dissections are not undertaken, the surgical category (Fig. 7-12). Initially defined as cancers
definition of a complete resection can still be met if a involving the lung apex, those associated with pain radiat-
“lobe-specific” nodal dissection is performed in which, in ing down the arm because of involvement of the brachial
addition to dissection of intrapulmonary and hilar nodes, plexus, or Horner syndrome because of involvement of the
at least three additional mediastinal nodal stations are stellate ganglion, it is now commonplace for all lesions
excised, with the specific nodes chosen depending on in the lung apices to be considered Pancoast tumors.
which lobe is resected, in all cases to include subcarinal A formal definition has been proposed by Detterbeck
nodes. In distinction, if tumor is identified in any of the (194) to include cancers arising in the lung apices that
tumor margins, evidence of extracapsular tumor is noted, involve structures of the apical chest wall. This includes all
or a residual positive node (as well as a malignant pleural lesions with evidence of chest-wall invasion, which may be
or pericardial effusion) is present, resections are consid- limited to the parietal pleura or include periostial or bony
ered “incomplete.” It should be noted that these guide- involvement of the upper ribs or apical vertebral bodies,
lines fail to include sublobar or bilobar resections, or the subclavian vessels, or nerve roots of the brachial plexus
classification of tumors that invade the visceral pleura. or stellate ganglion (194). Excluded by this definition are
Overall, rates of postoperative mortality have been esti- lesions involving the visceral pleura only or those involv-
mated to vary from 2% to 5%, with the risk noted to be ing only the second rib or below.
considerably worse for patients undergoing a pneumonec- Regardless of the definition, Pancoast tumors tradition-
tomy rather than a lobectomy (59). It has been shown ally have been first irradiated and then resected, provided
that both postoperative 30-day and in-hospital mortality no evidence is found of brachial plexus involvement. With
rates after pulmonary resection may be as low as 0.5% newer surgical approaches, however, and the use of com-
for pneumonectomies and 0.8% for lobectomies, res- bined chemotherapy and radiation therapy, it is now pos-
pectively, reflecting recent improvements in surgical sible to resect these lesions completely, even with evidence
technique (188). Although impressive, these data likely of vertebral body or neural foraminal involvement (194).
reflect the advantages of data obtained from highly As shown by Alifano et al. (195) in a retrospective study
specialized academic centers only (189). of 67 patients with superior sulcus tumors, a wide vari-
In part as a response to the current rates of postopera- ety of surgical techniques may be used to resect these
tive morbidity and mortality, considerable interest has lesions, including lobectomies, pneumonectomies, and
focused on the use of video-assisted surgery (VATS) both to even wedge resections encompassing patients with stage
diagnose and to resect lung cancers. This technique offers IIB to IIIB disease. Overall, complete surgical resections
the possibility of revising standard surgical approaches to were performed in 82%, with an operative mortality of
the entire gamut of diagnostic and therapeutic indications 8.9%. This is significant, as 5-year survival rates were
significantly higher in patients undergoing complete with only partial response (8.7 months) (p 0.01). Similar
rather than partial resections (44.9% vs. 0; p .000065). results have been reported in another study of 33 lung
It is noteworthy that additional postoperative therapy was tumors in 18 patients (including 5 patients in whom RFA
frequently administered in this study, most often includ- was performed by minithoracotomy) in which radiographic
ing radiation therapy, less often combined radiation and evidence of response was noted in 66% of tumors smaller
chemotherapy or chemotherapy alone. This may reflect the than 5 cm (202). As noted by Dupuy et al. (203), still better
fact that treatment of superior sulcus tumors has generally results may be obtained when RFA is followed by con-
not required assessment of mediastinal nodes with medi- ventional radiotherapy, with cumulative 2- and 5-year
astinoscopy or a search for distant metastases, although survival rates of 50% and 39%, respectively, in a study of
5-year survival in patients with N2 or N3 disease is less 24 consecutive patients deemed medially inoperable.
than 10%. An exception seems to be involvement of ipsi- Whereas treatment of larger lesions often proves problem-
lateral supraclavicular nodes; although by definition these atic, even lesions greater than 3 cm may be successfully
nodes are classified as N3 disease, in this setting, they treated with the use of overlapping ablations (204).
appear to have a prognostic significance closer to that of Typical patterns of response on follow-up CT scans
N1 disease. have been described (201). Initially, lesions tend to be
surrounded by a zone of ground-glass attenuation, which
usually resolves in less than 1 month. Subsequent cavita-
RADIOFREQUENCY ABLATION tion occurs in as many as one third of cases, especially
those lesions with bronchial communications, with bub-
Radiofrequency thermal ablation (RFA) is a relatively new ble-like lucencies alternatively developing in the remainder.
technique that uses an alternating RF current to cause coag- In most cases, follow-up CT demonstrates apparent tumor
ulation necrosis of tumors by means of frictional heating. enlargement with subsequent tumor regression. There may
It involves percutaneous placement within tumors of an also be evidence of focal pleural thickening, especially adja-
electrode made of an insulated metal shaft with an exposed cent to peripheral lesions, often associated with infarct-like
conductive tip attached to an external RF generator parenchymal opacities (Figs. 7-39 and 7-40). As noted
grounded with a reference electrode, usually in the form of a by Bojarski et al. (201), lesions that continue to enlarge
pad positioned on the opposite side of the body. The pri- after 6 months likely reflect local recurrence.
mary aim is to attain intratumoral temperatures between Complications are frequent and include fever, pneu-
50°C and 100°C, a lethal range for most tissues. Lesions mothoraces (occurring in approximately one third of cases,
within the lung prove to be especially amenable to this tech- only infrequently requiring chest tube placement), and
nique, as the surrounding normal lung acts to insulate effusions occurring in up to 30% of cases (196). Abscesses
lesions effectively, facilitating tissue coagulation (196). complicating RFA have also been described as occurring in
The main indication for the use of RFA is to treat other- up to 6% of cases. To date, treatment-related deaths have
wise unresectable or inoperable patients, for example, only rarely been reported, occurring in one case as a com-
those with comorbid cardiovascular or lung disease. plication of RFA because of hemoptysis more than 2 weeks
Although most such patients are traditionally treated with after treatment (202).
either radiation or chemotherapy alone or in combination,
RFA of lung malignancies represents an important alter-
native approach used both for attempted cure and for SCREENING
palliation, either alone or in combination with other
treatment modalities, including both radiotherapy and Few topics have proved more controversial than LDCT
brachytherapy. This latter category includes use of RFA to screening of lung cancer. Initially reported in the late
shrink tumors before radiotherapy, to treat local symptoms 1990s (205–207), numerous subsequent reports have
resulting from chest-wall or rib invasion, as well as to treat focused either on the clear and unequivocal benefits
disease recurrence in patients for whom further radiation of screening (208) or conversely on the unproved
or surgery is not indicated and unfounded nature of these claims (209). Although
To date, numerous studies have shown that RFA may be a detailed evaluation of the arguments pro and con
used successfully to treat otherwise unresectable and/or regarding LDCT is outside the scope of the present
inoperable parenchymal neoplasms, both primary and review, it is worthwhile to consider published data to
metastatic (197–201). Lee et al. (200), for example, in a gain some insight into the current status of lung cancer
study of 30 patients with either NSCLC (27 lesions) or screening.
pulmonary metastases (5 lesions), attained complete necro- The elements of a beneficial screening test have been
sis in 12 (38%) of 32 lesions, with partial necrosis numerously outlined and entail consideration of a num-
(defined as 50%) in the remaining 20 (62%) lesions. ber of factors, including those related to the disease itself,
Mean survival of patients with complete necrosis those related to the test, and those related to therapy (210).
(19.7 months) proved significantly longer than for those Factors related to the disease being screened include an
A B
C D
Figure 7-39 Radiofrequency ablation (RFA). A: Magnified view of a well-defined ground-glass nodule in the anterior aspect of the left
upper lobe in a 70-year-old woman previously irradiated for breast cancer, with osteonecrosis of the ribs and clavicle as well as aortic
stenosis, who was not considered a surgical candidate. Percutaneous needle biopsy showed a bronchoalveolar cell carcinoma. B–D:
Sequential magnified views at the same level as shown in A at 1, 6, and 12 months later, respectively. Note that in the immediate post-treat-
ment period (B), the lesion appears larger, with a zone of peripheral ground-glass attenuation seen associated with a central lucency, likely
due to necrosis. Note subsequent progressive retraction of this lesion (C and D) until it resembles a subpleural infarct. (Case courtesy of
JoAnne Shepard, MD, Massachusetts General Hospital, Boston, Massachusetts.)
assessment of the seriousness of the disease, whether a preclinical disease, its ability to detect tumors before the
high prevalence of detectability exists in the preclinical critical point of developing metastases, and whether the
stage, and whether minimal “pseudodisease” is present. test is safe and cost-effective. Finally, factors related to
This latter includes both the inadvertent diagnosis of therapy also must be considered and include whether
benign disease, as well as, most important for lung cancer effective therapy is available for preclinical disease and
screening, the key issue of potential overdiagnosis. Factors whether therapy is more effective for preclinical than for
related to the test itself include its accuracy for diagnosing symptomatic disease, as well as an assessment of related
A B
C D
Figure 7-40 Radiofrequency ablation (RFA). A: Magnified view through the right lower lobe showing a well-defined peripheral lesion in a
72-year-old man with a history of colon cancer, 2 years after successful liver resection for a hepatic metastasis, now with a new lung metas-
tasis documented by transthoracic needle biopsy. B: CT section at the same level as shown in A, 1 month later, shows typical evolution, with
evidence of central necrosis resulting in decreased density and cavitation, surrounded by a zone of ground-glass attenuation. Note that the
lesion has markedly increased in size when compared with initial CT appearance. C, D: Sequential magnified views at the same level as
shown in A and B, 6 and 12 months later, respectively, now show a focal subpleural density, simulating the appearance of a pulmonary
infarct (compare with Fig. 7-39). Although stability over a 6-month period is indicative of presumed scarring, in the absence of histologic ver-
ification, continued monitoring of this lesion is necessary to exclude residual and/or recurrent disease. (Case courtesy of JoAnne Shepard,
MD, Massachusetts General Hospital, Boston, Massachusetts.)
therapeutic morbidity and mortality. As will be discussed, overall estimated 10-year survival rate of 88% (95% CI,
for the specific case of lung cancer screening, most impor- 84–91). Still more striking, of 302 patients with clin-
tant have been the issues of potential overdiagnosis and ical stage I disease undergoing surgical resection within
issues related to the potential risks of therapy. 1 month of their diagnosis, the author’s estimated 10-year
It is important to put LDCT screening in context. survival proved to be 92% (95% CI, 88–95).
At present, routine radiographic screening of lung cancer A number of salient facts make this study of particular
has failed to gain general acceptance (211). The results of interest. First, using 5 mm as a cut-off for solid and mixed
four randomized controlled studies including 37,724 indi- solid/ground-glass lesions and 8 mm for pure ground-
viduals performed in the early 1980s failed to show evi- glass lesions, only 4,186 (14%) of 31,567 baseline studies
dence of improved disease-specific mortality, leading most were interpreted as positive, whereas only 1,460 (5%) of
critics to conclude that no indication exists for routine ra- 27,456 incidence studies were so interpreted. These num-
diographic screening (212,213). Although CT has been bers stand in striking distinction to other studies in which
conclusively shown to be more effective at detecting small the likelihood of a positive screening study has been
peripheral lung cancers than routine chest radiographs, it reported to be as high as 74% (59). It is also of interest
should be noted that the potential for routine chest radi- that the study was truly international in scope, As previ-
ographic screening for lung cancer is currently being ously noted, there is evidence that the prevalence and
re-evaluated. Recently, cancer-detection rates from baseline nature of cancers in different genetic populations may vary
screening radiographs have been reported from the considerably. In this regard, 20% of individuals enrolled
Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer were Japanese, including 10% of persons whose only expo-
Screening Trial (214). Of a total of 77,465 individuals sure was to secondhand smoke (7). The actual contribu-
randomly assigned to be screened with a single-view tion of this group to the overall number of lung cancer
posterior–anterior chest radiograph, 5,991 (8.9%) radio- cases is not reported.
graphs were deemed suggestive of lung cancer; of these, Also of interest is the fact that although most cancers
206 (3.4%) patients subsequently underwent biopsy, 126 proved to be adenocarcinomas, only 20 of 348 baseline
(61%) of whom proved to have lung cancer, resulting in a adenocarcinomas proved histologically to be bronchi-
PPV of 2.1%. Significantly, 44% of these cancers proved to oloalveolar cell cancers. Not surprisingly, only one addi-
be stage I NSCLCs, with the highest rates found in current tional case of BAC was diagnosed on follow-up scans,
smokers (6.3 per 1,000 screened individuals) compared raising the possibility that at least some of the smaller
with both ex-smokers (4.9 per 1,000 screened) and never ground-glass nodules identified on baseline studies may in
smokers (0.4 per 1,000 screened), respectively. Although fact represent truly indolent cancers for which biopsy may
still inconclusive, these data suggest that a role may not be indicated. This unexpectedly small number likely
well exist for routine radiographic screening for lung reflects the use of stringent criteria for documenting
cancer (214). BAC. As recently noted by these same investigators in an
To date, a number of large nonrandomized observa- evaluation of 65 individuals with resections for LDCT
tional studies have evaluated the use of LDCT screening screen-detected lung cancer, of 42 solitary adenocarcino-
for lung cancer. These studies have consistently shown that mas identified in this study, 37 of 42 proved to be invasive,
LDCT is capable of detecting early stage 1 lung cancers with only 5 representing true BAC (220). More impor-
before their identification with corresponding chest radi- tantly, fully 18 cases diagnosed as BAC by referring pathol-
ographs (59,208,215–219). As one example, in a single- ogists from outside institutions were reinterpreted as
arm prospective cohort study of 1,520 at-risk individuals, invasive carcinomas, casting doubt on the accuracy of this
in total, 68 lung cancers were detected in 66 participants, diagnosis in general pathology practice.
of which 61% were stage I NSCLCs with a mean diameter Finally, it is worth emphasizing that in the I-ELCAP
of 14.4 mm (range, 5 to 50 mm) (59). study cited above, of a total of 535 biopsies performed,
The most compelling study to date supporting a role 492 (92%) proved diagnostic, almost all presumably
for lung cancer screening has recently been reported by greater than 1 cm in size. Unfortunately, it is not clear how
the International Early Lung Cancer Action Program many of these proved to be solid versus nonsolid lesions,
Investigators (I-ELCAP) (208). In this study, including nor are data provided accounting for the 43 biopsies (8%)
31,567 baseline LDCT screening studies in asymptomatic that proved nondiagnostic. Also of interest is the fact that
persons at risk for lung cancer, with an additional 27,456 only 166 (34%) of the 484 proven lung cancers were eval-
interval screens, a total of 484 individuals proved to have uated with FDG-PET imaging, rendering the use of this
lung cancer. Of these, 412 (85%) had clinical stage I lung technique for screened cases unspecified.
cancer, of which 405 were identified at baseline (preva- Despite these considerations, it is clear that these data
lence of 1.3%) and 74 were identified on follow-up scans represent as strong an argument for LDCT screening of the
(incidence of 0.3%). Only an additional 5 symptomatic lung as has been made to date. This is especially true given
interval cancers were identified between screening studies. the truly small number of baseline and follow-up scans
Using statistical modeling, these authors reported an interpreted as positive, the high yield of biopsies, the small
number of BACS (rendering less significant the likelihood 92.4% and 96% for individual nodules when estab-
of overdiagnosis of at least this group of lesions), and the lished by observation or surgery, respectively (59). As
author’s reported operative mortality rate of only 2 (0.5%) documented in this study, 13 individuals underwent
of 411 resections. 15 surgical procedures for benign disease without surgical
Although most studies indicate a significant stage shift mortality. In another screening study of 1,520 participants,
in lung cancer detection, and hence a compelling rationale 55 (3.6%) underwent 60 operations for a variety of indica-
for pursuing LDCT screening, the ultimate value of early tions, including suggestive pulmonary nodules and
diagnosis has been questioned. Of particular concern has mediastinal adenopathy; of these, benign disease was iden-
been the significance of lesion size at the time of tified in 10 (18.1%) patients (225). Overall, complications
diagnosis, the possibility of overdiagnosis, and the risks occurred in 27% of patients, including 1.7% operative
associated with too numerous false-positive findings. mortality. Similar results have been reported by Pinsky
It appears that despite some conflicting data (50), a sig- et al. (226). In their study of 1,660 eligible subjects, includ-
nificant correlation does appear to exist between tumor ing 1,558 individuals undergoing baseline LDCT screen-
size at the time of diagnosis and tumor stage. In one study ings and 1,398 undergoing a 1-year incidence screen, a
of 28,689 baseline and repeated screening examinations, total of 522 (approximately 60%) subjects had at least one
including a total of 436 lung cancers diagnosed with positive finding; of these, 89 biopsies were performed on
staging based on postsurgical findings in 368 (84%) of 62 (12%) individuals, of whom 36 (58%) proved to have
426 cases of NSCLC and in 8 of 28 cases of SCLC, lesions lung cancer. In an additional 330 subjects, at least one fol-
categorized as 15 mm or less, 16 to 25 mm, 26 to 35 mm, low-up CT was obtained in 63%. Importantly, in this same
and 36 mm or greater, the percentages of cases with study, a wide variation was noted in the manner in which
no metastases (N0M0) were 91%, 83%, 68%, and 55%, nodules were subsequently evaluated, indicative of a lack
respectively (46,221). of a standardized well-accepted algorithm for workup of
Despite the finding of a close correlation between tumor suggestive findings identified on LDCT screening studies.
size and lymph node status, it remains to be determined The potential that false-positive findings will lead to unnec-
whether early detection actually leads to a decrease in essary interventions, including additional CT examinations
disease-specific mortality. Identifying early-stage lesions by as well as lung biopsies, clearly represents an important
itself may not prove significant unless accompanied by a concern.
subsequent decrease in the number of cases with advanced As already discussed (208), in an attempt to address
disease, raising the issue of potential overdiagnosis. By defi- this issue, Henschke et al. (227) proposed the use of
nition, overdiagnosis occurs when early-stage lung cancers 5 mm as a cut-off for deciding whether to obtain routine
identified by screening never eventuate in death, with yearly follow-up examinations versus more aggressive CT
patients dying instead of comorbid cardiovascular or other surveillance for solid and part-solid nodules. In this retro-
disease (222). Although definitive data regarding overdiag- spective review of two separate prevalence screenings of
nosis are currently unavailable, estimates of lung cancer 2,897 individuals performed over nearly a decade, the fre-
mortality from LDCT screening studies have been calcu- quency of malignancy that was or could have been diag-
lated. In one study reviewing data from two prospective nosed based on nodule growth rates was 0 of 378 when
observational trials, estimated lung cancer mortality ranged the largest lesion was smaller than 5 mm, versus 13 of 238
between 4.1 and 5.5 deaths per 1,000 person-years, similar when the largest lesion was between 5 and 9 mm. Based
to the lung cancer mortality identified in the prior Mayo on these data, these authors concluded that the risk of
Lung Project chest radiography study (222). Others have developing a lung cancer when the largest lesion was
contested the use of modeling in place of actual data as not smaller than 5 mm was sufficiently small to warrant
contributory (223). re-evaluation only on a yearly basis, effectively decreasing
Given the proposition that LDCT screening may lead to the number of referrals for subsequent evaluation of base-
a significant reduction in lung cancer deaths, it remains to line nodules by nearly 50%.
be determined whether the number of false-positive studies Currently, given the nature of the controversies surround-
offsets any positive effect of early lung cancer detection. ing LDCT screening, definitive evaluation must await the
With the exception of the recent I-ELCAP report cited previ- results of the ongoing prospective National Lung Cancer
ously (208), to date, nearly all reported studies have docu- Screening Trial (NLST). In the collaboration between the
mented the presence of a large number of benign American College of Radiology Imaging Network (ACRIN)
noncalcified nodules, the number increasing with the and the Lung Screening Study of the National Cancer
introduction of multidetector CT scanners (59,216,224). As Institute Prostate, Lung, Colon and Ovary (PLCO) trial
reported by Swensen et al. (59), in their 5-year prospective begun in 2002, to date, 53,418 participants have been
study of 1,520 individuals, although in total, 68 lung can- enrolled and randomized (228). It is anticipated that within
cers were identified, noncalcified nodules were identified the next several years, sufficient data will be available to put
in 1,118 (74%); overall, false-positive results were identi- into clearer perspective the value of LDCT screening. It is
fied in 69% of individuals, with false-positive rates of worth noting that this study focuses not only on LDCT but
also on the potential use of peripheral biomarkers. That 11. Dittman H, et al. (18F)FLT PET for diagnosis and staging of tho-
racic tumours. Eur J Nucl Med Mol Imaging. 2003;30:1407–1412.
some other correlative studies might be necessary to opti-
12. Buck AK, et al. Clinical relevance of imaging proliferative activ-
mize lung cancer screening is especially clear when it is ity in lung nodules. Eur J Nucl Med Mol Imaging. 2005;32:
considered that in all reported LDCT studies to date, the vast 522–523.
majority of lesions identified are peripheral adenocarcino- 13. Yap CS, et al. Evaluation of thoracic tumors with 18F-
fluorothymidine and 18F-fluorodeoxyglucose-positron emission
mas (59,217,218,221,229). Although the incidence of squa- tomography. Chest. 2006;129:393–401.
mous and large cell cancers has been declining, the relative 14. Beasley MB, Branbilla E, Travis WD. The 2004 World Health
lack of these tumors in reported LDCT screening studies is Organization Classification of lung tumors. Semin Roentgenol.
2005; 40:90–97.
evidence that CT by itself may not be an ideal method for 15. Gu CD, et al. Detection of micrometastatic tumor cells in pN0
lung cancer screening. As shown by McWilliams et al. (230), lymph nodes of patients with completely resected nonsmall cell
it is possible to alter the present LDCT imaging paradigm lung cancer: impact on recurrence and survival. Ann Surg.
2002;235:133–139.
substantially by coupling CT with a combination of sputum 16. Aguayo SM, Miller YE, Waldron JA Jr, et al. Brief report: idio-
cytology and autofluorescence bronchoscopy. In this study, pathic diffuse hyperplasia of pulmonary neuroendocrine cells
561 high-risk subjects were evaluated by using automated and airways disease. N Engl J Med. 1992;327:1285–1288.
17. Brown MJ, English J, Muller NL. Bronchiolitis obliterans due to
quantitative image cytometry as the first screening method; neuroendocrine hyperplasia: high-resolution CT-pathologic
among these, 423 (75%) were found to have sputum atypia, correlations. AJR Am J Roentgenol. 1997;168:1561–1562.
whereas noncalcified nodules were found in 259 (46%). 18. Lee JS, Brown KK, Cool C, et al. Diffuse pulmonary neuroen-
docrine cell hyperplasia: radiologic and clinical features.
Of 14 detected cancers, 13 had sputum atypia, with 9 of J Comput Assist Tomogr. 2002;26:180–184.
these cases detected by CT and 4 by bronchoscopy; only one 19. Travis WD, et al. Evolving concepts in the pathology and com-
case was detected solely by CT, whereas 29% of cancers puted tomography imaging of lung adenocarcinoma and bron-
chioloalveolar carcinoma. J Clin Oncol. 2005;23:3279–3287.
would have been missed if the diagnosis relied solely on CT. 20. Noguchi M, et al. Small adenocarcinoma of the lung: histologic
As important, use of sputum cytology in this study poten- characteristics and prognosis. Cancer. 1995;75:2844–2852.
tially could have reduced the number of LDCT studies by 21. Wislez M, et al. Clinical characteristics of pneumonic type ade-
nocarcinoma of the lung. Chest. 2003;123:1868–1877.
25%. Although these results are not immediately applicable 22. Travis WD, Linnoila RI, Tsokos MG. Neuroendocrine tumors
to all institutions lacking automated quantitative sputum of the lung with proposed criteria for large-cell neuroen-
cytology or autofluorescence bronchoscopy, the conclusion docrine carcinoma: an ultrastructural, immunohistochemical,
and flow cytometric study of 35 cases. Am J Surg Pathol.
that LDCT by itself may be inadequate for optimal screening 1991;15:529–553.
warrants further evaluation. It remains to be determined 23. Chong SM, et al. Neuroendocrine tumors of the lung: clinical,
whether the addition of either biomarkers and/or cytology pathologic, and imaging findings. Radiographics. 2006;26:41–58.
24. Munden RF, et al. Imaging of the patient with non-small cell
to the screening armamentarium will ultimately provide lung cancer: what the clinician wants to know. Radiology.
a statistically meaningful reduction in the mortality of 2006;238:803–818.
lung cancer. 25. Denoix PF. Enquete permanent dans les centres anticancereux.
Bull Inst Natl Hyg. 1946;1:70.
26. Little AG, Stitik FP. Clinical staging of patients with non-small
cell lung cancer. Chest. 1990;97:1431–1438.
27. American Joint Committee for Cancer Staging and End Results
REFERENCES Reporting. Clinical staging system for carcinoma of the lung.
Philadelphia: Lippincott; 1973.
1. Alberg AJ, Samet JM. Epidemiology of lung cancer. In: Kane K, 28. American Joint Committee for Cancer Staging and End-results
Bun PA, Kane MA, eds. Biology of lung cancer. New York: Marcel Reporting: Task Force on Lung Cancer. Staging of Lung Cancer
Dekker: 1998:11–51. 1979. Chicago: AJCC, 1979.
2. Jemal RC, Tiwari RC, Murray T. Cancer statistics, 2004. CA 29. Mountain CF. Revisions in the International System for Staging
Cancer J Clin. 2004;54:8–29. Lung Cancer. Chest. 1997;111:1710–1717.
3. Wingo PA, et al. Long-term trends in cancer mortality in the 30. Naruke T, et al. Prognosis and survival in resected lung carci-
United States, 1930–1998. Cancer 2003;97:31–33. noma based on the new international staging system. J Thorac
4. Alberg AJ, Samet JM. Epidemiology of lung cancer. Chest 2003; Cardioivasc Surg. 1988;96:440–447.
123:21S–49S. 31. Martini N, et al. Preoperative chemotherapy for stage IIIa (N2)
5. Yang P, et al. Clinical features of 5,628 primary lung cancer lung cancer: the Sloan-Kettering experience with 136 patients.
patients: experience at Mayo Clinic from 1997 to 2003. Chest. Ann Thorac Surg. 1993;55:1365–1374.
2005;128:452–462. 32. Kirsch MM, Sloan H. Mediastinal metastases in bronchogenic
6. Barnes DJ. The changing face of lung cancer. Chest. 2004; carcinoma: influence of postoperative irradiation, cell type, and
126:1718–1721. location. Ann Thorac Surg. 1982;33:459.
7. Patel JD, Bach PB, Kris MG. Lung cancer in US women: a con- 33. Pearson FG, DeLaure NC, Ilves Rea. Significance of positive
temporary epidemic. JAMA. 2004;291:1763–1768. superior mediastinal nodes identified at mediastinoscopy in
8. Toh C-K, et al. The impact of smoking status on the behaviour patients with resectable cancer of the lung. J Thorac Cardiovasc
and survival outcome of patients with advanced non-small Surg. 1982;83:1–11.
cell lung cancer: a retrospective analysis. Chest. 2004;126: 34. Smith RA. The importance of mediastinal lymph node invasion
1750–1756. by pulmonary carcinoma in selection of patients for resection.
9. Ferguson MK, Wang J, Hoffman PC. Sex associated differences Ann Thorac Surg. 1978;25:5.
in survival of patients undergoing resection of lung cancer. Ann 35. Martini N, et al. Results of resection in non-oat cell carcinoma
Thorac Surg. 2000;69:245–250. of the lung with mediastinal lymph node metastases. Ann Surg.
10. Potti A, Mukherjee S, Petersen R, et al. A genomic strategy to 1983;198:386–397.
refine prognosis in early-stage non-small cell lung cancer. 36. Boring CC, Squires SS, Tong T. Lung cancer. Cancer J Clin.
N Engl J Med. 2006;355:570–580. 1992;42:19–38.
37. Mathews MJ, et al. Frequency of residual and metastatic tumor 62. Duysinx B, et al. Evaluation of pleural disease with 18-
in patients undergoing curative surgical resection in lung cancer. fluorodeoxyglucose positron emission tomography imaging.
Chem Rep. 1973;4:63. Chest. 2004;125:489–493.
38. Winstanley DP. Fruitless resections. Thorax. 1968;23:327. 63. Riquet M, et al. Skip mediastinal lymph node metastasis and
39. Sider L, Horejs D. Frequency of extrathoracic metastases from lung cancer: a particular N2 subgroup with a better prognosis.
bronchogenic carcinoma in patients with normal-sized hilar Ann Thorac Surg. 2005;79:225–233.
and mediastinal lymph nodes on CT. AJR Am J Roentgenol. 64. Terzi A, et al. Early hilar lung cancer: clinical aspects and long
1988;151:893–895. term survival: identification of a subgroup of stage IA patients
40. Finke NM, et al. Autopsy results after surgery for non-small cell with more favorable prognosis. Lung Cancer J. 2000;27:
lung cancer. Mayo Clin Proc. 2004;79:1409–1414. 119–124.
41. Furak J, et al. Lung cancer and its operable brain metastasis: sur- 65. Mathur PN, et al. Treatment of early stage non-small cell lung
vival rate and staging problems. Ann Thorac Surg. 2005;79: cancer. Chest. 2003;123:176S–180S.
241–247. 66. Gabor S, et al. Invasion of blood vessels as significant prognos-
42. Decamp MM, Rice TW, Adelstein D, et al. Value of accelerated tic factor in radically resected T1–3N0M0 non-small-cell lung
multimodality therapy in stage IIIA and IIIB non-small cell lung cancer. Eur J Cardiovasc Thorac Surg. 2004;25:439–442.
cancer. J Thorac Cardiovasc Surg. 2003;126:17–27. 67. Chern MS, Wu MH, Chang CY. CT and MRI for staging of locally
43. Mountain CF. A new international staging system for lung can- advanced non-small cell lung cancer. Lung Cancer. 2003;
cer. Chest. 1986;89:225S–233S. 42:S5–S8.
44. Buccheri G, Ferrigno D. Prognostic value of stage grouping and 68. Pearlberg JL, et al. Limitations of CT in evaluation of neoplasms
TNM descriptors in lung cancer. Chest. 2000;117:1247–1255. involving the chest wall. J Comput Assist Tomogr. 1987;
45. Goldstraw P, Crowley JJ. The International Association for the 11:290–293.
Study of Lung Cancer international staging projecton lung 69. Libshitz HI. Computed tomography in bronchogenic carci-
cancer. J Thorac Oncol. 2006;1:281–286. noma. Semin Roentgenol. 1990;25:64–72.
46. Mulligan CR, Meram AD, Proctor CD, et al. Lung cancer staging: 70. Scott IR, et al. Resectable stage III lung cancer: CT, surgical, and
a case for a new T definition. Ann Thorac Surg. 2006;82:220–226. pathologic correlation. Radiology. 1988;166:75–79.
47. Wisnivesky JP, Yankelevitz D, Henschke CI. The effect of tumor 71. Pennes DR, et al. Chest wall invasion by lung cancer: limitations
size on curability of stage I non-small cell lung cancers. Chest. of CT evaluation. AJR Am J Roentgenol. 1985;144: 507–511.
2004;126:761–765. 72. Glazer HS, et al. Pleural and chest wall invasion in bron-
48. Gajra A, et al. Impact of tumor size on survival in stage IA non- chogenic carcinoma: CT evaluation. Radiology. 1985;157:
small cell lung cancer: a case for subdividing stage IA disease. 191–194.
Lung Cancer. 2003;42:51–57. 73. Shrager JB, Kaiser LR. Thoracoscopic lung biopsy: five com-
49. Birim O. Survival pathologic stage 1A non-small cell lung can- monly asked questions about video-assisted thoracic surgery.
cers. Ann Thorac Surg. 2005;79:1137–1141. Postgrad Med. 1999;106:139.
50. Patz EF Jr, et al. Correlation of tumor size and survival in 74. Murata K, et al. Chest wall and mediastinal invasion by lung
patients with stage IA non-small cell lung cancer. Chest. 2000; cancer: evaluation with multisection expiratory dynamic CT.
117:1568–1571. Radiology. 1994;191:251–255.
51. Black WC. Unexpected observations on tumor size and survival 75. Watanabe A, et al. Chest CT combined with artificial pneu-
in stage IA non-small cell lung cancer. Chest 2000;117: mothorax: value in determining origin and extent of tumor. AJR
1532–1534. Am J Roentgenol. 1991;156:707–710.
52. Lee KS, et al. T1 non-small cell lung cancer: Imaging and 76. Yokoi K, et al. Tumor invasion of the chest wall and medi-
histopathologic findings and their prognostic implications. astinum in lung cancer: evaluation with pneumothorax CT.
Radiographics. 2004;24:1617–1631. Radiology. 1991;181:147–152.
53. Flieder DB, et al. Tumor size is a determinant of stage distribu- 77. Glazer HS, et al. Indeterminate mediastinal invasion in bron-
tion in T1 non-small cell lung cancer. Chest. 2005;128: chogenic carcinoma: CT evaluation. Radiology. 1989;173:37–42.
2304–2308. 78. Baron RL, et al. Computed tomography in the preoperative
53a. Henschke CI, Yankelevitz DF, Miettinen OS. International Early evaluation of bronchogenic carcinoma. Radiology. 1982;145:
Lung Cancer Action Program Investigators: computed tomo- 727–732.
graphic screening for lung cancer: the relationship of disease 79. Webb WR, et al. CT and MR imaging in staging non-small cell
stage to tumor size. Arch Intern Med. 2006;166:321–325. bronchogenic carcinoma: report of the Radiologic Diagnostic
54. Carbone E, Asamura H, Kondo H, et al. T2 tumors larger than Oncology Group. Radiology. 1991;178:705–713.
five centimeters in diameter can be upgraded to T3 in non- 80. Gilbert A, et al. Tracheal sleeve pneumonectomy for carcinoma
small cell lung cancer. J Thorac Cardiovasc Surg. 2001;122: of the proximal left main bronchus. Can J Surg. 1984;27:
907–912. 583–585.
55. Deslauriers J, Brisson J, Cartier R, et al. Carcinoma of the lung: 81. Jensik RJ, Faber LP, Kittle CF, et al. Survival in patients undergo-
evaluation of satellite nodules as a factor influencing prognosis ing tracheal sleeve pneumonectomy for bronchogenic carci-
after resection. J Thorac Cardiovasc Surg. 1989;97:504–512. noma. J Thorac Cardiovasc Surg. 1982;84:489–496.
56. Watanabe A, Shimizu J, Oda M, et al. Proposals regarding some 81a. Bains MS. Surgical treatment of lung cancer. ACCP Section
deficiencies in the new international staging system for non- Report (part 2). Chest. 1991;100:826–837.
small-cell lung cancer. Jpn J Clin Oncol. 1991;21:160–168. 82. Mountain CF, Dresser CM. Regional lymph node classification
57. Shimizu N, Ando A, Date H, et al. Prognosis of undetected for lung cancer staging. Chest. 1997;111:1718–1723.
intrapulmonary metastases in resected lung cancer. Cancer 83. Mountain CF. Revisions in the International System for Staging
1993;71:3868–3872. Lung Cancer. Chest. 1997;111:1710–1717.
58. Martini N, Melamed MR. Multiple primary lung cancers. J 84. Glazer HS, et al. CT demonstration of calcified mediastinal
Thorac Cardiovasc Surg. 1975;70:606–612. lymph nodes: a guide to the new ATS classification. AJR Am J
59. Swensen SJ, Jett JR, Hartman TE, et al. CT screening for lung Roentgenol. 1986;147:17–20.
cancer: five-year prospective experience. Radiology. 2005;235: 85. Kop JP, et al. CT depiction of regional nodal stations for lung
259–265. cancer staging. AJR Am J Roentgenol. 2000;174:775–782.
60. Gupta NC, et al. Clinical role of F-18 fluorodeoxyglucose 86. Aquino SL, et al. Source and direction of thoracic lymphatics:
positron emission tomography imaging in patients with lung part I: upper thorax. J Comput Assist Tomogr. 2003;27:292–296.
cancer and suspected malignant pleural effusion. Chest. 2002; 87. Ginsberg RJ. Continuing controversies in staging NSCLC: an
122:1918–1924. analysis of the revised 1997 staging system. Oncology. 1998;
61. Schaffler GJ, et al. Non-small cell lung cancer: evaluation of 12(suppl 1):51–54.
pleural abnormalities on CT scans with T-18 FDG PET. 88. Wang K-P. Staging of bronchogenic carcinoma by bron-
Radiology. 2004;231:858–865. choscopy. Chest. 1994;106:588–593.
89. Riviera H. Anatomy of the human lymphatic system. Ann Arbor, MI: small cell lung cancer: experience from a UK cardiothoracic cen-
Edwards, 1938. tre. Clin Radiol. 2004;59:935–938.
90. Borrie J. Primary carcinoma of the bronchus: prognosis follow- 115. Walsh GL, et al. Is follow-up of lung cancer patients after resec-
ing surgical resection. Ann R Coll Surg Engl. 1952;10:165 tion medically indicated and cost-effective? Ann Thorac Surg.
91. Nohl-Oser HC. Lymphatics of the lung. In: Shields TW, ed. 1995;606:1563–1570.
General thoracic surgery. Philadelphia: Lea & Febiger; 1989. 116. Pfister DG, et al. American Society of Clinical Oncology treat-
92. Libshitz HI, McKenna RJ, Mountain CF. Patterns of mediastinal ment of unresectable non-small-cell lung cancer guideline:
metastases in bronchogenic carcinoma. Chest. 1986;90: update 2003. J Clin Oncol. 2004;22:330–353.
229–232. 117. Johnson BE. Second lung cancers in patients after treatment for
93. Nohl-Oser HC. An investigation of the anatomy of the lym- an initial lung cancer. J Natl Cancer Inst. 1998;90:1335–1345.
phatic drainage of the lungs. Ann R Coll Surg Engl. 1972;51:157. 118. Jennings SG, et al. Lung tumor growth: assessment with CT:
94. Quint LE, Glazer GM, Orringer MB, et al. Mediastinal lymph comparison of diameter and cross-sectional area with volume
node detection and sizing at CT and autopsy. AJR Am J measurements. Radiology. 2004;231:866–871.
Roentgenol. 1986;147:469–472. 119. Revel MP, et al. Are two-dimensional CT measurements of small
95. Genereux GP, Howie JL. Normal mediastinal lymph node size noncalcified pulmonary nodules reliable? Radiology. 2004;231:
and number: CT and anatomic study. AJR Am J Roentgenol. 453–458.
1984;142:1095–1100. 120. Webb WR, Jansen BC, Sollitto R, et al. Bronchogenic carcinoma:
96. Quinn DL, et al. Staging of non-small cell bronchogenic carci- staging with MR compared with staging with CT and surgery.
noma: relationship of the clinical evaluation to organ scans. Radiology. 1985;156:117–124.
Chest. 1986;89:270–275. 121. Tobler J, et al. Differentiation of proximal bronchogenic carci-
97. Ikezoe J, et al. Mediastinal lymph node metastases from nons- noma from postobstructive lobar collapse by magnetic reso-
mall cell bronchogenic carcinoma: reevaluation with CT. J nance imaging: comparison with computed tomography. Invest
Comput Assist Tomogr. 1990;14:340–344. Radiol. 1987;22:538–543.
98. Buy JN, et al. Computed tomography of mediastinal lymph 122. Herold CJ, Kuhlman JE, Zerhouni EA. Pulmonary atelectasis.
nodes in non small cell lung cancer: a new approach based on Radiology. 1991;178:715–720.
the lymphatic pathway of tumor spread. J Comput Assist Tomogr. 123. Suzuki K, et al. Magnetic resonance imaging and computed
1988;12:545–552. tomography in the diagnoses of brain metastases of lung cancer.
99. Eloubeidi MA, et al. Endoscopic ultrasound-guided fine needle Lung Cancer. 2004;46:357–360.
aspiration of mediastinal lymph node in patients with suspected 124. Elsayes KM, et al. Adrenal masses: MR imaging features with
lung cancer after positron emission tomography and computed pathologic correlation. Radiographics. 2004;24:S73–S86.
tomography scans. Ann Thorac Surg. 2005;79:263–268. 125. Outwater EK, Siegelman ES, Huang AB, et al. Adrenal masses:
100. Annema JT, Veselic M, Rabe KF. EUS-guided FNA of centrally correlation between CT attenuation value and chemical shift
located lung tumours following a non-diagnostic bron- ratio at MR imaging with in-phase and opposed-phase
choscopy. Lung Cancer. 2005;48:357–361. sequences. Radiology. 1996;200:749–752.
101. Herth FJF, Ernst A. Innovative bronchoscopic diagnostic 126. Musset D, et al. Primary lung cancer staging: prospective com-
techniques: endobronchial ultrasound and electromagnetic parative study of MR imaging with CT. Radiology. 1986;
navigation. Curr Opin Pulm Med. 2005;11:278–281. 160:607–611.
102. Yasufuku K, et al. Endobronchial ultrasound guided trans- 127. Haggar AM, et al. Chest-wall invasion by carcinoma of the lung:
bronchial needle aspiration for staging of lung cancer. Lung detection by MR imaging. AJR Am J Roentgenol. 1987;
Cancer. 2005;50:347–354. 148:1075–1078.
103. Society ATSER. Pretreatment evaluation of non-small-cell lung 128. Padovani B, et al. Chest wall invasion by bronchogenic carci-
cancer. Am J Respir Crit Care Med. 1997;156:320–332. noma: evaluation with MR imaging. Radiology. 1993;187:33–38.
104. Tahara RW, Lackner RP, Graver M. Is there a role for routine 129. Heelan RT, Demas BE, Caravelli JF, et al. Superior sulcus tumors:
mediastinoscopy in patients with peripheral T1 lung cancers? CT and MR imaging. Radiology. 1989;170:637–641.
Am J Surg. 2001;180:488–492. 130. Webb WR. MR imaging in the evaluation and staging of lung
105. Seely JM, et al. T1 Lung cancer: prevalence of mediastinal nodal cancer. Semin Ultrasound CT MR. 1988;9:53–66.
metastases and diagnostic accuracy of CT. Radiology. 1993; 131. Webb WR, Sostman HD. MR imaging of thoracic disease: clini-
186:129–132. cal uses. Radiology. 1992;182:621–630.
106. Dunnick NR, Korobkin M, Francis IR. Adrenal radiology: distin- 132. Thompson BH, Stanford W. MR imaging of pulmonary and
guishing benign from malignant masses. AJR Am J Roentgenol. mediastinal malignancies. Magn Reson Imaging Clin North
1996;167:861–867. Am. 2000;8:729–739.
107. Dunnick NR, Korobkin M. Imaging of adrenal incidentalomas: 133. Blair DN, et al. Normal brachial plexus: MR imaging. Radiology.
current status. AJR Am J Roentgenol. 2002;179:559–568. 1987;165:763–767.
108. Korobkin M, et al. Differentiation of adrenal adenomas form 134. Castagno AA, Shuman WP. MR imaging in clinically suspected
nonadenomas using CT attenuation values. AJR Am J Roentgenol. brachial plexus tumor. AJR Am J Roentgenol. 1987;149:
1996;166:531–536. 1219–1222.
109. Macari M, Rofsky NM, Naidich DP, Megibow AJ. Non-small cell 135. Gupta NC, et al. Solitary pulmonary nodules: detection of
lung carcinoma: usefulness of unenhanced helical CT of the malignancy with PET with 2-(F-18)-fluoro-2-deoxy-D-glucose.
adrenal glands in an unmonitored environment. Radiology. Radiology. 1992;184:441–444.
1998;209: 807–812. 136. Gupta NC, Maloof J, Gunel E. Probability of malignancy in soli-
110. Korobkin M, et al. Delayed enhanced CT for differentiation of tary pulmonary nodules using fluorin-18-FDG and PET. J Nucl
benign from malignant adrenal masses. Radiology. 1996;200: Med. 1996;37:943–948.
737–742. 137. Patz EF Jr, Lowe VJ, Hoffman JM. Focal pulmonary abnormali-
111. Blake MA, et al. Distinguishing benign from malignant adrenal ties; evaluation with F-18 fluorodeoxyglucose PET scanning.
masses: multi-detector row CT protocol with 10-minute delay. Radiology. 1993;188: 487–490.
Radiology. 2006;238:578–585. 138. Patz EF Jr, et al. Thoracic nodal staging with PET imaging with
112. Silvestri GA, Littenberg B, Colice GL. The clinical evaluation for 18FDG in patients with bronchogenic carcinoma. Chest.
detecting metastatic lung cancer: a meta-analysis. Am R Respir 1995;108:1617–1621.
Crit Care Med. 1995;152:225–230. 139. Duhaylongsod FG, et al. Detection of primary and recurrent
113. Tanaka K, et al. Extrathoracic staging is not necessary for non- lung cancer by means of F-18 fluorodeoxyglucose positron
small-cell lung cancer with clinical stage T1–2 N0. Ann Thorac emission tomography (FDG PET). J Thorac Cardiovasc Surg.
Surg. 1999;68:1039–1042. 1995;110:130–140.
114. Win T, et al. The value of performing head CT in screening for 140. Cook GJR. Maisey MN. The current status of clinical PET imag-
cerebral metastases in patients with potentially resectable non- ing. Clin Radiol. 1996;51:603–613.
141. Gambhir SS, et al. Decision tree sensitivity analysis for cost- 163. Weber WA, Peterson V, Schmidt B, et al. Positron emission
effectiveness of FDG-PET in the staging and management of tomography in non-small-cell lung cancer: predication of
non-small-cell lung carcinoma. J Nucl Med. 1996;37:1428–1436. response to chemotherapy by quantitative assessment of glu-
142. Hughes JMB. 18F-Fluorodeoxyglucose PET scans in lung cancer. cose use. J Clin Oncol. 2003;21:2651–2657.
Thorax. 1996;51(suppl 2):S16–S22. 164. Ryu JS, et al. FDG-PET in staging and restaging non-small cell
143. Bury T, et al. Evaluation of the solitary pulmonary nodule by lung cancer after neoadjuvant chemoradiotherapy: correlation
positron emission tomography imaging. Eur Respir J. 1996;9: with histopathology. Lung Cancer. 2002;35:179–187.
410–414. 165. Stroobants S, Verschakelen J, Vansteenkiste J. Value of FDG-PET
144. Gordon BA, Flanagan FL, Dehdashti F. Whole-body positron in the management of non-small cell lung cancer. Eur J Radiol.
emission tomography: normal variations, pitfalls, and technical 2003;45:49–59.
considerations. AJR Am J Roentgenol. 1997;169:1675–1679. 166. Cerfolio RJ, et al. The maximum standardized uptake values on
145. Pieterman RM, et al. Preoperative staging of non-small-cell lung positron emission tomography of a non-small cell lung cancer
cancer with positron-emission tomography. N Engl J Med. predict stage, recurrence, and survival. J Thorac Cardiovasc Surg.
2000;343:254–261. 2005;130:151–159.
146. Reed CE, et al. Results of the American College of Surgeons 167. Port JL, et al. Positron emission tomography scanning poorly
Oncology Group Z0050 Trial: the utility of positron emission predicts response to preoperative chemotherapy in non-small
tomography in staging potentially operable non-small cell lung cell lung cancer. Ann Thorac Surg. 2004;77:254–259.
cancer. J Thorac Cardiovasc Surg. 2003;126:1943–1951. 168. Patz EF Jr, Connolly JE, Herndon JE II. Prognostic value of tho-
147. van Tinteren H, et al. Effectiveness of positron emission tomog- racic FDG PET imaging after treatment for non-small cell lung
raphy in the preoperative assessment of patients with suspected cancer. AJR Am J Roentgenol. 2000;174:769–774.
non-small-cell lung cancer: the PLUS multicentre randomized 169. Pandit N, et al. Prognostic value of (F-18)FDG-PET imaging in
trial. Lancet. 2002;359:1388–1392. small cell lung cancer. Eur J Nucl Med Mol Imaging. 2003;30:
148. Sachs S, Bilfinger. The impact of positron emission tomography 78–84.
on clinical decision making in a university-based multidiscipli- 170. Bury T, Corhay JL, Duysinx B, et al. Value of FDG-PET in detect-
nary lung cancer practice. Chest. 2005;128:698–703. ing residual or recurrent non-small cell lung cancer. Eur Respir J.
149. Viney RC, et al. Randomized controlled trial of the role of 1999;14:1376–1380.
positron emission tomography in the management of stage 171. Simon GR, Wagner H. Small cell lung cancer. Chest. 2003;123:
I and II non-small-cell lung cancer. J Clin Oncol. 2004;22: 259S–271S.
2357–2362. 172. Stahel RA, Ginsberg R, Havemann K, et al. Staging and prognos-
150. Gould MK, et al. Test performance of positron emission tomog- tic factors in small cell lung cancer: a consensus report. Lung
raphy and computed tomography for mediastinal staging in Cancer. 1989;5:119–124.
patients with non-small-cell lung cancer: a meta-analysis. Ann 173. Stanley RJ. Screening revisited. AJR Am J Roentgenol. 2004;183:
Intern Med. 2003;139:879–892. 1537.
151. Boiselle PM, et al. Postpneumonectomy syndrome: another 174. Zwischenberger JB, Cox CS Jr. Staging and surgery for lung can-
twist. J Thorac Imaging. 1997;12:209–211. cer. J Thorac Imaging. 1991;84:641–648.
152. Dwamena BA, et al. Metastases from non-small cell lung cancer: 175. Spira A, Ettinger DS. Multidisciplinary management of lung
mediastinal staging in the 1990’s: meta-analytic comparison of cancer. N Engl J Med. 2004;350:379–392.
PET and CT. Radiology. 1999;213:530–536. 176. Meyer JA, Comis RL, Ginsber SJ. Selective surgical resection in
153. Birim O, et al. Meta-analysis of positron emission tomographic small cell carcinoma of the lung. J Thorac Cardiovasc Surg. 1982;
and computed tomographic imaging in detecting mediastinal 84:641–648.
lymph node metastases in nonsmall cell lung cancer. Ann 177. Meyer JA, Comis RL, Ginsber SJ, et al. The prospect of disease
Thorac Surg. 2005;79:375–382. control by surgery combined with chemotherapy in stage I and
154. Cerfolio RJ, et al. The accuracy of integrated PET-CT compared II small cell carcinoma of the lung. Ann Thorac Surg. 1983;36:
with dedicated PET alone for the staging of patients with 37–41.
nonsmall cell lung cancer. Ann Thorac Surg. 2004;78: 178. Ginsberg RJ. Surgery and small cell lung cancer: an overview.
1017–1023. Lung Cancer. 1989;5:22–236.
155. Metser U, et al. Tumor lesion detection: when is integrated 179. Davis S, Wright PW, Schulman SF, et al. Long term survival in
positron emission tomography/computed tomography more small cell carcinoma of the lung: a population experience. J Clin
accurate than side-by-side interpretation of positron emission Oncol. 1985;3:80–91.
tomography and computed tomography? J Comput Assist 180. Schumacher T, Brink L, Mix M, et al. FDG-PET imaging for the
Tomogr. 2005;29:554–559. staging and follow-up of small cell lung cancer. Eur J Nucl Med.
156. Heelan RT. Invited commentary. Ann Thorac Surg. 1995;60: 2001;28:483–488.
1573. 181. Hauber HP, Bohuslavizki KH, Liund CH, et al. Positron emis-
157. Roman CD, Martin WH, Delbeke D. Incremental value of sion tomography in the staging of small-cell lung cancer: a pre-
fusion imaging with integrated PET-CT in oncology. Clin Nucl liminary study. Chest. 2001;119:950–954.
Med. 2005;30:470–477. 182. Lee CB, et al. Current and evolving treatment options for limited
158. Halpern BS, et al. Presurgical staging of non-small cell lung stage small cell lung cancer. Curr Opin Oncol. 2006:18:162–172.
cancer: positron emission tomography, integrated positron 183. Jett J, et al. Guidelines on treatment of stage IIIB non-small cell
emission tomography/CT, and software image fusion. Chest. lung cancer. Chest. 2003;123:221S–225S.
2005;128:2289–2297. 184. Pitz CCM, et al. Results of surgical treatment of T4 non-small
159. Mavi A, et al. Fluorodeoxyglucose-PET in characterizing solitary cell lung cancer. Eur J Cardiovasc Thorac Surg. 2003;24:
pulmonary nodules, assessing pleural disease, and the initial 1013–1018.
staging, restaging, therapy planning and monitoring response of 185. Scott WJ, Howington J,. Movsas B. Treatment of stage II non-
lung cancer. Radiol Clin North Am. 2005;43:1–21. small cell lung cancer. Chest. 2003;123:188S–201S.
160. Juweid ME, Cheson BD. Current concepts: positron-emission 186. Robinson LA, Wagner H, Ruckdeschel JC. Treatment of stage
tomography and assessment of cancer therapy. N Engl J Med. IIIA non-small cell lung cancer. Chest. 2003;123:202S–220S.
2006;354:496–507. 187. Rami-Porta R, Wittekind C, Goldstraw P. Complete resection
161. Vansteenkiste J, et al. Positron emission tomography in prog- in lung cancer surgery: proposed definition. Lung Cancer.
nostic and therapeutic assessment of lung cancer: systematic 2005;49:25–33.
review. Lancet Oncol. 2004;5:531–540. 188. Watanabe S, et al. Recent results of postoperative mortality for
162. Vansteenkiste J, et al. Potential use of FDG-PET after induction surgical resections in lung cancer. Ann Thorac Surg. 2004;
chemotherapy in surgically staged IIIa-N2 non-small-cell lung 78:999–1003.
cancer: a prospective pilot study: the Leuven Lung Cancer 189. Bach PB, et al. The influence of hospital volume on survival
Group. Ann Oncol. 1998;9:1193–1198. after resection for lung cancer. N Engl J Med. 2001;345:181–188.
190. DAmico TA, et al. Molecular biologic substaging of stage I lung 209. Swensen S, et al. Consensus statement: CT screening for lung
cancer according to gender and histology. Ann Thorac Surg. cancer. J Thorac Imaging. 2005;20:321.
2000;69:882–886. 210. Obuchowski N, et al. Ten criteria for effective screening: their
191. McKenna RJ Jr, Houck W, Fuller CB. Video-assisted thoracic sur- application to multislice CT screening for pulmonary and col-
gery lobectomy: experience with 1,100 cases. Ann Thorac Surg. orectal cancers. AJR Am J Roentgenol. 2001;176:1357–1362.
2006;81:421–426. 211. Bach PB, et al. Screening for lung cancer: a review of the current
192. Nomori H, et al. What is the advantage of a thoracoscopic literature. Chest. 2003;123:722S–782S.
lobectomy over a limited thoracotomy procedure for lung can- 212. Eddy RJ. Cost-effectiveness of CT scanning compared with
cer surgery. Ann Thorac Surg. 2001;72:879–884. mediastinoscopy in the preoperative staging of lung cancer. Can
193. Daniel TM. A proposed diagnostic approach to the patient with Assoc Radiol J. 1989;40:189–193.
the subcentimeter pulmonary nodule: techniques that facilitate 213. Black WC, Welch HG. Screening for disease. AJR Am J
video-assisted thoracic surgery excision. Semin Thorac Cardiovasc Roentgenol. 1997;168:3–11.
Surg. 2005;17:115–122. 214. Oken MM, et al. Baseline chest radiograph for lung cancer
194. Detterbeck FC. Changes in the treatment of Pancoast tumors. detection in the randomized prostate, lung, colorectal and
Ann Thorac Surg. 2003;75:1990–1997. ovarian cancer screening trial. J Natl Cancer Inst.
195. Alifano M, et al. Surgical treatment of superior sulcus tumors: 2005;97:1832–1839.
results and prognostic factors. Chest. 2003;124:996–1003. 215. Henschke CI, Yankelevitz DF, Kostis WJ. CT screening for lung
196. Gandhi NS, Dupuy DE. Image-guided radiofrequency ablation cancer. Radiol Clin North Am. 2004;42:747.
as a new treatment option for patients with lung cancer. Semin 216. Henschke CI, et al. An update of CT screening for lung cancer.
Roentgenol. 2005;40:171–181. Semin Ultrasound CT MRI. 2005;26:348–356.
197. Kotaro Y, Susumu K, Yoshifumi S, et al. Thoracic tumors treated 217. Sone S, Yang S-G, Honda T, et al. Results of three-year mass
with CT-guided radiofrequency ablation: initial experience. screening program for lung cancer using a mobile low-
Radiology. 2004;231:850–857. dose spiral computed tomography scanner. Br J Radiol.
198. Steinke K, et al. Radiologic appearance and complications of 2001;84:25–32.
percutaneous computed tomography-guided radiofrequency- 218. Diederich S, et al. Screening for early lung cancer with low-dose
ablated pulmonary metastases from colorectal carcinoma. spiral computed tomography: results of annual follow-up exami-
J Comput Assist Tomogr. 2003;27:750–757. nations in asymptomatic smokers. Eur Radiol. 2004;14: 691–702.
199. Steinke K, King J, Glenn D, et al. Percutaneous radiofrequency abla- 219. Gohagan JK, et al. Final results of the Lung Screening Study, a
tion of lung tumors with expandable needle electrodes: tips from randomized feasibility study of spiral CT versus chest X-ray
preliminary experience. AJR Am J Roentgenol. 2004;183: 605–611. screening for lung cancer. Lung Cancer. 2005;47:9–15.
200. Lee JM, et al. Percutaneous radiofrequency ablation for inopera- 220. Flieder DB, Vasquez M, Carter D, et al. Pathologic findings of
ble non-small cell lung cancer and metastases: preliminary lung tumors diagnosed on baseline CT screening. Am J Surg
report. Radiology. 2004;230:125–134. Pathol. 2006;30:606–613.
201. Bojarski JD, Dupuy DE, MayoSmith WW. CT imaging findings 221. Henschke CI, et al. CT screening for lung cancer: significance of
of pulmonary neoplasms after treatment with radiofrequency diagnoses in its baseline cycle. Clin Imaging. 2006;30:11–15.
ablation: results in 32 tumors. AJR Am J Roentgenol. 2005;185: 222. Patz Jr EF, Swensen S. Herndon JE II. Estimate of lung cancer
466–471. mortality from low-dose spiral computed tomography screening
202. Herrera LJ, et al. Radiofrequency ablation of pulmonary malig- trials: implications for current mass screening recommenda-
nant tumors in nonsurgical candidates. J Thorac Cardiovasc Surg. tions. J Clin Oncol. 2004;22:2202–2206.
2003;125:929–937. 223. Gur D. Lung cancer screening: radiology’s opportunity here and
203. Dupuy DE, et al. Radiofrequency ablation followed by conven- now: editorial. Radiology. 2006;238:395–397.
tional radiotherapy for medially inoperable stage 1 non-small 224. Diederich S, Wormanns D. Impact of low-dose CT on lung
cell lung cancers. Chest. 2006;129:738–745. cancer screening. Lung Cancer. 2004;45:S13–S19.
204. Steinke K, et al. Percutaneous pulmonary radiofrequency abla- 225. Crestanello JA, et al. Thoracic surgical operations in patients
tion: difficulty achieving complete ablations in big lung lesions. enrolled in a computed tomographic screening trial. J Thorac
Br J Radiol. 2003;76:742–745. Cardiovasc Surg. 2004;128:254–259.
205. Kaneko M, et al. Peripheral lung cancer: screening and detection 226. Pinsky PF, et al. Diagnostic procedures after a positive spiral
with low-dose spiral CT versus radiography. Radiology. 1996; computed tomography lung carcinoma screen. Cancer. 2005;
201:798–802. 103:157–163.
206. Sone S, et al. Mass screening for lung cancer with mobile 227. Henschke CI, Yankelevitz DF, Naidich DP, et al. CT screening for
spiral computed tomography scanner. Lancet. 1998;351: lung cancer: suspiciousness of nodules according to size on
1242–1245. baseline scans. Radiology. 2004;231:164–168.
207. Henschke CI, Yankelevitz DF, McCauley D, et al. Early lung can- 228. Aberle D, et al. Imaging an cancer: research strategy of the
cer action project: overall design and findings from baseline American College of Radiology Imaging Network. Radiology.
screenings. Lancet. 1999;354:99–105. 2005;235:741–751.
208. Henschke CI, Yankelevitz DF, Libby DM, et al. The International 229. Henschke CI, et al. Early lung cancer action project: initial find-
Early Lung Cancer Action Program Investigators. Survival of ings on repeat screenings. Cancer. 2001;92:153–159.
patients with stage I lung cancer detected on CT screening. 230. McWilliams A, et al. Lung cancer screening: a different para-
N Engl J Med. 2006;355:1763–1771. digm. Am J Respir Crit Care Med. 2003;168:1167–1173.
Diffuse Lung Disease 8
GENERAL PRINCIPLES AND METHODOLOGY 672 CLINICAL UTILITY OF COMPUTED

TOMOGRAPHY 748
NORMAL LUNG ANATOMY 676 Diagnostic Sensitivity and Specificity of Computed
The Lung Interstitium 676 Tomography Compared with Chest
Bronchi and Pulmonary Vessels 677 Radiography 748
Secondary Pulmonary Lobules and Acini 678 High-Resolution Versus Thick-Section Computed
Lung Attenuation 682 Tomography Imaging 749
Diagnostic Accuracy of Computed Tomography
PATTERNS OF ABNORMALITY ON HIGH- Compared with Chest Radiography 750
RESOLUTION COMPUTED TOMOGRAPHY 682 Assessment of Disease Activity: Significance of
Linear and Reticular Opacities 682 Ground-Glass Opacities 752
Nodular Opacities 685 High-Resolution Computed Tomography–Guided
Ground-Glass Opacity 690 Lung Biopsy 755
Consolidation 691 High-Resolution Computed Tomography Findings
Decreased Attenuation 693 Sufficiently Diagnostic to Obviate Lung
Biopsy 757
CHRONIC INFILTRATIVE LUNG DISEASE 697
Diseases Characterized Primarily by Linear
and Reticular Opacities 697 Despite the well-established role of chest radiography in
Diseases Characterized Primarily by Nodular accurately and inexpensively displaying a wide range of pul-
Opacities 707 monary parenchymal pathology, equally well established
Diseases Characterized Primarily by limitations have been documented (1–3). In one study of
Ground-Glass Opacity 714 458 patients with histologically confirmed infiltrative lung
Diseases Characterized Primarily by disease, 44, or nearly 10%, had normal prebiopsy chest radi-
Consolidation 725 ographs (1). In a second study (2), nearly 16% of patients
Diseases Characterized Primarily who had pathologic proof of interstitial lung disease had
by Cysts 728 normal chest radiographs. When present, the radiographic
abnormalities are frequently nonspecific (4–6), and consid-
EMPHYSEMA 734 erable interobserver variation is found in the interpretation
of the findings, even among expert readers (4). A number of
ACUTE LUNG DISEASE IN THE studies have shown that CT, particularly high-resolution CT
IMMUNOCOMPROMISED PATIENT 738 (HRCT), is superior to chest radiography in the detection of
Pneumocystis jiroveci Pneumonia 740 parenchymal abnormalities, is more accurate in the differ-
Cytomegalovirus Pneumonia 743 ential diagnosis, and is subject to considerably less interob-
Invasive Aspergillosis 745 server variation in its interpretation (3,5–7).
Septic Emboli 747 Optimal assessment of diffuse lung disease on CT is
Noninfectious Complications 748 obtained by minimizing volume averaging by using thin
collimation (0.5- to 2-mm thick sections) and by enhancing The essential elements of HRCT are acquisition of thin
the contrast between adjacent structures by using a high sections (0.5 to 2 mm) and use of a high–spatial frequency
spatial frequency reconstruction algorithm. The combina- (edge-enhancing) reconstruction algorithm. Further im-
tion of thin-section and high-frequency reconstruction algo- provement in spatial resolution can be obtained with the
rithm is known as high-resolution computed tomography use of targeted reconstructions with field of view (FOV)
(HRCT). HRCT provides a depiction of lung morphology restricted to individual lungs (9). Routinely, images are
comparable to that of gross (macroscopic) lung pathology. acquired with the shortest possible scan time by using
Since it was first described in 1982 (8), until recently, HRCT a 512 512 matrix, with milliamperage (mA) typically
was typically obtained at 10- to 20-mm intervals through varying between 200 and 300, depending on the patient’s
the chest. The advent of multidetector CT scanners in the size. It should be noted that considerable dose reduction,
late 1990s has allowed volumetric imaging of the entire to as low as 40 mAs, still allows acquisition of inter-
lung by using high-resolution CT technique. Volumetric pretable images of the lung parenchyma on HRCT
imaging allows detection of abnormalities that might have (Fig. 8-1) (10). Although low-dose HRCT technique allows
been missed between HRCT sections and high-resolution better assessment of lung parenchyma than the chest radi-
multiplanar and three-dimensional reconstructions. Advan- ograph (7), it has been reported that mild ground-glass
tages derived from volumetric imaging, however, must be attenuation, emphysema, and fibrosis detectable on con-
weighed against added reconstruction and postprocessing ventional-dose HRCT can be missed on low-dose HRCT
times, as well as storage requirements for as many as several (Fig. 8-2) (7,10,11). As the frequency with which these
hundred additional images. Consideration must also be limitations occur clinically has not been definitively estab-
given to any potential added radiation dose, especially in lished, higher milliamperage settings are recommended in
young women and children. the initial assessment of most patients. However, reduced-
In this chapter, we review the CT techniques that are milliamperage HRCT should be used whenever feasible in
appropriate for the evaluation of diffuse lung disease, the the evaluation of response to treatment or disease progres-
various patterns of abnormality seen on CT, and the role of sion, especially in patients in whom radiation dose is a
CT in the assessment of these patients. The role of CT in major concern. This has become an especially important
the evaluation of focal lung disease was considered in issue as multiple follow-up CT studies are now increas-
Chapter 6. ingly used to monitor patients with diffuse lung disease.
Optimal technical parameters for individual images
have been described (7,9–12) but no general agreement
GENERAL PRINCIPLES AND exists as to what constitutes an acceptable high-resolution
METHODOLOGY study. Individual reports differ strikingly in overall tech-
nique, especially in determining the number and levels of
The diagnostic efficacy of CT in the evaluation of diffuse necessary scans and the indications for both prone and
lung disease is inextricably tied to scan technique. supine images as well as those obtained in expiration.
Accurate diagnosis requires that a variety of protocols that Two major alternative approaches are currently used in
specifically reflect a wide range of clinical indications be the assessment of patients with diffuse lung disease: HRCT
established. performed at preselected levels and volumetric HRCT
A B
Figure 8-1 Low-dose high-resolution CT (HRCT). A: HRCT (1.5-mm collimation) performed at the level of the aortic arch by using 120-kV,
200-mA, and 2-second scan. Mild parenchymal abnormalities are present in the subpleural lung regions, consisting of irregular interfaces
and small irregular lines. B: HRCT (1.5-mm collimation) obtained at the same level by using 120-kV, 40-mA, and 2-second scan shows the
findings equally well. Increased noise is present posteriorly in the paraspinal region, but this does not affect the diagnostic quality of the
image. The patient was a 71-year-old man with idiopathic pulmonary fibrosis.
Chapter 8: Diffuse Lung Disease 673
A B
Figure 8-2 Low-dose high-resolution CT (HRCT) in emphysema. A: HRCT (1-mm collimation) performed above the level of the aortic arch
by using 120 kV and 200 mA shows mild centrilobular emphysema (arrows). B: HRCT (1-mm collimation) obtained at the same level by using
120 kV and 40 mA shows considerable increase in image noise and no definite evidence of emphysema.
through the chest. In the first approach, HRCT is performed exposure and, in some cases, cost (16). The main disadvan-
at preselected levels chosen to maximize evaluation of the tages are that focal parenchymal abnormalities may be
different lung zones. There is no consensus, however, about overlooked, the patchy distribution of many diffuse infil-
the number of HRCT scans required for an adequate assess- trative lung diseases may not be appreciated, and small
ment of the lung parenchyma. Different investigators have nodules may be missed between high-resolution sections
used HRCT scans at 1-cm, 2-cm, and even 4-cm intervals (13–17). The second approach, possible since the advent of
with the patient supine, prone, or both supine and prone multidetector CT scanners, is to perform volumetric HRCT
(13–15). We consider scans obtained at 1-cm intervals, through the entire chest (18,19). This approach allows
from the lung apices to the bases, to be the most appropri- assessment of the entire lung parenchyma and therefore
ate scanning protocol for HRCT performed at preselected minimizes the risk of a falsely negative examination.
levels because it allows complete sampling of the lung and Volumetric CT also allows multiplanar reconstructions
lung disease, regardless of its distribution. The main advan- (Fig. 8-3) and generation of maximum (MIPs) and
tage of HRCT at preselected levels is to reduce radiation minimum (MinIPs) intensity projection images. The latter
A B
Figure 8-3 Cross-sectional and coronal images in a 49-year-old patient with sarcoidosis. A: A 1-mm high-resolution CT section through
the upper lobes obtained on a multidetector CT scanner shows nodular thickening of the bronchi (large straight arrow), vessels (arrowhead ),
major interlobar fissures (small straight arrows), and upwardly displaced right minor fissure (curved arrows). B: Coronal reformation demon-
strates the predominant distribution of the abnormalities in the middle and upper lung zones. It also shows nodular thickening of the
interlobular septa (arrowheads) and interlobar fissure (arrow).
have the advantage of retaining high spatial resolution radiation dose (21a), use of low-dose scanning coupled
while preserving the anatomic cross-sectional orientation with dose-modulation techniques available on newer CT
derived from thicker sections (Fig. 8-4). MIPs images have scanners can be used to minimize this problem. As a conse-
been shown to be particularly helpful in the detection quence, in our judgment, initial studies in patients
of small nodules, and MinIPs images, in the detection referred for suspected diffuse lung disease should be
of subtle emphysema (20,21). Although volumetric HRCT performed with volumetric data acquisition. In distinc-
has been reported to result in a considerably greater tion, follow-up examinations are best performed with a
A B
C D
Figure 8-4 Diffuse infiltrative lung disease: evaluation with maxi-

mum (MIP) and minimum (MinIP) intensity projection images. A:
A 1-mm high-resolution CT (HRCT) section through the lower lobes
shows pattern of subtle mosaic attenuation with suggestion of focal
air trapping within the middle lobe and lingula and the posterior
segments of both lower lobes. No obvious nodules or tree-in-bud
pattern identified. B, C: MIP and MinIP images, respectively,
obtained by using five contiguous 1-mm sections centered at the
same level as shown in A. On the MIP image, a pattern of diffuse
nodules with a tree-in-bud configuration is apparent (curved arrows
in B). Focal air trapping is slightly more apparent in the posterobasi-
lar segments of the lower lobes (straight arrows in B). Note that
whereas the nodules all but disappear on the MinIP image, mosaic
attenuation is much more apparent (arrows in C). These findings are
all consistent with infectious bronchiolitis with accompanying air
trapping, identifiable even without acquiring scans in expiration.
D, E: HRCT and corresponding MinIP image in a different patient
than in A–C. Note that in this case, whereas diffuse emphysema is
apparent on the HRCT image, visual assessment of the extent of
E emphysema is easier on the corresponding MinIP image.
dedicated high-resolution technique by using low-dose diagnosed in patients with large or small airway obstruc-
technique, whenever feasible, to minimize subsequent radi- tion or emphysema. In most lung regions of normal sub-
ation exposure. An optimal approach to high-resolution jects, lung parenchyma increases uniformly in attenuation
imaging ultimately depends on the clinical situation to during expiration, but in the presence of air trapping, lung
be evaluated. parenchyma remains lucent on expiration and shows little
The risk of cancer from radiation is influenced by age, change in volume (Fig. 8-5). Some investigators obtain
being greatest in children and young adults, particularly expiratory scans routinely in all patients who have HRCT,
women, and decreasing considerably in patients older whereas others limit their use to patients with inspiratory
than 50 years (16). Therefore, in younger patients, it is rec- scan abnormalities or suspected obstructive lung disease.
ommended that HRCT scans be obtained by using 1-mm Expiratory HRCT scans may be obtained during suspended
collimation at 1- or 2-cm intervals from the thoracic inlet respiration after forced exhalation (postexpiratory CT) or
to the diaphragm, prospectively reconstructed by using a during forced exhalation (dynamic expiratory CT) (24,25).
high-spatial-frequency algorithm. In patients older than The expiratory scans may be obtained at selected levels,
50 years and in patients in whom focal lung disease or typically three to five, or through the entire lung using
lung metastases are suspected, we routinely perform volu- HRCT technique on multidetector CT scanners.
metric high-resolution CT through the chest by using No predetermined “best” windows and levels exist
0.5 to 1 mm collimation. for imaging the lung. Precise settings are often a matter of
Scans are routinely performed with the patient supine. subjective preference. As a general guide, we recommend
In patients with known or suspected occupational expo- a window level of 600 to 700 HU and a window width
sure to asbestos, scans of the prone patient are often neces- of 1,000 to 1,500 HU for the assessment of the lung
sary to distinguish reversible gravity-dependent density parenchyma. However, as illustrated in this chapter, a wide
and atelectasis from fixed structural abnormalities in range of settings is used, depending on the particular nature
dependent portions of the lung bases, where pulmonary of the pathology to be illustrated. Narrowing the window
fibrosis is most likely to occur (13,17). Although some width enhances visual resolution. This may be particularly
investigators (6,12) obtain HRCT in the prone position helpful when assessing focal increases and decreases in lung
only when dependent lung density is problematic, others attenuation, as may be seen in patients with emphysema.
(13) recommend prone scanning routinely. It should be Conversely, wide window widths are often more informa-
noted that dependent density is a diagnostic dilemma only tive when the disease process occupies a large area and
in patients with normal lungs or subtle parenchymal involves a wide range of tissue-density alterations, as may
abnormalities. In patients with diffuse lung disease who occur with complex pleural–parenchymal disease. In these
have abnormalities visible on the chest radiograph, dep- cases, the panoramic view afforded by a wide window width
endent density is seldom a diagnostic problem, and there- improves visual comprehension (see Chapter 1 for a more
fore prone scans are usually unnecessary (22). detailed discussion). It should be emphasized that when-
As an adjunct to routine inspiratory images, expiratory ever possible, similar windows and levels should be used in
HRCT scans have proved useful in the evaluation of evaluating follow-up CT studies to minimize the risk of
patients with a variety of obstructive lung diseases (23–25). inadvertently misdiagnosing disease progression and/or
On expiratory scans, focal or diffuse air trapping may be regression.
A B
Figure 8-5 Air trapping on expiratory high-resolution CT (HRCT) in a patient with bronchiolitis obliterans. A: Inspiratory HRCT scan shows
inhomogeneous appearance of the lungs with localized areas of slightly decreased attenuation and vascularity. B: Expiratory scan shows
extensive bilateral air trapping throughout the lower lobes, middle lobe, and lingula.
NORMAL LUNG ANATOMY The Lung Interstitium

The lung is supported by a network of connective tissue
The lung is composed of anatomic units that demonstrate
fibers, termed the lung interstitium. The interstitium must
similar architecture at progressively smaller sizes, as
be strong enough to maintain the patency of alveoli,
described by fractal geometry; from large to small, these
airways, and vessels, but at the same time must be thin
units are lungs, lobes, segments, subsegments, secondary
enough to allow adequate gas exchange between air in the
lobules, and acini. At each level, these are organized
alveolar spaces and blood in pulmonary capillaries. The
around central or “core” supporting structures (airways
interstitium is organized into three fiber systems, the peri-
and pulmonary arteries) and within peripheral or “shell”
bronchovascular interstitium, the subpleural interstitium,
supporting structures (pleura and connective tissue
and the intralobular interstitium (29). Together, these
septa).
three form a continuous fiber skeleton that serves to
The lobes are marginated to a variable degree by
support lung parenchyma between the hila centrally and
the interlobar fissures. In normal subjects, three lobes
the pleural surfaces peripherally (Fig. 8-6), and which, to
(upper, middle, and lower) are present on the right, and
a large extent, is replicated from the level of the lobes
two (upper and lower) are present on the left. Lobes can
to the level of the secondary pulmonary lobules.
be found on CT by identifying the interlobular fis-
The peribronchovascular interstitium is a system of
sures or the avascular plane that mark their positions, or
fibers that invests bronchi and pulmonary arteries and
when fissures are incomplete or poorly localized, by
serves to support the “core” or medullary structures of the
identifying bronchial branches. Lobar bronchi have a
lung (Fig. 8-6). In the parahilar regions, the peribron-
consistent branching pattern in the large majority of
chovascular interstitium forms a strong connective tissue
patients; lobar bronchial anomalies are rare. Pulmonary
sheath that surrounds large bronchi and arteries; it contin-
artery branching patterns are more variable, and there-
ues into the lung periphery, investing centrilobular arteries
fore, less valuable in identifying individual pulmonary
and bronchioles, and continues to the level of the alveolar
lobes.
ducts and sacs (29). The subpleural interstitium is located
Pulmonary segments are less well marginated than
beneath the visceral pleura and envelops the lung in a
lobes and do not have easily definable boundaries. Indi-
fibrous sac from which connective tissue septa penetrate
vidual segments are best localized on the basis of segmen-
the lung parenchyma (Fig. 8-6). These septa include the
tal bronchi and by noting the relatively consistent location
interlobular septa, which are described in detail later. The
of segments within lobes and in relation to lobar fissures
intralobular interstitium is a network of thin fibers that
(26–28). Arterial anatomy is less valuable in identifying
forms a fine connective tissue mesh in the walls of alveoli,
segments, as arterial branching is somewhat more variable
and thus bridges the gap between the peribronchovascular
than bronchial branching (see Chapter 3). As with pul-
interstitium surrounding vessels and bronchi in the center
monary segments, subsegments are best identified on the
of lobules, and the interlobular septa and subpleural inter-
basis of bronchial anatomy. At a subsegmental level, lung
stitium in the lobular periphery (Fig. 8-6) (29). These
is composed of secondary pulmonary lobules and acini,
fibers keep capillaries and alveoli open for functional gas
which are described in detail later.
exchange.
Figure 8-6 Components of the lung interstitium. Taken

together, the peribronchovascular interstitium and
centrilobular interstitium correspond to the “axial fiber
system” described by Weibel (29). The subpleural inter-
stitium and interlobular septa correspond to Weibel’s
“peripheral fiber system.” The intralobular interstitium
is roughly equivalent to the “septal fibers” described by
Weibel.
Bronchi and Pulmonary Vessels particularly in dependent lung regions. Although the
presence of bronchi larger than their adjacent arteries is
Within the lung parenchyma, bronchi and pulmonary often assumed to indicate the presence of bronchial dilata-
arteries are closely associated and branch in parallel. Each tion, or bronchiectasis, bronchi may appear larger than
bronchus is positioned adjacent to a pulmonary artery of adjacent arteries in a significant number of normal sub-
similar diameter, and this relation is maintained from the jects. This occurs most commonly when the pulmonary
hila to the level of respiratory bronchioles in the lung artery divides before the adjacent bronchus, in patients
periphery. As indicated earlier, bronchi and arteries are scanned at altitude, and when using low window
encased by a network of fibers, the peribronchovascular levels (e.g., 450 HU rather than 700 HU) (24,33,34).
interstitium or bronchovascular sheath, which also ex- Bronchi may also appear spuriously enlarged when high-
tends from the pulmonary hila into the peripheral lung. In resolution prone images are obtained through the lung
addition to the artery and bronchus, this sheath contains bases in otherwise normal subjects. At least one bronchus
amorphous collagen, lymphatics, and small lymph nodes, with a diameter greater than the adjacent pulmonary artery
ranging in size from 1 mm in the lung periphery to 5 to has been reported in 7% of patients scanned at sea level
10 mm near the hila. It should be noted that lymphatics and in 59% of normal subjects scanned at 1,600 m
do not extend beyond the terminal bronchioles. altitude (Denver) (24,34).
Because some lung diseases produce thickening of the The outer walls of visible pulmonary artery branches
peribronchovascular interstitium in the central or parahilar should form a smooth and sharply defined interface with
lung, in relation to large bronchi and pulmonary vessels the surrounding lung, whether they are seen in cross
(Fig. 8-3), it is important to be aware of the normal CT section or along their length. The walls of large bronchi,
appearances of the parahilar bronchi and pulmonary ves- outlined by lung on one side and by air in the bronchial
sels (Fig. 8-7). (Interested readers are referred to Chapter 3 lumen on the other, normally appear to be smooth and of
for an in-depth discussion of normal bronchi.) uniform thickness. Thickening of the peribronchial and
When imaged at an angle to their longitudinal axis, cen- perivascular interstitium can result in irregularity of the
tral pulmonary arteries normally appear as rounded or interface between arteries and bronchi and the adjacent
elliptical opacities on HRCT, accompanied by uniformly lung (35,36).
thin-walled bronchi of similar shape (30–32). When Assessment of bronchial wall thickness on HRCT is
imaged along their axis, bronchi and vessels should appear quite subjective and is dependent on the window settings
roughly cylindrical, or show slight tapering as they branch, used (37,38). Also, because the apparent thickness of the
depending on the length of the segment that is visible; bronchial wall represents not only the wall itself, but also
tapering of a vessel or bronchus is most easily seen when the surrounding peribronchovascular interstitium, peri-
a long segment is visible. bronchovascular interstitial thickening results in apparent
The diameter of an artery and its neighboring bronchus bronchial wall thickening (so-called peribronchial cuffing)
should be approximately equal, although vessels may on HRCT.
appear slightly larger than their accompanying bronchus, The wall thickness of conducting bronchi and bronchi-
oles is approximately proportional to their diameter, at
least for bronchi distal to the segmental level. The normal
bronchial wall thickness at the level of the segmental
bronchi by using a window level of 700 HU and a
window width of 1,500 HU is approximately 1 mm. (33).
In general, the thickness of the wall of a bronchus or bron-
chiole less than 5 mm in diameter should be less than one
tenth of its diameter (Table 8-1) (39); however, precise
measurement of the wall thickness of small bronchi or
bronchioles is difficult, as wall thickness approximates
pixel size.
Bronchi and pulmonary arteries divide by dichotomous
branching. Approximately 23 generations of airway
branches are found, from the trachea to alveolar sacs, and
28 generations of pulmonary artery branches. Because a
bronchus is usually recognized only when its walls are
Figure 8-7 Normal appearances of large bronchi (arrows) and visualized, it is not possible to resolve airways with a
arteries (arrowheads). The diameters of vessels and their neighbor- wall thickness of less than 300 mm. This corresponds to
ing bronchi are approximately equal. The outer walls of bronchi bronchi of about 2 mm in diameter and not closer than
and pulmonary vessels are smooth and sharply defined. Bronchi
are not visible within the peripheral 2 cm of lung, although vessels 2 cm to the pleural surface, equivalent to between the
are well seen in this region. seventh and ninth generation of airways (32,40). Rarely,
Secondary Pulmonary Lobules and Acini

TABLE 8-1 The secondary pulmonary lobule, as defined by Miller (41),
RELATION OF AIRWAY DIAMETER TO WALL refers to the smallest unit of lung structure marginated by
THICKNESS connective tissue septa (Figs. 8-9 and 8-10) (41). Secondary
pulmonary lobules are irregularly polyhedral in shape and
Diameter Wall Thickness
somewhat variable in size, measuring approximately 1 to
(mm) (mm)
2.5 cm in diameter in most locations (42–44). In one
Lobar and segmental study, the average diameter of pulmonary lobules meas-
bronchi 5–8 1 ured in several adults ranged from 11 to 17 mm (44). Each
Subsegmental
secondary lobule is supplied by a small bronchiole and
bronchi/bronchiole 1.5–3 0.2–0.3
Lobular bronchiole 1 0.15 pulmonary artery and is variably marginated, in different
Terminal bronchiole 0.7 0.1 lung regions, by connective-tissue interlobular septa that
Acinar bronchiole 0.5 0.05 contain pulmonary vein and lymphatic branches (42). In
this book, we use the terms “secondary pulmonary lobule,”
“pulmonary lobule,” and “lobule” to refer to a secondary
lobule as defined by Miller.
smaller airways approximately 1 mm in diameter can be Secondary pulmonary lobules are made up of a limited
recognized coursing perpendicularly within the plane of a number of pulmonary acini, usually a dozen or fewer,
very thin section by virtue of their lower density and their although the number of lobular acini can vary from 3 to
proximity to an adjacent pulmonary artery, even though 24 (43,45). A pulmonary acinus is defined as the portion
their walls cannot be clearly seen. of the lung parenchyma distal to a terminal bronchiole
Blood vessels with diameters of 300 mm can be visual- and supplied by a first-order respiratory bronchiole or
ized, corresponding to 16th-generation arteries at the level bronchioles and comprising respiratory bronchioles, alve-
of the terminal and most proximal respiratory bronchiolar ducts, alveolar sacs, and alveoli (46). Because respira-
oles. Because vascular structures typically can be seen to tory bronchioles are the largest airways that have alveoli in
within a few millimeters of the pleural surfaces and can be their walls, an acinus is the largest lung unit in which
followed to the level of the secondary pulmonary lobule, all airways participate in gas exchange. Acini are usually
the arterial tree is a useful landmark for defining HRCT described as ranging from 6 to 10 mm in diameter (44,47).
lung anatomy (32,40).
The pulmonary veins follow a course independent of
the bronchial tree and lie between two pairs of bronchi High-Resolution Computed Tomography
and arteries. This position is maintained into the lung of the Secondary Lobule
periphery, where veins are seen within the interlobular
An understanding of secondary lobular anatomy and the
septa (Fig. 8-8).
appearances of lobular structures are keys to the inter-
pretation of HRCT. HRCT can show many features of the
secondary pulmonary lobule in both normal and abnor-
mal lungs, and many lung diseases, particularly interstitial
diseases, produce some characteristic changes in lobular
structures (43,48–51).
Interlobular Septa
Anatomically, secondary lobules are marginated by con-
nective tissue interlobular septa, which extend inward
from the pleural surface. These septa are part of the sub-
pleural interstitium, which extends over the surface of the
lung beneath the visceral pleura (Fig. 8-6) (29). A rich
lymphatic system, referred to as the superficial lymphatic
network, drains the visceral pleura and courses within the
interlobular septa in parallel with septal veins; these ulti-
Figure 8-8 Normal peripheral pulmonary veins, arteries, and
mately lead to lymphatics and nodes within the hila.
interlobular septa. High-resolution CT image demonstrates pul- It should be emphasized that not all interlobular septa
monary veins coursing within interlobular septa (arrowheads) and are equally well defined. The interlobular septa are thickest
centrilobular pulmonary arteries a few millimeters away from the
interlobular septa and pleura (arrows).
and most numerous in the apical, anterior, and lateral
Figure 8-9 A: Anatomy of the secondary pulmonary lob-

ule, as defined by Miller (4). Two adjacent lobules are
shown in this diagram. B: Radiographic anatomy of the
secondary pulmonary lobule. Radiograph of a 1-mm lung
slice taken from the lower lobe. Two well-defined second-
ary pulmonary lobules are visible. Lobules are marginated
by thin interlobular septa (S) containing pulmonary vein (V)
branches. Bronchioles (B) and pulmonary arteries (A) are
centrilobular. Bar 1 cm. (Courtesy of Dr. Harumi Itoh,
Kyoto University, Kyoto, Japan. From Itoh H, Murata K,
Konishi J, et al. Diffuse lung disease: pathologic basis
for the high-resolution computed tomography findings.
B J Thorac Imaging. 1993;8:176–188, with permission.)
aspects of the upper lobes, the anterior and lateral aspects are often visible in the lung periphery in normal subjects,
of the middle lobe and lingula, the anterior and diaphrag- but they tend to be inconspicuous; normal septa are most
matic surfaces of the lower lobes, and along the mediasti- often seen anteriorly (Fig. 8-11), along the mediastinal
nal pleural surfaces (52); thus secondary lobules are best pleural surfaces, or in the lower lobes or just above the
defined in these regions. Septa measure about 100 mm diaphragm (13,35,53). When visible, they are usually seen
(0.1 mm) in thickness in a subpleural location (32,40). extending to the pleural surface. In the central lung, septa
Within the central lung, interlobular septa are thinner and are thinner than they are peripherally and are infrequently
less well defined than peripherally, and lobules are more seen in normal lungs; often interlobular septa, which are
difficult to identify in this location. clearly defined in this region, are abnormally thickened.
Peripherally, interlobular septa measuring 100 mm or Occasionally, when interlobular septa are not clearly
0.1 mm in thickness are at the lower limit of HRCT resolu- visible, their locations can be inferred by locating septal
tion (40). On clinical scans in normal patients, a few septa pulmonary vein branches, approximately 0.5 mm in
Figure 8-10 A secondary lobule according to Miller is represented. The centrilobular structures are composed of the terminal bronchiole
and pulmonary artery, enclosed in a bronchovascular sheath. The bronchial artery runs in parallel to these structures within the connective
tissue sheath. A deep lymphatic system located in close apposition to the core structures drains the interstitium. Note that lymphatics
are not present around alveolar sacs. In the periphery of the lobule are the visceral pleura and septa, extending from the visceral pleura into
the lung substance. Within these peripheral structures run the pulmonary veins and the superficial lymphatic network that drains the pleura
via septal lymphatics that are perivenous in location. The functional units of the lung made up of collections of alveolar sacs are in effect
suspended between the core structures and the shell structures via a network of interconnecting elastic fibers. This basic organization is
self-replicated from the smallest to the largest functional lung unit.
Figure 8-11 Normal interlobular septa and peripheral vessels.

High-resolution CT image demonstrates normal interlobular septa
(arrowheads) in the lingula. Veins can be seen coursing in the septa,
whereas the centrilobular pulmonary arteries (arrows) are a few
millimeters away from the septa and pleura.
diameter. Veins can sometimes be seen as linear, arcuate, the lobule, as well as some supporting connective tissue
or branching structures or as a row or chain of dots, (i.e., the peribronchovascular interstitium) (Fig. 8-12)
surrounding centrilobular arteries, and approximately (29,32,40). The branching of the lobular bronchiole and
5 to 10 mm from them. artery is irregularly dichotomous (54). In other words,
when they divide, they divide into two branches that are
usually of different sizes; one branch is nearly the same
Centrilobular Region
size as the one it arose from, and the other is smaller
The central portion of the lobule, referred to as the (Fig. 8-9B). Thus often a single dominant bronchiole and
centrilobular region or lobular core (42), contains the artery occur in the center of the lobule, and give off smaller
pulmonary artery and bronchiolar branches that supply branches at intervals along their length.
Figure 8-12 Dimensions of secondary lobular structures (A) and their visibility on high-resolution CT (B).
The HRCT appearances and visibility of structures in the Generally speaking, lung opacity as seen on HRCT scans
centrilobular region are determined primarily by their size obtained at full inspiration appears relatively homoge-
(Figs. 8-9 and 8-12). Secondary lobules are supplied by neous. Measurements of lung attenuation in normal
arteries and bronchioles measuring approximately 1 mm subjects can range from 700 to 900 HU, corresponding
in diameter, whereas intralobular terminal bronchioles to lung densities of approximately 0.300 to 0.100 g/mL,
and arteries measure about 0.7 mm in diameter, and respectively (56,57). However, an attenuation gradient is
acinar bronchioles and arteries range from 0.3 mm to normally present, if measurements are made, with the
0.5 mm in diameter. Arteries of this size can be easily most dependent lung regions being the densest, and the
resolved by using HRCT technique (32,40). most nondependent lung regions being the least dense.
On clinical scans, a linear, branching, or dotlike opacity This gradient is largely caused by regional differences in
frequently seen within the center of a lobule, or within a blood volume and gas volume that, in turn, are deter-
centimeter of the pleural surface, represents the intralobu- mined by gravity, mechanical stresses on the lung, and
lar artery branch or its divisions (Figs. 8-8 and 8-11) (55). intrapleural pressures (57,58). Differences in attenuation
The smallest arteries resolved extend to within 3 to 5 mm between anterior and posterior lung have been measured
of the pleural surface or lobular margin and are as small as in supine patients, and values generally range from 50 to
0.2 mm in diameter (32,40,55). The visible centrilobular 100 HU (56,57,59,60), although gradients of more than
arteries are not seen to extend to the pleural surface in the 200 HU have been reported (60). The anteroposterior
absence of atelectasis. A similar branching structure, seen attenuation gradient was found to be nearly linear and
slightly farther from the pleural surface, can represent a was present regardless of whether the subject was supine
pulmonary vein. Although vein branches can sometimes or prone (60).
be seen in relation to interlobular septa, whereas artery Although most authors have reported that normal
branches are centrally located relative to surrounding anteroposterior lung-attenuation gradients are linear, with
septa, veins and arteries cannot always be distinguished. attenuation increasing gradually from anterior to posterior
Fortunately, differentiation between these is not often lung, the lingula and superior segments of the lower lobes
necessary in clinical practice. can appear relatively lucent in some normal subjects (61);
The thickness of the wall of an airway determines its focal lucency in these segments should be considered a
visibility. For a 1-mm bronchiole supplying a secondary normal finding. Although the reason for this is unclear,
lobule, the thickness of its wall measures approximately these slender segments may be less well ventilated than
0.15 mm; this is at the lower limit of HRCT resolution. The adjacent lung and therefore less well perfused, or some air
wall of a terminal bronchiole measures only 0.1 mm in trapping may be present.
thickness, and that of an acinar bronchiole, only 0.05 mm,
both of which are below the resolution of HRCT technique
for a tubular structure. In one in vitro study, only bronchi PATTERNS OF ABNORMALITY ON
having a diameter of 2 mm or more or having a wall thick- HIGH-RESOLUTION COMPUTED
ness of more than 100 mm (0.1 mm) were visible by using TOMOGRAPHY
HRCT (40); and resolution is certainly less than this on
clinical scans. It is important to remember that on clinical The diffuse parenchymal lung diseases include chronic and
HRCT, intralobular bronchioles are not normally visible, acute infiltrative lung diseases and emphysema. The defini-
and bronchi or bronchioles are not normally seen within tion of diffuse lung disease is somewhat arbitrary because
1 cm of the pleural surface. these conditions often have a patchy or focal distribution.
Pulmonary acini making up the lobules (43) and the Findings that indicate the presence of pulmonary parenchy-
intralobular interstitium (29), a fine network of very thin mal abnormalities consistent with diffuse lung disease can
fibers within the alveolar septa, are not visible on HRCT in be classified into four large categories based on their appear-
normal subjects. ances. These are linear and reticular opacities; nodules and
nodular opacities; increased lung opacity, including ground-
glass opacity and airspace consolidation; and abnormalities
Lung Attenuation
associated with decreased lung opacity, including cystic
Most of the lung, representing alveoli and their associated lesions, emphysema, and airway abnormalities.
pulmonary capillary bed, supplied by small airways and
branches of the pulmonary arteries and veins, appears
Linear and Reticular Opacities
featureless on HRCT, identifiable as homogeneous gray
“background” density. The thickness of alveolar walls Linear and reticular opacities result from thickening of the
normally measures approximately 20 to 30 mm and is pulmonary interstitium by fluid, fibrosis, or infiltration by
therefore an order of magnitude below the resolution of cells or other material. Linear and reticular opacities may
HRCT. On HRCT, lung parenchyma should be of greater be manifested by the interface sign, peribronchovascular
opacity than air, but this difference may vary with window interstitial thickening, interlobular septal thickening,
settings (see Chapter 1). parenchymal bands, subpleural interstitial thickening,
Figure 8-13 Linear and reticular opacities

visible on HRCT. (From Webb WR, Müller NL,
Naidich DP. High-resolution CT of the lung,
3rd ed. Philadelphia: Lippincott Williams &
Wilkins; 2001; with permission.)
intralobular interstitial thickening, honeycombing, irregu- The bronchi and pulmonary arteries are surrounded
lar linear opacities, and subpleural lines (Fig. 8-13). by a connective tissue sheath, termed the peribronchovascu-
The interface sign is characterized by the presence of lar interstitium, which extends from the level of the pul-
irregular interfaces between the aerated lung and bronchi, monary hila into the peripheral lung. Peribronchovascular
vessels, and visceral pleura (Fig. 8-14) (36). It is a common interstitial thickening occurs in many diffuse lung diseases.
and early finding of interstitial lung disease, being seen most The thickening can be smooth, nodular, or irregular.
frequently in patients with pulmonary fibrosis (36,62). Smooth peribronchovascular interstitial thickening is most
typical of patients with interstitial pulmonary edema (63)
or lymphangitic spread of carcinoma (64,65). Nodular
thickening of the peribronchovascular interstitium is par-
ticularly common in sarcoidosis and lymphangitic spread
of carcinoma (Fig. 8-15) (64–67). Irregular peribron-
chovascular interstitial thickening is most frequently seen
in patients with peribronchovascular and adjacent lung
fibrosis (62). In patients with pulmonary fibrosis and
peribronchovascular interstitial thickening, often the
dilatation of the bronchi is caused by traction by the
surrounding fibrosis. This is referred to as traction
bronchiectasis (Fig. 8-16) (68).
A characteristic finding in patients with interstitial lung
disease is the presence of thickening of the interlobular
septa. Normally only a few septa are seen. Thickened septa
(septal lines) are readily recognized in the lung periphery
as lines 1 to 2 cm in length, 0.5 to 2 cm apart, perpendicu-
Figure 8-14 Irregular interface sign. High-resolution CT at the lar to the pleura, and extending to the pleural surface
level of the bronchus intermedius demonstrates irregular inter- (Fig. 8-17). In the central lung regions, thickened septa
faces between the lung parenchyma and the parietal pleura outlining one or more adjacent secondary pulmonary lob-
(arrowheads), interlobar fissures (straight arrows), and interlobular
septa (curved arrows). The patient was a 53-year-old man with ules can be seen as polygonal arcades (36,65). Thickening
idiopathic pulmonary fibrosis. of the interlobular septa may be caused by edema, cellular
Figure 8-17 Interlobular septal thickening. High-resolution CT

at the level of the upper lobes demonstrates smooth asymmetric
bilateral interlobular septal thickening (arrows). The patient was a
Figure 8-15 Peribronchial and perivascular interstitial thicken- 69-year-old woman with interstitial pulmonary edema.
ing. High-resolution CT at the level of the right middle lobe
demonstrates nodular peribronchial (straight arrows) and perivas-
cular (curved arrows) thickening. Also noted are small nodules
along the interlobar fissures (arrowheads) and costal pleura. The occurs most commonly in patients with lymphangitic
patient was a 28-year-old man with sarcoidosis. carcinomatosis (Fig. 8-18) (64,83), Kaposi sarcoma (84),
sarcoidosis (Fig. 8-3) (85,86), silicosis, and coalworkers’
infiltration, or fibrosis. Septal thickening can be smooth, pneumoconiosis (87,88). Less common causes include
nodular, or irregular in contour (Table 8-2). Smooth septal
thickening is seen most commonly in patients with pul-
monary edema (63,69) or lymphatic spread of carcinoma
TABLE 8-2
(64,65); less common causes include lymphoma (70,71),
DIFFERENTIAL DIAGNOSIS OF INTERLOBULAR
leukemia (72,73), Churg-Strauss syndrome (74,75), acute
lung rejection (76), congenital lymphangiectasia (77), and SEPTAL THICKENING
Niemann-Pick disease (78). Smooth septal thickening Diagnosis Comments
may also be seen in association with ground-glass opacity,
a pattern termed crazy-paving; this pattern is typical of Pulmonary edema Common; predominant finding
in most; smooth
alveolar proteinosis but has a long differential diagnosis, Lymphangitic carcino- Common; predominant finding
which is reviewed later (79–82). Nodular septal thickening matosis, lymphoma, in most; smooth or nodular
leukemia
Benign lymphoprolif- Smooth or nodular; other abnormali-
erative disorders ties (e.g., ground-glass opacities,
(e.g., lymphocytic nodules) typically present
interstitial
pneumonia)
Sarcoidosis Common; usually nodular or irregular
Idiopathic pulmonary Sometimes visible; irregular;
fibrosis or other intralobular linear opacities usually
cause of usual predominate
interstitial
pneumonia
Nonspecific interstitial Associated with ground-glass opacity
pneumonia and reticulation
Asbestosis Sometimes visible; irregular
Pulmonary Smooth; associated with ground-glass
hemorrhage opacity
Acute lung rejection Common; smooth
Churg–Strauss Common; smooth
syndrome
Figure 8-16 Traction bronchiectasis. High-resolution CT at the Congenital Common; smooth
level of the lower lung zones shows extensive bilateral reticulation. lymphangiectasia
The bronchi within the areas of reticulation are dilated and dis- Niemann–Pick Common; smooth or nodular
torted and have a beaded appearance (arrows), characteristic of disease
traction bronchiectasis. The patient was a 76-year-old man with Amyloidosis Sometimes visible; smooth or nodular
idiopathic pulmonary fibrosis.
Figure 8-18 Interlobular septal thickening. High-resolution CT

at the level of the aortic arch demonstrates smooth (straight
arrows) and nodular (curved arrows) interlobular septal thickening. Figure 8-19 Intralobular linear opacities. High-resolution CT
Also noted are subpleural nodules. The patient was a 72-year-old through the lower lung zones demonstrates linear opacities sepa-
man with lymphangitic carcinomatosis caused by metastatic ade- rated by only a few millimeters, causing a fine reticular pattern.
nocarcinoma of the rectum. Also noted is irregular interlobular septal thickening. The patient
was a 58-year-old woman with idiopathic pulmonary fibrosis.
lymphoma (70,71), leukemia (72,73), lymphocytic inter- hemorrhage, some pneumonias (e.g., that due to Pneumocystis
stitial pneumonia (LIP) (89), and amyloidosis (90,91). jiroveci), lymphangitic carcinomatosis, and alveolar pro-
Irregular septal thickening is seen most commonly in teinosis (79,81). In these patients, traction bronchiectasis
patients with interstitial fibrosis, particularly idiopathic and other manifestations of fibrosis are absent.
pulmonary fibrosis (IPF), asbestosis, and sarcoidosis
(92,93), and is usually associated with other findings of
Nodular Opacities
fibrosis, such as intralobular linear opacities and architec-
tural distortion (49,93,94). Nodules of 1 to 10 mm in diameter are seen in a number
Intralobular linear opacities are caused by thickening of acute and chronic infiltrative lung diseases. The differ-
of the interstitium within the secondary pulmonary ential diagnosis of multiple small nodular opacities is
lobule. They result in a fine reticular pattern, a mesh of based on their appearance, whether well defined or poorly
small irregular lines separated by only a few millimeters defined, their distribution, and the presence of associated
(Fig. 8-19). Intralobular linear opacities are most com- findings. Three main types of distribution can be identi-
monly caused by fibrosis (Table 8-3). In IPF and interstitial fied on HRCT: perilymphatic, centrilobular, and random.
fibrosis associated with collagen-vascular diseases, the
intralobular linear opacities are characteristically most evi-
dent in the subpleural lung regions and in the lower lung
zones (Fig. 8-20) (62,95–97). A similar pattern and distri-
bution may be seen in asbestosis, although the diagnosis
can usually be suggested by the presence of associated
pleural thickening (13,94). Intralobular linear opacities
are also frequently seen in patients with nonspecific inter-
stitial pneumonia (NSIP) (97). NSIP can be idiopathic or
be seen in association with a variety of conditions, particu-
larly drug reactions and collagen-vascular diseases (97). In
NSIP, the intralobular linear opacities tend to be mild and
usually are seen in association with ground-glass opacities
(98,99). In chronic hypersensitivity pneumonitis, the
fibrosis is usually most severe in the mid lung zones and
is more random in distribution rather than having a
subpleural predominance (100), although in some cases, Figure 8-20 Intralobular linear opacities. High-resolution CT
the findings may be similar to those of IPF (101). Although (1-mm collimation) demonstrates linear opacities separated by only
intralobular linear opacities usually indicate fibrosis, a few millimeters (intralobular linear opacities) (arrows), causing a
fine reticular pattern. The abnormalities involve mainly the sub-
they may also occasionally be seen in the absence of fibro- pleural lung regions and are most severe in the right lung. The
sis, having been described in patients with pulmonary patient was a 70-year-old man with idiopathic pulmonary fibrosis.
TABLE 8-3 TABLE 8-4

DIFFERENTIAL DIAGNOSIS OF INTRALOBULAR DIFFERENTIAL DIAGNOSIS OF
INTERSTITIAL THICKENING PERILYMPHATIC NODULES
Diagnosis Comments Diagnosis Comments
Idiopathic pulmonary Present in vast majority of cases; Sarcoidosis Well defined; mainly peribronchovascular
fibrosis or other often associated with and along interlobar fissures; upper
cause of usual honeycombing lobe predominance
interstitial pneumonia Silicosis and coal- Centrilobular, subpleural, symmetric;
Nonspecific interstitial Common; ground-glass opacity workers’ posterior upper lobe predominance
pneumonia, desquama- usually visible pneumoconiosis
tive interstitial pneumo- Lymphangitic Septal, peribronchovascular; most
nia, acute interstitial carcinomatosis commonly bilateral symmetric; may
pneumonia be asymmetric, patchy, or unilateral
Hypersensitivity pneu- Common; associated with other Lymphocytic intersti- Well defined or poorly defined;
monitis (chronic) findings of fibrosis; often associ- tial pneumonia mainly peribronchial. Commonly have
ated with ground-glass opacities (LIP) and lymphoid ground-glass opacities; LIP commonly
and centrilobular nodules hyperplasia has cysts
Asbestosis Common; associated with other
findings of fibrosis and with
pleural thickening
Lymphangitic carcinoma- Smooth or nodular; associated
tosis; lymphoma; with septal thickening perilymphatic distribution are frequently associated
leukemia with nodular thickening of the bronchovascular bundles
Pulmonary edema Smooth; associated with septal (Figs. 8-3 and 8-15). Another sign helpful in assessing the
thickening and ground-glass perilymphatic distribution is the presence of subpleural
opacity
Pulmonary hemorrhage Smooth; associated with septal thick-
nodules in relation to the interlobar fissures, a characteristic
ening and ground-glass opacity finding in sarcoidosis, silicosis, and coalworkers’ pneumo-
Alveolar proteinosis Smooth; associated with septal coniosis (Fig. 8-22). The various diseases associated with a
thickening and ground-glass perilymphatic distribution of nodules can usually be distin-
opacity guished on HRCT by the different patterns of involvement
of the perilymphatic interstitium. In sarcoidosis, the nod-
ules are usually most numerous along the perihilar bronchi
A perilymphatic distribution along the bronchovascular and vessels and the major fissures (Fig. 8-3) (67,68,102). In
interstitium, interlobular septa, and subpleural lung regions patients with lymphangitic spread of tumor, when nodules
is characteristically seen in patients with sarcoidosis (67), are visible, they are most often visible within the thickened
lymphangitic carcinomatosis (64), silicosis and coalwork- peribronchovascular interstitium and interlobular septa. In
ers’ pneumoconiosis (Fig. 8-21) (Table 8-4) (87,88). In the majority of these patients, the main finding is septal
these conditions, the nodules tend to be well defined thickening (Fig. 8-18); peribronchovascular and subpleural
and usually measure 2 to 5 mm in diameter. Nodules in a nodules are typically not as profuse as in patients with
Figure 8-21 Appearance of small nodules with

a perilymphatic distribution. Nodules predominate
in relation to the perihilar peribronchovascular
interstitium, centrilobular interstitium, interlobular
septa, and subpleural regions. Conglomerate sub-
pleural nodules can form pseudoplaques. (From
Webb WR, Müller NL, Naidich DP. High-resolution
CT of the lung, 3rd ed. Philadelphia: Lippincott
Williams & Wilkins; 2001, with permission.)
Figure 8-22 Small nodules in a perilymphatic distribution. High-

resolution CT shows small nodules along the bronchi (straight Figure 8-23 HRCT appearances of centrilobular nodules.
arrows), vessels (curved arrows), and major and right minor interlo- Centrilobular nodules are usually separated from the interlobu-
bar fissures (arrowheads). Also noted are enlarged subcarinal lar septa and pleural surfaces by a distance of several millime-
and hilar lymph nodes. The patient was a 28-year-old man with ters; in the lung periphery, the nodules are usually centered
sarcoidosis. 5 to 10 mm from the pleural surface. Centrilobular nodules may
be associated with small pulmonary artery branches. Because of
the similar size of secondary lobules, centrilobular nodules often
appear to be evenly spaced. Although they are often ill defined,
sarcoidosis (103). Silicosis and coalworkers’ pneumoconio- this is not always the case. Either a single centrilobular nodule or
a centrilobular rosette of nodules may be seen. In occasional
sis are associated with the presence of small, well-defined cases, the air-filled centrilobular bronchiole can be recognized as
nodules, usually measuring from 2 to 5 mm in diameter, a rounded lucency within a centrilobular nodule. Tree-in-bud
which predominantly appear centrilobular and subpleural may be seen in patients with a centrilobular distribution, repre-
senting impaction of centrilobular bronchioles. (From Webb
in location on HRCT (87,88). WR, Müller NL, Naidich DP. High-resolution CT of the lung,
Centrilobular nodules are characterized on HRCT by 3rd ed. Philadelphia: Lippincott Williams & Wilkins; 2001, with
their location several millimeters away from the pleural permission.)
surfaces, interlobar fissures, and interlobular septa and
are further defined as appearing either poorly defined
ground-glass in density, or having a distinct “tree-in-bud”
configuration (Fig. 8-23). Predominantly poorly defined
centrilobular ground-glass nodules are most often a mani-
festation of subacute or chronic cellular bronchiolitis and
are typically seen in patients with subacute hypersensiti-
vity pneumonitis (14,104), less commonly in patients
with a variety of causes of bronchiolitis, including respira-
tory bronchiolitis (106,107), respiratory bronchiolitis–
interstitial lung disease (RB-ILD) (107,108), and rarely,
cryptogenic organizing pneumonia (COP) (also known
as bronchiolitis obliterans with organizing pneumonia;
BOOP) (Table 8-5) (Fig. 8-24) (110). In hypersensitivity
pneumonitis, the nodules may be diffuse throughout the
lungs or involve mainly the middle and lower lung
zones, whereas in respiratory bronchiolitis and RB-ILD, the
nodules usually involve mainly or exclusively the upper
lobes. Figure 8-24 Centrilobular nodules in hypersensitivity pneu-
monitis. High-resolution CT (1-mm collimation) at the level of the
In distinction to poorly defined ground-glass nodules bronchus intermedius demonstrates numerous poorly defined cen-
is the presence of branching nodular or linear opacities trilobular nodules. Note that the nodules are a few millimeters
resulting in a so-called tree-in-bud pattern (Figs. 8-25 and away from the pleura (arrowheads), including interlobar fissures
(arrows), a characteristic finding of centrilobular nodules. The
8-26) (111,112). The tree-in-bud pattern is characteristic of patient was a 35-year-old man with subacute hypersensitivity
acute cellular bronchiolitis and results from the presence pneumonitis (bird-breeder’s lung).
TABLE 8-5
DIFFERENTIAL DIAGNOSIS OF
CENTRILOBULAR NODULES
Diagnosis Comments
Hypersensitivity Poorly defined; similar size; bilaterally

pneumonitis symmetric; patchy or diffuse;
ground-glass opacities common
Respiratory bronchiolitis Poorly defined; bilaterally symmetric;
and respiratory bron- upper lobe predominance; patchy
chiolitis interstitial ground-glass opacity
lung disease
Infectious bronchiolitis Well defined or poorly defined;
and broncho- asymmetric, patchy distribution; Figure 8-26 Tree-in-bud appearance in infectious bronchiolitis.
pneumonia tree-in-bud pattern common High-resolution CT (1.0-mm collimation) at the level of the lung
bases shows centrilobular branching nodular and linear opacities,
Endobronchial spread Well defined; asymmetric, patchy
resulting in a tree-in-bud appearance (arrows). The patient was a
of tuberculosis and distribution; tree-in-bud pattern 20-year-old woman with recurrent respiratory infections.
Mycobacterium common
avium complex
(MAC)
Panbronchiolitis Well defined; symmetric; bronchiec- tious disease, it is significant that in most patients with
tasis common; tree-in-bud pattern
Asiatic panbronchiolitis, at least initially, a good response
common
Allergic bronchopul- Well defined or poorly defined; usually may be expected after antibiotic therapy (114a, 114b).
monary aspergillosis associated with central bronchiectasis Tree-in-bud opacities have been reported to occur distal to
Pulmonary Langerhans Centrilobular nodules commonly bronchiectasis of any cause (111,113), including in pa-
cell histiocytosis present; well defined; thin- or thick- tients with allergic bronchopulmonary aspergillosis (113).
walled cysts; round or bizarre shapes;
Small nodules in a random distribution in relation to
mid and upper lung zone distribu-
tion, relative sparing of lung bases structures of the lung and secondary lobule are seen
most commonly in miliary tuberculosis (112,115) and
miliary fungal infections (116) and less commonly in
of dilated centrilobular bronchioles impacted with sarcoidosis (Fig. 8-27). Hematogenous miliary metastases
mucus, fluid, and/or pus, often in association with peri- have a random distribution regarding the lobular struc-
bronchiolar inflammation. It is usually associated tures (Fig. 8-28) but tend to be most numerous in the
with infection of the small airways, the most common lung periphery and at the lung bases (117). Randomly dis-
causes being infectious bronchiolitis, bronchopneumo- tributed nodules can be seen in relation to interlob-
nia, and endobronchial spread of tuberculosis (43,105) or ular septa, small vessels, and pleural surfaces but do
Mycobacterium avium-intracellulare infection (111,113,114). not have consistent or predominant distribution to any
Typically focal in distribution, tree-in-bud opacities may of these. Lung involvement tends to be bilateral and
also be diffusely distributed throughout the lungs, as symmetric.
typically occurs in patients with Asian panbronchiolitis Distinction of perilymphatic, centrilobular, and random
(109). Although not typically characterized as an infec- distribution of small nodules is most readily accomplished
Figure 8-25 Tree-in-bud appearance in tuberculosis. High-

resolution CT (1.5-mm collimation) through the upper lobes
demonstrates bilateral cavities and sharply defined centrilobular
nodules. The nodules and branching lines have an appearance
that resembles a tree-in-bud (arrows).
characterization of centrilobular nodules either as having

a tree-in-bud configuration, indicative of acute cellular
bronchiolitis, or as poorly defined centrilobular nodules, as
frequently occurs in patients with subacute or chronic
cellular bronchiolitis (117a).
Nodules greater than 1 cm in diameter may result from
conglomeration of smaller nodules or from progressive
fibrosis, as may be seen in patients with silicosis (118), sar-
coidosis (119,120), or talcosis (121,122). More commonly,
larger nodules result from pulmonary metastases, septic
emboli (123), or Wegener granulomatosis (124,125).
Pulmonary metastases tend to have smooth margins and
are most numerous in the subpleural regions of the lower
Figure 8-27 Miliary tuberculosis. High-resolution CT through lung zones. Septic emboli have a similar distribution.
the upper lobes shows numerous widely scattered nodules, with- The nodules in Wegener granulomatosis may range from
out obvious relation to lobular anatomy. The patient was a 66-year- 2 mm to 7 cm in diameter, usually have irregular margins,
old woman with miliary tuberculosis. (Case courtesy of Dr. Isabela
Silva, Salvador, Brazil.) and frequently cavitate (124).
In immunocompromised patients nodules larger than
by adhering to an easily followed diagnostic algorithm. 1 cm in diameter may be caused by septic embolism
First proposed by Gruden et al. (114), this approach first (123), invasive aspergillosis, post-transplant lymphopro-
requires identifying whether pleural and/or perifissural liferative disorders, or lymphoma (126). Nodules caused
nodules are present. If subpleural nodules are absent, the by invasive pulmonary aspergillosis often have a character-
pattern is centrilobular. If numerous subpleural or fissural istic halo of ground-glass attenuation surrounding the
nodules are present, then the pattern is either perilym- nodule (Fig. 8-29). This halo is caused by hemorrhage.
phatic or random. These two patterns are further distin- Although the “halo sign” was first described in invasive
guished by noting additional features regarding their aspergillosis, it may also be seen in other hemorrhagic
anatomic distribution. Specifically, if nodules are patchy nodules, including candidiasis, Wegener granulomatosis,
in distribution and demonstrate a distinct peribroncho- metastatic angiosarcoma, and Kaposi sarcoma, as well as
vascular (axial), interstitial, and/or interlobular septal in nonhemorrhagic nodules such as bronchioloalveolar
predominance, then the nodules are perilymphatic. In carcinoma and metastases from carcinoma of the colon
distinction, if nodules are distributed in a diffuse and uni- (127,128). However, in the clinical context of a patient
form manner, or predominantly involve the lung bases, the with hematologic malignancy and severe neutropenia, the
pattern is random, with nodules likely hematogenous “halo sign” is highly suggestive of angioinvasive aspergillo-
in origin. A similar algorithmic approach allows further sis (129). Post-transplant lymphoproliferative disorders
Figure 8-29 Nodules with halo in invasive aspergillosis. High-

Figure 8-28 Miliary metastases. High-resolution CT through the resolution CT (1-mm collimation) through the upper lung zones
upper lobes shows numerous widely scattered nodules, without demonstrates nodules ranging from 5 to 25 mm in diameter. The
obvious relation to lobular anatomy. Incidental note is made of nodules have irregular margins and are surrounded by a halo of
mild emphysema. The patient was a 76-year-old man with miliary ground-glass attenuation (arrows). The patient was a 77-year-old
metastases from renal cell carcinoma. woman with leukemia and angioinvasive aspergillosis.
and lymphomas seen in patients with acquired immunod-

eficiency syndrome (AIDS) often are first seen with
pulmonary nodules, which may have smooth or irregular
margins (126). In immunocompromised patients, lym-
phoma may often involve the lungs without associated
hilar or mediastinal lymphadenopathy.
Ground-Glass Opacity
Ground-glass opacity is characterized by the presence of
hazy increased attenuation of the lung without obscuration
of the underlying bronchial and vascular outlines (130).
If the vessels are obscured, the term “consolidation” is used
Figure 8-31 Ground-glass attenuation in Pneumocystis pneu-
(131,132). Ground-glass opacity results from volume aver- monia (PCP). High-resolution CT (1.5-mm collimation) shows exten-
aging of abnormalities below the resolution of CT and may sive areas of ground-glass attenuation in the upper lobes. The
be caused by interstitial thickening, air-space filling, or patient was a 41-year-old man with acquired immunodeficiency
syndrome (AIDS) and PCP.
both (131). Although ground-glass opacity is a nonspecific
finding, it is an important sign because it usually indi-
cates the presence of active, potentially treatable disease The main differential diagnosis in patients with
(Fig. 8-30) (131,133). Furthermore, the differential diagno- chronic respiratory symptoms and exclusively ground-
sis can be narrowed considerably by assessing the HRCT glass opacities includes hypersensitivity pneumonitis
findings together with the clinical information (134). (15,141), NSIP (98,99,131,142), desquamative interstitial
The presence of ground-glass opacities as the predomi- pneumonia (DIP) (143), and RB-ILD (107,108,134)
nant or only finding in patients with AIDS is highly (Fig. 8-30) (Table 8-6). Hypersensitivity pneumonitis is
suggestive of Pneumocystis jiroveci pneumonia (Fig. 8-31) probably the most common cause of diffuse ground-glass
(Table 8-6) (84,135,136). In the non-AIDS immunocom- opacities in normal hosts (134). NSIP may be idiopathic
promised patient, ground-glass opacities are a less specific or associated with specific diseases, such as scleroderma
finding, being seen in patients with Pneumocystis jiroveci or other collagen-vascular diseases, and in drug reactions
pneumonia, cytomegalovirus pneumonia, cytotoxic drug (98,99,131,142). DIP and RB-ILD are uncommon, being
reaction, pulmonary edema, and pulmonary hemorrhage seen almost exclusively in cigarette smokers. Other
(Table 8-6) (137–140). chronic diseases that commonly have ground-glass opa-
cities but usually are associated with other findings
include IPF (62,95–99), COP (131,144), sarcoidosis
(120,145), LIP (89), chronic eosinophilic pneumonia
(80), Churg-Strauss syndrome (75), lipoid pneumo-
nia (146), bronchioloalveolar carcinoma (147), and
alveolar proteinosis (148).
The differential diagnosis of ground-glass opacities in
patients with chronic lung diseases is based on the distri-
bution of the ground-glass opacities and, more important,
on the presence of associated findings, such as reticulation
or centrilobular nodules. In IPF, NSIP, and DIP, the
ground-glass opacities tend to involve mainly the sub-
pleural lung regions and the lung bases. In virtually all
patients with IPF and in the majority with NSIP, ground-
glass opacities are seen in association with intralobular
linear opacities in a predominantly peripheral and basal
distribution, architectural distortion, and traction bron-
chiectasis (Fig. 8-32). Ground-glass opacities in hypersen-
sitivity pneumonitis may be diffuse or involve mainly the
mid and lower lung zones and are frequently associated
Figure 8-30 Ground-glass attenuation in desquamative intersti- with poorly defined centrilobular nodules and areas of air
tial pneumonia (DIP). High-resolution CT (1.5-mm collimation) trapping (15,149). Ground-glass opacities in sarcoidosis
through the lower lung zones demonstrates bilateral areas of are almost always associated with other findings, including
ground-glass attenuation with relative sparing of the right middle
lobe. Minimal, if any, evidence of fibrosis is seen. The patient was a hilar and mediastinal lymphadenopathy and perilym-
63-year-old woman with DIP. phatic nodules.
TABLE 8-6
DIFFERENTIAL DIAGNOSIS OF GROUND-GLASS OPACITY
Diagnosis Course Comments
Acute respiratory distress Acute Almost always present; consolidation

syndrome (ARDS), acute common; patchy or diffuse
interstitial pneumonia
Pulmonary edema Acute Perihilar or dependent; septal
thickening often present
Pulmonary hemorrhage Acute Patchy or diffuse; septal thickening
sometimes present
Pneumonia (e.g., Acute Common; diffuse or patchy; centrilobular
Pneumocystis jiroveci, nodules; consolidation or septal
cytomegalovirus) thickening may also be present
Nonspecific interstitial Chronic Common; patchy; subpleural; reticular
pneumonia opacities often associated; multiple
causes including collagen-vascular
diseases
Usual interstitial pneumonia Chronic Common in association with findings of
and idiopathic pulmonary fibrosis; uncommon as an isolated
fibrosis finding; subpleural and basal
predominance
Hypersensitivity pneumonitis Subacute, Very common; patchy or nodular; can be
chronic centrilobular; lobular areas of air
trapping often present
Respiratory bronchiolitis– Subacute, Usually present; patchy and localized;
interstitial lung disease chronic may be centrilobular; fibrosis uncommon
Desquamative interstitial Chronic Always present; diffuse or patchy;
pneumonia findings of fibrosis in approximately
50% of patients, usually mild
Cryptogenic organizing Subacute, Common; consolidation usually also
pneumonia (BOOP) chronic present; often predominant in peripheral
regions; can be nodular
Sarcoidosis Chronic Uncommon manifestation; due to
confluence of very small granulomas
Lymphoid interstitial Chronic Seen in majority of patients; often
pneumonia peribronchovascular nodules; about
50% have cystic lesions
Chronic eosinophilic Chronic Consolidation more common; patchy
pneumonia or nodular; peripheral predominance
Churg-Strauss syndrome Subacute, Patchy, bilateral, often mainly in lung
chronic periphery
Bronchioloalveolar carcinoma Chronic Diffuse, patchy, or centrilobular;
(diffuse) consolidation common
Lipoid pneumonia Subacute, Patchy or lobular; low-attenuation
chronic consolidation may be present
Alveolar proteinosis Subacute, Very common; patchy or diffuse; septal
chronic thickening common; fibrosis rare
Superimposition of interlobular septal thickening or pneumonitis (153), bronchioloalveolar carcinoma (147),

intralobular lines on ground-glass opacities may result and lipoid pneumonia (146). Clearly, the differential diag-
in a pattern known as crazy-paving (Fig. 8-33) (80,81). nosis of a crazy-paving pattern must be based on a consid-
This pattern was first described in patients with pulmonary eration of clinical as well as HRCT findings and knowledge
alveolar proteinosis (PAP) (81) and is quite typical of of whether symptoms are acute or chronic.
PAP, but may also be seen in patients with a variety of
other diseases, including pulmonary edema (Fig. 8-34)
Consolidation
(63), pulmonary hemorrhage (150), acute respiratory
distress syndrome (ARDS) (80,151), acute interstitial Parenchymal consolidation usually results from filling of
pneumonia (AIP) (80), bacterial pneumonia (80), the airspaces by fluid, protein, cells, or other material
Pneumocystis pneumonia (151), acute eosinophilic pneu- (43,131,154). Occasionally it may result from replacement
monia (152), Churg-Strauss syndrome (75), radiation of air in the alveoli by extensive interstitial disease, as may
Figure 8-34 Crazy-paving pattern in pulmonary edema. High-

resolution CT (1-mm collimation) shows bilateral patchy ground-
glass opacities with superimposed septal lines and fine intralobular
lines in the upper lobes (crazy-paving pattern). Also noted is a small
right pleural effusion. The patient was a 67-year-old man with
pulmonary edema secondary to left-sided heart failure.
Figure 8-32 Ground-glass opacities and mild reticulation in
nonspecific interstitial pneumonia (NSIP). High-resolution CT
image shows asymmetric bilateral ground-glass opacities. Also consist of extensive bilateral areas of ground-glass attenua-
noted are mild subpleural reticulation (straight arrows) and traction and consolidation (Fig. 8-35). AIP is essentially a
tion bronchiectasis (curved arrow). The patient was a 44-year-old
woman with NSIP. diagnosis of exclusion: ARDS in which no etiology is
found. COP, also known as BOOP, most frequently is ini-
be seen in patients with sarcoidosis, usual interstitial tially seen with bilateral areas of consolidation, which in
pneumonia (UIP), and NSIP (98,131). Essentially all con- approximately 60% to 80% of cases are peribronchial or
ditions leading to areas of ground-glass attenuation, when subpleural in distribution (Fig. 8-36) (110,144,158). In
severe enough, will cause consolidation, and both findings approximately 60% of patients with BOOP, some of the
are frequently present in the same patient (144,155). areas of consolidation involve mainly the perilobular
Diffuse diseases commonly associated with consolidation regions and may surround areas with ground-glass attenu-
include bacterial and fungal pneumonia, ARDS, AIP ation, resulting in a “reversed halo sign” (158,159).
(155), COP (144), and eosinophilic pneumonia (Table In the majority of patients with consolidation, the
8-7) (156). AIP is essentially the idiopathic form of ARDS. findings are readily apparent on the radiograph, and CT adds
It is uncommon and is classified as a subtype of idiopathic little additional information. However, because the distribu-
interstitial pneumonia (97). The clinical, histologic, and tion of the consolidation is better appreciated on CT, it
radiologic findings are those of ARDS (155,157) and sometimes can be helpful in the differential diagnosis.
Figure 8-33 Crazy-paving pattern in alveolar proteinosis. High- Figure 8-35 Consolidation in acute interstitial pneumonia (AIP).
resolution CT in a patient with alveolar proteinosis shows exten- High-resolution CT (1.5-mm collimation) demonstrates bilateral
sive bilateral ground-glass opacities with superimposed septal consolidation involving mainly the dependent regions of the lower
lines (arrows), resulting in a “crazy-paving” pattern. (Case courtesy lobes. Patchy areas of ground-glass attenuation are seen anteriorly.
of Dr. Paul Bresser, Amsterdam, The Netherlands.) The patient was an 83-year-old woman with AIP.
For example, in up to 50% of patients with chronic

eosinophilic pneumonia, the peripheral consolidation
(“reverse pulmonary edema pattern”) may not be apparent
on the radiograph but can be clearly seen on CT (160). CT
can also demonstrate localized collections of fat in patients
with patchy or diffuse consolidation caused by lipoid
pneumonia (161). It may also be helpful in patients
with suspected amiodarone lung by demonstrating the
characteristic increased attenuation caused by accumula-
tion of this iodinated agent within the lung parenchyma
(Fig. 8-37) (162,162a). It is also helpful in the diagnosis of
metastatic pulmonary calcification (163).
Decreased Attenuation
Decreased lung attenuation may result from lung destruc-
tion, as seen in cystic lung disease and emphysema, or
from decreased blood flow caused by pulmonary vascular
or airway abnormalities. Cysts are rounded air-containing
Figure 8-36 Consolidation in cryptogenic organizing pneumonia spaces surrounded by well-defined walls. They are com-
(idiopathic bronchiolitis obliterans with organizing pneumonia).
High-resolution CT (1-mm collimation) demonstrates asymmetric monly seen in IPF (93,95), interstitial fibrosis associated
bilateral consolidation. The consolidation is predominantly peri- with rheumatoid arthritis (93,164), asbestosis (13,93),
bronchial (arrows) and subpleural in distribution. The patient was and chronic hypersensitivity pneumonitis (Table 8-8)
an 81-year-old woman with cryptogenic organizing pneumonia.
TABLE 8-7
DIFFERENTIAL DIAGNOSIS OF CONSOLIDATION
Diagnosis Course Comments
Pneumonia (bacterial, fungal, viral, Acute Patchy, nodular, lobular, or diffuse

Mycoplasma, Pneumocystis) depending on the organism
Acute respiratory distress syndrome Acute Patchy or diffuse; tends to involve
(ARDS); acute interstitial mainly dependent lung
pneumonia (AIP)
Pulmonary edema Acute Predominantly dependent or
perihilar distribution
Pulmonary hemorrhage Acute Patchy or diffuse
Radiation pneumonitis Acute Extent usually corresponds to
radiation ports
Cryptogenic organizing pneumonia Subacute, Bilateral; predominantly peri-
(COP) (also known as bronchiolitis chronic bronchial, perilobular, and
obliterans with organizing peripheral distribution
pneumonia, BOOP)
Chronic eosinophilic pneumonia Subacute, Usually peripheral and upper lobe
chronic predominance
Churg–Strauss syndrome Subacute, Often has peripheral and upper
chronic lobe predominance
Bronchioloalveolar carcinoma Subacute, Diffuse, patchy, or nodular
(diffuse) chronic
Lymphoma Subacute, Diffuse, patchy, or nodular
chronic
Nonspecific interstitial pneumonia Subacute, Patchy; subpleural. Consolidation
chronic usually mild; ground-glass
opacity predominates
Lipoid pneumonia Subacute, Patchy; low attenuation within
chronic consolidation
Sarcoidosis Subacute, Confluence of very small
chronic granulomas
Alveolar proteinosis Subacute, Ground-glass opacity more
chronic common
A B
Figure 8-37 Amiodarone lung. A: Chest radiograph demon-

strates areas of parenchymal consolidation in the right upper,
middle, and lower lobes. Also noted are irregular linear opacities
in the right and left upper lobes. B: High-resolution CT (1.5-mm
collimation) demonstrates extensive areas of ground-glass atten-
uation in the right upper lobe and focal areas of consolidation in
the right upper and lower lobes, as well as a few irregular linear
opacities. C: Mediastinal windows demonstrate focal areas of
increased attenuation in the right upper and lower lobes (straight
arrows), as well as increased attenuation of the mediastinal lymph
nodes (curved arrows). Increased attenuation was also present
in the thyroid, heart, and liver (not shown). The patient was a
C 61-year-old man with amiodarone lung.
(93,101). In patients with pulmonary fibrosis, cystic spaces

represent honeycombing and are usually associated with
architectural distortion and traction bronchiectasis.
Honeycomb cysts typically occur in several layers and have
clearly definable walls 1 to 3 mm in thickness (Fig. 8-38).
The honeycombing in patients with IPF and asbestosis
usually is most severe in the subpleural lung regions and
in the lung bases. The honeycombing in chronic hypersen-
sitivity pneumonitis is also most marked in the subpleural
lung regions but tends to be most severe in the mid lung
zones, with relative sparing of the lung bases (93,100).
Honeycombing is seen in 10% to 30% of patients with
NSIP but tends to be less extensive than in IPF and typi-
cally associated with extensive ground-glass opacities and
fine intralobular lines (98,99,164a).
Cystic airspaces are a characteristic finding of lymphangi-
Figure 8-38 Cystic airspaces in idiopathic pulmonary fibrosis
oleiomyomatosis (Fig. 8-39) (165,166). In lymphangio- (IPF). High-resolution CT (1.5-mm collimation) through the lower
leiomyomatosis, the cysts are seen throughout the lungs lung zones demonstrates numerous cysts (straight arrows) involv-
without any zonal predominance. Furthermore, the cysts in ing mainly the subpleural lung regions. The cysts range from a few
millimeters to 1 cm in diameter. Note traction bronchiectasis
patients with lymphangioleiomyomatosis are not related to (curved arrow). The patient was an 81-year-old man with honey-
fibrosis and therefore are surrounded by relatively normal comb lung related to IPF.
lung parenchyma (165,166). Cystic spaces similar to those

seen in lymphangioleiomyomatosis can also be seen in
TABLE 8-8
patients with Langerhans cell histiocytosis (167–169). The
DIFFERENTIAL DIAGNOSIS OF CYSTIC
cysts in Langerhans histiocytosis characteristically involve
LUNG DISEASE the upper two thirds of the lungs, with relative sparing of
Diagnosis Comments the lung bases, often have bizarre shapes, and are com-
monly associated with nodules. Single or multiple cysts are
Idiopathic pulmonary Honeycombing is common also seen superimposed on ground-glass opacities in app-
fibrosis or other cause (70%–90% of cases); peripheral,
of usual interstitial basal, and subpleural
roximately 60% of patients with LIP (89), 10% of patients
pneumonia, such as predominance with subacute hypersensitivity pneumonitis (170), and 30%
rheumatoid arthritis of patients with Pneumocystis jiroveci pneumonia (171).
Nonspecific interstitial Honeycombing seen in 20%–40% Emphysema is defined as a permanent, abnormal
pneumonia of cases; usually mild; ground- enlargement of airspaces distal to the terminal bronchiole
glass opacities and intralobular
lines usually predominate
and accompanied by the destruction of the walls of the
Asbestosis Honeycombing seen in advanced involved airspaces (130). Emphysema is characterized by
disease; peripheral, basal, and the presence of areas of abnormally low attenuation with-
subpleural predominance out definable walls (172,173). It should be noted, how-
Hypersensitivity pneumo- Honeycombing seen in advanced ever, that when emphysema involves the entire secondary
nitis (chronic) disease; may be peripheral,
patchy, or diffuse; middle lung
lobule, the interlobular septa may give the appearance of
zones predominance common thin (1 mm) walls. Thin walls are also seen with bullae,
Sarcoidosis Honeycombing uncommon; may which are defined as sharply demarcated areas of emphy-
be peripheral or patchy; upper sema measuring 1 cm or more in diameter and possessing
lung predominance common a wall of less than 1 mm in thickness (174). Bullae may
Lymphangioleiomyomatosis Thin-walled cysts, usually round,
diffuse distribution
therefore resemble lung cysts, although the distinction can
Pulmonary Langerhans cell Thin- or thick-walled cysts; round usually be readily made because bullae are almost always
histiocytosis or bizarre shapes; centrilobular associated with extensive emphysema. Three main forms
nodules commonly present; of emphysema are recognized: centrilobular, typically
mid and upper lung zones, associated with cigarette smoking; panacinar, typically
relative sparing of lung bases
Lymphocytic interstitial Cysts in about 60% of cases;
associated with a1-antitrypsin deficiency; and paraseptal
pneumonia ground-glass opacities and small emphysema.
nodules in majority of patients Centrilobular emphysema can be diagnosed by the
Hypersensitivity Cysts in about 10% of cases; presence of small lucencies near the center of the second-
pneumonitis superimposed on ground-glass ary lobule (Fig. 8-40) (175). It is usually most severe in the
opacities and centrilobular
nodules
upper lobes. Panacinar (panlobular) emphysema involves
mainly the lower lobes and is characterized by the pres-
ence of diffuse areas of low attenuation with little inter-
vening normal lung (Fig. 8-41) (176). Paraseptal (distal
Figure 8-40 Centrilobular emphysema. High-resolution CT

Figure 8-39 Cystic airspaces in lymphangioleiomyomatosis. (1.5-mm collimation) demonstrates focal areas of decreased atten-
High-resolution CT (1-mm collimation) shows numerous cysts uation with no definable walls. Note that in several areas, small
randomly distributed throughout both lungs. The cysts have thin centrilobular vessels can be seen within the areas of decreased
walls (arrows). The patient was a 59-year-old woman with lymphan- attenuation (arrows). The patient was a 50-year-old smoker with
gioleiomyomatosis. centrilobular emphysema.
Figure 8-42 Mosaic perfusion. High-resolution CT demon-

strates a pattern of mosaic perfusion with focal areas of decreased
attenuation and small vessels (arrows), as well as areas with
increased attenuation associated with increased size and number
Figure 8-41 Panacinar emphysema. High-resolution CT (1-mm of vessels. The patient was a 52-year-old woman with mosaic
collimation) demonstrates panacinar emphysema in the native right
perfusion caused by chronic pulmonary thromboembolism.
lung and a normal transplanted left lung. Note the characteristic
simplification of the lung parenchyma in panacinar emphysema as
compared with the normal architecture in the transplanted lung.
different between normal and abnormal regions of lung
(Figs. 8-31 and 8-32) (69). A mosaic pattern of lung atten-
acinar) emphysema is characterized by areas of low attenuation caused by pulmonary vascular disease, such as
uation in the subpleural lung regions and adjacent to pulmonary arterial hypertension and chronic pulmonary
vessels and interlobular septa (173). thromboembolism, is associated with a decrease in the
Decreased attenuation may also be caused by decreased number and size of vessels in the areas of low attenuation
perfusion secondary to airway or vascular obstruction compared with areas with increased attenuation that
(177). Airway obstruction results in decreased ventilation have an increase in the number and size of vessels caused
and focal air trapping; areas of poorly ventilated lung are by flow redistribution (Fig. 8-42) (69). A similar pattern is
poorly perfused because of reflex vasoconstriction or seen in patients with flow redistribution caused by
because of a permanent reduction in the pulmonary capil- primary small airway disease (Fig. 8-43) (69,185). Dis-
lary bed. Common causes include bronchiolitis obliterans tinction between mosaic attenuation caused by small
(obliterative bronchiolitis) and other diseases associated airway disease and vascular disease requires careful analy-
with small airways obstruction, such as cystic fibrosis, sis of associated findings, such as enlargement of central
bronchiectasis of any cause, and asthma (74,178,179).
Decreased attenuation can also occur in association with
pulmonary vascular obstruction such as that caused by
chronic pulmonary thromboembolism (180,181). The
areas of decreased attenuation can be of varying sizes and
sometimes appear to correspond to lobules, segments,
lobes, or an entire lung (182,183). Redistribution of blood
flow away from the areas of vascular obstruction leads to
increased attenuation in areas of normal lung parenchyma.
This combination of areas with decreased attenuation
and areas with increased attenuation is known as mosaic
perfusion (184).
In some patients, it may be difficult to determine
whether a mosaic pattern of lung attenuation is caused by
flow redistribution to areas of normal lung or whether it is
caused by patchy areas of ground-glass attenuation result-
ing from infiltrative lung disease (177,185). This distinc-
Figure 8-43 Mosaic perfusion. High-resolution CT (1-mm colli-
tion, however, can usually be readily made by close atten- mation) shows extensive areas of decreased attenuation and vas-
tion to the size of the pulmonary vessels and assessment of cularity (arrows) in the right lung and left lower lobe and increased
presence of air trapping on expiratory CT. In patients with attenuation and increased size and number of blood vessels in the
left upper lobe, characteristic of mosaic perfusion. This patient
a patchy infiltrative process resulting in mosaic attenua- was a 20-year-old woman with bronchiolitis obliterans secondary
tion, the number and caliber of vessels are not appreciably to bone marrow transplantation.
pulmonary arteries, often seen in patients with chronic (187–189). Various classifications of CILDs have been sug-
thromboembolic pulmonary hypertension, and bronchial gested. These have been based on etiology, pathogenesis,
dilatation, often seen in patients with small airway disease or pathologic findings (97,190). From the radiologic point
(177). The distinction can also be facilitated by perform- of view, however, it is reasonable to classify the various
ing CT at suspended full expiration. In patients with diseases based on the predominant pattern of abnormality
airway disease, the lucent areas show little change in vol- on HRCT (Table 8-9).
ume and attenuation at end-expiration because of air trap-
ping (185,186). In patients with primary vascular disease,
because usually no air trapping is present, the attenuation Diseases Characterized Primarily by Linear
in both the oligemic and hyperemic lung tends to increase and Reticular Opacities
to a similar degree at end-expiration. It should be noted,
however, that in some cases, the distinction may not be Usual Interstitial Pneumonia and Idiopathic
possible based only on the HRCT findings (177).
Pulmonary Fibrosis
UIP is the most common and most important of the idio-
CHRONIC INFILTRATIVE LUNG DISEASE pathic interstitial pneumonias. The idiopathic interstitial
pneumonias are a heterogeneous group of interstitial
Many chronic diseases may cause diffuse infiltration of the inflammatory and fibrosing lesions that occur without any
lungs. Although they are often referred to as chronic inter- known cause. The American Thoracic Society and European
stitial diseases, the majority involve both the interstitium Respiratory Society multidisciplinary consensus statement
and the airspaces. Therefore the term chronic infiltrative in 2002 classified the idiopathic interstitial pneumonias
lung disease (CILD) is preferable (50,187). Although into seven distinct clinicopathologic entities in order of
more than 200 different CILDs are known, approximately decreasing frequency: usual interstitial pneumonia (UIP),
20 conditions account for the vast majority of cases nonspecific interstitial pneumonia (NSIP), cryptogenic
TABLE 8-9
CLASSIFICATION OF MOST COMMON CHRONIC INFILTRATIVE LUNG DISEASES BASED ON
PREDOMINANT PATTERN OF ABNORMALITY ON HIGH-RESOLUTION COMPUTED TOMOGRAPHY
Linear or Reticular Pattern Ground-glass Opacity
Usual interstitial pneumonia (UIP) Subacute hypersensitivity pneumonitis
Idiopathic is idiopathic pulmonary fibrosis (IPF) Nonspecific interstitial pneumonia (NSIP)
Collagen-vascular disease, mainly rheumatoid arthritis Idiopathic
Nonspecific interstitial pneumonia (NSIP) Collagen-vascular disease, mainly scleroderma
Idiopathic Drug reaction
Collagen-vascular disease, mainly scleroderma Respiratory bronchiolitis–interstitial lung disease (RB-ILD)
Drug reaction Desquamative interstitial pneumonia (DIP)
Chronic hypersensitivity pneumonitis Alveolar proteinosis
Asbestosis Bronchioloalveolar cell carcinoma
Lymphangitic carcinomatosis
Consolidation
Hydrostatic pulmonary edema
Cryptogenic organizing pneumonia (COP, BOOP)
Nodular Opacities Chronic eosinophilic pneumonia
Predominantly perilymphatic Lipoid pneumonia
Sarcoidosis Bronchioloalveolar cell carcinoma
Lymphangitic carcinomatosis Lymphoma
Silicosis/coalworkers’ pneumoconiosis
Cystic Pattern
Predominantly centrilobular
Pulmonary Langerhans cell histiocytosis
Subacute hypersensitivity pneumonitis
Lymphangioleiomyomatosis
Lymphocytic interstitial pneumonia
End-stage fibrosis (honeycomb lung)
Respiratory bronchiolitis–interstitial lung disease (RB-ILD)
Usual interstitial pneumonia (UIP)
Pulmonary Langerhans cell histiocytosis
Idiopathic is idiopathic pulmonary fibrosis (IPF)
Bronchioloalveolar cell carcinoma
Collagen-vascular disease, mainly rheumatoid arthritis
Random
Sarcoidosis
Hematogenous metastases
Asbestosis
Miliary disease
BOOP, bronchiolitis obliterans with organizing pneumonia.

organizing pneumonia (COP; also known as idiopathic

bronchiolitis obliterans organizing pneumonia, BOOP),
TABLE 8-10
acute interstitial pneumonia (AIP), respiratory bron-
HIGH-RESOLUTION COMPUTED TOMOGRAPHY
chiolitis-associated interstitial lung disease (RB-ILD),
desquamative interstitial pneumonia (DIP), and lymphoid
FINDINGS IN IDIOPATHIC PULMONARY
interstitial pneumonia (LIP) (97). A predominantly reticu- FIBROSIS
lar pattern is seen in virtually all patients with UIP and in Characteristic manifestations
some patients with fibrotic NSIP. The predominant pattern Intralobular linear opacities (reticulation)
in the majority of patients with cellular NSIP is ground- Honeycombing
glass; in COP, it is consolidation; in AIP, it is consolidation Traction bronchiectasis and bronchiolectasis
Subpleural and lower lung zone predominance
or ground-glass; in RB-ILD, it is centrilobular nodular or
ground-glass; in DIP, it is ground-glass; and in LIP, it is Other common findings
Ground-glass opacities (usually less extensive than reticulation)
ground-glass or small nodular pattern (97,107,191). These
Mediastinal lymphadenopathy (typically 15 mm diameter)
various entities are discussed under their respective most
frequent predominant pattern of presentation. The main
role of HRCT is to distinguish patients with typical findings
of UIP/IPF from those with the less specific findings seen also have honeycombing (50,95,192a). Honeycombing,
in the other idiopathic interstitial pneumonias (97). like the irregular linear pattern, is most severe in the
UIP is a pathologic entity characterized by a het- subpleural lung regions and in the lung bases (50,95).
erogeneous appearance with alternating areas of normal Honeycomb cysts usually range from 2 to 20 mm in
lung, interstitial inflammation, scattered fibroblastic foci, diameter (32,95). They typically appear to share walls on
fibrosis, and honeycombing, involving mainly the sub- HRCT and usually occur in several layers in the subpleural
pleural parenchyma and the lung bases (97). UIP is a lung (Fig. 8-46).
nonspecific reaction pattern that is usually idiopathic In the appropriate clinical setting, the presence of a
but that may also be seen in collagen-vascular diseases, bilateral, predominantly subpleural and basal reticular
particularly rheumatoid arthritis, asbestosis, and drug reac- pattern associated with traction bronchiectasis and bronchi-
tion (97). IPF is defined as a chronic fibrosing interstitial olectasis or honeycombing in the absence of small nodules,
pneumonia of unknown cause limited to the lungs and consolidation, or extensive ground-glass opacities allows
associated with a histologic pattern of UIP (97,192). The confident diagnosis of IPF on HRCT, precluding the need
diagnosis of IPF therefore is limited to patients with a for surgical lung biopsy (96,97,192a). In a prospective
histologic pattern of UIP in whom no underlying cause multicenter study, the positive predictive value of HRCT in
is found. diagnosing IPF by using these criteria was 96% (96). In the
IPF occurs most commonly in patients older than same study, these characteristic findings were present in
50 years and is slightly more common in men (97,192). 50% of patients. In the remaining patients, the pattern was
The onset of disease is typically insidious, with gradual less characteristic, or the distribution not predominantly
onset of dry cough and progressive dyspnea (97,192). subpleural and basal. Although in this prospective study,
These symptoms typically are present for more than none of the patients with IPF had normal HRCT findings,
6 months before presentation (97,192). Physical examina- the abnormalities in early IPF can be very mild (Fig. 8-47).
tion shows finger clubbing in 25% to 50% of cases, and A common finding seen in patients with IPF is the pres-
chest auscultation reveals Velcro-type fine end-inspiratory ence of ground-glass opacity (95,193). In the majority of
crackles most prevalent at the lung bases. The clinical patients, the extent of ground-glass opacity is less than the
course is invariably one of gradual deterioration. Clinical extent of reticulation. However, approximately one third
or radiologic improvement is rare. The median length of patients with IPF have an equivalent extent of reticula-
of survival from time of diagnosis varies between 2.5 and tion and ground-glass opacity, and 10% have predominant
3.5 years (97,192). ground-glass opacity (99). In these patients, the findings
The predominant abnormality on HRCT consists of can resemble those of other interstitial lung diseases,
symmetric bilateral intralobular linear opacities, resulting particularly NSIP and HP (Figs. 8-48 and 8-49) (194).
in a fine reticular pattern involving mainly the subpleural Ground-glass opacity may indicate the presence of active
lung regions and lung bases (Table 8-10; Fig. 8-44) (95–97). inflammation and potentially treatable disease or the pres-
The intralobular linear opacities correspond his- ence of fibrosis below the resolution of HRCT (131,145).
tologically to thin irregular areas of fibrosis (50,95). Also Ground-glass opacity should be considered to represent a
noted are irregular interfaces between the lung and pul- potentially reversible process only when no associated
monary vessels, bronchi, and pleural surfaces (36,62). HRCT findings of fibrosis are noted. Findings of fibrosis in
Fibrosis results in dilatation and tortuosity of bronchi and association with ground-glass opacity, thus suggesting an
bronchioles, findings referred to as traction bronchiectasis and irreversible process, include intralobular interstitial thick-
bronchiolectasis (Fig. 8-45) (68,145). The majority of patients ening, honeycombing, and traction bronchiectasis and
A B
C D
Figure 8-44 Idiopathic pulmonary fibrosis. See Color Figure 8-44C,D. A: High-resolution CT at the level of the lower lung zones demon-
strates linear opacities separated by only a few millimeters, causing a fine reticular pattern. Also noted are traction bronchiectasis (arrow)
and honeycombing (arrowheads). The abnormalities involve mainly the subpleural lung regions and are most severe in the middle lobe and
lingula. B: Coronal reformation demonstrates that the abnormalities involve mainly the subpleural regions (straight arrows) and the lower
lung zones (curved arrows). The patient was a 70-year-old man with idiopathic pulmonary fibrosis. C: Sagittal slice of the right lung from a
different patient shows severe fibrosis and honeycombing in the basal aspects of the lower and middle lobes. Less marked disease can be
seen in the subpleural region of the anterior portion of the upper lobe and the posterior portion of the lower lobe. D: Low-magnification
photomicrograph of lung from another patient shows characteristic spatial heterogeneity of idiopathic pulmonary fibrosis. Severe fibrosis
and early honeycomb change are present in one lobule and virtually normal lung in the adjacent one. Arrows indicate interlobular septum. (C
and D from Müller NL, Fraser RS, Lee KS, et al. Diseases of the lung: radiologic and pathologic correlations. Philadelphia: Lippincott Williams
& Wilkins; 2003, with permission.)
bronchiolectasis (Fig. 8-50) (145). Patients who have present with predominant reticulation and have progres-
predominantly ground-glass opacity on HRCT are more sive disease and a poor prognosis. Areas of ground-glass
likely to respond to treatment than are patients who opacity frequently can be seen to progress to fibrosis and
have predominantly reticulation (133,195). It should be honeycombing (Fig. 8-51) (196,197). An additional
emphasized, however, that most patients who have IPF unusual manifestation of IPF is the presence of punctuate
Figure 8-46 Honeycombing in idiopathic pulmonary fibrosis.

Figure 8-45 Idiopathic pulmonary fibrosis with traction High-resolution CT shows small honeycomb cysts (arrows) in the
bronchiectasis. High-resolution CT at the level of the lower lung subpleural regions of both lungs. The honeycomb cysts range
zones shows extensive reticulation throughout the left lower lobe from a few millimeters to approximately 1 cm in diameter and
with considerable associated volume loss. The bronchi within the have thin walls. The patient was a 63-year-old man with idiopathic
areas of reticulation are dilated and distorted and have a beaded pulmonary fibrosis.
appearance (arrows), a finding characteristic of traction bronchiec-
tasis. Mild subpleural reticulation is seen in the right lung. The
patient was a 76-year-old man with idiopathic pulmonary fibrosis.
In a small percentage of patients, acute exacerbation of
IPF develops, a condition characterized by marked exacer-
bation of dyspnea and a decrease in arterial oxygen
calcifications typically identifiable in the lung bases tension (PaO2) of more than 10 mm Hg within 1 month,
peripherally, representing focal ossification. in the absence of infection or heart failure (198,199).
The clinical course of IPF is typically one of gradual Histologically, these patients have diffuse alveolar damage
deterioration over several months or years, with progres- superimposed on the interstitial fibrosis or, less com-
sion of parenchymal abnormalities on serial HRCT scans. monly, extensive active fibroblastic foci (198,200). Acute
A B
Figure 8-47 Early idiopathic pulmonary fibrosis. A: High-resolution CT (1-mm collimation) obtained on a multidetector CT shows small
foci of ground-glass attenuation and mild reticulation (arrows) in the subpleural lung regions. B: Coronal reformation demonstrates that the
abnormalities (arrows) involve almost exclusively the lower lobes. The patient was a 60-year-old woman with idiopathic pulmonary fibrosis.
Figure 8-48 Idiopathic pulmonary fibrosis with predominantly

ground-glass opacities. High-resolution CT (1.3-mm collimation)
shows extensive ground-glass opacities throughout the lower Figure 8-50 Ground-glass opacities due to fibrosis. High-
lobes. Mild fine reticulation is present in the subpleural lung resolution CT shows bilateral ground-glass opacities. Note super-
regions. The patient was a 57-year-old man with idiopathic imposed fine reticular pattern and traction bronchiectasis (arrows),
pulmonary fibrosis. indicating the presence of fibrosis. The patient was a 62-year-old
man with idiopathic pulmonary fibrosis.
exacerbation is characterized on HRCT by the rapid

development of multifocal bilateral areas of ground-glass
opacity, consolidation, or both, superimposed on a back- hilar or mediastinal lymphadenopathy (202,203). Patients
ground of interstitial fibrosis (Fig. 8-52) (198,201,201a). with IPF are also at increased risk for tuberculosis, which
In this setting, the presence of extensive areas of ground- may appear as a solitary nodule, multiple nodules, or focal
glass correlates with a poor prognosis (198,201a) areas of consolidation (204).
Consolidation and nodules are uncommon in IPF and Enlarged mediastinal lymph nodes are present in 70% to
should suggest the presence of complications such as 90% of patients with IPF (205,206). They usually measure
superimposed infection or pulmonary carcinoma. Risk of between 10 and 15 mm in short axis diameter and involve
lung cancer is increased in patients who have IPF and can only one or two nodal stations (206). The prevalence of
appear as a nodule, as a focal area of consolidation, or as enlarged nodes correlates with the extent of parenchymal
abnormalities on HRCT (207). In the majority of patients,
the enlarged nodes become normal in size after administra-
tion of corticosteroids (208).
Asbestosis
Asbestosis is defined as interstitial fibrosis caused by
inhalation of asbestos fibers (209). The severity and rate of
progression of asbestosis correlate with exposure level
(190,209). It characteristically occurs 15 to 20 years after
exposure, with disease progressing even after exposure has
ceased. Initially, asbestosis affects respiratory bronchioles
with development of peribronchiolar fibrosis (94,210,211).
As the fibrosis progresses, it involves the alveolar walls
throughout the lobule and, eventually, the interlobular
septa. Honeycombing can be seen in advanced cases. The
fibrosis tends to be patchy and is usually most severe in
Figure 8-49 Idiopathic pulmonary fibrosis with predominately the subpleural regions of the lower lobes and posterior
ground-glass opacities. High-resolution CT (1-mm collimation) lungs (211). The fibrosis histologically resembles UIP, the
shows extensive bilateral ground-glass opacities throughout the
lower lobes. Localized lobular areas of normal or decreased atten- two main differences being the presence of asbestos bodies
uation (arrows) are present, a finding commonly seen in hypersen- and a paucity of fibroblastic foci in asbestosis (190).
sitivity pneumonitis. Mild reticulation is seen in the subpleural lung As pathologic evaluation is not usually obtained in
regions. No evidence of hypersensitivity pneumonitis was identi-
fied at surgical biopsy or clinical and laboratory evaluation. The individual cases, assessing pulmonary parenchymal fibro-
patient was a 67-year-old man with idiopathic pulmonary fibrosis. sis has necessitated the development of a constellation of
A B
Figure 8-51 Progression of ground-glass opacities to reticulation. A: High-resolution CT (1.3-mm collimation) shows bilateral ground-
glass opacities. Mild reticulation is present mainly in the subpleural regions. B: High-resolution CT at the same level 2 years later shows
replacement of ground-glass opacities by reticulation, indicating progression to fibrosis. Clinically the patient had deteriorated, with devel-
opment of more severe dyspnea and restrictive lung function. The patient was a 57-year-old woman with idiopathic pulmonary fibrosis.
clinical, functional, and radiographic findings to establish 200 admission chest radiographs interpreted according to
the diagnosis of asbestosis in vivo. The diagnosis can be the ILO classification of diffuse lung disease, 22 (11%)
inferred when a reliable history of nontrivial exposure to patients without occupational exposure had abnormali-
asbestos is known along with a combination of (a) restric- ties that otherwise might have been interpreted as indica-
tive lung disease on pulmonary function testing as well as tive of significant dust exposure (213). Additionally,
a diffusion abnormality; (b) the presence of rales at aus- radiographic–pathologic correlative studies have shown
cultation; and (c) abnormal chest radiograph, defined as that interstitial fibrosis may be present despite normal-
an International Labour Office (ILO) profusion (severity) appearing chest radiographs in up to 20% of patients
score of 1/1 or greater (212). It has been suggested that (214–216). Added to these difficulties are significant
among these, the findings on the chest radiograph are the problems with interobserver variability, specifically in the
most important (212). interpretation of opacities of mild profusion (217).
Unfortunately, there are limitations to the value of Several studies have shown that HRCT is more sensitive
radiographic interpretation in defining a population with than the radiograph in the detection of asbestosis
asbestos-induced abnormalities (Fig. 8-53). In a review of (53,218,219). In one study of 169 asbestos-exposed
A B
Figure 8-52 Acute exacerbation of idiopathic pulmonary fibrosis (IPF). A: High-resolution CT (1-mm collimation) at the level of the upper
lobes shows bilateral reticular pattern in the subpleural regions. The patient had mild symptoms. B: High-resolution CT at the same level
7 months later, when the patient had sudden worsening of symptoms and severe hypoxemia, shows extensive bilateral ground-glass opaci-
ties and small areas of consolidation. Surgical biopsy demonstrated diffuse alveolar damage superimposed on usual interstitial pneumonia,
a characteristic finding of acute exacerbation of IPF. The patient was a 64-year-old woman with idiopathic pulmonary fibrosis.
B C
Figure 8-53 Asbestosis. A: Posteroanterior radiograph shows minimal nodularity of the right hemidiaphragm. The lungs appear normal.
B, C: Sequential 1.5-mm target reconstructed sections through the right lower lobe obtained with the patient in a prone position show sub-
tle parenchymal changes consistent with asbestosis. In addition to pleural plaques, intralobular lines and dots (straight arrows in B and C)
are identifiable. These result in a prominent subpleural curvilinear line posteromedially (curved arrows in B and C).
individuals with normal or near normal chest radi- abnormal plain film findings (220). Based on the evidence
ographic findings (ILO score, 0/0 or 0/1), 57 (34%) had in the literature, a statement of the American Thoracic
findings suggestive of asbestosis on HRCT (219). HRCT Society in 2004 on the diagnosis of asbestosis and other
can also be valuable for eliminating the diagnosis of as- benign asbestos-related diseases emphasized that HRCT
bestosis in patients who have chest radiographic findings plays an important role “when radiographic findings are
that suggest this disease, but who have some other abnor- equivocal, when diminished pulmonary function is identi-
mality. Approximately 20% of patients suspected of having fied in association with otherwise normal plain chest
interstitial disease on the basis of plain radiographs have radiographic findings, and when extensive overlying pleu-
been shown by HRCT to have emphysema, prominent ves- ral abnormalities do not allow a clear interpretation of the
sels, pleural disease, or bronchiectasis as the cause of their parenchymal markings” (209).
In many cases, the fibrosis in asbestosis is mild and lim-

ited to the dependent regions of the lower lung zones. It
may be difficult, if not impossible, to distinguish mild
fibrosis in these areas from normal increase in attenuation
and atelectasis in the dependent lung. Therefore, it is
recommended that in patients being assessed for asbestosis,
HRCT scans be obtained in both supine and prone
positions (13,221). Although in most cases the scans should
be obtained at 10- or 20-mm intervals, a limited technique
consisting of four or five equally spaced scans obtained
from the level of the tracheal carina to the diaphragm has
been shown to have good sensitivity in the detection of
asbestosis and may be appropriate for screening (222).
Accurate depiction of asbestosis and asbestos-related pleural
disease has also been shown by using low-dose technique
(60 to 100 mAs) with automatic dose modulation and
continuous spiral CT through the chest performed on a Figure 8-54 Early parenchymal abnormalities in an asbestos-
exposed individual. High-resolution CT (1.5-mm collimation)
multidetector scanner (223). The main advantage of using performed with the patient prone shows subpleural dotlike and
continuous multidetector CT scanning is that it allows con- branching opacities (straight arrows). Confluence of dotlike opaci-
current search for malignant asbestos-related processes of ties results in subpleural curvilinear opacity (curved arrow). These
findings represent the earliest parenchymal abnormality in individ-
the lungs and pleura in these high-risk individuals (223). uals with asbestos exposure and reflect the presence of peribron-
Asbestosis can usually be readily recognized on HRCT chiolar fibrosis. Peribronchiolar fibrosis in patients with asbestos
by the presence of fibrosis associated with pleural plaques exposure is not by itself sufficient for a diagnosis of asbestosis.
The patient was a 56-year-old man with history of asbestos expo-
or diffuse pleural thickening (Table 8-11). The earliest pul- sure. (Case courtesy of Dr. Jorge Kavakama, São Paulo, Brazil.)
monary abnormalities on HRCT in patients with asbestos
exposure consist of subpleural nodules or dotlike opacities
1 mm or less in diameter (224). The nodules are situated a (209). It does not consider lesions of the membranous
few millimeters away from the pleura and occasionally bronchioles to be asbestosis. The mildest form of asbestosis
appear as fine branching structures (Fig. 8-54) (224). The histologically consists of involvement of the alveolated
subpleural nodules reflect the presence of centrilobular, walls of respiratory bronchioles and the alveolar ducts
peribronchiolar fibrosis (94,224). Although peribronchio- (209). Confluence of nodules leads to pleura-based nodu-
lar fibrosis is the earliest parenchymal finding seen in lar irregularities and subpleural curvilinear lines (Figs. 8-54
patients with asbestos exposure, it is not considered and 8-55) (224). As the fibrosis progresses, extending from
asbestosis. The 2004 statement of the American Thoracic the peribronchiolar region to involve the alveolar walls,
Society defines asbestosis specifically as interstitial fibrosis intralobular lines and irregular thickening of the interlobu-
caused by the deposition of asbestos fibers in the lung lar septa are seen on HRCT (Figs. 8-56 and 8-57) (224).
TABLE 8-11
FINDINGS IN ASBESTOSIS
Characteristic manifestations
Subpleural dotlike opacities in early disease (peribronchiolar
fibrosis)
Intralobular linear opacities (reticulation)
Subpleural lines
Parenchymal bands
Traction bronchiectasis
Subpleural and lower lung zone predominance
Parietal pleural thickening or plaques
Other findings Figure 8-55 Subpleural curvilinear lines in an asbestos-exposed
Honeycombing in advanced disease individual. High-resolution CT (1.5-mm collimation) performed
Traction bronchiolectasis with the patient prone shows curvilinear lines (curved arrows)
Ground-glass opacity in nondependent lung and dotlike opacities (straight arrows). The patient was a 68-year-
(relatively uncommon) old man with a history of asbestos exposure. (Case courtesy of
Dr. Jorge Kavakama, São Paulo, Brazil.)
bands are linear densities 2 to 5 cm in length coursing

through the lungs and usually attaching to the pleura
(Fig. 8-59) (53,94). Parenchymal bands may be caused by
thickened septa marginating several secondary lobules,
fibrosis along the bronchovascular sheaths, coarse scars, or
areas of atelectasis (94,227). Although parenchymal bands
are common in asbestosis, being seen in 60% to 80% of
patients (13,228), they are not specific, being present in
10% to 20% of patients with various lung diseases not
related to asbestos (227,228).
Figure 8-56 Asbestosis. High-resolution CT (1.5-mm collima- The HRCT findings of asbestosis resemble those of IPF.
tion) performed with the patient prone shows subpleural line The main differences are the more-marked basal and
(straight arrow), parenchymal band (arrowhead), nondependent subpleural distribution of disease in asbestosis (229) and
ground-glass opacities, and a few irregular linear opacities (curved
arrows). These findings are consistent with asbestosis. Also noted the greater prevalence of subpleural dotlike or branching
are bilateral diaphragmatic pleural plaques. The patient was a opacities, parenchymal bands, and subpleural lines and
71-year-old man with asbestosis. lower prevalence of honeycombing and traction bronchi-
olectasis in asbestosis (92). Ground-glass opacities are rela-
tively uncommon in patients with asbestosis and when
These findings reflect the presence of interstitial fibrosis present usually are not extensive (228). They correlate with
(i.e., asbestosis). With further progression of fibrosis, the presence of mild alveolar wall fibrosis or edema (94).
architectural distortion, traction bronchiectasis, and even- The diagnosis of asbestosis on HRCT is based on clini-
tually honeycombing are seen (Fig. 8-58) (93,225). cal history of exposure, findings indicative of bilateral
Pathologically and on HRCT, the findings are most severe pulmonary fibrosis and bilateral pleural plaques or diffuse
in the subpleural lung regions of the lower lung zones pleural thickening. The specificity of the diagnosis
(Fig. 8-57) (190,209). Other common findings in asbesto- increases with the number of parenchymal abnormalities
sis include curvilinear subpleural lines and parenchymal present on HRCT (225). In a study correlating HRCT and
bands. In patients with asbestosis, curvilinear subpleural histopathologic findings, in 25 patients with asbestosis
lines may be caused by early peribronchiolar fibrosis and 5 without, the HRCT findings in patients with con-
combined with collapse of alveoli caused by fibrosis (224), firmed asbestosis consisted of interstitial lines (84%),
fine honeycombing (226), or atelectasis (Fig. 8-55) (220). parenchymal bands (76%), architectural distortion of
They are, however, a nonspecific finding, being seen in the secondary pulmonary lobules (56%), subpleural lines
dependent lung regions in up to 20% of subjects without (44%), and honeycombing (32%) (225). The likelihood
history of exposure to asbestos (53,227). Parenchymal of asbestosis being present histopathologically increased
A B
Figure 8-57 Asbestosis. A: High-resolution CT (1-mm collimation) performed on a multidetector CT scanner shows irregular linear opaci-
ties (arrows) in the subpleural lung regions. B: Coronal reformation demonstrates predominant distribution of the irregular linear opacities in
the subpleural regions and lower lung zones. Also noted is diffuse left pleural thickening. The patient was a 76-year-old man with asbestosis.
mediastinal and hilar lymph nodes (231,232). It is most

commonly secondary to adenocarcinoma arising from the
breast, lung, gastrointestinal tract, or prostate, or less com-
monly from an unknown primary (231).
The characteristic abnormalities consist of thickening of
the interlobular septa and thickening of the peribron-
chovascular and subpleural interstitium with preservation
of normal lung architecture (Fig. 8-60) (64,65,233). The
thickened interlobular septa are most commonly identi-
fied as lines 1 to 2 cm in length contacting the pleural
surface or as polygonal arcades in the more central lung
regions (36,64,65). The thickened septa may be smooth or
have a beaded or nodular appearance caused by tumor
growth along the lymphatics (Fig. 8-61) (64,65,233).
Similarly, the thickening of the peribronchovascular inter-
stitium may be smooth or nodular (Fig. 8-62) (64).
Figure 8-58 Asbestosis with honeycombing. High-resolution
CT in a 59-year-old patient with advanced asbestosis demon- Thickening of the interstitium along the centrilobular
strates extensive bilateral honeycombing involving mainly the sub- artery is frequently seen in the areas with interlobular
pleural lung regions. septal thickening, resulting in a prominent centrilobular
dot. Smooth or nodular thickening of the subpleural
interstitium is also frequently present, being particularly
from 60% in patients with one of these abnormalities on well seen in the region of the interlobar fissures. Discrete
HRCT, 80% in patients with two, and 100% in patients small nodules may also be present (36,83). A distinctive
with three or more abnormalities. However, the sensitivity feature of lymphangitic carcinomatosis is the preservation
decreased from 88% for a diagnosis based on the presence of normal lobular architecture, allowing easy distinction
of only one of the five abnormalities, to 78% for two from pulmonary fibrosis. Although lymphangitic carci-
abnormalities, and 56% for three abnormalities (225). nomatosis is most commonly bilateral and diffuse,
The extent of parenchymal abnormalities on HRCT corre- it may occasionally be focal or unilateral in distribution
lates with the severity of functional impairment (230). (Fig. 8-63) (64,65).
Abnormalities characteristic of lymphangitic carcino-
matosis have been described on HRCT in patients with
Lymphangitic Carcinomatosis
normal or nonspecific findings on chest radiographs
Pulmonary lymphangitic carcinomatosis refers to the (64,65). In our experience, the presence of thickened and
spread of tumor within pulmonary lymphatics. It usually beaded interlobular septal lines on HRCT in the clinical
results from a hematogenous spread to lung, with subse- setting of known malignancy is essentially diagnostic of
quent interstitial and lymphatic invasion, but can also lymphangitic carcinomatosis. Although similar findings
occur because of direct lymphatic spread of tumor from have been described in cases of viral pneumonia and,
Figure 8-59 Parenchymal bands. High-resolution CT (1-mm colli-

mation) shows left upper lobe parenchymal bands (straight arrows) Figure 8-60 Lymphangitic carcinomatosis. High-resolution CT
coursing toward the pleura. Note associated volume loss with (1-mm collimation) shows bilateral thickening of the interlobular
forward displacement of the left major fissure. Irregular linear opaci- septa (arrows). The septal thickening is smooth and involves
ties and irregular septal thickening (curved arrows) are present in the mainly the anterior aspects of the upper lobes. The patient was a
right upper lobe. The patient was a 65-year-old man with asbestosis. 52-year-old man with metastatic adenocarcinoma of the colon.
A B
Figure 8-61 Lymphangitic carcinomatosis. See Color Figure 8-61B. A: High-resolution CT (1-mm collimation) shows nodular thickening of
the interlobular septa (arrows), more extensive in the left lung. The patient was a 74-year-old man with metastatic adenocarcinoma, primary
unknown. B: Pathologic section from another patient with lymphangitic carcinomatosis shows nodular thickening of interlobular septa due to
tumor infiltration (straight arrows). Also note tumor growth in the interstitium along the centrilobular bronchovascular bundles (curved arrow).
more important, have been reported to occur in patients involves primarily the lung and lymphatic systems of the
with sarcoidosis, when identified in the appropriate clini- body (235). Pulmonary manifestations are present in
cal setting, these changes can be sufficiently characteristic more than 90% of patients, in at least 10% to 20% of
in selected cases to obviate either transbronchial or surgi- whom permanent sequelae develop (235).
cal biopsies (35). Approximately 60% to 70% of patients with sarcoidosis
have a characteristic radiologic appearance consisting of
bilateral hilar and paratracheal lymphadenopathy with
Diseases Characterized Primarily
or without concomitant parenchymal abnormalities (236–
by Nodular Opacities 238). In 25% to 30% of cases, however, the radiologic
findings are nonspecific or atypical, and in 5% to 10% of
Sarcoidosis patients, the chest radiograph is normal despite the pres-
Sarcoidosis is a multisystem disorder of unknown etiology ence of pulmonary granulomas (190,236,239).
characterized by the presence of widespread, noncaseating The variable and often nonspecific radiographic findings
granulomas (190,234). The granulomas may resolve are surprising, given the characteristic pathologic appear-
spontaneously or progress to fibrosis. Sarcoidosis most ance and distribution of sarcoidosis. Sarcoid granulomas,
commonly affects patients between 20 and 40 years of age the hallmark of the disease, are distributed mainly along
and shows a slight female predominance (190,234,235). It the lymphatics in the bronchovascular sheath and, to a
lesser extent, in the interlobular septa and subpleural lung
regions (50,190,234). This distribution is one of the most
helpful features in recognizing sarcoidosis pathologically
Figure 8-62 Lymphangitic carcinomatosis. High-resolution CT Figure 8-63 Unilateral lymphangitic carcinomatosis. High-
(1.5-mm collimation) demonstrates marked thickening of the peri- resolution CT (1-mm collimation) shows thickening of the interlobular
bronchial and perivascular interstitium of the right lung compared septa (arrows) in the left upper and lower lobes. Incidental note is
with the left lung. The patient was a 67-year-old woman with lym- made of mild emphysema. The patient was a 57-year-old man with
phangitic carcinomatosis involving almost exclusively the right lung. metastatic adenocarcinoma.
and is responsible for the high rate of success in diagnosis

by bronchial and transbronchial biopsies (190). This distri-
TABLE 8-12
bution is difficult to appreciate on the radiograph because
of the superimposition of parenchymal structures but can
be clearly identified on HRCT (50,67,240).
FINDINGS IN SARCOIDOSIS
The characteristic CT findings of sarcoidosis consist of Characteristic manifestations
small nodules and nodular thickening along the bron- Small peribronchial and perivascular nodules
chovascular bundles, interlobular septa, interlobar Small nodules along interlobar fissures and interlobular septa
Upper lobe and perihilar predominance
fissures, and subpleural lung regions (see Figs. 8-3 and
Symmetric hilar and mediastinal lymphadenopathy
8-64) (Table 8-12) (67,85,240). The thickening of the
interstitium surrounding vessels and bronchi is most Other common findings
Small subpleural nodules
marked in the perihilar regions of the middle and upper Large nodules (1 cm) or consolidation
lung zones (Figs. 8-3 and 8-64). Thickening of the peri- Ground-glass opacity
bronchovascular interstitium is also present in the Findings of fibrosis: traction bronchiectasis, honeycombing
centrilobular regions, leading to centrilobular nodules Conglomerate masses associated with traction bronchiectasis
Figure 8-64 Sarcoidosis. See Color Figure 8-64C. A: High-

resolution CT performed on a multidetector scanner at the level of
the left upper lobe bronchus shows nodular thickening along the
bronchi (straight arrow), vessels (curved arrows), and interlobar fis-
sures (arrowheads). Also noted is subcarinal and hilar lymphadenopa-
thy. B: Coronal reformation demonstrates that the abnormalities are
located mainly in the middle and upper lung zones. Nodular thicken-
ing is seen along the bronchi (straight arrows) and vessels (curved ar-
rows). Also noted are small nodules along the major interlobar fis-
sures (white arrowheads), right minor fissure (black arrowhead), and
costal pleura. The patient was a 28-year-old man with sarcoidosis. C:
Pathologic specimen from another patient shows sarcoid granulo-
mas along interlobular septa (straight arrows), bronchovascular
B sheaths (curved arrow), and subpleural region (arrowhead).
and branching structures. The nodules seen on HRCT rep- Although the distribution of abnormalities in sarcoido-
resent coalescence of microscopic granulomas and usually sis is typically perilymphatic, the appearance can be quite
have irregular margins (50,67). Irregular or nodular thick- variable (Fig. 8-65). The nodules may be small and result
ening of the interlobular septa is frequently present but is in a diffuse miliary pattern or involve mainly the periph-
usually not extensive. eral lung regions (Fig. 8-66). Confluence of numerous
A B
C D
Figure 8-65 The many faces of sarcoidosis. A: High-resolution CT (HRCT) target reconstruction through the right upper lobe shows a
small cluster of peribronchovascular nodules (arrow) associated with nodularity of the major fissure. This pattern, although superficially
resembling tree-in-bud, is characteristic of early sarcoidosis. B: HRCT target reconstruction in a different patient than in A shows greater
profusion of nodules, which nonetheless are still clearly clustered in a distinctive peribronchovascular distribution. Note the absence of
obvious secondary lobular septal thickening, although sarcoidosis is a perilymphatic disease. C: HRCT through the upper lung zones demon-
strates irregular nodules in the peribronchial (straight arrows) and subpleural lung regions, particularly along the interlobar fissures (curved
arrows). The patient was a 32-year-old man with stage 2 sarcoidosis. D: HRCT target reconstruction through the left upper lobe shows a
miliary pattern. Note that despite marked profusion, still these nodules tend to cluster, especially peripherally.
Figure 8-68 Consolidation in sarcoidosis. High-resolution CT

(1 mm) performed at the level of the tracheal carina shows dense
symmetric bilateral upper lobe consolidation and a few small nod-
ules. The patient was a 62-year-old woman with sarcoidosis.
Figure 8-66 Micronodules in sarcoidosis. High-resolution CT
(1.5 mm) performed at the level of the lower lung zones with the
patient prone shows numerous micronodules in the outer third submucosal granulomas. The air trapping is usually
of both lungs and along the interlobar fissures. The patient was a mild and contributes little to functional impairment in
47-year-old man with sarcoidosis. (Case courtesy of Dr. Isabela sarcoidosis (245).
Silva, Salvador, Brazil.)
As fibrosis develops, irregular reticular opacities,
including intralobular lines and irregular septal thicken-
granulomas may lead to nodules or large opacities measuring, become a prominent feature (67,240). Fibrosis is
ing 1 to 4 cm in diameter (Fig. 8-67). These are seen in associated with distortion of lung architecture and char-
15% to 25% of patients and frequently contain air acteristic posterior displacement of the upper lobe
bronchograms (97,102,240,241). Areas of ground-glass bronchi, indicating loss of volume in the posterior seg-
attenuation may be present in patients with sarcoidosis. ments of the upper lobes (Fig. 8-70) (240,246). Fibrosis
Correlation with pathologic specimens has shown that also leads to abnormal conglomeration of dilated dis-
these represent extensive interstitial sarcoid granulomas torted parahilar bronchi, a finding known as traction
below the resolution of HRCT rather than alveolitis bronchiectasis (Fig. 8-71) (68,246). The fibrosis involves
(102,131,242). Similarly, areas of consolidation (“alveolar mainly the upper lobes, resulting in superior retraction of
sarcoid”) also represent conglomeration of interstitial the hila and bronchi and compensatory overinflation of
granulomas (Fig. 8-68) (242). the lower lobes (Fig. 8-70). Honeycombing may also be
Expiratory CT images in patients with sarcoidosis seen but is usually mild and present only in patients with
frequently show patchy air trapping at the level of the severe fibrosis and central conglomeration of bronchi
secondary lobules (Fig. 8-69) (243,244). The air trapping
is presumably secondary to bronchiolar obstruction by
Figure 8-67 Large nodules in sarcoidosis. High-resolution CT Figure 8-69 Air trapping in sarcoidosis. High-resolution CT
(1.5 mm), performed at the level of the aortic arch, shows large (1 mm) performed at the level of the dome of the right hemidi-
irregular nodules and masses along the bronchi (straight arrows), aphragm demonstrates bilateral areas of air trapping (arrows).
vessels, and interlobular septa (curved arrows). A few small nod- Also noted are several small nodules, mainly in the subpleural
ules can be seen adjacent to the major fissures (arrowheads). The regions (arrowheads). The patient was a 50-year-old woman with
patient was a 59-year-old man with sarcoidosis. sarcoidosis.
A B
Figure 8-70 Fibrosis and active disease in sarcoidosis. A: High-resolution CT performed on a multidetector scanner shows posterior
displacement of the upper lobe bronchi (arrows), which are dilated and have a beaded appearance characteristic of traction bronchiectasis.
Also noted are extensive peribronchial thickening, architectural distortion, ground-glass opacities, and a few small nodules. B: Coronal ref-
ormation demonstrates cephalad displacement of the bronchi (straight arrows) due to the predominantly upper lobe distribution of fibrosis
in sarcoidosis. The upper lobe volume loss is associated with tenting of both hemidiaphragms (arrowheads). Also noted are extensive
ground-glass opacities, a few small nodules, and nodular thickening of the interlobular septa (curved arrow). These findings are consistent
with active disease. The patient was a 53-year-old man with sarcoidosis.
(85). It is subpleural in distribution and usually involves with only hilar adenopathy apparent on the radiograph
mainly the middle and upper lung zones, with relative (Fig. 8-72) (67,240). CT may also be helpful in assessing
sparing of the lung bases (93). Occasionally the honey- the presence and extent of complications of sarcoidosis
combing can be diffuse or involve mainly the lower lobes (249). Superimposed bacterial infection and mycetomas
and resemble IPF (246,247). within cavities and areas of bronchiectasis are more readily
Enlarged hilar and mediastinal nodes are seen on CT in detected on CT than on the radiograph (249).
80% to 90% of patients with sarcoidosis, a higher prevalence The distribution of sarcoid granulomas along the lym-
than is apparent on the chest radiograph (67,206,248). phatics is similar to that seen in pulmonary lymphangitic
Calcification of hilar and mediastinal lymph nodes is rela- carcinomatosis. Both conditions may cause a beaded
tively common, being seen in eight (44%) of 18 patients thickening of the bronchovascular bundles and interlobu-
with long-standing sarcoidosis in one study (120). lar septa (65,85). However, in sarcoidosis, the septal thick-
CT may demonstrate parenchymal abnormalities in ening is usually less extensive than that seen in pulmonary
patients with a normal chest radiograph and in patients lymphangitic carcinomatosis, characteristically involves
mainly the middle and upper lung zones, and is often as-
sociated with distortion of the lobular architecture
(67,250). The nodules in sarcoidosis often have irregular
margins, whereas in lymphangitic carcinomatosis, they are
smooth. In some patients, however, the pattern of
parenchymal involvement may be similar (6).
Silicosis and Coalworkers’ Pneumoconiosis

Silicosis is a chronic lung disease characterized by the
development of progressive parenchymal nodules and pul-
monary fibrosis secondary to inhalation of crystalline free
silica (190). It is seen most commonly in heavy metal or
hard rock miners. Prolonged exposure results in the forma-
tion of small, discrete, hyalinized nodules predominantly
in the upper lobes. The nodules are sharply defined, tend to
Figure 8-71 End-stage fibrosis in sarcoidosis. High-resolution localize around respiratory bronchioles, and are clearly sep-
CT demonstrates perihilar cystic appearance caused by markedly
ectatic bronchi. Central conglomeration of ectatic bronchi caused arable from surrounding alveoli, which may either be nor-
by perihilar fibrosis is characteristic of end-stage sarcoidosis. mal or show evidence of emphysema (234). The diagnosis
A B
C D
Figure 8-72 Sarcoidosis: clinical role of CT. A: Posteroanterior chest radiograph shows typical pattern of bilateral hilar and right paratra-
cheal lymphadenopathy in this asymptomatic 20-year-old woman. A slight nonspecific asymmetric accentuation of markings appears in the
right lung. B: One-millimeter section shows characteristic appearance of clusters of small peribronchovascular and subpleural nodules.
Despite asymmetry, these findings in conjunction with the radiograph and clinical history were sufficient to obviate biopsy. Clinical follow-up
after steroids confirmed resolution of findings consistent with clinical diagnosis of sarcoidosis. C, D: One-millimeter sections imaged with
lung and mediastinal windows, respectively, in a different patient than shown in A and B. In this case, despite mild thickening of the right
upper lobe bronchial wall, routine forceps transbronchial biopsy was nondiagnostic. Follow-up transbronchial (Wang) core needle biopsy,
however, yielded noncaseating granulomata.
of silicosis is usually made by correlating clinical history centrilobular in location (190). The characteristic radiologic
with characteristic plain radiographic findings, demons- findings in patients with coalworkers’ pneumoconiosis are
trating either simple silicosis (characterized by multiple similar to those of silicosis and consist of small, well-
small nodular opacities) or complicated silicosis (charac- circumscribed nodules usually measuring 2 to 5 mm in
terized by large coalescent pulmonary opacities resulting in diameter and involving mainly the upper and posterior lung
so-called progressive massive fibrosis). Hilar lymph node zones (232,253). Large opacities, also known as conglomer-
enlargement and calcification are often evident on the radiate masses or progressive massive fibrosis, may be seen.
ograph. On CT, evidence of mediastinal nodal calcification On HRCT, as on the radiograph, the most characteristic
is seen in up to 95% of patients (251). In the vast majority findings in simple silicosis and coalworkers’ pneumoco-
of cases, the nodal calcification is diffuse or speckled (251). niosis consist of small nodules that are seen predomi-
In 5% or fewer of patients with silicosis, the nodal calci- nantly in the posterior aspects of the upper lung zones
fication is peripheral (“egg-shell”) (251,252). Although (Fig. 8-73) (88,254–256). These nodules are characteristi-
relatively uncommon, egg-shell nodal calcification is cally in a centrilobular distribution, although they may
highly suggestive of silicosis (252). also frequently be seen in the subpleural lung regions
Coalworkers’ pneumoconiosis results from inhalation of (88,254,255). The nodules vary in size but usually meas-
coal dust. Histologically, the characteristic findings consist ure between 2 to 5 mm in diameter and may be calcified.
of coal dust macules and focal emphysema, both being Progressive massive fibrosis in silicosis and coalworkers’
required for diagnosis (190). Macules correspond to the pneumoconiosis is always associated with a background of
accumulation of pigmented macrophages in the walls of res- small nodules visible on CT (88,256). The conglomerate
piratory bronchioles and adjacent alveoli and are therefore masses are usually oval, have irregular margins, and are
A B
Figure 8-73 Silicosis. A: High-resolution CT performed on a multidetector scanner shows well-defined bilateral upper lobe nodules most
numerous in the dorsal half of the lungs. Note early conglomeration of nodules (arrows). Also seen are a few subpleural nodules and
emphysema. B: Sagittal reformation of the right lung better demonstrates that the nodules involve mainly the upper and middle lung zones
and are most numerous in the dorsal half of the lung. The patient was a 77-year-old man with silicosis.
associated with distortion of lung architecture and fre- silicosis and coalworkers’ pneumoconiosis (87,118,257).
quently surrounding emphysema (Fig. 8-74). In one study, 13 (41%) of 32 patients exposed to silica
HRCT has been shown to be superior to chest radiogra- who had chest radiographs interpreted as normal had
phy in the detection of small nodules in patients with evidence of silicosis on HRCT (118). In a second study,
A B
Figure 8-74 Progressive massive fibrosis in silicosis. A: High-resolution CT (HRCT) targeted to the right upper lobe demonstrates con-
glomerate mass of fibrosis with surrounding architectural distortion and emphysema. Also noted are a few subpleural nodules. B: HRCT
targeted to the right mid lung zone demonstrates a few nodules mainly in the right lower lobe. Also noted is evidence of emphysema.
nodules were detected on HRCT in 11 (23%) of 48 miners granulomatous lung disease caused by inhalation of anti-
exposed to coal dust who had no evidence of pneumoco- gens contained in a variety of organic dusts (190,262). It is
niosis on chest radiographs (ILO profusion score 1/0) most commonly caused by thermophilic actinomycetes
(257). Because of the relatively high false-negative and present in moldy hay (farmer’s lung), and contaminated
false-positive radiographic interpretations, particularly in humidifiers or air conditioners (humidifier lung) or pro-
the detection of early disease, it has been suggested that teins contained in birds’ serum, droppings, and feathers
HRCT be used instead of chest radiography as a standard (bird fancier’s lung) (190).
screening method to distinguish between normal and Heavy exposure to the inciting antigen may lead to the
early coalworkers’ pneumoconiosis (257). acute form of HP, characterized by dyspnea and fever
CT also can provide useful information regarding the developing within 4 to 8 hours after exposure (190,262).
stage of the disease in patients with silicosis and coalwork- The chest radiograph may be normal or show bilateral
ers’ pneumoconiosis by allowing detection of coalescence consolidation. The consolidation resolves within a few
of nodules and the development of conglomerate masses days. The radiograph may return to normal or reveal a fine
that may not be apparent on plain radiographs (254,256, nodular or reticulonodular pattern characteristic of the
258). It is also useful in the assessment of superimposed subacute phase of HP (190,263). In patients with heavy
complications such as silicotuberculosis. exposure, the diagnosis is readily made by clinical history
and identification of serum precipitins (262). In many
patients, however, the presentation is more insidious, with
Hematogenous Metastases
repeated exposures to relatively low doses of antigen lead-
In most patients, hematogenous tumor metastases to the ing to slowly progressive dyspnea over several weeks or
lungs result in the presence of localized tumor nodules months (subacute HP) (190,262). Continued, usually
rather than the extensive lymphatic invasion seen in lym- low-level, antigen exposure over months or years may
phangitic carcinomatosis. Hematogenous metastases usually result in pulmonary fibrosis (chronic HP). The diagnosis
result in multiple, large, well-defined nodules. Less com- of subacute and chronic HP is often not suspected clini-
monly, they may appear as multiple small nodules, mimick- cally before HRCT or lung biopsy (190).
ing the appearance of diffuse interstitial disease on chest The most common histologic features of HP are cellular
radiographs. On HRCT, as on the radiograph, hematogenous bronchiolitis and peribronchiolar interstitial alveolitis
metastases tend to involve mainly the lower lung zones (190). Poorly defined, non-necrotizing granulomas are
and frequently have a peripheral distribution (259,260). found in about two thirds of cases. Except for the presence
Hematogenous metastases are usually randomly distributed of granulomas, the histologic findings can resemble those
with respect to lobular anatomy (Fig. 8-28) (117). of NSIP (190). Progression to fibrosis may result in a his-
Assessment for pulmonary metastases requires volu- tologic picture resembling UIP or end-stage fibrosis (190).
metric scanning through the lungs. On multidetector CT The fibrosis tends to involve mainly the middle and lower
scanners, the volumetric scanning can be performed by lung zones (264,265).
using HRCT technique (0.5- to 1-mm-thick sections) or HRCT has been shown to be particularly helpful in the
thicker sections. It is important to note, however, that with assessment of patients with subacute HP (266,267).
thicker sections, metastases may commonly be missed. In Characteristic HRCT findings include bilateral ground-
one study of five autopsy lungs, 1,013 metastatic 0.5- to glass opacities and poorly defined centrilobular nodules
30-mm nodules were detected at helical CT with 1-mm (Table 8-13) (266–268). These correlate with the presence
collimation and histopathologically diagnosed as metas- of interstitial alveolitis and cellular bronchiolitis, respec-
tases (261). The same lungs were assessed by using 5-mm tively (267). The areas of ground-glass attenuation may be
section thickness scans obtained on a multidetector row diffuse but frequently have a mid and lower lung zone
CT at various sets of helical pitch and table speed and predominance (Fig. 8-75) (141,266). Localized areas of
using 5-mm collimation on a single-detector row CT scan- decreased attenuation and perfusion are often present in
ner. Regardless of varying pitch or detector collimation, conjunction with areas of ground-glass attenuation. The
multi- and single-detector row CT scans obtained with areas of low attenuation in HP often have a lobular distri-
5-mm section thickness had similar sensitivity and bution and show air trapping on CT scans performed at
detected only approximately 65% of the nodules (261). end-expiration (Fig. 8-76) (149). Ground-glass opacities
may also be seen in conjunction with centrilobular nodules
(Fig. 8-77) (15,266). The centrilobular nodules are typi-
cally poorly defined, measure less than 5 mm in diameter
by Ground-Glass Opacity and are also most numerous in the middle and lower lung
zones. In some patients, centrilobular nodules may be the
Hypersensitivity Pneumonitis only finding or the predominant abnormality seen in HP
Hypersensitivity pneumonitis (HP) (extrinsic allergic alve- (Fig. 8-78) (15,268). Approximately 10% of patients with
olitis) is an immunologically mediated diffuse interstitial subacute HP have thin-walled lung cysts ranging from 3 to
TABLE 8-13
FINDINGS IN HYPERSENSITIVITY PNEUMONITIS
Subacute
Patchy bilateral or diffuse ground-glass opacities
Small ill-defined centrilobular nodules
Lobular areas of decreased attenuation
Lobular areas of air trapping on expiratory scans
Thin-walled cysts (in 10% of patients)
Chronic
Findings of fibrosis
Intralobular linear opacities
Irregular interlobular septal thickening
Honeycombing
Traction bronchiectasis or bronchiolectasis
Ground-glass opacity and/or centrilobular nodules
Figure 8-75 Hypersensitivity pneumonitis. High-resolution CT
Patchy distribution of abnormalities image of the lower lung zones shows extensive bilateral ground-
Middle or lower lung zone predominance of fibrosis glass opacities. Also noted are lobular areas of decreased attenua-
tion and vascularity (arrows). The patient was a 55-year-old man
with hypersensitivity pneumonitis.
25 mm in maximal diameter and 1 to 15 in number in a architectural distortion, and, eventually, honeycombing

random distribution (170). The pathogenesis of these cysts (100,267). The distribution of fibrosis in the transverse
is uncertain, but they are presumably secondary to partial plane is variable, being patchy in distribution in some
small-airway obstruction by the bronchiolitis associated cases and predominantly subpleural in others (100).
with HP (170). Honeycombing, when present, is usually subpleural
The most common finding in patients with chronic (15,100). The fibrosis in HP usually involves mainly the
bird fancier’s lung is fibrosis. However, several studies have middle or lower lung zones (Fig. 8-79) (93,100,269). In
shown that the most common finding in patients with the majority of patients with chronic HP, evidence of acute
chronic farmer’s lung, including life-long nonsmokers, is or subacute disease is also seen on HRCT, consisting of
emphysema (141,269). The emphysema in these patients areas of ground-glass attenuation and centrilobular nod-
tends to involve mainly the lower lung zones and can be ules involving mainly the middle and lower lung zones
centrilobular, paraseptal, bullous, or cicatricial (269). (100,268).
The HRCT findings of fibrosis in chronic HP consist Several studies have shown that HRCT may demonstrate
of intralobular linear opacities, giving a reticular pattern, parenchymal abnormalities in patients with HP and
A B
Figure 8-76 Air trapping in hypersensitivity pneumonitis (HP). A: High-resolution CT at the level of the aortic arch shows bilateral
ground-glass opacities consistent with subacute HP and poorly defined localized areas of decreased attenuation and vascularity. Also noted
is mild subpleural reticulation (curved arrows) reflecting the presence of fibrosis secondary to prolonged repeated exposure to the inciting
antigen. B: Expiratory high-resolution CT image obtained at the same level as A shows bilateral areas of air trapping (straight arrows). The
patient was a 47-year-old woman with hypersensitivity pneumonitis (bird fancier’s lung).
A B
C D
Figure 8-77 Hypersensitivity pneumonitis (HP): CT spectrum. A: Target reconstructed high-resolution CT (HRCT) image through the right
upper lobe shows characteristic appearance of uniformly distributed centrilobular ground-glass nodules in the absence of reticulation or
cavitation. This finding is characteristic of subacute HP. B: HRCT demonstrates bilateral ground-glass opacities and poorly defined small
nodular opacities in a patient with subacute HP. Incidental note is made of right mastectomy. C: HRCT through the lower lung zones in a
42-year-old woman with subacute HP demonstrates extensive bilateral areas of ground-glass attenuation. D: HRCT through the upper lobes
in a different patient from that in A or B also shows pattern of diffuse ground-glass attenuation. In this case, close inspection reveals that
areas of ground-glass attenuation actually represent myriad small centrilobular nodules (compare with B). Note, in addition, focal areas of
apparent air trapping, a feature commonly identified in patients with diffuse HP.
normal chest radiographs (15,268,270). However, HP may diagnosis. In total, 199 of these patients were proven to
also be present in patients with normal HRCT. In one study have HP. Of the 199 patients with HP, 171 (86%) had an
of 11 subjects with relatively mild HP, only 1 (9%) had abnormal chest radiograph, and 183 (92%) had ground-
abnormal chest radiograph, and 5 (45%) had abnormal glass opacities and/or centrilobular nodules evident on
HRCT (14). It should be noted that in this study, HRCT HRCT (271).
scans were performed at 4-cm intervals (14). It cannot be The characteristic findings frequently allow confident
overemphasized that optimal assessment of infiltrative diagnosis of HP on HRCT (Fig. 8-80) (93,101). The main
lung disease requires HRCT scans at 1-cm intervals or volu- differential diagnosis of patients with a several-month
metric HRCT through the chest on a multidetector-row CT history of dry cough, progressive shortness of breath, and
scanner. In one multicenter study, the authors performed HRCT showing bilateral areas of ground-glass attenua-
HRCT at 10-mm intervals in 661 consecutive patients with a tion includes HP, NSIP, DIP, and alveolar proteinosis
condition for which HP was considered in the differential (81,143,266). Diagnosis of HP is based on careful clinical
history, physical examination, HRCT, presence of serum

precipitating antibodies to known antigens, and fiberoptic
bronchoscopy, usually with bronchoalveolar lavage (BAL)
and transbronchial biopsy (TBBx) (271,272). Final diag-
nosis can be made based on a combination of exposure
history, characteristic HRCT findings, presence of positive
serum precipitins to known antigens, and BAL showing
lymphocytosis (271). TBBx is routinely performed in some
centers (272). Surgical biopsy is required only in patients
in whom the HRCT, BAL, and TBBx fail to yield a confident
diagnosis (271,272).
The chronic form of HP can usually be readily distin-
guished from other causes of pulmonary fibrosis (93,101).
The main differential diagnosis is with IPF. The distinction
Figure 8-78 Hypersensitivity pneumonitis. High-resolution CT can usually be made by the presence of centrilobular nod-
at the level of the right upper lobe bronchus shows poorly defined ules and the relative sparing of the lung bases seen in HP
nodular opacities. The nodules are in clusters and a few millime-
ters away from the pleura (straight arrows), including interlobar as compared with the basal predominance seen in IPF.
fissures (curved arrows). This centrilobular distribution of nodules However, in approximately 10% of cases, the findings of
in hypersensitivity pneumonitis correlates with the presence of chronic HP may be identical to those of IPF (93,101). A
bronchiolitis and peribronchiolar alveolitis seen histologically. The
patient was a 35-year-old woman with hypersensitivity pneumoni- definitive diagnosis requires correlation with clinical his-
tis (bird fancier’s lung). tory, BAL, TBBx, and, in some cases, surgical lung biopsy.
A B
C
Figure 8-79 Chronic hypersensitivity pneumonitis (HP). A–C: High-resolution CT (1-mm collimation) sections through the upper, middle,
and lower lung zones, respectively, demonstrate irregular linear opacities mainly involving the middle lung zones. Also noted are areas of
ground-glass attenuation, presumably caused by superimposed subacute HP. The predominant distribution in the mid lung zones with rela-
tive sparing of the lung bases in fibrosis caused by HP allows distinction from idiopathic pulmonary fibrosis in the majority of cases.
A B
Figure 8-80 Hypersensitivity pneumonitis (HP): clinical role of CT. A: High-resolution CT section shows pattern of apparent diffuse
ground-glass attenuation. Small centrilobular nodules are apparent, however, throughout both lungs in association with focal areas of air
trapping. In this case, detailed clinical history revealed that the patient raised pigeons. B: Follow-up CT scan several weeks after initiation
of steroid therapy shows near-complete resolution of the abnormalities. In the appropriate clinical setting, CT findings may be sufficiently
specific of HP to warrant empiric therapy.
Nonspecific Interstitial Pneumonia

NSIP is a chronic interstitial lung disease characterized
histologically by varying degrees of alveolar wall inflamma-
tion and fibrosis (97,273). Lung biopsies may show
predominantly interstitial inflammation (cellular NSIP), or
fibrosis (fibrotic NSIP), or a combination of inflammation
and fibrosis. It is distinguished from UIP by a homoge-
neous appearance and the lack of fibroblastic foci (273).
NSIP is the second most common of the idiopathic intersti-
tial pneumonias, the most common one being UIP
(97,273). Several studies have shown that NSIP accounts
for approximately 30% to 35% of idiopathic interstitial
pneumonias and UIP for 55% to 60% (274–276).
Distinction from UIP is important because NSIP has a con-
siderably greater likelihood of response to treatment and a
better prognosis than UIP (97).
Figure 8-81 Nonspecific interstitial pneumonia (NSIP). High-
NSIP may be idiopathic or be a manifestation of colla- resolution CT (1-mm collimation) at the level of the basal segments
gen-vascular disease, HP, or drug reaction (97,190,277). It of the lower lobes shows extensive bilateral ground-glass opaci-
may precede the diagnosis of collagen-vascular disease by ties with minimal superimposed fine reticulation. The patient was a
54-year-old woman with idiopathic NSIP.
several months or several years (97).
The clinical manifestations usually consist of dry cough,
progressive dyspnea, and fatigue (97). NSIP tends to be
seen initially at a younger age than IPF (median age, 40 to
TABLE 8-14
50 years) (97). The clinical duration of symptoms before
diagnosis ranges from 6 months to 3 years (278,279). It is
equally common in men and women (278). Finger FINDINGS OF NONSPECIFIC INTERSTITIAL
clubbing, which occurs in 10% to 35% of patients, is less PNEUMONIA
common than it is in IPF (97). The other clinical manifes- Characteristic manifestations
tations are similar to those of IPF. Bilateral ground-glass opacities (main pattern)
The most common HRCT manifestation consists of Superimposed fine reticulation (common)
bilateral symmetric ground-glass opacities (Fig. 8-81) Peripheral and lower lung zone predominance
(Table 8-14) (98,99,280). The majority of patients have a Less common manifestations
fine reticular pattern superimposed on the ground-glass Airspace consolidation
Extensive fibrosis
opacities, traction bronchiectasis, and architectural distor-
Honeycombing
tion (Figs. 8-82 and 8-83) (98,99,280). The prevalence of
A B
Figure 8-82 Nonspecific interstitial pneumonia (NSIP). A: High-resolution CT performed on a multidetector scanner shows extensive
bilateral ground-glass opacities. Also noted are mild reticulation (curved arrows) and traction bronchiectasis (straight arrows), mainly in the
subpleural regions of the lower lobes. B: Coronal reformation demonstrates bilateral ground-glass opacities, mild reticulation, and traction
bronchiectasis (arrows). Note that the abnormalities involve almost exclusively the lower lobes. The patient was a 77-year-old man with idio-
pathic NSIP.
reticular opacities reported in various studies ranges from lung zones, and in 50% to 70% of patients, they
50% to 100% of cases (98,99,280). Honeycombing is involve predominantly the peripheral lung regions (98,
seen in 10% to 30% of patients and tends to be mild, 99,280).
involving less than 10% of the parenchyma (98,99,280). The HRCT manifestations of NSIP can mimic those of
Areas of consolidation and centrilobular nodules have UIP, HP, and COP (280). The presence of predominantly
been reported in a small percentage of patients in some ground-glass opacities allows distinction of NSIP from UIP
studies (98,280) but were not seen in one other large in the majority of cases (99). However, it may be impossi-
study (99). The abnormalities in NSIP may be diffuse, but ble to distinguish fibrotic NSIP with a predominantly
in 60% to 90% of cases, they involve mainly the lower reticular pattern from UIP (Fig. 8-84) (99,280). HP can
Figure 8-84 Nonspecific interstitial pneumonia (NSIP). High-

resolution CT (1-mm collimation) shows a reticular pattern in the
subpleural regions of the right lung (curved arrows) and more
extensively in the left lung. Note associated traction bronchiecta-
Figure 8-83 Nonspecific interstitial pneumonia (NSIP). High- sis (straight arrows). The findings are compatible with fibrotic NSIP
resolution CT (1-mm collimation) shows bilateral ground-glass opac- and with usual interstitial pneumonia (UIP). Surgical biopsy demon-
ities. A reticular pattern involves mainly the subpleural regions strated fibrotic NSIP. The patient was a 54-year-old man with
(arrows). The patient was a 79-year-old woman with idiopathic NSIP. idiopathic NSIP.
usually be distinguished from NSIP by the presence of

centrilobular nodules and lobular areas of air trapping.
COP can usually be distinguished by the predominance of
consolidation, which typically involves mainly the peri-
bronchial and subpleural lung regions.
Although the presence of extensive bilateral ground-
glass opacities with superimposed fine reticulation in a
patient with clinical findings consistent with chronic inter-
stitial lung disease should suggest the possibility of NSIP,
a confident diagnosis is seldom possible on HRCT
(143,191,280). Even a histologic diagnosis of NSIP does
not establish a final diagnosis. NSIP is a common reaction
pattern to various drugs, is commonly associated with
collagen-vascular diseases, particularly scleroderma, and
can be a histologic manifestation of HP (97,190). These
conditions must be excluded by careful clinical assessment Figure 8-85 Respiratory bronchiolitis–interstitial lung disease
before making a diagnosis of idiopathic NSIP. (RB-ILD). High-resolution CT shows bilateral ground-glass opaci-
ties and poorly defined centrilobular nodules (straight arrows).
Also noted is mild emphysema (curved arrow). The patient was a
45-year-old two-pack-a-day cigarette smoker with RB-ILD.
Respiratory Bronchiolitis–Interstitial
Lung Disease
and centrilobular emphysema. A small percentage of
Respiratory bronchiolitis (RB), also known as smokers’ patients have a reticular pattern due to fibrosis (107,108).
bronchiolitis, is a common incidental histologic finding in The fibrosis in RB-ILD is mild and tends to involve mainly
cigarette smokers (186,281). By definition, these patients the subpleural regions and lower lung zones (Fig. 8-86)
have no symptoms related to the bronchiolitis (97). A (107,108).
small percentage of smokers with RB present with with
clinical findings mimicking those of interstitial lung
disease and are referred to as having respiratory bronchi-
Desquamative Interstitial Pneumonia
olitis–interstitial lung disease (RB-ILD) (97,282). DIP is an uncommon condition, characterized histologi-
RB is characterized histologically by the presence of cally by the presence of numerous macrophages filling
mild chronic inflammation and fibrosis in the walls of the the alveolar airspaces, mild inflammation of the alveolar
respiratory bronchioles and by the accumulation of pig- walls, and minimal fibrosis (97,190). Approximately 90%
mented macrophages within respiratory bronchioles and
adjacent alveolar ducts and alveoli (186,281). Patients
with RB-ILD have more extensive peribronchiolar and
alveolar wall inflammation than do asymptomatic patients
(97,282). Patients who have RB-ILD are typically young,
usually in their 30s and 40s, current smokers with average
exposures of more than 30 pack-years of cigarette smoking
(97). The patients complain of chronic cough and progres-
sive shortness of breath. The prognosis is good. Smoking
cessation usually leads to amelioration of symptoms with-
out need for additional treatment.
In the majority of patients who have RB, the histologic
abnormalities are too mild to be detected on HRCT
(107,283). When present, HRCT findings consist of poorly
defined centrilobular nodules and multifocal ground-
glass opacities (107,283,284). These findings can be dif-
fuse but tend to involve mainly the upper zones. The most
common HRCT findings of RB-ILD consist of ground-
Figure 8-86 Respiratory bronchiolitis–interstitial lung disease
glass opacities and poorly defined centrilobular nodules (RB-ILD). High-resolution CT shows bilateral ground-glass opaci-
(Fig. 8-85) (107,108,285). The abnormalities can be ties most extensive in the lingula and left lower lobe. Also noted
diffuse or involve mainly the upper or lower lung zones are a few irregular lines in the subpleural regions consistent with
mild fibrosis (straight arrows) and mild emphysema (curved
(107,108). Other common findings are thickening of arrows). The patient was a 60-year-old male cigarette smoker with
bronchial walls, presumably related to chronic bronchitis, RB-ILD.
of patients who have DIP are current cigarette smokers

(107,286). Less commonly, DIP can be seen in association
with a variety of conditions in nonsmokers, including drug
reactions, asbestosis, and human immunodeficiency virus
(HIV) infection (190). In some patients who have DIP, the
macrophage accumulation may have a peribronchiolar
predominance similar to that seen in RB-ILD, the only dis-
tinction being the presence of more diffuse involvement of
the airspaces in DIP (287,288). A continuum of the extent
of airspace macrophage accumulation exists between RB-
ILD and DIP, sometimes making it difficult to distinguish
between these two entities. Because of the frequent associ-
ation with cigarette smoking and the histologic overlap,
DIP is considered by many to represent the end of a
spectrum of RB-ILD, with RB, RB-ILD, and DIP represent- Figure 8-88 Desquamative interstitial pneumonia (DIP). High-
ing different degrees of small-airway and parenchymal resolution CT in a 40-year-old man with DIP shows bilateral areas
reaction to cigarette smoke (107,287,288). of ground-glass attenuation in a patchy distribution.
DIP occurs most commonly in patients between 30 and
50 years of age (143,286). The clinical symptoms usually
consist of slowly progressive exertional dyspnea and dry subpleural lung regions and lung bases (Fig. 8-89).
cough (97). The most common finding on chest radio- Traction bronchiectasis and honeycombing are uncom-
graphs in patients who have DIP is the presence of mon (143,290). Definitive diagnosis of DIP, like that of
ground-glass opacities in the lower lung zones (286). RB-ILD, requires surgical biopsy. The prognosis is good,
However, in 3% to 22% of patients who have biopsy- the vast majority of patients improving with smoking
proven DIP, the chest radiograph has been reported to be cessation and corticosteroids (97).
normal (286,289).
The predominant abnormality on HRCT is the presence
Collagen–Vascular Diseases
of bilateral ground-glass opacities (143,191). The ground-
glass attenuation is predominantly peripheral and basal in Interstitial lung disease occurs in the majority of patients
distribution in 60% of cases, patchy in 20%, and diffuse in with progressive systemic sclerosis (PSS) and in a small
20% of patients (Figs. 8-87 and 8-88) (143). A subpleural percentage of other collagen-vascular diseases. The HRCT
and basal predominance is present in 60% to 70% of and histologic findings in patients with PSS are usually
patients. Intralobular linear opacities resulting in a fine those of NSIP (291–293). In one review of the surgical
reticular pattern are seen in 60% to 80% of patients lung biopsy specimens of 80 patients with PSS and
(143,191). These are usually mild and involve mainly the pulmonary disease, 62 (78%) had NSIP, 6 (7%) had UIP,
6 (7%) had end-stage lung disease, and 6 (7%) had other
patterns (291). Similarly, the HRCT findings resemble
those of NSIP rather than UIP (292). In one study compar-
ing the HRCT features of 225 patients with PSS and
pulmonary disease with those of 40 patients with biopsy-
proven IPF and 27 with biopsy-proven idiopathic NSIP, no
difference was found in the pattern or extent of lung
disease in PSS with that of NSIP (292). The extent of
ground-glass attenuation was approximately 50% of the
lung parenchyma in PSS and NSIP and 23% in UIP, and
the pattern of reticulation was similarly fine in PSS and
NSIP and coarser in UIP (292).
The most common pulmonary manifestation of PSS on
HRCT consists of bilateral symmetric ground-glass opaci-
ties (Fig. 8-90) (142,292,294). The majority of patients
have a fine reticular pattern superimposed on the ground-
glass opacities, traction bronchiectasis, and architectural
distortion (Fig. 8-91). Honeycombing has been reported in
Figure 8-87 Desquamative interstitial pneumonia (DIP). High- 20% to 60% of patients with PSS and pulmonary
resolution CT at the level of the bronchus intermedius shows bilat-
eral ground-glass opacities involving mainly the peripheral lung parenchymal abnormalities but tends to be mild, usually
regions. The patient was a 57-year-old woman with DIP. involving less than 10% of the lungs (142,292,294). The
Figure 8-89 Desquamative interstitial pneumonia (DIP). High-

resolution CT in a 53-year-old man with DIP shows diffuse areas
of ground-glass attenuation. Also noted is evidence of mild fibro-
sis in the subpleural lung regions of the upper lobes.
abnormalities usually have a distinct basal, posterior, and from 12 to 40 mm (mean, 23 mm), a finding that was not
peripheral predominance (142,294). seen in a control group of 13 patients who had a variety of
The interstitial abnormalities in patients with PSS have a other parenchymal and airway abnormalities (296).
considerably better prognosis than do those of IPF (295). The most common pulmonary complication of
Patients with PSS and predominantly ground-glass attenua- polymyositis/dermatomyositis is aspiration pneumonia,
tion on HRCT are more likely to respond to treatment than which is seen in 15% to 20% of patients (190). Clinically
are patients with predominantly reticulation (295). significant interstitial disease occurs in 5% to 20% of
Other findings on CT in patients who have PSS include patients (190). Similar to PSS, the most common interstitial
diffuse pleural thickening, seen in one third of cases (142); lung disease in polymyositis/dermatomyositis is NSIP. In a
asymptomatic esophageal dilatation, present in 40% to review of the surgical lung biopsy specimens of 22 patients
80% of cases (Figs. 8-90 and 8-91) (119,296); and enlarged with pulmonary disease secondary to polymyositis/
mediastinal nodes, seen in approximately 60% of cases dermatomyositis, 18 (82%) had NSIP, 2 (11%) had UIP,
(296). The presence of esophageal dilatation may be helpful 1 had organizing pneumonia, 1 had diffuse alveolar dam-
in the differential diagnosis of PSS from other diffuse ILDs. age, and 1 had unclassifiable interstitial pneumonia (297).
The coronal luminal diameter of the esophagus in patients The HRCT findings are similar to those of idiopathic NSIP
who have PSS, as shown on CT, has been reported to range and PSS and consist mainly of ground-glass opacities
involving predominantly the lower lung zones (297). Fine
superimposed reticulation and traction bronchiectasis are
Figure 8-91 Interstitial lung disease in progressive systemic

Figure 8-90 Interstitial lung disease in progressive systemic sclerosis. High-resolution CT (1-mm collimation) at the level of the
sclerosis. High-resolution CT (1-mm collimation) shows bilateral dome of the right hemidiaphragm shows extensive bilateral
ground-glass opacities. Also noted is a dilated esophagus contain- ground-glass opacities with superimposed reticulation and honey-
ing a fluid level (arrow). The patient was a 73-year-old woman with combing. Also noted is a dilated esophagus (arrow) containing a
progressive systemic sclerosis and nonspecific interstitial pneumo- fluid level. The patient was a 40-year-old woman with progressive
nia (NSIP). systemic sclerosis and interstitial fibrosis.
common. Other findings, including consolidation and hon- Drug-Induced Lung Disease
eycombing, are uncommon.
In approximately 5% of patients with rheumatoid Drug reactions are common, and their frequency is
arthritis, clinically significant interstitial lung disease increasing (190). It has been estimated that as many as
develops (295,298). Contrary to the other collagen-vascu- 5% of all hospitalizations are the result of untoward
lar diseases, the most common interstitial lung disease effects of drug therapy, and about 0.3% of deaths that
seen in rheumatoid arthritis is UIP (Fig. 8-92). In one occur in the hospital are believed to be drug related (302).
study of 63 patients with rheumatoid arthritis–related The diagnosis of drug-induced lung disease is primarily
interstitial lung disease, approximately 41% had an HRCT clinical, being based on a temporal relation between
pattern of UIP; 30%, NSIP; 18%, bronchiolitis; 8%, organ- chemotherapy and a characteristic reaction after an appro-
izing pneumonia (BOOP); 1 patient had a pattern of priate latent period and exclusion of other causes of lung
diffuse alveolar damage, and 1 had LIP (299). Biopsy in disease. The most common histopathologic patterns of
17 patients showed good correlation between the HRCT drug-induced lung injury are noncardiogenic pulmonary
and the histologic patterns. The predominant UIP pattern edema, NSIP, organizing pneumonia, eosinophilic pneu-
on HRCT and histologically in patients with rheumatoid monia, and pulmonary hemorrhage (303).
arthritis was also shown in a study of 29 patients, 19 of Noncardiogenic pulmonary edema (ARDS) has been
whom had histologic confirmation from lung biopsy described most commonly after administration of
(164,300,301). Although the HRCT pattern of UIP in chemotherapeutic agents, particularly bleomycin, busul-
patients with rheumatoid arthritis is identical to those of fan, carmustine, and mitomycin (303). The clinical and
patients with IPF, patients with collagen vascular dis- radiologic findings are similar to those of ARDS from
ease–associated UIP have fewer fibroblastic foci and better other causes. HRCT usually demonstrates extensive bilat-
prognosis than do those with IPF (300). eral ground-glass opacities and dependent areas of
Other common findings seen in patients with rheuma- airspace consolidation (304–306). Superimposed reticu-
toid arthritis are bronchiectasis unrelated to fibrosis and lation may also be seen (304).
bronchiolitis obliterans. Bronchiectasis or bronchiolectasis NSIP is seen in association with a variety of drugs,
has been reported in approximately 30% of patients, the most common being methotrexate, amiodarone,
centrilobular nodules in 20%, and mosaic perfusion in and carmustine (162a,305). The main HRCT manifestations
approximately 20% of patients (164,299–301). In one consist of patchy bilateral or diffuse ground-glass opacities
study, 10 (43%) of 23 patients who underwent expiratory (277,305,306). Superimposed intralobular linear opacities,
HRCT had air trapping (299). Extrapulmonary findings architectural distortion, and traction bronchiectasis are seen
include enlargement of the main pulmonary artery, in patients with associated fibrosis (Fig. 8-93) (162a,304).
reported in approximately 30% of cases, pleural thickening In one review of the HRCT findings in 60 patients with
or effusion in 20%, and lymph node enlargement in 10% drug-induced pneumonitis, intralobular linear opacities
to 20% of cases (164,299–301). were seen in the vast majority of patients with drug
Figure 8-92 Interstitial lung disease in rheumatoid arthritis.

High-resolution CT (1-mm collimation) demonstrates irregular Figure 8-93 Nonspecific interstitial pneumonia secondary to
linear opacities, mild honeycombing, and areas of ground-glass nitrofurantoin. High-resolution CT (1-mm collimation) at the level
attenuation. The abnormalities are patchy in distribution but of the main bronchi shows bilateral ground-glass opacities, mild
involve predominantly the subpleural lung regions. The patient subpleural reticulation, and traction bronchiectasis (arrow). The
was a 63-year-old man with interstitial lung disease associated with patient was a 73-year-old man with nonspecific interstitial pneu-
rheumatoid arthritis. monia secondary to nitrofurantoin.
reactions secondary to antineoplastic agents but were

uncommon in other drug reactions (307).
Organizing pneumonia (also known as BOOP-like
reaction) has been reported most frequently in association
with methotrexate, cyclophosphamide, gold, nitrofuran-
toin, amiodarone, bleomycin, and busulfan (277,308,
308a). HRCT usually shows bilateral symmetric or asym-
metric areas of consolidation, most commonly in a pre-
dominantly peribronchial and/or subpleural distribution
(Fig. 8-94) (304,305,309). Less common manifestations
include poorly defined nodular areas of consolidation,
centrilobular nodules and branching opacities (“tree-in-
bud” pattern), and peripheral reticulation (304,306).
Figure 8-95 Eosinophilic pneumonia secondary to phenytoin.
Eosinophilic pneumonia drug reaction is seen most High-resolution CT (1-mm collimation) shows bilateral areas of
commonly in association with methotrexate, sulfasalazine, consolidation in the peripheral regions of the apical segments of
para-aminosalicylic acid, nitrofurantoin, minocycline, and the upper lobes. The patient was a 47-year-old man with
eosinophilic pneumonia secondary to phenytoin.
nonsteroidal anti-inflammatory drugs (305,310). Chest
radiography and HRCT show bilateral airspace consolida-
tion and ground-glass opacities, typically involving mainly concurrent pathologic patterns (190) and that HRCT is of
the peripheral lung regions and upper lobes (Fig. 8-95) limited value in distinguishing between the various specific
(277,305,310). In a review of the HRCT findings in types of histopathologic patterns of drug-induced lung
14 patients with biopsy-proven drug-induced eosinophilic disease (277,303,304). Only occasionally, HRCT may
pneumonia, the most common parenchymal abnormali- demonstrate findings that are highly specific for the diag-
ties were airspace consolidation seen in all 14 patients, nosis. These include increased attenuation in amiodarone
ground-glass opacities in 11, small nodules in 8, and septal lung and areas of fat attenuation in lipoid pneumonia. Ami-
thickening in 6 patients (310). The abnormalities involved odarone is a tri-iodinated antiarrhythmic drug. Amiodarone
mainly the peripheral lung regions in 10 patients (72%) pulmonary toxicity can result in high-attenuation parenchy-
and the upper lung zones in 7 patients (50%) (310). mal opacities because of its incorporation into the type II
HRCT is helpful in assessing the presence and extent of pneumocytes (Fig. 8-37) (162a,309,313). High attenuation
pulmonary disease and may demonstrate parenchymal is typically also present in the liver and, less commonly, in
abnormalities in patients with proven drug reaction and the mediastinal lymph nodes (Fig. 8-37).
normal radiographs (311,312). It is important to note,
however, that drug reactions can be associated with several
Alveolar Proteinosis
PAP is a rare lung disease characterized by filling of the air-
spaces with a lipoproteinaceous material (314). The majority
of cases are idiopathic. Less commonly, alveolar proteinosis
may result from heavy exposure to dusts (silicoproteinosis),
infection (e.g., Pneumocystis jiroveci), malignancy (leukemia
or lymphoma), or immune deficiency (190). Idiopathic PAP
is seen most commonly in men 30 to 50 years of age. The
patients may be asymptomatic or have slowly progressive
dyspnea, fatigue, and nonproductive cough (190). The
radiographic findings usually consist of bilateral patchy or
diffuse consolidation (315,316).
The characteristic HRCT findings of alveolar proteinosis
consist of bilateral ground-glass opacities associated
with smoothly thickened interlobular septa (Fig. 8-96)
(81,317). The ground-glass attenuation often has a geo-
graphic distribution with sharp demarcation between
Figure 8-94 Organizing pneumonia (BOOP-like reaction) second- normal and abnormal parenchyma (79,81). Thickening of
ary to sulfasalazine. High-resolution CT (1-mm collimation) at the
level of the aortopulmonary window shows patchy bilateral areas of the interlobular septa is seen only in areas with ground-glass
consolidation in a predominantly peribronchial distribution (straight attenuation and is presumably caused by mild edema or
arrows). Also noted is consolidation in a perilobular (curved arrow) fibrosis (79,81). The combination of geographic distribu-
and subpleural distribution in the dorsal region of the right upper
lobe. The patient was a 55-year-old woman with organizing pneu- tion and interlobular septal thickening in areas of ground-
monia (BOOP-like reaction) secondary to sulfasalazine. glass attenuation results in a “crazy-paving” appearance
A B
Figure 8-96 Alveolar proteinosis. A: High-resolution CT (HRCT) at the level of the bronchus intermedius demonstrates extensive areas of
ground-glass attenuation and consolidation. Note sharp demarcation between normal and abnormal parenchyma, particularly in the left upper
lobe. Also noted is fine reticulation. B: HRCT targeted to the right lower lobe and performed 1 year later demonstrates interval improvement.
Again noted are areas of ground-glass attenuation and fine reticulation. The reticulation is caused by smoothly thickened interlobular septa.
(Fig. 8-33) (81,317). Other common HRCT findings include ducts and alveoli with or without bronchiolar intraluminal
ill-defined nodular opacities and areas of airspace consoli- polyps (97,318,319). The condition is also commonly
dation (79). In a small percentage of patients, alveolar known as BOOP (318,320). The clinical, functional, and
proteinosis results in fibrosis with the development of radiologic findings are primarily the result of an organizing
irregular linear opacities and architectural distortion on pneumonia. Furthermore, the term cryptogenic organizing
HRCT (317). The abnormalities usually involve all lung pneumonia avoids confusion with airway diseases such as
zones to a similar extent (317). In one review of 139 CT constrictive bronchiolitis obliterans, which can be prob-
scans from 27 patients, ground-glass opacities were seen on lematic with the term BOOP. Therefore the American
at least one scan in 100% of the patients, interlobular septal Thoracic Society/European Respiratory Society Inter-
thickening in 85%, airspace consolidation in 78%, and national Multidisciplinary Consensus Classification of the
evidence of fibrosis in 7% (317). Idiopathic Interstitial Pneumonias recommended that the
The crazy-paving pattern on HRCT in the appropriate term BOOP be replaced by organizing pneumonia and idio-
clinical setting is suggestive of PAP. This pattern, however, pathic BOOP by cryptogenic organizing pneumonia (COP)
can also be seen in several other airspace and interstitial (97). Although COP is by definition idiopathic, organizing
diseases, including NSIP, ARDS, pulmonary hemorrhage, pneumonia (BOOP-like reaction) may also be seen in a
exogenous lipoid pneumonia, and pulmonary edema (Fig. variety of conditions, including pulmonary infection, drug
8-34) (63,80,82,151). These diseases, however, usually reactions, collagen-vascular diseases, and after toxic fume
have characteristic clinical features that allow ready dis- inhalation (97,192a,318). These conditions must be ex-
tinction from PAP (317). CT may also be helpful in the cluded before making a diagnosis of COP.
detection of focal pneumonia (79). In patients with alveo- Patients with COP typically have a relatively short
lar proteinosis, it may be difficult to distinguish infection history (median, 3 months) of cough and dyspnea (97).
from consolidation caused by the underlying disease on They often have intermittent fever and malaise. Although
the chest radiograph. CT may confirm a clinically sus- the clinical and functional findings may resemble those of
pected infection by demonstrating focal areas of dense IPF, the duration of symptoms in patients who have COP
consolidation or abscess formation (79). is shorter, systemic symptoms are more common, and
finger clubbing is rarely seen (97). The patients usually
respond well to corticosteroid therapy and have a good
prognosis (97).
by Consolidation The main radiologic manifestations of COP consist of
patchy unilateral or bilateral areas of airspace consolidation
Cryptogenic Organizing Pneumonia/ (97,318,321). Irregular reticular opacities may be present,
Bronchiolitis Obliterans Organizing Pneumonia but they are rarely a major feature. Small nodular opacities
COP is a condition characterized histologically by the pres- may be seen as the only finding or, more commonly, in
ence of patchy organizing pneumonia involving alveolar association with areas of airspace consolidation.
TABLE 8-15
FINDINGS IN CRYPTOGENIC ORGANIZING
PNEUMONIA (IDIOPATHIC BOOP)
Characteristic findings
Patchy bilateral airspace consolidation
Subpleural and/or peribronchovascular distribution (60%–80%)
Perilobular distribution (60%)
Air bronchograms (95%–100%)
Other common findings
Ground-glass opacities
Small nodular opacities, often centrilobular
Unilateral or bilateral pleural effusions
Less common findings
Consolidation surrounding ground-glass opacities
(“reversed halo” sign)
Large irregular nodules or masses with or without air
bronchograms
Figure 8-97 Cryptogenic organizing pneumonia (idiopathic
bronchiolitis obliterans with organizing pneumonia, BOOP). High- Irregular linear opacities
resolution CT at the level of the right upper lobe bronchus shows
bilateral areas of consolidation in a predominantly peripheral dis-
tribution. The patient was an 81-year-old woman with cryptogenic
organizing pneumonia. involve mainly the peripheral regions of the secondary
lobule, resulting in a perilobular pattern (Fig. 8-99) (158).
In approximately 20% of patients these polygonal or ring-
The characteristic HRCT findings consist of patchy bilat- like areas of consolidation surround central ground-glass
eral areas of airspace consolidation in a predominantly opacity (reversed halo sign) (Fig. 8-100) (159). The con-
subpleural and/or peribronchial distribution (Figs. 8-97 solidation may show no zonal predominance or involve
and 8-98) (Table 8-15) (97,110,191). Consolidation is mainly the lower lobes and typically contains air bron-
seen in approximately 80% of patients with COP and a chograms.
predominantly subpleural and/or peribronchial distribu- Ground-glass opacities are present in 80% to 90% of
tion in 60% to 80% of patients (144,158,191). In about patients and small, ill-defined nodules, often in a centrilob-
60% of patients, some of the areas of consolidation ular distribution, in 30% to 50% of cases (110,144,158).
Figure 8-98 Cryptogenic organizing pneumonia (idiopathic

bronchiolitis obliterans with organizing pneumonia, BOOP). High- Figure 8-99 Cryptogenic organizing pneumonia (idiopathic
resolution CT at the level of the lower lung zones shows bilateral bronchiolitis obliterans with organizing pneumonia, BOOP). High-
areas of consolidation in a predominantly peribronchial distribu- resolution CT at the level of the aortic arch shows bilateral areas of
tion (arrows) and focal ground-glass opacities. Also noted are a consolidation in a predominantly perilobular (arrows) and sub-
few irregular linear opacities and mild architectural distortion. pleural distribution. Also noted are focal ground-glass opacities.
The patient was a 49-year-old man with cryptogenic organizing The patient was a 46-year-old man with cryptogenic organizing
pneumonia. pneumonia.
of pulmonary edema,” involving mainly the upper lobes

(323). This peripheral distribution, however, is evident on
the chest radiograph in only 50% of cases (324). In the
remaining cases, the radiographic findings are nonspecific
and consist of unilateral or patchy bilateral consolidation.
On HRCT, a peripheral distribution of consolidation is
seen in virtually all patients, even when it is not apparent
on the radiograph (Fig. 8-101) (160,325). The consolida-
tion usually is most severe in the upper lobes (160). Less
common manifestations include ground-glass opacities,
bandlike opacities, and nodular opacities (160,325).
The combination of peripheral consolidation and
eosinophilia is highly suggestive of chronic eosinophilic
pneumonia (156). Less common causes of peripheral con-
solidation and eosinophilia include simple pulmonary
Figure 8-100 Cryptogenic organizing pneumonia (idiopathic eosinophilia, Churg-Strauss syndrome, and drug-induced
bronchiolitis obliterans with organizing pneumonia, BOOP) with
nodular consolidation and reversed halo sign. High-resolution CT
eosinophilic lung disease (Fig. 8-95) (156,310,326). As
at the level of the aortic arch shows nodular areas of consolidation mentioned previously, an identical peripheral distribution
(straight arrows) in the subpleural lung regions. Ground-glass of consolidation may also be seen in patients with COP
opacities can be seen surrounding some of the nodular areas of
consolidation (halo sign). A focal area of consolidation in the right
(110,158). However, the consolidation in COP often
upper lobe (curved arrow) surrounds central ground-glass opacity involves mainly the lower lung zones, whereas the con-
(reversed halo sign). The patient was a 52-year-old woman with solidation in chronic eosinophilic pneumonia usually
cryptogenic organizing pneumonia.
has a predominantly upper lobe distribution. Less
commonly, peripheral distribution similar to chronic
Less commonly, patients may have multiple large nodules eosinophilic pneumonia may be seen in DIP (143) and
or masses (Fig. 8-100) (322). The nodules usually have sarcoidosis (327).
irregular margins and, in approximately 50% of cases, con-
tain air bronchograms (322).
Lipoid Pneumonia
Irregular linear opacities, usually mild, are seen in 5%
to 15% of patients (110,144) and small pleural effusions Exogenous lipoid pneumonia results from aspiration of
are present in 30% to 35% of cases (110,144). The effu- mineral, vegetable, or animal oil. The initial reaction to
sions are usually small and may be unilateral or bilateral animal oils and some vegetable oils is a hemorrhagic
(110,144). bronchopneumonia (328). Aspiration of mineral oil is
In the appropriate clinical context, in a patient with associated with minimal acute reaction. The chronic phase
consolidation that is increasing over several weeks despite is characterized by an infiltrate consisting of lipid-laden
antibiotics, an HRCT pattern of consolidation in a pre-
dominantly peribronchial or subpleural distribution is
highly suggestive of COP (97). The diagnosis of COP
requires exclusion of possible causes, such as infection,
drug-induced lung disease, and inhalation injury.
Chronic Eosinophilic Pneumonia

Chronic eosinophilic pneumonia is an idiopathic condi-
tion characterized histologically by filling of the airspaces
with eosinophils and an amorphous proteinaceous exu-
date (190). The patients usually have a several-month his-
tory of cough, low-grade fever, weight loss, and dyspnea
(190). Chronic eosinophilic pneumonia is seen most
commonly in middle-aged women, and the majority of
patients have peripheral eosinophilia. Approximately
50% of patients are asthmatic. Chronic eosinophilic pneu-
monia resolves rapidly with corticosteroid therapy. Figure 8-101 Chronic eosinophilic pneumonia. High-resolution
The characteristic radiographic findings of chronic CT at the level of the bronchus intermedius shows bilateral areas
of consolidation and ground-glass opacities in the subpleural lung
eosinophilic pneumonia consist of peripheral, nonseg- regions. The patient was a 49-year-old woman with chronic
mental areas of consolidation, the “photographic negative eosinophilic pneumonia.
A B
Figure 8-102 Lipoid pneumonia. A: High-resolution CT demonstrates focal area of consolidation in the left upper lobe. B: Identical
image as A, imaged with mediastinal windows, demonstrates fat within the areas of consolidation (arrow). Also noted is a fluid-filled esoph-
agus. The patient was an 81-year-old woman in whom lipoid pneumonia was caused by aspiration of mineral oil taken as a laxative.
macrophages and multinucleated giant cells surrounded along the bronchioles. The initial abnormalities consist
by varying amounts of histiocytic infiltration or fibrosis of nodules measuring 1 to 10 mm in diameter (232). The
(328). Radiographic findings include a solitary nodule nodules may cavitate (190). Over time, the nodules
simulating bronchogenic carcinoma, multiple masses, undergo fibrosis and result in a characteristic stellate
localized areas of consolidation, and extensive bilateral border extending into the surrounding interstitium
consolidation (329,330). (190,232). As the granulomatous reaction progresses, it
The most common HRCT findings consist of bilateral results in destruction of the bronchiolar wall and pro-
areas of consolidation and single or multiple masses gressive dilatation of the bronchiolar lumen with the
(161,329). Patients with acute lipoid pneumonia usually formation of cysts (332). PLCH typically involves mainly
present with bilateral consolidation, sometimes associated the upper and middle lung zones (232).
with pleural effusion (329). The majority of patients with PLCH is usually limited to the lungs, but it can also
chronic lipoid pneumonia have irregular single or multi- involve other organs, especially bone and lymph nodes.
ple masses (161,329). In approximately 80% of patients, PLCH primarily afflicts young adults between the ages of
focal areas of fat attenuation can be seen on CT in patients 20 and 40 years and occurs almost exclusively in current
with single or multiple masses (Fig. 8-102) (161,329) and or previous cigarette smokers (190,333). The clinical
in patients with bilateral areas of consolidation due to presentation is variable, with up to 25% of patients being
lipoid pneumonia (329). Less common manifestations of asymptomatic. The most common clinical manifestations
chronic lipoid pneumonia include a reticular pattern are nonproductive cough and dyspnea (190,331).
(161,329) and ground-glass opacities in association with Pneumothorax develops in about 25% of patients and
interlobular septal thickening and intralobular lines may be the first manifestation of PLCH. The prognosis is
(crazy-paving pattern) (146). good, with the disease frequently stabilizing or resolving
after cessation of smoking or spontaneously even in
patients who continue to smoke (190,334). However,
Diseases Characterized Primarily by Cysts
with continued cigarette smoking, the disease often
progresses and may result in end-stage fibrosis and respi-
Pulmonary Langerhans Cell Histiocytosis
ratory failure (190).
Pulmonary Langerhans cell histiocytosis (PLCH), previ- Radiographically, PLCH most commonly causes a retic-
ously also known as histiocytosis X or eosinophilic ulonodular pattern primarily involving the mid and upper
granuloma, is a chronic, progressive disorder character- lung zones (335,336). Cysts and honeycombing have been
ized by peribronchiolar proliferation of Langerhans reported on the radiograph in between 1% and 15% of
cells (190,331). Langerhans cells are distinctive special- cases. Lung volumes are normal or increased.
ized cells of monocyte–macrophage lineage that contain HRCT is superior to the radiograph in demonstrating
rod- or racket-shaped organelles (Birbeck granules) the presence and distribution of parenchymal findings in
(190). PLCH is a granulomatous process that extends PLCH and closely reflects the macroscopic pathologic
A B
Figure 8-103 Pulmonary Langerhans cell histiocytosis. A: High-resolution CT (HRCT) at the level of the aortic arch shows numerous thin-
walled cysts (straight arrows) and small nodules (curved arrows) in the upper lobes. Although some of the cysts are closely related to
pulmonary arteries and may be difficult to distinguish from bronchiectasis, several of the cysts are in the subpleural lung regions and there-
fore are easily distinguished from dilated bronchi. B: HRCT through the lung bases shows very few cysts. The relative sparing of the lung
bases is characteristic of pulmonary Langerhans cell histiocytosis. The patient was a 30-year-old man.
findings, therefore frequently allowing a confident, spe- have smooth or, more commonly, irregular margins, usually
cific diagnosis (5,51,331). measure less than 10 mm in diameter, and have a centrilob-
The characteristic HRCT manifestations of PLCH consist ular distribution (167,331). Occasionally the nodules may
of cysts and nodules predominantly in the upper and mid- be cavitated or greater than 10 mm in diameter (Fig. 8-105).
dle lung regions, with relative sparing of the lung bases HRCT shows no consistent central or peripheral predomi-
(Figs. 8-103 and 8-104) (Table 8-16) (167,169,331). The nance of lesions (5,167), but in nearly all cases, the lung
cysts usually measure less than 1 cm in diameter and are bases and the costophrenic sulci are relatively spared
round or ovoid (5,167,331). When confluent, the cysts can (51,169). The distribution of the abnormalities throughout
become larger than 2 cm in diameter and often exhibit the lungs can be better appreciated on coronal reformations
bizarre configurations. The cyst walls are usually thin and from volumetric helical HRCT performed through the lungs
smooth but can also be several millimeters in thickness by using a multidetector CT scanner than on cross-sectional
and have a nodular appearance (167,331). The nodules can HRCT images (Fig. 8-106) (18,19).
A B
Figure 8-104 Pulmonary Langerhans cell histiocytosis (PLCH): disease spectrum. A: One-millimeter section through the upper lobes
shows a pattern of ill-defined ground-glass centrilobular nodules consistent with respiratory bronchiolitis and a few scattered thick-walled
cysts (arrows) consistent with early PLCH, subsequently documented by transbronchial biopsy. B: One-millimeter section through a differ-
ent patient than shown in A demonstrates a more typical appearance of scattered thick-walled irregular cysts in the absence of significant
reticulation or nodularity. C: High-resolution CT section through the right lung in another patient with documented PLCH. Innumerable
well-defined, uniformly thick-walled cysts are scattered throughout the lung, some in bizarre, branching patterns mimicking the appear-
ance of bronchiectasis (arrow). Presumably this appearance is caused by cyst fusion. Despite extensive disease, no obvious reticulation or
fibrosis is apparent. D: Pathologic section from another patient with PLCH showing multiple cysts throughout the lungs, frequently ex-
hibiting unusual branching configurations. Note relative sparing of the lung bases. (D courtesy of Dr. Carlos R.R. de Carvalho, University of
São Paulo, São Paulo, Brazil.) (continued )
C D
The prevalence of nodules and cysts in PLCH is influ- 20% of patients (5,338). The ground-glass opacities in
enced by the stage of the disease. A predominantly nodular patients with PLCH have been shown to reflect the pres-
pattern is seen in the early stages of PLCH, and a predomi- ence of respiratory bronchiolitis and DIP-like changes on
nantly cystic pattern, in the later stages (167,168,331). It has biopsy specimens and to correlate with the cumulative
been postulated that the following sequence of abnormali- exposure to cigarettes smoked (338). Mosaic perfusion on
ties occurs in PLCH: nodules, cavitary nodules, thick-walled inspiratory scans and air trapping on expiratory scans may
cysts, and confluent cysts (167,168,337). The nodules also be seen (339).
have been shown to correlate with active peribronchiolar The characteristic findings of cysts and nodules
granulomas (337). Cystic lesions may represent ectatic throughout the middle and upper lung zones with relative
bronchioles (332) or, less commonly, cavitary granulomas sparing of the lung bases allow confident diagnosis of
or fibrous cystic lesions (337). PLCH on HRCT in the vast majority of cases (5,340,341).
Because the vast majority of patients are cigarette smok-
ers, emphysema is commonly present, although usually
mild. Ground-glass opacities are seen in approximately
TABLE 8-16
FINDINGS IN PULMONARY LANGERHANS
CELL HISTIOCYTOSIS
Thin-walled cysts, some with bizarre shapes
Nodules, usually smaller than 10 mm, centrilobular
Upper and middle lung zone predominance
Relative sparing of lung bases
Other common findings Figure 8-105 Pulmonary Langerhans cell histiocytosis. High-
Ground-glass opacities resolution CT demonstrates several small nodules in the upper
Fine reticular opacities lobes. Cavitation is present in one of the nodules in the right
Emphysema upper lobe (arrow). The patient was a 30-year-old man with pul-
monary Langerhans cell histiocytosis.
A B
Figure 8-106 Pulmonary Langerhans cell histiocytosis. A: High-resolution CT performed on a multidetector CT scanner shows numerous
thin-walled cysts (straight arrows) and small nodules (curved arrows) in the upper lobes. Also noted are irregular linear opacities consistent
with fibrosis, focal ground-glass opacities, and emphysema (arrowheads). B: Coronal reformation demonstrates the characteristic distribu-
tion of disease mainly in the upper and middle lung zones with relative sparing of the lung bases. On both the cross-sectional and coronal
images, it is difficult to distinguish some of the cysts due to Langerhans cell histiocytosis from emphysema. The patient was a 30-year-old
woman. (Case courtesy of Dr. Eduardo Sabbagh, Santiago, Chile.)
Centrilobular emphysema can usually be readily distin- pneumothorax (345,346). Other common manifestations
guished by the lack of visible walls and the presence of include hemoptysis, chylous effusion, renal angiomy-
vessels within the focal areas of lung destruction (340). olipoma, and chylous ascites (190,347).
The cysts in lymphangioleiomyomatosis are typically seen The chest radiograph is normal in 20% to 30% of
throughout the lungs without any zonal predominance patients with LAM (165,166,348). The most common
and are not associated with nodules (165,166). The cysts radiographic finding consists of diffuse bilateral reticula-
of honeycombing in IPF typically involve mainly the sub- tion (166,348). The lung volumes can be normal or
pleural and basilar lung regions and are associated with increased. Other common findings include recurrent
reticulation and decreased lung volumes. Upper lobe pneumothorax and pleural effusion (165,190,347).
cystic lesions identical to those in PLCH may be seen in The characteristic CT findings consist of thin-walled
Pneumocystis jiroveci pneumonia (342). Distinction can cysts, usually measuring a few millimeters to 2 cm in
be usually be readily made by clinical history and the pres- diameter, distributed diffusely throughout the lungs
ence of extensive ground-glass opacities in patients with (Fig. 8-107) (Table 8-17) (165,166,349). Initially, only a
Pneumocystis pneumonia. few scattered cysts measuring less than 1 cm in diameter
may be identifiable. With progression, the cysts become
larger, measuring up to 5 cm in diameter, and uniformly
Pulmonary Lymphangioleiomyomatosis
distributed throughout the lungs. The cysts tend to involve
Lymphangioleiomyomatosis (LAM) is a rare disease char- the central and peripheral lung regions and all lung zones
acterized by proliferation of abnormal smooth muscle to a similar extent (Fig. 8-108). In the vast majority of
cells along distal airways (resulting in airway obstruction, cases, the intervening lung appears strikingly normal,
cyst formation, and pneumothorax), vessels (resulting despite almost total replacement of lung tissue in cases
in pulmonary hemorrhage), and lymphatics (resulting in with advanced disease (165,166). Not surprisingly, similar
lymphatic obstruction, chylothorax, and lymph node CT findings have been reported to occur in patients with
enlargement) (190,343,344). It occurs almost exclusively tuberous sclerosis (165,346).
in women of childbearing age (190,343). It can occur spo- Occasionally ground-glass opacities, septal lines, or cen-
radically or in association with tuberous sclerosis (344). trilobular nodules may be seen (346,350,351). Pulmonary
The most common symptoms are dyspnea and recurrent hemorrhage occurs in approximately 10% of women with
TABLE 8-17
FINDINGS IN LYMPHANGIOLEIOMYOMATOSIS
Thin-walled lung cysts, usually round
Diffuse distribution throughout the lungs
Common associated findings
Pneumothorax
Chylous pleural effusion
Lymphadenopathy
Renal angiomyolipoma
Uncommon findings
Ground-glass opacities
Figure 8-107 Lymphangioleiomyomatosis (LAM). High-resolution Mild septal thickening
CT at the level of the aortic arch shows bilateral thin-walled cysts Centrilobular nodules
(arrows) randomly distributed throughout the upper lobes. The
patient was a 58-year-old woman with LAM.
LAM and presumably accounts for the ground-glass opaci- The most common extrapulmonary manifestations of
ties (346). Pulmonary edema and hemorrhage may result LAM are pneumothorax, chylothorax, mediastinal and/
from involvement of the venules, leading to vascular occlu- or retrocrural lymphadenopathy, and renal angiomyolipo-
sion, pulmonary venous hypertension, and hemoptysis. mas (346,352). Other findings include thoracic duct
Small centrilobular nodules result from peribronchiolar dilatation, abdominal lymphangioleiomyomas, and chy-
accumulation of smooth muscle cells (346). lous ascites (344).
A B
Figure 8-108 Lymphangioleiomyomatosis (LAM). See Color Figure 8-108C. A: High-resolution CT (1-mm collimation) performed on a multi-
detector CT scanner at the level of the bronchus intermedius shows numerous thin-walled cysts diffusely distributed throughout both lungs
without any central or peripheral predominance. B: Sagittal reformation image of the right lung shows diffuse distribution of the cysts in the
cephalocaudad plane without any upper lobe or basal predominance. The diffuse distribution of cysts throughout the lungs allows distinction of
LAM from pulmonary Langerhans cell histiocytosis (compare with Figs. 8-103 and 8-106). C: Pathologic section from another patient with LAM
shows diffuse distribution of cysts throughout the lung.
End-Stage Pulmonary Fibrosis (Honeycombing)

Honeycombing is the result of several CILDs with replace-
ment of normal lung by cystic spaces separated by areas of
dense fibrosis (234). The cystic spaces represent dilated
respiratory bronchioles and alveolar ducts (232). The
same process of fibrosis around membranous bronchioles
and bronchi results in traction bronchiolectasis and
bronchiectasis (232).
Honeycombing produces a characteristic cystic appear-
ance on HRCT that allows a confident diagnosis of lung
fibrosis (32,95). On HRCT, the cystic spaces of honey-
combing usually measure 0.5 to 1 cm in diameter and
have clearly definable walls 1 to 3 mm thick (32,95).
Honeycomb cysts typically share walls and occur in several
contiguous layers (Fig. 8-109). Honeycombing is usually
associated with other findings of lung fibrosis, such as
architectural distortion, intralobular interstitial thickening,
traction bronchiectasis, traction bronchiolectasis, and
irregular linear opacities.
The presence of honeycombing on HRCT is indicative
of significant lung fibrosis and in most cases should lead
to a diagnosis of UIP and a consideration of its most-
C common causes, particularly IPF and collagen-vascular
Figure 8-108 (continued) diseases, most notably rheumatoid arthritis (93,190). Less
common causes of honeycombing include asbestosis, HP,
sarcoidosis, DIP, and NSIP (93,190,191).
Honeycombing in patients with IPF and asbestosis is
usually most severe in the subpleural lung regions and at
The presence of many small, thin-walled cysts scattered
the lung bases (93,96). In the majority of patients first
through both lungs in a young woman is highly suggestive
seen with clinical features of IPF, the presence of a
of LAM. In selected cases, differentiation between LAM
and pulmonary emphysema may present difficulties
because of the seeming lack of an identifiable cyst wall in
both disorders. However, uniform distribution throughout
both lungs strongly favors the diagnosis of LAM. The lack
of walls and the residual core lobular structures within
areas of emphysema usually allows ready distinction
between these two entities (340). The cystic changes in
LAM may be identical to those of Langerhans cell histiocy-
tosis. However, the cysts in Langerhans histiocytosis tend
to show relative sparing of the lung bases, whereas the
cysts in LAM involve all lung zones to a similar extent.
Other helpful findings are the frequent presence of
bizarre-shaped cysts and nodules in Langerhans cell histio-
cytosis. The accuracy of HRCT in distinguishing PLCH,
LAM, and emphysema is excellent. In one study, two
observers were confident of the diagnosis of LAM in 79%
of cases and were correct in all of these (340). In a second
study including various chronic cystic lung diseases, two
independent observers made a confident diagnosis of LAM
Figure 8-109 Honeycombing in end-stage idiopathic pulmonary
in 12 (67%) of 18 patients. This diagnosis was correct in fibrosis (IPF). High-resolution CT at the level of the bronchus inter-
88% of cases (341). Misdiagnosis occurred in cases of medius shows honeycombing (arrows) in the subpleural regions of
PLCH presenting with cysts distributed diffusely through- the right lung and left lower lobe and diffusely in the lingula. Also
noted is associated volume loss of the left lung. The findings are
out the lungs and in cases with LAM associated with characteristic of end-stage fibrosis in IPF. The patient was a
centrilobular nodules (341). 67-year-old man with idiopathic pulmonary fibrosis.
A B
Figure 8-110 End-stage fibrosis in asbestosis. A: High-resolution CT demonstrates extensive bilateral honeycombing involving mainly
the subpleural lung regions. The appearance is identical to that of end-stage idiopathic pulmonary fibrosis. B: Mediastinal windows demon-
strate bilateral pleural plaques (arrows) characteristic of asbestos exposure. The patient was a 59-year-old with end-stage fibrosis caused by
asbestosis.
predominantly subpleural and basal distribution of fibro- if the primary focus of destruction centers on airspaces
sis on HRCT can be sufficiently characteristic to obviate surrounding respiratory bronchioles, as typically occurs in
biopsy (Figs. 8-46 and 8-109) (96,97,353). Although hon- smokers; panacinar (panlobular) emphysema if airspaces
eycombing in asbestosis may have a similar distribution, are evenly destroyed throughout the acinus, as occurs in
the diagnosis can usually be confidently made on HRCT patients with a1-antitrypsin deficiency; paraseptal (distal
based on the presence of associated pleural plaques or acinar) emphysema, if selective involvement is found of
diffuse pleural thickening (Fig. 8-110) (93). The honey- the distal acini, which, when confluent, leads to the forma-
combing in chronic HP may be most marked in the tion of bullae, with resultant spontaneous pneumothorax,
subpleural lung regions but is more often patchy in distri- as occurs especially in younger patients (355). Irregular
bution, and tends to be most severe in the mid lung zones airspace enlargement, also known as irregular or paracica-
with relative sparing of the lung bases (Fig. 8-79) tricial emphysema, is seen in association with focal scars
(93,100). Honeycombing in sarcoidosis usually has upper or with diffuse cicatrizing lung diseases, including tubercu-
lobe predominance (85,93). Characteristic findings of losis, chronic sarcoidosis, and certain pneumoconioses,
end-stage fibrosis in sarcoidosis include conglomerate especially silicosis. It cannot be overstated that differentia-
masses, perihilar cystic changes caused by markedly tion among these various forms of emphysema may be
ectatic bronchi, and mild subpleural honeycombing exceedingly difficult, even for the pathologist, especially
(85,93). The diagnosis can usually be readily made by the when the emphysema is severe and when more than one
characteristic predominance of fibrosis in the perihilar type is present at the same time, as occurs not infrequently.
regions of the mid and upper lung zones (Figs. 8-70 and The only direct sign of emphysema on the radiograph is
8-71). The ability to make a confident diagnosis on HRCT the presence of bullae (356). Indirect signs include focal
in the majority of patients with honeycomb lung is of sig- absence of pulmonary vessels and reduction in vessel
nificant clinical utility. Areas with honeycomb lung do caliber with tapering toward the lung periphery (357–359).
not yield a specific histologic diagnosis and therefore These findings, however, are nonspecific. Limitations of
should be avoided on surgical lung biopsy (2). chest radiography include low sensitivity in the detection
of mild emphysema, low specificity, considerable interob-
server disagreement in the interpretation of findings, and
EMPHYSEMA inability to quantify the severity of emphysema (173,356).
Emphysema is characterized on HRCT by the presence
Emphysema is defined as condition of the lung character- of localized areas of low attenuation without definable
ized by an abnormal enlargement of the airspaces distal to walls and easily separable from surrounding lung
the terminal bronchioles, accompanied by destructive parenchyma (Table 8-18) (360–362). The pattern and
changes of the alveolar walls (190,354). Emphysema is distribution of emphysema on HRCT are influenced by
usually classified anatomically based on the distribution the type of emphysema (173,175,360). Centrilobular
of abnormalities within the acinus or secondary lobule emphysema is characterized by the presence of multi-
and includes centrilobular (proximal acinar) emphysema, ple small, round areas of abnormally low attenuation,
TABLE 8-18
FINDINGS IN EMPHYSEMA
Centrilobular emphysema
Multiple, small, focal lucencies
Usually no visible walls
Often small vessels present within focal lucencies
Upper lobe predominance
Often associated with paraseptal emphysema or bullae
Panacinar emphysema Figure 8-111 Centrilobular emphysema. High-resolution-CT
Widespread lucent lung containing small pulmonary vessels shows focal areas of decreased attenuation with no definable walls
Diffuse or lower lobe predominance (straight arrows). Note that in several areas, small centrilobular
Focal lucencies and bullae less common than with vessels can be seen within the areas of decreased attenuation
(curved arrows). The patient was a 50-year-old man with centrilob-
centrilobular emphysema
ular emphysema.
Bronchiectasis (in 40% of patients with a1-antitrypsin
deficiency)
Paraseptal emphysema
Multiple, subpleural lucencies in a single layer (32,40). The areas of lucency typically have no identifi-
Usually less than 1 cm able walls. However, in patients who have centrilobular
May be associated with centrilobular emphysema
or bullae
emphysema, bullae within the lung may have visible
thin walls.
Panacinar emphysema is characterized by uniform
destruction of the secondary lobule, leading to widespread
areas of abnormally low attenuation (Fig. 8-41)
several millimeters in diameter, distributed throughout (176,363,364). A decrease in the number and size of
the lung, but usually having upper lobe predominance pulmonary vessels is noted in the affected lung. Panacinar
(Figs. 8-111 and 8-112). The areas of lucency often appear emphysema is almost always most severe in the lower-lung
to be grouped near the centers of secondary pulmonary zones (Fig. 8-113). Pulmonary vessels in the affected lung
lobules, surrounding the centrilobular artery branches appear fewer and smaller than normal and may be quite
A B
Figure 8-112 Centrilobular emphysema. A: High-resolution CT performed on a multidetector CT scanner shows focal areas of emphy-
sema (arrows) in the left upper lobe and more-confluent emphysema in the right upper lobe. B: Coronal reformation demonstrates that the
emphysema (arrows) involves almost exclusively the upper lobes and is most severe in the lung apices. The patient was a 67-year-old man
with centrilobular emphysema.
A B
Figure 8-113 Panacinar emphysema. A: High-resolution CT image from a multidetector CT scan at the level of the lower lung zones
shows severe diffuse emphysema. B: Coronal reformation demonstrates that the emphysema involves almost exclusively the lower lobes.
Minimal emphysema is seen in the upper lobes. The lower-lobe changes are characteristic of panacinar emphysema. The patient was a 53-
year-old man with panacinar emphysema due to a1-antitrypsin deficiency.
inconspicuous. Associated paraseptal emphysema and (MinIP) technique (21). MinIP is a software program that
bullae are much less common than in centrilobular identifies only areas of lung with the lowest attenuation
emphysema. Panacinar emphysema is frequently associ- and thus suppresses normal lung and pulmonary vessels.
ated with a1-antitrypsin deficiency (363). In approxi- The technique involves obtaining contiguous thin sec-
mately 40% of patients with a1-antitrypsin deficiency, tions (e.g., 1 mm) through a volume of lung tissue. The
bronchiectasis also develops (364,365). Mild and even thin-section images are reconstructed individually or as a
moderately severe panacinar emphysema can be very sub- slab several millimeters in thickness.
tle and difficult to detect on HRCT (362). Furthermore, The overall extent of emphysema may be difficult to
diffuse panacinar emphysema unassociated with focal estimate visually because of the wide range of volumes
areas of lung destruction or bullae may be difficult to dis-
tinguish from diffuse small airways obstruction and air
trapping resulting from bronchiolitis obliterans (366).
Paraseptal emphysema involves the periphery of the
secondary lobule, near the interlobular septa and sub-
pleural lung regions (175,362). Paraseptal emphysema
is the easiest form of emphysema to recognize on HRCT
(Fig. 8-114) (362).
Several studies have shown that HRCT correlates closely
with the macroscopic pathologic findings of emphysema
(361,367,368) and that it is superior to chest radiography
and pulmonary function tests in the detection and
quantification of severity of emphysema (356,369,370).
However, mild emphysema can be missed on HRCT
(21,362). Therefore, although CT is undoubtedly the most
sensitive method to diagnose emphysema in vivo, it does
not detect the early stages of emphysema and cannot be Figure 8-114 Paraseptal emphysema. High-resolution CT at the
used to rule out the diagnosis definitively (173,356). level of the aortic arch shows extensive emphysema in the sub-
pleural lung regions. The appearance is characteristic of paraseptal
Assessment of mild emphysema on CT scanning is emphysema. Also noted is mild associated centrilobular emphy-
improved by using the minimum intensity projection sema. The patient was a 36-year-old woman.
represented in the different images. This difficulty can be

circumvented by using various techniques that objectively
quantify emphysema on CT. These include use of a thresh-
old value below which emphysema is considered to be
present (density mask or pixel index) (367,371), assess-
ment of the range of lung densities represented in a lung
slice (histogram analysis) (372), and assessment of overall
lung density, often in combination with volumetric
imaging (373,374). It has been shown that a threshold
of –910 HU, which highlights all pixels with attenuation
values less than –910, correlates best with the extent of
emphysema when CT is performed by using 7- to 10-mm
collimation and that –950 HU correlates best with the
extent of emphysema on HRCT (367,371).
Although CT is the most accurate method for diagnos-
ing emphysema in vivo in most patients, the diagnosis can
be readily made by a combination of clinical history, Figure 8-115 Bullous emphysema. High-resolution CT at the
level of the inferior pulmonary veins shows several large bullae in
pulmonary function tests, and chest radiographs. The right lung. Note associated increased attenuation of the right mid-
main indication for the use of CT is in the preoperative dle lobe due to compressive atelectasis. Relatively mild emphysema
assessment of patients being considered for bullectomy is present in the left lung. The patient was a 77-year-old man.
(375,376) or for lung volume reduction surgery (LVRS)
(377). Bullectomy is most effective in patients with large
bullae and absence of generalized emphysema and who both objective and subjective measures of lung perform-
have rapid progression of dyspnea and demonstrate ance in properly selected patients (380,381). Inclusion
restrictive lung function caused by compression of normal criteria include severe emphysema with hyperinflation,
areas of lung. Large bullae can compress the remaining forced expiratory volume in 1 second between 10% and
lung parenchyma and cause further functional and clinical 40% of predicted, residual volume equal to or greater than
impairment. CT allows assessment not only of the extent 180% of predicted, and total lung capacity equal to or
of bullae but also of the degree of compression and the greater than 110% of predicted (377). The best surgical
severity of emphysema in the remaining lung (Fig. 8-115). candidates are patients with hyperinflation caused by
Bullectomy is applicable only to a small, highly selected severe emphysema rather than airway disease, heteroge-
group of patients. A much more common procedure for the neous distribution of the emphysema with specific target
treatment of severe emphysema is LVRS. The procedure areas for surgical resection in the upper lobes, and evidence
consists of bilateral wedge resection of the most severely of compression of relatively normal lung (Fig. 8-116)
involved lung (378,379). LVRS has been shown to improve (378,382). CT allows excellent assessment of the presence,
A B
Figure 8-116 Emphysema in a patient being considered for lung-volume reduction surgery (LVRS). A: High-resolution CT at the level of the
aortic arch shows severe upper lobe emphysema. B: High-resolution CT at the level of the lung bases shows mild emphysema. The patient
was a 61-year-old woman who had marked hyperinflation, forced expiratory volume in 1 second between 10% and 40% of predicted, and
residual volume greater than 180% predicted. Given the severity and distribution of emphysema on high-resolution CT, she was considered a
good surgical candidate for LVRS.
A B
Figure 8-117 Bullae in recurrent pneumothorax. A: High-resolution CT image at the level of the lung apices from a CT scan performed
on a multidetector scanner shows several apical bullae (arrows). B: Coronal reformation demonstrates that the only abnormalities are apical
bullae (arrows). The patient was an 18-year-old man with recurrent pneumothorax.
extent, and localization of areas of emphysema before LVRS injuries (140). Although HRCT plays a limited role in the
(377,383). The analysis may be based on subjective assess- differential diagnosis of acute lung disease in the immuno-
ment of the images or the use of objective quantification of competent host, it is being used with increasing frequency in
areas with decreased attenuation by using a threshold value the assessment of acute lung disease in the immunocompro-
for emphysema (377,384,385). mised patient (139,388,389).
HRCT may also be indicated for assessing patients who Pulmonary lung disease is a major cause of morbidity
have dyspnea and decreased carbon monoxide diffusing and mortality in AIDS and in non-AIDS immunocom-
capacity without evidence of obstruction of airflow (370). promised patients. Early diagnosis is important in direct-
In such patients, HRCT can help differentiate emphysema ing treatment (390) and reducing the mortality rate
from pulmonary, vascular, or interstitial lung disease from pulmonary complications in immunocompromised
(370). CT may also be useful in the early detection of patients (391).
apical subpleural bullae in patients with idiopathic The clinical and chest radiographic findings associated
spontaneous pneumothorax (386,387). This form of with acute infectious and/or infiltrative lung disease in
pneumothorax occurs most often in tall, young adults and immunocompromised patients have been extensively
is thought to be caused by rupture of a subpleural bulla. described, but in many cases, a timely diagnosis proves
Subpleural emphysema or bullae have been demonstrated elusive. Although the chest radiograph remains the first
in the vast majority of these patients surgically and on and foremost imaging modality in these patients, it may
HRCT, including lifelong nonsmokers (386,387). The be normal in up to 10% of patients with proven pul-
emphysema in these patients tends to involve mainly the monary disease, and it seldom allows a confident specific
periphery of the upper lung zones, a distribution consis- diagnosis (392,393). A number of studies have shown that
tent with paraseptal emphysema (Fig. 8-117) (386,387). HRCT may demonstrate parenchymal abnormalities in
patients with normal or questionable radiographic find-
ings and may allow confident diagnosis in patients with
ACUTE LUNG DISEASE IN THE nonspecific radiographic findings (Figs. 8-118 and 8-119)
IMMUNOCOMPROMISED PATIENT (394–398). The HRCT findings in immunocompromised
patients have been shown to reflect the pathologic find-
Acute lung disease may result from a variety of causes ings closely (399–402). Because HRCT is superior to the
including infection, cardiogenic and noncardiogenic edema, radiograph in demonstrating the presence, distribution,
hemorrhage, drug reactions, aspiration, and inhalation and extent of parenchymal abnormalities, HRCT may also
A B
Figure 8-118 Drug reaction. A 28-year-old woman had progressive shortness of breath after bone marrow transplant. A: Chest radiograph
was interpreted prospectively as being normal. B: High-resolution CT demonstrates areas of ground-glass attenuation mainly in the lower
lobes. Based on the CT findings, the patient underwent lung biopsy, which showed interstitial pneumonitis consistent with drug reaction. The
final clinical diagnosis was drug reaction to carmustine.
be helpful as a guide to the optimal type and site of lung malignancy such as leukemia or lymphoma, chemotherapy
biopsy (395,403). of tumors, and transplantation. Patients with AIDS are
Although a number of different diseases may cause dif- more susceptible than nonimmunocompromised individ-
fuse pulmonary infiltrates in the immunocompromised uals to develop community-acquired pneumonias, but the
host, the majority of cases are caused by infection most common organism to lead to life-threatening compli-
(393,404). The type of complication is influenced by the cations in these patients is Pneumocystis jiroveci pneumonia.
specific immunologic abnormality. The most common Although patients with AIDS may have invasive aspergillo-
causes of immunosuppression are AIDS, hematologic sis, this complication is seen much more commonly in
A B
Figure 8-119 Accuracy of CT versus radiography in diffuse AIDS-related lung disease. Infectious bronchiolitis mimicking Pneumocystis
pneumonia (PCP) on chest radiography. A: Chest radiograph in a 32-year-old man with a productive cough demonstrates diffusely increased
coarse reticular lung markings, mimicking interstitial infiltrate seen in PCP (compare with C). B: A 1-mm section through the lung bases clearly
demonstrates bronchial wall thickening (large white arrows), bronchiectasis (open arrow), and centrilobular impacted bronchioles (black ar-
rows). Note the absence of interstitial disease or ground-glass attenuation. Sputum culture grew mixed pyogenic organisms; no other infec-
tions were diagnosed at bronchoscopy. C: Chest radiograph in a 44-year-old man with PCP and a dry cough demonstrates a similar pattern of
coarse interstitial markings. D: A 1.5-mm section through the lung bases, however, depicts a distinctly different pattern of focal ground-glass
attenuation (arrows) consistent with alveolitis and reticular markings (open arrows), reflecting thickened interlobular septa (continued ).
C D
patients with leukemia who are undergoing immunosup- been shown that the degree of hypoxemia in patients
pressive therapy. Cytomegalovirus pneumonia, conversely, with PCP correlates with the degree of inflammatory
is seen most commonly in patients undergoing organ reaction and not with the number of organisms (409).
transplantation. The type of complication is also influ- Immunocompromised non-AIDS patients with PCP tend
enced by the severity of immunosuppression. For example, to have more severe inflammatory reaction and therefore
in patients undergoing solid organ or hematopoietic stem be more hypoxemic than patients with AIDS (409).
cell (bone marrow) transplantation, opportunistic infec- Radiographically, the typical features of PCP consist of
tion is seen most commonly during the first month after bilateral perihilar ground-glass or fine reticulonodular
transplantation, whereas post-transplant lymphoprolifera- opacities that become more diffuse and progress to consol-
tive disorders and bronchiolitis obliterans occur several idation with increased severity of disease (406,410).
months to years after transplantation. It is essential, there- Atypical features include asymmetric distribution, apical
fore, to interpret the HRCT findings in the appropriate disease, nodular opacities, cavitary nodules, and cysts
clinical context. In this section, we review the CT manifesta- (411–413). In the majority of cases of PCP, the diagnosis
tions of the most common intrathoracic complications can be readily be made by a combination of clinical and
seen in immunocompromised patients. radiographic findings. However, in 5% to 10% of patients
with PCP, the radiograph is normal (410,414,415). HRCT
can be helpful in the assessment of symptomatic patients
Pneumocystis jiroveci Pneumonia
with normal or questionable radiographic findings and
Pneumocystis jiroveci pneumonia (PCP) is a common oppor- in the assessment of patients with atypical presentation.
tunistic infection in immunocompromised AIDS and Several studies have shown that the vast majority of
non-AIDS patients (405,406). The incidence and severity of
PCP in patients with AIDS have markedly decreased since
the advent of highly active antiretroviral therapy (HAART)
(405,406). However, PCP remains a significant problem,
especially as a presenting illness in patients not yet known
to be infected with HIV (405,406). In contradistinction to
the decreasing incidence in AIDS, the incidence of PCP in
immunocompromised patients who do not have AIDS
increased considerably in the last decade, and the mortality
in this population ranges from 30% to 60% (406). In con-
trast to the usual insidious onset seen in patients with
AIDS, PCP in the non-AIDS immunocompromised host is
usually acute and manifested by an abrupt onset of fever,
dyspnea, and respiratory insufficiency (407,408). Non-
AIDS immunocompromised patients with PCP have fewer Figure 8-120 Pneumocystis pneumonia (PCP). High-resolution
CT image at the level of the aortopulmonary window shows exten-
organisms than do patients with AIDS (409), and therefore sive bilateral ground-glass opacities. The patient was a 39-year-old
induced sputum and BAL are less likely to be positive. It has man with AIDS and PCP.
masses (Fig. 8-122). Less-common manifestations include

nodules, centrilobular nodular opacities, asymmetric con-
TABLE 8-19
solidation, interlobular septal thickening, intralobular
linear opacities, pneumothorax, pleural effusion, and
FINDINGS IN PNEUMOCYSTIS JIROVECI lymphadenopathy (Fig. 8-123) (84,171,412). The paren-
PNEUMONIA chymal abnormalities in PCP can be diffuse but often
Characteristic findings involve mainly the perihilar regions or upper lobes
Patchy or diffuse bilateral ground-glass opacity (135,171,416).
Central, perihilar, or upper lobe predominance The HRCT findings in patients who have PCP reflect
Less common findings the stage of disease (205,418). Initially, scattered foci of
Consolidation ground-glass opacity or airspace consolidation can be iden-
Thin-walled cysts tified. With time, interstitial abnormalities develop and
Pneumothorax related to cysts
may predominate (Fig. 8-124). In treated patients who
Reticulation and septal thickening (resolving disease)
Bronchiectasis or bronchiolectasis have resolving or subacute infection, reticular opacities rep-
resenting thickened interlobular septa and intralobular
Uncommon findings
Thick-walled, irregular, septated cavities
lines can be seen in association with ground-glass opacity
Small nodules, centrilobular or diffuse (e.g., crazy-paving pattern). Less frequently, PCP results in
Large nodules or masses mild, peripheral bronchiectasis and/or bronchiolectasis,
Lymphadenopathy presumably the result of PCP bronchiolitis (419).
Pleural effusion The accuracy of CT in the diagnosis of PCP and other
pulmonary complications of AIDS was assessed in a study
of 122 patients (84). The most common pulmonary
patients with PCP and normal radiographs have abnormal complication in these patients was PCP, present in 35 cases.
and characteristic findings on HRCT (397,416,417). Based on the CT findings, the two observers made a confi-
The characteristic HRCT findings of PCP consist of dent diagnosis of PCP in 25 cases, and this diagnosis was
symmetric bilateral ground-glass opacities (Fig. 8-120) correct 94% of the time. The diagnosis of PCP in these
(Table 8-19) (84,135,171). A distinct “mosaic” pattern can cases was based on the presence of areas of ground-glass
often be identified, with areas of normal lung intervening opacity. False-positive diagnoses in 6% of the cases were
between scattered, focal ground-glass opacities (Fig. 8-121) due to motion or poor inspiratory effort, resulting in appar-
(135,171). With more severe disease, extensive bilateral ent areas of ground-glass opacity. Although the appearance
consolidation develops. Cystic changes are identified on of ground-glass opacity can be seen in immunosuppressed
HRCT in approximately 30% of patients who have PCP patients as a result of other pneumonias, they are much less
and AIDS (84,171,412) but are uncommon in non-AIDS common. In a second study, the authors compared the
patients. The cysts can have thin or thick walls, and sensitivity and specificity of chest radiography and CT in
as they enlarge, they may form multiseptated cystic the detection of pulmonary infections and tumors in
A B
Figure 8-121 Pneumocystis pneumonia (PCP) with mosaic pattern. A: High-resolution CT image at the level of the aortic arch from a multi-
detector CT scan shows extensive bilateral ground-glass opacities. Scattered areas of normal or less involved lung between the ground-glass
opacities result in a “mosaic” pattern. B: Coronal reformation shows mosaic pattern throughout all lung zones. The patient was a 36-year-old
man with acquired AIDS and PCP.
Figure 8-122 Pneumocystis jiroveci pneumonia with cyst forma-

tion. High-resolution CT image at the level of the aortic arch shows
bilateral ground-glass opacities, small foci of consolidation, and
several thick-walled cysts (arrows) in the left upper lobe. The patient
was a 60-year-old woman with AIDS and Pneumocystis pneumonia.
Figure 8-123 Unusual manifestations of Pneumocystis pneu-

monia (PCP). A: High-resolution CT with 1.0-mm collimation in a
34-year-old man with AIDS-related lymphoma demonstrates pro-
fusion of 2- to 3-mm nodules, consistent with miliary PCP. At
transbronchial biopsy, poorly formed granulomas were identified
in the lung interstitium, associated with P. jiroveci organisms in the
airspaces. PCP occasionally incites a granulomatous response,
even during acute infection. B: A 34-year-old man with PCP
appreciated as poorly marginated subcentimeter lung nodules
(arrows). C: Dense, segmental consolidation in the lingula, middle
lobe, and right lower lobe is appreciated at high-resolution CT in
C this 37-year-old man with PCP.
A B
Figure 8-124 Acute and subacute Pneumocystis pneumonia (PCP) in a 36-year-old man with AIDS. A: A 1.5-mm section through the
lower lobes shows patchy ground-glass opacities (arrows). Fine reticular changes are superimposed on these areas of involved lung,
attributable to interstitial and interlobular infiltration by inflammatory cells. B: Follow-up CT at the same level 3 weeks after initiation of
antibiotic therapy demonstrates regression of ground-glass attenuation as the alveolitis resolves, and replacement by coarse reticulation in
the same areas (arrow) as the predominant finding.
139 patients who had AIDS (397). Of the 106 patients who suggestive of PCP, in immunocompromised patients who
had intrathoracic complications, 90% were correctly identi- do not have AIDS, ground-glass opacity is a less specific
fied on the radiograph and 96% on CT. Of 33 patients who finding. In a study in which the HRCT findings were
did not have intrathoracic disease, 73% were correctly correlated with pathologic specimens in 33 immunocom-
identified at radiography and 86% at CT. A confident first- promised non-AIDS patients who had acute pulmonary
choice diagnosis by two independent observers was most complications, 14 had predominantly ground-glass opac-
often correct in Kaposi sarcoma (31 of 34 interpretations, ity (137). In these 14 patients, PCP accounted for the areas
91%), PCP (33 of 38 interpretations, 87%), and lymphoma of ground-glass opacity in three cases, cytotoxic drug
(4 of 4 interpretations, 100%) (397). reaction in four, COP in four, lymphoma in two, and CMV
HRCT is particularly helpful in detecting PCP when pneumonia in one.
clinical and plain film assessment is inconclusive. In one
prospective study of 51 AIDS patients who had a high clin-
Cytomegalovirus Pneumonia
ical probability of PCP, and in whom chest radiographic
findings were normal, equivocal, or nonspecific, the sensi- Cytomegalovirus (CMV) is a major cause of pneumonia in
tivity of HRCT in detecting PCP was 100%, with a speci- solid organ and hematopoietic stem cell (bone marrow)
ficity of 89% and an accuracy of 90% (416). In this study, transplant patients (421). Infection typically occurs 1 to
patchy or nodular ground-glass attenuation visible on 6 months after transplantation (402,421,422). The vast
HRCT was considered to indicate possible PCP. In a majority of episodes of CMV disease in these patients result
second study (417), HRCT findings were compared with from reactivation of latent virus in seropositive recipients
BAL results in 13 HIV-positive patients who had a strong (421). Recent studies have also shown an increase in CMV
clinical suspicion of PCP and a normal chest radiograph. pneumonia in nontransplanted adult leukemia patients
The patients all had a CD4 count of less than 200 cells per (423). CMV was also an important cause of morbidity
mm3, a nonproductive cough or nonpurulent sputum, and mortality in AIDS patients before the introduction
fever, and dyspnea or decreased exercise tolerance. Four of highly active antiretroviral therapy (HAART) (418,
patients had patchy ground-glass opacities visible on 424,425). Since the introduction of HAART, the incidence
HRCT. All four proved to have PCP on BAL. None of the of CMV infection in AIDS has declined dramatically.
nine patients who were negative for PCP on BAL had However, it still can be seen in patients who did not receive
ground-glass opacity on HRCT. Based on these data, in our HAART and in patients in whom resistance or intolerance
judgment and that of others, the presence of bilateral to HAART has developed (424).
ground-glass opacities on HRCT in an HIV-positive patient The clinical manifestations of CMV pneumonia include
is sufficiently suggestive of the diagnosis of PCP either to fever, cough, and hypoxemia (421). The radiographic man-
base initiation of empiric therapy (420) or to pursue defin- ifestations are variable and may consist of a reticular
itive diagnosis with bronchoscopy. or reticulonodular pattern, ground-glass opacities, airspace
It should be noted that although the presence of areas consolidation, small nodules, or a combination of these
of ground-glass opacity in patients with AIDS is most patterns (426–428).
A B
Figure 8-125 Cytomegalovirus pneumonia. High-resolution CT at the level of the aortic arch (A) and at the level of the bronchus inter-
medius (B) shows bilateral ground-glass opacities, poorly defined centrilobular nodules (curved arrows), and small foci of consolidation (straight
arrows). The patient was a 23-year-old man with cytomegalovirus pneumonia after hematopoietic stem cell (bone marrow) transplantation.
The HRCT manifestations of CMV pneumonia usually

consist of variable mixtures of three patterns: bilateral
TABLE 8-20
ground-glass opacities, areas of consolidation, and small
centrilobular nodular opacities (Fig. 8-125) (Table 8-20)
(429,430). In a review of the findings in 32 patients with
FINDINGS IN CYTOMEGALOVIRUS
CMV pneumonia, ground-glass opacities were seen in 66%, PNEUMONIA
areas of consolidation in 59%, and nodules in 59% (429). Characteristic findings
A small percentage of patients had only ground-glass opac- A mixture of patterns including bilateral ground-glass
ities, only consolidation, or only small nodules. Similar opacities, consolidation, and multiple centrilobular
findings have been described in other studies (430,431). nodules
Less common manifestations of CMV pneumonia include Less common findings
tree-in-bud pattern, thickening of the bronchovascular Only ground-glass opacities, consolidation, or nodules
Masslike areas of consolidation
bundles, and pleural effusions (429–431). Areas of dense
Tree-in-bud pattern
consolidation and masslike opacities have been described Pleural effusion
in patients with AIDS (Fig. 8-126) (400). In the proper
A
Figure 8-126 Spectrum of findings in AIDS-related cytomegalovirus (CMV) pneumonitis. A: A 1.5-mm section in a 32-year-old man with a
CD4 level of 35 cells/mm3 demonstrates areas of ground-glass attenuation in the lower portions of the lungs, consistent with alveolitis.
Within these areas, interlobular septal thickening is present (black arrows), best appreciated in the right lower lobe. Although the common-
est manifestation of CMV pneumonia in AIDS, this appearance overlaps with other causes of alveolitis, especially that due to Pneumocystis
jiroveci pneumonia. B: A 5-mm section in a 34-year-old man with AIDS. An irregularly marginated mass is present in the right lower lobe.
A transthoracic biopsy specimen (not shown) showed a solid mass of necrotic atypical lymphocytes with numerous cytomegalic inclusion
bodies. C: Enlargement of a 1-mm section through the right lung base in another patient with CMV infection manifesting as airway disease.
Thick-walled, ectatic bronchi are noted, with peripheral ill-defined centrilobular small nodules, consistent with impacted bronchi.
B C
clinical context, HRCT may be helpful in suggesting the dyspnea. The radiographic findings are nonspecific, consist-
diagnosis of CMV pneumonia; however, the CT findings by ing of ill-defined nodules or masses and subsegmental and
themselves are nonspecific and resemble those described in segmental consolidation (437,438).
other viral, bacterial, and fungal infections seen in the The characteristic HRCT findings of angioinvasive
immunocompromised host (388,432,433). aspergillosis consist of nodules surrounded by a halo of
ground-glass opacity (“halo sign”) or pleura-based, wedge-
shaped areas of consolidation (Figs. 8-127 and 8-128)
Invasive Aspergillosis
(Table 8-21) (436,439). The nodules with a halo sign
Invasive aspergillosis is a relatively common pulmonary correspond to foci of necrotic lung surrounded by viable but
complication in immunocompromised patients, character- hemorrhagic parenchyma (232,402). The wedge-shaped
ized by extension of Aspergillus into normal lung tissue areas of consolidation correspond to subsegmental or seg-
(232). It occurs almost exclusively in immunosuppressed mental hemorrhage or infarction (232,402). The nodules
patients with severe neutropenia and is particularly com- usually range from a few millimeters to 3 cm in diameter. In
mon in patients with acute leukemia and after solid organ hospitalized patients with severe neutropenia, the presence
and hematopoietic stem cell (bone marrow) transplanta- of nodules with a halo sign is highly suggestive of angio-
tion (421,434). Although it may be seen in patients who invasive aspergillosis (129,440).
have AIDS, it is uncommon and usually associated with Although the presence of nodules with a halo sign in
neutropenia or steroid therapy (435). Invasive aspergillo- neutropenic patients is highly suggestive of aspergillosis,
sis can be classified into two main subtypes: angioinvasive this finding is seen in only 60% to 80% of patients with
aspergillosis, which accounts for 70% to 85% of cases, and angioinvasive aspergillosis (128,129,440,441). Further-
acute bronchiolitis and bronchopneumonia (airway inva- more, a halo of ground-glass opacity has also been
sive aspergillosis), which accounts for 15% to 30% of cases described in other infections seen in these patients, includ-
(232,399). ing Candida, mucormycosis, cytomegalovirus, and herpes
simplex pneumonia, and in organizing pneumonia
(128,436,441). It has recently been suggested that the
Angioinvasive Aspergillosis
detection of angioinvasive pulmonary aspergillosis in
Angioinvasive aspergillosis is characterized histologically patients with antibiotic-resistant fever of unknown origin
by invasion and occlusion of small to medium-sized pul- under immunosuppression can be improved by showing
monary arteries by fungal hyphae (232,436). This leads to direct vessel involvement at a peripheral level with high-
the formation of necrotic hemorrhagic nodules or pleura- resolution (1.5-mm collimation) multidetector CT angiog-
based, wedge-shaped hemorrhagic infarcts. The clinical raphy (442). The diagnosis of vessel involvement was based
manifestations usually include fever, cough, and progressive on demonstration of an interruption of pulmonary arteries
A B
Figure 8-127 Angioinvasive aspergillosis. A: High-resolution CT image at the level of the lung apices from a multidetector CT scanner
shows large right upper lobe nodule surrounded by halo or ground-glass opacity (arrows). Also noted is mild emphysema. B: Coronal refor-
mation demonstrates two large upper lobe nodules with halo sign (straight arrows) and smaller nodule (curved arrow) in the right lower
lobe. The patient was a 59-year-old woman with acute myelogenous leukemia and angioinvasive aspergillosis.
directly at the border of the suspected lesion without visual-

ization of the vessel inside the lesion or immediately
TABLE 8-21
peripheral to the lesion on multiplanar and maximum
intensity projection (MIP) reconstructions. Fourteen focal
pulmonary lesions were detected by CT in eight patients. In
FINDINGS IN INVASIVE PULMONARY
four of five lesions with histologically proven fungal ASPERGILLOSIS
angioinvasion, vascular occlusion was detected on CT Angioinvasive aspergillosis
angiography. Only two of these five lesions had a CT halo Nodules or focal consolidation with a halo sign
sign. In all nine lesions in which angioinvasive aspergillosis Wedge-shaped pleural-based areas of consolidation
was excluded, CT angiography showed patent vessels. The Cavitary nodules with an air-crescent sign (usually late)
Aspergillus bronchiolitis and bronchopneumonia
Patchy peribronchial or lobular consolidation
Small centrilobular nodules
Tree-in-bud pattern
latter included one patient with a CT halo sign shown

histologically to be due to diffuse alveolar damage and a
patient with a nodule due to mucormycosis (442).
Another common finding in angioinvasive aspergillosis
is the development of a crescent-shaped air density within a
nodule or parenchymal consolidation (“air crescent” sign)
(436,443). This sign results from separation of fragments
of necrotic lung (pulmonary sequestra) from adjacent
parenchyma (436). It is usually a late sign typically being
seen during convalescence (e.g., 1–10 days after recovery
from neutropenia), although it may occasionally be seen at
the time of initial diagnosis (Fig. 8-129) (436,443). The air-
Figure 8-128 Angioinvasive aspergillosis. High-resolution CT crescent sign is seen in approximately 50% of patients, and
demonstrates irregularly marginated soft tissue nodules in the
upper lobes surrounded by a halo of ground-glass attenuation. develops almost exclusively in nodules greater than 1 cm in
The patient was a 77-year-old woman with leukemia. diameter and in areas of consolidation (443).
Figure 8-129 Angioinvasive aspergillosis. High-resolution CT

image at the level of the bronchus intermedius shows right lower
lobe nodule (curved arrow) with eccentric cavitation and air- Figure 8-130 Aspergillus bronchopneumonia (airway-invasive
crescent sign, cavitating left lower lobe nodule (straight arrow), aspergillosis). High-resolution CT demonstrates focal areas of peri-
and small foci of ground-glass attenuation and consolidation. The bronchial consolidation (arrows). The patient was a 55-year-old
patient was a 33-year-old man with angioinvasive aspergillosis man with acute myelogenous leukemia who had undergone
after hematopoietic stem cell transplantation. hematopoietic stem cell transplantation.
Several studies have shown that HRCT is superior to scans consistent with Aspergillus bronchiolitis or bron-
chest radiography in the assessment of neutropenic chopneumonia and in only two (18%) of 11 patients
patients with suspected pulmonary infection (398,444, who had HRCT findings consistent with angioinvasive
445). It may demonstrate abnormalities in patients with aspergillosis (446).
normal radiographs and show characteristic nodules with
a halo sign in patients with normal or nonspecific radio-
graphic findings (398,444,445). Septic Emboli
Septic emboli are seen most commonly in intravenous
Aspergillus Bronchiolitis and Bronchopneumonia drug abusers and in immunocompromised patients with
(Airway Invasive Aspergillosis) central venous lines. The diagnosis is usually suspected
based on clinical history and radiographic findings of
Aspergillus bronchiolitis and bronchopneumonia, also peripheral nodules with various degrees of cavitation.
known as airway-invasive aspergillosis, are characterized
histologically by necrosis and a neutrophilic infiltrate that
is centered along the small bronchi and bronchioles
(232). Invasion of pulmonary arteries may be present but
is much less conspicuous than with angioinvasive disease.
The characteristic HRCT findings consist of patchy
peribronchial or lobular areas of consolidation, centrilob-
ular nodules, and branching nodular and linear opacities
(tree-in-bud pattern) (Figs. 8-130 and 8-131) (399,436,
446). Histologically, the peribronchial and lobular areas of
consolidation represent bronchopneumonia, and the nod-
ules represent Aspergillus bronchiolitis, with a variable
degree of peribronchiolar organizing pneumonia and
hemorrhage (399,436,446). Although the majority of
patients have both centrilobular nodules and areas of
consolidation, some patients may have predominantly
or exclusively centrilobular nodules, and others, predomi- Figure 8-131 Aspergillus bronchiolitis and bronchopneumonia
(airway-invasive aspergillosis). High-resolution CT at the level of
nantly consolidation (399). the bronchus intermedius shows centrilobular nodular opacities
In patients who have Aspergillus bronchopneumonia, (arrows) in the right lower lobe, left lower lobe, and, to a lesser
BAL is more likely to be positive than in patients who extent, in the lingula. Also noted are focal ground-glass opacities
and small areas of consolidation. The patient was a 38-year-old
have angioinvasive disease. In one study, BAL was positive woman with acute lymphoblastic leukemia and Aspergillus bron-
for fungi in eight of ten (80%) patients who had CT chiolitis and bronchopneumonia.
A B
Figure 8-132 Septic embolism. A: CT image at the level of the lung apices from a 5-mm helical scan shows bilateral cavitating nodules
(straight arrows), focal ground-glass opacities, and right pneumothorax. B: CT image at the level of the right middle lobe bronchus shows
pleural-based area of consolidation (curved arrow) in the right lower lobe, small patchy areas of consolidation in both lungs, and right pneu-
mothorax. The patient was a 41-year-old intravenous drug user with septic embolism.
However, in some patients, the findings may not be readily CLINICAL UTILITY OF COMPUTED
apparent on the radiograph and be well demonstrated on TOMOGRAPHY
CT (123,447,448). The CT and HRCT findings consist of
peripheral nodules, most numerous in the lower lung The clinical value of CT and its role in the evaluation of
zones (123,447). The nodules usually measure between patients with suspected or known diffuse lung diseases can
1 and 3 cm in diameter and involve mainly the subpleural be reviewed by addressing the following questions:
lung regions (Fig. 8-132) (448). The nodules frequently
show varying stages of cavitation, presumably caused by 1. To what extent does CT augment the diagnostic sensitiv-
intermittent seeding of the lungs by infected material. ity and specificity of chest radiography?
Septic emboli may also result in wedge-shaped pleural- 2. What are the advantages and disadvantages of HRCT
based ground-glass opacities or consolidation (Fig. 8-132). over thick-section CT?
These may represent focal areas of pneumonia, hemor- 3. What is the diagnostic accuracy based on the HRCT
rhage, or infarction. After the administration of a bolus of findings as compared with the chest radiograph in dif-
intravenous-contrast medium, the perimeter of an infarct fuse lung disease?
characteristically enhances, presumably because of collat- 4. What is the role for HRCT in the assessment of disease
eral blood flow from adjacent bronchial arteries, whereas activity and prognosis?
the center of the lesion remains lucent. Cystic changes 5. How does CT compare with other modalities in predict-
within an area of infarction may signify either necrosis or ing the accuracy of transbronchial versus open-lung
infection. These findings, however, are not entirely diag- biopsy in the diagnosis of diffuse lung disease?
nostic, because pneumonias may occasionally result in a 6. Can HRCT findings be sufficiently diagnostic to obviate
similar appearance. The specificity of the finding is lung biopsy?
increased when a vessel can be identified at the apex of the
infarct (449). Other common complications of septic Diagnostic Sensitivity and Specificity
emboli include pneumothorax and empyema (Fig. 8-132).
of Computed Tomography Compared
with Chest Radiography
Noninfectious Complications
Chest radiographs are important in the assessment of
A large number of noninfectious complications may be patients suspected of having diffuse lung disease: they are
seen in immunocompromised patients, including pul- inexpensive, readily available, and often can provide infor-
monary edema, drug-induced lung disease, diffuse alveolar mation that is sufficient for diagnosis or management.
hemorrhage, post-transplantation lymphoproliferative dis- Chest radiographs, however, have well-known limitations
order, bronchiolitis obliterans, and recurrence or extension in both sensitivity and specificity for evaluating patients
of an underlying neoplasm. CT has been shown to be with diffuse infiltrative lung diseases (1–3). Ten percent to
particularly helpful in the assessment of patients with 16% of patients with pathologically proven infiltrative
clinically suspected drug-induced lung disease and bron- lung disease have normal chest radiographs (Fig. 8-118)
chiolitis obliterans. (1–3). Furthermore, between 10% and 20% of patients
with radiographic findings suggestive of infiltrative lung patients with documented diffuse lung disease. In each case,
disease have normal lungs (1–3,213). the observers provided their most likely diagnosis as well as
The sensitivity of CT for detecting diffuse lung disease their degrees of confidence. The highest confidence level in
has been compared with that of plain chest radiography diagnosis was reached in 49% of interpretations, based only
in a number of studies. Without exception, these have on the three available HRCT sections, as compared with 31%
shown that CT, and particularly HRCT, are more sensitive of interpretations based on the corresponding three 10-mm
than chest radiography for detecting both acute and sections and 43% of interpretations based on a complete set
chronic diffuse lung diseases. Diseases in which HRCT has of contiguous 10-mm-thick sections through the thorax. The
been shown to be more sensitive than plain radiographs correct diagnoses were made in 92%, 96%, and 94% of
include IPF (62,450,451), sarcoidosis (67,240), HP these readings, respectively. Surprisingly, the confidence in
(14,15), asbestosis (13,219), coalworkers’ pneumoconio- diagnosis and the diagnostic accuracy did not improve sig-
sis (88), silicosis (118), scleroderma (294), rheumatoid nificantly when the findings on thick-section CT were added
lung disease (452), PLCH (167,169), lymphangioleiomy- to those of HRCT.
omatosis (166, 349), emphysema (370), and pulmonary Similarly, Remy-Jardin et al. (454), in an assessment of
complications in immunocompromised patients (394, 150 patients evaluated by both 10-mm CT sections and
395,397). HRCT obtained at identical levels, found that HRCT was
It should be noted, however, that a negative HRCT study clearly superior to thick-section CT in the identification of
does not rule out parenchymal lung disease (3,453). Based nodular and linear interfaces, septal and paraseptal lines,
on the results of several studies, which included assessment small cystic airspaces, bronchiectasis, and pleural thicken-
of chest radiographs and HRCT, it has been estimated that ing (Fig. 8-133). These authors also confirmed that reliable
the sensitivity of HRCT performed at 10-mm intervals for identification of ground-glass attenuation required the use
detecting biopsy-proven CILDs is 94%, as compared with of HRCT. They also demonstrated that acquiring HRCT
80% for chest radiographs (453). In one study, the authors sections at 15-mm intervals was significantly more accu-
compared the sensitivity of CT scans and chest radiographs rate in evaluating the full range of abnormalities when
in 106 patients with pulmonary complications in AIDS and compared with a few selected HRCT sections.
33 patients with no active intrathoracic disease (397). Of Several studies have focused on the potential role of low-
the 106 patients with intrathoracic complications, 96% dose CT scans in the evaluation of patients with diffuse
were correctly identified at CT as compared with 90% on parenchymal lung disease (Fig. 8-1) (7,10,455). HRCT scans
the chest radiograph. performed at 10- and 20-mm intervals result in a skin-
CT also has been shown to have a greater specificity, entrance dose equivalent to 12% and 6%, respectively, of
allowing more-confident distinction of patients with nor- the radiation dose of thick-section CT performed through-
mal lungs from those with infiltrative lung disease. In one out the chest (455). HRCT scans performed at 20-mm inter-
study that included 86 patients with CILD and 14 normal vals by using low-dose technique (40 mAs) result in an aver-
patients, the specificity in identifying normal persons age skin dose comparable to that of routine chest
based on CT findings was 96% compared with 82% on radiography. One group of investigators compared the diag-
chest radiographs (3). Similarly, HRCT has been shown to nostic accuracy of low-dose HRCT (80 mAs, 120 kVp),
have a greater specificity in excluding parenchymal disease conventional-dose HRCT (300 mAs, 120 kVp), and con-
in patients with AIDS (397). ventional chest radiographs in 10 normal controls and
Based on the greater sensitivity and specificity of CT as 50 patients with CILD (7). For each HRCT technique, only
compared with the chest radiograph, CT is indicated in the three images were used, obtained at the levels of the aortic
assessment of patients with normal or questionable radi- arch, tracheal carina, and 1 cm above the right hemidi-
ographic findings who have symptoms or pulmonary aphragm. A correct first-choice diagnosis was made signifi-
function findings suggestive of diffuse lung disease. cantly more often with either HRCT technique than with
radiography. A high confidence level in making a diagnosis
was reached in 42% of radiographic examinations, 61%
High-Resolution Versus Thick-Section
of low-dose, and 63% of conventional-dose HRCT exam-
Computed Tomography Imaging inations (p 0.05), and these were correct in 92%, 90%,
Several studies have demonstrated that HRCT is superior to and 96% of studies, respectively. Although conventional-
CT performed by using 5- to 10-mm-thick sections in the as- dose HRCT was slightly more accurate than low-dose HRCT,
sessment of diffuse lung disease (13,17,88,454). Aberle et al. this difference was not statistically significant. The limited
(13) demonstrated that HRCT performed in both the supine low-dose, thin-section CT used in this study was signifi-
and prone positions was able to identify subtle parenchymal cantly superior to chest radiography and resulted in an
reticulation in 96% of patients with asbestosis compared equivalent effective radiation dose (0.03 mSv). It should be
with only 83% of cases on thick-section CT. Leung et al. (17) noted, however, that mild ground-glass attenuation and
randomly compared three selected HRCT sections with emphysema detectable on conventional-dose HRCT may be
thick-section CT scans obtained at the same three levels and missed on low-dose HRCT (Fig. 8-2) (10). The optimal dose
with complete thick-section CT studies of the lungs, in 75 for HRCT has yet to be established.
A B
Figure 8-133 High-resolution CT (HRCT) versus conventional CT. A: A 7-mm collimation spiral CT at the level of the right upper lobe
bronchus demonstrates no definite abnormality. B: HRCT (1-mm collimation) at the same level demonstrates irregular interfaces (straight ar-
rows) and localized areas of ground-glass attenuation mainly in the subpleural lung regions (curved arrow). Surgical lung biopsy demon-
strated idiopathic pulmonary fibrosis with mild fibrosis and active alveolitis.
HRCT (1- to 2-mm collimation) is recommended in the agreement, both in CILDs (3,5–7,10,454) and in acute
assessment of chronic and acute infiltrative lung diseases lung diseases in immunocompromised AIDS (397) and
because it is superior to thick-section CT. As previously non-AIDS patients (Figs. 8-119 and 8-134) (396). The
discussed, two major alternative approaches can be used: majority of these studies have used an approach empha-
HRCT performed at preselected levels, preferably 1-mm- sizing blinded interpretations of a series of radiographic
thick sections at 10-mm intervals, and volumetric HRCT and CT findings, in which individual observers listed their
through the chest. The main advantage of HRCT at prese- first three diagnostic choices in decreasing order of proba-
lected levels is to reduce radiation exposure (16). The main bility, with degrees of confidence in those diagnoses usu-
disadvantages are that focal parenchymal abnormalities ally expressed on a 3-point scale. In the study by
may be overlooked, the patchy distribution of many diffuse Mathieson et al. (6), the accuracies of chest radiographs
infiltrative lung diseases may not be appreciated, and small and CT in making specific diagnoses were compared in
nodules may be missed between high-resolution sections 118 patients with various diffuse infiltrative lung diseases.
(13–17). The second approach, possible since the advent of A confident diagnosis was made more than twice as often
multidetector CT scanners, is to perform volumetric HRCT on CT than on chest radiographs (49% vs. 23%, respec-
through the entire chest (18,19). This approach allows tively), and this diagnosis was more often correct on CT
assessment of the entire lung parenchyma and therefore (93%) than on the radiograph (77%) (p 0.001). An
minimizes the risk of a falsely negative examination. In approximately twofold improvement in diagnostic accu-
addition, volumetric HRCT allows use of advanced imaging racy of HRCT (53% of cases) compared with chest radio-
techniques such as multiplanar reconstructions (MPRs) graphs (27%) was also reported by Grenier et al. (5), in
and MIPs or MinIPs. As a consequence, despite entailing a study of 140 patients with 18 different infiltrative lung
some increase in radiation exposure, potential advantages diseases. Slightly smaller improvement in the diagnostic
in our judgment make this technique the preferred method accuracy of CT compared with chest radiography was
of initial examination in all patients with suspected diffuse reported by others (3,456).
lung disease. In distinction, unless otherwise indicated, In an attempt to refine the diagnostic accuracy further,
follow-up CT studies are best performed with a dedicated Grenier et al. (119) used bayesian analysis to determine
HRCT technique by using low-dose technique whenever the relative value of clinical data, chest radiographs, and
feasible. HRCT in patients with chronic diffuse infiltrative lung
disease. For this study, two samples from the same popula-
tion of patients with 27 different diffuse lung diseases were
Diagnostic Accuracy of Computed consecutively assessed, an initial training set of 208 cases
Tomography Compared with Chest for the development of the decision aid and a subsequent
prospectively evaluated set of 100 “test” cases. Of the 208
Radiography
initial cases, a correct diagnosis based on clinical data
Several studies have documented that CT is superior to the alone was obtained in 29% of cases, radiography alone in
plain radiograph in allowing correct diagnosis of specific 9%, and HRCT in 36% of cases. This increased to 54%
lung diseases, with considerably better interobserver when clinical and radiographic findings were combined
and to 80% when all three were analyzed together (p probably reflected a population with advanced disease, as
0.01). With prior and conditional probabilities determi- virtually all of these patients had diffuse radiographic
ned from the initial set, the frequency of a high- abnormalities. Not surprisingly, the value of HRCT dimin-
confidence-level correct diagnosis based on bayesian ished with less common diseases: in this regard, it is
analysis in the 100 test cases from clinical, radiographic, significant that 23 (68%) of 34 misclassified patients in
and HRCT data alone was 27%, 4%, and 49%, respectively. this study had diseases classified as “miscellaneous.”
The combination of the three tests allowed a high degree Kang et al. (397) compared the diagnostic accuracy of
of confident diagnosis in 61% of cases. HRCT made the CT with that of chest radiography in the detection and
greatest contribution to the diagnosis of sarcoidosis, diagnosis of thoracic complications in patients with AIDS.
Langerhans cell histiocytosis, HP, lymphangitic carcino- CT was superior to chest radiography in the exclusion of
matosis, and silicosis. Although only minor improvement disease in normal controls, in the detection of parenchy-
was seen in the diagnosis of IPF, as the authors noted, this mal abnormalities, and in the differential diagnosis. Of
A B
Figure 8-134 “Dirty lung” emphysema. A, B:

Posteroanterior and lateral radiographs show
diffuse, ill-defined reticular opacities throughout
both lungs. C, D: Retrospectively targeted high-
resolution CT images of the right mid and lower
lung, respectively, clearly reveal extensive emphy-
sema without evidence of diffuse interstitial disease.
The reticular patterns seen on the chest radiographs
are caused by the septa-separating bullae, most
easily identified laterally and posteriorly (arrows). In
addition to peripheral bullae, discrete areas of
markedly low tissue attenuation without clearly
definable walls can be identified within the lung
parenchyma, findings compatible with diffuse
emphysematous disease. Note that the intervening
lung parenchyma is normal and that intrapulmonary
C, D vessels are well defined with smooth contours.
89 patients with one proven pulmonary complication in radiographic findings are nonspecific, and further evalua-
AIDS, the observers were confident in their first-choice tion is considered appropriate.
diagnosis in 34% of cases at chest radiography and in 47%
at CT. This diagnosis was correct in 67% of confident
Assessment of Disease Activity: Significance
radiographic interpretations as compared with 87% of
interpretations at CT (p 0.01). The confident first-choice
of Ground-Glass Opacities
diagnosis at CT was most often correct in Kaposi sarcoma, Although HRCT can play an important role in the evalua-
in Pneumocystis jiroveci pneumonia, and in the AIDS- tion of disease activity in patients with diffuse lung disease,
related lymphoma. Gruden et al. (416) found that HRCT assessment of the true clinical value of HRCT must be indi-
had a sensitivity of 100% and a specificity of 89% in the vidualized by disease entity. Needless to say, a potential
diagnosis of PCP in 51 patients with AIDS and a high clin- role for HRCT in managing therapy is important because
ical pretest probability of PCP and normal, equivocal, or determining when and whether to treat patients, particu-
nonspecific chest radiographs. larly with corticosteroids, often proves to be a frustrating
Although all studies comparing CT with chest radiogra- clinical dilemma.
phy have demonstrated that CT is superior, some variation HRCT findings in CILDs that have been shown often to
has occurred in the reported accuracy of CT. This may be be reversible include ground-glass opacities, parenchymal
related to a different patient population or to a different consolidation, and small nodules. Conversely, reticulation
CT technique. The studies reporting the highest diagnostic and honeycombing typically are associated with irreversible
accuracy of CT as compared with chest radiography have fibrosis and poor prognosis.
used either HRCT combined with thick-section CT or, Leung et al. (131), in a study of 22 patients with a variety
exclusively, 1- to 2-mm collimation HRCT sections of CILDs in which ground-glass opacities were the pre-
(5–7,119). The studies reporting lower diagnostic accuracy dominant or exclusive CT finding, showed that 18 (82%)
of CT used 3-mm-thick (3) or 5-mm-thick sections (456). had active potentially reversible lung disease on lung
It cannot be overemphasized that accurate assessment of biopsy (Fig. 8-135) (131). Similarly, Remy-Jardin et al.
diffuse lung diseases requires the use of HRCT (1- to 2-mm (145), in a study of 26 patients with various diffuse infiltra-
collimation scans). tive lung diseases, demonstrated that ground-glass opaci-
The various studies have shown that HRCT allows a ties corresponded to inflammation in 24 (65%) of 37
greater diagnostic accuracy than chest radiography and biopsy sites. As emphasized by these authors, ground-glass
thick-section CT in the CILDs and in the acute infiltrative opacities are most valuable when seen in the absence of
lung diseases seen in immunocompromised patients. other findings indicative of fibrosis. In this same study, in
Therefore HRCT is indicated to make a specific diagnosis, 11 (85%) of the 13 cases in which ground-glass opacities
or to limit the differential diagnosis, in patients with represented fibrosis rather than active disease, they were
abnormal chest radiographs, in whom the clinical and associated with traction bronchiectasis or bronchiolectasis,
A B
Figure 8-135 Reversible ground-glass opacities in nonspecific interstitial pneumonia (NSIP). A: High-resolution CT at the level of the
bronchus intermedius shows extensive bilateral ground-glass opacities and mild peripheral reticulation. The patient was dyspneic and had
severe restrictive lung function. B: High-resolution CT 16 months later, after treatment with corticosteroids, shows considerable reduction in
the extent of ground-glass opacities. Subpleural reticulation persists. The patient, a 71-year-old man with NSIP, had improved clinically, was
no longer dyspneic at rest, and had less severe restrictive lung function.
markers of interstitial fibrosis. It cannot be overemphasized score (p 0.03) were able to predict mortality. A pretreat-
that ground-glass opacities can be considered a reliable ment CT-fibrosis score of 2.0 or more demonstrated 80%
indicator of active inflammation only when associated sensitivity and 85% specificity in predicting survival.
findings of fibrosis are not present in the same areas Those patients who did not respond to initial steroid ther-
(Figs. 8-30 and 8-87) (145). apy demonstrated a worse long-term survival and greater
Wells et al. (133) assessed the predictive value of HRCT likelihood of decreased pulmonary function (195).
in 76 patients with idiopathic interstitial fibrosis. They Zisman et al. (457) prospectively assessed IPF patients
found that response to therapy in previously untreated receiving treatment with cyclophosphamide in whom
patients was significantly greater in patients who had pre- prior steroid therapy had proved unsuccessful. In total,
dominantly ground-glass opacity. Four-year survival was 19 patients were evaluated by using both CRP and HRCT
highest in patients who had predominantly ground-glass scores, including one responder, seven patients who
opacity, and higher in patients who had equivalent remained stable, and 11 who deteriorated. Of a total of
ground-glass and reticular opacities than in those who had 11 parameters evaluated, including age, CRP score, year of
predominantly reticular abnormalities, independent of the onset, and a variety of pulmonary function tests, only
duration of symptoms or severity of pulmonary function HRCT-fibrosis score and percentage of lymphocytes on
abnormalities (p 0.001). It should be noted however, BAL proved to be predictors of response to cyclophos-
that only 8 (11%) of the 76 patients had a predominant phamide therapy. Similarly, the HRCT-fibrosis score was
pattern of ground-glass opacity, and 18 (24%) had an one of the few parameters to be a significant predictor of
equivalent distribution of areas of ground-glass opacity survival (p 0.03), along with BAL lymphocytosis, level
and reticulation; the majority of patients had predominant of dyspnea, and CRP score. Based on these data, these
reticulation and did not respond to corticosteroid treat- authors concluded not only that cyclophosphamide ther-
ment. Furthermore, in retrospect, this study published in apy was of limited value for the treatment of IPF but also
1993 probably included patients who had DIP and NSIP. that in patients who had high HRCT-fibrosis scores, the
It is likely that the majority of patients who had predomi- prognosis was sufficiently poor to warrant withholding
nant ground-glass opacity had DIP or NSIP. therapy (457).
The prognostic significance of areas of ground-glass Several studies have demonstrated that NSIP has a
opacity on HRCT in patients with IPF compared with better prognosis than IPF and that patients with NSIP and
those with DIP was assessed in a study of 23 patients predominantly ground-glass opacities have a better prog-
(12 with IPF and 11 with DIP) by Hartman et al. (196). nosis than do patients with predominantly reticulation
Eleven patients with IPF and 11 with DIP received treat- (97,458,459). Kim et al. (460) evaluated 13 patients with
ment with corticosteroids. In 9 (75%) of the 12 patients biopsy-proven NSIP with serial high-resolution CT and
with IPF, the ground-glass opacities increased in extent or pulmonary function tests. In all patients, the initial high-
progressed to reticulation or honeycombing on follow-up resolution CT showed areas of ground-glass opacity and,
HRCT as compared with only 2 (18%) of 11 patients with to a lesser extent, reticulation. On the follow-up high-
DIP. Thus although ground-glass opacity may reflect the resolution CT, a significant reduction in the extent
presence of active inflammation, in most patients with IPF, of ground-glass attenuation was found (p 0.005).
it progresses to fibrosis despite treatment (196). An improvement also was seen in forced vital capacity
Gay et al. (195) assessed HRCT and histologic factors (p 0.003) on follow-up that correlated with the extent
predictive of response to therapy in 38 patients who had of reduction in ground-glass attenuation on CT. Similar
biopsy–proven IPF. In each case, a pretreatment clinical, results were described by Arakawa et al. (297) in a study
radiographic, and physiologic score (CRP) was generated. of 14 patients with NSIP associated with polymyositis or
The HRCT scans were independently scored by four radi- dermatomyositis. Screaton et al. (459) assessed the prog-
ologists for ground-glass (CT-alveolitis score) and linear nostic value of HRCT in 38 patients with biopsy-proven
opacity (CT-fibrosis score) on a scale of 0 to 4. Biopsy NSIP. The predominant pattern of abnormality on the
samples were scored for inflammation and fibrosis. initial HRCT scan was reticulation and honeycombing in
All patients were treated with 3 months of high-dose 32 (84%) cases and ground-glass in 6 (16%) cases. On
corticosteroids, and the CRP scoring was repeated. follow-up CT, the disease extent reduced by more than
Patients were divided into three groups: responders with 10% in 13 (34%) and increased by more than 10% in
a greater than 10-point decrease in CRP (n 10); stable 6 (16%) patients. Patients with predominant ground-glass
with 10-point change in CRP (n 14); and nonrespon- opacities on the initial HRCT were more likely to respond
ders with a more than 10-point increase in CRP or death to treatment than were patients with predominant reticu-
(n 14). Those responding to steroids were treated for lation/honeycombing (459). It should be noted, however,
18 months in a tapering fashion. The CT-alveolitis scores that an increase in the extent of ground-glass opacities
were higher and CT-fibrosis scores were lower in re- after 6 months of treatment with corticosteroids and other
sponders than in nonresponders. However, only the immunosuppressants is associated with a poor prognosis.
CT-fibrosis score (p 0.009) and pathology fibrosis Flaherty et al. (458), in a study of 80 patients with UIP
and 29 patients with NSIP, showed a more extensive granulomata. These account for the predominant finding
fibrotic pattern in UIP than in NSIP on the initial HRCT. on HRCT that consists of nodules in a characteristic peri-
After a 6-month follow-up, a progression in the extent of lymphatic distribution. Even when areas of ground-glass
ground-glass opacity by more than 10% was seen in a attenuation are present on HRCT, these have been shown
higher proportion of patients with UIP than with NSIP. to represent volume averaging of interstitial granulomata
Additionally, an increase in ground-glass opacity of below the resolution of HRCT (67,131,242). In patients
greater than 10% was associated with an increase in the with sarcoidosis, nodules, areas of ground-glass attenua-
risk of subsequent death (relative risk, 2.88). tion, and parenchymal consolidation have been shown to
Sequential CT findings have been assessed in 27 patients be reversible in most patients after treatment with corti-
with HP (bird breeders’ lung) (15). Patients in whom the costeroids (Fig. 8-136) (120,240). Reticulation, architec-
main CT abnormalities consisted of diffuse micronodules, tural distortion, cystic spaces, and traction bronchiectasis
ground-glass attenuation, and focal air trapping in the are irreversible (120,240). It should be noted, however,
absence of diffuse reticulation or honeycombing uniformly that although certain findings on HRCT may indicate
improved after cessation of exposure to the inciting antigen, potentially reversible disease, it is controversial whether
whereas those with a predominantly fibrotic pattern HRCT is helpful in predicting which patients with
showed little if any change. Similar to patients with IPF, the sarcoidosis are most likely to respond to corticosteroid
prognostic significance of ground-glass attenuation in itself therapy (461). Further studies are required to determine
was less significant than the extent of underlying fibrosis. the role of the HRCT in the assessment of disease activity
Although alveolitis has been postulated to represent and prognosis in patients with sarcoidosis.
the initial stage in the development of sarcoid granulo- The histopathologic marker of disease activity in PLCH
mata, it is seldom present at the time when the patients is the presence of Langerhans cell granulomas in the lung
are evaluated clinically (242). The predominant histologic (462). Strong correlation has been demonstrated between
finding at the time of diagnosis consists of noncaseating the presence and extent of nodular lesions visible on HRCT
Figure 8-136 Ground-glass reversibility in sarcoidosis.

A: High-resolution CT (HRCT) in a 32-year-old man with sar-
coidosis demonstrates extensive bilateral areas of ground-glass
attenuation and interlobular septal thickening (straight arrows).
Small nodules are present, particularly in the left lower lobe.
Also noted is right peribronchial fibrosis with traction bronchiec-
tasis and architectural distortion (curved arrows). B: HRCT
2 years later demonstrates marked improvement. Note almost
complete resolution of ground-glass attenuation and inter-
lobular septal thickening, as well as decreased numbers of small
B nodules. However, the peribronchial fibrosis persists.
and the density of florid granulomatous lesions in lung High-Resolution Computed

tissue, in keeping with an active disease process (337). Tomography–Guided Lung Biopsy
However, thin-walled cysts in pulmonary Langerhans
histiocytosis may reflect the presence of irreversible fibrosis One of the most important indications for the use of
or represent cavitary but still inflammatory granulomas. HRCT is as a potential guide for the site and type of lung
Therefore a cystic-scan pattern, the persistence of an biopsy. Many diffuse lung diseases are patchy in distribu-
active, albeit limited granulomatous process (that may tion, with areas of abnormal lung being interspersed
result in progressive respiratory insufficiency), cannot be among relatively normal areas of lung parenchyma.
excluded (337). Furthermore, both active and fibrotic disease can be pres-
In patients with AIDS, bilateral areas of ground-glass ent in the same lung and indeed in the same pulmonary
attenuation are virtually diagnostic of PCP (84,416,463). segment (234,464). Because of its ability to visualize, char-
Less commonly, they may be caused by cytomegalovirus acterize, and determine the distribution of parenchymal
pneumonia, pyogenic pneumonia, or be seen around disease, HRCT also provides a unique insight into the
hemorrhagic nodules in Kaposi sarcoma (84,400). likely efficacy of TBBx or surgical lung biopsy (via thoraco-
HRCT may also be helpful in assessing disease activity in tomy or video-assisted thoracoscopy) in patients with
mycobacterial infections, in both immunocompromised either acute or chronic diffuse lung disease and the opti-
and nonimmunocompromised patients. Im et al. (112) mal biopsy site (Figs. 8-137 and 8-138). In recent years,
assessed sequential CT scans before and after antitubercu- video-assisted thoracoscopic lung biopsy (VAT) has
lous therapy in 26 patients with active tuberculosis. The replaced thoracotomy as the method of choice for lung
most common findings at presentation consisted of biopsy in diffuse lung disease because of equivalent diag-
centrilobular nodules and linear branching structures corre- nostic accuracy, lower morbidity, and lower cost (465).
sponding pathologically to the presence of mucus-filled Because the operator’s field of view is rather limited when
small airways. In virtually all cases, sequential studies performing this procedure, HRCT has proven extremely
showed these opacities to be reversible within 5 months helpful in directing the surgeon to the most appropriate
after the start of treatment. Furthermore, in 11 of 12 patients biopsy site (Fig. 8-138).
with recent reactivation, CT accurately differentiated old Although TBBx is frequently used in an attempt to
fibrotic lesions from new active ones. The authors con- diagnose diffuse lung disease, the limitations of TBBx
cluded that HRCT was a reliable method for determining
disease activity in patients with tuberculosis.
Figure 8-138 High-resolution CT (HRCT) as a guide to video-

assisted thoracoscopic biopsy. HRCT scan in a 58-year-old man
with progressive shortness of breath demonstrates a reticular
pattern and areas of ground-glass attenuation involving mainly the
Figure 8-137 High-resolution CT (HRCT) as a guide to trans- subpleural regions of the right lower lobe, left lower lobe, and
bronchial biopsy. HRCT in a 52-year-old woman demonstrates lingula. Also noted is honeycombing in these areas. Small focal
nodular thickening of the bronchovascular bundles and of the areas of ground-glass attenuation are present in the right middle
interlobar fissures. The findings are most severe in the upper lobe. Biopsy of the right middle lobe in this patient confirmed the
lobes, particularly the left upper lobe, with minimal abnormalities diagnosis of idiopathic pulmonary fibrosis but gave no indication
seen in the lower lobes. The peribronchovascular distribution of of the extensive honeycombing present in this patient. Biopsy of
the abnormalities suggested that transbronchial biopsy would be the lower lobes, conversely, may have revealed only honeycomb-
likely to yield a definitive diagnosis. Given the distribution of ing but may not have yielded a specific diagnosis. In patients with
abnormalities, a biopsy of the left upper lobe was considered chronic infiltrative lung disease, open or video-assisted thoraco-
more likely to be diagnostic than a biopsy in either lower lobe. scopic biopsy should be performed in areas with relatively mild
Transbronchial biopsy of the left upper lobe was diagnostic of involvement. Areas with honeycombing should be avoided, as
sarcoidosis. they may only yield nonspecific end-stage fibrosis.
for establishing the etiology of diffuse pulmonary disease patients with IPF, owing to the predominantly subpleural
have been well documented. In a classic study, Wall et al. distribution of the fibrosis.
(466) showed that TBBx was diagnostic in only 20 (38%) Given the limitations of both transbronchial and sur-
of 53 patients with radiographic evidence of diffuse lung gical biopsy techniques, it is not surprising that HRCT
disease. In the remaining 33 cases, TTBx were reported has emerged as an important tool for assessing patients
either as normal or nonspecific, whereas surgical lung with suspected DILD before lung biopsy. HRCT is of con-
biopsy resulted in specific diagnoses in 92% of cases. siderable value in determining the most appropriate sites
Similar results have been reported by Wilson et al. (467) for biopsy by showing the regions most likely to be active
in a study of 127 patients with a variety of parenchymal and therefore most likely to be diagnostic (138,403,464).
abnormalities. They found that TBBx allowed a “specific” By using HRCT, areas of end-stage honeycombing can be
diagnosis in only 52% of patients with diffuse infiltrative avoided. Second, HRCT can play a decisive role in select-
processes. Also, diagnoses suggested by TTBx may bear ing among TBBx, lavage, and/or surgical biopsy as the
little relation to diagnoses subsequently made on surgical most efficacious method for obtaining a histologic diag-
biopsy (466), and a nonspecific TBBx diagnosis, such as nosis. Of particular value is the identification of peri-
“interstitial pneumonia” or “interstitial fibrosis,” should bronchial abnormalities, as characteristically occur in
be considered potentially misleading (2,468). patients with sarcoidosis (67,240) and lymphangitic
In patients with chronic diffuse lung disease, TTBx is carcinomatosis (64,65). As shown by Lenique et al.
most accurate in patients with sarcoidosis or lymphangitic (469), the demonstration of abnormal airways in
carcinomatosis (466); these entities preferentially involve patients with sarcoidosis correlates with the likelihood of
peribronchial tissues and therefore are most accessible to obtaining a histologic diagnosis.
TBBx (466,469). In this regard, however, it should be noted Not surprisingly, HRCT has proved more efficacious
that despite the efficacy of TTBx, especially in patients with than chest radiographs in predicting the likely efficacy of
sarcoidosis, a significantly greater diagnostic yield may TBBx for diagnosing DILD. Mathieson et al. (6) compared
result when TBBx is coupled with transbronchial needle the accuracy of plain radiographs and CT in determining
aspiration (TBNA) (Fig. 8-72). As documented by Morales whether TTBx or surgical biopsy would be most appropri-
et al. (470), in their study of 51 consecutive patients with ate in patients with CILD. By using CT, three observers
combined TBBx and TBNA, overall in 33% of cases, a histo- correctly predicted that a TTBx would be likely to be diag-
logic diagnosis of sarcoidosis was established only with nostic in 87% of patients in which this was appropriate.
TBNA. Interestingly, in patients with stage 1 disease, TBBx They correctly predicted the need for surgical biopsy in
was diagnostic in 60%, and TBNA was diagnostic in 53%; 99% of cases. By comparison, plain radiographs proved
the combined yield equaled 83%. Surprisingly, even in significantly less valuable (p 0.001).
patients with stage 2 sarcoidosis, TBBx was diagnostic in HRCT has also been shown to play an important
only 76% of cases. These data strongly suggest a role for CT potential role in the prebronchoscopic assessment of
before bronchoscopy to optimize selection of sites for immunocompromised patients. This has been documented
biopsy, both in the lung and in the mediastinum. by Janzen et al. (403) in a retrospective study evaluating,
Although the accuracy of TBBx has improved over the with both HRCT and bronchoscopy, 33 consecutive
past decade, especially in establishing such diagnoses immunocompromised non-AIDS patients (including 20
as PLCH, PAP, eosinophilic lung disease, Goodpasture hematopoietic stem cell transplant patients) with acute
syndrome, and Wegener granulomatosis, these entities pulmonary disease. Bronchoscopy provided a specific
represent a distinct minority of cases (189). More impor- diagnosis in 17 (52%) of 33 patients. Significantly, bron-
tant, little improvement has been seen in the ability choscopy proved diagnostic more often in patients with
of TBBx or BAL to assess patients with pulmonary fibro- HRCT disease involving the central versus the peripheral
sis (468). third of the lungs (70% vs. 23%; p 0.02). Results also
Surgical lung biopsy, whether by thoracotomy or VAT, proved more diagnostic in patients with infectious versus
is often diagnostic, with accuracies greater than 90% gen- noninfectious disease (71% vs. 17%; p 0.005). Based on
erally reported (2,465,466), but this procedure is also these findings, the authors concluded that HRCT should
subject to sampling error, as biopsies taken from a small precede bronchoscopy in immunocompromised patients
region of lung may not reflect the state of the diseased to determine optimal sites for biopsy as well as to predict
lung as a whole. This is most important when attempts likely results of bronchoscopy.
are made to assess disease activity in patients with diffuse Similar studies have also concluded a role for HRCT in
fibrotic lung diseases, such as IPF or collagen-vascular evaluating hematopoietic stem cell transplant patients
disease (468). It has been emphasized that the role of (394,398). Mori et al. (398), in a study of 33 febrile
the surgeon at the time of biopsy is to obtain representa- hematopoietic stem cell recipients, found that CT showed
tive tissue, while avoiding areas of extensive honeycomb- nodules in 20 of 21 episodes of documented fungal infec-
ing (2). However, this may be difficult, especially in tion but none in 9 bacteremic episodes. From these data,
the authors concluded that CT studies showing the pres- results of two prospective studies. The first was by Raghu et
ence of nodules in this population could be taken as pre- al. (353), who prospectively evaluated 59 consecutive
sumptive evidence of fungal infection, warranting empiric patients with new-onset chronic interstitial lung disease
therapy without the need for bronchoscopy. In a similar who had final diagnosis proven by surgical biopsy. The sen-
prospective study of 36 symptomatic episodes in 33 sitivity and specificity of the HRCT diagnosis of IPF without
hematopoietic stem cell transplant patients, Barloon et al. any clinical information were 78.5% and 90%, respectively.
(394) reported that in comparison with plain radiographs, By comparison, the sensitivity and specificity of the diagno-
HRCT resulted in a change of management in a total of 11 sis of IPF on clinical grounds, which included review of the
(50%) of 22 patients, including establishing the need for HRCT findings by the pneumonologist, were 62% and 97%,
bronchoscopy and/or surgical lung biopsy in 6 patients. respectively. The second study was by Hunninghake et al.
From the data in the literature, it can be concluded that (96), who prospectively evaluated 91 patients who had sus-
CT is indicated as a guide for the need of lung biopsy and pected IPF. In this study, a confident diagnosis of IPF on
as a guide for the optimal type and site of biopsy. HRCT had a specificity of 95% and a positive predictive
value of 96%. The positive predictive value of a confident
clinical diagnosis by expert pulmonologists was 87% (96).
High-Resolution Computed Tomography
The confident diagnosis on HRCT in the study by
Findings Sufficiently Diagnostic to Obviate Hunninghake et al. (96) was based on the presence of retic-
Lung Biopsy ulation and honeycombing involving predominantly the
In our judgment, a number of diseases have appearances peripheral lung regions and lower lung zones in the absence
sufficiently characteristic in cross section to warrant empiric of small nodules, extensive ground-glass opacities, and
therapy in the absence of tissue confirmation in the appro- pleural abnormalities. On multivariate analysis of specific
priate clinical setting (Table 8-22) (471). HRCT features recorded by the four experienced chest
HRCT findings have been shown to be highly accurate in radiologists who participated in the study, the only independ-
making a diagnosis of IPF (93,96,451). Based on the data in ent predictors of IPF were lower-lung honeycombing (odds
the literature, the joint statement of the American Thoracic ratio, 5.4) and upper-lung irregular lines (odds ratio, 6.3)
Society/European Respiratory Society concluded that in (472). It should be noted, however, that a confident
more than 50% of cases suspected to be IPF, the presence of diagnosis of IPF based on the HRCT findings is possible only
typical clinical and HRCT features of IPF, when identified by in 50% to 60% of cases. In the remaining patients, the
expert clinicians and radiologists, is sufficiently characteris- HRCT findings are not characteristic enough to make a
tic to allow a confident diagnosis and to obviate surgical confident diagnosis, and surgical biopsy is required for
lung biopsy (97). This conclusion was based mainly on the a definitive diagnosis.
TABLE 8-22
SPECIFIC ETIOLOGIES POTENTIALLY ESTABLISHED BY HIGH-RESOLUTION COMPUTED TOMOGRAPHY
Diagnosis CT Findings
Idiopathic pulmonary fibrosis Reticular pattern and/or honeycombing involving the lung bases and periphery
Subacute hypersensitivity pneumonitis Diffuse ground-glass centrilobular nodules in the absence of reticulation or a tree-in-bud
appearance; exposure history
Silicosis Well-defined (often calcified) nodules predominantly in the posterior portions of both upper
lobes, usually in association with pericicatricial emphysema; exposure history
Asbestosis Reticulation and/or honeycombing involving the lung bases and associated with pleural
plaques or diffuse thickening; exposure history
Sarcoidosis Perilymphatic nodules associated with symmetric hilar and mediastinal lymphadenopathy
Lymphangitic carcinomatosis Beaded interlobular septal thickening in a patient with a known primary tumor
Lymphangioleiomyomatosis Thin-walled cystic spaces uniformly distributed throughout both lungs in the absence of
reticular changes or nodules; woman of childbearing age
Pulmonary Langerhans cell histiocytosis Bilateral bizarre-shaped cysts with or without associated centrilobular nodules with relative
sparing of the lower third of the lungs; smoker or ex-smoker
Pneumocystis jiroveci pneumonia Symmetric bilateral ground-glass opacities with or without associated cystic changes in
HIV patient
HIV, human immunodeficiency virus.

HRCT findings of diffuse centrilobular ground-glass two independent radiologists made a confident diagnosis
opacities are suggestive of subacute HP and are often the of PLCH in 84% of cases, LAM in 79%, and emphysema in
first clue to the diagnosis. The diagnosis can often be 95%. When confident, the observers were correct in 100%
confirmed by the clinical history and presence of positive of the cases (340). Koyama et al. (341) reviewed the HRCT
precipitating antibodies to the offending antigen, preclud- findings in 92 patients with chronic cystic lung diseases.
ing the need for lung biopsy (268,271,473). Lacasse et al. Two independent radiologists made a diagnosis with a
(271) assessed various diagnostic criteria in 400 consecu- high degree of confidence in 105 (57%) of 184
tive patients in whom HP was considered in the differential interpretations. The confident diagnosis was correct in
diagnosis. They identified six significant clinical predictors 98 (93%) of 105 interpretations. The confident diagnosis
of HP: (a) exposure to a known offending antigen, (b) pos- (average of two observers) was correct in all cases (100%)
itive precipitating antibodies to the offending antigen, of IPF, 100% of cases of emphysema, 88% of cases of LAM,
(c) recurrent episodes of symptoms, (d) inspiratory crackles and 88% of cases of PLCH. In both these studies, the
on physical examination, (e) symptoms occurring 4 to diagnosis was based exclusively on the HRCT findings. It
8 hours after exposure, and (f) weight loss. In this study, cannot be overemphasized that a confident definitive diag-
HRCT and BAL defined the presence or absence of HP. nosis can be made without biopsy only if both the HRCT
Patients underwent surgical lung biopsy when the HRCT, and clinical findings are highly suggestive of a specific
BAL, and other diagnostic procedures failed to yield a diagnosis.
diagnosis. They demonstrated that the diagnosis of HP can More problematic is the question of whether biopsies
often be made or rejected with confidence, based on clini- should be obtained in patients with otherwise unexplained
cal and HRCT findings without BAL or biopsy (271). respiratory symptoms with normal HRCT studies. Padley et
Likewise, the presence of characteristic findings on al. (3), in a study of 100 patients with documented DILD,
HRCT allows confident diagnosis of pneumoconiosis in noted that in 18% of patients, HRCT studies were inter-
patients with a well-documented history of occupational preted as normal (3). Similarly, in a prospective study of
exposure, particularly in workers exposed to coal dust, 25 patients with biopsy-proven IPF, Orens et al. (475)
silica, or asbestos. reported finding 3 (12%) patients with normal HRCT exam-
In our opinion, the finding of clusters of ill-defined inations. However, all of the patients included in the study
peribronchovascular and subpleural nodules, especially by Orens et al. were also part of the prospective study by
when bilateral, predominantly upper lobe, and associated Hunninghake et al. (96). None of the cases was interpreted
with central airways abnormalities, in the appropriate clini- as normal by expert CT readers who reviewed the HRCT
cal setting, should allow a definitive diagnosis of sarcoidosis findings independently and were blinded to any clinical
to be made without the necessity for histologic confirma- information. In one multicenter study, the authors per-
tion (471,474). Although a number of other diseases, most formed HRCT at 10-mm intervals in 661 consecutive
notably lymphangitic carcinomatosis, may result in a peri- patients with a condition for which HP was considered in
lymphatic distribution of disease typical of sarcoidosis, the differential diagnosis. In total, 199 of these patients
differentiation is usually easily made, even in the absence of were proven to have HP. Of the 199 patients with HP, 171
clinical correlation. For example, in a study comparing CT (86%) had an abnormal chest radiograph, and 183 (92%)
scans in 40 patients who had lymphangitic carcinomatosis had ground-glass opacities and/or centrilobular nodules ev-
with 41 patients who had sarcoidosis, Honda et al. (103) ident on HRCT (271). From these data, it can be concluded
showed that findings of thickened interlobular septa and that a normal HRCT performed at 10-mm intervals does not
extensive involvement of the subpleural interstitium were definitely exclude the possibility of parenchymal disease.
significantly more common in patients who had lymphan-
gitic carcinomatosis (p 0.0001), whereas bilateral distri-
bution of disease was significantly more common in REFERENCES
patients who had sarcoidosis (p 0.001).
HRCT can also prove diagnostic in patients who have 1. Epler GR, McLoud TC, Gaensler EA, et al. Normal chest
roentgenograms in chronic diffuse infiltrative lung disease.
LAM and PLCH, provided the HRCT shows characteristic N Engl J Med. 1978;298:934–939.
abnormalities in the appropriate clinical setting (5,340, 2. Gaensler EA, Carrington CB. Open biopsy for chronic diffuse infil-
341). LAM is typically seen in women of childbearing age trative lung disease: clinical, roentgenographic, and physiological
correlations in 502 patients. Ann Thorac Surg. 1980;30:411–426.
and is characterized by the presence of small, thin-walled 3. Padley SPG, Hansell DM, Flower CDR, et al. Comparative
cysts scattered throughout both lungs. PLCH is seen accuracy of high resolution computed tomography and chest
almost exclusively in cigarette smokers and is characterized radiography in the diagnosis of chronic diffuse infiltrative
lung disease. Clin Radiol. 1991;44:222–226.
by the presence of cysts of various shapes and nodules in- 4. McLoud TC, Carrington CB, Gaensler EA. Diffuse infiltrative
volving the middle and upper lung zones and sparing lung disease: a new scheme for description. Radiology. 1983;149:
of the lung bases. Bonelli et al. (340) reviewed the 353–363.
5. Grenier P, Valeyre D, Cluzel P, et al. Chronic diffuse interstitial
HRCT findings in 34 patients with various cystic lung lung disease: diagnostic value of chest radiography and high-
diseases and emphysema. Based on the HRCT findings, resolution CT. Radiology. 1991;179:123–132.
6. Mathieson JR, Mayo JR, Staples CA, et al. Chronic diffuse infil- 29. Weibel ER. Looking into the lung: what can it tell us? AJR Am J
trative lung disease: comparison of diagnostic accuracy of CT Roentgenol. 1979;133:1021–1031.
and chest radiography. Radiology. 1989;171:111–116. 30. Webb WR. High-resolution computed tomography of the lung:
7. Lee KS, Primack SL, Staples CA, et al. Chronic infiltrative lung normal and abnormal anatomy. Semin Roentgenol. 1991;26:
disease: comparison of diagnostic accuracies of radiography and 110–117.
low- and conventional-dose thin-section CT. Radiology 1994; 31. Webb WR. High resolution lung computed tomography: nor-
191:669–673. mal anatomic and pathologic findings. Radiol Clin North Am.
8. Todo G, Itoh H, Nakano Y, et al. High-resolution CT for the 1991;29:1051–1063.
evaluation of pulmonary peripheral disorders. Jpn J Clin Radiol. 32. Webb WR, Stein MG, Finkbeiner WE, et al. Normal and diseased
1982;27:1319–1326. isolated lungs: high-resolution CT. Radiology. 1988;166:81–87.
9. Mayo JR, Webb WR, Gould R, et al. High-resolution CT of the 33. Kim JS, Müller NL, Park CS, et al. Bronchoarterial ratio on thin
lungs: an optimal approach. Radiology. 1987;163:507–510. section CT: comparison between high altitude and sea level.
10. Zwirewich CV, Mayo JR, Müller NL. Low-dose high-resolution J Comput Assist Tomogr. 1997;21:306–311.
CT of lung parenchyma. Radiology. 1991;180:413–417. 34. Lynch DA, Newell JD, Tschomper BA, et al. Uncomplicated
11. Majurin ML, Varpula M, Kurki T, et al. High-resolution CT of the asthma in adults: comparison of CT appearance of the lungs in
lung in asbestos-exposed subjects: comparison of low-dose and asthmatic and healthy subjects. Radiology. 1993;188:829–833.
high-dose HRCT. Acta Radiol. 1994;35:473–477. 35. Zerhouni E. Computed tomography of the pulmonary pare-
12. Murata K, Khan A, Rojas KA, et al. Optimization of computed nchyma: an overview. Chest. 1989;95:901–907.
tomography technique to demonstrate the fine structure of the 36. Zerhouni EA, Naidich DP, Stitik FP, et al. Computed tomogra-
lung. Invest Radiol. 1988;23:170–175. phy of the pulmonary parenchyma: part 2, interstitial disease.
13. Aberle DR, Gamsu G, Ray CS, et al. Asbestos-related pleural and J Thorac Imaging 1985;1:54–64.
parenchymal fibrosis: detection with high-resolution CT. 37. King GG, Muller NL, Whittall KP, et al. An analysis algorithm for
Radiology. 1988;166:729–734. measuring airway lumen and wall areas from high-resolution
14. Lynch DA, Rose CS, Way D, et al. Hypersensitivity pneumonitis: computed tomographic data. Am J Respir Crit Care Med. 2000;
sensitivity of high-resolution CT in a population-based study. 161:574–580.
AJR Am J Roentgenol. 1992;159:469–472. 38. Webb WR, Gamsu G, Wall SD, et al. CT of a bronchial phantom:
15. Remy-Jardin M, Remy J, Wallaert B, et al. Subacute and chronic factors affecting appearance and size measurements. Invest
bird breeder hypersensitivity pneumonitis: sequential evalua- Radiol. 1984;19:394–398.
tion with CT and correlation with lung function tests and 39. Weibel ER, Taylor CR. Design and structure of the human lung.
bronchoalveolar lavage. Radiology. 1993;198:111–118. In: Fishman AP, ed. Pulmonary diseases and disorders, 2nd ed.
16. Mayo JR, Aldrich J, Muller NL. Radiation exposure at chest CT: a New York: McGraw-Hill; 1988:11–60.
statement of the Fleischner Society. Radiology. 2003;228:15–21. 40. Murata K, Itoh H, Todo G, et al. Centrilobular lesions of the
17. Leung AN, Staples CA, Müller NL. Chronic diffuse infiltrative lung: demonstration by high-resolution CT and pathologic cor-
lung disease: comparison of diagnostic accuracy of high- relation. Radiology. 1986;161:641–645.
resolution and conventional CT. AJR Am J Roentgenol. 1991;157: 41. Miller WS. The lung. Springfield: Charles C Thomas; 1947.
693–696. 42. Heitzman ER, Markarian B, Berger I, et al. The secondary
18. Johkoh T, Muller NL, Nakamura H. Multidetector spiral high- pulmonary lobule: a practical concept for interpretation of radi-
resolution computed tomography of the lungs: distribution of ographs, II: application of the anatomic concept to an under-
findings on coronal image reconstructions. J Thorac Imaging. standing of roentgen pattern in disease states. Radiology.
2002;17:291–305. 1969;93:514–520.
19. Remy-Jardin M, Campistron P, Amara A, et al. Usefulness of 43. Itoh H, Murata K, Konishi J, et al. Diffuse lung disease: patho-
coronal reformations in the diagnostic evaluation of infiltrative logic basis for the high-resolution computed tomography find-
lung disease. J Comput Assist Tomogr. 2003;27:266–273. ings. J Thorac Imaging. 1993;8:176–188.
20. Bhalla M, Naidich DP, McGuinness G, et al. Diffuse lung 44. Osborne DR, Effmann EL, Hedlund LW. Postnatal growth and
disease: assessment with helical CT: preliminary observations size of the pulmonary acinus and secondary lobule in man. AJR
of the role of maximum and minimum intensity projection Am J Roentgenol. 1983;140:449–454.
images [see comments]. Radiology. 1996;200:341–347. 45. Reid L. The secondary pulmonary lobule in the adult human
21. Remy-Jardin M, Remy J, Gosselin B, et al. Sliding thin slab, lung, with special reference to its appearance in bronchograms.
minimum intensity projection technique in the diagnosis of Thorax. 1958;13:110–115.
emphysema: histopathologic-CT correlation. Radiology. 1996; 46. Reid L, Simon G. The peripheral pattern in the normal bron-
200:665–671. chogram and its relation to peripheral pulmonary anatomy.
21a. Leswick DA, Webster ST, Wilcox BA, et al. Radiation cost of heli- Thorax. 1958;13:103–109.
cal high-resolution chest CT. Radiology. 2005;184:742–745 47. Gamsu G, Thurlbeck WM, Macklem PT, et al. Peripheral bron-
22. Volpe J, Storto ML, Lee K, et al. High-resolution CT of the lung: chographic morphology in the normal human lung. Invest
determination of the usefulness of CT scans obtained with the Radiol. 1971;6:161–170.
patient prone based on plain radiographic findings. AJR Am 48. Colby TV, Swensen SJ. Anatomic distribution and histopatho-
J Roentgenol. 1997;169:369–374. logic patterns in diffuse lung disease: correlation with HRCT.
23. Arakawa H, Webb WR. Air trapping on expiratory high- J Thorac Imaging. 1996;11:1–26.
resolution CT scans in the absence of inspiratory scan abnormal- 49. Johkoh T, Müller NL, Ichikado K, et al. Perilobular pulmonary
ities: correlation with pulmonary function tests and differential opacities: high-resolution CT findings and pathologic correla-
diagnosis. AJR Am J Roentgenol. 1998;170:1349–1353. tion. J Thorac Imaging. 1999;14:172–177.
24. Park CS, Müller NL, Worthy SA, et al. Airway obstruction in 50. Müller NL, Miller RR. Computed tomography of chronic diffuse
asthmatic and healthy individuals: inspiratory and expiratory infiltrative lung disease: part 1. Am Rev Respir Dis. 1990;142:
thin-section CT findings. Radiology. 1997;203:361–367. 1206–1215.
25. Stern EJ, Webb WR, Gamsu G. Dynamic quantitative computed 51. Müller NL, Miller RR. Computed tomography of chronic diffuse
tomography: a predictor of pulmonary function in obstructive infiltrative lung disease: part 2. Am Rev Respir Dis. 1990;142:
lung diseases. Invest Radiol. 1994;29:564–569. 1440–1448.
26. Boyden EA. The nomenclature of the bronchopulmonary seg- 52. Reid L, Rubino M. The connective tissue septa in the foetal
ments and their blood supply. Dis Chest. 1961;39:1–6. human lung. Thorax. 1959;14:3–13.
27. Naidich DP, Zinn WL, Ettenger NA, et al. Basilar segmental 53. Aberle DR, Gamsu G, Ray CS. High-resolution CT of benign
bronchi: thin-section CT evaluation. Radiology. 1988;169:11–16. asbestos-related diseases: clinical and radiographic correlation.
28. Osborne D, Vock P, Godwin JD, et al. CT identification of bron- AJR Am J Roentgenol. 1988;151:883–891.
chopulmonary segments: 50 normal subjects. AJR Am 54. Itoh H, Tokunaga S, Asamoto H, et al. Radiologic-pathologic
J Roentgenol. 1984;142:47–52. correlations of small lung nodules with special reference
to peribronchiolar nodules. AJR Am J Roentgenol. 1978;130: 82. Rossi SE, Erasmus JJ, Volpacchio M, et al. “Crazy-paving” pat-
223–231. tern at thin-section CT of the lungs: radiologic-pathologic
55. Webb WR. High-resolution CT of the lung parenchyma. Radiol overview. Radiographics. 2003;23:1509–1519.
Clin North Am. 1989;27:1085–1097. 83. Ren H, Hruban RH, Kuhlman JE, et al. Computed tomography
56. Millar AB, Denison DM. Vertical gradients of lung density in of inflation-fixed lungs: the beaded septum sign of pulmonary
healthy supine men. Thorax. 1989;44:485–490. metastases. J Comput Assist Tomogr. 1989;13:411–416.
57. Vock P, Malanowski D, Tschaeppeler H, et al. Computed tomo- 84. Hartman TE, Primack SL, Müller NL, et al. Diagnosis of thoracic
graphic lung density in children. Invest Radiol. 1987;22: complications in AIDS: accuracy of CT. AJR Am J Roentgenol.
627–631. 1994;162:547–553.
58. Murata K, Itoh H, Senda M, et al. Stratified impairment of pul- 85. Lynch DA, Webb WR, Gamsu G, et al. Computed tomography
monary ventilation in “diffuse panbronchiolitis”: PET and CT in pulmonary sarcoidosis. J Comput Assist Tomogr. 1989;13:
studies. J Comput Assist Tomogr. 1989;13:48–53. 405–410.
59. Genereux GP. Computed tomography and the lung: review of 86. Müller NL, Mawson JB, Mathieson JR, et al. Sarcoidosis: correla-
anatomic and densitometric features with their clinical application of extent of disease at CT with clinical, functional, and radi-
tion. J Can Assoc Radiol. 1985;36:88–102. ographic findings. Radiology. 1989;171:613–618.
60. Rosenblum LJ, Mauceri RA, Wellenstein DE, et al. Density 87. Remy-Jardin M, Beuscart R, Sault MC, et al. Subpleural micro-
patterns in the normal lung as determined by computed tomog- nodules in diffuse infiltrative lung diseases: evaluation with
raphy. Radiology. 1980;137:409–416. thin-section CT scans. Radiology. 1990;177:133–139.
61. Webb WR, Stern EJ, Kanth N, et al. Dynamic pulmonary CT: 88. Remy-Jardin M, Degreef JM, Beuscart R, et al. Coal worker’s
findings in normal adult men. Radiology. 1993;186:117–124. pneumoconiosis: CT assessment in exposed workers and corre-
62. Nishimura K, Kitaichi M, Izumi T, et al. Usual interstitial lation with radiographic findings. Radiology. 1990;177:363–371.
pneumonia: histologic correlation with high-resolution CT. 89. Johkoh T, Müller NL, Pickford HA, et al. Lymphocytic intersti-
Radiology. 1992;182:337–342. tial pneumonia: thin-section CT findings in 22 patients.
63. Storto ML, Kee ST, Golden JA, et al. Hydrostatic pulmonary Radiology. 1999;212:567–572.
edema: high-resolution CT findings. AJR Am J Roentgenol. 90. Graham CM, Stern EJ, Finkbeiner WE, et al. High-resolution CT
1995;165:817–820. appearance of diffuse alveolar septal amyloidosis. AJR Am
64. Munk PL, Müller NL, Miller RR, et al. Pulmonary lymphangitic J Roentgenol. 1992;158:265–267.
carcinomatosis: CT and pathologic findings. Radiology. 1988; 91. Pickford HA, Swensen SJ, Utz JP. Thoracic cross-sectional imag-
166:705–709. ing of amyloidosis. AJR Am J Roentgenol. 1997;168:351–355.
65. Stein MG, Mayo J, Müller N, et al.. Pulmonary lymphangitic 92. Akira M, Yamamoto S, Inoue Y, et al. Resolution CT of
spread of carcinoma: appearance on CT scans. Radiology. 1987; asbestosis and idiopathic pulmonary fibrosis. AJR Am J
162:371–375. Roentgenol. 2003;181:163–169.
66. Bessis L, Callard P, Gotheil C, et al. High-resolution CT of 93. Primack SL, Hartman TE, Hansell DM, et al. End-stage lung dis-
parenchymal lung disease: precise correlation with histologic ease: CT findings in 61 patients. Radiology. 1993;189:681–686.
findings. Radiographics. 1992;12:45–58. 94. Akira M, Yamamoto S, Yokoyama K, et al. Asbestosis: high-
67. Müller NL, Kullnig P, Miller RR. The CT findings of pulmonary resolution CT-pathologic correlation. Radiology. 1990;176:
sarcoidosis: analysis of 25 patients. AJR Am J Roentgenol. 389–394.
1989;152:1179–1182. 95. Müller NL, Miller RR, Webb WR, et al. Fibrosing alveolitis:
68. Westcott JL, Cole SR. Traction bronchiectasis in end-stage CT-pathologic correlation. Radiology. 1986;160:585–588.
pulmonary fibrosis. Radiology. 1986;161:665–669. 96. Hunninghake GW, Zimmerman MB, Schwartz DA, et al. Utility
69. Primack SL, Müller NL, Mayo JR, et al. Pulmonary parenchymal of a lung biopsy for the diagnosis of idiopathic pulmonary fi-
abnormalities of vascular origin: high-resolution CT findings. brosis. Am J Respir Crit Care Med. 2001;164:193–196.
Radiographics. 1994;14:739–746. 97. American Thoracic Society/European Respiratory Society.
70. Au V, Leung AN. Radiologic manifestations of lymphoma in the International Multidisciplinary Consensus Classification of the
thorax. AJR Am J Roentgenol. 1997;168:93–98. Idiopathic Interstitial Pneumonias. Am J Respir Crit Care Med.
71. Lee KS, Kim Y, Primack SL. Imaging of pulmonary lymphomas. 2002;165:277–304.
AJR Am J Roentgenol. 1997;168:339–345. 98. Johkoh T, Muller NL, Colby TV, et al. Nonspecific interstitial
72. Heyneman LE, Johkoh T, Ward S, et al. Pulmonary leukemic in- pneumonia: correlation between thin-section CT findings and
filtrates: high-resolution CT findings in 10 patients. AJR Am pathologic subgroups in 55 patients. Radiology. 2002;225:
J Roentgenol. 2000;174:517–521. 199–204.
73. Okada F, Ando Y, Kondo Y, et al. Thoracic CT findings of adult 99. MacDonald SL, Rubens MB, Hansell DM, et al. Nonspecific
T-cell leukemia or lymphoma. AJR Am J Roentgenol. 2004; interstitial pneumonia and usual interstitial pneumonia: com-
182:761–767. parative appearances at and diagnostic accuracy of thin-section
74. Silva CI, Colby TV, Muller NL. Asthma and associated condi- CT. Radiology. 2001;221:600–605.
tions: high-resolution CT and pathologic findings. AJR Am 100. Adler BD, Padley SP, Müller NL, et al. Chronic hypersensitivity
J Roentgenol. 2004;183:817–824. pneumonitis: high-resolution CT and radiographic features in
75. Worthy SA, Müller NL, Hansell DM, et al. Churg-Strauss 16 patients. Radiology. 1992;185:91–95.
syndrome: the spectrum of pulmonary CT findings in 17 patients. 101. Lynch DA, Newell JD, Logan PM, et al. Can CT distinguish
AJR Am J Roentgenol. 1998;170:297–300. hypersensitivity pneumonitis from idiopathic pulmonary fibro-
76. Gotway MB, Dawn SK, Sellami D, et al. Acute rejection follow- sis? AJR Am J Roentgenol. 1995;165:807–811.
ing lung transplantation: limitations in accuracy of thin-section 102. Nishimura K, Itoh H, Kitaichi M, et al. Pulmonary sarcoidosis:
CT for diagnosis. Radiology. 2001;221:207–212. correlation of CT and histopathologic findings. Radiology.
77. Nobre LF, Muller NL, de Souza Junior AS, et al. Congenital pul- 1993;189:105–109.
monary lymphangiectasia: CT and pathologic findings. J Thorac 103. Honda O, Johkoh T, Ichikado K, et al. Comparison of high reso-
Imaging. 2004;19:56–59. lution CT findings of sarcoidosis, lymphoma, and lymphangitic
78. Rodrigues R, Marchiori E, Muller NL. et al. Niemann-Pick carcinoma: is there any difference of involved interstitium?
disease: high-resolution CT findings in two siblings. J Comput J Comput Assist Tomogr. 1999;23:374–379.
Assist Tomogr. 2004;28:52–54. 104. Akira M, Kita N, Higashihara T, et al. Summer-type hypersensitiv-
79. Godwin JD, Müller NL, Takasugi JE. Pulmonary alveolar pro- ity pneumonitis: comparison of high-resolution CT and plain
teinosis: CT findings. Radiology. 1988;169:609–613. radiographic findings. AJR Am J Roentgenol. 1992;158:1223–1228.
80. Johkoh T, Itoh H, Müller NL, et al. Crazy-paving appearance at 105. Lee KS, Kim YH, Kim WS, et al. Endobronchial tuberculosis: CT
thin-section CT: spectrum of disease and pathologic findings. features. J Comput Assist Tomogr. 1991;15:424–428.
Radiology. 1999;211:155–160. 106. Remy-Jardin M, Remy J, Gosselin B, et al. Lung parenchymal
81. Murch CR, Carr DH. Computed tomography appearances of changes secondary to cigarette smoking: pathologic-CT correla-
pulmonary alveolar proteinosis. Clin Radiol. 1989;40:240–243. tions. Radiology. 1993;186:643–651.
107. Heyneman LE, Ward S, Lynch DA, et al. Respiratory bronchioli- 129. Kami M, Kishi Y, Hamaki T, et al. The value of the chest
tis, respiratory bronchiolitis-associated interstitial lung disease, computed tomography halo sign in the diagnosis of invasive
and desquamative interstitial pneumonia: different entities or pulmonary aspergillosis: an autopsy-based retrospective study
part of the spectrum of the same disease process? AJR Am of 48 patients. Mycoses. 2002;45:287–294.
J Roentgenol. 1999;173:1617–1622. 130. Austin JH, Müller NL, Friedman PJ, et al. Glossary of terms for
108. Park JS, Brown KK, Tuder RM, et al. Respiratory bronchiolitis- CT of the lungs: recommendations of the Nomenclature
associated interstitial lung disease: radiologic features with Committee of the Fleischner Society. Radiology. 1996;200:
clinical and pathologic correlation. J Comput Assist Tomogr. 327–331.
2002;26:13–20. 131. Leung AN, Miller RR, Müller NL. Parenchymal opacification in
109. Nishimura K, Kitaichi M, Izumi T, et al. Diffuse panbronchioli- chronic infiltrative lung diseases: CT-pathologic correlation.
tis: correlation of high-resolution CT and pathologic findings. Radiology. 1993;188:209–214.
Radiology. 1992;184:779–785. 132. Remy-Jardin M, Remy J, Giraud F, et al. Computed tomography
110. Müller NL, Staples CA, Miller RR. Bronchiolitis obliterans assessment of ground-glass opacity: semiology and significance.
organizing pneumonia: CT features in 14 patients. AJR Am J Thorac Imaging. 1993;8:249–264.
J Roentgenol. 1990;154:983–987. 133. Wells AU, Hansell DM, Rubens MB, et al. The predictive value of
111. Aquino SL, Gamsu G, Webb WR, et al. Tree-in-bud pattern: appearances of thin-section computed tomography in fibrosing
frequency and significance on thin section CT. J Comput Assist alveolitis. Am Rev Respir Dis. 1993;148:1076–1082.
Tomogr. 1996;20:594–599. 134. Miller WT Jr, Shah RM. Isolated diffuse ground-glass opacity
112. Im JG, Itoh H, Shim YS, et al. Pulmonary tuberculosis: CT in thoracic CT: causes and clinical presentations. AJR Am
findings—early active disease and sequential change with anti- J Roentgenol. 2005;184:613–622.
tuberculous therapy. Radiology. 1993;186:653–660. 135. Bergin CJ, Wirth RL, Berry GJ, et al. Pneumocystis carinii pneumo-
113. Collins J, Blankenbaker D, Stern EJ. CT patterns of bronchiolar nia: CT and HRCT observations. J Comput Assist Tomogr.
disease: what is “tree-in-bud”? AJR Am J Roentgenol. 1998;171: 1990;14:756–759.
365–370. 136. Hidalgo A, Falco V, Mauleon S, et al. Accuracy of high-resolution
114. Gruden JF, Webb WR, Naidich DP, et al. Multinodular disease: CT in distinguishing between Pneumocystis carinii pneumonia
anatomic localization at thin-section CT: multireader evalua- and non-Pneumocystis carinii pneumonia in AIDS patients. Eur
tion of a simple algorithm. Radiology. 1999;210:711–720. Radiol. 2003;13:1179–1184.
114a. Nishimura K, Kitaichi M, Izumi T, et al. Diffuse panbronchioli- 137. Brown MJ, Miller RR, Müller NL. Acute lung disease in the
tis: correlation of high resolution CT and pathologic findings. immunocompromised host: CT and pathologic examination
Radiology. 1992;184:779–785. findings. Radiology. 1994;190:247–254.
114b. Kudoh S. Erythromycin treatment in diffuse panbronchiolitis. 138. Primack SL, Müller NL. High-resolution computed tomography
Curr Opin Pulmon Med.. 1998;4:116–121 in acute diffuse lung disease in the immunocompromised
115. Oh YW, Kim YH, Lee NJ, et al. High-resolution CT appear- patient. Radiol Clin North Am, 1994;32:731–744.
ance of miliary tuberculosis. J Comput Assist Tomogr. 1994;18: 139. Franquet T, Müller NL, Lee KS, et al. High-resolution CT
862–866. and pathologic findings of noninfectious pulmonary complica-
116. McGuinness G, Naidich DP, Jagirdar J, et al. High resolution CT tions after hematopoietic stem cell transplantation. AJR Am
findings in miliary lung disease. J Comput Assist Tomogr. 1992; J Roentgenol. 2005;184:629–637.
16:384–390. 140. Hiorns MP, Screaton NJ, Muller NL. Acute lung disease in the
117. Murata K, Takahashi M, Mori M, et al. Pulmonary meta- immunocompromised host. Radiol Clin North Am. 2001;39:
static nodules: CT-pathologic correlation. Radiology. 1992;182: 1137–1151, vi.
331–335. 141. Cormier Y, Brown M, Worthy S, et al. High-resolution computed
117a. Raoof S, Amchentsev A, Vlahos I, et al. Pictorial essay: multi- tomographic characteristics in acute farmer’s lung and in its
nodular disease: a high-resolution CT scan diagnostic algo- follow-up. Eur Respir J. 2000;16:56–60.
rithm. Chest 2006;129:1–11 142. Remy-Jardin M, Remy J, Wallaert B, et al. Pulmonary involve-
118. Bégin R, Ostiguy G, Fillion R, et al. Computed tomography scan ment in progressive systemic sclerosis: sequential evaluation
in the early detection of silicosis. Am Rev Respir Dis. 1991;3:1. with CT, pulmonary function tests, and bronchoalveolar lavage.
119. Grenier P, Chevret S, Beigelman C, et al. Chronic diffuse infiltra- Radiology. 1993;188:499–506.
tive lung disease: determination of the diagnostic value of clini- 143. Hartman TE, Primack SL, Swensen SJ, et al. Desquamative inter-
cal data, chest radiography, and CT with Bayesian analysis. stitial pneumonia: thin-section CT findings in 22 patients.
120. Murdoch J, Müller NL. Pulmonary sarcoidosis: changes on fol- 144. Lee KS, Kullnig P, Hartman TE, et al. Cryptogenic organizing
low-up CT examinations. AJR Am J Roentgenol. 1992;159: pneumonia: CT findings in 43 patients. AJR Am J Roentgenol.
473–477. 1994;162:543–546.
121. Padley SPG, Adler BD, Staples CA, et al. Pulmonary talcosis: CT 145. Remy-Jardin M, Giraud F, Remy J, et al. Importance
findings in three cases. Radiology. 1993;186:125–127. of ground-glass attenuation in chronic diffuse infiltrative
122. Ward S, Heyneman LE, Reittner P, et al. Talcosis associated with lung disease: pathologic-CT correlation. Radiology. 1993;189:
IV abuse of oral medications: CT findings. AJR Am J Roentgenol. 693–698.
2000;174:789–793. 146. Franquet T, Giménez A, Bordes R, et al. The crazy-paving pattern
123. Kuhlman JE, Fishman EK, Teigen C. Pulmonary septic emboli: in exogenous lipoid pneumonia: CT-pathologic correlation. AJR
diagnosis with CT. Radiology. 1990;174:211–213. Am J Roentgenol. 1998;170:315–317.
124. Sheehan RE, Flint JD, Muller NL. Computed tomography fea- 147. Tan RT, Kuzo RS. High-resolution CT findings of mucinous
tures of the thoracic manifestations of Wegener granulomatosis. bronchioloalveolar carcinoma: a case of pseudopulmonary
J Thorac Imaging. 2003;18:34–41. alveolar proteinosis. AJR Am J Roentgenol. 1997;168:99–100.
125. Lee KS, Kim TS, Fujimoto K, et al. Thoracic manifestation of 148. Lee KN, Levin DL, Webb WR, et al. Pulmonary alveolar pro-
Wegener’s granulomatosis: CT findings in 30 patients. Eur teinosis: high-resolution CT, chest radiographic, and functional
Radiol. 2003;13:43–1351. correlations. Chest. 1997;111:989–995.
126. Carignan S, Staples CA, Müller NL. Intrathoracic lymphoprolif- 149. Hansell DM, Wells AU, Padley SP, et al. Hypersensitivity pneu-
erative disorders in the immunocompromised patient: CT find- monitis: correlation of individual CT patterns with functional
ings. Radiology. 1995;197:53–58. abnormalities. Radiology. 1996;199:123–128.
127. Gaeta M, Blandino A, Scribano E, et al. Computed tomography 150. Primack SL, Miller RR, Müller NL. Diffuse pulmonary hemor-
halo sign in pulmonary nodules: frequency and diagnostic rhage: clinical, pathologic, and imaging features. AJR Am
value. J Thorac Imaging. 1999;14:109–113. J Roentgenol. 1995;164:295–300.
128. Kim Y, Lee KS, Jung KJ, et al. Halo sign on high resolution CT: 151. Murayama S, Murakami J, Yabuuchi H, et al. “Crazy paving
findings in spectrum of pulmonary diseases with pathologic appearance” on high resolution CT in various diseases. J Comput
correlation. J Comput Assist Tomogr. 1999;23:622–626. Assist Tomogr. 1999;23:749–752.
152. Cheon JE, Lee KS, Jung GS, et al. Acute eosinophilic pneumo- 176. Spouge D, Mayo JR, Cardoso W, et al. Panacinar emphysema:
nia: radiographic and CT findings in six patients. AJR Am CT and pathologic correlation. J Comput Assist Tomogr. 1993;17:
J Roentgenol. 1996;167:1195–1199. 710–713.
153. Logan PM. Thoracic manifestations of external beam radio- 177. Worthy SA, Müller NL, Hartman TE, et al. Mosaic attenuation
therapy. AJR Am J Roentgenol. 1998;171:569–577. pattern on thin-section CT scans of the lung: differentiation
154. Naidich DP, Zerhouni EA, Hutchins GM, et al. Computed among infiltrative lung, airway, and vascular diseases as a
tomography of the pulmonary parenchyma: part 1: distal air- cause. Radiology. 1997;205:465–470.
space disease. J Thorac Imaging. 1985;1:39–53. 178. Hansell DM, Rubens MB, Padley SP, et al. Obliterative bronchi-
155. Primack SL, Hartman TE, Ikezoe J, et al. Acute interstitial pneu- olitis: individual CT signs of small airways disease and func-
monia: radiographic and CT findings in nine patients. tional correlation. Radiology. 1997;203:721–726.
Radiology. 1993;188:817–820. 179. Roberts HR, Wells AU, Milne DG, et al. Airflow obstruction in
156. Johkoh T, Muller NL, Akira M, et al. Eosinophilic lung diseases: bronchiectasis: correlation between computed tomography fea-
diagnostic accuracy of thin-section CT in 111 patients. tures and pulmonary function tests. Thorax. 2000;55: 198–204.
Radiology. 2000;216:773–780. 180. King MA, Ysrael M, Bergin CJ. Chronic thromboembolic pul-
157. Greene R. Adult respiratory distress syndrome: acute alveolar monary hypertension: CT findings. AJR Am J Roentgenol. 1998;
damage. Radiology. 1987;163:57–66. 170:955–960.
158. Ujita M, Renzoni EA, Veeraraghavan S, et al. Organizing pneu- 181. Oikonomou A, Dennie CJ, Muller NL, et al. Chronic throm-
monia: perilobular pattern at thin-section CT. Radiology. 2004; boembolic pulmonary arterial hypertension: correlation of
232:757–761. postoperative results of thromboendarterectomy with preoper-
159. Kim SJ, Lee KS, Ryu YH, et al. Reversed halo sign on high- ative helical contrast-enhanced computed tomography. J
resolution CT of cryptogenic organizing pneumonia: diagnos- Thorac Imaging. 2004;19:67–73.
tic implications. AJR Am J Roentgenol. 2003;180:1251–1254. 182. King MA, Bergin CJ, Yeung DWC, et al. Chronic pulmonary
160. Mayo JR, Müller NL, Road J, et al. Chronic eosinophilic pneu- thromboembolism: detection of regional hypoperfusion with
monia: CT findings in six cases. AJR Am J Roentgenol. 1989;153: CT. Radiology. 1994;191:359–363.
727–730. 183. Schwickert HC, Schweden F, Schild HH, et al. Pulmonary arter-
161. Lee KS, Müller NL, Hale V, et al. Lipoid pneumonia: CT find- ies and lung parenchyma in chronic pulmonary embolism:
ings. J Comput Assist Tomogr. 1995;19:48–51. preoperative and postoperative CT findings. Radiology. 1994;
162. Kuhlman JE. The role of chest computed tomography in the 191:351–357.
diagnosis of drug-related reactions. J Thorac Imaging. 1991;6: 184. Martin KW, Sagel SS, Siegel BA. Mosaic oligemia simulating
52–61. pulmonary infiltrates on CT. AJR Am J Roentgenol. 1986;147:
162a. Silva CI, Müller NL. Drug-induced lung diseases: most com- 670–673.
mon reaction patterns and corresponding high-resolution CT 185. Stern EJ, Swensen SJ, Hartman TE, et al. CT mosaic pattern
manifestations. Semin Ultrasound CT MR. 2006;27:111–116. of lung attenuation: distinguishing different causes. AJR Am
163. Hartman TE, Müller NL, Primack SL, et al. Metastatic pul- J Roentgenol. 1995;165:813–816.
monary calcification in patients with hypercalcemia: findings 186. Müller NL, Miller RR. Diseases of the bronchioles: CT and
on chest radiographs and CT scans. AJR Am J Roentgenol. histopathologic findings. Radiology. 1995;196:3–12.
1994;162:799–802. 187. Carrington CB, Gaensler EA. Clinical-pathologic approach
164. Akira M, Sakatani M, Hara H. Thin-section CT findings in to diffuse infiltrative lung disease. Monogr Pathol. 1978;19:
rheumatoid arthritis-associated lung disease: CT patterns and 58–87.
their courses. J Comput Assist Tomogr. 1999;23:941–948. 188. Corrin B. Pathology of interstitial lung disease. In: Harasawa
164a. Sumikawa H, Johkoh T, Ichikado K, et al. Usual interstitial pneu- M, Fukuchi Y, Morinari H, eds. Interstitial pneumonia of un-
monia and chronic idiopathic interstitial pneumonia: analysis known etiology. Tokyo: University of Tokyo Press; 1989:149.
of CT appearance in 92 patients. Radiology. 2006;241:258–266. 189. Coultas DB, Zumwalt RE, Black WC, et al. The epidemiology of
165. Lenoir S, Grenier P, Brauner MW, et al. Pulmonary lymphan- interstitial lung diseases. Am J Respir Crit Care Med. 1994;150:
giomyomatosis and tuberous sclerosis: comparison of radi- 967–972.
ographic and thin-section CT findings. Radiology. 1990;175: 190. Travis WD, Colby TV, Koss MN, et al. Non-neoplastic disorders of
329–334. the lower respiratory tract. Washington: Armed Forces Institute of
166. Müller NL, Chiles C, Kullnig P. Pulmonary lymphangiomy- Pathology; 2002.
omatosis: correlation of CT with radiographic and functional 191. Johkoh T, Müller NL, Cartier Y, et al. Idiopathic interstitial
findings. Radiology. 1990;175:335–339. pneumonias: diagnostic accuracy of thin-section CT in 129
167. Brauner MW, Grenier P, Mouelhi MM, et al. Pulmonary histio- patients. Radiology. 1999;211:555–560.
cytosis X: evaluation with high resolution CT. Radiology. 192. American Thoracic Society (ATS) and the European Respiratory
1989;172:255–258. Society (ERS). Idiopathic pulmonary fibrosis: diagnosis and
168. Brauner MW, Grenier P, Tijani K, et al. Pulmonary Langerhans treatment: international consensus statement. Am J Respir Crit
cell histiocytosis: evolution of lesions on CT scans [see com- Care Med. 2000;161:646–664.
ments]. Radiology. 1997;204:497–502. 192a. Lynch DA, Travis WD, Müller NL, et al. Idiopathic interstitial
169. Moore AD, Godwin JD, Müller NL, et al. Pulmonary histiocy- pneumonias: CT features. Radiology. 2005;236:10–21.
tosis X: comparison of radiographic and CT findings. Radiology. 193. Lee JS, Im JG, Ahn JM, et al. Fibrosing alveolitis: prognostic
1989;172:249–254. implication of ground-glass attenuation at high-resolution CT.
170. Franquet T, Hansell DM, Senbanjo T, et al. Lung cysts in sub- Radiology. 1992;184:415–454.
acute hypersensitivity pneumonitis. J Comput Assist Tomogr. 194. Souza CA, Müller NL, Flint JD, et al. Idiopathic pulmonary
2003;27:475–478. fibrosis: spectrum of high-resolution CT findings. AJR Am
171. Kuhlman JE, Kavuru M, Fishman EK, et al. Pneumocystis carinii J Roentgenol. 2005;185:1531–1539.
pneumonia: spectrum of parenchymal CT findings. Radiology. 195. Gay SE, Kazerooni EA, Toews GB, et al. Idiopathic pulmonary
1990;175:711–714. fibrosis: predicting response to therapy and survival. Am J
172. Foster WL, Pratt PC, Roggli VL, et al. Centrilobular emphy- Respir Crit Care Med. 1998;157:1063–1072.
sema: CT-pathologic correlation. Radiology. 1986;159:27–32. 196. Hartman TE, Primack SL, Kang EY, et al. Disease progression in
173. Thurlbeck WM, Müller NL. Emphysema: definition, imaging, usual interstitial pneumonia compared with desquamative in-
and quantification. AJR Am J Roentgenol. 1994;163: 1017–1025. terstitial pneumonia: assessment with serial CT [see comments].
174. Tuddenham WJ. Glossary of terms for thoracic radiology: rec- Chest 1996;110:378–382.
ommendations of the Nomenclature Committee of the 197. Terriff BA, Kwan SY, Chan-Yeung MM, et al. Fibrosing alveoli-
Fleischner Society. AJR Am J Roentgenol. 1984;143:509–517. tis: chest radiography and CT as predictors of clinical and func-
175. Bergin CJ, Müller NL, Miller RR. CT in the qualitative assess- tional impairment at follow-up in 26 patients. Radiology.
ment of emphysema. J Thorac Imaging. 1986;1:94–103. 1992;184:445–449.
198. Akira M, Hamada H, Sakatani M, et al. CT findings during 220. Friedman AC, Fiel SB, Radecki PD, et al. Computed tomogra-
phase of accelerated deterioration in patients with idiopathic phy of benign pleural and pulmonary parenchymal abnormali-
pulmonary fibrosis. AJR Am J Roentgenol. 1997;168:79–83. ties related to asbestos exposure. Semin Ultrasound CT MR.
199. Kondoh Y, Taniguchi H, Kawabata Y, et al. Acute exacerbation 1990; 11:393–408.
in idiopathic pulmonary fibrosis: analysis of clinical and 221. Gamsu G, Aberle DR, Lynch D. Computed tomography in the
pathologic findings in three cases. Chest. 1993;103:1808–1812. diagnosis of asbestos-related thoracic disease. J Thorac Imaging.
200. Rice AJ, Wells AU, Bouros D, et al. Terminal diffuse alveolar 1989;4:61–67.
damage in relation to interstitial pneumonias: an autopsy 222. Murray K, Gamsu G, Webb WR, et al. High-resolution CT sam-
study. Am J Clin Pathol. 2003;119:709–714. pling for detection of asbestos-related lung disease. Acad Radiol.
201. Ambrosini V, Cancellieri A, Chilosi M, et al. Acute exacerbation 1995;2:111–115.
of idiopathic pulmonary fibrosis: report of a series. Eur Respir J. 223. Remy-Jardin M, Sobaszek A, et al. Asbestos-related pleuropul-
2003;22:821–826. monary diseases: evaluation with low-dose four-detector row
201a. Silva CIS, Müller NL, Fujimoto K, et al. Acute exacerbation spiral CT. Radiology. 2004;233:182–190.
of chronic interstitial pneumonia: high-resolution computed 224. Akira M, Yokoyama K, Yamamoto S, et al. Early asbestosis:
tomography and pathologic findings. J Thorac Imaging evaluation with high-resolution CT. Radiology. 1991;178:
2007:in press. 409–416.
202. Lee HJ, Im JG, Ahn JM, et al. Lung cancer in patients with idio- 225. Gamsu G, Salmon CJ, Warnock ML, et al. CT quantification of
pathic pulmonary fibrosis: CT findings. J Comput Assist Tomogr. interstitial fibrosis in patients with asbestosis: a comparison of
1996;20:979–982. two methods. AJR Am J Roentgenol. 1995;164:63–68.
203. Sakai S, Ono M, Nishio T, et al. Lung cancer associated with dif- 226. Yoshimura H, Hatakeyama M, Otsuji H, et al. Pulmonary
fuse pulmonary fibrosis: CT-pathologic correlation. J Thorac asbestosis: CT study of subpleural curvilinear shadow (work in
Imaging. 2003;18:67–71. progress). Radiology. 1986;158:653–658.
204. Chung MJ, Goo JM, Im JG. Pulmonary tuberculosis in patients 227. Bergin CJ, Castellino RA, Blank N, et al. Specificity of high-
with idiopathic pulmonary fibrosis. Eur J Radiol. 2004; resolution CT findings in pulmonary asbestosis: do patients
52:175–179. scanned for other indications have similar findings? AJR Am
205. Bergin C, Castellino RA. Mediastinal lymph node enlargement J Roentgenol. 1994;163:551–555.
on CT scans in patients with usual interstitial pneumonitis. AJR 228. Al-Jarad N, Strickland B, Pearson MC, et al. High-resolution
Am J Roentgenol. 1990:154:251–254. computed tomographic assessment of asbestosis and crypto-
206. Niimi H, Kang EY, Kwong JS, et al. CT of chronic infiltrative genic fibrosing alveolitis: a comparative study. Thorax. 1992;
lung disease: prevalence of mediastinal lymphadenopathy. 47:645–650.
J Comput Assist Tomogr. 1996;20:305–308. 229. Copley SJ, Wells AU, Sivakumaran P, et al. Asbestosis and
207. Jung JI, Kim HH, Jung YJ, et al. Mediastinal lymphadenopathy idiopathic pulmonary fibrosis: comparison of thin-section CT
in pulmonary fibrosis: correlation with disease severity. features. Radiology. 2003;229:731–736.
J Comput Assist Tomogr. 2000;24:706–710. 230. Sette A, Neder JA, Nery LE, et al. Thin-section CT abnormalities
208. Franquet T, Gimenez A, Alegret X, et al. Mediastinal lym- and pulmonary gas exchange impairment in workers exposed
phadenopathy in cryptogenic fibrosing alveolitis: the effect of to asbestos. Radiology. 2004;232:66–74.
steroid therapy on the prevalence of nodal enlargement. Clin 231. Heitzman ER. The lung: radiologic-pathologic correlations. St.
Radiol. 1998;53:435–438. Louis: Mosby; 1984.
209. American Thoracic Society. Diagnosis and initial management 232. Müller NL, Fraser RS, Lee KS, et al. Diseases of the lung: radiologic
of nonmalignant diseases related to asbestos. Am J Respir Crit and pathologic correlations. Philadelphia: Lippincott Williams &
Care Med. 2004;170:691–715. Wilkins; 2003.
210. Bégin R, Ostiguy G, Filion R, et al. Recent advances in the early 233. Johkoh T, Ikezoe J, Tomiyama N, et al. CT findings in lymphan-
diagnosis of asbestosis. Semin Roentgenol. 1992;27:121–139. gitic carcinomatosis of the lung: correlation with histologic
211. Craighead JE, Abraham JL, Churg A, et al. The pathology of findings and pulmonary function tests. AJR Am J Roentgenol.
asbestos-associated disease of the lungs and pleural cavities: 1992;158:1217–1222.
diagnostic criteria and proposed grading schema. Arch Pathol 234. Colby T, Carrington C. Interstitial lung disease. In: Thurlbeck
Lab Med. 1982;106:544–596. W, Churg A, eds. Pathology of the lung. New York: Thieme
212. American Thoracic Society. The diagnosis of nonmalignant Medical Publishers, 1995;589–737.
diseases related to asbestos. Am Rev Repir Dis. 1986;134: 235. American Thoracic Society (ATS), the European Respiratory
363–368. Society (ERS) and the World Association of Sarcoidosis and
213. Epstein DM, Miller WT, Bresnitz EA, et al.. Application of ILO Other Granulomatous Disorders. Statement on sarcoidosis:
classification to a population without industrial exposure: find- Joint Statement of the American Thoracic Society (ATS), the
ings to be differentiated from pneumoconiosis. AJR Am European Respiratory Society (ERS) and the World Association
J Roentgenol. 1984;142:53–58. of Sarcoidosis and Other Granulomatous Disorders (WASOG)
214. Gaensler EA, Carrington CB, Coutu RE, et al. Pathological, adopted by the ATS Board of Directors and by the ERS
physiological, and radiological correlations in the pneumoco- Executive Committee, February 1999. Am J Respir Crit Care
nioses. Ann N Y Acad Sci. 1972;200:574–607. Med. 1999; 160:736–755.
215. Kipen HM, Lilis R, Suzuki Y, et al. Pulmonary fibrosis in 236. Hillerdal G, Neu E, Osterman K, et al. Sarcoidosis: epidemiol-
asbestos insulation workers with lung cancer: a radiological ogy and prognosis: a 15-year European study. Am Rev Respir
and histopathological evaluation. Br J Indust Med. 1987;44: Dis. 1984;130:29–32.
96–100. 237. McLoud TC, Epler GR, Gaensler EA, et al. A radiographic classi-
216. Rockoff SD, Schwartz A. Roentgenographic underestimation fication for sarcoidosis: physiologic correlation. Invest Radiol.
of early asbestosis by International Labor Organization classifi- 1982;17:129–138.
cation: analysis of data and probabilities. Chest.1988; 238. Scadding JG, Mitchell DN. Sarcoidosis. London: Chapman &
93:1088–1091. Hall; 1985.
217. Ducatman AM, Yang WN, Forman SA. “B-readers” and asbestos 239. Crystal RG, Bitterman PB, Rennard SI, et al. Interstitial lung
medical surveillance. J Occup Med. 1988;30:644–647. diseases of unknown cause: disorders characterized by chronic
218. Friedman AC, Fiel SB, Fisher MS, et al. Asbestos-related pleural inflammation of the lower respiratory tract. N Engl J Med.
disease and asbestosis: a comparison of CT and chest radiogra- 1984; 310:154–166.
phy. AJR Am J Roentgenol. 1988;150:268–275. 240. Brauner MW, Grenier P, Mompoint D, et al. Pulmonary sar-
219. Staples CA, Gamsu G, Ray CS, et al. High resolution computed coidosis: evaluation with high-resolution CT. Radiology. 1989;
tomography and lung function in asbestos-exposed workers 172:467–471.
with normal chest radiographs. Am Rev Respir Dis. 1989;139: 241. Bergin CJ, Bell DY, Coblentz CL, et al. Sarcoidosis: corre-
1502–1508. lation of pulmonary parenchymal pattern at CT with
results of pulmonary function tests. Radiology. 1989;171: 268. Buschman DL, Gamsu G, Waldron JA, et al. Chronic hypersensi-
619–624. tivity pneumonitis: use of CT in diagnosis. AJR Am J Roentgenol.
242. Müller NL, Miller RR. Ground-glass attenuation, nodules, alve- 1992;159:957–960.
olitis, and sarcoid granulomas [Editorial]. Radiology. 1993; 269. Malinen AP, Erkinjuntti-Pekkanen RA, Partanen PL, et al. Long-
189:31–32. term sequelae of farmer’s lung disease in HRCT: a 14-year
243. Davies CW, Tasker AD, Padley SP, et al. Air trapping in sarcoido- follow-up study of 88 patients and 83 matched control farmers.
sis on computed tomography: correlation with lung function. Eur Radiol. 2003;13:2212–2221.
Clin Radiol. 2000;55:217–221. 270. Genereux GP. The Fleischner lecture: computed tomography of
244. Gleeson FV, Traill ZC, Hansell DM. Evidence of expiratory CT diffuse pulmonary disease. J Thorac Imaging. 1989;4:50–87.
scans of small-airway obstruction in sarcoidosis. AJR Am 271. Lacasse Y, Selman M, Costabel U, et al. Clinical diagnosis of hy-
J Roentgenol. 1996;166:1052–1054. persensitivity pneumonitis. Am J Respir Crit Care Med. 2003;168:
245. Hansell DM, Milne DG, Wilsher ML, et al. Pulmonary sarcoidosis: 952–958.
morphologic associations of airflow obstruction at thin-section 272. Glazer CS, Rose CS, Lynch DA. Clinical and radiologic manifes-
CT. Radiology. 1998;209:697–704. tations of hypersensitivity pneumonitis. J Thorac Imaging. 2002;
246. Abehsera M, Valeyre D, Grenier P, et al. Sarcoidosis with pul- 17:261–272.
monary fibrosis: CT patterns and correlation with pulmonary 273. Katzenstein AL, Fiorelli RF. Nonspecific interstitial pneumo-
function. AJR Am J Roentgenol. 2000;174:1751–1757. nia/fibrosis: histologic features and clinical significance. Am J
247. Padley SP, Padhani AR, Nicholson A, et al. Pulmonary sarcoido- Surg Pathol. 1994;18:136–147.
sis mimicking cryptogenic fibrosing alveolitis on CT. Clin Radiol. 274. Nagai S, Kitaichi M, Itoh H, et al. Idiopathic nonspecific inter-
1996;51:807–810. stitial pneumonia/fibrosis: comparison with idiopathic pul-
248. Sider L, Horton ES Jr. Hilar and mediastinal adenopathy in sar- monary fibrosis and BOOP. Eur Respir J. 1998;12:1010–1019.
coidosis as detected by computed tomography. J Thorac Imaging. 275. Nicholson AG, Colby TV, du Bois RM, et al. The prognostic
1990;5:77–80. significance of the histologic pattern of interstitial pneumonia
249. Hamper UM, Fishman EK, Khouri NF, et al. Typical and atypical in patients presenting with the clinical entity of cryptogenic
CT manifestations of pulmonary sarcoidosis. J Comput Assist fibrosing alveolitis. Am J Respir Crit Care Med. 2000;162:
Tomogr. 1986;10:928–936. 2213–2217.
250. Bergin C, Roggli V, Coblentz C, et al. The secondary pulmonary 276. Travis WD, Matsui K, Moss J, et al. Idiopathic nonspecific intersti-
lobule: normal and abnormal CT appearances. AJR Am tial pneumonia: prognostic significance of cellular and fibrosing
J Roentgenol. 1988;151:21–25. patterns. Am J Surg Pathol. 2000;24:19–22.
251. Ooi CG, Khong PL, Cheng RS, et al. The relationship between 277. Muller NL, White DA, Jiang H, et al. Diagnosis and manage-
mediastinal lymph node attenuation with parenchymal lung ment of drug-associated interstitial lung disease. Br J Cancer.
parameters in silicosis. Int J Tuberc Lung Dis. 2003;7:1199–1206. 2004;91(suppl 2):S24–S30.
252. Müller NL, Fraser RS, Colman N, et al. Radiologic diagnosis of 278. Cottin V, Donsbeck AV, Revel D, et al. Nonspecific interstitial
diseases of the chest. Philadelphia: Saunders; 2001. pneumonia: individualization of a clinicopathologic entity in a
253. Kim KI, Kim CW, Lee MK, et al. Imaging of occupational lung series of 12 patients. Am J Respir Crit Care Med. 1998;158:
disease. Radiographics. 2001;21:1371–1391. 1286–1293.
254. Bégin R, Bergeron D, Samson L, et al. CT assessment of silico- 279. Daniil ZD, Gilchrist FC, Nicholson AG, et al. A histologic pat-
sis in exposed workers. AJR Am J Roentgenol. 1987;148: tern of nonspecific interstitial pneumonia is associated with a
509–514. better prognosis than usual interstitial pneumonia in patients
255. Bégin R, Ostiguy G, Groleau S, et al. Computed tomographic with cryptogenic fibrosing alveolitis. Am J Respir Crit Care Med.
scanning of the thorax in workers at risk of or with silicosis. 1999;160:899–905.
Semin Ultrasound CT MR. 1990;11:380–392. 280. Hartman TE, Swensen SJ, Hansell DM, et al. Nonspecific inter-
256. Bergin CJ, Müller NL, Vedal S, et al. CT in silicosis: correlation stitial pneumonia: variable appearance at high-resolution chest
with plain films and pulmonary function tests. AJR Am CT. Radiology. 2000;217:701–705.
J Roentgenol. 1986;146:477–483. 281. Colby TV. Bronchiolitis: pathologic considerations. Am J Clin
257. Savranlar A, Altin R, Mahmutyazicioglu K, et al. Comparison of Pathol. 1998;109:101–109.
chest radiography and high-resolution computed tomography 282. Myers JL, Veal CF, Shin MS, et al. Respiratory bronchiolitis caus-
findings in early and low-grade coal worker’s pneumoconiosis. ing interstitial lung disease: a clinicopathologic study of six
Eur J Radiol. 2004;51:175–180. cases. Am Rev Respir Dis. 1987;135:880–884.
258. Ooi GC, Tsang KW, Cheung TF, et al. Silicosis in 76 men: quali- 283. Remy-Jardin M, Remy J, Boulenguez C, et al.. Morphologic
tative and quantitative CT evaluation: clinical-radiologic corre- effects of cigarette smoking on airways and pulmonary
lation study. Radiology. 2003;228:816–825. parenchyma in healthy adult volunteers: CT evaluation and cor-
259. Davis SD. CT evaluation for pulmonary metastases in patients relation with pulmonary function tests. Radiology. 1993;186:
with extrathoracic malignancy. Radiology. 1991;180:1–12. 107–115.
260. Hirakata K, Nakata H, Nakagawa T. CT of pulmonary metastases 284. Gruden JF, Webb WR. CT findings in a proved case of respiratory
with pathological correlation. Semin Ultrasound CT MR. 1995; bronchiolitis. AJR Am J Roentgenol. 1993;161:44–46.
16:379–394. 285. Holt RM, Schmidt RA, Godwin JD, et al. High resolution CT in
261. Kozuka T, Johkoh T, Hamada S, et al. Detection of pulmonary respiratory bronchiolitis-associated interstitial lung disease.
metastases with multi-detector row CT scans of 5-mm nominal J Comput Assist Tomogr. 1993;17:46–50.
section thickness: autopsy lung study. Radiology. 2003;226: 286. Carrington CB, Gaensler EA, Coute RE, et al. Natural history
231–234. and treated course of usual and desquamative interstitial pneu-
262. Bourke SJ, Dalphin JC, Boyd G, et al. Hypersensitivity pneumonia. N Engl J Med. 1978;298:801–809.
monitis: current concepts. Eur Respir J Suppl. 2001;32:81s–92s. 287. Katzenstein AL, Myers JL. Idiopathic pulmonary fibrosis: clinical
263. Cook PG, Wells IP, McGavin CR. The distribution of pulmonary relevance of pathologic classification. Am J Respir Crit Care Med.
shadowing in farmer’s lung. Clin Radiol. 1988;39:21–27. 1998;157:1301–1315.
264. Mindell HJ. Roentgen findings in farmer’s lung. Radiology. 288. Yousem SA, Colby TV, Gaensler EA. Respiratory bronchiolitis-
1970;97:341–346. associated interstitial lung disease and its relationship to
265. Yoshizawa Y, Ohtani Y, Hayakawa H, et al. Chronic hypersensi- desquamative interstitial pneumonia. Mayo Clin Proc. 1989;64:
tivity pneumonitis in Japan: a nationwide epidemiologic survey. 1373–1380.
J Allergy Clin Immunol. 1999;103:315–320. 289. Feigin DS, Friedman PJ. Chest radiography in DIP: a review of
266. Hansell DM, Moskovic E. High-resolution computed tomogra- 37 patients. AJR Am J Roentgenol. 1980;134:91–99.
phy in extrinsic allergic alveolitis. Clin Radiol. 1991;43:8–12. 290. Akira M, Yamamoto S, Hara H, et al. Serial computed tomo-
267. Silver SF, Müller NL, Miller RR, et al. Hypersensitivity pneu- graphic evaluation in desquamative interstitial pneumonia.
monitis: evaluation with CT. Radiology. 1989;173:441–445. Thorax. 1997;52:333–337.
291. Bouros D, Wells AU, Nicholson AG, et al. Histopathologic sub- 315. Prakash UBS, Barham SS, Carpenter HA, et al. Pulmonary alveo-
sets of fibrosing alveolitis in patients with systemic sclerosis lar phospholipoproteinosis: experience with 34 cases and a
and their relationship to outcome. Am J Respir Crit Care Med. review. Mayo Clin Proc. 1987;62:499–518.
2002;165:1581–1586. 316. Shah PL, Hansell D, Lawson PR, et al. Pulmonary alveolar pro-
292. Desai SR, Veeraraghavan S, Hansell DM, et al. CT features of teinosis: clinical aspects and current concepts on pathogenesis.
lung disease in patients with systemic sclerosis: comparison Thorax. 2000;55:67–77.
with idiopathic pulmonary fibrosis and nonspecific interstitial 317. Holbert JM, Costello P, Li W, et al. CT features of pulmonary
pneumonia. Radiology. 2004;232:560–567. alveolar proteinosis. AJR Am J Roentgenol. 2001;176:1287–1294.
293. Nicholson AG, Colby TV, Wells AU. Histopathological ap- 318. Epler GR, Colby TV, McLoud TC, et al. Bronchiolitis obliterans
proach to patterns of interstitial pneumonia in patient with organizing pneumonia. N Engl J Med. 1985;312:152–158.
connective tissue disorders. Sarcoidosis Vasc Diffuse Lung Dis. 319. Davidson AG, Heard BE, McAllister WC, et al. Cryptogenic
2002;19: 10–17. organizing pneumonitis. Q J Med. 1983;52:382–393.
294. Schurawitzki H, Stiglbauer R, Graninger W, et al. Interstitial 320. Myers JL, Colby TV. Pathologic manifestations of bronchiolitis,
lung disease in progressive systemic sclerosis: high-resolution constrictive bronchiolitis, cryptogenic organizing pneumonia
CT versus radiography. Radiology. 1990;176:755–759. and diffuse panbronchiolitis. Clin Chest Med. 1993;14:611–623.
295. Wells AU, Hansell DM, Corrin B, et al. High resolution com- 321. Müller NL, Guerry-Force ML, Staples CA, et al. Differential diag-
puted tomography as a predictor of lung histology in systemic nosis of bronchiolitis obliterans with organizing pneumonia
sclerosis. Thorax. 1992;47:508–512. and usual interstitial pneumonia: clinical, functional, and radi-
296. Bhalla M, Silver RM, Shepard JO, et al. Chest CT in patients ologic findings. Radiology. 1987;162:151–156.
with scleroderma: prevalence of asymptomatic esophageal di- 322. Akira M, Yamamoto S, Sakatani M. Bronchiolitis obliterans
latation and mediastinal lymphadenopathy. AJR Am J organizing pneumonia manifesting as multiple large nodules or
Roentgenol. 1993; 161:269–272. masses. AJR Am J Roentgenol. 1998;170:291–295.
297. Arakawa H, Yamada H, Kurihara Y, et al. Nonspecific interstitial 323. Gaensler EA, Carrington CB. Peripheral opacities in chronic
pneumonia associated with polymyositis and dermatomyositis: eosinophilic pneumonia: the photographic negative of pul-
serial high-resolution CT findings and functional correlation. monary edema. AJR Am J Roentgenol. 1977;128:1–13.
Chest. 2003;123:1096–1103. 324. Jederlinic PJ, Sicilian L, Gaensler EA. Chronic eosinophilic
298. Jurik AG, Davidsen D, Graudal H. Prevalence of pulmonary pneumonia: a report of 19 cases and a review of the literature.
involvement in rheumatoid arthritis and its relationship to Medicine. 1988;67:154–162.
some characteristics of the patients: a radiological and clinical 325. Ebara H, Ikezoe J, Johkoh T, et al. Chronic eosinophilic pneu-
study. Scand J Rheumatol. 1982;11:217–224. monia: evolution of chest radiograms and CT features. J Comput
299. Tanaka N, Kim JS, Newell JD, et al. Rheumatoid arthritis- Assist Tomogr. 1994;18:737–744.
related lung diseases: CT findings. Radiology. 2004;232:81–91. 326. Silva CIS, Müller NL, Fujimoto K, et al. Churg-Strauss syn-
300. Flaherty KR, Colby TV, Travis WD, et al. Fibroblastic foci drome: high-resolution CT and pathologic findings. J Thoracic
in usual interstitial pneumonia: idiopathic versus collagen vas- Imaging. 2005;20:74–80.
cular disease. Am J Respir Crit Care Med. 2003;167:1410–1415. 327. Glazer HS, Levitt RG, Shackelford GD. Peripheral pulmonary
301. Remy-Jardin M, Remy J, Cortet B, et al. Lung changes in rheuma- infiltrates in sarcoidosis. Chest 1984;86:741–744.
toid arthritis: CT findings. Radiology. 1994;193: 375–382. 328. Fraser RS, Colman N, Müller NL, et al. Synopsis of diseases of the
302. Rosenow EC 3rd. Drug-induced pulmonary disease. Dis Mon. chest. Philadelphia: Elsevier Saunders; 2005.
1994;40:253–310. 329. Baron SE, Haramati LB, Rivera VT. Radiological and clinical
303. Camus P, Kudoh S, Ebina M. Interstitial lung disease associated findings in acute and chronic exogenous lipoid pneumonia.
with drug therapy. Br J Cancer. 2004;91(suppl 2):S18–S23. J Thorac Imaging. 2003;18:217–224.
304. Cleverley JR, Screaton NJ, Hiorns MP, et al. Drug-induced lung 330. Kennedy JD, Costello P, Balikian JP, et al. Exogenous lipoid
disease: high-resolution CT and histological findings. Clin pneumonia. AJR Am J Roentgenol. 1981;136:1145–1149.
Radiol. 2002;57:292–299. 331. Abbott GF, Rosado-de-Christenson ML, Franks TJ, et al. From
305. Erasmus JJ, McAdams HP, Rossi SE. Drug-induced lung injury. the archives of the AFIP: pulmonary Langerhans cell histiocyto-
Semin Roentgenol. 2002;37:72–81. sis. Radiographics. 2004;24:821–841.
306. Rossi SE, Erasmus JJ, McAdams HP, et al. Pulmonary drug toxi- 332. Kambouchner M, Basset F, Marchal J, et al. Three-dimensional
city: radiologic and pathologic manifestations. Radiographics. characterization of pathologic lesions in pulmonary Langerhans
2000;20:1245–1259. cell histiocytosis. Am J Respir Crit Care Med. 2002;166:
307. Akira M, Ishikawa H, Yamamoto S. Drug-induced pneumonitis: 1483–1490.
thin-section CT findings in 60 patients. Radiology. 2002;224: 333. Howarth DM, Gilchrist GS, Mullan BP, et al. Langerhans cell
852–860. histiocytosis: diagnosis, natural history, management, and out-
308. Epler GR. Drug-induced bronchiolitis obliterans organizing come. Cancer. 1999;85:2278–2290.
pneumonia. Clin Chest Med. 2004;25:89–94. 334. Mogulkoc N, Veral A, Bishop PW, et al. Pulmonary Langerhans’
308a. Dodd JD, Lee KS, Johkoh T, Müller NL. Drug-associated cell histiocytosis: radiologic resolution following smoking ces-
organizing pneumonia: high-resolution CT findings in 9 sation. Chest. 1999;115:1452–1455.
patients. J Thorac Imaging. 2006;21:22–26. 335. Friedman PJ, Liebow AA, Sokoloff J. Eosinophilic granuloma of
309. Ellis SJ, Cleverley JR, Muller NL. Drug-induced lung disease: lung: clinical aspects of primary pulmonary histiocytosis in the
high-resolution CT findings. AJR Am J Roentgenol. 2000;175: adult. Medicine. 1981;60:385–396.
1019–1024. 336. Lacronique J, Roth C, Battesti JP, et al. Chest radiological fea-
310. Souza CA, Müller NL, Johkoh T, et al. Drug-induced tures of pulmonary histiocytosis X: a report based on 50 adult
eosinophilic pneumonia: high-resolution CT findings in 14 pa- cases. Thorax. 1982;37:104–109.
tients. AJR Am J Roentgenol. 2006;186:368–373. 337. Soler P, Bergeron A, Kambouchner M, et al. Is high-resolution
311. Bellamy EA, Husband JE, Blaquiere RM, et al. Bleomycin-re- computed tomography a reliable tool to predict the histopatho-
lated lung damage: CT evidence. Radiology. 1985;156:155–158. logical activity of pulmonary Langerhans cell histiocytosis? Am J
312. Padley SPG, Adler B, Hansell DM, et al. High-resolution com- Respir Crit Care Med. 2000;162:264–270.
puted tomography of drug-induced lung disease. Clin Radiol. 338. Vassallo R, Jensen EA, Colby TV, et al. The overlap between respi-
1992;46:232–236. ratory bronchiolitis and desquamative interstitial pneumonia in
313. Marchiori E, Souza AS Jr, Franquet T, et al. Diffuse high- pulmonary Langerhans cell histiocytosis: high-resolution CT, his-
attenuation pulmonary abnormalities: a pattern-oriented diag- tologic, and functional correlations. Chest. 2003;124:1199–1205.
nostic approach on high-resolution CT. AJR Am J Roentgenol. 339. Stern EJ, Webb WR, Golden JA, et al. Cystic lung disease associ-
2005;184:273–282. ated with eosinophilic granuloma and tuberous sclerosis: air
314. Trapnell BC, Whitsett JA, Nakata K. Pulmonary alveolar pro- trapping at dynamic ultrafast high-resolution CT. Radiology.
teinosis. N Engl J Med. 2003;349:2527–2539. 1992;182:325–329.
340. Bonelli FS, Hartman TE, Swensen SJ, et al. Accuracy of high- individuals with alpha-1 antitrypsin deficiency. Am J Respir Crit
resolution CT in diagnosing lung diseases. AJR Am J Roentgenol. Care Med. 2003;168:818–900.
1998;170:1507–1512. 364. Guest PJ, Hansell DM. High resolution computed tomography
341. Koyama M, Johkoh T, Honda O, et al. Chronic cystic lung dis- (HRCT) in emphysema associated with alpha-1-antitrypsin defi-
ease: diagnostic accuracy of high-resolution CT in 92 patients. ciency. Clin Radiol. 1992;45:260–266.
AJR Am J Roentgenol. 2003;180:827–835. 365. Shin MS, Ho KJ. Bronchiectasis in patients with alpha 1-anti-
342. Feuerstein I, Archer A, Pluda JM, et al. Thin-walled cavities, trypsin deficiency: a rare occurrence? Chest. 1993;104:1384–1386.
cysts, and pneumothorax in Pneumocystis carinii pneumonia: 366. Copley SJ, Wells AU, Muller NL, et al. Thin-section CT in
further observations with histopathologic correlation. Radiology. obstructive pulmonary disease: discriminatory value. Radiology.
1990;174:697–702. 2002;223:812–819.
343. Glassberg MK. Lymphangioleiomyomatosis. Clin Chest Med 367. Gevenois PA, Yernault JC. Can computed tomography quantify
2004;25:573–582, vii. pulmonary emphysema? Eur Respir J. 1995;8:843–848.
344. Niku S, Stark P, Levin DL, et al.. Lymphangioleiomyomatosis: 368. Kuwano K, Matsuba K, Ikeda T, et al. The diagnosis of mild
clinical, pathologic, and radiologic manifestations. J Thorac emphysema: correlation of computed tomography and pathol-
Imaging. 2005;20:98–102. ogy scores. Am Rev Respir Dis. 1990;141:169–178.
345. Hancock E, Osborne J. Lymphangioleiomyomatosis: a review of 369. Gurney JW, Jones KK, Robbins RA, et al. Regional distribution of
the literature. Respir Med. 2002;96:1–6. emphysema: correlation of high-resolution CT with pulmonary
346. Pallisa E, Sanz P, Roman A, et al. Lymphangioleiomyomatosis: function tests in unselected smokers. Radiology. 1992;183:
pulmonary and abdominal findings with pathologic correla- 457–463.
tion. Radiographics. 2002;22:S185–S198. 370. Klein JS, Gamsu G, Webb WR, et al. High-resolution CT diagno-
347. Kelly J, Moss J. Lymphangioleiomyomatosis. Am J Med Sci. sis of emphysema in symptomatic patients with normal chest
2001;321:17–25. radiographs and isolated low diffusing capacity. Radiology.
348. Chu SC, Horiba K, Usuki J, et al. Comprehensive evaluation of 1992;182:817–821.
35 patients with lymphangioleiomyomatosis. Chest. 1999;115: 371. Müller NL, Staples CA, Miller RR, et al. “Density mask”: an
1041–1052. objective method to quantitate emphysema using computed
349. Aberle DR, Hansell DM, Brown K, et al. Lymphangiomyomatosis: tomography. Chest. 1988;94:782–787.
CT, chest radiographic, and functional correlations. Radiology. 372. Gould GA, Macnee W, Mclean A, et al. CT measurements of
1990;176:381–387. lung density in life can quantitate distal airspace enlargement:
350. Kitaichi M, Nishimura K, Itoh H, et al. Pulmonary lymphangi- an essential defining feature of human emphysema. Am Rev
oleiomyomatosis: a report of 46 patients including a clinico- Resp Dis. 1988;137:380–392.
pathologic study of prognostic factors. Am J Respir Crit Care 373. Brown MS, McNitt-Gray MF, Goldin JG, et al. Automated meas-
Med. 1995;151:527–533. urement of single and total lung volume from CT. J Comput
351. Lim KE, Tsai YH, Hsu YL, et al. Pulmonary lymphangioleiomy- Assist Tomogr. 1999;23:632–640.
omatosis: high-resolution CT findings in 11 patients and 374. Kauczor H-U, Heussel CP, Fischer B, et al. Assessment of lung
compared with the literature. Clin Imaging. 2004;28:1–5. volumes using helical CT at inspiration and expiration: compar-
352. Sherrier RH, Chiles C, Roggli V. Pulmonary lymphangioleiomy- ison with pulmonary function tests. AJR Am J Roentgenol.
omatosis: CT findings. AJR Am J Roentgenol. 1989;153:937–940. 1998;171:1091–1095.
353. Raghu G, Mageto YN, Lockhart D, et al. The accuracy of the clin- 375. Gaensler EA, Jederlinic PJ, FitzGerald MX. Patient work-up for
ical diagnosis of new-onset idiopathic pulmonary fibrosis and bullectomy. J Thorac Imaging. 1986;13:75–93.
other interstitial lung disease: a prospective study. Chest. 376. Morgan MDL, Strickland B. Computed tomography in the as-
1999;116:1168–1174. sessment of bullous lung disease. Br J Dis Chest. 1984;78:10–25.
354. American Thoracic Society. Chronic bronchitis, asthma, and 377. Kazerooni EA, Whyte RI, Flint A, et al. Imaging of emphysema
pulmonary emphysema: statement by the Committee on and lung volume reduction surgery. Radiographics. 1997;17:
Diagnostic Standards for Nontuberculosis Respiratory Diseases. 1023–1036.
Am Rev Respir Dis. 1962;85:762–768. 378. Brenner M, Yusen R, McKenna R Jr, et al. Lung volume reduction
355. Snider GL, Kleinerman J, Thurlbeck WM, et al. The definition of surgery for emphysema. Chest. 1996;110:205–218.
emphysema: report of a National Heart, Lung, and Blood 379. Cooper JD, Trulock EP, Triantafillou AN, et al. Bilateral
Institute, Division of Lung Diseases workshop. Am Rev Respir pneumectomy (volume reduction) for chronic obstructive pul-
Dis. 1985;132:182–185. monary disease. J Thorac Cardiovasc Surg. 1995;109:106–116;
356. Muller NL, Coxson H. Chronic obstructive pulmonary disease, discussion 116–109.
4: imaging the lungs in patients with chronic obstructive pul- 380. Fishman A, Martinez F, Naunheim K, et al. A randomized trial
monary disease. Thorax. 2002;57:982–985. comparing lung-volume-reduction surgery with medical ther-
357. Miniati M, Filippi E, Falaschi F, et al. Radiologic evaluation of apy for severe emphysema. N Engl J Med. 2003;348:2059–2073.
emphysema in patients with chronic obstructive pulmonary 381. Meyers BF, Yusen RD, Guthrie TJ, et al. Results of lung volume
disease: chest radiography versus high resolution computed reduction surgery in patients meeting a national emphysema
tomography. Am J Respir Crit Care Med. 1995;151:1359–1367. treatment trial high-risk criterion. J Thorac Cardiovasc Surg.
358. Sutinen S, Christoforidis AJ, Klugh GA, et al. Roentgenologic 2004;127:829–835.
criteria for the recognition of nonsymptomatic pulmonary 382. Slone RM, Gierada DS. Radiology of pulmonary emphysema
emphysema: correlation between roentgenologic findings and and lung volume reduction surgery. Semin Thorac Cardiovasc
pulmonary pathology. Am Rev Respir Dis. 1965;91:69–76. Surg. 1996;8:61–82.
359. Thurlbeck WM, Simon G. Radiographic appearance of the chest 383. Hamacher J, Bloch KE, Stammberger U, et al. Two years’ out-
in emphysema. AJR Am J Roentgenol. 1978;130:429–440. come of lung volume reduction surgery in different morphologic
360. Foster WL, Gimenez EI, Roubidoux MA, et al. The emphysemas: emphysema types. Ann Thorac Surg. 1999;68:1792–1798.
radiologic-pathologic correlations. Radiographics. 1993;13: 384. Bae KT, Slone RM, Gierada DS, et al. Patients with emphysema:
311–328. quantitative CT analysis before and after lung volume reduction
361. Hruban RH, Meziane MA, Zerhouni EA, et al. High resolution surgery: work in progress. Radiology. 1997;203:705–714.
computed tomography of inflation fixed lungs: pathologic- 385. Holbert JM, Brown ML, Sciurba FC, et al. Changes in lung vol-
radiologic correlation of centrilobular emphysema. Am Rev ume and volume of emphysema after unilateral lung reduction
Respir Dis. 1987;136:935–940. surgery: analysis with CT lung densitometry. Radiology. 1996;
362. Miller RR, Müller NL, Vedal S, et al. Limitations of computed 201:793–797.
tomography in the assessment of emphysema. Am Rev Respir 386. Bense L, Lewander R, Eklund G, et al. Nonsmoking, non-alpha-
Dis. 1989;139:980–983. 1-antitrypsin deficiency induced emphysema in non-smokers
363. American Thoracic Society/European Respiratory Society with healing spontaneous pneumothorax, identified by com-
Statement. Standards for the diagnosis and management of puted tomography of the lungs. Chest. 1993;103:433–438.
387. Lesur O, Delorme N, Fromaget JM, et al. Computed tomogra- tion in patients with and without AIDS. Am Rev Respir Dis.
phy in the etiologic assessment of idiopathic spontaneous 1989;140:1204–1209.
pneumothorax. Chest. 1990;98:341–347. 410. DeLorenzo LJ, Huang CT, Maguire GP, et al. Roentgenographic
388. Franquet T, Muller NL, Gimenez A, et al. Infectious pulmonary patterns of Pneumocystis carinii pneumonia in 104 patients with
nodules in immunocompromised patients: usefulness of com- AIDS. Chest. 1987;91:323–327.
puted tomography in predicting their etiology. J Comput Assist 411. Barrio JL, Suarez M, Rodriguez JL, et al. Pneumocystis carinii
Tomogr. 2003;27:461–468. pneumonia presenting as cavitating and noncavitating solitary
389. Marchiori E, Muller NL, Soares Souza A Jr, et al. Pulmonary dis- pulmonary nodules in patients with the acquired immunodefi-
ease in patients with AIDS: high-resolution CT and pathologic ciency syndrome. Am Rev Respir Dis. 1986;134:1094–1096.
findings. AJR Am J Roentgenol. 2005;184:757–764. 412. Boiselle PM, Crans CA Jr, Kaplan MA. The changing face of
390. Catterall JR, McCabe RE, Brooks RG, et al.. Open lung biopsy in Pneumocystis carinii pneumonia in AIDS patients. AJR Am
patients with Hodgkin’s disease and pulmonary infiltrates. Am J Roentgenol. 1999;172:1301–1309.
Rev Respir Dis. 1989;139:1274–1279. 413. Sandhu JS, Goodman PC. Pulmonary cysts associated with
391. Ramsey PG, Rubin RH, Tolkoff-Rubin NE, et al. The renal trans- Pneumocystis carinii pneumonia in patients with AIDS. Radiology.
plant patient with fever and pulmonary infiltrates: etiology, 1989;173:33–35.
clinical manifestations, and management. Medicine. (Baltimore) 414. Boiselle PM, Tocino I, Hooley RJ, et al. Chest radiograph interpre-
1980;59:206–222. tation of Pneumocystis carinii pneumonia, bacterial pneumonia,
392. Murray JF, Mills J. Pulmonary infectious complications of and pulmonary tuberculosis in HIV-positive patients: accuracy,
human immunodeficiency virus infection. Am Rev Respir Dis. distinguishing features, and mimics. J Thorac Imaging. 1997;12:
1990;141:1356–1372. 47–53.
393. Rosenow EC 3rd, Wilson WR, Cockerill FR 3rd. Pulmonary dis- 415. Kennedy CA, Goetz MB. Atypical roentgenographic manifesta-
ease in the immunocompromised host. 1. Mayo Clin Proc. 1985; tions of Pneumocystis carinii pneumonia. Arch Intern Med.
60:473–487. 1992;152:1390–1398.
394. Barloon TJ, Galvin JR, Mori M, et al. High-resolution ultrafast 416. Gruden JF, Huang L, Turner J, et al. High-resolution CT in
chest CT in the clinical management of febrile bone marrow the evaluation of clinically suspected Pneumocystis carinii
transplant patients with normal or nonspecific chest pneumonia in AIDS patients with normal, equivocal, or non-
roentgenograms [see comments]. Chest. 1991;99:928–933. specific radiographic findings. AJR Am J Roentgenol. 1997;169:
395. Graham NJ, Müller NL, Miller RR, et al. Intrathoracic complica- 967–975.
tions following allogeneic bone marrow transplantation: CT 417. Richards PJ, Riddell L, Reznek RH, et al. High resolution com-
findings. Radiology. 1991;181:153–156. puted tomography in HIV patients with suspected Pneumocystis
396. Janzen DL, Padley SP, Adler BD, et al. Acute pulmonary complica- carinii pneumonia and a normal chest radiograph. Clin Radiol.
tions in immunocompromised non-AIDS patients: comparison 1996;51:689–693.
of diagnostic accuracy of CT and chest radiography. Clin Radiol. 418. McGuinness G, Gruden JF. Viral and Pneumocystis carinii infec-
1993;47:159–165. tions of the lung in the immunocompromised host. J Thorac
397. Kang EY, Staples CA, McGuinness G, et al. Detection and differ- Imaging. 1999;14:25–36.
ential diagnosis of pulmonary infections and tumors in patients 419. McGuinness G, Naidich DP, Garay SM, et al. AIDS associated
with AIDS: value of chest radiography versus CT. AJR Am bronchiectasis: CT features. J Comput Assist Tomogr. 1993;17:
J Roentgenol. 1996;166:15–19. 260–266.
398. Mori M, Galvin JR, Barloon TJ, et al. Fungal pulmonary infec- 420. Miller RF, Mitchell DM. AIDS and the lung: update 1995, 1:
tions after bone marrow transplantation: evaluation with radi- Pneumocystis carinii pneumonia. Thorax. 1995;50:191–200.
ography and CT. Radiology. 1991;178:721–726. 421. Kotloff RM, Ahya VN, Crawford SW. Pulmonary complications
399. Logan PM, Primack SL, Miller RR, et al. Invasive aspergillosis of solid organ and hematopoietic stem cell transplantation. Am
of the airways: radiographic, CT, and pathologic findings. J Respir Crit Care Med. 2004;170:22–48.
Radiology. 1994;193:383–388. 422. Soubani AO, Miller KB, Hassoun PM. Pulmonary complications
400. McGuinness G, Scholes JV, Garay SM, et al. Cytomegalovirus of bone marrow transplantation. Chest. 1996;109:1066–1077.
pneumonitis: spectrum of parenchymal CT findings with patho- 423. Nguyen Q, Estey E, Raad I, et al. Cytomegalovirus pneumonia
logic correlation in 21 AIDS patients. Radiology. 1994;192: in adults with leukemia: an emerging problem. Clin Infect Dis.
451–459. 2001;32:539–545.
401. Staples CA, Kang EY, Wright JL, et al. Invasive pulmonary 424. Drew WL. Cytomegalovirus disease in the highly active anti-
aspergillosis in AIDS: radiographic, CT, and pathologic findings. retroviral therapy era. Curr Infect Dis Rep. 2003;5:257–265.
Radiology. 1995;196:409–414. 425. Hoover DR, Saah AJ, Bacellar H, et al. Clinical manifestations of
402. Worthy SA, Flint JD, Müller NL. Pulmonary complications after AIDS in the era of Pneumocystis prophylaxis. N Engl J Med.
bone marrow transplantation: high-resolution CT and patho- 1993;329:1922–1926.
logic findings. RadioGraphics. 1997;17:1359–1371. 426. Leung AN, Gosselin MV, Napper CH, et al. Pulmonary infections
403. Janzen DL, Adler BD, Padley SP, et al. Diagnostic success of after bone marrow transplantation: clinical and radiographic
bronchoscopic biopsy in immunocompromised patients with findings. Radiology. 1999;210:699–710.
acute pulmonary disease: predictive value of disease distribu- 427. Olliff JF, Williams MP. Radiological appearances of cytomega-
tion as shown on CT. AJR Am J Roentgenol. 1993;160:21–24. lovirus infections. Clin Radiol. 1989;40:463–467.
404. Williams DM, Krick JA, Remington JS. Pulmonary infection in 428. Schulman LL. Cytomegalovirus pneumonitis and lobar consoli-
the compromised host: part I. Am Rev Respir Dis. 1976;114: dation. Chest. 1987;91:558–561.
359–394. 429. Franquet T, Lee KS, Muller NL. Thin-section CT findings in 32
405. Morris A, Lundgren JD, Masur H, et al. Current epidemiology immunocompromised patients with cytomegalovirus pneumo-
of Pneumocystis pneumonia. Emerg Infect Dis. 2004;10: nia who do not have AIDS. AJR Am J Roentgenol. 2003;181:
1713–1720. 1059–1063.
406. Thomas CF Jr, Limper AH. Pneumocystis pneumonia. N Engl 430. Moon JH, Kim EA, Lee KS, et al. Cytomegalovirus pneumonia:
J Med. 2004;350:2487–2498. high-resolution CT findings in ten non-AIDS immunocompro-
407. Godeau B, Coutant-Perronne V, Le Thi Huong D, et al. mised patients. Korean J Radiol. 2000;1:73–78.
Pneumocystis carinii pneumonia in the course of connective tis- 431. Kang EY, Patz EF Jr, Müller NL. Cytomegalovirus pneumonia in
sue disease: report of 34 cases. J Rheumatol. 1994;21:246–251. transplant patients: CT findings. J Comput Assist Tomogr.
408. Sepkowitz KA, Brown AE, Telzak EE, et al. Pneumocystis carinii 1996;20:295–299.
pneumonia among patients without AIDS at a cancer hospital. 432. Gasparetto EL, Escuissato DL, Inoue C, et al. Herpes simplex
JAMA. 1992;267:832–837. virus type 2 pneumonia after bone marrow transplantation:
409. Limper AH, Offord KP, Smith TF, et al. Pneumocystis carinii high-resolution CT findings in 3 patients. J Thorac Imaging.
pneumonia: differences in lung parasite number and inflamma- 2005;20:71–73.
433. Oikonomou A, Muller NL, Nantel S. Radiographic and high-res- 455. Mayo JR, Jackson SA, Müller NL. High-resolution CT of the
olution CT findings of influenza virus pneumonia in patients chest: radiation dose. AJR Am J Roentgenol. 1993;160:479–481.
with hematologic malignancies. AJR Am J Roentgenol. 2003;181: 456. Nishimura K, Izumi T, Kitaichi M, et al. The diagnostic accuracy
507–511. of high-resolution computed tomography in diffuse infiltrative
434. Denning DW. Invasive aspergillosis. Clin Infect Dis. 1998;26: lung diseases. Chest. 1993;104:1149–1155.
781–803; quiz 804–785. 457. Zisman DA, Lynch JP 3rd, Toews GB, et al. Cyclophosphamide
435. Klapholz A, Salomon N, Perlman DC, et al. Aspergillosis in the treatment of idiopathic pulmonary fibrosis: a prospective
in the acquired immunodeficiency syndrome. Chest. 1991;100: study in patients who failed to respond to corticosteroids. Chest.
1614–1618. 2000;117:1619–1626.
436. Franquet T, Muller NL, Gimenez A, et al. Spectrum of pulmonary 458. Flaherty KR, Mumford JA, Murray S, et al. Prognostic implica-
aspergillosis: histologic, clinical, and radiologic findings. tions of physiologic and radiographic changes in idiopathic
Radiographics. 2001;21:825–837 interstitial pneumonia. Am J Respir Crit Care Med. 2003;168:
437. Gefter WB. The spectrum of pulmonary aspergillosis. J Thorac 543–548.
Imaging. 1992;7:56–74. 459. Screaton NJ, Hiorns MP, Lee KS, et al. Serial high resolution CT
438. Thompson BH, Stanford W, Galvin JR, et al. Varied radiologic in non-specific interstitial pneumonia: prognostic value of the
appearances of pulmonary aspergillosis. Radiographics. 1995;15: initial pattern. Clin Radiol. 2005;60:96–104.
1273–1284. 460. Kim EY, Lee KS, Chung MP, et al. Nonspecific interstitial
439. Kuhlman JE, Fishman EK, Siegelman SS. Invasive pulmonary pneumonia with fibrosis: serial high-resolution CT findings
aspergillosis in acute leukemia: characteristic findings on CT, with functional correlation. AJR Am J Roentgenol. 1999;173:
the CT halo sign, and the role of CT in early diagnosis. Radiology. 949–953.
1985;157:611–614. 461. Remy-Jardin M, Giraud F, Remy J, et al. Pulmonary sarcoidosis:
440. Blum U, Windfuhr M, Buitrago-Tellez C, et al. Invasive pul- role of CT in the evaluation of disease activity and functional
monary aspergillosis. MRI, CT, and plain radiographic findings impairment and in prognosis assessment. Radiology. 1994;
and their contribution for early diagnosis. Chest. 1994;106: 191:675–680.
1156–1161. 462. Travis WD, Borok Z, Roum JH, et al. Pulmonary Langerhans cell
441. Primack SL, Hartman TE, Lee KS, et al. Pulmonary nodules and granulomatosis (histiocytosis X): a clinicopathologic study of
the CT halo sign. Radiology. 1994;190:513–515. 48 cases. Am J Surg Pathol. 1993;17:971–986.
442. Sonnet S, Buitrago-Tellez CH, Tamm M, et al. Direct detection 463. Naidich DP, McGuinness G. Pulmonary manifestations of
of angioinvasive pulmonary aspergillosis in immunosuppressed AIDS: CT and radiographic correlations. Radiol Clin North Am.
patients: preliminary results with high-resolution 16-MDCT 1991;29:999–1017.
angiography. AJR Am J Roentgenol. 2005;184:746–751. 464. Müller NL, Staples CA, Miller RR, et al. Disease activity in idio-
443. Kim MJ, Lee KS, Kim J, et al. Crescent sign in invasive pul- pathic pulmonary fibrosis: CT and pathologic correlation.
monary aspergillosis: frequency and related CT and clinical Radiology. 1987;165:731–734.
factors. J Comput Assist Tomogr. 2001;25:305–310. 465. Bensard DD, McIntyre RC Jr, Waring BJ, et al. Comparison
444. Caillot D, Casasnovas O, Bernard A, et al. Improved manage- of video thoracoscopic lung biopsy to open lung biopsy in
ment of invasive pulmonary aspergillosis in neutropenic the diagnosis of interstitial lung disease. Chest. 1993;103:
patients using early thoracic computed tomographic scan and 765–770.
surgery. J Clin Oncol. 1997;15:139–147. 466. Wall CP, Gaensler EA, Carrington CB, et al. Comparison of
445. Won HJ, Lee KS, Cheon JE, et al. Invasive pulmonary aspergillo- transbronchial and open biopsies in chronic infiltrative lung
sis: prediction at thin-section CT in patients with neutropenia: diseases. Am Rev Respir Dis. 1981;123:280–285.
a prospective study. Radiology. 1998;208:777–782. 467. Wilson RK, Fechner RE, Greenberg SD, et al. Clinical implica-
446. Brown MJ, Worthy SA, Flint JD, et al. Invasive aspergillosis in tions of a “nonspecific” transbronchial biopsy. Am J Med.
the immunocompromised host: utility of computed tomogra- 1978;65:252–256.
phy and bronchoalveolar lavage. Clin Radiol. 1998;53:255–257. 468. Raghu G. Idiopathic pulmonary fibrosis: a rational clinical
447. Huang RM, Naidich DP, Lubat E, et al. Septic pulmonary approach. Chest. 1987;92:148–154.
emboli: CT-radiographic correlation. AJR Am J Roentgenol. 469. Lenique F, Brauner MW, Grenier P, et al. CT assessment of
1989;153:41–45. bronchi in sarcoidosis: endoscopic and pathologic correlations.
448. Iwasaki Y, Nagata K, Nakanishi M, et al. Spiral CT findings in Radiology. 1995;194:419–423.
septic pulmonary emboli. Eur J Radiol. 2001;37:190–194. 470. Morales CF, Patefield AJ, Strollo PJ Jr, et al. Flexible trans-
449. Balakrishnan J, Meziane MA, Siegelman SS, et al. Pulmonary bronchial needle aspiration in the diagnosis of sarcoidosis.
infarction: CT appearance with pathologic correlation. J Comput Chest. 1994;106:709–711.
Assist Tomogr. 1989;13:941–945. 471. Gruden JF, Naidich DP. High-resolution CT: can it obviate
450. Staples CA, Müller NL, Vedal S, et al. Usual interstitial pneumo- biopsy? Clin Pulm Med. 1998;5:23–35.
nia: correlation of CT with clinical, functional, and radiologic 472. Hunninghake GW, Lynch DA, Galvin JR, et al. Radiologic find-
findings. Radiology. 1987;162:377–381. ings are strongly associated with a pathologic diagnosis of usual
451. Tung KT, Wells AU, Rubens MB, et al. Accuracy of the typical interstitial pneumonia. Chest. 2003;124:1215–1223.
computed tomographic appearances of fibrosing alveolitis. 473. Rose C, King TE. Controversies in hypersensitivity pneumonitis
Thorax. 1993;48:334–338. [editorial]. Am Rev Respir Dis. 1992;145:1–2.
452. Fujii M, Adachi S, Shimizu T, et al. Interstitial lung disease in 474. Swensen SJ, Aughenbaugh GL, Myers JL. Diffuse lung disease:
rheumatoid arthritis: assessment with high-resolution com- diagnostic accuracy of CT in patients undergoing surgical biopsy
puted tomography. J Thorac Imaging. 1993;8:54–62. of the lung. Radiology. 1997;205:229–234.
453. Padley SPG, Adler B, Müller NL. High-resolution computed 475. Orens JB, Kazerooni EA, Martinez FJ, et al. The sensitivity of
tomography of the chest: current indications. J Thorac Imaging. high-resolution CT in detecting idiopathic pulmonary fibrosis
1993;8:189–199. proved by open lung biopsy: a prospective study. Chest.
454. Remy-Jardin M, Remy J, Deffontaines C, et al. Assessment of dif- 1995;108:109–115.
fuse infiltrative lung disease: comparison of conventional CT
and high-resolution CT. Radiology. 1991;181:157–162.
Pleura, Chest Wall, 9

and Diaphragm
PLEURA 769 PLEURA

Computed Tomography Technique 769
Interlobar Fissures 772
Inferior Pulmonary Ligaments and Phrenic Nerves 778 Computed Tomography Technique
Pleural Surfaces and Adjacent Chest Wall 781 Multidetector CT (MDCT) scanning has simplified our
Pleural Effusions: Clinical Characterization 784 approach to imaging diseases related to the pleura, chest
Asbestos-Related Pleural Disease 803 wall, and diaphragm. The ability to acquire contiguous
Malignant Pleural Disease 810 thin (1 to 3 mm) sections throughout the chest in a single
Benign Fibrous Tumors of the Pleura 831 breath-hold has made visualization of otherwise difficult
anatomic regions, such as the diaphragm or thoracic
CHEST WALL 832 inlet, far easier to evaluate (3–6). The use of multiplanar
Sternum 832 coronal and sagittal reconstructions, in particular, has
Ribs 835 proved of value in assessing the relation between lesions
Axilla and Chest Wall 843 and fissures, for example, or evaluating the chest wall and
Magnetic Resonance Evaluation of the Pleura diaphragm after blunt trauma. The ability to generate con-
and Chest Wall 858 tiguous high-resolution axial images, either prospectively
or especially retrospectively, may also prove of value in
DIAPHRAGM 860 detecting subtle pleural and extrapleural abnormalities,
Anatomy 860 for example, in detecting asbestos-related pleural plaques
Computed Tomography Appearances 862 (7). In cases in which radiation exposure must be mini-
Diaphragmatic Defects 863 mized, it has been shown that low-dose (mAs of 80 to
Peridiaphragmatic Fluid Collections 867 100) technique is an acceptable method for evaluating
Diaphragm as Pathway for the Spread of Disease 871 pleural disease (8).
Magnetic Resonance Evaluation of the Diaphragm 874 As is extensively illustrated, the administration of
intravenous contrast can play an indispensable role,
especially in differentiating between pleural and paren-
Very early in its development, the value of computed chymal processes (9–12). In cases in which the major
tomography (CT) was recognized in assessing pleural and indication for CT is evaluation of complex pleuro-
chest-wall diseases (1,2). This reflects in part the wide parenchymal disease, a bolus of intravenous (IV) contrast
range of pathology that affects these areas, as well as the should be administered, with sections obtained volumet-
accepted limitations of chest radiography, especially in rically, when available. This technique allows optimal
the assessment of complex pleural and parenchymal dis- visualization of parenchymal vasculature during the
ease. In this chapter, the value and limitations of CT in the phase of maximal pulmonary artery and vein enhance-
assessment of diffuse and focal pleural disease, chest-wall ment. It has been suggested that images be ob-
lesions, and the diaphragm are discussed and illustrated. tained between 20 and 60 seconds after IV contrast
Potential applications of magnetic resonance (MR) and administration to optimize visualization of the paren-
especially fluorodeoxyglucose–positron emission tomog- chyma, specifically when the lung is consolidated or
raphy (FDG-PET) scanning also are addressed. collapsed (13).
Localization The cross-sectional appearance of pleural pathology,

especially loculated pleural fluid collections, will also
CT is of greatest efficacy in (a) confirming the presence of a be affected by pleural adhesions along the lesion
lesion; (b) determining its precise location and extent as margins. Pleurodesis restricts the mobility of the pleural
either primarily pulmonary or pleural; and (c) further layers; the result may be acute angulation between the
characterizing the nature of the pathology by means of pleural lesion and/or fluid, and the adjacent chest wall
attenuation coefficients. (Fig. 9-1).
Peripheral lesions are generally classified as extrapleural, Pulmonary parenchymal lesions, when peripheral,
pleural, or parenchymal, and are usually characterized radi- may abut the pleura; this typically results in acute angula-
ographically by the angle (either acute or obtuse) formed tion between the lesion and the chest wall. However, if
by the interface between the lesion and the adjacent pleura. sufficiently large, parenchymal lesions may result in ob-
Unfortunately, although CT is far superior to routine chest tuse angles between the lesion and the chest wall, usually
radiography in detecting the presence of pathology, consid- as the result of visceral and parietal pleural infiltration
erable overlap is found in the cross-sectional appearance of (Figs. 9-1 and 9-4) (16). It cannot be overemphasized
these lesions, as shown in Figure 9-1. that evaluation of peripheral lung lesions that only
Extrapleural lesions usually displace the overlying pari- abut pleural surfaces is extremely limited; the config-
etal and visceral pleura, resulting in an obtuse angle uration of peripheral lung cancer in relation to the
between the lesion and the chest wall (Fig. 9-2). Associated chest wall, for example, is of little use in determining
changes, such as rib destruction or muscle infiltration, whether histologic invasion of the pleura is present. As
help to confirm the site of origin as extrapleural, although is discussed in greater detail later, exclusion of tumor
these signs are often absent. Extrapleural lesions may pro- infiltration into the pleura or chest wall often requires
lapse into the adjacent lung, resulting in acute angulation biopsy.
between the lesion and the chest wall; however, this is Another potential pitfall in differentiating parenchy-
uncommon. mal from pleural disease is the presence of fluid within
Pleural lesions arising from the visceral or parietal pre-existing parenchymal cavities (17). As documented
pleura usually remain confined to the pleural space and by Zinn et al. (17), fluid within bullae may precisely
have a configuration similar to that of extrapleural lesions. mimic the appearance of a loculated pleural fluid collec-
However, pedunculated pleural lesions, especially those tion. In selected cases, differentiation may be possible
arising from the visceral pleura, are an important excep- only by reference to previous chest radiographs docu-
tion (14,15). As shown in Figure 9-1, these may prolapse menting the presence of prior bullous lung disease
or invaginate into the adjacent pulmonary parenchyma. If (Fig. 9-5). Therefore it is apparent that although the
the lesion is small, the appearance will mimic a peripheral, mechanics of pathology vary, the result is that consider-
subpleural parenchymal nodule; if the lesion is larger and able overlap may exist in the cross-sectional appearance
broad-based, the appearance of a pedunculated pleural of extrapleural, pleural, and peripheral subpleural
lesion may mimic larger intraparenchymal subpleural parenchymal lesions (Fig. 9-1).
lesions (Fig. 9-3).
Figure 9-1 Schematic drawing of the cross-sectional

appearance, typical and atypical, of extrapleural, pleural,
and peripheral parenchymal lesions. Extrapleural lesions dis-
place the overlying parietal and visceral pleura (a), resulting
in an obtuse angle between the lesion and the chest wall.
Associated chest-wall pathology (for example, rib erosion)
helps define the lesion as extrapleural. Pleural lesions gener-
ally remain confined between the layers of the pleura and
cause obtuse angles between the lesion and the chest wall
(b). Pleural lesions, however, may be pedunculated (c), in
which case, they may prolapse into the pulmonary
parenchyma, resulting in acute angles between the lesion
and the chest wall. Additionally, pleural fibrosis may result in
fusion of the parietal and visceral pleura (d), leading to
abnormal configurations of pleural lesions and/or loculation
of pleural fluid. Parenchymal lesions, if subpleural, abut the
pleural (e, f), resulting in acute angulation with the chest
wall. Clearly, although the mechanics vary, a considerable
overlap exists in the cross-sectional appearance of
extrapleural, pleural, and parenchymal lesions. In practical
terms, a biopsy should be performed on all peripheral soft
tissue lesions, regardless of site of origin.
B C
Figure 9-2 Rib metastasis in a patient with lung cancer. A: 5-mm unenhanced CT section demonstrates a lytic destructive lesion centered
around a right-sided rib. An associated large soft tissue extraosseous component is seen. A dependent effusion is present. Note the obtuse
angles between the lesion and the adjacent lung. B: Coronal 5-mm reconstruction performed from a 1-mm volumetric data set
retrospectively reconstructed on a multidetector scanner. Multiple right-sided destructive rib metastases with extraosseous soft tissue are
present. Pericardial metastases are also noted medially (arrow). C: A maximum intensity projection of the entire thoracic volume facilitates
visualization of multiple rib osseous abnormalities simultaneously. Cosmetic breast implants are incidentally present.
Figure 9-3 Benign fibrous tumor of the pleura. Section through the
left midlung demonstrates a well-defined tumor mass on the left side.
Note that, unusual for a lesion arising from the pleura, the lesion
forms acute margins with the chest wall. At surgery, this feature was
confirmed to be due to the pedunculated nature of the lesion with a
narrow focal pleural attachment. Pedunculation is not uncommon in
fibrous tumors of the pleura, accounting for lesion mobility and variable
anatomic location on radiographs and CT.
771
More recently, as shown by Nandalur et al. (24), in a retro-

spective evaluation of 145 pleural effusions, including
101 exudates and 44 transudates, although a significant
difference in CT density could be identified between exu-
dates and transudates [17.1 HU vs. 12.5 HU (p 0.001)],
the overall accuracy of attenuation values for identifying
exudates proved only moderate, with the most important
limitation due to overlap between transudates and exu-
dates in effusions in the 10- to 20-HU range.
In the setting of an acute bleed, CT may allow identifi-
cation of hemorrhagic pleural effusions (Fig. 9-7). Rarely,
CT also may be of value in detecting so-called milk of
calcium effusions resulting from chronic inflammation
Figure 9-4 Chest-wall invasion: bronchogenic carcinoma with (25). Although CT easily detects the presence of even small
extensive localized spread. Enhanced 5-mm section through the
lung apices demonstrates an enhancing mass invading the chest
pneumothoraces (Fig. 9-8), considerable limitations apply
wall, destroying the ribs, and eroding the medial aspect of the ver- in the evaluation of soft tissue masses, for which specific
tebral body. histologic diagnosis can be made only rarely. From a prac-
tical standpoint, a biopsy should be performed on any
Tissue Density Characteristics peripheral soft tissue mass, regardless of its probable site
of origin.
As compared with routine chest radiographs, a major value
of both CT and MRI is their ability to provide improved
contrast resolution. This has proven to be immensely valu-
Interlobar Fissures
able in assessing pleural pathology, especially in detecting
the presence of pleural fluid (18,19). Less commonly, CT The interlobar fissures represent invaginations of the
may help in differentiating a lipoma from a soft tissue visceral pleura, which, to a variable degree, separate the
mass or cyst (Fig. 9-6) (20). Unfortunately, although CT is pulmonary lobes. Knowledge of fissural anatomy is essen-
extremely accurate in detecting the presence of pleural tial in the localization and diagnosis of both pleural and
fluid, CT densitometry is of less clinical value in differenti- parenchymal abnormalities. Interlobar fissures can be
ating among the various etiologies of pleural effusions. localized on CT in nearly all cases, based on knowledge of
Specifically, CT numbers do not allow differentiation their anatomy, although their appearances vary depending
between transudative and exudative effusions and cannot on whether thick collimation (5 to 10 mm) or thin colli-
even be used reliably to detect chylous effusions (21–23). mation (1 to 3 mm) is used (26–31).
A B
Figure 9-5 Infected bulla. CT sections through the region of the lingula initially (A) and 3 months later after antibiotic therapy (B). On the
initial examination, a soft tissue neoplastic mass was suspected in the medial lingula, with possible areas of cavitation. However, after ther-
apy with antibiotics, complete clearance of this area was demonstrated, revealing a large paraseptal bulla that had become fluid filled and
infected (B). Emphysema with numerous other paraseptal and centrilobular bullae is evident. Fluid-filled bullae may also mimic the appear-
ance of empyema when they lie in an elongated subpleural distribution and should always be considered in patients with extensive bullous
disease.
Chapter 9: Pleura, Chest Wall, and Diaphragm 773
A
Figure 9-6 Pleural/Chest-wall lipoma. A: In a 52-year-old male patient, posteroanterior chest radiograph shows an ill-defined soft tissue
density in the right upper lung field. The obtuse angle margins to the lung parenchyma suggest a pleural or extrapleural lesion. The
presence of bilateral mid-thoracic thickening, gynecomastia, and a further lobulated density in the right apex suggests the possibility of
a pleural lipoma. B: A 5-mm noncontrast CT confirms the presence of pure fat-density pleural lesion invaginating into the chest wall. The
appearance is pathognomonic and requires no further diagnostic evaluation.
A B
Figure 9-7 Hemorrhagic effusions. A: Pleural hemorrhage. CT section shows a large right-sided pleural fluid collection within which a
clear fluid–fluid level is seen (arrows). The density within the dependent fluid collection measured within the range of acute hemorrhage,
subsequently verified via thoracentesis. B: Non–contrast-enhanced CT section through a different patient than in A, after trauma. A massive
hemothorax is identifiable as inhomogeneous areas of high contrast. Note that in this case, a marked shift of the mediastinum and a small
right pleural effusion are present.
A B
Figure 9-8 Pneumothorax: CT evaluation. A: CT section in a patient previously diagnosed with a spontaneous pneumothorax (PTX) shows
evidence of a residual large right PTX, despite the presence of a right-sided chest tube (arrow). In this case, the persistent air leak was due
to underlying bullous emphysema, identifiable as a cluster of cysts involving the medial aspect of the middle lobe (curved arrow). Not sur-
prisingly, these were not initially identifiable on the admission radiograph. B: CT section in an acquired immunodeficiency syndrome (AIDS)
patient shows ill-defined areas of ground-glass attenuation, especially in the left upper lobe. Note the presence of a subtle, “occult” PTX
(arrows). These two cases point out the potential contribution of CT both to detect and to characterize PTXs.
The Major (Oblique) Fissure incomplete laterally, with contiguity of the parenchyma of
the upper and middle lobes (32).
The major fissures serve to separate the lower lobe from the The minor fissure is rarely visible on routine axial
upper lobe on the left, and from the upper and middle images, as its position generally parallels the scan plane.
lobes on the right. However, many patients have incom- Characteristically, however, the position of the minor
plete major fissures, with some contiguity between the fissure can be inferred by identifying a broad avascular
parenchyma of adjacent lobes. In anatomic studies, the right band in the anterior portion of the right lung, anterior to
major fissure is complete in only 30% of patients (32,33), the major fissure, at the level of the bronchus intermedius
although in an additional 30%, near complete separation is (Fig. 9-9). This avascular region represents peripheral lung
present. The right major fissure is more often complete infe- on each side of the fissure, lying in or near the plane of
riorly; complete separation of the right lower and middle scan. The avascular band is often triangular, with one cor-
lobes is present in about 53% of patients (32,33). On the ner of the triangle at the pulmonary hilum, and the other
left, the major fissure has been reported to be complete in two laterally, but considerable variation in the appearance
27% to 60% of patients, although the extent of contiguity of the minor fissure can be seen owing to variations in its
between lobes is often minimal (32,33). orientation and curvature. Less frequently, the avascular
On CT obtained by using thin collimation or high-reso- region of the minor fissure appears rectangular or round or
lution technique, the major fissures usually appear as thin, elliptical. This appearance has been labeled the “right
well-defined lines, surrounded by a plane of avascular lung midlung window.” (28).
measuring about 1 cm in thickness (Fig. 9-9). If the fissure The appearance of the minor fissure on scans obtained
is invisible on high-resolution CT (HRCT), it may be by using thin-section or HRCT technique is quite variable
incomplete. In some cases, cardiac motion during the scan (Figs. 9-10 and 9-11) (26,35,36). Depending on its orienta-
results in a confusing artifact termed the “double-fissure” tion and contour, the fissure may appear as (a) a thick or
sign (34). When present, this artifact is easily identifiable, thin linear opacity, directed from anterior to posterior, and
as the fissure is visible in two locations on the same scan. either medial or lateral in location; (b) a thick or thin
Recognition of this artifact is most important in cases in linear opacity, extending from medial to lateral, paralleling
which detailed high-resolution imaging of the airways and and anterior to the major fissure; (c) an ill-defined opacity
parenchyma is performed. With MDCT scanners, it is now resulting from the fissure lying in the plane of scan; (d)
possible to obtain high-quality sagittal multiplanar recon- a circle or ring; or (e) a combination of these findings
structions through the entire length of the major fissures, (Figs. 9-10 and 9-11). Interestingly, in one study, the minor
in selected cases simplifying localization of parenchymal fissure appeared to be absent in 20% of cases and incom-
abnormalities. plete in 72% (35). Similar findings using 1.5-mm sections
have been reported by Frija et al. (26), who also found that
the minor fissure was incomplete in 76% of cases.
The Minor (Horizontal) Fissure
Recently it has become apparent that whereas the fis-
The minor or horizontal fissure separates the superior sures characteristically have smooth margins along most of
aspect of the right middle lobe from the right upper lobe. their length, it is common to identify focal areas of nodu-
Incompleteness of the minor fissure is common, occurring lar thickening, especially in the mid portions of the major
in 78% to 88% of patients. The minor fissure is most often fissures and anterior aspect of the minor fissure (Fig. 9-12)
A B
Figure 9-9 Normal pleural fissures. A, B: The 10-mm sec-

tions are through the bronchus intermedius and origin of the
middle lobe bronchus, respectively. The appearance on CT of the
fissures is variable, depending on their axis relative to the plane of
cross section. As shown in A and B, the major fissures most often
are identifiable as broad avascular bands within the pulmonary
parenchyma (straight arrows in A and B). The minor fissure
appears as a broad, triangular, or ovoid band in the anterior por-
tion of the right lung, typically located at the level of the bronchus
intermedius (curved arrows in A). C: A 1.5-mm-thick section at
the same level as shown in B. Note that with thin sections, the
C fissures now appear as sharply etched thin lines (arrows).
A B
Figure 9-10 Major and minor fissures; appearance on high-resolution CT. A–D: On sequential high-resolution CT images, the major
fissures (black arrows) and minor fissure (white arrows) are visible as thin white lines. The major fissures tend to bow anteriorly. The appear-
ance of the minor fissure is variable, depending on its orientation to the plane of scan. In this patient (A), the minor fissure outlines a roughly
circular portion of right middle lobe. RUL, right upper lobe; LUL, left upper lobe; RML, right middle lobe; RLL, right lower lobe; LLL, left
lower lobe. E: In another patient, the minor fissure (arrow) is incomplete, being visible only in its lateral aspect (continued ).
C D
E Figure 9-10 (continued)
B
Figure 9-11 Diagrammatic representation of the minor fissure, its orientation relative to the scan plane, and its appearance on high-
resolution CT. If the minor fissure often angles caudally (A), the major and minor fissures can have a similar appearance, with the major
fissure being posterior and the minor fissure anterior; in this situation, the lower lobe is most posterior, the upper lobe is most anterior, and
the middle lobe is in the middle. If the minor fissure is concave caudally (B), the minor fissure can be seen in two locations or can appear
ring-shaped or triangular, with the middle lobe between the fissure lines or in the center of the ring, and the upper lobe anterior to the most
anterior part of the fissure. RUL, right upper lobe; RML, right middle lobe, RLL, right lower lobe.
A, B C
D, E F
G H
Figure 9-12 Perifissural opacities. A–F: Magnified contiguous 1-mm images through the lower aspect of the right major fissure show
typical high-resolution CT appearance of focal fissural nodularity, most likely due to scarring with focal fibrosis. This appearance should not
be misinterpreted as a pulmonary nodule. G, H: Magnified views in a different patient from the one in A through F show characteristic
triangular scar in the lateral and inferior portion of the minor fissure (arrow in G). Although the fissure itself is not well seen on these images,
the configuration and location of this density are characteristic.
(37). These pseudotumors should not be mistaken for true region of the inferior pulmonary ligament to join the
nodules, as long-term follow-up CT scans fail to document major fissure.
any change in the vast majority of cases. Differentiation Other accessory fissures that have been described
between these pseudo fissural nodules and true lung nod- include (a) the left minor fissure, anatomically present in
ules is best accomplished with contiguous high-resolution approximately 15% of normal lungs, separating the ante-
1-mm images. rior segment of the left upper lobe from the lingula (Fig.
9-15) (41); (b) a left “azygos” fissure, analogous to the
more typical right azygos lobe, in which a malpositioned
Accessory Fissures
left superior intercostal vein is found (476); (c) the supe-
Any segment of lung can be separated from adjacent seg- rior accessory fissure, demarcating the superior segment
ments by an accessory fissure; numerous accessory fissures from the remainder of the lower lobes (40); and (d) a
have been identified (38–41). As many as 50% of lungs rudimentary fissure occasionally identifiable subtending
show some accessory fissure, although these are rarely of an anomalous cardiac bronchus (Fig. 9-16). Identification
clinical significance. of these accessory fissures rarely presents diagnostic diffi-
An azygos fissure and azygos lobe are most commonly culties; their recognition is of obvious benefit in identify-
seen, occurring in fewer than 0.4% of subjects. The azy- ing pathology related to the fissures (Fig. 9-13), including
gos fissure consists of four layers of pleura (two parietal loculated interfissural fluid collections. Although identifi-
and two visceral), as the azygos vein originates in an cation of benign processes involving both normal and
extrapulmonary location (Fig. 9-13) (39). The azygos fis- accessory fissures is generally straightforward, it is unfortu-
sure limits the lateral margin of the azygos lobe, which nate that identification of transfissural extension of tumor
frequently extends well behind the trachea and some- is more problematic. This is even more pertinent in cases
times the esophagus. The azygos fissure extends from the in which fissures are incomplete (Fig. 9-17).
brachiocephalic vein, anteriorly, to a position adjacent to
the right posterolateral aspect of the T4 or T5 vertebral
Inferior Pulmonary Ligaments
body; it is usually identifiable on CT as a thin, curved
line. An azygos lobe represents parts of the apical or pos-
and Phrenic Nerves
terior segments of the right upper lobe. Although the The inferior pulmonary ligaments represent reflections of
bronchial supply of an azygos lobe is variable, an apical the parietal pleura that extend from just below the inferior
or anterior subsegmental branch of the apical segment is margins of the pulmonary hila caudally and posteriorly to
always present (42). the diaphragm, and serve to anchor the lung to the medi-
The inferior accessory fissure separates the medial astinum (43,44). The ligament can terminate before reach-
basilar segment of either lower lobe from the remaining ing the diaphragm, or extend over the medial diaphrag-
basal segments. It is present anatomically in 30% to 45% matic surface. The inferior pulmonary ligament is often
of lobes, although it is less commonly seen on CT (Fig. contiguous with the intersublobar septum, dividing the
9-14) (40). It extends laterally and anteriorly from the medial from the posterior basal lung segments.
A B
Figure 9-13 Azygos fissure: 5-mm sections. Note that the azygos fissure courses from the lateral aspect of the right brachiocephalic vein
anteriorly to the right superior intercostal vein posteriorly. As the fissure contains four pleural layers, it is normally thicker than the two-
pleural-layer major and minor fissures—compare with the contralateral major fissure (A). The fissure is thickest inferiorly where the actual
azygos vein runs (B).
A B
C D
Figure 9-14 Right inferior accessory fissure. Medial basilar segment pneumonia. A: Chest radiograph (CXR) in a febrile patient demon-
strates a focus of infective consolidation silhouetting the diaphragm, with a well-demarcated lateral margin. After antibiotic therapy and
consolidation clearance, note the typical obliquely oriented inferior accessory fissure demonstrated on repeat CXR (arrow) (B). A 5-mm axial
CT (C) with 5-mm coronal multiplanar reconstructions (MPRs) (D) performed for persistent symptoms confirm the typical thin curved fissure
line separating the medial basilar segment from the remainder of the basilar segments without residual lung pathology (arrow).
A B
Figure 9-15 Left minor fissure: 1-mm noncontiguous sections demonstrate a left-sided minor fissure (A) analogous to the right minor
fissure slightly more inferiorly (B, arrow). Left-sided minor fissures are variable in appearance and frequently incomplete. Their presence may
alter the chest radiograph and CT appearances of pathology; in particular, the classic appearances of left upper lobe collapse may become
a mirror image of right upper lobe collapse.
In a review of 100 CT studies using 10-mm sections, inferior margins of these ligaments are variable; in their
Cooper et al. (45) identified at least one of these ligaments most caudal extension, they may assume a triangular config-
in 42% of cases. With 10-mm sections, Rost and Proto (44) uration as they reflect onto the diaphragm (Fig. 9-18).
could identify the left pulmonary ligament in 67% of cases It should be emphasized that the inferior pulmonary
and the right inferior pulmonary ligament in 37% of cases. ligaments must be distinguished from the phrenic nerves,
It is expected that with use of high-resolution imaging, these which may run nearby (Figs. 9-18 and 9-19). Although
ligaments can be identified if desired in nearly all cases. The the left phrenic nerve generally can be identified as a 1- to
Figure 9-17 Incomplete fissures: CT detection. High-resolution

Figure 9-16 Accessory fissure: 5-mm coronal multiplanar recon- image through the carina demonstrates that the fissures are
struction demonstrates an accessory fissure (arrow) dividing a incomplete bilaterally. A small amount of linear atelectasis is inci-
right lower lobe superomedial rudimentary lobe supplied by an dentally noted in the superior segment of the right lower lobe.
unusually patent cardiac bronchus from the rest of the right lower The fissures may be incomplete in 30% to 40% of patients and can
lobe. The use of multidetector CT has enabled us to visualize many account for variations in the spread of both neoplastic and benign
accessory fissures and airways that often have no set defined disease and their radiographic manifestations. The presence of in-
terminology. (Courtesy of Jane P. Ko, MD, New York University, complete fissures is a contributory factor to the maintenance of
New York.) aeration in lobes whose airways are proximally occluded.
A B, C
Figure 9-18 A–C: High-resolution CT scans at 1-cm intervals showing the left inferior pulmonary ligament. A: The left inferior pulmonary
vein and left major fissure are visible at the most cephalad level. B, C: The left inferior pulmonary ligament is visible as a triangular structure
arising immediately caudal to the inferior pulmonary vein. This relation is characteristic. A thin soft tissue septum within the adjacent lung
represents the intersublobar septum.
3-mm rounded structure lying adjacent to the pericard- External to the parietal pleura is a layer of loose areolar
ium, identification of the right phrenic nerve typically connective tissue, which separates the parietal pleura from
proves more difficult (46). The relation between the right the endothoracic fascia. This fatty layer averages 250 mm
inferior pulmonary ligament and right phrenic nerve has in thickness in most locations (49,50), but can be
been examined in detail (47). With anatomic specimens, it markedly thickened over the lateral or posterolateral ribs,
has been shown that the right inferior pulmonary liga- resulting in extrapleural fat pads several millimeters in
ment appears as a thin, high-attenuation line frequently thickness (53,54). The thoracic cavity is lined by the fibro-
identifiable above or at the level of the diaphragm, usually elastic endothoracic fascia (approximately 250 mm in
extending from the region of the esophagus. Accurate thickness) (49,50), which covers the surface of the inter-
identification of the inferior pulmonary ligaments is costal muscles and intervening ribs, blends with the peri-
important, as alterations in the normal appearance of chondrium and periosteum of the costal cartilages and
these ligaments typically are produced by pleural effu- sternum anteriorly, and posteriorly, and is continuous
sions, pneumothoraces, and lobar collapse (35,40,48). with the prevertebral fascia, which covers the vertebral
bodies and intervertebral disks.
External to the endothoracic fascia are a layer of fatty
Pleural Surfaces and Adjacent Chest Wall
connective tissue and the three layers of the intercostal
A number of structures, arranged in layers, surround muscles. The innermost intercostal muscle (intercostales
the lung and line the inner aspect of the thoracic cavity intimi) passes between the internal surfaces of adjacent
(Fig. 9-20). Some can be identified on CT, and knowledge ribs and is relatively thin; it is separated from the inner
of their anatomy is helpful in understanding normal CT and external intercostal muscles by the intercostal vessels
findings and the appearances of pleural diseases. and nerve. Although the innermost intercostal muscles are
The combined thickness of the layers of visceral and incomplete in the anterior and posterior thorax, other
parietal pleura that surround the lung, and the fluid- muscles (the transversus thoraces and subcostalis) can
containing pleural space, is approximately 200 to 400 mm occupy the same relative plane and are considered by some
(0.2 to 0.4 mm). The parietal pleura consists of four layers authors to be parts of the innermost intercostal muscles.
and measures approximately 100 to 200 mm; the visceral Anteriorly, the transversus thoracis muscle consists of four
pleura is anatomically similar to but somewhat thicker or five slips, which arise from the xiphoid process or lower
than the parietal pleura (49,50). The width of the pleural sternum and pass superolaterally from the second to sixth
space, in studies of frozen thoraces, has been estimated to costal cartilages. The internal mammary vessels lie external
be 10 to 20 mm (51,52). to the transversus thoracis.
A B
C D
Figure 9-19 Spiral CT in two patients showing the relationsships between the inferior pulmonary ligaments, phrenic nerves, and major
fissures. A–C: The inferior pulmonary ligaments arise below the level of the inferior pulmonary veins (A), and extend inferiorly (B and C),
lying near the esophagus. The right phrenic nerve (C) results in a similar opacity anterior to the inferior pulmonary ligament, adjacent to the
inferior vena cava. The right major fissure is visible more anteriorly. D: In another patient, two adjacent scans through the left lung show the
location of the inferior pulmonary ligament (IPL), and the pleural reflections associated with the left phrenic nerve (LPN), just above the left
hemidiaphragm. The major fissure is more anterior.
Posteriorly, the subcostal muscles are thin, variable Similarly, on conventional CT, a visible soft tissue stripe,
muscles, which extend from the inner aspect of the angle internal to a rib, is used to diagnose pleural thickening or
of the lower ribs, crossing one or two ribs and intercostal effusion.
spaces, to the inner aspect of a rib below. Normal extrapleural fat can be seen on HRCT internal to
ribs in several locations, and can mimic pleural thickening
(Fig. 9-23) (7,53,54). The normal layer of extrapleural fat
Computed Tomography Appearances
between the parietal pleura and the endothoracic fascia is
On HRCT in normal individuals, a 1- to 2-mm-thick soft tis- notably thicker adjacent to the lateral ribs than in other
sue density line is visible in the anterolateral and posterolat- sites (7,53,54). It is most abundant over the posterolateral
eral intercostal spaces, at the point of contact between lung fourth to eighth ribs and can result in fat pads several mil-
and chest wall (Figs. 9-21 and 9-22). This line primarily limeters thick, which extend into the intercostal spaces. In
represents the innermost intercostal muscle but also reflects normal subjects, these costal fat pads can be difficult to dis-
the combined thicknesses of visceral and parietal pleura, the tinguish from pleural thickening or plaques when extended
fluid-filled pleural space, endothoracic fascia, and fat layers. window settings (width, 2,000 HU) are used, but are very
Although the pleurae, extrapleural fat, and endothoracic low in attenuation and difficult to see with soft tissue win-
fascia pass internal to the ribs, they are not visible in this dow settings. In one study, extrapleural fat could be seen by
location on HRCT unless they are abnormally thickened. using extended windows in 12 of 15 normal subjects (7).
Figure 9-20 A, B: Diagrammatic represen-

tations of structures of the chest wall and the
pleural surface. All structures internal to the
innermost intercostal muscle pass internal to
B the ribs.
Figure 9-21 Normal CT of the pleural surface and chest wall.

On a high-resolution CT scan, a thin line visible in the intercostal
spaces (arrows) primarily represents the innermost intercostal
muscle, but is also made up of the combined thicknesses of
visceral and parietal pleura, the fluid-filled pleural space, endotho-
racic fascia, and fat layers. Although the pleurae, extrapleural
fat, and endothoracic fascia pass internal to the ribs, they are not
visible in this location on high-resolution CT unless they are abnor-
mally thickened.
A B
Figure 9-22 CT appearance of the normal pleura and chest wall. A: On a section of a cadaver, the parietal pleura and endothoracic fascia
are visible as a white line (white arrows) at the internal surface of the chest wall. The innermost intercostal muscle (black arrow) lies external
to these. Although the pleura and fascia pass internal to the ribs, they are not visible in this location on high-resolution CT (HRCT). In the
paravertebral regions, the internal intercostal muscle is absent. Intercostal vessels (open arrow) lie within the paravertebral fat. B: HRCT
of the cadaver at the same level, photographed by using mediastinal window settings. A 1-mm to 2-mm line (white arrows) at the internal
surface of the chest wall, in the intercostal spaces, represents the combined thicknesses of the parietal pleura, endothoracic fascia, and
the innermost intercostal muscle. In a paravertebral location, the innermost intercostal muscle is absent, and the combined parietal pleura
and fascia are invisible or result in a very thin line. Intercostal vessels (open arrows) are visible within the fat external to the intercostal
muscle or fascia.
The transversus thoracis and subcostalis muscles can thickening (Figs. 9-25 and 9-26). Continuity of these
also mimic pleural thickening in some patients. Anteriorly, opacities with the azygos or hemiazygos veins can
at the level of the heart and adjacent to the lower sternum sometimes allow them to be correctly identified (7).
or xiphoid process, the transversus thoracis muscles are Furthermore, when viewed by using lung window
nearly always visible internal to the anterior ends of ribs or settings, intercostal vein segments do not indent the lung
costal cartilages (Fig. 9-24) (7). Posteriorly, at the same surface; pleural plaques of the same thickness nearly
level, a 1- to 2-mm-thick line is sometimes seen internal to always will.
one or more ribs, representing the subcostalis muscle; this
muscle is present in only a small percentage of patients
Pleural Effusions: Clinical Characterization
(Fig. 9-25) (7). In contrast to pleural thickening, these
muscles are smooth, uniform in thickness, and symmetric Pleural effusions have traditionally been defined as either
bilaterally. transudative or exudative (10,55). Transudative effusions
Segments of intercostal veins are commonly visible in result from systemic abnormalities that lead to an imbal-
the paravertebral regions and can mimic focal pleural ance in hydrostatic and osmotic forces, resulting in the
A B
Figure 9-23 Normal extrapleural fat. A: A cadaver section shows a fat pad (large arrows) internal to the posterior rib. More laterally, only
a thin white stripe (small arrow) is seen internal to rib, representing the parietal pleura and endothoracic fascia. B: High-resolution CT of the
cadaver by using a wide window width. The fat pad (arrow), internal to the most posterior rib segment, is several millimeters thick. Note
that fat pads are not visible internal to more lateral ribs.
Figure 9-24 High-resolution CT in a normal subject shows the

transversus thoraces muscles (large arrows). These lie internal to
the innermost intercostal muscles and internal mammary vessels
(small arrows).
formation of protein-poor pleural fluid (Figs. 9-27 to 9-30). criteria: (a) pleural fluid/serum total protein ratio greater
Common causes include congestive heart failure, cirrhosis, than 0.5; (b) pleural fluid lactate dehydrogenase
and nephrotic syndrome. In distinction, exudative effusions (LDH)/serum LDH ratio greater than 0.6; or (c) pleural
are usually associated with conditions that result either in fluid LDH greater than two thirds of the upper limit of
increased permeability of abnormal pleural capillaries or normal for serum LDH. Transudative effusions are those
in lymphatic obstruction with resultant formation of pro- that do not meet these criteria. With these criteria,
tein-rich fluid (Fig. 9-31). Although the list of causes of an although the overall sensitivity has been reported to be as
exudative effusion is quite long, most important are those high as 98%, specificity has been reported to be only
caused by infection and malignancy (55). 82% (55). One cause of lower specificity is that transuda-
Traditionally, differentiation between transudative and tive effusions in patients with congestive heart failure
exudative effusions has usually required thoracentesis. may be misinterpreted as exudative, especially after
With original criteria established by Light et al. (56), diuresis. Alternative measurements have been evaluated,
exudative effusions must meet one of the following including quantifying pleural fluid cholesterol and
Figure 9-26 Normal enhancing intercostal arteries and veins

Figure 9-25 Normal subcostalis and transversus thoraces mus- adjacent to bilateral simple effusions, mimicking pleural enhance-
cles. High-resolution CT in a normal subject at the level of the ment. These structures may be normally visualized along the
heart shows a well-defined linear opacity (large arrow), internal to posteromedial surface of the extrapleural space, most prominently
the posterior rib, and separated from it by a fat pad. This repre- on studies acquired early in arterial enhancement, such as this
sents the subcostalis muscle. Its location and appearance are char- 2-mm collimation CT pulmonary angiography (CTPA) study with
acteristic. Anteriorly (small arrow), the transversus thoracic muscle clot in the left lower lobe artery. As opposed to true pleural
is visible passing internal to calcified costal cartilage. Paravertebral enhancement, these structures are irregular, discontinuous, and do
vein segments are also seen, joining azygos and hemiazygos veins. not indent the pleural surface.
A B
Figure 9-27 Transudative effusions. A: Contrast-enhanced CT

shows simple left-sided transudative effusion. Note that neither
the parietal nor visceral pleura is thickened and that the effusion
has a typically meniscoid configuration conforming to the shape of
the posterior pleural space. B: Contrast-enhanced CT in a different
patient from the one in A shows a larger effusion, this time causing
marked compression atelectasis of the right lower lobe (arrow).
Despite the large size of this effusion, again note lack of enhance-
ment of the pleural surfaces, consistent with a transudative effu-
sion. C: Contrast-enhanced CT section in a different patient from
that in A or B again shows a large right-sided effusion. In this case,
the effusion actually extends to the left hemithorax (arrow), pre-
sumably the result of fusion of pleural membranes with resultant
direct communication between the right and left pleural spaces.
Despite the large size of this effusion, there is no evidence of un-
derlying compression atelectasis (compare with B). Again note lack
of evidence of either visceral or parietal pleural thickening in this
patient with a transudative effusion. Although an absence of pleu-
ral thickening is compatible with either a transudative or exudative
effusion (including those associated with metastatic disease), this
C effectively eliminates an empyema.
A B
Figure 9-28 Fissural pseudotumor: CT appearance. A, B: Sequential CT scans through the middle thorax show typical appearance of a
loculated pseudotumor in the superior portion of the right major fissure. Characteristically, these conform to the expected position of the
fissure, the posterior margin of which can be identified superiorly (arrow in A).
A B
Figure 9-29 Loculated effusions: CT evaluation. A, B: 5-mm unenhanced images in a patient with lung cancer. The presence of discrete
pleural collections with intervening areas of normal or thickened pleura without fluid indicates loculation, a feature of exudative effusions.
Significant intrafissural or antidependent collections also indicate the complex nature of fluid. CT-guided drainage may be advantageous in
these cases to avoid tethered atelectatic lung and guide catheter placement into the largest, most interconnecting locule.
bilirubin and assessing the serum–pleural fluid albumin lung abscess, and are further subdivided as either simple
gradient; to date, most of these have not proved of rou- or complicated. Complicated parapneumonic effusions
tine clinical use (57). are usually defined by one of the following criteria: (a)
In addition to being classified as either transudative or elevated pleural LDH greater than 1,000; (b) low pH
exudative, effusions are also defined as either parapneu- (usually 7.1); or (c) low pleural glucose (40 mg/dL). In
monic or as an empyema (Fig. 9-32). Parapneumonic distinction, effusions are considered empyemas only when
effusions are exudative effusions associated with an under- cultures are positive.
lying parenchymal infection, most often pneumonia or a The incidence of exudative parapneumonic effusions is
dependent to some degree on the infecting organism,
ranging from about 10% for pneumonias caused by
Streptococcus pneumoniae to more than 50% for those
caused by Staphylococcus pyogenes (58). Regardless of the
infecting organism, parapneumonic effusions have been
noted to follow a natural history. Three stages have been
described, each of which is pathophysiologically and ther-
apeutically distinct (59).
1. Exudative or simple parapneumonic stage. In this stage, an
underlying pneumonic process causes inflammation of
the visceral pleura, resulting in the accumulation of thin,
uninfected pleural fluid, usually the result of increased
capillary permeability, with resultant protein loss. A
thoracentesis at this stage reveals a simple exudative
parapneumonic effusion. Such uncomplicated or simple
exudative parapneumonic effusions resolve without
drainage, provided that the underlying cause of infection
Figure 9-30 Pleural adhesions: CT evaluation. Patient with trans-
udative effusions associated with heart failure after valvular is adequately treated with appropriate antibiotics (10).
surgery. The presence of pleural adhesions does not necessarily 2. Fibrinopurulent or complicated parapneumonic stage. In this
indicate an exudative effusion in the absence of pleural enhance- stage, large numbers of polymorphonucleocytes (PMNs)
ment or thickening. In this case, the atelectatic middle lobe and
right lower lobe are tethered posteriorly by a pre-existing postin- and bacteria accumulate in the pleural space, and sheets
flammatory adhesion, whereas the atelectatic “free” left lower of fibrin are deposited over the visceral and parietal
lobe floats on the left effusion. The presence of pleural adhesions pleura. As a consequence, a progressive tendency toward
may contribute to chronic atelectasis and scarring, including the
nonexpanding “trapped” lung that occasionally occurs after pro- fluid loculation exists as fluid resorption is impaired,
longed atelectasis. presumably because of decreased lymphatic drainage
A B
Figure 9-31 Exudative effusion: metastatic breast cancer. A: Contrast-enhanced CT in a patient with known metastatic breast cancer.
A right-sided effusion is present, associated with mild thickening and enhancement of the parietal pleura (arrows; compare with Fig. 9-27).
Note presence of minimally enlarged mediastinal nodes. The finding of thickened parietal pleura is indicative of an exudative effusion and
may be due to an empyema. B: Exudative effusion from metastatic breast carcinoma in a different patient demonstrates that pleural
thickening, suggestive of the exudative nature of an effusion, may be apparent on unenhanced images alone, although the sensitivity and
specificity of this finding increase with intravenous contrast administration.
A B
Figure 9-32 Complicated parapneumonic effusions. A: Contrast-

enhanced CT section shows loculated right effusion, causing
compression atelectasis of adjacent portions of the right middle and
lower lobes. In this case, smooth enhancement of both the parietal
(arrow) and visceral (curved arrow) pleura gives rise to the so-called
“split pleura” sign. Thoracentesis in this case confirmed that this was
a complicated parapneumonic effusion requiring tube drainage.
B: Contrast-enhanced CT in a different patient from that in A shows
similar if more extensive changes than in A. A massive left-sided
effusion is present with evidence of uniformly thickened and enh-
ancing parietal pleura. Although thoracentesis in this case was
consistent with a simple parapneumonic effusion (pH, 7.32; protein,
5; glucose, 97; and LDH, 432), sputum cultures proved positive
for Mycobacterium avium-intercellulare; subsequent pleural biopsy
revealed noncaseating granulomas. C: Contrast-enhanced section in
a different patient from that in A or B again demonstrates marked
pleural thickening and enhancement, resulting in the split pleura
appearance. In this patient with an incidental marked pectus excava-
tum, the chronicity of this empyema is reflected by the presence of
extrapleural fat noted not only posteromedially but also most
C notably centrally over the right hemidiaphragm.
(Fig. 9-32). If the underlying infection is not adequately stage may be quite rapid, occasionally occurring in the
treated, the pleural fluid becomes infected. Furthermore, course of therapy with closed pleural tube drainage. In
there is a tendency for progressive thickening of the most cases, however, the organizing phase typically
extrapleural subcostal tissues to develop as well, pre- occurs within 2 to 3 weeks after initial pleural fluid for-
sumably secondary to spread of infection and edema to mation. Eventually, especially if the infection is inade-
the adjacent chest wall tissues (Figs. 9-32 and 9-33). A quately treated, the pleura may calcify, appearing initially
thoracentesis performed at this stage reveals features as small, punctate foci involving both the visceral and
characteristic of a complicated parapneumonic effusion parietal pleura, progressing to form a calcified rind of
as defined earlier. There is a tendency toward an increas- pleura (Fig. 9-33). Often the result of tuberculous
ing white blood cell count, decreasing glucose levels, empyema in the past after pneumothorax therapy, tuber-
and a decreased pH. Although controversial, in many culosis of the pleura still remains problematic, even in
centers, complicated parapneumonic effusions are the era of antituberculous chemotherapy (60,61). Re-
treated with immediate closed-tube drainage. gardless of the etiology, with the formation of a fibrotho-
3. Organizing stage. In this stage, there is an ingrowth of rax, contraction of the involved hemithorax and an
fibroblasts along the fibrin sheets lining the visceral and expansion of the extrapleural fat occur. The underlying
parietal pleura. The result is pleural fibrosis, which acts as lung can no longer expand. Expansion of the extrapleural
an inelastic membrane trapping the adjacent lung (10). fat is especially characteristic and may be accompanied by
Progression from the fibrinopurulent to the organizing periosteal changes in the adjacent ribs (Fig. 9-33).
A B
C D
Figure 9-33 Chronic tuberculous empyema. A: Contrast-enhanced section shows thickening and calcification of the pleura associated
with marked expansion of the extrapleural space with fat (black arrow). Thickening of the cortex of the adjacent ribs and marked volume
loss of the ipsilateral hemithorax are apparent. These findings are all characteristic of chronic pleural and periosteal inflammation, in this
case, caused by tuberculosis. Note the small residual collection of fluid within the pleural space (white arrow). This represents a potential
source of reactivation. B: Non–contrast-enhanced section in a different patient from that in A shows more extensive pleural calcifications,
again associated with pleural thickening and expansion of the extrapleural space with fat. This appearance is also consistent with chronic
inflammation, in this case, also caused by prior tuberculosis. C, D: Contrast-enhanced 5-mm sections in a different patient from that in A or
B. Bilateral thick rinds of pleural calcification are secondary to prior tuberculous pleural disease (C, mediastinal windows). When bilateral,
tuberculous calcific pleural disease may be difficult to differentiate from asbestos-related disease. The presence of upper lobe scarring with
a large granuloma helps confirm the postgranulomatous nature of pleural disease (D, lung windows).
the parietal and visceral pleura, persistent pleural effusions

could be identified in 22 (15%) of cases. This finding is
especially important, given the propensity for residual infec-
tion to result in either a bronchopleural fistula (Figs. 9-35
and 9-36) or a chest-wall infection (empyema necessitatis)
(Figs. 9-37 and 9-38) (63). Similar findings have been
reported for tuberculous empyemas (60,64).

of Complex Pleuro-Parenchymal Disease
CT has been documented to be of considerable value in
assessing all aspects of complex pleuro-parenchymal dis-
ease (8,11,12,60,61,65–68). In our experience, CT has
Figure 9-34 Chronic tuberculous empyema. Contrast-enhanced been of greatest value in (a) differentiation of pleural from
CT section shows large loculated right-sided effusion with exten- parenchymal disease; (b) characterization of underlying
sive rind of calcification. A discrete fat-fluid level (arrow) can be
identified within the pleural space, which can occasionally be iden- parenchymal disease, including identification of necrotiz-
tified in patients with long-standing chronic tuberulous empyemas. ing pneumonias, lung abscesses, and pulmonary infarcts;
Note as well the presence of expansion of the extrapleural fat and (c) characterization of pleural fluid as either free or locu-
ipsilateral volume loss, all consistent with chronic inflammation.
lated, as well as characterization of the appearance of the
pleural membranes themselves; and (d) assessment and
Despite extensive pleural calcification, residual pleural guidance of therapy.
fluid may be identified in a surprisingly high percentage of It cannot be overemphasized that optimal evaluation of
cases, especially when evaluated by CT (Figs. 9-33 and 9-34) complex pleuro-parenchymal disease requires the use of a
(62). In one study of 140 patients with calcification of both bolus of IV contrast media (Fig. 9-39) (9). Besides allowing
A B
Figure 9-35 Tuberculous empyema—bronchopleural fistula. A,

B: Identical sections imaged with narrow and wide windows, re-
spectively, show appearance of chronic loculated tuberculous
effusion with calcifications of the parietal and visceral pleural sur-
faces, expansion of the extrapleural fat, and ipsilateral volume
loss. In this case, an obvious bronchopleural fistula, best seen with
wide windows (arrow in B), is associated with extensive consolida-
tion of both the middle lobe and left upper lobe. A small left effu-
sion is also present. C: CT section in a different patient from that in
A and B also shows evidence of a bronchopleural fistula (curved
arrow) appearing in this patient with history of chronic tuberculous
empyema. As shown in these cases, there is a propensity for
chronic tuberculous empyemas to reactivate, resulting in either a
bronchopleural fistula or chest-wall infection (empyema necessi-
C tatis; see Fig. 9-37).
Figure 9-36 Pleurodesis sequela. CT evaluation. A: 5-mm contrast-

enhanced CT section in a 45-year-old patient with metastatic breast
carcinoma, multiloculated effusions, and underlying pulmonary
atelectasis. More inferior CT section (B) demonstrates that multiple
high-density areas of residual talc are present. These appearances may
mimic prior tuberculous disease but are usually in the dependent pleu-
ral spaces and loculations. Unfortunately, distinction from concurrent
pleural metastatic disease can become problematic, especially as areas
of pleural talc may demonstrate elevated standard uptake values on
B positron emission tomography examinations for many years.
A B
Figure 9-37 Tuberculous empyema: chest-wall infection. A: Contrast-enhanced section shows evidence of extensive lytic destruction of
posterior ribs associated with a multiloculated chest-wall abscess (a). B: Contrast-enhanced CT section in a different patient from that in A
shows presence of bilateral chest-wall abscesses, again associated with lytic rib destruction on the right side (arrow). Both of these patients
proved to have prior pleural disease with evidence of reactivation and extension into the chest wall (so-called empyema necessitatis). In the
case illustrated in B, presumably this was associated with hematogenous seeding, resulting in tuberculous osteomyelitis on the right side.
A B
Figure 9-38 Empyema necessitatis: MRI evaluation. The 56-year-old patient with chronic renal failure and immune compromise status
demonstrated CT appearances of multiloculated pleural effusions similar to the prior case (not shown). On axial T2-weighted MRI (A) a
high-signal-intensity pleural fluid collection is seen with soft tissue elements in the pleura. The collection transgresses the chest wall into
the posterior subcutaneous tissues. Rim enhancement of this infected self-decompressing pleural collection is demonstrated on accom-
panying T1 fat-suppressed gadolinium-enhanced image (B). Unusual to this case, the ultimate etiology proved to be secondary to
Aspergillus.
A B
C D
Figure 9-39 Complex pleuroparenchymal disease: role of intravenous contrast administration. A, B: Non–contrast-enhanced
sections show apparent uniform soft tissue density throughout the left lower lobe. C, D: Sections obtained at the same level as A and B,
after a bolus of intravenous contrast media. Evidence of extensive necrotizing pneumonitis appears throughout the left lower lobe. Contrast
within intraparenchymal vessels is easily identified because of the uniform low density of the surrounding lung, a sign of extensive edema
and/or necrosis. This case clearly demonstrates the value of contrast enhancement for differentiating pleural from parenchymal disease, as
well as a means for further characterizing parenchymal abnormalities when present.
precise localization of fluid collections, intravenous con- (Fig. 9-42). This differentiation may prove difficult even at
trast media may enhance both pleural membranes, surgery. In addition to occasional cases in which empye-
especially when inflamed, and adjacent lung tissue. In most mas and lung abscesses appear to share similar characteris-
cases in which suppurative lung disease is present, charac- tics, fluid within pre-existing pulmonary cavities also
teristic patterns of enhancement may be identified, allow- may pose problems. Even the split-pleura sign may be mis-
ing more precise determination of the nature and extent of leading, as fluid within a bulla may have an identical
underlying lung disease. appearance (Fig. 9-5) (17).
It should also be noted that, rarely, tumor may also
mimic the appearance of loculated pleural fluid (Fig. 9-43).
Differentiation of Pleural from
The key to diagnosis in these cases is to note the presence
Parenchymal Disease of nodularity along the apparent pleural surfaces after
The role of CT in differentiating pleural from parenchymal contrast enhancement.
disease has been well established (10,11,13,69). Lung
abscesses, for example, characteristically appear spherical
Characterization of Underlying
with an irregularly thick wall and cause little compression
Parenchymal Disease
of the adjacent pulmonary parenchyma. By comparison,
empyemas are usually lenticular with a smooth wall, con- In addition to differentiating pleural from parenchymal
form to the shape of the chest wall, and, if sufficiently pathology, CT offers a unique opportunity to characterize
large, cause compression of the adjacent lung (Fig. 9-40). lung disease. Given the spectrum of potential parenchymal
Unfortunately, not all cases fall into such easily classifiable causes of parapneumonic effusions, identification of the
subgroups. In our experience, in a small but significant underlying disease process may have both prognostic and
percentage of cases, even using strict CT criteria, differenti- therapeutic value.
ation between lung abscesses and loculated pleural fluid CT is especially valuable for identifying lung abscesses.
collections/empyemas may be difficult (Fig. 9-41). In Lung abscesses are part of a spectrum of pulmonary
selected cases, of course, both lung abscesses and empye- suppurative processes characterized by necrosis and cavita-
mas may coexist, further complicating interpretation tion, most commonly associated with Staphylococcus aureus,
Pseudomonas aeruginosa, Klebsiella pneumoniae, and anaer-
obes (70). The most common pathogenic mechanisms in
their development are aspiration of oral flora, necrosis
within an antecedent pneumonia, bronchial obstruction,
septic emboli, and penetrating trauma. The necrosis and
Figure 9-40 CT differentiation between loculated pleural fluid

collections/empyemas and lung abscess. Empyemas (E) typically
appear lenticular with a smooth wall, conform to the shape of
the chest wall, and, if sufficiently large, cause compression of the
adjacent lung with displacement of vessels and airways. Unlike
infected bullae, empyemas may extend past the margins of
the adjacent major fissure (MF). In distinction, lung abscesses (LA)
lie within the substance of the lung. Initially, lung abscesses are
smooth walled, and before communication either with an airway Figure 9-41 Lung abscess versus empyema: limitation in CT
or the pleural space, contain no air (arrow). At this stage, adminis- diagnosis. Contrast-enhanced CT section through the lower lobes
tration of intravenous contrast may be invaluable, owing to shows a well-defined, smooth-walled fluid collection with a small
marked enhancement of the abscess wall. Later, after communica- air-fluid level within marginating the posterior mediastinum on the
tion with an adjacent bronchus (B), however, abscesses generally right (arrow). Despite the suggestion of a split-pleura sign (curved
cause little displacement of surrounding structures within the arrow) and the appearance of oblique angles along the margins
lung. Although as drawn, lung abscesses appear to form acute of this lesion, at surgery, this proved to be a thick-walled lung
angles with the adjacent pleura and chest wall, in our experience, abscess associated with adjacent pleural thickening but without
absence of this sign is unreliable. evidence of an empyema.
A, B C
D, E F
Figure 9-42 CT differentiation: lung abscess versus empyema. A–F: Sequential enlargements of views through the left lung base show
findings suggestive of both a lung abscess and empyema. Superiorly a slightly irregular cavity is seen, with an air-fluid level within what
appears to be the peripheral portion of the left lower lobe (arrow in B); more inferiorly, however, this cavity merges with what appears to be
a loculated pleural effusion with an air-fluid level (asterisk in F). Note that medially, there are multiple small thick-walled fluid collections
(arrow in C), which also have features suggestive of both lung abscesses and loculated effusions. In this case, a presumed lung abscess had
ruptured into the pleural space, resulting in a large bronchopleural fistula. After therapy with a chest tube, these findings resolved. This case
illustrates that despite strict adherence to established criteria, differentiation between lung abscesses and empyemas (especially those
associated with a bronchopleural fistula) may be difficult if not impossible. In most of these cases, successful treatment requires that both
abnormalities are assumed to be present.
cavitation that frequently are associated with these infec- Each of these forms of suppurative lung disease has a
tions tend to take several forms. Development of multiple characteristic CT appearance. Lung abscesses are easily rec-
small areas of necrosis, or cavitation within a larger area of ognized as homogeneous areas of low density surrounded
necrosis, is generally termed necrotizing pneumonitis (70). by a markedly enhancing wall, usually associated with a
By comparison, a lung abscess is characterized by a domi- smooth inner contour (Fig. 9-44). Because of the presence
nant focus of suppuration surrounded by a containing wall of hypertrophied bronchial arteries, the wall of most lung
composed of well-vascularized fibrous and granulation abscesses tends to be extremely hypervascular (Fig. 9-45).
tissue (Fig. 9-40), supplied by hypertrophied bronchial This is especially the case in patients with superimposed
arteries. This accounts for the association between lung fungal infections. For this reason, optimal evaluation of
abscesses and hemoptysis. Although the abscess is initially lung abscesses is accomplished by using a bolus injection
self-contained, communication is eventually established of IV contrast, optimally timed to coincide with the early
with nearby airways or with the adjacent pleural space into arterial phase of enhancement. After communication with
which the abscess contents are expelled. Rarely, these infec- the pleural space or adjacent airways, lung abscesses cavi-
tions lead to pulmonary gangrene, characterized by still tate; at this time, their walls become thick and irregular
more extensive necrosis with resultant sloughing of lung (Fig. 9-46). In contrast, necrotizing pneumonitis is charac-
tissue. As previously discussed, all of these manifestations terized by multiple, poorly defined foci of low density,
of pulmonary suppuration are frequently accompanied by unassociated with enhancing margins (Figs. 9-39 and
parapneumonic effusions or empyema. 9-47). When extensive, this appearance merges with that of
A B
C D
Figure 9-43 Lung cancer mimicking loculated pleural fluid. A,

B: Posteroanterior and lateral radiographs show a well-defined
mass on the right side, thought possibly to represent loculated
fluid within the minor fissure. C, D: Sequential sections show
an apparently well defined fluid collection present anteriorly, asso-
ciated with a small quantity of free pleural fluid posteriorly. This
appearance mimics that of a fluid collection within the minor fis-
sure. In retrospect, slight nodular thickening of the “pleural sur-
faces” is seen (arrows in C and D). After unsuccessful attempts at
thoracentesis, this patient underwent an exploratory thoracotomy,
at which time, this mass proved to be a necrotic non–small cell
lung cancer. E: Contrast-enhanced CT section in a different
patient from that in A to D also shows an apparent loculated pleu-
ral effusion on the right side. Note again, however, subtle nodular
thickening of the “pleura” (arrow). This also proved to be a
E necrotic non–small cell lung cancer.
A B
Figure 9-44 Lung abscess. A, B: Identical sections imaged with wide and narrow windows, respectively, after a bolus of intravenous
contrast media show characteristic appearance of a lung abscess within a focal area of parenchymal consolidation (arrow in B).
Figure 9-45 Lung abscess: CT demonstration of hypertro-

phied bronchial arteries. A: Posteroanterior radiograph ob-
tained after catheterization of the right bronchial artery. A focal
area of hyperemia can be identified on the right, associated with
considerable hypertrophy of the distal bronchial arteries (arrow).
B, C: Enlargements of sequential sections through the right
lower lobe in the same patient. In this case, CT scans were
obtained with the catheter still in place within the right bronchial
artery, using a slow drip infusion of intravenous contrast media.
Hypertrophied bronchial arteries can be identified within the
right hilum (small arrows in B and C), leading to a well-defined,
smooth-walled fluid collection in the right lower lobe (large
arrows). Note that the wall of this lesion is dramatically enhanc-
ing, consistent with a bronchial arterial blood supply. Surgically
A documented lung abscess.
B C
796
may be considered (71). CT may also be valuable in detect-

ing complications that may occur within abscesses, includ-
ing the presence of intracavitary fungus balls. It should be
noted that although cavitary lung neoplasms may mimic
the appearance of a lung abscess, in most cases, only
minimal enhancement of the tumor wall is seen, as
compared with lung abscesses, especially in squamous cell
carcinomas (72).
Pleural Effusions: Computed Tomography

Evaluation
Free Versus Loculated Effusions

Free pleural fluid has a characteristic appearance when
seen in cross section. Fluid typically looks “meniscoid,”
occupying the posterior pleural space in patients scanned
in the supine position (Fig. 9-27). As effusions increase in
size, they conform to the natural boundaries of the pleura.
If sufficiently large, effusions usually result in compression
Figure 9-46 Lung abscess: communication with the pleural
space. Enlargement of a contrast-enhanced CT section through atelectasis of the underlying lung (Fig. 9-27); this is not
the right middle lung shows a typical lung abscess (curved arrow). invariable, however, with large effusions often causing
In this case, the abscess is beginning to communicate with the minimal, if any, change to the adjacent lung parenchyma
adjacent pleural space (arrowheads), within which a small, locu-
lated pleural fluid collection is beginning to form (straight arrow). (Fig. 9-27). Rarely, direct communication may be identi-
Note the presence of air within both the pleural fluid collection fied between the right and left pleural spaces, presumably
and the abscess. (Case courtesy of Peter Nardi, MD, Brooklyn, the result either of congenital communications or after
New York.)
prior surgery (Fig. 9-27). In a surprising number of cases,
small effusions may prove difficult to differentiate from
pulmonary gangrene. Although in typical cases, differentia- pleural thickening and/or fibrosis. In these cases, scans
tion between patients with lung abscesses and necrotizing obtained with the patient in the lateral decubitus and/or
pneumonitis is not associated with significant differences prone position can be extremely helpful. Fissural pseu-
in therapy, this differentiation is of importance in those dotumors are also common and are usually secondary to
cases for which percutaneous drainage of lung abscesses congestive heart failure (73).
A B
Figure 9-47 Necrotizing pneumonia. A, B: Sequential contrast-enhanced sections show extensive consolidation throughout the entire
right lung. Note that the right lung has heterogeneous low density, with scattered small fluid collections easily identified (arrows in A and B)
associated with small air-fluid collections. These findings are consistent with a necrotizing pneumonia associated with multiple small
“microabscesses.” Vessels are easily identified traversing the right lung (so-called “positive angiogram” sign). Note that only a small quan-
tity of pleural fluid is left after drainage with a pleural tube. The pleural surfaces are clearly enhancing. Both blood cultures and pleural
cultures were positive for Streptococcus pneumoniae. It cannot be overemphasized that this appearance is distinctly separate from that
seen in patients with compressive atelectasis caused by an effusion in which no underlying infection is found. In these cases, after intra-
venous contrast enhancement, uniform density is seen throughout the collapsed lung (compare with Figs. 9-27B and 9-32).
Radiographically, fissural fluid usually is easily identi- in 96% of 25 patients with empyema who underwent con-
fied because (a) the fluid collection lies in the expected trast-enhanced CT scans. Whereas 86% showed thickening
region of the fissures; and (b) unless the fissure lies exactly of the parietal pleura, 60% showed thickening of the
perpendicular to the plane of the radiograph, the margins extrapleural subcostal tissues, and 35% showed increased
of the fluid collection appear hazy or poorly defined. attenuation of the extrapleural fat (Figs. 9-32 and 9-48)
These principles apply to the appearance of fluid within (76). Of 14 patients with complicated parapneumonic
fissures on CT (Figs. 9-28 through 9-30) (74). effusions, the thickness of the parietal pleura and extra-
pleural tissues averaged 3 mm and 3.5 mm, respectively,
Computed Tomography Characterization whereas in eight patients requiring decortication, the
Numerous reports have examined the CT characteri- average thickness of these layers was 4 mm and 4.5 mm,
stics in patients with pleural effusions (11,75–77). In an respectively. In distinction, none of these findings were
early retrospective evaluation of 48 patients with pleural present in the 20 patients with transudative effusions (76).
effusions who had a sonographically directed thoracentesis, Aquino et al. (77) also evaluated CT findings in 80 con-
Himmelman and Callen (78) found a significant correla- secutive patients (86 effusions) in whom correlative
tion between pleural fluid loculation and exudative pleural thoracenteses were performed. Specific CT signs evaluated
fluid chemistries. In addition, patients with loculated effu- included evidence of both parietal and visceral pleural
sions tended to have larger effusions and longer hospitaliza- thickening (further classified as either focal or diffuse,
tions, and more frequently required tube drainage. Seven of smooth or irregular), loculation, and evidence of expan-
nine empyemas (78%) and 10 of 28 exudates (36%) were sion or edema of the extrapleural space. Of 59 proved
loculated. In 30% of cases, pleural fluid loculation was exudative effusions, 36 (61%) showed evidence of parietal
identified, only by CT in 8%. None of these findings pleural thickening. All cases of empyema (n 10) and
was noted in any patient with a transudative effusion. 56% of patients with exudative parapneumonic effusions,
Waite et al. (76) reported evidence of pleural thickening including all five patients with complicated parapneu-
after contrast enhancement on CT in nearly all patients monic effusions, had evidence of parietal pleural thicken-
with either empyemas or complicated parapneumonic ing. Overall, the specificity of this finding for exudative
effusions. Among 35 patients with documented thoracic effusions proved to be 96%, with a corresponding posi-
empyema, 30 patients with malignant pleural effusions, tive predictive value for 97% (77). Also of value was the
and 20 patients with transudative effusions, Waite et al. finding of thickening and increased attenuation within
found that enhancement of the parietal pleura was present the extrapleural fat, identifiable in 8 of 10 patients with
A B
Figure 9-48 Complicated parapneumonic effusion. A: Fibrinopurulent stage. Contrast-enhanced CT scan in a patient who had received
antibiotics before admission. A moderate-sized right pleural effusion is associated with volume loss and consolidation in the right lower lobe
(curved arrow). The parietal pleura is slightly thickened and is clearly enhancing (arrows). Thoracentesis yielded clear pleural fluid without
organisms. pH, 6.72; glucose, 22 mg/dL; protein, 5.4 g/dL; LDH, 518 IU; WBC, 3400/mm3 with 70% polymorphonuclear leukocytes. This
patient was subsequently successfully treated with closed-tube drainage. B: Fibrinopurulent stage. Contrast-enhanced CT scan in a differ-
ent patient from that shown in A. A loculated pleural effusion is present on the right side, compressing the adjacent right lower lobe.
Marked thickening of both the visceral and parietal pleura (arrows) is present. In addition, there is considerable thickening of the
extrapleural subcostal tissues and increased attenuation of the extrapleural fat (curved arrow), presumably secondary to spread of infection
or edema to the adjacent chest-wall tissues.
empyemas. In distinction, the size or shape of effusions did only for effusions with a pH less than 7.0 (80). Pleural
not prove of diagnostic value. Based on these data, Aquino fluid loculation is also cited as a reason for drainage.
et al. (77) drew the following conclusions: (a) pleural The role of CT for distinguishing between simple para-
thickening in association with an effusion in a patient with pneumonic effusions and empyemas is controversial. Evans
pneumonia indicates the presence of an exudative effusion; et al. (13), for example, argued that CT is not required in
and (b) conversely, the absence of pleural thickening the management of most patients with parapneumonic
makes the likelihood of either a complicated parapneu- effusions or empyemas, especially when ultrasound is used
monic effusion or empyema extremely unlikely. to confirm the presence and to identify optimal sites for
More recently, in a prospective study of 211 consecutive chest tube placement (13). Most other investigators have
effusions with documented etiologies, Arenas-Jimenez et al. reported a role for contrast-enhanced CT studies (77), but
(11) re-evaluated the ability of CT to differentiate between it should be noted that CT is limited in differentiating
tranudative and exudative effusions. Similar to previous between parapneumonic effusions and empyemas. As
reports, these authors noted that pleural fluid loculation, documented by Arenas-Jiminez et al. (11), although the
pleural thickening, pleural nodules, and increased density finding of diaphragmatic, mediastinal, or circumferential
in the adjacent extrapleural fat were findings identified pleural thickening in patients with effusions due to sus-
only in patients with exudative effusions. Of these, pleural pected infection proves a reliable means for diagnosing
thickening proved to be the most sensitive [sensitivity, empyema, this occurs in a small minority of cases. In dis-
42%; specificity, 100%; positive predictive value (PPV), tinction, the presence of fluid loculation, thickening of the
100%; and negative predictive value (NPV), 25.8%]. costal or visceral pleural, or abnormalities involving the
In this same study, when effusions due to proved extrapleural fat, although suggestive, are seen in patients
infection were evaluated separately, only visceral pleural with both parapneumonic effusions and empyemas. In
thickening and increased density in the extrapleural fat particular, although costal pleural thickening has been pre-
proved of diagnostic value, with the PPV of both findings viously reported to occur in between 96% and 100% of
of approximately 75%. Visceral pleural thickening was cases (76,77,81), in Arenas-Jiminez’ study (11), this find-
identified in 20% of patients with infectious etiology, ing was identified in only 75% of cases. Similarly, although
similar to findings previously reported by Aquino et al. the finding of increased attenuation of extrapleural fat
(77). As is discussed later, this same finding is also seen is often identified in patients with empyemas, these
in patients with malignant effusions, rendering it less findings are also nonspecific, with sensitivities ranging
significant (69). from 30% to 72% (11,76,77,82), presumably reflecting
Although the thickness of the parietal pleura appears to differences in timing, with CT studies performed earlier in
correlate with the stage of parapneumonic effusions, little the course of disease being less specific.
correlation is found between this appearance and the abil- In cases for which pleural drainage is indicated, a num-
ity to predict which patients ultimately will require either ber of alternative approaches are available, including
chest tube insertion, thoracoscopy, or decortication. In a repeated thoracenteses as well as image-guided percuta-
prospective study of serial CT scans in 10 patients after neous catheter drainage performed by using fluoroscopy,
radiologic catheter drainage of empyemas, Neff et al. (79) ultrasound, or CT, as well as video-assisted thoracotomy
found that although the pleura was noticeably thickened (VATS) and thoracotomy with drainage or, if indicated,
4 weeks after catheter removal in all patients, at 12 weeks, decortication. Choice between these modalities reflects
the pleura was essentially normal in four patients and individual practitioners’ experience.
showed only minimal or mild residual pleural thickening To date, numerous studies have shown an advantage
in the remainder (79). to image-guided catheter placement over blind place-
ment of large-bore chest tubes (67). In addition to ensur-
ing accurate tube placement, image-guided drainage has
Assessment and Guidance of Therapy
the additional advantage of directing individual drainage
In cases in which exudative effusions are the result of of multiple loculations when present. Although image
infection, differentiation between parapneumonic effu- guidance can be successfully performed by using either
sions and empyemas is critical to guide appropriate man- ultrasound or CT, ultrasonic guidance has the advantage
agement. As a general rule, antibiotic therapy is the initial of being relatively inexpensive and portable, allowing
treatment of choice in patients with uncomplicated para- studies to be performed even in intensive care units.
pneumonic effusions (67). Although consensus indicates However, ultrasound is far more operator dependent
that patients with empyemas require drainage, in the than is CT. Furthermore, despite claims to the contrary,
absence of frankly purulent effusions, indications for fluid CT is far more accurate in detecting underlying lung
drainage are less clear. In this setting, pleural fluid pH has pathology, as well as in identifying multiple areas of loc-
usually been the decisive factor, although a precise thresh- ulation, especially when these are paramediastinal.
old is still debated, with most investigators using a thresh- Additionally, CT is not limited by the presence of band-
old pH of 7.21 to 7.29, whereas others argue for drainage ages, drains, or tubes.
Merriam et al. (83), by using a combination of imaging placed within the major fissures or within the lungs, as
modalities, retrospectively reviewed the outcome of percu- well as those inadequately positioned to ensure proper
taneous catheter drainage of pleural fluid collections in drainage (Fig. 9-49) (87,88). In one retrospective study, CT
18 patients, including 16 patients with documented emp- revealed 28 malpositioned chest tubes among 76 chest
yemas, nine of whom had previous unsuccessful surgical tubes placed in 54 patients, including 20 intraparenchy-
chest tube drainage. Twelve (80%) of 15 patients who mal and 3 intrafissural tubes, as well as 5 chest tubes in
had an adequate trial of guided drainage were cured. extrathoracic locations, including 4 in the mediastinum
VanSonnenberg et al. (84) reported similar results by using and 1 in the chest wall (66). In distinction, frontal radi-
both CT and ultrasound to drain 17 patients in whom ographs disclosed malpositioned tubes in none. Similar
previous conventional chest tube drainage had proven findings have been documented by Stark et al. (87), who
unsuccessful. In this study, 15 (88%) of patients were suc- reported, in a study of 26 patients with tube thoracos-
cessfully drained, averting the need for open drainage tomies for treatment of empyema, that frontal radiographs
(84). By using fluoroscopic guidance, Westcott (85) suc- were of value in identifying malpositioned chest tubes in
cessfully drained 11 of 12 patients with documented only 1 of 21 patients.
empyemas; in five of these patients, percutaneous fluid CT also can help detect more serious complications of
drainage was used as the sole means of drainage (85). chest tube placement, including significant chest wall
Percutaneous catheter placement may also be accompa- hemorrhage (Fig. 9-50). In addition to assessing the ade-
nied by the installation of fibrinolytic agents, including quacy of chest tube placement, CT has also been proven
streptokinase, urokinase, or tissue plasminogen activator, effective in identifying residual changes caused by previous
with relatively few adverse reactions (86). pleural tubes. CT may play an especially important role in
CT is of proven efficacy for both directing and assessing the evaluation of patients who have undergone surgical
the response of patients to therapy. By identifying areas of therapy, including thoracoplasties (Fig. 9-51), open
loculated fluid, CT can be used to guide the appropriate drainage procedures including Eloesser window thoracos-
placement of chest tubes in cases for which tube thoracos- tomies (Fig. 9-52) (89), and even in patients who have
tomy is indicated. In patients with chest tubes, CT can be been treated with oleothoraces or plombage (Fig. 9-53)
especially valuable by disclosing chest tubes inadvertently (90–92). It should be noted that in selected cases,
A, B
Figure 9-49 Chest-tube placement: CT

evaluation. A–D: Enlargements of sequen-
tial images in a patient with atelectasis of
the middle lobe shows a chest tube enter-
ing from the lateral chest wall, traversing
the right upper lobe to end with its distal
tip in the anterior mediastinal fat (arrows).
The clinical significance of such misplace-
ment is variable and relates predominantly
to whether functional drainage of pleural
fluid or air exists. Despite the apparent
intraparenchymal course of many chest
drains, significant parenchymal lacerations
C, D are unusual.
A B
Figure 9-50 Complication of closed chest-tube placement: CT evaluation. A: Posteroanterior radiograph obtained shortly after placement
of a left-sided chest tube shows nonspecific increased density on the left, suggestive of a loculated pleural collection. B: Non–contrast-
enhanced CT section shows that the pleural tube is displaced medially by a large, high-density, extrapleural fluid collection compressing and
displacing the adjacent lung and pleura (arrows and asterisk), consistent with extrapleural hemorrhage. At surgery, these findings were con-
firmed to be secondary to traumatic injury to an intercostal artery that presumably occurred at the time of chest-tube insertion.
A B
Figure 9-51 Postsurgical assessment: thoracoplasty. 5-mm un-

enhanced images in a patient after prior thoracoplasty for
advanced pleuroparenchymal tuberculosis. The resection of multi-
ple ribs with deformity of the chest and collapse of the ipsilateral
hemithorax is best noted on bone windows (A). The patient ini-
tially had fevers and weight loss; the presence of advanced
bronchiectasis with infective consolidation, including a cavity
that was colonized with Aspergillus, is confirmed on correspon-
C ding lung windows (B and C).
A B
Figure 9-52 Postsurgical assessment: Eloesser window thora-

costomy, with 5-mm sections just inferior to the carina (lung
windows, A) and at a level in the inferior hemithorax (lung and
mediastinal windows, B and C) in a patient after Eloesser flap tho-
racostomy. The purpose of this procedure is to provide continued
drainage of the pleural space in patients with intractable chronic
empyemas. Despite the presence of surgically created open com-
munication to the exterior (B and C), the lung does not collapse,
largely because of the marked thickening of the visceral pleura
with adjacent chronically scarred and atelectatic parenchyma
(A and B). Packing material in the pleural space is periodically
replaced to absorb pleural secretions. Centrilobular emphysema is
C noted in the contralateral lung.
A B
Figure 9-53 Postoperative evaluation: tuberculous empyemas. A: Non–contrast-enhanced CT section shows multiple rounded lucencies
within the right hemithorax in a patient previously treated with plombage. Note the marked distortion of the ribs on the right side, consis-
tent with prior thoracoplasty. B: Non–contrast-enhanced section through the middle lung, in a different patient from that in A, shows that
the entire left hemithorax is filled with uniform fat density, surrounded by a dense rim of calcification. This is the result of prior oleothorax.
Similar to plombage, oleothorax was formerly used in the therapy of tuberculosis and may occasionally still be identified in elderly patients.
A B
Figure 9-54 Lung abscess: MR evaluation. A, B: T1- and T2-weighted MR scans, respectively, through the mid thorax show extensive
consolidation throughout the entire left lung. A well-defined, smooth-walled, fluid-filled cavity can be identified, within which a discrete air–fluid
level is seen (arrows in A and B). A central mass is also present, obliterating the left lower lobe bronchus and significantly narrowing the left
upper lobe bronchus. In this case, MR clearly discloses the presence of a lung abscess developing distal to a central obstructing lesion.
especially in those for whom there is a contraindication effusions; and (c) diffuse pleural fibrosis (98). Although
for the use of IV contrast media, MR may be of value in these may be associated with underlying parenchymal
differentiating pleural from parenchymal pathology, as abnormalities, they frequently occur independently (99).
well as detecting lung abscesses (Fig. 9-54). Of these, CT has proven most valuable in identifying pleu-
Although most investigators advocate the use of closed ral plaques and diffuse pleural thickening (100–106). The
pleural drainage in the initial management of complicated subject of CT findings in asbestosis is covered in detail in
parapneumonic effusions, adequate treatment of empye- Chapter 8.
mas usually requires surgical intervention. This is especially
true in patients in whom initial trials with antibiotics and
Circumscribed Pleural Plaques
closed pleural drainage have proved unsuccessful. In one
series of 70 patients with thoracic empyemas, closed tube Pleural plaques are the most common benign pleural
thoracostomy was successful in only 35% of cases, whereas manifestations of asbestos exposure (Figs. 9-55 and 9-56).
rib resection proved curative in 91% (93). Recently, atten- They occur after a latency period of between 20 and
tion has focused on the use of VATS for both drainage 40 years, and generally are asymptomatic. Presumably
and decortication in patients with empyema (94,95). these are the result of parietal pleural irritation caused by
Advantages cited for VATS compared with open thoract- asbestos fibers protruding from the visceral pleural sur-
omy are less pain and shorter hospital stays, especially face; it has also been hypothesized that fibers could pass
when performed early in the course of infection (76). In from the lung via chest wall lymphatics to the parietal
one retrospective study, VATS debridement and decortica- pleura (98). Typically, plaques appear as discrete, ele-
tion were successfully performed in 65 of 70 consecutive vated, sharply defined foci of pleural thickening up to
cases (96). Unfortunately, CT has proved relatively insensi- 15 mm thick (Fig. 9-55) (103,107). Although usually
tive for predicting the outcome of therapy, especially which found posterolaterally along the inferior costal margins as
patients will ultimately require decortication. The presence well as along the diaphragm, rarely plaques may involve
of thickened costal pleura in children in particular has the visceral pleura within fissures (108). Characteristically
proved of limited value, as this may resolve with return bilateral, some have observed a distinct unilateral, left-
of normal lung function when monitored for up to sided predominance (109). Calcifications occur in appro-
18 months (97). ximately 10% of plaques (105). These may appear
punctate, linear, or occasionally “cakelike,” especially
when located along the diaphragmatic surfaces. Less com-
Asbestos-Related Pleural Disease
monly, calcified plaques may be pedunculated, in which
Benign pleural manifestations of asbestos exposure include case they may be mistaken for intraparenchymal nodules
(a) circumscribed pleural plaques; (b) benign exudative (Fig. 9-56). Histologically, plaques are composed of
A B
C D
Figure 9-55 Asbestos-related pleural plaque disease: range of appearances using 5-mm axial CT sections with mediastinal windows in
four different patients. A: Typical bilateral distribution of disease is demonstrated, including a combination of calcified and noncalcified
posteromedial and anterolateral plaques bilaterally. B: Predominantly calcified pleural plaques. Note the presence of extrapleural fat along
the posteromedial left hemithorax, indicating the chronicity of the process. C: Calcified pleural plaques along the diaphragmatic pleural
surfaces, including the posterior right hemidiaphragm. This region corresponds to the bare area of the liver and therefore in the absence of
peritoneum in this region, calcifications identified in this area must be pleural in nature. This feature may be of use in isolated calcifications
occasionally noted in this area. D: Unilateral predominantly soft tissue plaques. A unilateral distribution of plaque disease may make the dif-
ferentiation from postgranulomatous or post-traumatic pleural thickening problematic; however, such unilateral predominance of asbestos-
related pleural disease is well described, particularly on the left side.
predominantly acellular bundles of collagen usually Benign Exudative Effusions

described as having an undulated or “basketweave” con-
figuration (99, 109, 110). Although plaques have been Benign exudative effusions generally occur considerably
attributed to other causes, including previous empyema earlier than other pleural manifestations of asbestos-
and hemothorax, correlation between the presence of related pleural disease, usually within 10 to 20 years after
bilateral plaques and asbestos exposure is sufficiently exposure (98). They may be unilateral or bilateral, and
high to warrant defining them as markers of previous recur in up to 30% of patients. As these effusions are usu-
dust exposure. A history of asbestos exposure can be ally nondescript and self-limited, their true incidence is
elicited in more than 80% of individuals with difficult to determine. Epler et al. (111) reported identify-
pleural plaques (99,109,110). Pleural plaques are always ing 35 cases (3.1%) of benign asbestos effusion among a
benign. survey group of 1135 asbestos-exposed workers (111).
A B
Figure 9-56 Benign pleural plaque mimicking a nodule. A chest

radiograph performed in a 55-year-old man as part of a preoperative
evaluation for coronary artery disease demonstrated an ill-defined
nodule in the left mid lung (A, arrow) that was considered suggestive
of a neoplasm. CT performed using 5-mm sections through this
region demonstrated no pulmonary abnormality (B). A thin pleural
plaque (B and C, arrow in C) accounted for the pulmonary “pseudo-
nodule.” In the absence of chest-wall invasion or the presence of
C pleural fluid, such plaques are almost always benign.
Although the relationship between benign exudative effu- et al. (107) defined diffuse pleural thickening as a smooth,
sions and the subsequent development of malignancy is noninterrupted pleural density extending over at least one
unclear, it is unlikely that they represent a significant risk fourth of the chest wall, with or without costophrenic angle
factor for mesothelioma (112). Nonetheless, as noted by obliteration. Unlike pleural plaques, diffuse pleural fibrosis
Gefter et al. (113), extra caution is probably warranted in presumably involves both the visceral and parietal pleural
those individuals with histories of asbestos exposure in surfaces. The exact mechanism by which this occurs is con-
whom a benign exudative effusion is identified. troversial. It has been postulated that diffuse fibrosis results
from an extension of underlying parenchymal disease to
involve the adjacent visceral pleura (114). The frequency
Diffuse Pleural Thickening
with which this occurs has been challenged, however.
In addition to pleural plaques and benign exudative effu- As documented by McLoud et al. (107), in a study of 185
sions, diffuse pleural thickening may also result from individuals with diffuse pleural thickening, this radio-
asbestos exposure (Fig. 9-57) (98). Radiographic differenti- graphic appearance proved to be the residue of a prior
ation between these entities may be problematic. McLoud benign asbestos effusion in 31% of cases, and the result of
documented by Sargent et al. (53), in a study of 30 patients

with known asbestos exposure in whom radiographic inter-
pretations were equivocal for the presence of pleural
plaques, in 14 (48%) cases, CT confirmed that the
changes were caused by increased amounts of subpleural fat
(Fig. 9-58) (53). Similar findings have been reported by
Friedman et al. (104). In a study of 60 individuals with his-
tories of occupational exposure to asbestos comparing the
value of HRCT scans with routine radiographs for diagnos-
ing pleural abnormalities, these authors showed the positive
predictive value of chest radiographs to be 79%, compared
with a positive predictive value of 100% for HRCT. Of
eight cases, false-positive chest radiographs proved to be
secondary to extra subpleural fat in seven, and a prominent
intercostal muscle in one.
The role of HRCT in the evaluation of benign asbestos-
related pleural disease has also been evaluated by Aberle
et al. (101). In a study of 29 subjects with histories of occu-
pational exposure to asbestos, pleural thickening was
identified in 100% of cases by using HRCT, compared with
Figure 9-57 Asbestos-related pleural disease: diffuse pleural
fibrosis. Enlargement of a CT section through the left lung base 93% by using routine 10-mm-thick CT sections. In addition
shows a markedly thickened pleural rind, associated with marked to identifying pleural plaques, CT can be of value by
expansion of the extrapleural fat (arrows). In addition, a small, disclosing diffuse pleural fibrosis. Gamsu et al. (105) sug-
residual pleural fluid collection is identifiable (curved arrow). The
presence of residual fluid suggests the possibility of a mesothe- gested a useful CT definition of diffuse thickening: a con-
lioma; however, pleural aspiration and biopsy showed only dense tinuous sheet of increased density at least 5 cm broad, 8 to
fibrosis and noninfected nonmalignant fluid. It is probable that 10 cm long, and more than 3 mm thick. It should be
with CT, small, residual pleural fluid collections will be recognized
more commonly in patients with asbestos-related pleural fibrosis. emphasized that care must be taken in assessing pleural
thickness, especially when using HRCT. As shown by Im
et al. (7), in normal patients, identification of the usually
confluent pleural plaques in 25%. In this study, only 10% pencil-thin line of the normal pleural surfaces may be
of cases with diffuse pleural fibrosis proved to have under- obscured, especially in the paravertebral regions, because of
lying diffuse parenchymal fibrosis with extension to the a normal increase in the amount of extrapleural soft tissue
visceral and parietal pleura. Differentiation between pleural caused by the incorporation of adjacent intercostal vessels
plaques and diffuse pleural thickening is important, as (Figs. 9-21 to 9-23, 9-26). As a consequence, the diagnosis
the latter may be associated with significant alterations in of diffuse pleural thickening with HRCT requires that
pulmonary function (115,116). abnormalities be identified at several levels, preferably
identifiable in other than just the paraspinal region.
Adequate visualization of both pleural and parenchy-
Computed Tomography Evaluation of Benign
mal changes in asbestos-exposed patients can be achieved
Asbestos-Related Pleural Disease
with a low-dose technique. As documented by Remy-
As early as the late 1970s, Kreel and Katz (2,117,118) Jardin et al. (8), no significant differences were observed in
showed CT to be significantly more sensitive than plain the identification of parietal pleural plaques or most find-
radiographs in the detection of asbestos-related pleural ings consistent with parenchymal fibrosis when low-dose
disease. In their series of 36 patients with known asbestos (60 to 100 mAs, depending on individual body habitus)
exposure, 27 (75%) individuals were found to have images were obtained on a four-detector CT scanner using
abnormal pleural thickening on CT, whereas 24 (66%) 4 2.5-mm collimation to reconstruct contiguous 5-mm
individuals had abnormalities detected on chest radio- sections. It may be anticipated that similar, if not superior,
graphs (118). In this same series, CT was 50% more results can be obtained on newer-generation MDCT
sensitive in detecting pleural calcifications than were scanners, allowing the reconstruction of contiguous 1-mm
conventional radiographs. images while still using low-dose technique.
It is apparent that a significant advantage exists in visual- The appearance of diffuse pleural thickening generally
izing the pleural surfaces without superimposition of densi- is easily differentiated from mesothelioma. Rabinowitz
ties (119,120). Not only does CT allow a more precise et al. (122), however, drew attention to a variant of
assessment of the extent of pleural disease, but in a signi- asbestos-related pleural fibrosis that closely mimics the
ficant percentage of cases, CT can clarify otherwise appearance of malignant mesothelioma. In their series of
indeterminate chest radiographic findings (100,103,121). As 40 patients with known asbestos exposure, seven had
Figure 9-58 Prominent extrapleural fat: CT evaluation. Chest radi-

ograph initially interpreted as demonstrating pleural thickening
(arrows; A) in a patient with asbestos-exposure history. The 5-mm
axial images after intravenous contrast (B) and 3-mm coronal multi-
planar reconstruction (C) demonstrate that the finding is due to
extrapleural fat, asymmetrically prominent on the right side (arrows).
The presence of extrapleural fat should be considered when the
apparent chest radiographic thickening is bilateral, most marked in
the middle third of the hemithorax without effusions or costophrenic
angle involvement. The presence of notable subcutaneous body fat
and male gynecomastia, not prominent in this case, are ancillary
A supporting findings.
B C
diffusely thickened, nodular pleural surfaces, indistin- described in this century, round atelectasis has been recog-
guishable from malignant mesotheliomas. None of these nized with increasing frequency, especially with increasing
patients, however, had evidence of malignant transforma- use of CT to evaluate asbestos-exposed patients (103,105).
tion detected by multiple biopsies or by surgery. The Radiologically, round atelectasis usually is an incidental
significance of this type of fibrosis has yet to be estab- finding on routine chest radiographs (Figs. 9-59 and 9-60)
lished, as long-term follow-up studies of these patients (124,125). A distinct preponderance in men has been iden-
were not undertaken (Fig. 9-57). tified (126). Typically, conventional radiographs reveal a
sharply defined pleural-based mass, ranging between 2 and
7 cm in size, usually located posteriorly in the lower lobes
Round Atelectasis
adjacent to an area of pleural thickening. Air bronchograms
A variety of terms, including folded lung, atelectatic pseudo- may be present within. Characteristically, the mass is associ-
tumor, shrinking pleuritis with atelectasis, pleuroma, and ated with vessels and bronchi that have a curvilinear appear-
round atelectasis (the most commonly used descriptive), ance, coursing like a comet tail toward the hilum. These
refer to what is most often a manifestation of asbestos- findings are associated with focal volume loss and often
related parenchymal and pleural disease (123). First with hyperlucency of adjacent lung segments.
A B
Figure 9-59 Rounded atelectasis: MR evaluation. T1-weighted (A) and T2-weighted (B) images demonstrate a complex mass of interme-
diate T1 and relatively high T2 signal intensity (arrows). Contrary to opinion, foci of rounded atelectasis do not represent areas of parenchy-
mal fibrosis, but instead, infolded lung. It is, therefore, not surprising that considerable signal is present within these lesions, particularly on
T2-weighted sequences. In distinction, note that the adjacent pleura, composed as it is of dense, mature fibrous tissue, generates no signal
at all (curved arrows). Potentially in equivocal cases, MR may be of value in confirming that the pleural thickening seen on CT does not
represent tumor infiltration.
CT findings in round atelectasis have been extensively that causes extensive focal pleural fibrosis (Fig. 9-60). In an
reviewed (103,105,127–132). The major CT sign is a evaluation of 74 patients with rounded atelectasis evaluated
rounded or wedge-shaped peripheral lung mass forming by Hillerdal (126), 64 patients gave a history of asbestos ex-
an acute angle with the adjacent pleura, which is almost posure. Of the remaining 10 cases, 2 occurred after trauma,
invariably focally thickened. Additionally, vessels and and 4 occurred after a pleural exudate. Round atelectasis
bronchi can be identified curving toward the mass, creat- also has been described in association with histoplasmosis
ing a comet-tail appearance analogous to that seen on rou- (139) as well as tuberculosis, including in patients who
tine chest radiographs (Figs. 9-59 and 9-60). Minor signs have undergone therapeutic pneumothoraces (130).
include focal emphysema, punctate areas of calcification, In addition to the typical appearances described, several
and uniform enhancement after administration of IV con- variant forms of round atelectasis have been described. In
trast media. Unfortunately, hypervascular lesions have patients with extensive pleural fibrosis, in particular, a pat-
been described in patients with asbestos-related pleural tern of single or multiple fibrous strands within the lung,
disease and lung cancers (133,134). Other tumors, albeit radiating toward a focal area of pleural thickening, has
exceedingly rare, also have been described in association been described in the absence of a definable parenchymal
with asbestos-related disease (135,136). mass. These so-called “crow’s feet” have been interpreted
Histologically, round atelectasis occurs adjacent to areas by some as the earliest manifestation of round atelectasis
of both parietal and visceral pleural fibrosis (123). Round (Fig. 9-61) (126). As noted by Lynch et al. (129), other
atelectasis probably results from at least one of two mecha- benign lesions also occur in association with asbestos
nisms. As hypothesized by Blesovsky (137), atelectasis exposure, including benign pleural-based masses, intrafis-
probably results as a consequence of asbestos-induced sural pleural plaques, and masslike fibrotic sheets, espe-
pleural fibrosis. In support of this explanation, extensive cially adjacent to the hemidiaphragms. These authors
fibrotic changes involving the adjacent visceral pleura have calculated that CT will detect asymptomatic benign masses
been verified pathologically (123,131). Alternatively, it is in up to 10% of individuals with histories of significant
likely that in some cases, round atelectasis results from exposure to asbestos.
compression of the lung caused by a pleural effusion, caus- In most cases, especially when a history of prior asbestos
ing infolding and distortion of the adjacent lung (138). exposure exists, the CT appearance of round atelectasis is
Interestingly, parenchymal changes have been reported to sufficiently characteristic to obviate histologic verification,
resolve after decortication (123). although careful follow-up surveillance is mandatory to
Although generally associated with asbestos-related pleu- ensure an absence of growth. Alternatively, it has been
ral disease, round atelectasis may result from any process suggested that suspicious lesions may also be evaluated
A B
C D
Figure 9-60 Round atelectasis caused by prior tuberculosis. A: Posteroanterior chest radiograph shows ill-defined nodular density in the
left middle lung field (arrow). No other apparent pleural or parenchymal abnormality can be identified. B, C: Sequential images confirm
focal, ill-defined density in the peripheral portion of the left upper lobe, extending to the pleural surface. Note the curvilinear configuration
of the proximal left upper lobe vessels (arrows in B and C), the so-called “comet tail sign.” D: Identical section as in C, imaged with narrow
windows, shows the presence of dense calcifications within the parenchymal mass; focal pleural thickening can be identified adjacent to this
mass, associated with a broad band of soft tissue leading from the calcifications to the pleural surface (arrow). Note the bilateral hilar calcifi-
cations. In this case, round atelectasis has resulted from prior granulomatous infection with focal pleural and parenchymal scarring.
Figure 9-61 Round atelectasis: early recognition by CT. High-

resolution CT scan shows multiple curvilinear strands within the right
lower lobe, all oriented toward a slightly irregular and thickened pleu-
ral surface. These so-called “crow’s feet” have been described as the
earliest sign of developing round atelectasis (arrows).
with FDG-PET scans. As reported by McAdams et al. (140), in patients without direct pleural invasion. Mechanisms
in their study of nine patients with 10 lesions, none proved that have been proposed to account for paramalignant
positive, leading these authors to conclude that round effusions include (a) tumor obstruction of both central
atelectasis is not metabolically active on FDG imaging. lymphatics and airways, with resultant pneumonia or
These findings, while suggestive, clearly require further atelectasis; (b) systemic effects of the disseminated tumor;
validation. and (c) results of radiation or drug therapy (58).
In those cases in which there is evidence of a change The definitive diagnosis of malignant pleural disease
in size on follow-up CT studies, histologic evaluation is usually requires pleural fluid cytologic examination, pleu-
indicated, even though it is well documented that areas of ral biopsy, or even exploratory thoracotomy (147,148). In a
round atelectasis may continue to enlarge with time. It review of recent literature, Sahn (58) documented that
should also be noted that although rare, an association has the diagnostic yield from pleural fluid cytologic exami-
been made between round atelectasis and malignant pleu- nation was 66%, and from pleural biopsy was 46%
ral mesothelioma (MPM) (141). In particular, findings (58). Combining procedures resulted in a 73% diagnostic
other than those described as typically seen in patients yield. A cytologic diagnosis of pleural malignancy is most
with round atelectasis, such as diffuse pleural thickening commonly made in patients with adenocarcinoma (sensi-
or nodularity or pleural effusions, should suggest the pos- tivity, 70%), less commonly with squamous cell carcinoma
sibility of coexistent malignancy. (sensitivity, 20%) or metastatic sarcomas (sensitivity, 25%),
Although MR has been used to evaluate round atelecta- and least likely with mesothelioma (sensitivity, 10%),
sis, to date no consistent pattern of signal intensity within although with current methods of immunohistochemical
areas of rounded atelectasis has been found (142). staining, cytologic diagnosis of MPM has substantially
However, MR may disclose the densely fibrotic nature of improved (55,149). The diagnosis of lymphomatous
the adjacent visceral pleural thickening, eliminating possi- involvement of the pleura is enhanced by performing flow
ble confusion with tumor infiltration into the pleura cytometry to establish the presence of a clonal cell popula-
(Fig. 9-59). More recently, it has been shown that whereas tion (150). A diagnosis of pleural malignancy may be elu-
CT remains superior to MR for visualizing calcified plaque, sive. Ryan et al. (151) emphasized that pleural effusions
MRI may be superior to CT for assessing pleural thicken- caused by malignancy may go undiagnosed even with care-
ing, pleural effusion, and increase in the amount of extra- ful examination of the pleura, lungs, and mediastinum at
pleural fat (143). Despite these findings, indications for thoracotomy. In their retrospective study of the outcome of
the use of MRI for evaluating patients with suspected 51 patients with pleural effusions of indeterminate cause at
benign asbestos-related pleural disease remain limited. thoracotomy, 25% were found later to have malignant
pleural disease.
Malignant effusions usually imply a poor prognosis,
Malignant Pleural Disease
although long-term survival has been reported, particu-
Evaluation of patients with suspected pleural malignancy larly in patients with metastatic breast carcinoma (58). In
is an important indication for the use of CT. This is largely addition to radiation and chemotherapy, specific therapies
a reflection of the frequency of malignant pleural disease that have been used to control malignant effusions include
estimated to occur in up to 50% of both outpatients and repeated thoracentesis, tube thoracostomy with or without
hospitalized patients presenting with effusions (12). In the use of sclerosing agents, and pleurectomy (152).
their series of 96 patients with carcinomatous involvement Although controversial in most centers, malignant effu-
of the pleura, for example, Chernow and Sahn (144) sions are initially treated by tube thoracostomy. Sclerosing
showed that an effusion provided the basis for the first agents that have been evaluated include tetracycline,
diagnosis of cancer in 44 (46%) of 96 patients. bleomycin, quinacrine, and talc, among others (152,153).
Carcinoma of the lung is the most common cause of As shown by Martini et al. (154), in patients in whom
pleural malignancy, constituting between 35% and 50% of initial tube thoracostomy is unsuccessful, pleurectomy
cases in most series. Breast cancer is also common, in may be effective, although this procedure can result in con-
some series equaling the incidence of lung cancer as a siderable morbidity and mortality.
cause of malignant effusions. In approximately 7% of
cases, the primary tumor site is unknown at the time of
initial diagnosis; these frequently prove to be adenocarci- Imaging Malignant Pleural Disease: Overview
nomas of “unknown origin” (58). Most unilateral malig- with Computed Tomography and
nant effusions are the result of pulmonary arterial tumor
Fluorodeoxyglucose–Positron Emission
emboli seeding the visceral pleura, with subsequent spread
Tomography Correlation
to the ipsilateral parietal pleura (Fig. 9-62) (145,146). In
distinction, bilateral malignant pleural effusions generally CT remains the imaging modality of choice for assessing the
result from tumor spread to the liver, with subsequent presence and extent of pleural tumor. Numerous articles
hematogenous seeding. Pleural fluid may also accumulate have documented the CT appearance of pleural tumors,
A, B C
D, E F
Figure 9-62 Early pleural metastases. Enlargement of representative 5-mm sections through the left lung in a patient with pancreatic
adenocarcinoma (A–C). Images demonstrate the presence of small nodules studding along the left major fissure, left lateral chest wall, and
are also identified over the left hemidiaphragm. Each nodule in isolation along the pleural surface is not suggestive; however, the presence
of multiple such nodules, particularly as the same areas were normal 3 months earlier (D–F), is highly suggestive of metastatic disease, even
in the absence of an effusion. Note that most metastatic malignant disease is the product of pulmonary artery seeding of the visceral pleura.
both benign and malignant (11–13,69,75,155–158). Speci- for CT findings of costal, mediastinal, and/or diaphrag-
fic CT signs that have been described for both primary and matic nodularity and circumferential pleural thickening.
metastatic disease include circumferential thickening of the However, in distinction to the results reported by Leung
pleura and focal and/or diffuse nodularity of the pleura. et al. (69), in this series, CT proved to have a much lower
Leung et al. (69), in an evaluation of 74 patients with sensitivity (7.5% vs. 41%) for circumferential pleural
proven diffuse pleural disease, including 39 cases of pleural thickening (11). Overall, only 52% of malignant effusions
malignancy, showed CT findings to be highly specific. With in this study proved to have findings other than pleural
the criteria of circumferential pleural thickening, nodular fluid, leading these investigators to the appropriate conclu-
pleural thickening, parietal pleural thickening greater than sion that the finding of a simple effusion, per se, does not
1 cm, and mediastinal pleural thickening alone or in com- rule out malignant disease (11). In keeping with this
bination, CT correctly diagnosed pleural malignancy in conclusion is the finding that, in studies comparing CT
28 of 39 cases (sensitivity, 72%; specificity, 83%). Signifi- findings with thoracoscopy, thoracoscopy has proven to
cantly, in this series, circumferential pleural thickening have greater sensitivity for detecting pleural nodules (156).
proved to be 100% specific in predicting malignant pleural Given the limitations of CT to diagnose malignant
disease. pleural disease definitively, it is not surprising that consid-
Similarly, Arenas-Jiminez et al. (11), in their study of erable attention has focused on alternate imaging modali-
211 patients with pleural effusions, including 93 with doc- ties. Of these, the most important has been the introduc-
umented malignant effusions (only two of which proved tion of FDG-PET imaging (159–169). To date, FDG-PET
to be due to malignant mesotheliomas), reported similar scanning has been shown to be a sensitive method for
findings with specificities ranging between 96% and 100% both detecting and characterizing pleural abnormalities,
both benign and malignant. Although benign causes of Bronchogenic Carcinoma

FDG uptake are numerous and include infection, nonspe-
cific pleuritis, and recent surgery, among others, the degree As previously noted, bronchogenic carcinoma is the most
of uptake as measured by the SUV is rarely more than 2.0 frequent cause of a malignant pleural effusion, especially
(163). Transudative effusions are rarely FDG positive. In adenocarcinoma, occurring in up to 15% of patients first
distinction, as is discussed in greater detail later in the seen with non–small cell lung cancer. As discussed in detail
appropriate sections, patients with malignant pleural dis- in Chapter 7, evidence of pleura involvement significantly
ease, including both metastatic disease and mesothe- alters the staging, prognosis, and therapy of these tumors.
liomas, have markedly abnormal FDG-PET studies. Several The pathways by which the pleura become involved with
reports have documented that FDG-PET can accurately dif- tumor have been described by Heitzman et al. (170).
ferentiate benign pleural plaques or nonspecific pleuritis Tumor may extend to the pleura secondary to reversal of
from malignant pleural disease (164–166). In addition, the centripetal flow of lymph (commonly caused by pri-
FDG-PET has proved to have greater sensitivity than either mary central tumors) or tumorous involvement of hilar
CT or MR in detecting remote disease, including both nodes. Central venous obstruction also may be present.
mediastinal nodes and extrathoracic disease. As reported Alternatively, peripheral tumors may directly invade the
by Duysinx et al. (162), in a study of 98 consecutive pati- adjacent pleura, either by tumor growth along peripheral
ents with either pleural thickening or an exudative effu- perivascular–lymphatic sheaths or by direct invasion of
sion, FDG-PET revealed uptake in 61 of 63 patients with the adjacent pleura with subsequent pleural seeding (Fig.
subsequently documented malignant pleural disease, 9-63). Evidence of pleural seeding may be discovered
while correctly classifying 31 of 35 benign lesions. in sites far removed from the primary tumor, including
the mediastinal pleura (Fig. 9-63). Interestingly, not
A B
Figure 9-63 Pleural seeding: CT evaluation. Contrast-enhanced CT

in a patient with advanced metastatic prostate carcinoma and symp-
toms of cord compression. A: A 5-mm section through the upper lung
demonstrates irregular enhancing nodular disease of the pleura
surrounding an effusion. Additional enhancing soft tissue is destroying
the posterior aspect of the vertebral body, invasion of the epidural
space accounting for the neurologic symptoms. At the lung base (B),
a large enhancing pleural deposit is seen, but also discontinuity of dis-
ease with intervening areas of normal pleura, a characteristic of
metastatic pleural disease. A coronal 5-mm multiplanar reconstruction
(MPR) (C) demonstrates the overall extent of disease; despite the
presence of a chest drain, pleural fluid causes mediastinal shift. The
large basilar pleural metastasis is again seen, as are smaller peripher-
ally enhancing lesions in the medial left upper hemithorax.
Extrathoracic extension along the left lateral chest wall and axillary
C nodes also is seen.
infrequently, tumor may involve the pleura without pro- (Figs. 9-64 to 9-66). It has been suggested that this prob-
ducing pleural fluid; these tumors may also involve the lem may be solved either with images obtained after
chest wall. therapeutic pneumothorax or by using inspiration-
Unfortunately, assessment of pleural involvement is expiration images. Although the presence of air within
limited when tumors, both peripheral and central, abut the pleural space helps to outline the pleural surfaces
the pleura but do not appear directly to invade either the (Fig. 9-62), adherence of lung to the pleural surfaces is
chest wall or mediastinum (171–175). In most cases, nonspecific. This may result from either tumor infiltra-
definitive evaluation usually still requires pleural biopsy tion of the parietal pleura and chest wall or previous
A B
C D
Figure 9-64 Peripheral lung cancer: evaluation with CT with 5-mm sections (A–C) through the a peripheral cavitary lung carcinoma with
apparent pleural thickening on posteroanterior and lateral radiographs (not shown). Mediastinal windows (B) suggest a connection to the
right upper lobe airways (arrow in B); this fistulization is confirmed on lung windows of the same section (C). The mass clearly abuts the
pleura, which is thickened. A small loculated lateral pleural fluid collection (arrow, A) and other areas of preserved extrapleural fat (B) sug-
gest that no gross chest wall involvement is present. These findings were confirmed at surgery. A CT-PET evaluation of the same patient
(D) confirms biologic activity limited to the rim-enhancing tumor.
B C
Figure 9-65 Neoplastic pleural invasion. A–C: Axial, coronal, and sagittal CT images through the right upper lobe, respectively, show ev-
idence of a peripheral tumor abutting the pleura without evidence of rib or chest-wall invasion. No effusion is present. Correlative fluo-
rodeoxyglucose–positron emission tomography (FDG-PET) study (not shown) showed marked uptake of tracer in the tumor anatomically
corresponding to CT findings without evidence of pleural involvement. In this case, neither CT nor FDG-PET allows definitive staging, as
neither can determine whether tumor involves the parietal pleura. At surgery, this proved to be a non–small cell carcinoma limited by the
visceral pleura only.
adhesions. Similarly, some have suggested that the use of extrapleural fat, usually associated with some scarring and
three-dimensional surface reconstructions may be of volume loss in the adjacent lung parenchyma (Fig. 9-67).
value to identify pleural extension of tumor. To date, The apical pleura itself is rarely sufficiently thickened to
none of these applications has gained widespread clini- cause the radiographic changes to which it is typically
cal acceptance. attributed.
For similar reasons, CT is also limited in assessing Another problem is encountered in evaluating patients
patients with superior sulcus tumors because of problems with central endobronchial lesions with secondary
related to cross-sectional imaging through the lung apices obstructive pneumonitis or apparent lobar collapse:
(see Fig 7-31). It should be noted, however, that CT is of Differentiation of infiltrate from tumor may be difficult.
value for assessing patients with asymmetric apical pleural As in patients with Pancoast tumors, MR is usually more
thickening. Typically the result of prior tuberculous infec- accurate than CT in predicting parietal pleural and chest-
tion, “apical pleural thickening” is a misnomer; in reality, wall involvement in patients with underlying atelectasis
it is the result of marked expansion of the apical (Fig. 9-68) (176–180).
A B
C D
Figure 9-66 Evolving bronchopleural fistula: CT evaluation. A:

5-mm CT section demonstrates complete occlusion of the left
lower lobe bronchus by a tumor mass and collapse of the left
lower lobe. Three months later, after chemotherapy, CT at the
same level (B) demonstrates that although some improved aera-
tion of the left lower lobe airways occurred, a new pleural cavity
is seen inferomedially. Axial (C), coronal (D), and 3D volume-
rendered images (E) demonstrate a connection between the left
lower lobe airway and an infected pleural cavity. Additional exten-
sive pleural disease (not shown) complicated surgical intervention;
however, the use of thin sections and multiplanar reconstructions
can be invaluable in both the characterization of bronchopleural
E fistulas and planning their surgical management.
It has been shown that in select cases FDG-PET imaging measuring more than 10 HU with or without solid pleural
may be of value for evaluating the pleura in patients with abnormalities in the absence of rib or chest-wall abnor-
suspected pleural metastases. In one study of 92 patients malities) and truly negative in 29%, the sensitivity of
either with documented non–small cell lung cancer, FDG-PET was 100%, the specificity was 71%, the PPV was
whereas contrast-enhanced CT was interpreted as indeter- 63%, the NPV was 100%, and the accuracy was 84%,
minate in 71% of cases (defined as having effusions respectively (164). This led these authors to conclude that
A B
Figure 9-67 Apical pleural thickening shown in 3-mm sections from a contrast-enhanced CT pulmonary angiogram (CTPA) study. CT
sections through both lung apices (A and B) show characteristic appearance of so-called “apical pleural thickening.” What appears radio-
logically to be thickened apical pleura is in reality due to marked expansion of the apical extrapleural fat, usually in association with some
element of volume loss and scarring in the adjacent lung parenchyma. This is exactly analogous to what occurs in patients with prior pleural
infection (compare with Figs. 9-33 to 9-35), usually resulting from tuberculosis. The postgranulomatous etiology is confirmed in this case by
the presence of left apical calcifications adjacent to focal scarring and the extrapleural fat.
a negative FDG-PET scan in association with an indetermi- astinal adenopathy and for identification of extrathoracic
nate CT study was diagnostic of a benign pleural process, disease (181). Of particular note is that FDG-PET is equally
whereas a positive FDG-PET scan was consistent with nondiagnostic as either CT or MR in assessing direct pari-
malignant disease. Similar findings have been reported etal pleural invasion in patients with bulky subpleural
by Erasmus et al. (160), who noted a sensitivity of 92%, a lesions.
specificity of 67%, an NPV of 100%, and a PPV of 63%,
respectively.
Malignant Pleural Mesothelioma
Despite these data, the role of FDG-PET specifically to
evaluate the pleura in patients with non–small cell lung Although MPM is a rare neoplasm representing fewer than
cancer remains limited, with the main indication for 5% of pleural malignancies, its incidence is increasing
FDG-PET still primarily for an initial assessment of medi- worldwide, especially in Europe and Australia, as well as
A B
Figure 9-68 Peripheral lung cancers: MR evaluation. A: T1-weighted sagittal MR image shows a peripheral tumor that abuts the pleural
surface. Note the complete preservation of the extrapleural fat adjacent to the tumor (arrow), indicating that the chest wall is not grossly in-
filtrated by tumor. B: T1-weighted sagittal image in a different patient with a superior sulcus tumor. In this case, tumor has clearly invaded
the adjacent chest wall, obliterating the normal extrapleural fat planes (arrow). Although CT is superior to MR in evaluating bony pathology,
because of its greater contrast resolution, MR has proven more valuable than CT in detecting tumor infiltration into the soft tissues of the
chest wall, especially infiltration into chest-wall fat and adjacent muscles.
throughout the developing world (182). The association factor-1 (TTF-1) are indicative of adenocarcinoma, instead.
between malignant mesotheliomas and asbestos exposure Closed-needle biopsy may also be value, prospectively
is well documented (183). Unlike the association between diagnosing 25 (83.3%) of 30 patients in one series, result-
asbestos and bronchogenic carcinoma, the development ing in local seeding in only 20% of cases (185).
of malignant mesotheliomas does not appear to be dose In association with cytologic diagnosis, recent attention
related; tumors have been documented after only relatively has also focused on the potential use of serum markers,
trivial environmental or household exposure. It is proba- especially serum mesothelin-related protein (SMRP), which
bly for this reason that malignant mesotheliomas are is elevated in 84% of patients with malignant mesothe-
frequently identified in patients without evidence of either liomas, including more than 60% at the time of diagnosis
pleural plaques or pleuropulmonary fibrosis (103). The (184). In distinction, SMRP is positive in fewer than 2% of
risk of mesothelioma is related to the type of fiber to patients with other diseases involving the pleura, both
which patients are exposed. Crocidolite poses a far greater benign and malignant, making SMRP a highly specific indi-
risk for development of a mesothelioma than either cator of mesothelioma. SMRP has also been shown to
amosite or chrysolite (183). The latency period for the increase with disease progression, making it a potential
development of malignant mesothelioma is long, gener- marker for monitoring response to therapy (182). Other
ally between 20 and 30 years. Approximately 80% of serum markers that have been evaluated include osteopon-
mesotheliomas are pleural, whereas 20% are peritoneal tin, which has been shown to be of value to distinguish
(183). Some 80% occur in men, with patients typically individuals with exposure to asbestos without cancer from
first seen with dyspnea and chest pain accompanied by a those with exposure and mesotheliomas, with a sensitivity
pleural effusion (184). of 84.6% (186).
The diagnosis of mesothelioma is often difficult. In It should be emphasized that although these tech-
particular, differentiating mesothelial neoplasia from niques allow a confident diagnosis in the majority of cases,
mesothelial hyperplasia on the one hand and metastatic definitive staging with accurate estimation of local disease
adenocarcinoma on the other may be problematic (182). extent is possible only during thoracotomy.
Traditionally, the diagnosis of MPM has required extensive Histologically, three forms of diffuse malignant mesothe-
histopathologic examination to provide adequate tissue, lioma have been recognized: epithelial (55% to 65%),
either via VATS or surgical thoracotomy (185). However, mixed (20% to 25%), and sarcomatoid (10% to 15%). Of
over the past decade, there has been growing acceptance of these, most common are epithelioid tumors, with sarcoma-
cytology and/or fine-needle aspiration in cases without toid tumors having the worst prognosis: overall, survival in
effusion to establish the diagnosis, with reported sensitivi- patients with MPM is poor, with medial survival of 8 to
ties of 33% to 84% of cases (149). In addition to the find- 18 months generally cited (187). Mesotheliomas arise on
ing of malignant cells, specific immunochemical markers the parietal pleural surface and proceed to grow by contigu-
for mesothelioma include calretinin, which identifies cells ous spread throughout the pleura, including the mediasti-
as mesothelial in origin, and epithelial membrane antigen nal pleural surfaces and the fissures (Figs. 9-69 through
(also known as CA15–3) (182). Other markers of proven 9-74). Local invasion is common, with direct spread to
value include cytokeratin and mesothelin, whereas carci- the chest wall and hemidiaphragms in particular noted
noembryonic antigen (CEA) and thyroid transcription (Figs. 9-69 through 9-74). Mesotheliomas also may
A B
Figure 9-69 Malignant mesothelioma. A, B: Sequential contrast-enhanced sections show nodular thickening involving both pleural
surfaces (arrows in A and B) associated with loculated pleural fluid. The mediastinal pleura is subtly involved as well (curved arrow in A).
Note presence of focal chest-wall invasion anteriorly (curved arrow in A). Considerable volume loss on the left is typical of diffuse neoplastic
involvement of the pleura.
A B
Figure 9-70 Malignant mesothelioma. A, B: Sequential contrast-enhanced sections show extensive pleural thickening and nodularity,
more pronounced inferiorly, affecting all pleural surfaces, including the mediastinal pleura. In comparison to the case shown in Figure 9-69,
only a small quantity of pleural fluid is present. This is atypical, as most mesotheliomas are associated with large pleural effusions. Note that
in this case, contralateral calcified pleural plaques appear, consistent with prior exposure to asbestos. Extensive invasion of the chest wall
and diaphragm in this case renders this tumor unresectable.
metastasize hematogenously to mediastinal lymph nodes, particular the potential role of extrapleural pneumonec-
as well as to the contralateral lung (183). A rare variant— tomy followed by chemoradiation to allow long-term
localized (solitary) mediastinal malignant mesotheliomas— survival in a selected population of patients (190,191) and
has also been described (188). This entity presents as a well- to bring these into line with staging systems using the T, N,
defined focal pleural mass that is easily mistaken for a and M descriptors used for other neoplasms, including
solitary fibrous tumor of the pleura. As discussed later, with non–small cell lung cancer, a new staging system was pro-
the introduction of FDG-PET scanning, it is becoming posed by the International Mesothelioma Interest Group in
increasingly clear that a far greater number of cases of MPM 1995 and subsequently adopted by the American Joint
than previously thought result in extrathoracic disease, Committee on Cancer (AJCC) and the Union International
occurring in up to 25% of cases (167). Contre le Cancer (UICC) (192,193). With TNM descriptors,
At least five staging systems of MPMs have been variously MPM has been grouped into four stages (Tables 9-1 to
proposed (189). Reflecting recent advances in therapy, in 9-3). Most important is the determination of T status. T1
describes early tumor confined to the ipsilateral hemitho-
rax, primarily involving the parietal pleura without evidence
of pleural adhesions. In distinction, T2 designates tumor
usually involving the diaphragmatic muscle that cannot be
resected without removing the underlying lung. T3 tumors
are those with evidence of locally advanced tumor extend-
ing into the endothoracic fascia, mediastinal fat, or the
surface of the pericardium that may nonetheless still be
amenable to surgical resection. T4 lesions are those with
locally advanced, surgically unresectable tumor caused by
diffuse extension into the chest wall, direct extension
through the diaphragm into the abdomen, or direct ext-
ension with encasement of mediastinal organs or the con-
tralateral pleural space.
Both N and M status are defined equivalent to their des-
ignation for staging non–small cell lung cancer (192). In
addition to stratifying patients into homogeneous group-
ings, the better to assess the impact of prospective clinical
trials, this staging system is designed to identify potentially
surgically resectable tumors. Reflecting recent advances in
therapy, in particular, the potential role of extrapleural
Figure 9-71 Malignant mesothelioma: chest-wall involvement.
Contrast-enhanced CT scan shows extensive tumor involving the pneumonectomy followed by chemoradiation allows
left lower lobe, ribs, and chest wall. This appearance is nonspecific long-term survival in a selected population of patients
and initially suggested a primary lung cancer with secondary (190,191).
chest-wall invasion. Malignant mesothelioma was biopsy proven.
A B
Figure 9-72 Malignant pleural mesothelioma with intraperitoneal mesothelioma. A: Contrast-enhanced 7-mm section shows diffuse
nodular enhancing tissue encasing the contracted right hemithorax. Bilateral effusions are present, and fine linear pleural calcifications pos-
teromedially attest to prior asbestos exposure. B: The upper abdominal image demonstrates extensive nodular thickening of the greater
omentum anterior to the colon, with numerous enhancing nodular foci in the dependent subhepatic ascites. Intraperitoneal mesothelioma is
a rare manifestation of asbestos-related disease and usually occurs as a sequela of direct extension transdiaphragmatically, possibly through
congenital pleuroperitoneal connections.
A B
Figure 9-73 Malignant mesothelioma: MR evaluation. A, B: T1- and T2-weighted images, respectively, through the lower chest in a
patient with documented malignant mesothelioma. The pleura is circumferentially thickened and nodular, with evidence of tumor infiltration
into the major fissure (arrows in A and B). The multiplanar capacity of MRI may be of particular advantage in malignant mesothelioma in eval-
uating structures such as the chest wall, diaphragm, and lung apex, the evaluation of which may be problematic on CT.
Imaging Features
The appearance on CT of both “benign” and malignant
mesotheliomas has been well described (103,105,122,
194–202). Characteristically, mesothelioma permeates the
pleural space, causing the pleura to become markedly
thickened, irregular, and nodular (Figs. 9-69 through
9-74). The tumor often encircles the lung, which may then
become entrapped (Figs. 9-69 through 9-74). Effusions are
present in up to 80% of cases. Ancillary findings include
mediastinal and hilar lymphadenopathy and pulmonary
nodules, which have been reported to be present in up to
60% of cases (105). Rarely, foci of calcification that have
proved to represent areas of osteogenic sarcomatous
degeneration may be identified (203).
In a review of 50 cases of documented diffuse malig-
Figure 9-74 Malignant mesothelioma. Contrast-enhanced sec- nant mesothelioma, Kawashima and Libshitz (200)
tion shows diffuse nodular enhancing tissue encasing the right found 92% of cases had pleural thickening, 86% had
hemithorax, which is contracted. The appearance is highly sugges-
tive of malignant mesothelioma, although it may also occur in
thickening of the pleural surfaces of the interlobar
metastatic adenocarcinoma. The presence of fissural involvement fissures, and 74% had pleural effusions, but only 20%
is the only site where the definitive presence of visceral pleural dis- had evidence of pleural calcifications (200). Similarly, in
ease can be radiologically demonstrated. The tumor is hence at
least a T2 lesion with a significantly worse prognostic outcome
a recent study of 84 patients with proved MPM, Sahin
than parietal involvement alone. et al. (202) also found that the most common findings
TABLE 9-1
MESOTHELIOMA STAGING SYSTEM: T DESIGNATION
Status Region involved Characteristics
T1a Limited to ipsilateral parietal pleura, including No involvement of visceral pleura

mediastinal and diaphragmatic pleura
T1b Ipsilateral parietal pleura, including mediastinal
and diaphragmatic pleura. Scattered foci of
tumor also involving visceral pleura
T2 Each ipsilateral pleural surfacea At least one of the following:
• Involvement of diaphragmatic muscle
• Confluent visceral pleural tumor (including fissures) or extension
of tumor from visceral pleura into underlying pulmonary
parenchyma
T3 Locally advanced but potentially At least one of the following:
resectable tumor; each ipsilateral • Involvement of endothoracic fascia
pleural surfacea • Extension into mediastinal fat
• Solitary, completely resectable focus of tumor extending into
soft tissues of chest wall
• Nontransmural involvement of pericardium
T4 Locally advanced technically At least one of the following:
unresectable tumor; each • Diffuse extension or multifocal masses of tumor in chest wall, with
ipsilateral pleural surfacea or without associated rib destruction
• Direct transdiaphragmatic extension of tumor to peritoneum
• Direct extension of tumor to contralateral pleura
• Direct extension of tumor to one or more mediastinal organs
• Direct extension of tumor into spine
• Tumor extending through to internal surface of pericardium with
or without pericardial effusion, or tumor involving myocardium
aParietal,
mediastinal, diaphragmatic, and visceral pleura.
See reference 192.
tively evaluated 41 consecutive patients with MPM by

using both CT and MR, and found that although both CT
TABLE 9-2
and MR had more than 90% sensitivity for detecting signs
MESOTHELIOMA STAGING SYSTEM: N AND M
of unresectability, corresponding specificity measured only
DESIGNATIONS between 25% and 50%. This remained true regardless of
Designation Description whether the location of the tumor evaluated was along the
diaphragm, chest wall, or mediastinum (205). Although
NX Regional lymph nodes not assessable these authors did note a potential benefit of MR as com-
N0 No regional lymph node metastases
N1 Metastases in ipsilateral bronchopulmonary
pared with CT due to improved visualization of the pleural
or hilar lymph nodes surfaces, owing to the ability to acquire images in multiple
N2 Metastases in subcarinal or ipsilateral planes (Fig. 9-73), this advantage has been lessened in the
mediastinal lymph nodes, including era of MDCT.
ipsilateral internal mammary nodes Given the limitations of both CT and MR to stage
N3 Metastases in contralateral mediastinal,
contralateral internal mammary, and
MPM accurately, recent attention has focused on the
ipsilateral or contralateral supraclavicular potential use of FDG-PET (Fig. 9-75) (167,168,206,207).
lymph nodes To date, FDG-PET has proved most useful in initial
MX Distant metastases not assessable staging for identifying suggestive nodes otherwise unsus-
M0 No distant metastases pected on CT and MRI, and occult extrathoracic disease:
M1 Distant metastases present
in distinction, FDG-PET has proved of far less value for
See reference 192. assessing T stage because of poor spatial resolution
(167). Although primary tumor in MPM is characteristi-
cally FDG-avid, FDG-PET is inaccurate in differentiating
included pleural thickening present in more than 90% of locally advanced T3 tumor, for which surgery may be
cases, with interfissural involvement, pleural effusions, indicated, from diffuse T4 lesions in which surgery is
and ipsilateral volume loss noted in more than 70% of contraindicated. Flores et al. (208), for example, by using
cases. Malignant mesotheliomas may rarely present as FDG-PET alone, reported a sensitivity of only 19% for
more localized masses, making CT distinction from detecting T4 disease. In this same study, the sensitivity
benign fibrous tumors of the pleura extremely difficult. It of FDG-PET for detecting nodal disease proved to be
should also be noted that no CT-specific signs exist for 11% only.
mesothelioma; almost any malignancy involving the Similar results have been reported by Erasmus et al.
pleura can simulate the appearance of a mesothelioma (207). With state-of-the-art integrated PET-CT imaging to
(136,204). evaluate retrospectively 29 patients with MPM deemed
Current options for managing MPM include surgery potential candidates for EPP, the sensitivity, specificity,
[including diagnostic thoracoscopy, palliative surgery PPV, NPV, and accuracy of CT-PET for T4 disease were
including partial pleurectomy, and curative surgery with 67%, 93%, 86%, 82%, and 83%, respectively (207).
extrapleural pneumonectomy (EPP)]; chemotherapy, espe- CT-PET proved particularly limited in assessing transdi-
cially with premetrexed or gemcitabine in combination aphragmatic extent of tumor. In this same study, CT-PET
with cis and/or carboplatin); and intensity-modulated also proved of limited value for staging N disease, with a
radiotherapy, largely used for local control (182). At pres- sensitivity, specificity, PPV, NPV, and accuracy of CT-PET
ent, the main value of both CT and MR is to attempt to for surgically staged N2 disease of 38%, 78%, 60%, 58%,
(a) delineate the intra- and extrathoracic extent of disease; and 59%, respectively. Based on these data, these authors
(b) determine the best therapeutic approach, in particular concluded that accurate preoperative staging in patients
the feasibility of performing an extrapleural pneumonec- considered for extrapleural pneumonectomy should
tomy; and (c) monitor the results of therapy. include extended surgical staging, including cervical
Unfortunately, attempts to stage patients accurately by mediastinoscopy, laparoscopy, and peritoneal lavage.
using either CT or MR have proved of limited value. As Perhaps most striking has been the use of FDG-PET to
outlined by Patz et al. (205), radiologic criteria for deter- identify otherwise unsuspected distant metastases (Fig. 9-75)
mining resectability include preservation of extrapleural (207–209). In the study of Erasmus et al. (207), in 7 (24%)
fat planes, absence of an extrapleural soft tissue mass, of 29 patients, CT-PET identified extrathoracic metastases
smooth undersurface of the diaphragm, and normal CT that were otherwise unsuspected.
attenuation values and MR signal characteristics of ad- In addition to initial staging, imaging also plays a cri-
jacent structures. By comparison, radiologic features tical role in assessing response to therapy. Unlike with
indicative of unresectability include tumor encasing the non–small cell lung cancer, assessing response in patients
diaphragm, invasion of extrapleural soft tissues or fat, or with mesotheliomas has proved extremely problematic.
invasion, separation, or destruction of adjacent osseous This is because, unlike lung cancer, MPM tends to spread
structures (205). With these criteria, these authors prospec- nonuniformly throughout the pleural space and chest
A, B C
D, E F
Figure 9-75 Malignant mesothelioma: CT–positron emission tomography (PET) correlations. See Color Figure 9-75C,F,I. A–C: Images
from a 62-year-old with documented epithelioid mesothelioma. Coronal CT (A), positron emission tomography (PET) (B), and integrated CT-
PET (C) show diffuse uptake of (18F)-fluoro-2-deoxy-D-glucose (FDG) in primary left pleural tumor and focal area of increased uptake in the
abdomen (arrow), localized to an abdominal lymph node, precluding patient from extrapleural pneumonectomy (EPP). (Case courtesy of
Jeremy Erasmus, MD, M.D. Anderson Hospital, Houston, Texas.) D–F: A 63-year-old man with an epithelioid mesothelioma with bone occult
metastases. Coronal CT (D), PET (E), and integrated CT-PET (F) images show circumferential FDG uptake indicative of extensive pleural in-
volvement. Note focal increased uptake in chest wall at sites of a prior video-assisted thoracotomy (VATS) procedure seen only with PET and
on integrated CT-PET images. G–I: A 65-year-old man with documented malignant pleural mesothelioma. Sagittal CT (G), PET (H), and inte-
grated CT-PET (I), respectively, show extensive FDG uptake throughout the pleura. Note the evidence of transdiaphragmatic invasion
shown to better advantage on the PET and integrated CT-PET images compared with the sagittal CT image, a finding subsequently docu-
mented at laparoscopy. (D–I, Case courtesy of Mylene Truong, M.D., Anderson Hospital, Houston, Texas.)
G, H I
wall, making bidimensional measurements particularly Metastatic Disease

difficult (210). A lack of consistent landmarks, especially
below the carina, also makes follow-up assessment Pleural metastases, apart from those caused by bron-
difficult. Recent advances in chemotherapy, in particular, chogenic carcinoma, occur most frequently from primary
use of pemetrexed and/or gemcitabine in combination neoplasms of the breast, gastrointestinal tract (including
with cisplatin, resulting in significant improvement in pancreas), kidneys, and ovaries (Figs. 9-76 through 9-78)
survival in patients with MPM, have made the develop- (145,146,148). Invasive thymoma also frequently involves
ment of reliable methods of measuring treatment
response imperative (184). In an attempt to overcome
previous limitations for assessing disease response, modi- TABLE 9-3
fied RECIST criteria have recently been proposed MESOTHELIOMA: STAGING
(211,212). As currently proposed, the modified RECIST
Stage Status Lymph node Metastases
criteria use unidimensional measurement of tumor
thickness assessed perpendicular to the chest wall or Ia T1a N0 M0
mediastinum with two measurements obtained at each of Ib T1b N0 M0
three separate sites at least 1 cm apart (212). In addition, II T2 N0 M0
III Any T3 Any N1 or N2 M0
unidimensional measurements of enlarged nodes or IV Any T4 Any N3 Any M1
discrete chest-wall masses are obtained.
Figure 9-76 Pleural metastases: breast cancer. A: Contrast-

enhanced section shows subtle asymmetric thickening and nodularity
along the posterior right pleural surface (arrow) and anterior mediasti-
nal pleural surface in this patient with documented pleural metastases
from breast cancer. B, C: Contrast-enhanced CT scans through the
carina and the lower lobes in a patient after a right mastectomy. In
addition to a moderate-sized right pleural effusion, enlarged internal
mammary lymph nodes can be identified on the right (small arrow).
The pleural surface is nodular and thickened superiorly (large arrow in
B). Inferiorly, discrete pleural masses can be seen easily separable
from the surrounding effusion (arrows in C). Metastatic breast cancer
A was removed at biopsy.
B C
the pleura, usually by contiguous invasion (Figs. 9-79 and large pleural effusions in which the foci of malignancy
9-80). Malignant thymomas may also cause pleural seed- are difficult to identify. In a study of 86 patients evalu-
ing, resulting in the development of discrete pleural ated with contrast-enhanced CT reported by O’Donovan
masses, often at a distance far removed from tumor within and Eng (155), evidence was found of either smooth or
the anterior mediastinum (Figs. 9-79 and 9-80; see also irregular pleural thickening in 62% of cases. Although
Chapter 4, Figs. 4-22E and 4-30 to 4-33) (213,214). loculated fluid was present in 40%, focal soft tissue
In patients with metastatic disease, pleural effusions may masses in association with fluid could be identified in
develop for a variety of reasons. These include increased per- only 10%.
meability of the capillaries supplying tumor implants, as MR findings in patients with metastatic disease have
well as increased capillary permeability caused by pleuritis been described (Fig. 9-78) (215). In a study of 45 patients
associated with obstructive pneumonitis if present, direct evaluated with both CT and MR, Falaschi et al. (215) noted
tumor erosion of pleural blood and lymphatic vessels, and that although little difference was noted between the mor-
decreased removal of pleural fluid caused by mediastinal phologic appearance of tumors on both CT and MR, high
lymph node infiltration (58). signal intensity was observed on T2-weighted images in all
The same wide variation in the appearance on CT malignant lesions, but in only two benign lesions (sensi-
described for mesotheliomas may be seen with metastatic tivity, 100%; specificity, 87%). It would appear that, at
pleural disease (see Figs. 9-76 through 9-78). Pleural least for a subset of lesions that are predominantly fibrous
metastases may cause marked thickening and nodularity in nature, including pleural plaques, fibrothoraces, and
of the pleura associated with only a small quantity of fibromas, MR may play a limited role as an ancillary
pleural fluid. Alternatively, pleural metastases may cause method of evaluation.
A B
Figure 9-77 Peripherally enhancing pleural metastases: Contrast-enhanced CT sections in two different patients, demonstrating periph-
eral rim enhancement of pleural masses. This unusual appearance is usually the result of metastatic disease from a vascular primary, includ-
ing renal cell carcinoma among others. In (A), this proved secondary to melanoma; in (B) the enhancing lesion in the costophrenic sulcus
surrounded by pleural fluid is secondary to an aggressive poorly differentiated lung cancer.
Figure 9-78 Pleural metastases: MR evaluation. T1-weighted section in a patient with radiographic evidence of a nonspecific right-sided
effusion (not shown). A pleural effusion is apparent on the right side, compressing the adjacent right lower lobe (arrow). Additionally, an
enlarged anterior mediastinal lymph node is embedded in fat (curved arrow), strongly suggesting malignant disease. Adenocarcinoma of
the pleura was documented.
A B
C D
Figure 9-79 Invasive thymoma with pleural metastases: CT appearances. A 60-year-old woman with dyspnea. A, B: The 5-mm postcon-
trast CT images demonstrate a homogeneous prevascular mass (A). The left hemidiaphragm is elevated, suggestive of phrenic nerve paral-
ysis from tumor that extended along the mediastinum. The elongated mass in the left posterior costophrenic sulcus is a pleural deposit;
these lesions are often termed “drop metastases.” C, D: Enhanced CT images from a different patient from that in A and B. The entire right
hemithorax demonstrates multiple soft tissue nodules and masses. The primary mass is identified in the right paracardiac region (arrow).
Both the mass and a right supradiaphragmatic metastasis demonstrate punctate calcifications, a relatively common feature of invasive
thymoma. The liver is indented by a large soft tissue mass (D). The demonstrated scalloping indicates that the lesion is extrinsic to the liver;
in this case, it is secondary to a massive pleural deposit arising above the diaphragm and extrinsically compressing the liver. At surgery,
diaphragmatic but no hepatic invasion was identified (D).
Figure 9-80 Pleural splenosis: CT diagnosis. A 45-year-old

patient was evaluated for pleura-based nodules detected on chest
radiograph. The 5-mm noncontrast CT images through the lower
chest demonstrated multiple left-sided pleura-based nodules (A). A
gunshot fragment is seen in the posterior pleural space (B) as well
as multiple subcutaneous gunshot fragments. The pancreatic tail is
in contact with the posterior abdominal wall (C), highlighting the
absence of the spleen and its hilum, in which the pancreatic tail nor-
mally terminates. The patient had forgotten to mention a splenec-
tomy after a prior gunshot injury, confirming the CT diagnosis of
pleural splenosis. Pleural splenosis may mimic pleural metastatic
disease; however, usually the absence of effusions and the spleen
confirms the diagnosis. In problematic cases, sulfur colloid or heat-
A denatured red blood cell scintigraphy is confirmatory.
B C
Pleural Lymphoma effusion; and direct pleural infiltration as a result of

involvement of adjacent ribs with subsequent extension
It has been estimated that approximately 10% of malignant through the chest wall and pleura, or secondary to involve-
pleural effusions are the result of lymphoma (58). Pleural ment of the adjacent lung (58,153).
effusions are more likely to occur in non-Hodgkin lym- Lymphoma frequently causes subpleural deposits of
phomas (NHL) than with Hodgkin disease (HD) and are tumor in the form of either nodules or plaques (Fig. 9-82)
more likely to occur in patients with extensive disease (Fig. (219). In a retrospective radiographic review of 112 nonse-
9-81) (181,216,217). Although distinctly uncommon at the lected patients with documented histiocytic lymphoma,
time of initial diagnosis, in up to 30% of patients with HD, Burgener and Hamlin (220) showed that pleural involve-
effusions may eventually develop (218). In patients with ment was present in 18% of cases, including four cases with
lymphoma, pleural effusions result from a variety of causes, localized pleural plaques. In another series of 71 patients
including impaired lymphatic drainage caused by obstruc- with both documented HD (n 47) and NHL (n 24)
tion of hilar and mediastinal lymph nodes; obstruction evaluated by CT, solid pleural manifestations were present
of the thoracic duct, frequently associated with a chylous in 31%.
A B
Figure 9-81 Pleural lymphoma: effusions. A 50-year-old with a history of treated non-Hodgkin lymphoma evaluated for dyspnea with
a CT-pulmonary angiogram study to exclude pulmonary embolus (A and B). Note that the central right arteries are extrinsically narrowed by
the large mediastinal mass that also occludes the right lower lobe airways. A large effusion is present without significant pleural enhance-
ment or nodularity. Aspiration of the pleural fluid confirmed a recurrence of a large B-cell lymphoma. Although effusions associated with
lymphoma are typically nondescript, airway occlusion in this case is slightly unusual, more characteristic of bronchogenic malignancy.
C: Contrast-enhanced CT in a different patient, demonstrating a simple effusion associated with posterior mediastinal adenopathy, an
appearance suggestive of the established diagnosis of non-Hodgkin lymphoma (continued).
C Figure 9-81 (continued )
A B
Figure 9-82 Pleural lymphoma. A: Contrast-enhanced sections

show nondescript right-sided pleural effusion; note the absence of
pleural thickening or nodularity. B: Section at the level of the lower
thorax shows markedly enlarged posterior mediastinal nodes
(arrows). This combination of effusion with enlarged retrocrural
and retroperitoneal nodes is especially suggestive of non-Hodgkin
lymphoma, subsequently proven. C: Contrast-enhanced section in
a different patient from that shown in A and B shows extensive
posterior adenopathy in a patient with known non-Hodgkin lym-
phoma associated with an otherwise nondescript pleural effusion.
This appearance is also characteristic and should suggest the cor-
C rect diagnosis in most cases.
Documentation of the presence of pleural disease can therapy in this study had their radiation ports changed on
be of particular benefit in the development of appropriate the basis of MR findings.
treatment strategies, both after initial diagnosis and as a
means for follow-up (221).
Postpneumonectomy Space
In selected cases, the ability to identify chest-wall in-
volvement may have important therapeutic and prognostic Postoperative complications are frequent after pneumonec-
implications. In this regard, MR has been shown to be more tomy, both in the immediate and the remote postsurgical
accurate than CT for predicting the presence and assessing period (223). Altogether, it has been estimated that pneu-
the extent of chest-wall invasion. In one retrospective study monectomy has a combined morbidity and mortality of
comparing CT with MR in the evaluation of 57 patients, 22 approximately 30% (190).
of whom had chest-wall or pleural involvement, chest-wall CT is especially efficacious in evaluating the postpneu-
disease was identified on MR in 20 patients compared with monectomy space (Figs. 9-83 through 9-85; and Fig. 7-27)
only seven with CT, whereas pleural involvement was de- (224). In one series of 22 patients after pneumonectomy
tected on MR in 14 patients compared with only 5 with CT evaluated by CT, residual fluid was present in 13 (60%) of
(222). Most important, 3 of 15 patients receiving radiation 22 cases, even years after surgery, whereas in 9 (40%) of 22
A B
C D
Figure 9-83 Postpneumonectomy space: CT/MR correlations. A, B: Contrast-enhanced CT scans through the carina and lower lobes,
respectively, show typical appearance after a recent pneumonectomy in a patient with lung cancer. A large air-fluid level can be seen in A.
The pleural surfaces appear smooth, without evidence of focal masses or nodules. C, D: T1- and T2-weighted images, respectively, obtained
at the same level through the lower chest in the same patient as shown in A and B, several months later. The pneumonectomy space is
now entirely filled with fluid. The pleural surfaces are smooth without evidence of nodules or masses. Note that the pleural surfaces are
most easily evaluated on the T1-weighted scans (arrows in C).
A B
Figure 9-84 Postpneumonectomy space empyema: CT evaluation. A: Contrast-enhanced CT scans obtained through the midthorax in a
patient after right-sided pneumonectomy years before for treatment of bronchiectasis. An air-fluid level is present within the pneumonec-
tomy space, consistent with a fistula. Sections at the level of the carina and right mainstem bronchus show no evidence of airway pathology
(not shown). Increased soft tissue density is apparent in the region of the esophagus, which is difficult to identify (straight arrow). A sugges-
tion of a potential communication exists with the pneumonectomy space at this level (curved arrow). Note that the fluid level within the
pneumonectomy space has extremely high density (open arrow). B: Coned-down view from an esophragm showing fistulous communication
between the esophagus (arrow) and the pneumonectomy space.
A B
Figure 9-85 Postoperative assessment—tumor recurrence. A: Contrast-enhanced section at the level of the aortic arch in a patient after
left pneumonectomy for non–small cell lung cancer. Evidence is seen of an ill-defined soft tissue mass in the anterior mediastinum, consis-
tent with tumor recurrence, subsequently proved by thoracentesis. B: Contrast-enhanced CT in a different patient from that in A, after an
extrapleural pneumonectomy for mesothelioma. Markedly enlarged anterior mediastinal nodes are apparent consistent with residual/recur-
rent tumor. The appearance of the left hemithorax after an extrapleural pneumonectomy is quite similar to that seen in patients at post-
pneumonectomy, as is the appearance of recurrent tumor.
cases, CT showed the postpneumonectomy space to be pleura, these neoplasms represent only about 10% of
completely obliterated (224). Other findings identified primary pleural neoplasms (231). Unlike malignant meso-
were consequent to rotation and ipsilateral displacement theliomas, no known association exists between localized
of mediastinal and hilar structures, as well as hyperaera- pleural mesotheliomas and exposure to either asbestos or
tion of the contralateral lung. history of tobacco use. Usually presenting as incidental
To date, CT has proved most useful in the assessment of asymptomatic masses, these tumors have been associated
tumor recurrence (Fig. 9-85). This can be identified in with a number of extrathoracic abnormalities, including
most cases as mediastinal or hilar masses or, less com- pulmonary hypertrophic osteoarthropathy and digital
monly, as discrete peripheral masses projecting into the clubbing, which have been reported to occur in as many as
lower-density fluid within the pneumonectomy space 20% of cases. Along with nonspecific joint pains, these
(225,226). Other complications for which CT may be of findings have been noted to resolve after excision of these
use include the development of a postpneumonectomy tumors. Other less common symptoms include those
empyema secondary to bronchopleural fistulae (Fig. 9-84) referable to large masses in the chest, including chest pain
(227). CT also may also be valuable in diagnosing the and hemoptysis. An association with hypoglycemia has
so-called “right pneumonectomy syndrome” (228). In this also been noted (230).
condition, occurring sometimes years after surgery, symp- Pathologically, these tumors are composed of a mix-
toms of dyspnea and recurrent pulmonary infections ture of spindle-shaped fibroblast-like cells within a vari-
develop in the left lung owing to compression of the distal able amount of collagenous stroma. Interestingly, the cell
trachea or left mainstem bronchus, caused by the marked of origin of these tumors is not definitively known. In
shift of the mediastinum to the right side. 30% to 50% of cases, these tumors are pedunculated.
MR has also been used to evaluate the postpneumonec- Diagnosis usually requires surgical excision; transthoracic
tomy space (Fig. 9-83) (229). Not surprisingly, the main needle biopsy rarely provides sufficient tissue for defini-
advantage of MR is to delineate hilar and mediastinal structive diagnosis.
tures from recurrent tumor masses, especially in patients CT findings have been described (231–233). These
for whom contrast-enhanced CT is contraindicated. include well-defined, usually lobulated lesions ranging in
size from solitary nodules to massive neoplasms occupy-
ing most of the pleural cavity (Figs. 9-3 and 9-87). Pleural
Benign Fibrous Tumors of the Pleura
effusions are conspicuously absent, given the large size of
Primary neoplasms of the pleura are rare tumors that are these lesions. Calcifications within these tumors are also
generally divided into two broad subgroups: localized, rarely observed. Lee et al. (231), in a series of nine cases,
generally benign lesions (Fig. 9-86); and diffuse, invariably noted marked heterogeneous contrast enhancement in all
malignant lesions (230). Variously referred to as localized eight evaluated after the administration of IV contrast
pleural mesotheliomas or solid fibrous tumors of the (231). This they attributed to the presence of thin-walled
A B
Figure 9-86 Pleural hemangioendothelioma—CT evaluation. A, B: The 7-mm sections of contrast-enhanced images demonstrate a lobu-
lated relatively low-density but enhancing tumor encasing the right hemithorax and narrowing the central airways. The appearances are
nonspecific and although extremely large could be due to metastatic adenocarcinoma or mesothelioma. Biopsy in this case resulted in the
extremely rare diagnosis of pleural hemangioendothelioma, a tumor of vascular origin. Despite its supposed benign or intermediate grade
of neoplasia, in view of the high operative mortality of these vascular tumors, the prognosis is very poor. Additionally, histologic differentia-
tion from its frankly malignant counterpart, the pleural epithelioid angiosarcoma, is subjective and based on gradation of nuclear atypia,
which may be largely influenced by preoperative sampling in these heterogeneous tumors.
A C
B
Figure 9-87 Benign fibrous tumor of the pleura: spectrum of contrast-enhanced CT appearances in three different patients: 3- to 5-mm
sections. A: Fairly typical appearance of a solitary benign fibrous tumor with compression of the adjacent enhancing parenchyma (arrow).
These lesions are often large at presentation if symptomatic, as in this case, or detected incidentally if small. The lesion demonstrates typi-
cal heterogeneity of enhancement that reflects the lesion’s heterogeneous and variable composition of fibrous and connective stromal tis-
sues. As such, there is great variability of these lesions’ signal intensities on MR imaging. B: Further typical enhancement of a suspected be-
nign fibrous tumor of the pleura, thought to arise from a narrow pedicle off the supradiaphragmatic pleura. Although this lesion proved to
have the benign histologic features characteristic of a fibrous tumor, the lesion proved entirely intraparenchymal, a rare related variation
termed “intrapulmonary fibrous tumor.” C: Rarely benign fibrous tumors may exhibit atypia and are hence of indeterminate neoplastic po-
tential. In this case, at histology, spindle cell features were identified. Note no reliable distinguishing features to differentiate this lesion
from its entirely benign counterparts demonstrated in A and B. Effusions as seen in this case may be present in all forms.
venous sinusoids within tumors associated with thick- CHEST WALL

walled arteries and arterioles and dilated veins (231). In
distinction, areas of low attenuation corresponded to foci
Sternum
of myxoid or hemorrhagic degeneration.
Although usually benign, localized malignant mesothe- The sternum and sternoclavicular joints are difficult to
liomas do occur. Okike et al. (234), in a study of 60 patients evaluate with plain radiographs, mainly because overlying
with localized pleural mesotheliomas identified over a structures are difficult to exclude from view on frontal and
25-year period, 8 (13%) proved malignant (234). Similarly, oblique projections. The sternoclavicular joints are angled
Briselli et al. (235), in a review of more than 360 cases, obliquely, making their visualization particularly difficult.
found 12% to be malignant (235). Unfortunately, to date, Several reports have outlined in detail the normal CT
no clinical or radiographic features allowing preoperative appearance of the sternum and its articulations, including
differentiation between malignant and benign localized normal variants (Fig. 9-88) as well as pathologic changes
fibrous tumors of the pleura have been reported. Although (Figs. 9-89 to 9-95 and 9-98)(236–241).
resection is usually definitive, in approximately 15% of The value of CT in assessing lesions of the sternoclavic-
cases, local recurrence will be observed. ular joints and sternum has been documented (238,
B C
D E
Figure 9-88 Normal anatomy of the manubriosternal and sternoclavicular joints. A: Schematic drawing of the manubrium and sternoclav-
icular joints. A denotes a section through the distal clavicles above the level of the sternum. B, C: Coronal and parasagittal multiplanar
reconstructions through the sternoclavicular joints generated from 3-mm helically acquired source images reconstructed every 2 mm
provide good anatomic resolution of the manubrium (asterisks in B and C) as well as the distal clavicles (arrows in B and C). Coronal (D) and
sagittal (E) reconstructions of the bony thorax can be performed from contiguous 1-mm volumetric data sets acquired at 0.8-mm intervals
by using a 16-detector-row CT scanner. The normal anatomy of the manubriosternal junction can be evaluated in this patient with sickle cell
disease, demonstrating characteristic endplate depressions of the vertebral bodies. F: Oblique volume-rendered and surface-shaded pro-
jection using a similar data set in a different patient, demonstrating an alternative method for evaluating the normal anatomy of the
manubriosternal and sternoclavicular joints. This patient also demonstrates a large subscapular osteochondroma well highlighted by the
same image rendering. G: Axial image of same patient as in F, demonstrating the characteristic cartilage cap of an osteochondroma (arrow)
(continued).
F G
Figure 9-88 (continued )
A B
C D
Figure 9-89 Axial displaced sternal fracture. A–D: Serial axial 5-mm sections through the thorax on bone windows in a road accident vic-
tim demonstrate no fracture plane; however, the sternum is not visualized on one slice (C). Such abnormalities can be easy to overlook, par-
ticularly when reviewing many body parts on several window settings. Sagittal multiplanar image (E) may be very useful in evaluating the
sternum, as in this case, demonstrating an axial plane displaced fracture of the sternum. F: A 15-mm coronal maximum intensity projection
demonstrates the displacement of the fracture fragments and also the normal anatomic alignment of the manubriosternal joints.
E
242–250). In their series of 17 patients with sternoclavi- cases of nonspecific synovitis and osteoarthritis, and
cular abnormalities evaluated by CT, Destouet et al. (236) two cases of osteomyelitis.
demonstrated pathology better, or provided additional In our experience, CT has proven especially valuable (a)
diagnostic information in the majority of cases. These in assessing the sternum after sternotomy, both to rule out
included six cases of sternoclavicular joint dislocation, two postoperative infections and to evaluate the use of rotated
pectoral flaps after sternal debridement (Figs. 9-91 and
9-92) (238–240,242–244); (b) in assessing the sternum
and sternoclavicular joints, especially in intravenous drug
users, to rule out infection, including associated media-
stinitis (Fig. 9-93) (236,239,240,242,250,251); (c) in eval-
uating post-traumatic abnormalities (Fig. 9-94) (60); and
(d) in evaluating patients for whom a clinical suspicion
exists of a possible sternal lesion, especially in patients
with known breast cancer, as well as in evaluating sternal
pathology after irradiation (252).
Ribs
Although the ability of CT to assess rib pathology is
somewhat limited by the oblique orientation of ribs
relative to the CT scan plane, it is possible to identify
particular ribs accurately, provided there is detailed
knowledge of the CT appearances of their costovertebral
Figure 9-90 Renal osteodystrophy: 3-mm axial CT section and costotransverse articulations, as well as the relation
through the manubriosternal joints. Extensive osseous sclerosis is
present, consistent with renal osteodystrophy in this patient, who
of the clavicle to the first rib (253). This anatomy is illus-
also demonstrates a stent in the left brachiocephalic vein to treat a trated in Figures 9-95 through 9-97. In brief, the heads of
stricture associated with repeated central venous cannulations. the first, 10th, 11th, and 12th ribs articulate only with
The joint space of the manubriosternal joints is widened, consis-
tent with resorption. This feature of hyperparathyroidism may also
the body of the corresponding vertebra, whereas the
be appreciated in the chest at the acromioclavicular joints. heads of the second to ninth ribs articulate with the
A B
C
Figure 9-91 Poststernotomy infection: CT sinogram. A, B: Enlargements of sequential CT sections through the sternum in a patient after
sternotomy. This study was performed after injection of dilute contrast media into a catheter left in place after surgery. A space can be seen
between the two halves of the sternum filled with contrast; additionally, pockets of air and contrast can be identified anteriorly in the chest
wall (arrows in A and B) as well as posteriorly, adjacent to the sternum (curved arrow in A). Note that there is no evidence of a significant
fluid collection or contrast within the anterior mediastinum. C: Corresponding sinogram. Note that CT is far more precise in delineating the
exact number and locations of fluid collections.
articular facet on the lateral aspect of the body of the 1. Identify the first rib. The first rib is most easily identi-
corresponding vertebra, the intervening disc, and the fied on the axial image showing the middle third of the
demifacet on the vertebral body above. The first 10 ribs clavicle.
also articulate with their corresponding transverse 2. Identify the next two or three ribs on the same section by
processes. The intervening space between the neck of the counting posteriorly along the rib cage. Each more poste-
rib and the transverse process is called the costotransverse rior rib is numbered one higher than the rib anterior to it.
foramen. Although only a short segment of each rib is vi- 3. Proceed sequentially through the remaining images,
sualized on an axial CT image because of the caudal slope concentrating on the costovertebral articulations. Each
of the ribs, the head and neck of a rib are usually in the subsequent and numerically higher thoracic vertebra
same horizontal plane as the pedicle and the transverse and the corresponding rib are enumerated.
process of the corresponding vertebra. Knowledge of 4. Localize individual rib or pleural lesions by sequentially
these relations allows individual ribs to be identified enumerating the ribs along the rib cage, proceeding
in most cases. As outlined by Bhalla et al. (253), this anteriorly from the spine. Each successive rib encoun-
involves the following steps: tered represents the numerically preceding rib.
Text continues on page 843.
836
A B
Figure 9-92 Sternal dehiscence and chronic infection: CT evalua-

tion. A–C: The 5-mm enhanced CT sections after sternotomy for
ascending aortic aneurysm graft repair. Mediastinal windows demon-
strate soft tissue thickening around the sternotomy site but no dis-
crete fluid collection. Bone windows at the same level (B) and slightly
more inferior (C) demonstrate separation of the sternotomy with
frayed nonunion (B) and fragmentation (C). The appearances suggest
chronic dehiscence and infection. Evaluation of such abnormalities by
C CT is complemented by bone isotope and leukocyte scintigraphy.
Figure 9-93 Sternoclavicular osteomyelitis. CT section through the

sternoclavicular joints shows lytic destruction of the lateral aspect of
the sternum (arrow) as well as the distal clavicle, associated with con-
siderable soft tissue thickening in a known intravenous drug addict.
Needle aspiration was positive for staphylococcal infection.
Figure 9-94 Sternal fracture. A: 5-mm section after contrast in a trauma patient
demonstrates high-attenuation soft tissue density in the anterior mediastinum (arrow)
separated by normal fat from the heart and at other levels (not shown) from heart and
pericardium. The diagnostic possibilities for such a distribution of hematoma is due to
sheared small veins in the mediastinum or a sternal fracture. B: A 2-mm reconstruction
retrospectively performed through the sternum demonstrates a coronally oriented frac-
ture. Although such injuries are not of immediate hemodynamic concern, subtle sternal
fractures are often missed and are an indication of high-impact trauma that may be associ-
B ated with cardiac contusion and arrhythmogenic sequela.
Figure 9-95 Sternalis muscle: CT

findings. A–E: The sternalis muscle is
an unusual normal-variant muscle that
is often unilateral in appearance. Its
presence can cause the spurious
appearance of an irregular nodular
density in the medial breast on cranio-
caudal mammograms. On isolated
axial CT images, this structure may
appear as a paramedian soft tissue
nodular structure just anterior to
the sternum (arrows, A); however,
the appearance on serial images or
sagittal reconstructions can be dem-
onstrated to reflect an elongated
accessory muscle (arrows, B). The
10- to 30-mm maximum intensity pro-
jection oblique and coronal recon-
structions demonstrate the muscle to
be present on only the left side in this
case, characteristically lying anterior
to the medial margin of the pectoralis
A, B muscle (arrows, D, E).
A B
C, D E
Figure 9-96 Diagram of costovertebral articulations at T4 and T5, with a sagittal schematic representation of the corresponding axial
planes in A, B, and C. (From Leitman BS, Naidich DP, McCauley SI. Computed tomography of pectoral flaps. J Comput Assist Tomogr.
A B
Figure 9-97 Representative axial CT images through the planes

indicated in Figure 9-96. A: Through plane A, head (h), neck (n), and
tubercle (t) of the left fourth rib; transverse process (tp), and pedi-
cle (P) of the T4 vertebra: costotransverse foramen (arrowhead). B:
Through plane B, T4–T5 facet joints (arrowheads). C: Through plane
C, T4–T5 disk space (d), head of left fifth rib (h). Note partial
volume effect of corresponding pedicle and the transverse process.
(From Leitman BS, Naidich DP, McCauley SI. Computed tomogra-
phy of pectoral flaps. J Comput Assist Tomogr. 1988;12:392–393,
C with permission.)
B C
Figure 9-98 Ankylosing spondylitis: CT evaluation. A: 5-mm section through the lung apex in a patient with marked restrictive lung
physiology demonstrates mild parenchymal scarring that was initially attributed to remote postinflammatory sequelae. Review of bone
windows demonstrated multilevel osseous fusion of the costovertebral, costotransverse, and manubriosternal articulations, consistent with
ankylosing spondylitis (see selected bone-window enlargements, B and C). Ankylosing spondylitis can cause marked respiratory impairment,
not only by thoracic osseous restriction but also associated with marked upper lobe scarring that can mimic postgranulomatous inflammation;
however, parenchymal calcifications are absent.
Figure 9-99 Characterization by CT of chest radiograph pulmonary

“nodule”: pseudonodule due to prominent first costal cartilage calcifi-
cation. A: Posteroanterior chest radiograph demonstrates asymmetric
right apical nodular density that is suspected to reflect a pulmonary
nodule. B: Serial 2-mm reconstructions through the right lung apex
demonstrate that the suspected nodule is due to inferior invagination
into the lung parenchyma of asymmetric prominence of the first costo-
chondral junction. Thin sections are essential to avoid partial-volume
effects, which may cause a similar pseudonodule on thicker sections.
C: Target sagittal reconstruction confirms the absence of a pulmonary
nodule. Coronal or sagittal reconstructions may be particularly useful
with suspicion of a combination of pulmonary abnormality and osseous
A adjacent incidental changes.
C
Figure 9-100 Characterization by CT of chest radiograph pulmonary

“nodule”/pseudonodule due to bone island. A: Posteroanterior chest
radiograph demonstrates a suspected pulmonary nodule in the right
upper lung (arrow). B: 5-mm unenhanced section with lung windows (B)
demonstrates no pulmonary nodule at the appropriate level. C: Small
focus of sclerosis (arrow) is identified in the third anterior rib, accounting
for the pulmonary pseudonodule. Bone islands in ribs are typically well
A defined and elongated along the long axis of the rib (continued).
C Figure 9-100 (continued )
A B
Figure 9-101 Rib trauma evaluation by CT. A: Coronal 5-mm

multiplanar reconstruction (MPR) demonstrates a right-sided pleu-
ral collection surrounding atelectatic lung parenchyma. Right lat-
eral rib fractures are appreciated; a chest drain is in the subcuta-
neous tissues. B: Off-axis 5-mm MPR along the course of the chest
drain demonstrates that it is entirely within the soft tissues.
Anteroposterior radiographs had unsurprisingly suggested ade-
quate placement. C: Maximum intensity projection of the entire
chest volume, provided as an overview, shows that nearly all the
right-sided ribs are fractured both laterally and posteromedially,
resulting in a flail chest. When multiple rib fractures are present,
it is important to identify the number of contiguous ribs fractured
in two or more sites; if this is four or more, a radiologic flail seg-
ment is present, requiring clinical evaluation to exclude physio-
logic compromise. C
In a retrospective review of 12 cases with documented

rib pathology, Bhalla et al. (253) successfully localized 21
lesions, including 17 costal and 5 thoracic spine lesions,
by using this approach. It is obviously apparent that
knowledge of normal rib and costal cartilage anatomy is
essential for identifying pathologic changes involving
these structures (Figs. 9-99 through 9-101).
Axilla and Chest Wall

Detailed evaluation of the soft tissue structures of the
chest wall requires thorough familiarity with normal
cross-sectional anatomy. Anatomic relations in the region
of the axilla are especially important, given the signifi-
cance of this area for identifying chest-wall neoplasia, es-
pecially manifestations of breast cancer and lymphoma,
as well as identifying lesions that affect the brachial
plexus. Sequential cross-sectional anatomic illustrations
of the axilla are shown in detail in Figures 9-102 through
9-104. As described by Fishman et al. (254), the axilla is a
pyramid-shaped space between the upper arm and the
chest wall, with the apex directed superiorly and the base
directed downward. The boundaries of the axilla include Figure 9-103 The structures of the normal axilla, with corre-
(a) an anterior wall, formed by the pectoralis major and sponding characteristic transaxial segments labeled 1 to 8. Corac,
coracobrachialis; Pect major, pectoralis major; L dorsi, latissimus
minor muscles, subclavius muscles, clavipectoral fascia, dorsi; T major, teres major. (From Fishman EK, Zinreich ES, Jacobs
and the suspensory ligament of the axilla; (b) a posterior CG, et al. CT of the axilla: normal anatomy and pathology.
Radiographics. 1986;6:475–502, with permission.)
wall, formed by the subscapularis, latissimus dorsi, and

teres major muscles; (c) a medial wall, formed by the first
through the fifth ribs, their intercostal spaces, and the
serratus anterior muscle; (d) a lateral wall, formed by
the humerus and the coracobrachial and biceps brachii
muscles; (e) an apex, formed by the clavicle and upper
border of the scapula and outer border of the first rib; and
finally (f ) a base, formed by the anterior and posterior
axillary folds, the serratus anterior muscle, and the chest
wall. In the CT assessment of axillary structures, it must be
kept in mind that when patients are scanned with their
arms up, the axilla is open laterally.
Important landmarks within the axilla include the axil-
lary artery and vein and the cords of the brachial plexus, all
of which are enclosed in a connective tissue sheath, the
axillary sheath. Of these, only the artery and vein can be
identified as discrete structures on CT. The axillary vein
courses cephalad and lies successively on the anterior,
medial, and inferior sides of the axillary artery. With the
arm raised, characteristically the vein tends to lie anterior
to the artery throughout its course. In addition to these
structures, numerous lymph node chains are also present
within the axilla. Although variably described, these usu-
ally include (a) nodes located between the inferolateral
Figure 9-102 The boundaries of the axilla. Lat wall, lateral wall; margin of the pectoralis minor muscle and the latissimus
ant wall, anterior wall; P major, pectoralis major; med wall, medial dorsi muscle; (b) nodes located behind the pectoralis mus-
wall; L dorsi, latissimus dorsi; post wall, posterior wall; T major, cle in the axilla; and (c) nodes located between the supero-
teres major; Cora c, coracobrachialis. (From Fishman EK, Zinreich
ES, Jacobs CG, et al. CT of the axilla: normal anatomy and pathol- medial margin of the pectoralis muscle and the thoracic
ogy. Radiographics. 1986;6:475–502, with permission.) inlet (254,255). Additional nodes identifiable by CT
Text continues on page 851.
Figure 9-104 Labeled normal CT anatomy of the axilla and thoracic musculature. A–F: Select 2-mm axial noncontrast CT sections from
the cervicothoracic junction extending inferiorly. Arms are elevated per typical CT acquisition. G–L: 2-mm coronal reconstructions from the
same data set progressing from posterior to anterior. M, N: 2-mm sagittal reconstructions again from the same data set as A–L, progressing
from lateral (M) to more medial through the axilla (N).
844


H
J
L
N
A B
Figure 9-105 A: Recurrent breast cancer. Enlargement of a CT section through the right axilla in a patient after a right mastectomy.
A poorly marginated soft tissue mass can be identified within the axilla (arrows) associated with subcutaneous nodules. In addition, nodular
tumor implants can be associated with a large pleural effusion; mediastinal lymphadenopathy is present as well. B: Contrast-enhanced sec-
tion in a different patient from that in A shows extensive soft tissue infiltration of the anterior chest wall. Predominantly involving the left
breast, tumor can be seen to infiltrate the anterior chest wall muscles on the left, as well as extending into the subcutaneous tissues of the
right breast. Right-sided effusion proved to be caused by metastatic disease.
include interpectoral lymph nodes, normally seen only as recurrence, 16 (49%) patients had areas of disease identified
small dots in the interpectoral fat (256). by CT that were clinically unsuspected. In exceptional cases,
A wide variety of diseases, both benign and neoplastic, CT can provide specific tissue diagnoses. In patients with
may result in the development of axillary or chest-wall chest-wall lipomas, for example, CT may play an invaluable
adenopathy or masses (252,257–262). These generally are role by excluding more significant pathology, especially
nonspecific and frequently require biopsy (Figs. 9-105 in patients with previously documented malignancies
through 9-115). CT findings in patients after mastectomy (Fig. 9-6) (20,264–266). As illustrated in Figure 9-110, how-
and radiation therapy have been well described (252). ever, care must be taken to ensure that no solid or soft tissue
In selected cases, CT may prove of value by identifying elements are present. In the face of these or evidence of
otherwise unsuspected tumor recurrence. As shown by growth, a biopsy of these lesions should be performed.
Lindfors et al. (263), CT can detect foci of tumor recurrence CT can also be of specific value in identifying patients
unsuspected on physical examination. In their study of with chest-wall abscesses, especially those associated with
42 patients with local and/or regional recurrence of breast tuberculosis (Figs. 9-37 and 9-38) (259). CT also has
cancer, of 33 patients with clinical evidence of chest-wall Text continues on page 856.
A B
Figure 9-106 Axillary adenopathy: cutaneous T-cell lymphoma. A, B: 5-mm sections after intravenous contrast. Numerous subcutaneous
nodules are present, particularly well noted along the posterior chest wall. Additionally, bilateral axillary nodes are present. A subpectoral
node is highlighted (curved arrow). A larger lesion in the right chest wall bridges the anatomic compartments of the right axilla and the pre-
pectoral subcutaneous space, highlighting the difficulty in definitively determining the origin of larger lesions, despite the use of CT.
C
Figure 9-107 Focal soft tissue masses in the chest wall and axilla: CT appearances. A: CT scan at the level of the thoracic inlet shows
a well-defined soft tissue mass posteriorly (arrow). Benign desmoid, biopsy proven. B: Enhanced CT at the level of the aortic arch.
Asymmetry of the left posterior paravertebral muscles is present secondary to a well-defined mass bulging the intermuscular fat planes.
Biopsy-proven schwannoma. C: Contrast-enhanced CT through the lung bases demonstrates a soft tissue mass in the left chest wall (m).
Biopsy-proven metastatic non–small cell lung cancer. The similarity in appearances of these cases demonstrates that for many chest-wall
masses, although CT characterizes the anatomic location, it does not always obviate diagnostic biopsy.
A B
Figure 9-108 Diffuse soft tissue masses in the chest wall and
axilla: CT appearances in different patients. A: Polymyositis. CT scan
through the upper chest shows ill-defined soft tissue stranding in
both axillary regions, with additional linear calcification along the
posterior chest wall musculature (arrow). B: Neurofibromatosis.
Unenhanced CT images performed through the upper abdomen on
a chest examination. Multiple small cutaneous nodular densities are
present, as well as a thoracolumbar scoliosis. C: Mycosis fungoides
lymphoma. Contrast-enhanced CT through the lung apex demon-
strates multiple soft tissue nodules in the subcutaneous tissues of
the left chest wall; these appearances are relatively characteristic for
cutaneous T-cell lymphomas. Unlike in focal chest wall masses, CT
may often suggest a relatively specific diagnosis for diffuse or multi-
C focal soft tissue abnormalities.
A B
Figure 9-109 Superior vena cava (SVC) occlusion in a patient with chronic renal failure. Enhanced CT sections through the superior chest
(A, B) demonstrate multiple cutaneous enhancing collaterals. The SVC is occluded; high-density material within reflects pacing wires and
indwelling catheters causing the occlusion. A pacing device is also present in the chest wall. Depending on the length of occlusion, chest-wall
collaterals may, as in this case, communicate to the more caudal SVC via posterior intercostal anastomoses that supply the azygos system and
arch or also, as in this case, to the inferior vena cava via collaterals to the hepatic veins in the left lobe of the liver (C). D: A 15-mm slab coro-
nal maximum intensity projection (MIP) generated from a 1-mm contiguous data set through the entire volume demonstrates the posterior
intercostal vein anastomoses. E: MIP of the entire thoracic volume demonstrates the true extent of thoracic collaterals. When unenhanced
examinations are performed, limited collaterals visualized discontinuously as they pass in and out of the axial plane may mimic soft tissue
nodes or nodules (continued ).
C D
E Figure 9-109 (continued )
A B
Figure 9-110 Liposarcoma. A–D: Enlarge-

ments of sections through the right axilla and
chest wall show a fat-density mass causing
separation of the chest-wall muscles. Although
most of this mass appears composed of well-
differentiated fat, at least one soft tissue
nodule can be identified within this lesion
(arrow in D; compare with Fig. 9-6). Despite the
benign appearance of this lesion, growth led to
excisional biopsy, which documented well-
C D differentiated liposarcoma.
A B
Figure 9-111 Focal bone lesions of the thorax in two different patients. As with focal soft tissue lesions, the appearances are frequently
nonspecific. A: Chondrosarcoma of the right scapula. B: Aneurysmal bone cyst within a rib.
A B
C D
Figure 9-112 Multifocal osseous lesions of the thorax: CT findings. A, B: Multiple hereditary exostoses, 5-mm CT images through the
upper thorax. A cartilage-capped exostosis from a left anterior rib and an expansile lesion in a right anterior rib are relatively specific for the
diagnosis. The presence of anterior exostoses along the anterior aspect of the scapula can be associated with painful impingement against
the chest wall in the descriptively termed “snapping scapula syndrome.” C, D: Rheumatoid arthritis: 5-mm axial sections in a different pa-
tient from that in A and B demonstrating extensive fluid and soft tissue density around the glenohumeral joints, reflecting a combination of
joint effusion, bursitis, and pannus (C, mediastinal windows). Bone windows at the same level (D) demonstrate extensive erosion of the left
humeral head near the joint margin. E: Sclerotic metastases, prostate. Metastatic osseous disease may be sclerotic, lytic, or mixed and is
rarely specific for a single diagnosis. MRI is more sensitive at detecting the soft tissue disease that may be associated, particularly if cord
compression is suspected (continued ).
proved to be of diagnostic aid in detecting chest-wall infil- In distinction, CT is especially sensitive in detecting
tration in patients with actinomycosis (Fig. 9-113) chest-wall collaterals that may develop in patients with
(267–269). In selected cases, MR may also be of value to central tumors causing superior vena caval obstruction
delineate precisely the extent of chest-wall involvement (270). In general, these are easily distinguished from tran-
(Fig. 9-114). sient thoracic venous collaterals that may be seen normally
The role of CT for detecting neoplastic infiltration of in occasional patients after the bolus administration of IV
the chest wall, especially in patients with lung cancer, has contrast (271). It should be emphasized that as noted by
previously been discussed (Figs. 9-4 and 9-65). The value Engel et al. (272), identification of venous obstruction is
of CT in detecting other infiltrating lesions, including contingent on recognizing both decreased opacification of
mesothelioma (Figs. 9-68 and 9-70) and lymphoma, has central veins and enlarged collaterals: neither in itself is
already been described. In general, CT is effective only sufficient for diagnosis of venous obstruction, as collateral
in patients with relatively extensive tumor, as was empha- veins are frequently identifiable owing to flow dynamics in
sized previously; CT is considerably less accurate in ass- patients receiving rapid infusions of contrast media. As
essing patients with only minimal or subtle chest-wall shown by Gosselin and Rubin (273), in selected cases, CT
infiltration (Fig. 9-115). angiography may augment conventional CT evaluation of
A B
Figure 9-113 Actinomycosis. A: Posteroanterior (PA) radiograph shows a poorly defined, nonspecific area of increased density on the
right (arrow). B: CT scan through the sternum increased soft tissue density (arrows) involving the chest wall, corresponding to the ill-defined
density seen on the PA radiograph. Actinomycosis was biopsy proven. (Case courtesy of Robert Maissel, MD, Queens, New York.)
A B
Figure 9-114 Actinomycosis: MR evaluation. A: Posteroanterior chest radiograph shows evidence of asymmetric increased density in the
left apex. B: Coronal T2-weighted image shows diffuse increased signal intensity within the chest wall and lower neck. Changes caused by
documented actinomycosis. (Case courtesy of Sanford A. Rubin, MD, Galveston, Texas.)
Figure 9-115 Chest-wall invasion: lymphoma. A: Contrast-enhanced

section shows bulky necrotic anterior mediastinal mass with invasion of
both the anterior chest wall and sternum. B, C: MRI of a different patient
with lymphoma. Both T2-weighted images (B) and T1 fat-saturated
gadolinium-enhanced MR images demonstrate invasion of the right
parasternal chest wall by an anterior mediastinal lymphoma. Involvement
of the chest wall by lymphoma is usually obvious, but when subtle, MRI
may confer an advantage compared with CT staging alone secondary to
A superior contrast resolution.
B C
the thoracic spine, scapular injuries, and pulmonary lacera-

tions and contusions (Fig. 9-116) (278). CT frequently
detects small, incidental pneumothoraces after multisystem
trauma (277,279). CT may also be of value in assessing
patients with multisystem trauma to identify possible tho-
racic sources of infection, including both lung abscesses
and empyemas. CT also can play a role in evaluating late
sequelae of trauma, including the finding of post-traumatic
lung herniation (Figs. 9-117 and 9-118) (280).

Figure 9-116 Pulmonary contusion. Section through the upper of the Pleura and Chest Wall
lobes in a patient after a motor vehicle accident with extensive
subcutaneous emphysema on the right. An unusual-shaped air- Compared with other uses, relatively less attention has
space in the right lung extends to the pleural surface, consistent focused on the use of MRI to evaluate the pleura, chest
with pulmonary laceration caused by accompanying rib fractures wall, and diaphragm (13,168,232,281–288). Although
(not shown). Note the presence of bilateral effusions, larger on the
left side, without evidence of a pneumothorax. some correlations have been found in both in vitro and in
vivo between MR signal intensities and pleural fluid com-
position, MRI has not yet proven to be a reliable means
venous obstruction by allowing identification of abnormal for differentiating among various etiologies of pleural
temporal relations after contrast enhancement. effusions (289–292). Although MRI can differentiate
CT plays an especially important role in assessing between free and loculated effusions, as well as identify-
patients with trauma, including iatrogenic causes such as ing coexistent underlying lung disease, no compelling
in-dwelling chest-wall catheters, chest tubes, and pacemak- clinical role exists for the use of MRI in evaluating pleural
ers, as well as postsurgical changes (4,6,66,274–276). CT fluid collections.
may detect unsuspected small pneumothoraces (Fig. 9-8) The main advantage of MRI for evaluating the chest wall
or hemothoraces, especially after rib fractures or abdomi- is increased contrast resolution when compared with CT.
nal trauma (277), and also allows identification of stern- To date, this has proved most useful for evaluating neoplas-
oclavicular dislocation and sternal fractures (Figs. 9-94 and tic chest-wall invasion (Figs. 9-68, 9-114, and 9-115). As
9-95), including retrosternal hematomas, injuries involving previously discussed, CT is of only limited use in assessing
A B
Figure 9-117 Postoperative iatrogenic lung hernia. See Color Figure 9-117B. A: 5-mm section through the left lung apex demonstrates a
large anterior herniation after prior anterior thoracotomy and rib resection. B: A surface shaded volume-rendered projection formed from
contiguous 1-mm data volumetric data sets elegantly displays the anatomic detail of the hernia with respect to the chest wall. Such hernias
are a potential site of weakness for future pulmonary injury but when small are unlikely to cause significant respiratory compromise, despite
possible paradoxic respiratory motion.
B
A
C
Figure 9-118 Costochondral injury and pulmonary contusion:
evaluation by MRI of a 25-year-old with acute on chronic weight-
lifting injury to the right first costochondral junction. T1 fat-
suppressed images before (A, sagittal) and after intravenous
gadolinium administration (B, sagittal; C, coronal) demonstrate an
appearance resembling an enhancing soft tissue mass in the chest
wall surrounding the right parasternal region. Axial short inversion
time inversion-recovery (STIR) images demonstrate high signal inten-
sity in the soft tissues and a masslike area in the pulmonary
parenchyma (D), with more inferior thickening and fluid (E). The
appearances spontaneously resolved within 4 weeks and likely
reflected a combination of acute chest wall trauma with an inflam-
matory vascular response and an associated adjacent traumatic E
contusion.
parietal pleural and chest-wall invasion. By comparison, DIAPHRAGM

MRI has proven significantly more accurate in assessing
tumor infiltration, especially in patients with superior sul-
cus tumors (293). In a study of 31 patients evaluated with Anatomy
both thin-section CT scans and MR. Heelan et al. (177) The diaphragm is a musculotendinous layer that, in addi-
showed that MR was 94% accurate in detecting tumor tion to its function in respiration, serves to separate the
invasion into the chest wall as compared with 63% accu- thorax and abdomen. It consists of a fibrous central tendon,
racy for CT. In a study of 10 patients with superior sulcus divided, into middle, right, and left leaflets (Fig. 9-119), and
tumors evaluated with MR, McLoud et al. (158,178,294) muscle fibers, which inset into the central tendon, and arise
also showed MRI to be of value in detecting chest-wall and from all parts of the inner aspect of the body wall (297).
mediastinal invasion. MRI proved to be particularly sensi- Based on the origin of its muscle fibers, the diaphragm is
tive for detecting tumor infiltration into the soft tissues of considered to consist of an anterior or sternocostal part
the chest wall caused by marked differences in the signal and a posterior or lumbar part, which may be functionally
intensity of tumor, as compared with both fat and muscle, distinct (298).
on T2-weighted sequences in particular. Unfortunately, in The sternocostal or anterior part of the diaphragm
this same series, MR failed to identify subsequently con- arises from the inner aspect of the sternum and xiphoid
firmed rib destruction in five patients because of poor process and the seventh through 12th ribs and costal carti-
visualization of cortical bone (158,178,294). MRI has also lages, and inserts into the anterior portions of the middle,
been proven more sensitive than CT in detecting chest-wall right, and left leaflets (Fig. 9-120). The xiphoid process
involvement by malignant lymphoma (222,295). In one and costal cartilages form an inverted V- or U-shaped arch
study of 28 patients examined with both CT and MRI, CT when viewed from the front, and the anterior diaphragm
detected chest-wall invasion in 7 sites in 4 patients, MR assumes a similar shape (299–301).
showed chest-wall lesions in 14 sites in 7 patients (295). The posterior diaphragm is more properly referred to as
MRI has also been used to evaluate a variety of other the lumbar part of the diaphragm (302,303). It is made up
chest-wall masses, both benign and malignant, including of the right and left diaphragmatic crura, and fibers arising
lipomatous lesions and soft tissue hemangiomas, as well from the right and left medial and lateral arcuate liga-
as abnormalities involving the sternum and clavicles ments (304). The crura originate from the anterolateral
(260,261,287). Although MR has been shown to help surfaces of the bodies of the first three lumbar vertebrae on
identify chest-wall infections, MR is less accurate in the right and the first two lumbar vertebrae on the left. The
detecting osteomyelitis compared with CT (296). medial and lateral arcuate ligaments represent thickenings
Figure 9-119 Schematic diagram of the lumbar portion of the diaphragm, as viewed from below. This portion is composed of the crura,
which arise from the anterolateral surfaces of the first three lumbar vertebrae on the right and the first two lumbar vertebrae on the left,
respectively, and fibers that arise from the medial and lateral arcuate ligaments. These ligaments represent thickenings of the thoracolum-
bar fascia overlying the anterior surface of the psoas and quadratus muscles. Note that as drawn, no clear line of demarcation separates the
crura from the remainder of the posterior portions of the diaphragm.
A B
C D
Figure 9-120 Normal anatomy. A–D: Sequential 7-mm sections through the diaphragm, beginning at the level of the esophageal hiatus
and proceeding downward. The hemidiaphragms can be visualized as separate structures only when marginated centrally by peritoneal or
retroperitoneal fat (white arrows in A–D), or peripherally by air within the lungs, or by extraperitoneal fat. Visualization is lost when the
diaphragms abut a structure of similar density (for example, the liver or spleen) (straight black arrows in A and B). The position of the
diaphragm can still be inferred, however, from knowledge of characteristic anatomic relationships. At all levels, the lung and pleura lie adja-
cent and peripheral to the diaphragm; abdominal viscera, fat, and retroperitoneal structures lie adjacent and central to the hemidiaphragms.
Note posteriorly the appearance of a normal esophageal hiatus defined by the medial margins of the crura, bilaterally (arrowheads in B).
Figure 9-121 Retrocrural anatomy. CT depiction of serial 3-mm unenhanced images. The crura of the diaphragm are usually thin, smooth
structures on axial CT images, reflecting the spinal attachments of the diaphragm. Occasionally, particularly on the right side, these may
appear nodular and be confused with adenopathy. Review of serial images as depicted here should confirm the benign nature of this normal
variation of the diaphragmatic attachments. When nodular, the area may also appear metabolically active on PET, particularly if the patient
experiences anxiety-related hyperventilation before the study. Note that genuine lymph nodes in the retrocrural space are still above the
diaphragm and therefore within the thorax for staging purposes.
of the thoracolumbar fascia overlying the anterior surfaces

of the psoas and quadratus lumborum muscles (302,303).
The medial arcuate ligament arises from the lateral margin
of the L1 vertebral body and inserts on the transverse
process of L1. The lateral arcuate ligament arises from the
transverse process of the L1 vertebral body and inserts on
the 12th rib (Figs. 9-119 through 9-121). Muscle fibers aris-
ing from both the crura and arcuate ligaments arch for-
ward to insert into the posterior aspects of the middle,
right, and left leaflets.
Computed Tomography Appearances

Near its dome, the diaphragm lies in or near the plane of
scan and is impossible to identify as a specific structure
because of volume averaging (Fig. 9-120). Furthermore,
visualization of the normal diaphragm, measuring only a
Figure 9-122 Normal anatomy. CT section through the base of
few millimeters in thickness, is difficult when the dia- the lungs shows that the right hemidiaphragm can be visualized
phragm abuts structures of similar density, such as the liver almost along its entire course owing to marked fatty infiltration of
and spleen. The diaphragm can be visualized as a separate the liver (arrows).
structure only when it is imaged at an angle to its surface,
its outer aspect is marginated by air in the lungs or of a tumor (Figs. 9-123 and 9-124). Presumably, this appear-
extraperitoneal fat, and its inner aspect is marginated by ance is caused by infoldings or invaginations of contracted
intraperitoneal or retroperitoneal fat or when an alteration and foreshortened muscle bundles. Caskey et al. (308) have
appears in the density of adjacent viscera, as may occur studied age-related changes in the appearance of the
in patients with marked fatty infiltration of the liver hemidiaphragms. In a study of 120 individuals evaluated
(Fig. 9-122). by CT, these authors failed to note any appreciable change
A number of normal variants have been described in the thickness of the diaphragm with age, even when
(302,305–307). The significance of these lies primarily in the appearance was correlated with other indicators of
their not being confused with significant pathology. In physical condition, including skeletal muscle status, obesity,
some patients, and in some locations, the diaphragm may pulmonary emphysema, and the presence of an esophageal
assume a nodular configuration, mimicking the appearance hiatal hernia.
Figure 9-123 The costal diaphragm:

normal anatomy. A–D: Enlargements of
sequential 10-mm-thick sections through
A, B the costal diaphragm. This portion of the
diaphragm arises from the posterior sur-
faces of the lower six costal cartilages and
inserts into the anterolateral border of the
central tendon. Additionally, fibers attach
the middle leaflet to the xiphoid (arrows in
C and D). Note the close association
between the inferior portion of the peri-
cardium and the central tendon (curved
arrow in A). Most commonly, the costal
portion of the diaphragm appears as a
continuous, slightly undulating line that is
continuous across the midline with the lat-
eral diaphragmatic attachments (curved
arrow in B). This results when the middle
C, D leaflet lies superior to the xiphoid.
A B
Figure 9-124 Diaphragmatic pseudotumors. A, B: Ten-millimeter-thick CT sections through the costal and lumbar portions of the
diaphragm, respectively. Note the marked lobularity of the diaphragmatic surfaces both anteriorly (curved arrows in A and B) and postero-
laterally (arrows in B). This appearance is caused by infoldings of contracted and foreshortened muscle fibers resulting from a deep inspira-
tory effort. This phenomenon appears at least in part to be age related.
Variations in the degree of inspiration may have a pro- intrathoracic pressure, as opposed to intra-abdominal
found effect on the cross-sectional appearance of the pressure on the undersurface of the diaphragms, forcing
diaphragm (309). Specifically on scans obtained in deep them up into the thorax.
inspiration, nodularity of the diaphragm is accentuated The precise nature of these posterior defects has been
(305–307,309). These changes are reversible when expira- questioned. In an attempt to correlate the appearance of
tory scans are obtained. The diameter of the crura also the diaphragms with age, Caskey et al. (308) described
varies with lung volume, increasing at full inspiration. three types of diaphragmatic defects: type 1, in which
Recently it has been shown that after bilateral lung a localized defect can be identified in the thickness of
volume–reduction surgery, the total surface area of the the diaphragm without loss of diaphragmatic continuity;
diaphragm, as measured by CT, increased by approxi- type 2, in which an apparent defect can be identified
mately 17%, paralleling postsurgical improvement in where muscle fibers appear to separate into layers parallel
forced expiratory volume (FEV1) and decrease in func- to the diaphragmatic contour; and type 3, any defect in
tional residual capacity (FRC) (310). which a portion of the diaphragm appears absent, typi-
cally associated with protrusion of omental fat (308). As
shown by these authors, the appearance of diaphragmatic
Diaphragmatic Defects
defects appears related to both age and the presence of
The most common pathways that allow communication pulmonary emphysema. In a review of 120 scans, these
between the abdomen and thorax are the aortic and authors found that none of their patients in their 20s or
esophageal hiatuses. Another potential pathway for 30s demonstrated defects of any type, whereas 56% of
spread of disease between the abdomen and thorax patients in their 60s and 70s proved to have defects. A sig-
is focal defects in the hemidiaphragms. Posteriorly, nificant association between diaphragmatic defects and
the most common of these defects is caused by persist- pulmonary emphysema was also noted, especially in men.
ence of the embryonic pleuroperitoneal canal (so-called Based on these findings, these authors have suggested that
Bochdalek hernia) (Fig. 9-125). Although it has been many of the diaphragmatic defects identified, especially
reported that these defects occur in up to 90% of cases on posteriorly in older patients, represent acquired defects
the left side, it has been shown that right-sided defects are occurring in areas of structural weakness, perhaps them-
also common (Figs. 9-125 and 9-126). In an evaluation selves embryologic in origin.
of 940 patients studied with CT, Gale (311) identified Regardless of their origin, congenital diaphragmatic
60 Bochdalek hernias in 52 patients for a surprisingly defects are easily identified with CT (Figs. 9-125 and
high prevalence of 6% (311). As surprising, left-sided 9-126) (5,315–318). Commonly identified on plain chest
hernias proved only twice as common as right-sided her- radiographs as areas of eventration, these defects character-
nias. Herniation of abdominal structures through this istically are associated with protrusion of omental or even
defect may occur and may simulate intrathoracic masses retroperitoneal fat. Herniation of intraabdominal fat
(311–314). Herniation is presumably facilitated by lower and/or viscera may also occur in the anterior portions
A B
C D
Figure 9-125 Bochdalek hernia. Examples in different patients imaged by CT. A: Typical appearance of interruption of the posterior
hemidiaphragms with partial herniation of fat and the kidneys bilaterally. Although classically described on the left side, bilateral defects
usually containing fat alone are not infrequent. B, C: Different patient initially with a suspected nodule identified only on the lateral chest
radiograph (arrow, B) confirmed on CT as due to focal herniation of fat through a small Bochdalek hernia (C). D: Fat-containing Bochdalek
hernia (arrow) appears to “float” in a right-sided unrelated effusion. E, F: A 3-mm axial CT (E) demonstrates a small herniation of stomach
through a Bochdalek hernia lateral to the hiatus. Note the interrupted diaphragm. The anatomy of this herniation is best depicted on a
3-mm coronal multiplanar reconstruction (F). G, H: Coronal and sagittal volume-rendered CT images created from a contiguous 1-mm
volumetric data set demonstrate a large right-sided fat-containing Bochdalek hernia. The sagittal projection confirms failure of the
diaphragm to reach the posterior chest wall and herniation of fat and small associated vessels through the defect (arrow).
E F
G H
A B
Figure 9-126 Morgagni hernia. Coronal 5-mm maximum intensity projections reconstructed from a 1-mm volumetric data set. Coronal
(A) and right parasagittal (B) reconstructions demonstrate intra-abdominal peritoneal fat and small vessels herniating through a defect in
the anteromedial right diaphragm, consistent with a Morgagni hernia. Note that on the parasagittal reconstruction, the sternal slip of the
right hemidiaphragm does not reach the anterior chest wall, and herniation occurs through this region. Morgagni hernias generally become
problematic, requiring urgical intervention with bowel-content herniation.
(costal portions) of the diaphragms, resulting in a para- CT findings in patients with diaphragmatic rupture
cardiac mass (so-called Morgagni hernia) (Fig. 9-126) have been extensively reported (Figs. 9-127 to 9-130)
(315). CT is efficacious, first, in defining such masses as (4,274,275,319–324). CT can identify discontinuities in the
fatty, and second, in demonstrating continuity between course of the diaphragm, especially along the posterolateral
this fat and abdominal fat. The actual point of the di- aspect, an appearance that, when sufficiently severe,
aphragmatic defect may be definable by using parasagittal may aptly be described as the “absent diaphragm sign”
reconstructions. (Fig. 9-127 to 9-130). Superiorly, near the dome of the
diaphragm, CT identification of diaphragmatic discontinu-
ity is considerably more difficult. In these cases, recognition
Diaphragmatic Rupture
of diaphragmatic rupture usually necessitates identification
In addition to congenital abnormalities, defects in the of intrathoracic herniation of fat or abdominal contents. As
diaphragm may also be the result of both blunt and pene- noted by Shakelton et al. (324), identification in these cases
trating trauma. Diaphragmatic rupture usually occurs on is simplified when herniated organs manifest a “waist” or
the left side (in up to 95% of patients) and may go undiag- collar sign. Rarely, diaphragmatic rupture will be accompa-
nosed for years after both blunt and penetrating injuries nied by rupture of the adjacent pericardium, resulting in
(4,274,275,319–324). Chest radiographic diagnosis of herniation of abdominal contents into the pericardial sac
diaphragmatic rupture may be difficult, as this entity is (Fig. 9-130). Thickening of the diaphragm is another
frequently misdiagnosed as an elevated hemidiaphragm, recently described finding in patients with traumatic
left lower lobe atelectasis, left pleural effusion, or left sub- diaphragmatic rupture (322).
phrenic fluid collection and/or abscess. As documented Unfortunately, it is well established that CT is of only
by Lee et al. (325), in a study of 50 patients with acute limited use in the diagnosis of diaphragmatic rupture.
diaphragmatic rupture caused by blunt injury, nearly all As reported by Shapiro et al. (328), in an evaluation of
had evidence of other associated thoracic, abdominal, 12 patients with traumatic diaphragmatic rupture, only
and/or pelvic injuries. Unfortunately, despite the long time five (42%) CT studies were positive. Similar findings have
interval, chronic traumatic diaphragmatic rupture may been reported by Murray et al. (329), who demonstrated
eventuate in patients having symptoms of acute intestinal a sensitivity of only 61% and a specificity of 87% for the
obstruction secondary to infarction and/or strangulation CT diagnosis of diaphragmatic rupture. Reliance on axial
of herniated bowel or viscera (326,327). imaging is the major limitation of CT; in our experience,
A B
C D
Figure 9-127 Diaphragmatic rupture: CT evaluation with 5-mm unenhanced CT sections through the left chest and left hemidiaphragm.
Selected 1-mm section from a 1-mm contiguous set of images retrospectively reconstructed through the left diaphragm. A: Dependent viscera
sign is demonstrated: the small bowel contents in this case are in direct contact with the posterior ribs. B, C: The intra-abdominal contents an-
teriorly are herniated posterosuperiorly through a narrow defect in the diaphragm, resulting in the so-called “collar” or “waist” sign (arrows).
Note the normal position of the right hemidiaphragm crus and the absence of the left crus from its paravertebral location: the so-called
“absent hemidiaphragm sign.” D: Thin sections may be particularly useful to identify the separated diaphragmatic components (arrow) that
may be much more difficult to appreciate on thicker sections and also permit the performance of multiplanar reconstructions.
the use of multiplanar reconstructions around the dia- through the diaphragm at the level of the lateral arcuate
phragm has proven to have only a marginal value, likely ligament and upper portion of the quadratus muscle, the
caused by poor spatial resolution and the presence of lungs and pleura always lie peripheral to the diaphragm.
frequent artifacts that themselves may mimic diaphrag- The lung is invested with visceral pleura. Just below
matic disruption. Some have suggested that better results the most inferior portion of the lung there is a space—
would be obtained by using MR (330). In selected cases, generally a potential space—referred to as the posterior
however, to date, this application has not received wide- pleural recess. This space is defined anteriorly and poste-
spread clinical acceptance. riorly by layers of parietal pleura. Fibers of the lumbar
portion of the diaphragm extend below the posterior
pleural recess, ending at the lateral arcuate ligament at
Peridiaphragmatic Fluid Collections
the level of the 12th rib.
Accurate localization of peridiaphragmatic fluid requires Identification of the posterior pleural recesses is critical
detailed knowledge of normal cross-sectional anatomy when evaluating peridiaphragmatic fluid collections, as
through the lung bases and upper abdomen. As shown in this space represents the most inferior, dependent recess of
Figure 9-131, a schematic drawing of a sagittal section the pleural space: it is this space that will be filled first by
A B
C D
Figure 9-128 Diaphragmatic rupture: CT evaluation. Posteroanterior and lateral radiographs (A and B) suggested a possible partial
eventration of the right diaphragm in a relatively typical anteromedial location. C, D: 5-mm enhanced CT sections, acquired through the liver
at a later date for unrelated abdominal suspected pathology, demonstrated subtle periportal edema limited to the anteromedial hepatic
parenchyma (arrows in C and D). Surgical exploration confirmed the presence of partial disruption of the right hemidiaphragm and hernia-
tion of part of the liver through a diaphragmatic defect, causing mild obstructive edema. A remote forgotten traumatic event (see rib frac-
ture; E) was the suspected etiology. The case demonstrates that on the right side in the absence of bowel herniation, direct visualization of
diaphragmatic discontinuity adjacent to the liver may be problematic, even with CT.
E Figure 9-128 (continued )
A B
Figure 9-129 Right-sided traumatic diaphragmatic rupture. A–C:

The 10-mm-thick CT sections through the right upper quadrant from
above downward show characteristic appearance of herniation of
bowel and fat into the right hemithorax (arrow in A) secondary to
traumatic rupture of the right hemidiaphragm. The right hemidi-
aphragm itself can be identified in B (arrows), appearing somewhat
thickened and displaced medially. Diaphragmatic discontinuity can
C be identified in C (arrows).
869
A B
Figure 9-130 Diaphragmatic and pericardial rupture. A: Contrast-enhanced section shows apparent air collection in front of the right
ventricle (asterisk). Note the presence of moderate right effusion as well. B: Sagittal reconstruction shows that the air in front of the ventri-
cle represents bowel that has herniated through both the diaphragm and pericardium (arrow).
free pleural fluid (331). Although the posterior pleural 2. The “displaced crura sign” results from the interposition
recess is usually a potential space, on occasion it can be of pleural fluid between the crus and adjacent vertebra,
identified with CT in normal individuals. with resultant anterior displacement of the crus.
As illustrated in Figure 9-132, peridiaphragmatic fluid 3. The “interface sign,” first described by Teplick et al.
can be localized to one of four potential spaces: the pleural (333), is based on the nature of the interface between
cavity, the lung, the peritoneum, or the retroperitoneum. the liver and adjacent fluid. This interface is hazy in the
Four main criteria have been proposed to differentiate case of pleural fluid, but distinct in the case of ascites.
pleural from peritoneal fluid (331–334). 4. The “bare area sign” is based on recognition that identi-
fication of the bare area of the liver indicates fluid
1. The “diaphragm sign.” As already discussed, direct visu-
within the peritoneum (Fig. 9-134).
alization of the diaphragm itself allows accurate locali-
zation of peridiaphragmatic fluid collections, as fluid Additional CT findings have been described as potential
peripheral to the diaphragm is intrathoracic, whereas pitfalls in differentiating pleural from peritoneal fluid. A
fluid central to the diaphragm is intra-abdominal combination of atelectasis and subpulmonic pleural fluid
(Fig. 9-133). in particular may be misleading because subsegmental
atelectasis may result in a curvilinear band at the lung base
that superficially may simulate the hemidiaphragm (335).
In these cases, variability in the appearance of pleural fluid
at the lung base probably reflects differences in the
Figure 9-132 Schematic drawing of the potential spaces in which

peridiaphragmatic fluid may collect. Accurate identification of the
Figure 9-131 Schematic diagram of a sagittal section through diaphragm is critical. Fluid within the pleural spaces or lung lies
the lateral arcuate ligament. The lung is invested with visceral peripheral to the hemidiaphragms; intraperitoneal or retroperitoneal
pleura. Below the inferior edge of the lung is a potential space, the fluid lies central to the hemidiaphragms. On the right side, intraperi-
posterior pleural recess, which is defined anteriorly and posteriorly toneal fluid is restricted medially at the level of the right coronary
by layers of parietal pleura. The diaphragm always lies central to ligament. Ao, aorta; E, esophagus; IVC, inferior vena cava; LLL, left
the lung and pleura when seen in cross section. lower lobe; RLL, right lower lobe.
Underneath the diaphragm, both peritoneal and retro-

peritoneal fluid may abut the diaphragm (336,337).
Retroperitoneal fluid will conform to the anatomic space
and/or structures involved (e.g., posterior pararenal versus
perirenal spaces). The psoas muscles are also retroperitoneal,
and, as discussed earlier, are intimately related to the arcuate
ligaments. Generally, differentiation between retroperitoneal
and intraperitoneal fluid is not difficult. This is especially
true in cases in which retroperitoneal fluid accumulates sec-
ondary to a leaking abdominal aortic aneurysm.
In distinction, intraperitoneal fluid will collect in the
peritoneal spaces and will therefore be restricted by
the peritoneal reflections. These are best visualized after
intraperitoneal injection of dilute contrast media (Fig.
9-135). On the right side, peritoneal fluid is restricted
Figure 9-133 Peridiaphragmatic fluid localization. Contrast- posteromedially by the right coronary ligament. This
enhanced section through the dome of the liver (asterisk) shows results in a characteristic medial tapering of intraperitoneal
evidence of both ascites and right pleural effusion. In this case, fluid (the “bare area sign”) (Figs. 9-134 and 9-135). It
ascites clearly lies central to the diaphragm (arrows), whereas the
effusion and associated linear band of atelectatic lung lie periph- should be remembered, however, that superiorly, ascites,
eral and posterior to the diaphragm (curved arrow). when massive, may collect under the domes of the dia-
phragm on both the left and right sides (i.e., above the level
of the coronary ligaments), thereby limiting considerably
anatomy of the pulmonary ligaments. If incomplete the value of the bare area sign. On the left side, intra-
attachments of these ligaments exist, the lower lobes are abdominal (intraperitoneal) fluid is easy to identify
free to float on pleural fluid, provided the lung is not stiff- because it will characteristically be central to the left
ened by disease and no adhesions are present between the hemidiaphragm and, additionally, will surround the
lung and the adjacent pleura. In distinction, when these spleen. As noted previously, identification of intraperi-
attachments are complete, the base of the lung may toneal fluid around the spleen is facilitated by recognition
become tethered inferiorly, despite the presence of even of a bare area, analogous to that of the liver (338).
sizable pleural effusions; the result may be the appearance Despite clarification of normal cross-sectional anatomy
of a “pseudodiaphragm sign.” of the pleural, peritoneal, and retroperitoneal spaces,
Potential difficulty may also be encountered in identify- certain cases remain difficult to evaluate. This is especially
ing basilar lung consolidation and/or atelectasis with or true when pleural fluid is sufficiently massive to cause
without associated pleural fluid (331). Superficially, areas inversion of the hemidiaphragms, because pleural fluid
of dense parenchymal consolidation may mimic an effu- may then simulate the appearance of intraabdominal fluid
sion, especially when small. (339,340). In these cases and others, multiplanar recon-
structions may prove helpful by further delineating
anatomic relations (Fig. 9-136).
Diaphragm as Pathway for the

Spread of Disease
The diaphragm usually serves to localize disease in either
the thorax or abdomen. However, extension through the
diaphragm can occur via several pathways. Direct contigu-
ous spread most commonly occurs through the pre-
existing normal channels of communication, that is, the
aortic and esophageal hiatuses.
Sliding hiatal hernias typically are manifest as widening
of the esophageal hiatus on cross section (see Chapter 4,
Figs. 4-149 and 4-150). Measurements of the standard width
of the esophageal hiatus, defined as the distance between
Figure 9-134 Peridiaphragmatic fluid localization. Contrast- the medial margins of the crura, have been reported (308,
enhanced section through the liver in a patient with ascites shows 341). The width of the esophageal hiatus clearly enlarges
fluid marginating the right lobe of the liver, clearly lying central
to the diaphragm (arrows). Note that the fluid does not extend to with age, especially in men (308). Measuring an average
cover the bare area of the liver. 1.25 cm in 20-year-old individuals, the esophageal hiatus
A B
C D
Figure 9-135 Peridiaphragmatic fluid localization: intraperi-

toneal contrast injection. A–E: Sequential images through the
lower thorax and upper abdomen in a patient with known
ovarian carcinoma evaluated after the intraperitoneal injection
of dilute contrast media allow precise visualization of all the
potential areas in which intraperitoneal fluid may accumulate.
Note lack of contrast in the region of the bare area of the liver
E (arrow in B).
progressively widens each decade, measuring more than the contrast-filled stomach may occasionally be identified
3 cm on average in individuals in their 70s. Hiatal as well (342).
hernias typically offer little difficulty in diagnosis. They are In addition to hiatal hernias, the esophageal hiatus serves
frequently associated with an apparent increase in mediasti- as a passage for the spread of malignancy, especially that
nal fat surrounding the distal esophagus, secondary to arising in the esophagus and stomach (see Fig. 4-136).
herniation of omentum through the phrenicoesophageal Esophageal varices also may traverse the esophageal hiatus
ligament. Fluid within herniated peritoneum anterior to (343); in these cases, CT is especially valuable in detecting
A B
C D
Figure 9-136 Peridiaphragmatic fluid collections: multiplanar reconstructions. A, B: Sequential contrast-enhanced sections through
upper abdomen show evidence of extensive ascites surrounding both the liver and spleen. C, D: Coronal and sagittal reconstructions help
to find further the precise location of intraperitoneal fluid (asterisks in C and D) owing to the sharp margin of the diaphragm.
paraesophageal varices (see Figs. 4-147 and 4-148). Fre- disease, especially mesotheliomas (Fig. 9-70). In this set-
quently these are first seen as nonspecific mediastinal soft ting, the diaphragm may become involved by tumor; the
tissue masses, necessitating differentiation from enlarged result is that tumor tracks along the diaphragm to gain
paraesophageal lymph nodes or other posterior mediastinal entry into the abdomen. A similar appearance has been
masses. Despite the accuracy of endoscopy and esophagog- described in patients with invasive thymomas and lym-
raphy to detect esophageal varices, paraesophageal varices phomas and may be expected to be seen in any tumor that
have previously required angiography for definite diagnosis. originates in viscera adjacent to the diaphragms (Fig. 9-78
More rarely, the esophageal hiatus serves as the pathway and 9-138) (214,346). Unfortunately, as previously dis-
for the spread of pancreatic fluid collections (Fig. 9-137). cussed, both CT and MRI are of limited value in identifying
The key to evaluating these fluid collections is analysis extension of tumor to the diaphragm in patients with
of contiguous sections from the lung bases to the upper malignant pleural mesothelioma. In one study evaluating
abdomen. It should be noted that fluid in the lesser sac the efficacy of CT with laparoscopy, CT correctly identified
might appear to extend past the crural margins, occasion- only 3 of 10 patients with documented transdiaphragmatic
ally even mimicking the appearance of thrombosis of the extension of tumor (347). Similar limitations have been
inferior vena cava (344). described for MRI and FDG-PET (167,347).
Although disease usually spreads between the abdomen Analogous to the spread of tumor, the diaphragm may
and thorax by means of a diaphragmatic defect (either con- also serve as a means of extension of infection. In our
genital or acquired), the diaphragm itself, on occasion, experience, this is most frequently secondary to tuberculo-
may actually serve as a pathway for disease, both inflamma- sis, although a similar appearance may be seen with other
tory and malignant (214,345,346). This is most typically infections, including actinomycosis. Infections may also
encountered in patients with diffuse malignant pleural occur after abdominal surgery (Fig. 139).
A B
C D
Figure 9-137 Mediastinal pseudocyst. In this case, the key to the diagnosis is evaluation of sequential images through the thoracoab-
dominal junction. A: Section shows a large fluid collection in the posterior mediastinum, adjacent to the esophagus. B: Section just above
the esophageal hiatus. The fluid collection seen in A is still present, although smaller, lying anterior to the crura and lateral to the distal
esophagus (arrow). C: Section at the level of the pancreas. A small fluid collection can be identified in the region of the head of the pancreas
(arrow). In this case, fluid has tracked from the retroperitoneum, through the esophageal hiatus, to localize in the posterior mediastinum.
Communication of these fluid collections could be established by reviewing sequential images (not all are shown). D: Oblique view from an
endoscopic retrograde cholangiopancreatogram confirming the diagnosis.
Finally, tumors may arise within the diaphragm (348– isolated reports have shown the efficacy of MR in evaluating
351). This is exceptionally rare and frequently presents diag- thoracic and abdominal wall infections, diaphragmatic rup-
nostic dilemmas. Such cases invariably require surgery for ture, Morgagni hernias, mediastinal pseudocysts, and even
definitive diagnosis and treatment. diaphragmatic endometriosis, no large series has ever been
reported comparing MR with CT in the evaluation of peridi-
aphragmatic fluid collections (Figs. 9-140 through 9-143)
(296,330,349,354). The utility of MR for routine evaluation
of the Diaphragm
of diaphragmatic abnormalities is especially limited in the
To date, the role of MR in evaluating disease in and around era of MDCT with high-quality multiplanar reconstructions
the diaphragm has been limited (352,353). Although easily acquired.
A B
Figure 9-138 Metastatic renal cell carcinoma. A–C: Sequential

10-mm-thick CT sections through the left upper quadrant from
above downward show a large renal cell carcinoma arising in the
left kidney (arrow in C), clearly extending to and invading the left
C hemidiaphragm (arrows in A and B).
A B
Figure 9-139 Diaphragm as pathway for spread of infection. A, B: Sequential contrast-enhanced sections through the upper abdomen in
a patient after recent cholecystectomy with spiking fevers. An ill-defined fluid collection appears in the right upper quadrant (asterisk in B)
that extends superiorly to the diaphragm, which is thickened (arrow in A). A focal area of consolidation is also apparent in the right lower
lobe. CT-guided transabdominal aspiration confirmed the presence of pus.
A B
Figure 9-140 Normal diaphragm: MR evaluation. A, B: Sequential T1-weighted axial spin-echo scans through the upper abdomen show
typical appearance of the diaphragms, recognizable as a distinctly low-signal-intensity line, marginated centrally and peripherally by high-
signal-intensity fat (arrows in A and B). Note that, because of increased contrast resolution, the line of the diaphragm can also be
appreciated, marginating the lateral aspect of the liver (curved arrows in A and B; compare with Fig. 9-120).
Figure 9-142 Diaphragmatic tumor invasion: MR evaluation.

Coronal T1-weighted scan through the liver shows a large, necrotic
tumor mass in the liver extending superiorly to involve the
Figure 9-141 Peridiaphragmatic fluid localization/ascites: eval- diaphragm (arrow). Note the presence of basilar consolidation, as
uation with MR. Coronal T1-weighted MR scan clearly shows the well, in the right lower lobe (curved arrow). Given the inherent
presence of ascites surrounding both the liver and spleen (arrows), advantages of multiplanar imaging and superior contrast resolu-
recognizable as a low-signal-intensity fluid collection. tion, MR should be of benefit in difficult cases for defining the true
extent of peridiaphragmatic tumor.
C
E
F
Figure 9-143 Chest-wall tumor: hepatocellular carcinoma recurrence and suspected diaphragm invasion from biopsy-tract tumor
seeding—CT and MR evaluation. A: Enlargement from a posteroanterior radiograph demonstrates a small effusion with rib destruction
(arrows) and pulmonary atelectasis. B: A 5-mm unenhanced CT section through the lung base demonstrates the chest-wall mass destroying
the lateral ribs (short arrow), an associated right effusion, and suggests possible diaphragmatic invasion, although this cannot be directly
visualized. Patient has undergone prior liver transplantation (long arrow demonstrates inferior vena cava anastomotic sutures plane). C:
Coronal half-Fourier single shot turbo spin-echo (HASTE) T2-weighted image highlights the inferior extrinsic indentation of the hepatic
parenchyma without diaphragmatic invasion. T1 fat-suppressed sequence immediately after (D) and delayed (E) after contrast administra-
tion shows initial enhancement of compressed normal atelectatic parenchyma lateral to the mass. The mass lesion itself enhances later. (F)
Short inversion time inversion-recovery (STIR) sequence is useful for demonstrating pleural and peritoneal fluid present in this case.
REFERENCES 32. Raasch BN, et al. Radiographic anatomy of the interlobar

fissures: a study of 100 specimens. AJR Am J Roentgenol. 1982;
138:1043.
1. Pugatch RD, et al. Differentiation of pleural and pulmonary 33. Yamashita H. Roentgenologic anatomy of the lung. Stuttgart: Thieme
lesions using computed tomography. J Comput Assist Tomogr. Medical Publishers; 1978.
1978;2:601–606. 34. Mayo JR, Müller NL, Henkelman RM. The double-fissure sign: a
2. Kreel L. Computed tomography of the lungs and pleura. Semin motion artifact on thin-section CT scans. Radiology. 1987;165:
Roentgenol. 1978;13:213–225. 580–581.
3. Kwan SW, et al. Primary interpretation of thoracic MDCT 35. Berkman YM, et al. Anatomy of the minor fissure: evaluation
images using coronal reformations. AJR Am J Roentgenol. 2005; with thin-section CT. Radiology. 1989;170:647–651.
185:1500–1508. 36. Otsuji H, et al. Right upper lobe versus right middle lobe: differ-
4. Mirvis SE. Imaging of acute thoracic injury: the advent of MDCT entiation with thin-section high-resolution CT. Radiology.
screening. Semin Ultrasound CT MRI. 2005;26:305–331. 1989;172:653–656.
5. Ravenel JG, McAdams HP. Multiplanar and three-dimensional 37. Ahn MI, et al. Abstract: perifissural opacities (PFOs) detected at
imaging of the thorax. Radiol Clin North Am. 2003;41:475. chest CT screening for lung cancer. Chicago: Radiologic Society
6. Rivas LA, et al. Multislice CT in thoracic trauma. Radiol Clin of North America; 2004.
North Am. 2003;41:599. 38. Takasugi JE, Godwin JD. Left azygos lobe. Radiology. 1989;171:
7. Im JG, et al. Costal pleura: appearances at high-resolution CT. 133–134.
Radiology. 1989;171:125–131. 39. Speckman JM, Gamsu G, Webb WR. Alterations in CT mediasti-
8. Remy Jardin M, et al. Asbestos-related pleuropulmonary dis- nal anatomy produced by an azygos lobe. AJR Am J Roentgenol.
eases: evaluation with low-dose four-detector row spiral CT. 1981;137:47–50.
Radiology. 2004;233:182–190. 40. Godwin JD, Tarver RD. Accessory fissures of the lung. AJR Am J
9. Bressler EL, et al. Bolus contrast medium enhancement for dis- Roentgenol. 1985;144:39–47.
tinguishing pleural from parenchymal lung disease: CT features. 41. Austin JHM. The left minor fissure. Radiology. 1986;161: 433–
J Comput Assist Tomogr. 1987;11:436–440. 436.
10. Medford A, Maskell N. Pleural effusion. Postgrad Med J. 2005; 42. Boyden EA. The distribution of bronchi in gross anomalies of
81:702–710. the right upper lobe, particularly lobes subdivided by the azygos
11. Arenas JJ, et al. Evaluation of CT findings for diagnosis of pleural vein and those containing pre-eparterial bronchi. Radiology.
effusions. Eur Radiol. 2000;10:681–690. 1952;58:797.
12. Traill Z, Davies RJ, Gleeson FV. Thoracic computed tomography 43. Rabinowitz JG, Cohen BA, Mendelson DS. The pulmonary
in patients with suspected malignant pleural effusions. Clin ligament. Radiol Clin North Am. 1984;22:659–672.
Radiol. 2001;56:193–196. 44. Rost RC Jr, Proto AV. Inferior pulmonary ligament: computed
13. Evans AL, Gleeson FV. Radiology in pleural disease: state of the tomographic appearance. Radiology. 1983;148:479–483.
art. Respirology. 2004;9:300–312. 45. Cooper C, et al. CT appearance of the normal inferior pulmonary
14. Berne AS, Heitzman ER. The roentgenographic signs of pedun- ligament. AJR Am J Roentgenol. 1983;141:237–240.
culated pleural tumors. AJR Am J Roentgenol. 1962;87:892–895. 46. Taylor GA, et al. CT demonstration of the phrenic nerve. J
15. Lewis MI, et al. The case of the moving intrathoracic mass. Comput Assist Tomogr. 1983;7: 411–414.
Chest. 1985;88:897–898. 47. Berkmen YM, et al. Right phrenic nerve: anatomy, CT appearance,
16. Williford ME, et al. Computed tomography of pleural disease. and differentiation from the pulmonary ligament. Radiology.
AJR Am J Roentgenol. 1983;140:909–914. 1989;173:43–46.
17. Zinn WL, et al. Fluid within preexisting pulmonary air-spaces: 48. Godwin JD, Vock P, Osborne DR. CT of the pulmonary liga-
a potential pitfall in the CT differentiation of pleural from ment. AJR Am J Roentgenol. 1983;141:231–236.
parenchymal disease. J Comput Assist Tomogr. 1987;11:441–448. 49. Bernaudin J-F, Fleury J. Anatomy of the blood and lymphatic
18. Griffin DJ, et al. Observations on CT differentiation of pleural circulation of the pleural serosa. In: Chretien J, Bignon J, Hirsch
and peritoneal fluid. J Comput Assist Tomogr. 1984;8:24–28. A, eds. The pleura in health and disease. New York: Marcel Dekker;
19. Federle MP, Mark AS, Guillaumin ES. CT of subpulmonic pleu- 1985:101–124.
ral effusions and atelectasis: criteria for differentiation from 50. Policard A, Galy R. La Plevre. Paris: Masson; 1942:23–33.
subphrenic fluid. AJR Am J Roentgenol. 1986;146:685–689. 51. Agostoni E, Miserocchi G, Bonanni MV. Thickness and pressure
20. Epler GR, et al. Pleural lipoma: diagnosis by computed tomog- of the pleural liquid in some mammals. Respir Physiol. 1969;
raphy. Chest. 1986;90:265–268. 6:245–256.
21. Levi C, et al. The unreliability of CT numbers as absolute values. 52. Staub NC, Wiener-Kronish JP, Albertine KH. Pleural liquid and
AJR Am J Roentgenol. 1982;139:443–447. solute exchange. In: Chretien J, Bignon J, Hirsch A, eds. The
22. Vock P, et al. Analysis of the density of pleural fluid analogs by pleura in health and disease. New York: Marcel Dekker; 1985:170.
computed tomography. Invest Radiol. 1984;19:10–15. 53. Sargent EN, et al. Subpleural fat pads in patients exposed to
23. Rawkin RN, Raval B, Finley R. Case report: primary chyloperi- asbestos: distinction from non-calcified pleural plaques.
toneum: combined lymphangiographic and CT diagnosis. Radiology. 1984;152:273–277.
J Comput Assist Tomogr. 1980;4:869–870. 54. Vix VA. Extrapleural costal fat. Radiology. 1974;112:563–565.
24. Nandalur KR, et al. Accuracy of computed tomography attenua- 55. Light RW. Pleural effusion. N Engl J Med. 2002;346:1971–1977.
tion values in the characterization of pleural fluid: an ROC 56. Light RW, Macgrefor I, Luchsinger PC, et al. Pleural effusions:
study. Acad Radiol. 2005;12:987–991. the diagnostic separation of transudates from exudates. Ann
25. Im J-G, Chung JW, Han MC. Milk of calcium pleural collections: Intern Med. 1972;77:507–513.
CT findings. J Comput Assist Tomogr. 1993;17:613–616. 57. Bartter T, Akers SM, Pratter MR. The evaluation of pleural effu-
26. Frija J, Yana C, Laval-Jeantet M. Anatomy of the minor fissure: sion. Chest. 1994;106:1209–1214.
evaluation with thin-section CT. Radiology. 1989;173:571–572. 58. Sahn SA. The pleura. Am Rev Respir Dis. 1988;138:184–234.
27. Marks BW, Kuhns IR. Identification of the pleural fissures with 59. Andrews NC, et al. American Thoracic Society. Management of
computed tomography. Radiology. 1982;143:139–141. non-tuberculous empyema: a statement of the Subcommittee
28. Goodman LR, et al. The right mid-lung window. Radiology. 1982; on Surgery. Am Rev Respir Dis. 1962;85:935–936.
143:135–138. 60. Choi JA, et al. CT manifestations of late sequelae in patients
29. Genereux GP. The posterior pleural reflections. AJR Am J with tuberculous pleuritis. AJR Am J Roentgenol. 2001; 176:
Roentgenol. 1983;141:141–149. 441–445.
30. Proto AV, Ball JB. Computed tomography of the major and 61. Kim DJ, et al. Chronic tuberculous empyema: relationships
minor fissures. AJR Am J Roentgenol. 1983;140:439–448. between preoperative CT findings an postoperative improve-
31. Proto AV, Ball JB. The superolateral major fissures. AJR Am J ment measured by pulmonary function testing. Clin Radiol.
Roentgenol. 1983;140:431–437. 2005;60:503–507.
62. Schmitt WGH, Hubener KH, Rucker HC. Pleural calcification 92. Mullin DM., et al. Computed tomography of oleothorax.
with persistent effusion. Radiology. 1983;149:633–638. Comput Tomogr 1987:10:197–199.
63. Bhatt GM. Case report: CT demonstration of empyema necessi- 93. Orringer MB. Thoracic empyema; back to basics. Chest. 1988;
tatis. J Comput Assist Tomogr. 1985;1985:1108–1109. 93:901–902.
64. Hulnick DH, Naidich DP, McCauley DI. CT of pleural tubercu- 94. Wurnig PN, et al. Video-assisted thoracic surgery for pleural
losis. Radiology. 1983;149:759–765. empyema. Ann Thorac Surg. 2006;81:309–313.
65. Ellis JRC, Gleeson FV. Non-traumatic thoracic emergencies: 95. Cassina PC, et al. Video-assisted thoracoscopy in the treatment
imaging and treatment of thoracic fluid collections (including of pleural empyema: stage-based management and outcome.
pneumothorax). Eur Radiol. 2002;12:1922–1930. J Thorac Cardiovasc Surg. 1999;117:234–238.
66. Lim KE, et al. Diagnosis of malpositioned chest tubes after 96. Kim BY, et al. Video-assisted thoracoscopic decortication for
emergency tube thoracostomy: is computed tomography more management of postpneumonic pleural empyema. Am J Surg.
accurate than chest radiograph? Clin Imaging. 2005;29:401–405. 2004;188:321–324.
67. Ozkan OS, Omen MN, Akhan O. Percutaneous management of 97. Satish B, Bunker M, Seddon P. Management of thoracic
parapneumonic effusions. Eur J Radiol. 2005;55:311–320. empyema in childhood: does the pleural thickening matter.
68. Perkins GD, Thickett D. CT-guided biopsy for diagnosis of Arch Dis Child. 2003;88:918–921.
malignant disease in pleural effusions. Lancet. 2003;362:173. 98. Rudd RM. New developments in asbestos-related pleural dis-
69. Leung AN, Müller NL, Miller RR. CT in differential diagnosis of ease. Thorax. 1996;51:210–216.
diffuse pleural disease. AJR Am J Roentgenol. 1990;154:487–492. 99. Sison RF, et al. Pulmonary disease associated with pleural
70. Bartlett JG. Anaerobic bacterial pneumonitis. Am Rev Respir Dis. “asbestos” plaques. Chest.1989;95:835.
1979;119:19–23. 100. Gefter WB, Epstein DM, Miller WT. Radiographic evaluation of
71. Robinson BR. The treatment of lung abscess. Curr Concepts. asbestos-related chest disorders. CRC Crit Rev Diagn Imaging.
1985;87:709–710. 1984;21:133–181.
72. Viamonte MJ, Parks RE, Smoak WM. Guided catheterization of 101. Aberle DR, et al. Asbestos-related pleural and parenchymal
the bronchial arteries, Part 2: pulmonary and mediastinal fibrosis: detection with high-resolution CT. Radiology. 1988;166:
neoplasms. Radiology. 1965;1965:205–230. 729–734.
73. Baron MG. Radiologic notes in cardiology: interlobar effusion. 102. Aberle DR, Hamsu G, Ray CS. High-resolution CT of benign
Circulation. 1971;44: 475–483. asbestos-related diseases: clinical and radiologic correlation.
74. Pecorari A, Weisbrod GL. Computed tomography of pseudotu- AJR Am J Roentgenol. 1988;151: 883–891.
moral pleural fluid collections in the azygoesophageal recess. 103. Aberle DR, Balmes JR. Computed tomography of asbestos-
J Comput Assist Tomogr. 1989;13:803–805. related pulmonary and pleural diseases. Clin Chest Med.
75. Maffessanti M, Tommasi M, Pelegrini P. Computed tomography 1991;12:115–131.
of free pleural effusions. Eur J Radiol. 1987;7:87–90. 104. Friedman AC, et al. Asbestos-related pleural disease and
76. Waite RJ, et al. Parietal pleural changes in empyema: appear- asbestosis: a comparison of CT and chest radiography. AJR Am
ances at CT. Radiology. 1990;175:145–150. J Roentgenol. 1988;150:269–275.
77. Aquino SL, Webb WR, Gushiken BJ. Pleural exudates and tran- 105. Gamsu G, Aberle DR, Lynch D. Computed tomography in the
sudates: diagnosis with contrast-enhanced CT. Radiology. 1994; diagnosis of asbestos-related thoracic disease. J Thorac Imaging.
192:803–808. 1989;4:61–67.
78. Himmelman RB, Callen PW. The prognostic value of loc- 106. Al Jarad N, et al. Assessment of asbestos induced pleural disease
ulations in parapneumonic pleural effusions. Chest. 1986;90: by computed tomography: correlation with chest radiograph
852–856. and lung function. Respir Med. 1991;85:203–208.
79. Neff CC, et al. CT follow-up of empyemas: pleural peels resolve after 107. McLoud TC, et al. Diffuse pleural thickening in an asbestos-
percutaneous catheter drainage. Radiology. 1990;176:195–197. exposed population: prevalence and causes. AJR Am J Roentgenol.
80. Heffner JE, et al. Pleural fluid chemical analysis in parapneu- 1985;144:9–18.
monic effusions: a meta-analysis. Am J Respir Crit Care Med. 108. Rupp SB, Jolles H. Calcified plaque in the superior portion of
1995;151:1700–1708. the major fissure: an unusual manifestation of asbestos expo-
81. Donnelly LF, Klosterman LA. CT appearance of parapneumonic sure. Chest. 1989;96:1436–1437.
effusions in children: findings are not specific for empyema. AJR 109. Craighead JE, Abraham JL, Churg A, et al. The pathology of
Am J Roentgenol. 1997;169:179–182. asbestos-associated diseases of the lungs and pleural cavities:
82. Takasugi J.E, Godwin JD, Teefey SA. The extrapleural fat in diagnostic criteria and proposed grading schema. Arch Pathol
empyema: CT appearance. Br J Radiol. 1991;64:580–583. Lab Med. 1982;106:544–596.
83. Merriam MA, et al. Radiographically guided percutaneous 110. Churg A. Asbestos fibers and pleural plaques in a general
catheter drainage of pleural fluid collections. AJR Am J Roentgenol. autopsy population. Am J Pathol. 1982;109:88–96.
1988;151:1113–1116. 111. Epler GR, McLoud TC, Gaensler EA. Prevalence and incidence of
84. vanSonnenberg E, Nakamoto SK, Mueller PR, et al. CT and benign asbestos pleural effusion in a working population.
ultrasound-guided catheter drainage of empyemas after chest- JAMA. 1982;247:617–622.
tube failure. Radiology. 1984;151:349–353. 112. Halvorsen RA, et al. CT differentiation of pleural effusion from
85. Westcott J. Percutaneous catheter drainage of pleural effusion ascites: an evaluation of four signs using blinded analysis of
and empyema. AJR Am J Roentgenol. 1985;144:1189–1193. 52 cases. Invest Radiol. 1986;21:391–395.
86. Bouros D, et al. Intrapleural streptokinase versus urokinase in 113. Gefter WB, Conant EF. Issues and controversies in the plain-film
the treatment of complicated parapneumonic effusions: a diagnosis of asbestos-related disorders in the chest. J Thorac
prospective, double-blind study. Am J Respir Crit Care Med. Imaging. 1988;3:11–28.
1997;155:291–295. 114. Becklake MR. State of the art: asbestos-related disease of the
87. Stark DD, Federle MP, Goodman PC. CT and radiographic lung and other organs: their epidemiology and implications for
assessment of tube thoracostomy. AJR Am J Roentgenol. 1983; clinical practice. Am Rev Respir Dis. 1976;114:187–227.
141:253–258. 115. Jones RN, McLoud T, Rockoff SD. The radiographic pleural
88. Webb WR. LaBerge JM. Radiographic recognition of chest tube abnormalities in asbestos exposure: relationship to physiologic
malposition in the major fissure. Chest. 1984;85:81–83. abnormalities. J Thorac Imaging. 1988;3:57–66.
89. Thourani VH, et al. Twenty-six years of experience with the 116. Schwartz DA, et al. Asbestos-induced pleural fibrosis and
modified Eloesser flap. Ann Thorac Surg. 2003;76:401–405. impaired lung function. Am Rev Respir Dis. 1990;141:321–326.
90. Panicek DM, et al. Chest tube tracks. Radiographics. 1987; 117. Kreel L. Computer tomography in the evaluation of pulmonary
7:321–342. asbestosis: preliminary experiences with the EMI general pur-
91. Shapiro MP, Gale ME, Daly BD. Eloesser window thoracostomy pose scanner. Acta Radiol. (Stockh) 1976;17: 4.
for treatment of empyema: radiographic appearance. AJR Am 118. Katz D, Kreel L. Computed tomography in pulmonary asbesto-
J Roentgenol. 1988;150:549–552. sis. Clin Radiol. 1979;30:207–213.
119. Chapman SJ, et al. Benign asbestos pleural diseases. Curr Opin 151. Ryan CJ, et al. The outcome of patients with pleural effusions
Pulmon Med. 2003;9:266–271. of indeterminate cause at thoracotomy. Mayo Clin Proc. 1981;
120. Cugell DW, Kamp WD. Asbestos and the pleura: a review. Chest. 56:145–149.
2004;125:1103–1117. 152. Tattersall MHN, Boyer MJ. Management of malignant effusions.
121. Bégin R, et al. Radiographic assessment of pleuropulmonary Thorax. 1990;45:81–82.
disease in asbestos workers: posteroanterior, four view films, 153. Miller KS, Sahn S. Chest tubes: indications, technique, manage-
and computed tomograms of the thorax. Br J Ind Med. ment and complications. Chest. 1987;1987:258–264.
1984;41:373–383. 154. Martini N, Beatatie EJ. Indications for pleurectomy in malignant
122. Rabinowitz JG, et al. A comparative study of mesothelioma and effusion. Cancer. 1975;35:734–738.
asbestosis using computed tomography and conventional chest 155. O’Donovan PB, Eng P. Pleural changes in malignant pleural
radiography. Radiology. 1982;144:453–460. effusions: appearance on computed tomography. Clev Clin J
123. Menzies R, Fraser R. Round atelectasis: pathologic and patho- Med. 1994;61:127–131.
genetic features. Am J Surg Pathol. 1987;11:674–681. 156. Colt HG. Thoracoscopic management of malignant effusions.
124. Schneider H, Felson B, Gonzalez L. Rounded atelectasis. AJR Am Clin Chest Med. 1995;16:505–518.
J Roentgenol. 1980;134:225–232. 157. Müller NL. Imaging the pleura. Radiology. 1993;186:297–309.
125. Mintzer RA, et al. Rounded atelectasis and its association with 158. McLoud TC, Flower CDR. Imaging the pleura: sonography, CT,
asbestos-induced pleural disease. Radiology. 1981;139:567–570. and MR imaging. AJR Am J Roentgenol. 1991;156:1145–1153.
126. Hillerdal G. Rounded atelectasis: clinical experience with 159. Bury T, et al. Evaluation of pleural diseases with FDG-PET imag-
74 patients. Chest. 1989;95:836–941. ing: preliminary report. Thorax. 1997;52:187–189.
127. Doyle TC, Lawler GE. CT features of rounded atelectasis of the 160. Erasmus JJ, et al. FDG PET of pleural effusions in patients with
lung. AJR Am J Roentgenol. 1984;143:225–228. non-small cell lung cancer. AJR Am J Roentgenol. 2000;175:
128. McHugh K, Blaquiere RM. CT features of rounded atelectasis. 245–249.
AJR Am J Roentgenol. 1989;153:257–260. 161. Gerbaudo VH, et al. Metabolic significance of the pattern, inten-
129. Lynch DA. Asbestos-related focal lung masses: manifestations sity and kinetics of F-18-FDG uptake in malignant pleural
on conventional and high-resolution CT scans. Radiology. 1988; mesothelioma. Thorax. 2003;58:1077–1082.
169:603–607. 162. Duysinx B, et al. Evaluation of pleural disease with 18fluoro-
130. Batra P, et al. Rounded atelectasis. J Thorac Imaging. 1996;11: deoxyglucose positron emission tomography imaging. Chest.
187–197. 2004.
131. Ren H, et al. Computed tomography of rounded atelectasis. 163. Mavi A, et al. Fluorodeoxyglucose-PET in characterizing solitary
J Comput Assist Tomogr. 1988;12:1031–1034. pulmonary nodules, assessing pleural diseases, and the initial
132. Ren H, et al. Case report: computed tomography of rounded staging, restaging, therapy planning, and monitoring response
atelectasis. J Comput Assist Tomogr. 1989;12:1031–1034. of lung cancer. Radiol Clin North Am. 2005;43:1.
133. Taylor PM. Dynamic contrast enhancement of asbestos-related 164. Schaffler GJ, et al. Non-small cell lung cancer: evaluation of
pulmonary pseudotumors. Br J Radiol. 1988;61:1070–1072. pleural abnormalities on CT scans with18F FDG PET. Radiology.
134. Coleman BG, et al. Case report: CT features of an unusual 2004;231:858–865.
hypervascular lung carcinoma complicating chronic asbestosis 165. Kramer H, et al. PET for the evaluation of pleural thickening
related pleural disease. J Comput Assist Tomogr. 1985;9:554–557. observed on CT. J Nucl Med. 2004;45:995–998.
135. Sinner WWN. Pleuroma, a cancer mimicking atelectatic pseudotu- 166. Gupta NC, et al. Clinical role of F-18 fluorodeoxyglucose
mor of the lung. Fortschr Geb Rontgenst Nuklarmed Erganzungsband. positron emission tomography imaging in patients with lung
1980;133:578–585. cancer and suspected malignant pleural effusion. Chest. 2002;
136. Reifsnyder AC, et al. Case report: malignant fibrous histiocy- 122:1918–1924.
toma of the lung in a patient with a history of asbestos expo- 167. Truong MT, Marom EM, Erasmus JJ. Preoperative evaluation of
sure. AJR Am J Roentgenol. 1990;154:65–66. patients with malignant pleural mesothelioma: role of inte-
137. Blesovsky A. The folded lung. Br J Dis Chest. 1966;60:19–22. grated CT-PET imaging. J Thorac Imaging. 2006;21:146–153.
138. Hanke R, Kretzschmar R. Rounded atelectasis. Semin Roentgenol. 168. Wang ZY, et al. Malignant pleural mesothelioma: evaluation
1980;15:174–182. with CT, MR imaging, and PET. Radiographics. 2004;24:105–119.
139. Stancato-Pasik A, Mendelson DS, Marom Z. Case report: 169. Shim SS, et al. Integrated PET/CT and the dry pleural dissemina-
rounded atelectasis caused by histoplasmosis. AJR Am J tion of peripheral adenocarcinoma of the lung: diagnostic
Roentgenol. 1990;155:275–276. implications. J Comput Assist Tomogr. 2006;30:70–76.
140. McAdams HP, et al. Evaluation of patients with round atelecta- 170. Heitzman ER, et al. Annual oration: pathways of tumor spread
sis using 2-(18F)fluoro-2-deoxy-D-glucose PET. J Comput Assist through the lung: radiologic correlations with anatomy and
Tomogr. 1998;22:601–604. pathology. Radiology. 1982;144:3–14.
141. Munden RF, Libshitz HI. Rounded atelectasis and mesothe- 171. Webb WR, Jeffrey RB, Godwin JD. Thoracic computed tomogra-
lioma. AJR Am J Roentgenol. 1998;170:1519–1522. phy in superior sulcus tumors. J Comput Assist Tomogr. 1981;
142. Verschakelen JA, et al. Rounded atelectasis of the lung: MR 5:361–365.
appearance. AJR Am J Roentgenol. 1989;152:965–966. 172. Glazer HS, et al. Pleural and chest wall invasion in bronchogenic
143. Weber MA, et al. Asbestos-related pleural disease: value of dedi- carcinoma: CT evaluation. Radiology. 1985;157:191–194.
cated magnetic resonance imaging techniques. Invest Radiol. 173. Pennes DR, et al. Chest wall invasion by lung cancer: limita-
2004;39:554–564. tions of CT evaluation. AJR Am J Roentgenol. 1985;144:507–511.
144. Chernow B, Sahn SA. Carcinomatous involvement of the 174. Shin MS, et al. Case report: pitfalls in CT evaluation of chest
pleura: an analysis of 96 patients. Am J Med. 1977;63:695–702. wall invasion by lung chest. J Comput Assist Tomogr. 1986;
145. Sahn SA. Pleural diseases related to metastatic malignancies. Eur 10:136–138.
Respir J. 1997;10:1907–1913. 175. Pearlberg JL, et al. Limitations of CT in evaluation of neo-
146. Meyer PC. Metastatic carcinoma of the pleura. Thorax. plasms involving chest wall. J Comput Assist Tomogr. 1987;
1966;21:437–443. 11:290–293.
147. Scerbo J, Keltz H, Stone DJ. A prospective study of closed pleu- 176. Grenier P, et al. Preoperative thoracic staging of lung cancer: CT
ral biopsies. JAMA. 1971;218:377–380. and MR evaluation. Diagn Intervent Radiol. 1989;173(P):69.
148. Salyer WR, Eggleston JC, Erozan YS. The efficacy of pleural 177. Heelan RT, et al. Superior sulcus tumors. Radiology. 1989;
needle biopsy and of pleural fluid cytology in the diagnosis of 170:637.
malignant neoplasm involving the pleura. Chest. 1975;67:5. 178. McLoud TC, et al. MR imaging of superior sulcus carcinoma.
149. Whitaker D. The cytology of malignant mesothelioma. J Comput Assist Tomogr. 1989;13:233–239.
Cytopathology. 2000;11:139–151. 179. Webb WR, Jensen BJ, Sollito R, et al. Bronchogenic carcinoma:
150. Moriarty AL, et al. Immunophenotyping of cytologic specimens staging with MR compared with staging with CT surgery.
by flow cytometry. Diagn Cytopathol. 1993;9:252–258. Radiology. 1985;156:117.
180. Haggar AM, et al. Chest-wall invasion by carcinoma of the lung: patient with malignant pleural mesothelioma. J Thorac Cardiovasc
detection by MR imaging. AJR Am J Roentgenol. 1987; 148: Surg. 2003;1:11–16.
1075–1078. 209. Schneider DB, Clary-Mcay C, Challa S, et al. Positron emission
181. Pieterman RM, van Putten JW, Meuzelaar JJ, et al. Preoperative tomography with F18-fluorodeoxyglucose in the staging and
staging of non-small-cell lung cancer with positron-emission preoperative evaluation of malignant pleural mesothelioma.
tomography. N Engl J Med. 2000;343:254–261. J Thorac Cardiovasc Surg. 2000;1:128–133.
182. Robinson BWS, Musk AW, Lake RA. Malignant mesothelioma. 210. van Klaveren RJ, et al. Inadequacy of the RECIST criteria for
Lancet. 2005;366:397–408. response evaluation in patients with malignant pleural
183. Dunn MM. Asbestos and the lung. Chest. 1989;95:1304–1308. mesothelioma. Lung Cancer. 2004;43:63–69.
184. Robinson BWS, Lake RA. Advances in malignant mesothelioma. 211. Byrne MJ, Nowak AK. Modified RECIST criteria for assessment
N Engl J Med. 2005;353:1591–1603. of response in malignant pleural mesothelioma. Ann Oncol.
185. Metintas M, et al. CT-guided pleural needle biopsy in the diag- 2004;15:257–260.
nosis of malignant mesothelioma. J Comput Assist Tomogr. 1995; 212. Nowak AK. CT, RECIST, and malignant pleural mesothelioma.
19:370–374. Lung Cancer. 2005;49:S37–S40.
186. Pass HI, et al. Asbestos exposure, pleural mesothelioma, and 213. Zerhouni EA, et al. Invasion thymomas: diagnosis and evalua-
serum osteopontin levels. N Engl J Med. 2005;353:1564–1573. tion by computed tomography. J Comput Assist Tomogr. 1982;
187. Aisner J. The current approach to malignant mesothelioma of 6:92–100.
the pleura. Chest. 1995;107(suppl):S332–S344. 214. Scatariage JC, et al. Transdiaphragmatic extension of invasion
188. Erdogan E, et al. Incidental localized (solitary) mediastinal thymoma. AJR Am J Roentgenol. 1985;144:31–35.
malignant mesothelioma. Br J Radiol. 2005;78:858–861. 215. Falaschi F, et al. Usefulness of MR signal intensity in distin-
189. VanSchil P. Malignant pleural mesothelioma: staging systems. guishing benign from malignant pleural disease. AJR Am J
Lung Cancer. 2005;49:S45–S48. Roentgenol. 1996: 166:963–968.
190. Sugarbaker DJ, Garcia JP. Multimodality therapy for malignant 216. Blank N, Casttellino RA. The intrathoracic manifestations of the
pleural mesothelioma. Chest. 1997;112(4 suppl):S272–S275. malignant lymphomas and leukemia. Semin Roentgenol.
191. Sugarbaker DJ, et al. Extrapleural pneumonectomy in the multi- 1980;15:227–245.
modality therapy of malignant pleural mesothelioma: results in 217. Xaubet A, Diumenjo MC, Marin A, et al. Characteristics and
120 consecutive patients. Ann Surg. 1996;224:288–294. prognostic value of pleural effusions in non-Hodgkin’s
192. Rusch VW. A proposed new international TNM staging system lymphomas. Eur J Respir Dis. 1985;66:135–140.
for malignant pleural mesothelioma from the international 218. Fisher AMH, Kendall B, Van Leuven BD. Hodgkin’s disease:
mesothelioma interest group. Lung Cancer. 1996;14:1–12. a radiologic survey. Clin Radiol. 1962;13:115–127.
193. Patz EF, Rusch VW, Heelan R. The proposed new international 219. Stolberg HO, et al. Hodgkin’s disease of the lung: radiologic-
TNM staging system for malignant pleural mesothelioma: appli- pathologic correlation. AJR Am J Roentgenol. 1964;192:96–115.
cation to imaging. AJR Am J Roentgenol. 1996;166:323–327. 220. Burgener FH, Hamlin DJ. Intrathoracic histiocytic lymphoma.
194. Dedrick CG, et al. Computed tomography of localized pleural AJR Am J Roentgenol. 1981;136:499–504.
mesothelioma. AJR Am J Roentgenol. 1985;144:275–280. 221. Bernadreschi P, Bonechi I, Urbano U. Recurrent pleural effusion
195. Alexander E, et al. CT of malignant pleural mesothelioma. AJR as manifesting feature of primitive chest wall Hodgkin’s disease.
Am J Roentgenol. 1981;137:287–291. Chest. 1988;94:424–426.
196. Grant DC, et al. Computed tomography of malignant pleural 222. Carlson SE, Bergin CJ, Hoppe RT. MR imaging to detect chest
mesothelioma. J Comput Assist Tomogr. 1983;7:626–632. wall and pleural involvement in patients with lymphoma: effect
197. Mirvis S, et al. CT of malignant pleural mesothelioma. AJR Am J on radiation therapy planning. AJR Am J Roentgenol. 1993;
Roentgenol. 1983;140:665–670. 160:1191–1195.
198. Salonen O, et al. Computed tomography of pleural lesions with 223. Tsukada G, Stark P. Pictorial essay: postpneumonectomy com-
special reference to the mediastinal pleura. Acta Radiol. 1986; plications. AJR Am J Roentgenol. 1997;169:1363–1370.
27:527–531. 224. Biondetti PR, et al. Evaluation of the post-pneumonectomy
199. Rusch VW, Godwin JD, Shuman WP. The role of computed space by computed tomography. J Comput Assist Tomogr. 1982;
tomography scanning in the initial assessment and the follow- 6:238–242.
up of malignant pleural mesothelioma. J Thorac Cardiovasc Surg. 225. Glazer HS, et al. Utility of CT in detecting postpneumonectomy
1988;96:171–177. carcinoma recurrence. AJR Am J Roentgenol. 1984;142:487–494.
200. Kawashima A, Libshitz HI. Malignant pleural mesothelioma: CT 226. Peters JC, Desai KK. CT demonstration of postpneumonectomy
manifestations in 50 cases. AJR Am J Roentgenol. 1990;155: tumor recurrence. AJR Am J Roentgenol. 1983;141:259–262.
965–969. 227. Heater K, Revzani L, Rubin JM. CT evaluation of empyema in
201. Uri AJ, et al. Diffuse contralateral pulmonary metastases in the postpneumonectomy space. AJR Am J Roentgenol. 1985;
malignant mesothelioma: an unusual radiographic presenta- 145:39–40.
tion. Chest. 1988;93:433–434. 228. Shepard J-AO, et al. Right-pneumonectomy syndrome: radio-
202. Sahin AA, et al. Malignant pleural mesothelioma caused by logic findings and CT correlation. Radiology. 1986;161:
environmental exposure to asbestos or erionite in rural Turkey: CT 661–664.
findings in 84 patients. AJR Am J Roentgenol. 1993;161:533–537. 229. Laissy J-P, et al. MR appearance of the normal chest after
203. Raizon A, et al. Calcification as a sign of sarcomatous degenera- pneumonectomy. J Comput Assist Tomogr. 1989;13: 248–252.
tion of malignant pleural mesotheliomas: a new CT finding. 230. Robinson LA, Reilly RB. Localized pleural mesothelioma: the
J Comput Assist Tomogr. 1996;20:42–44. clinical spectrum. Chest. 1994;106:1611–1615.
204. Taylor RD, et al. Metastatic renal cell carcinoma mimicking 231. Lee KS, et al. CT findings in benign fibrous mesothelioma of
pleural mesothelioma. Thorax. 1987;42:901–992. the pleura: pathologic correlation in nine patients. AJR Am J
205. Patz EF, et al. Malignant pleural mesothelioma: value of CT and Roentgenol. 1992;158:983–986.
MR imaging in predicting resectability. AJR Am J Roentgenol. 232. Ferretti GR, et al. Localized benign fibrous tumors of the pleura:
1992;159:961–966. MR appearance. J Comput Assist Tomogr. 1997;21:115–120.
206. Benamore RE, O’Doherty MJ, Entwisle JJ. Use of imaging in the 233. de Perrot M, et al. Review: solitary fibrous tumors of the pleura.
management of malignant pleural mesothelioma. Clin Radiol. Ann Thorac Surg. 2002;74:285–293.
2005;60:1237–1247. 234. Okike N, Bernatz PE, Woolner LB. Localized mesothelioma of
207. Erasmus JJ, et al. Integrated computed tomography-positron the pleura. J Thorac Cardiovasc Surg. 1978;75:363–372.
emission tomography in patients with potentially resectable 235. Briselli M, Mark EJ, Dickersin GR. Solitary fibrous tumors of the
malignant pleural mesothelioma: staging implications. J Thorac pleura: eight new cases and review of 60 cases in the literature.
Cardiovasc Surg. 2005;129:1364–1370. Cancer. 1981;47:2678–2689.
208. Flores RM, Akhurst T, Goen R, et al. Positron emission tomogra- 236. Destouet JM, et al. Computed tomography of the sternoclavicular
phy defines metastatic disease but not locoregional disease in joint and sternum. Radiology. 1981; 138:123–128.
237. Goodman LR, Teplick SK, Kay H. Computed tomography of the 270. Kim HC, et al. Systemic-to-pulmonary venous shunt in superior
normal sternum. AJR Am J Roentgenol. 1983;141:219–223. vena cava obstruction: depiction on computed tomography
238. Goodman LR, et al. Complications of median sternotomy: com- venography. Acta Radiol. 2004;45:269–274.
puted tomographic evaluation. AJR Am J Roentgenol. 1983; 271. Gerard PS, et al. Transient thoracic venous collaterals: an inci-
141:225–230. dental CT finding. J Comput Assist Tomogr. 1986;10:75–77.
239. Hatfield MK, et al. Computed tomography of the sternum and 272. Engel IA, et al. CT diagnosis of mediastinal and thoracic inlet
its articulations. Skeletal Radiol. 1984;11:197–203. venous obstruction. AJR Am J Roentgenol. 1983;141:521–526.
240. Stark P, Jaramillo D. Pictorial essay. CT of the sternum. AJR Am J 273. Gosselin MV, Rubin GD. Altered intravascular contrast material
Roentgenol. 1986;147:72–77. flow dynamics: clues for refining thoracic CT diagnosis. AJR Am
241. Stark P, et al. Episternal ossicles. Radiology. 1987;165:143–144. J Roentgenol. 1997;169:1597–1603.
242. Jafri SZ, et al. Computed tomography of chest wall masses. 274. Lomoschitz FM, et al. Imaging of chest trauma: radiological pat-
Radiographics.1989;9:51–68. terns of injury and diagnostic algorithms. Eur J Radiol. 2003;
243. Carter AR, et al. Thoracic alterations after cardiac surgery. AJR 48:61–70.
Am J Roentgenol. 1983;140:475–481. 275. Scaglione M, Pinto F, Grassi R, et al. Diagnostic sensitivity of
244. Leitman BS, Naidich DP, McCauley SI. Computed tomography computerized tomography in closed trauma of the diaphragm:
of pectoral flaps. J Comput Assist Tomogr. 1988;12:392–393. retrospective study of 35 consecutive cases. Radiol Med. (Torino)
245. Aoki J, Mosser RP, Kransdork MJ. Chondrosarcoma of the ster- 2000;99:46–50.
num: CT features. J Comput Assist Tomogr. 1989;13:806–810. 276. Tello R, et al. Value of spiral CT in hemodynamically stable
246. Levinsohn EM, Bunnell WP, Wuan HA. Computed tomography patients following blunt chest trauma. Comput Med Imaging
in the diagnosis of dislocations of the sternoclavicular joint. Graph. 1998;22:447–452.
Clin Orthop. 1979;140:12–16. 277. Wolfman NT, et al. Occult pneumothorax in patients with
247. Burnstein MI, Pozniak A. Case report: computed tomography abdominal trauma: CT studies. J Comput Assist Tomogr. 1993;
with stress maneuver to demonstrate sternoclavicular joint 17:56–59.
dislocation. J Comput Assist Tomogr. 1990;14:159–160. 278. Kerns SB, Gay SB. Pictorial essay: CT of blunt chest trauma. AJR
248. Sartoris DJ, et al. Sternoclavicular hyperostosis: a review and Am J Roentgenol. 1990;154:55–60.
report of 11 cases. Radiology. 1986;158:125–128. 279. Tocino IM, et al. CT detection of occult pneumothorax in head
249. Rafii M, Firooznia H, Golimbu C. Computed tomography of trauma. AJR Am J Roentgenol. 1984;143:987–990.
septic joints. Comput Tomogr. 1985:9:51–60. 280. Bhalla M, et al. Lung hernia: radiographic features. AJR Am J
250. Alexander PW, Shin SI. CT manifestations of sternoclavicular Roentgenol. 1990;154:51–53.
pyarthrosis in patients with intravenous drug abuse. J Comput 281. Chern MS, Wu MH, Chang CY. CT and MRI for staging of locally
Assist Tomogr. 1990.;14:104–106. advanced non-small cell lung cancer. Lung Cancer. 2003;
251. Pollack MS. Staphylococcal mediastinitis due to sternoclavicular 42:S5–S8.
pyarthrosis: CT appearance. J Comput Assist Tomogr. 1990: 282. Kodalli N, Erzen Yuksel M. Evaluation of parietal pleural inva-
14:924–928. sion of lung cancers with breathhold inspiration and expiration
252. Jung JI, et al. Thoracic manifestations of breast cancer and its MRI. Clin Imaging. 1999;23:227–235.
therapy. Radiographics. 2004;24:1269–1285. 283. Kinoshita T, Ishii K, Miyasato S. Localized pleural mesothe-
253. Bhalla M, et al. Counting ribs on chest CT. J Comput Assist lioma: CT and MR findings. Magn Reson Imaging. 1997;
Tomogr. 1990;14:590–594. 15:377–379.
254. Fishman EK, et al. CT of the axilla: normal anatomy and pathol- 284. Hergan K, et al. MR and CT anatomy of the axilla. Acta Radiol.
ogy. Radiographics. 1986;6:475–502. 1997;38:198–205.
255. Goldberg RP, Austin RM. Computed tomography of axillary and 285. Heelan RT, et al. Magnetic resonance imaging of the postp-
supraclavicular adenopathy. Clin Radiol. 1985;36:593–596. neumonectomy chest: normal and abnormal findings. J Thorac
256. Holbert BL, Holbert JM, Libshitz HI. CT of interpectoral lymph Imaging. 1997;12:200–208.
nodes. AJR Am J Roentgenol. 1987;149:687–688. 286. Frola C, et al. Transudative vs exudative pleural effusions: differ-
257. Gross JL, et al. Soft-tissue sarcomas of the chest wall: prognostic entiation using Gd-DTPA-enhanced MRI. Eur Radiol. 1997;7:
factors. Chest. 2005;127:902–908. 860–864.
258. Wong KS, et al. Thoracic wall lesions in children. Pediatr 287. Brossmann J, et al. Sternoclavicular joint: MR imaging-anatomic
Pulmonol. 2004;37:257–263. correlation. Radiology. 1996;198:193–198.
259. Morris BS, Maheshwari M, Chalwa A. Chest wall tuberculosis: 288. Gierada DS, et al. Diaphragmatic motion: fast gradient-recalled-
a review of CT appearances. Br J Radiol. 2004;77:449–457. echo MR imaging in healthy subjects. Radiology. 1995;194:
260. Tateishi U, et al. Chest wall tumors: radiologic findings and 879–884.
pathologic correlation, Part 1: benign tumors. Radiographics. 289. Brown JJ, et al. Magnetic resonance relaxation times of percuta-
2003;23:1477–1490. neously obtained normal and abnormal body fluids. Radiology.
261. Tateishi U, et al. Chest wall tumors: radiologic findings and 1985;727–731.
pathologic correlation, Part 2: malignant tumors. Radiographics. 290. Terrier F, et al. MR imaging of body fluid collections. J Comput
2003.23:1491–1508. Assist Tomogr. 1986;10: 953–962.
262. Jeung MY, et al. Imaging of chest wall disorders. Radiographics. 291. Davis SD, et al. MR imaging of pleural effusions. J Comput Assist
1999;19:617–637. Tomogr. 1990;14:192–198
263. Lindfors KK, et al. Evaluation of local and regional breast cancer 292. Lorigan JG, Libshitz HI. MR imaging of malignant pleural
recurrence. AJR Am J Roentgenol. 1985;145:833–837. mesothelioma. J Comput Assist Tomogr. 1989;13:617–620.
264. Faer MJ, Burnam RE, Beck CL. Transmural thoracic lipoma: 293. Webb RW, et al. CT and MR in staging non-small cell bron-
demonstration by computed tomography. AJR Am J Roentgenol. chogenic carcinoma: report of the Radiological Diagnostic
1978;130:161–163. Oncology Group. Radiology. 1989;173(P):69.
265. Sullivan WT. Extrapleural fat prolapse mimicking recurrent 294. Mcloud TC. CT and MR in pleural disease. Clin Chest Med.
bronchogenic carcinoma. Comput Tomogr. 1986;10:277–279. 1998;19:261.
266. Buxton RC, et al. Atypical transmural thoracic lipoma: CT diag- 295. Bergin CJ, et al. MR evaluation of chest wall involvement in
nosis. J Comput Assist Tomogr. 1988;12:196–198. malignant lymphoma. J Comput Assist Tomogr. 1990;14:
267. Webb WR, Sagel SS. Actinomycosis involving the chest wall: 928–933.
CT findings. AJR Am J Roentgenol. 1982;139:1007–1009. 296. Sharif HS, et al. MR imaging of thoracic and abdominal wall
268. Allen HA, Scatarige JC, Kim MH. Actinomycosis: CT findings in infections: comparison with other imaging procedures. AJR Am
six patients. AJR Am J Roentgenol. 1987;149:1255–1258. J Roentgenol. 1990;154:989–995.
269. James R, Heneghan NA, Lipansky V. Thoracic actinomycosis. 297. Panicek DM, et al. The diaphragm: anatomic, pathologic, and
Chest. 1985;87:536–637. radiologic considerations. Radiographics. 1988;8:385.
298. De Troyer A, et al. The diaphragm: two muscles. Science. 1981; 328. Shapiro MJ, et al. The unreliability of CT scans and initial chest
213:237. radiographs in evaluating blunt trauma induced diaphragmatic
299. Patterson NW, Teates CD. CT measurements of the anterior por- rupture. Clin Radiol. 1995;51:27–30.
tions of the diaphragm with illustrative abnormal cases. Comput 329. Murray JG, et al. Acute rupture of the diaphragm due to blunt
Radiol. 1985;9:61–65. trauma: diagnostic sensitivity and specificity of CT. AJR Am J
300. Kleinman PK, Raptopoulos V. The anterior diaphragmatic Roentgenol. 1996;166:1035–1039.
attachments: an anatomic and radiologic study with clinical 330. Mirvis SE, et al. MR imaging of traumatic diaphragmatic rup-
correlates. Radiology. 1985;155:289–293. ture. J Comput Assist Tomogr. 1988;12:147–149.
301. Gale EM. Anterior diaphragm: variations in the CT appearance. 331. Naidich DP, et al. Computed tomography of the diaphragm:
Radiology. 1986;161:635–639. peridiaphragmatic fluid localization. J Comput Assist Tomogr.
302. Naidich DP, et al. Computed tomography of the diaphragm: 1983;7:641–649.
normal anatomy and variants. J Comput Assist Tomogr. 1983;7: 332. Dwyer A. The displaced crus: a sign for distinguishing between
633–640. pleural fluid and ascites on computed tomography. J Comput
303. Shin MS, Berland LL. Computed tomography of retrocrural Assist Tomogr. 1978;2:598–599.
spaces: normal, anatomic variants, and pathologic conditions. 333. Teplick JG, et al. The interface sign: a computed tomographic
AJR Am J Roentgenol. 1985;145:81–86. sign for distinguishing pleural and intraabdominal fluid.
304. Silverman PM, Cooper C, Zeman RK. Lateral arcuate ligaments Radiology. 1982;144:359–362.
of the diaphragm: anatomic variations at abdominal CT. 334. Alexander ES, Proto AU, Clark RA. CT differentiation of sub-
Radiology. 1992;185:105–108. phrenic abscess and pleural effusion. AJR Am J Roentgenol.
305. Williamson BRJ, et al. Variation in the thickness of the diaphrag- 1983;140:47–51.
matic crura with respiration. Radiology. 1987;163:683. 335. Silverman PM, Baker ME, Mahoney BS. Atelectasis and subpul-
306. Nightingale RC, Dixon AK. Crural change with respiration: monic fluid: a CT pitfall in distinguishing pleural from peri-
a potential mimic of disease. Br J Radiol. 1984;57:101–102. toneal fluid. J Comput Assist Tomogr. 1985;9:763–766.
307. Anda S, et al. CT appearance of the diaphragm varying with 336. Rubenstein WA, et al. The perihepatic spaces: computed
respiratory phase and muscular tension. J Comput Assist Tomogr. tomographic and ultrasound imaging. Radiology. 1985;149:
1986;10:744–745. 231–239.
308. Caskey CI, et al. Aging of the diaphragm: a CT study. Radiology. 337. Dodds WJ, et al. Anatomy and imaging of the lesser peritoneal
1989;171:385–389. sac. AJR Am J Roentgenol. 1985;14:567–575.
309. Rosen A, et al. CT appearance of diaphragmatic pseudotumors. 338. Vibhakar SD, Bellon EM. The bare area of the spleen: a constant
J Comput Assist Tomogr. 1983;7:995–999. CT feature of the ascitic abdomen. AJR Am J Roentgenol.
310. Cassart M, et al. Effects of lung volume reduction surgery for 1984;141:953–955.
emphysema on diaphragm dimensions and configuration. Am 339. Katzen BT, et al. Pseudo-mass of the liver due to pleural effusion
J Respir Crit Care Med. 2001;163:1042–1043. and inversion of the diaphragm. AJR Am J Roentgenol. 1978;
311. Gale ME. Bochdalek hernia: prevalence and CT characteristics. 131:1077–1078.
Radiology. 1985;156:449–452. 340. Dallemand S, Twersky J, Gordon DH. Pseudomass of the left
312. Demartini WJ, House AJS. Partial Bochdalek’s herniation: com- upper quadrant from inversion of the left hemidiaphragm: CT
puterized tomographic evaluation. Chest. 1980;77:702–704. diagnosis. Gastrointest Radiol. 1982;7:57–59.
313. Curley FJ, Hubmayr RD, Raptopoulos V. Bilateral diaphragmatic 341. Ginalski JM, Schnyder P, Moss AA. Incidence and significance of
densities in a 72 year old woman. Chest. 1984;86:915–917. a widened esophageal hiatus at CT scan. J Clin Gastroenterol.
314. Weinshelbaum AM, Weinshelbaum EI. Incarcerated adult 1984;6:467–470.
Bochdalek hernia with splenic infarction. Gastrointest Radiol. 342. Godwin JD, MacGregor JM. Extension of ascites into the chest
1982;7:287–289. with hiatal hernia: visualization on CT. AJR Am J Roentgenol.
315. LaRosa JR DV, Esham RH, Morgan SL. Diaphragmatic hernia of 1987;148:31–32.
Morgagni. South Med J. 1999;92:409–411. 343. Balthazar EJ, et al. CT evaluation of esophageal varices. AJR Am
316. Pettiaux N, Cassart M, Paiva M, et al. Three-dimensional recon- J Roentgenol. 1987;148:131–135.
struction of human diaphragm with use of spiral computed 344. Raval B, Hall JT, Jackson H. CT diagnosis of fluid in the lesser
tomography. Appl Physiol. 1997;82:998–1002. sac mimicking thrombosis of inferior vena cava. J Comput Assist
317. Sener RN, et al. Bilateral large Bochdalek hernias in an adult: Tomogr. 1985;9:956–958.
CT demonstration. Clin Imaging. 1995;19:40–42. 345. Peterson MW, et al. Case report: CT findings in transdiaphrag-
318. Wilbur AC, Gorodetsky A, Hibbeln JF. Imaging findings of adult matic empyema necessitatis due to tuberculosis. J Comput Assist
Bochdalek hernias. Clin Imaging. 1994;18:224–229. Tomogr. 1987;1987:704–706.
319. Larici AR, Gotway MB, Litt HI, et al. Helical CT with sagittal and 346. Shuman LS, Libshitz HI. Pictorial essay: solid pleural manifesta-
coronal reconstructions: accuracy for detection of diaphrag- tions of lymphoma. AJR Am J Roentgenol. 1984;142:269–273.
matic injury. AJR Am J Roentgenol. 2002;179:451–457. 347. Rice DC, et al. Extended surgical staging for potentially resec-
320. Bergin D, Ennis R, Keogh C, et al. The “dependent viscera” sign table malignant pleural mesothelioma. Ann Thorac Surg. 2005;
in CT diagnosis of blunt traumatic diaphragmatic rupture. AJR 80:1988–1993.
Am J Roentgenol. 2001;177:1137–1140. 348. Estaun JE, Alfageme AG, Banuelos GS. Radiological appearance
321. Shanmuganathan K, Killeen K, Mirvis SE, et al. Imaging of of diaphragmatic mesothelial cysts. Pediatr Radiol. 2003;33:
diaphragmatic injuries. J Thorac Imaging. 2000;15:104–111. 855–858.
322. Leung JCM, et al. Thickening of the diaphragm: a new com- 349. Posniak HV, Keshavarzian A, Jabamoni R. Diaphragmatic end-
puted tomography sign of diaphragm injury. J Thorac Imaging. ometriosis: CT and MR findings. Gastrointest Radiol. 1990;
1999;14:126–129. 15:349–351.
323. Killeen KL, Mirvis SE, Shanmuganathan K. Helical CT of 350. Anderson LS IV. Tumors of the diaphragm. AJR Am J Roentgenol.
diaphragmatic rupture caused by blunt trauma. AJR Am 1973;119:259–265.
J Roentgenol. 1999;173:1611–1616. 351. Müller NL. CT features of cystic teratoma of the diaphragm.
324. Shackleton KL, Stewart ET, Tahylor AJ. Traumatic diaphragmatic J Comput Assist Tomogr. 1986;10:325–326.
injuries: spectrum of radiographic findings. Radiographics. 352. Gierada DS, et al. Fast gradient echo magnetic resonance imag-
1998;18:49–59. ing of the normal diaphragm. J Thorac Imaging. 1997;
325. Lee WC, et al. Rupture of the diaphragm after blunt trauma. Eur 12:70–74.
J Surg. 1994;160: 479–483. 353. Shanmuganathan K, et al. MR imaging evaluation of hemidi-
326. Aronchick JM, et al. Chronic traumatic diaphragmatic hernia: the aphragms in acute blunt trauma: experience with 16 patients.
significance of pleural effusion. Radiology. 1988;168: 675–678. AJR Am J Roentgenol. 1996;167:397–402.
327. Radin DR, Ray MJ, Halls JM. Strangulated diaphragmatic hernia 354. Yeager BA, et al. Magnetic resonance imaging of Morgagni
with pneumothorax. AJR Am J Roentgenol. 1986;146:321–322. hernia. Gastrointest Radiol. 1987;12:296–298.
5636_Naidich_IDX_pp885-898 12/8/06 12:13 PM Page 885
Index
Note: Page numbers followed by f indicate figures, page numbers followed by t indicate tabular matter.
A Airway(s), 453–555. See also Bronchus(i); pulmonary, 219

Abscess(es) Trachea negative, 224–225
lung. See Lung(s), abscess of anatomy of pitfalls in interpretation of,
mediastinal, 428–429, 429f cross-sectional, 463, 465–467f 247–258
in tuberculosis, 429f central in pulmonary embolism detection,
Acquired immunodeficiency syndrome CT/bronchoscopic correlations in, 221–222
(AIDS) 471–503 MR
laryngotracheobronchial papillomato- tumor of, 472, 475f of aorta, 89–96, 90–91f
sis in, 491, 493f CT of coronary, 25–27
Pneumocystis carinii pneumonia in. See clinical indications for, 453 for pulmonary embolism detection,
Pneumocystis carinii (jiroveci) pneu- contrast media in, 456–457, 457f 234, 235f
monia neoplastic obstruction on, 457, 457f in pulmonary embolism detection,
pulmonary complications of, 738, 739f technique of, 453–459, 454–459f 221–222, 222f
diagnostic strategies for, 738–739, diameter of Angiolipoma, 403
739f wall thickness and, 677, 678t Angiosarcoma, 51, 408
Actinomycosis dimensions of, 463t Angioscopy
chest wall involvement in, 851–856, disease of. See specific disorders virtual endoluminal, 265, 266f
856–857f normal Ankylosing spondylitis, 840f
Adenocarcinoma anatomy of, 460–463f, 460–469, Aorta
CT-pathologic correlations in, 562–563, 465–471f acute syndromes involving, 131–163.
563–569f pathology of See also specific disorders
of gastroesophageal junction, 416f CT classification of, 472, 473f ascending
of lung, 472, 479f. See also Lung peripheral, 512–543 aneurysm or dissection of
cancer CT evaluation of, 513–543, 517t surgery for, 173–176, 176f
as bronchoalveolar cell carcinoma, visibility on, 515–517f, 518 composite graft replacement of, 175
622–623, 624f terminology associated with, 462–463, branching of, 101–103, 102–103f
CT in, 623f 463f descending thoracic
incidence of, 622 Allergic bronchopulmonary aspergillosis surgery of, 176–181, 177–184f
“mixed type,” 622, 623f, 623t bronchiectasis in, 527–529, 528f disease of
WHO classification of, 625t Alveolar proteinosis, 684, 724–725 acquired, 119–131
metastatic to pleura, 810, 811f, 825f ground-glass attenuation in, 724, 725f congenital, 104–119. See also specific
Adenoid-cystic carcinoma American Thoracic Society lymph node disorders
of trachea, 489–491, 490f stations, 355, 357–358t, 360t endografts of
Adenoma Amiodarone lung, 693, 694f endoleaks in, 184f, 185
adrenal Amyloidosis, 501–502, 502f MR of, 89–96, 90–91f
metastatic lung cancer vs. cardiac, 37–38 normal anatomy of, 99–100f, 99–104
on CT, 646, 646f tracheobronchial, 502, 502f normal diameter of, 100, 101f
on MRI, 649, 650f–651f Aneurysm(s) penetrating ulcer of. See Atherosclerotic
parathyroid, 351, 351f, 353f of aberrent right subclavian artery, 106, ulcer, penetrating
Adhesive atelectasis, 505, 508f, 512 107–108f postoperative, 168–173, 169–174f
Adrenal adenoma aortic. See Aortic aneurysm(s) surgical advances and, 167–185
metastatic lung cancer vs. of pulmonary artery, 271–274, pseudoaneurysm of. See Aortic
on CT, 646, 646f 272–273f pseudoaneurysm
on MRI, 649, 650f–651f Angiography stented, 182–185
Adrenal disease contrast-enhanced magnetic resonance, thoracic
benign, 645f, 646 12–13, 14f surgery of
Adrenal lesions coronary magnetic resonance, 13–14, elephant trunk technique in,
algorithm for evaluating, 646, 646f 15f 176–178, 177–180f
Air trapping CT, 88–90 trauma to, 163–167, 165–167f
in bronchiectasis, 513, 516f coronary, 25–27 clinical features of, 163–165,
in sarcoidosis, 710, 710f multidetector, 222–223, 223f 164f
Aorta (continued) in intramural hematoma, 163, 163f middle lobe, 505f

CT evaluation in, 165–166, MR evaluation of, 130 cicatrization, 507f
165–166f with penetrating ulcer, 151f nonobstructive, 510–512
MR evaluation in, 166–167, 167f Aortic regurgitation, 55–56, 56f causes of, 510t
radiographic evaluation in, 165 Aortic root obstructive (resorption), 508–510
treatment of, 167 aneurysm or dissection of mechanisms and causes of, 508–510,
Aortic aneurysm(s) surgery for, 173–176, 175–176f 509t
infected, 186–190 Aortic stenosis, 55, 55f right upper lobe, 503–505, 503–505f
MR evaluaion of, 130–131, 131f Aortic valve cicatrization, 503, 507–508f
pulmonary artery obstruction in, 278, bicuspid, 61f, 61–62 round, 807–810, 808–809f
279f pathology of, 55–56, 55–56f Atherosclerotic ulcer
ruptured, 132–135, 132–135f, 418, Aortitis penetrating, 146–150f, 148–155
420f clinical features of, 185 clinical features of, 148–149,
contrast CT for assessment of, 133f, CT evaluation of, 185, 186–187f 148–150f
134 infectious, 186–191, 190f CT evaluation of, 150, 151f
missing calcium sign in, 134, 135f MR evaluation of, 185–186, intramural hematoma and, 150, 152f
MRI evaluation of, 135, 136f 187–189f MR evaluation of, 150, 152f
multidetector row helical CT of, 134, Aortopulmonary window natural history and treatment of,
135f CT of, 298f, 299 150–155, 153–155f
thoracic Arrhythmogenic right ventricular car- Atrial septal defect, 62, 63f
clinical features of, 119–123, diomyopathy, 38–40, 39t, 40f Atrium
120–124f Arteries left
Crawford classification of, 123–124, coronary, 23, 23f anatomy of, 74–75, 75–77f
125–127f anomalies of, 70–74, 70–74f Attenuation
imaging features of, 124–129 imaging of, 24–26f, 25–27 decreased
Aortic arch(es) great in diffuse lung disease, 693–697,
anomalies of, 104f, 104–112 complete transposition of, 67–68, 694–696f, 695t
with left-sided arch, 105–106, 67–68f Axilla
105–108f congenitally corrected transposition anatomy of, 843, 844–850f
with right-sided arch, 106–110, of, 69f, 69–70 boundaries of, 843, 843f
108–110f radicular, 130 and chest wall, 843–858
symptomatic, 110–112, 111f pulmonary. See Pulmonary artery(ies) CT of, 851–857f, 851–858
CT of, 297–299, 298f Arteriovenous malformation, 584–585f, landmarks of, 843, 844–850f
CT section through, 195, 195f 585–586 soft tissue masses of, 852–853f
double, 104, 104f, 110, 111f Arteritis structures of, 843, 843f
embryology of, 104, 104f, 109f great cell, 186, 189f Azygoesophageal recess, 298f, 299,
and great veins, 191, 193f Artery of Adamkiewicz, 130 300–301f
interrupted, 112, 112f Asbestos Azygos fissure, 778, 778f
MR of, 296–297f benign pleural disease due to, 803–810, Azygos vein, 191–199, 194f
surgery of, 173–176, 175–176f 804–806f
Aortic coarctation. See Coarctation of CT evaluation of, 806–807, 807f B
aorta malignant mesothelioma and, Bare area sign
Aortic dissection, 135–148 806–807, 817 in peridiaphragmatic fluid localization,
acute Stanford type A, 137f, 137–138 Asbestosis, 701–706, 703f 870, 871–872f
acute Stanford type B, 138, 138f honeycombing in, 701, 706f Bicuspid aortic valve, 61f, 61–62
classification of, 137f, 137–138 HRCT in, 704t, 704–705, 704–706f Biopsy
clinical features of, 135–136 linear opacities in, 704, 704f in diffuse lung disease, 755f, 755–757
clinical presentation and imaging uti- Aspergillosis pleural
lization in, 138–140 airway invasive, 747, 747f CT-guided, 813–815f
CT evaluation of, 140–142, 140–143f allergic bronchopulmonary in solitary pulmonary nodule assess-
angiographic techniques in, bronchiectasis in, 527–529, 528f ment, 607f, 607–610
140–142 angioinvasive, 745–747, 746–747f transbronchial, 607, 607f
differentiation of true versus false fungal, 498 Bird fancier’s lung. See Hypersensitivity
lumen in, 142–145, 144f halo sign in, 689, 689f pneumonitis
results of, 145 invasive, 745–747 Black blood imaging
CT in, 140–142, 140–142f halo sign in, 586–589, 588f in MR angiography, 96
International Registry of Acute Aortic HRCT findings in, 745, 746t Blood vessels
Dissection Study, 138–140, pulmonary nodules in, 689, 689f pulmonary
147–148 Aspergillus bronchiolitis, 747, 747f anatomy of, 677–678, 677–678f
mortality associated with, 147–148 Asthma, 543, 543f Boerhaave syndrome, 427, 427f
MR evaluation of, 145–147, 146–147f Atelectasis Bone lesions
pathophysiology of, 136, 137f adhesive, 505, 508f, 512 of thorax, 855–856f
persistent perigraft flow after, 172, 172f cicatrization, 503, 507–508f, 510–511 Bone marrow transplantation
ruptured, 147, 148f compression (passive), 503–505, pulmonary complications of, 738,
Aortic pseudoaneurysm 507–508f, 511–512 739f
after aortic surgery, 168, 171–172f, 173, left upper lobe, 506f BOOP. See Bronchiolitis obliterans, with
183f lower lobe, 505, 506f organizing pneumonia
Index 887
Bowel disease with poorly defined centrilobular nod- Cancer. See also specific types
inflammatory, 503 ules, 538–539, 539f risk of
Brachiocephalic veins, 191–194f postinfectious, 537f due to radiation, 675
Breast cancer respiratory, 539, 539f Carcinoid tumor(s), 469, 470f, 569–572,
pleural involvement in, 810, 824f smokers’ 571–572f
recurrent interstitial lung disease associated hilar, 572, 573–574f, 578f
axilla and chest wall assessment in, with, 720, 720f thymic, 332–333f
851f with tree-in-bud pattern, 533–538, Carcinoma
Bronchial dilatation 538f bronchoalveolar cell, 622–623, 624f
in bronchiectasis, 513–518, 517f Bronchiolitis obliterans, 696 bronchogenic. See Lung cancer
Bronchial tapering consolidation in, 692–693f dormant scar, 604
lack of HRCT in, 675, 675f esophageal, 413–418, 415–417f, 419f
in bronchiectasis, 515–517f, 518 with intraluminal polyps, 531, 532f of lung. See Lung cancer
Bronchial tumor. See Endobronchial mosaic perfusion in, 696f, 696–697 renal cell
tumor with organizing pneumonia (BOOP), metastatic to diaphragm, 875f
Bronchial wall thickening 536f, 725–727, 726–727f thrombus in vena cava from, 207,
in bronchiectasis, 518–519 CT findings in, 539–540, 540–541f 208f
Bronchiectasis Bronchoalveolar cell carcinoma, 622–623, thyroid
in allergic bronchopulmonary 624f CT evaluation of, 344–346, 346f,
aspergillosis, 527–529, 528f Bronchogenic carcinoma. See Lung cancer 349f
in atypical mycobacterial infection, Bronchogenic cyst tracheal, 491, 492f
525, 525f intrapulmonary, 574, 578f Carcinomatosis
bronchial dilatation in, 513–518, 517t Bronchopleural fistula, 790, 790f, 815f lymphangitic, 706–707, 706–707f
bronchial wall thickening in, 518–519 Bronchopneumonia Cardiac bronchus, 469, 471f
causes of, 512t, 512–513 Aspergillus bronchiolitis and, 747, 747f Cardiac masses, 44–54, 46t
CT technique in, 455–456, 455–456f, Bronchoscopy Cardiac strain
517, 518f in hemoptysis evaluation, 545 right
in cystic fibrosis, 517f, 525–527 in immunocompromised patient CT features of, 242–246, 243–246f
diagnostic criteria for hRCT assessment before, 756 Cardiomyopathy
on CT, 513–523, 517t virtual CT-guided ultrathin, 459, 459f arrhythmogenic right ventricular,
differential diagnosis of, 529, 530f Bronchus(i). See also Airway(s) 38–40, 39t, 40f
extent and severity of, 521–523 anatomic variants of, 467–469, dilated, 36–37
high-resolution CT findings in, 468–471f hypertrophic, 34–36, 35t, 36f
513–517, 514t, 514–515f anatomy of, 460–467, 464t, 677f, iron overload, 38
lack of bronchial tapering in, 515–517f, 677–678 restrictive, 37f, 37–40, 39t, 40f
518 left-sided, 467 Carotid artery(ies), 101, 103f
mucoid impaction in, 513, 516f, right-sided, 463–467 redundant, 103, 103f
519–520, 519–520f, 521–522f bridging, 469, 470f Castleman disease, 308, 384t, 384–385,
in mycobacterial infection, 525, 525f cardiac, 469, 471f 385f
pitfalls in diagnosis of, 523, 524, 524f central tumor of, 472, 475f, 477–478f, Catheter drainage
radiographic findings in, 513 480f percutaneous
rheumatoid arthritis, 723 mainstem for pleural effusion or pneumothorax
traction, 523, 524f, 683, 684f, 700f, and trachea CT and, 800, 800f
710, 711f abnormalities involving, 486t, Centrilobular emphysema, 695, 695f,
Bronchiolar disease 486–503 734–735, 735f
ancillary signs of, 520–521 tuberculosis of, 494 Centrilobular nodules
Bronchiole(s) nomenclature for, 462–463, 463f, 464t differential diagnosis of, 688, 688f
disease of. See Bronchiolitis segmental and subsegmental HRCT appearance of, 687–688f, 688
Bronchiolitis, 529–543 nomenclature and variants of, 464t poorly defined, bronchiolar disease
Aspergillus, 747, 747f tracheal, 469, 470f associated with, 538–539, 539f
cellular, 531, 532t Bulla(ae) Centrilobular region of lung, 681f,
classification of loculated pleural fluid vs., 770, 772f 681–682
clinicopathologic, 529–533 Bullectomy Cerebral perfusion
using CT patterns, 533t, for emphysema retrograde
533–543 CT assessment before, 737f, for aortic surgery, 168
constrictive (obliterative), 531–533, 737–738 Chemodectoma
534–537f, 541–543, 542f Bypass grafts mediastinal, 308, 435f, 435–436, 436t
with decreased attenuation, 541 assessment of, 31–34, 32t, 33f Chest tube(s)
etiologies and clinical conditions asso- for pleural drainage
ciated with, 530, 531t C CT and, 800–801f
follicular, 535f Cabrol procedure venous injury due to, 207, 208f
with ground-glass attenuation or con- with aortic inclusion graft, 172–173 Chest wall, 832–860
solidation, 539–540, 540–541f for aortic root replacement, 175, 175f actinomycosis invading, 851–856,
histologic patterns of, 532f, 533t Calcification 856–857f
infectious pleural, 803, 805f axilla and, 843–858
tree-in-bud pattern in, 533–538, Calcium scoring CT of, 782–784, 783–785f, 851–857f,
538f, 687, 688f coronary artery imaging, 25 851–858
Chest wall (continued) post processing, 6, 7–8f multidetector (MDCT), 87

lipoma of, 772, 773f spatial resolution in, 5–6 of airways, 453–455, 454–455f, 458,
lung cancer invading, 772f temporal resolution in, 5 458f
lymphoma invading, 857f for classification of airway pathology, angiography using, 222–223, 223f
MRI evaluation of, 857f, 858–860 472, 473f multidetector
soft tissue masses of, 852–853f in classification of bronchiolitis, 533t, of pleura, 769
structure of, 781–784, 783f 533–543 multidetector row helical
surfaces of pleura and, 781–784, 783f contrast in. See Contrast of artery of Adamkiewicz, 130
CT appearances of, 782–784, in cystic fibrosis, 526–527 of ruptured aortic aneurysm, 134,
783–785f in differentiation of causes of 135f
Children bronchiectasis, 529, 530f multidetector scanners, 2, 2t, 5, 6, 7–8f
thymus in, 309, 310–312f in diffuse lung disease in myasthenia gravis, 330–331
Chondrosarcoma chest radiography versus, in neoplasms of trachea and mainstem
mediastinal, 412 748–749 bronchi, 491–494, 493–494f
Chordoma diagnostic accuracy of, 750–752 of neuroblastoma, 433–434
mediastinal, 411f, 412 clinical utility of, 748–758 of penetrating atherosclerotic ulcer,
Cicatrization atelectasis, 503, 507–508f of esophagus, 412–428 146–149, 148f, 150f
Cine sequences in carcinoma, 414t, 414–418 of pleural effusions, 786–787f,
in cardiac MRI imaging, 10, 11f in evaluation of metastatic lung cancer, 797–803
Coalworker’s pneumoconiosis, 712–714 644–647, 645–646f of primary aortic intramural
Coarctation of aorta, 64, 65f, 144f in evaluation of nodal disease in lung hematoma, 153–154, 153–155f
clinical features of, 112, 113f cancer, 641–643f, 641–644 pulmonary angiography, 219
CT evaluation of, 117–118f in focal lung disease, 557–559, pitfalls in interpretation of,
MR evaluation of, 114–117, 116–117f 558f 247–258
treatment of, 112–114, 113–116f helical/spiral (volumetric) in pulmonary embolism
Coccidiodomycosis of mediastinum, 291, 292f detection, 229–231, 232f
lymph node enlargement in, 389f, in hemoptysis, 543–546, 544t technique of, 225–232
389–390 high-resolution (HRCT) in ruptured aortic aneurysms, 132–135,
Collagen-vascular diseases of airways, 455–456, 455–456f 132–135f
pulmonary interstitial involvement in, in asbestos-related pleural disease, simplified obstruction index, 246
721–723, 722–723f 806–807, 807f technical parameters in, 2t, 2–3
Colon cancer in bronchiectasis evaluation, thick-section
thymic rebound in, 315, 318f 513–517, 514t, 514–515f, 543 versus HRCT
Compression atelectasis, 503–505, in diffuse lung disease, 672–675, in diffuse lung disease, 749–750,
507–508f, 511–512 672–675f 750f
Computed tomography (CT) etiologies established by, 757t of thoracic aneurysms, 124–130,
in acute pulmonary embolism findings of 127–129f
imaging appearances of, 237–240, to obviate lung biopsy, 757–758 of thymoma, 319–325, 321–326f
238–241f general principles and methodology Computer-assisted diagnosis
in adenocarcinoma of lung, 623f of, 672–675 of lung cancer, 611–612f, 611–613
of airways, 453–459 lung biopsy guided by, 755f, Consolidation
in allergic bronchopulmonary as- 755–757 differential diagnosis of, 692, 693t
pergillosis, 527–528 patterns of abnormalities on, in diffuse lung disease, 691–693,
angiography using, 88–89 682–697 692–693f
in pulmonary embolism detection, of pleura, 782, 784f diseases characterized by, 725–728
220–225 in sarcoidosis, 708t, 708–710, Contrast
of aortic arch, 297–299, 298f 708–711f in CT
of aortic dissection, 140–142, 140–143f in silicosis, 712–714, 713f administration of
in aortic trauma, 165–166, 166–167f versus thick-section CT for cardiac imaging, 6
in aortitis, 185, 186–187f in diffuse lung disease, 749–750, for angiography of aorta, 88–89
of aortopulmonary window, 298f, 299 750f for assessing pleural lesions, 769
for assessment of morphology of lymph low-dose in lung cancer nodal evaluation,
nodes, 354–355f, 359f, 362, 362f in lung cancer screening, 589–593, 590–596f
in asthma, 543, 543f 663–664 in neoplastic airway obstruction,
of axilla and chest wall, 851–857f, in lung cancer 457, 457t
851–858 evaluation of, 589–593, 590–596f for pulmonary embolism detection,
in bronchiolar disease with poorly de- screening for, 660–664 226–227, 227f
fined centrilobular nodules, staging of, 636 in MR angiography, 12–14, 14–15f
538–539, 539f techniques for, 637–648 in MRI
in bronchiolitis obliterans, 540, of mediastinal lymph nodes, 354–356f, in pulmonary nodule evaluation,
541–542f 356–359, 359f, 360t 593–596, 597f
bronchoscopic correlation with of mediastinum, 639–641, Costochondral injury
in central airways, 471–503 639–641f evaluation by MRI, 859f
in cardiac imaging, 2–6 indications for, 290 Crazy-paving appearance
contrast administration in, 6 normal anatomy on, 294–295f, in alveolar proteinosis, 684, 691, 692f
electrocardiogram referencing with, 294–303, 298f, 300–301f Cross-clamping
3f, 3t, 3–5, 4–5f techniques for, 291–293, 292f spinal cord ischemia and, 168
Index 889
Crow’s feet Draped aorta sign Esophagus

round atelectasis and, 808, 809f in ruptured aortic aneurysm, 134, 134f benign tumors of, 419–422, 421f
Cryptogenic organizing pneumonia Drug-induced lung disease, 723–724, carcinoma of, 413–418, 415–417f,
(COP). See Bronchiolitis obliter- 723–724f 419f
ans, with organizing pneumonia staging of, 414t, 414–418
(BOOP) E CT evaluation of, 412–428
CT. See Computed tomography (CT) Ebstein anomaly, 64 dilatation of, 422, 423–424f
Cushing syndrome Echocardiography fibrovascular polyps of, 422
thymic carcinoid tumor in, 332, 333f transesophageal, 87 intrinsic vs. extrinsic lesions of,
Cystic fibrosis Edema 418–419, 420–421f
bronchiectasis in, 517f, 525–527 pulmonary perforation of, 427f, 427–428
CT in, 526–527 crazy-paving pattern in, 691, 691f Extramedullary hematopoiesis, 436–437,
Cystic hygroma, 403–405, 405–406f Electrocardiogram referencing 437–438f
Cystic thymoma, 320, 322–323f MRI and, 8, 8f
Cyst(s) Electrocardiographic triggering F
bronchogenic, 392–394, 393t, in CT angiography, 89 Farmer’s lung. See Hypersensitivity pneu-
393–395f Elephant trunk technique monitis
diseases characterized by, 728–734 in surgery of thoracic aorta, 176–178, Fat
esophageal duplication, 394, 396f 177–180f extrapleural, 781
mediastinal, 392–398, 393t Eloesser window thoracostomy CT appearance of, 782, 784f
differential diagnosis of, 307, 307t CT assessment after, 800, 802f mediastinal masses containing, 307,
neurenteric, 394–395 Embolism 307t, 398–403, 399–402f
pericardial, 42–43, 395, 397f pulmonary. See Pulmonary embolism MR evaluation of, 403
thoracic duct, 396 septic, 747–748, 748f in solitary pulmonary nodule,
thymic, 331f, 334, 334f Emphysema 573–574, 575–576f
Cytomegalovirus (CMV) infection centrilobular, 695, 695f, 734–735, FDG-positron emission tomography
in AIDS, 744–745, 744–745f 735f (FDG-PET)
HRCT findings in, 744t, 744–745, “dirty lung,” 750, 751f in lung cancer
744–745f HRCT in, 673f, 734–736, 735t, for assessing prognosis at diagnosis,
in non-AIDS immunocompromised 735–738f 657
patient, 743–744 panacinar (panlobular), 695, 696f, for assessing response to initial ther-
735–736, 736f apy, 657
D paraseptal, 736, 736f for evaluation, 652–658,
Dermoid cysts Empyema(s) 655–657f
mediastinal. See Teratoma, mediastinal CT characteristics of, 799 for post-treatment evaluation and
Desquamative interstitial pneumonia lung abscess vs., 793–795f restaging, 657f, 657–658
(DIP), 720–721, 721–722f tuberculous, 789–791f for staging, 652–658, 655–657f
ground-glass attenuation in, 690, 690f postoperative CT evaluation of, in malignant pleural disease,
Diaphragm, 860–874 802f 810–812
anatomy of, 860–862, 860–862f Empyema necessitatis, 790, 792f in residual lymphoma after treatment,
CT appearances of, 862–863, 862–863f Endobronchial obstruction 382
defects in, 863–866, 864–866f nontumoral, 509–510, 510t in solitary pulmonary nodule evalua-
hernia of, 863–866, 866f Endobronchial stents, 472, 481–482f, tion, 596–601, 598–601f
MRI evaluation of, 874, 876–877f 486f Fiberoptic bronchoscopy
as pathway for disease spread, 871–874, Endobronchial tumor in lung nodules, 607–608
874–875f carcinoid, 491, 492f Fibroelastoma
and peridiaphragmatic fluid collections, central, 639f–641f, 639–641 papillary, 47, 48f
867–871, 870–873f lobar atelectasis due to, 508 Fibroma, 48–49, 49f
pseudotumors of, 862, 863f metastasis of, 519, 520f Fibromatosis
rupture of, 866–867, 867–870f Endobronchial ultrasound, 484–485, mediastinal, 408
Diaphragm sign 484–485f Fibrosarcoma(s), 51
in peridiaphragmatic fluid localization, Endografts mediastinal, 409–410
870, 871f of aorta Fibrosing mediastinitis
Diffuse lung disease, 671–768. See also spe- endoleaks in, 184f, 185 pulmonary artery obstruction due to,
cific disorders Endomyocardial disease, 38 277–278, 278f
HRCT in, 672–675, 672–675f Eosinophilic pneumonia Fibrous tumor
clinical utility of, 748–758 chronic, 727, 727f of mediastinum
limitations of radiographs in, 671 Esophageal duplication cyst, solitary, 408–409
nodular opacities in, 685–690, 394, 396f pleural
686–689f Esophageal fistula benign, 770, 771f
video-assisted thoracoscopic surgery tuberculous, 418, 421f Fistula
(VATS) in, 755f, 755–756 Esophageal hiatus bronchopleural, 790, 790f, 815f
Diffuse panbronchiolitis, 533–534, 538f disease spread via, 872, 873, 874f esophageal, tuberculous, 418, 421f
Displaced crura sign Esophageal varices, 422–423, 424–425f Fluorodeoxyglucose-positron emission
in peridiaphragmatic fluid localization, Esophagitis, 422, 422f tomography. See FDG-positron
870 Esophagogastrectomy emission tomography
Dormant scar carcinoma, 604 CT evaluation of, 412–413f (FDG-PET)
Focal lung disease, 557–619 H Honeycombing

calcification in, 567–573 Halo sign in asbestosis, 701, 706f
characterization of in invasive aspergillosis, 586–589, 588f, in idiopathic pulmonary fibrosis, 694,
definition of, 561 689, 689f 694f, 700f, 733–734, 733–734f
clinical evaluation of, 561–562 reverse, 540, 541f Hypersensitivity pneumonitis
CT densitometry in evaluation of, 562, Hamartoma, 573, 575–576f acute, 714
562t, 570f endobronchial, 472, 477f, 491 bronchiolar disease in, 535f
CT halo sign in, 586–589, 588f Heart centrilobular nodules in, 687, 687f
definition of, 561 chambers of, 18–21, 20f, 22f chronic, 715, 717, 717f
detection of, 557–561 normal anatomy of, 18–24, 19–20f ground-glass attenuation in, 714–715,
CT in, 557–559, 558–559f valves of, 21–23 716–718f
MRI in, 560–561 Heart disease HRCT in, 714, 715t, 715–718f
feeding vessel sign in, 583–585, congenital, 59–74, 61t Hypertension
583–585f ischemic, 24–34 chronic thromboembolic pulmonary,
location of, 583 valvular, 54–59 258–266, 259–260f
morphologic features of, 583, 589 Hemangioendothelioma main pulmonary artery findings and,
morphological evaluation of, 562–589 epithelioid, 408 263–264, 264f
positive bronchus sign in, 586–588f pleural, 831f parenchymal findings in, 264–265,
size of pulmonary nodule in, 581, 582f Hemangioma, 49–50, 51f 265f
subsolid versus solid lung nodules in, mediastinal, 308, 405–408, 407f Hyperthyroidism
562–567 Hemangiopericytoma, 408 thymic hyperplasia in, 314, 314f
Foreign body Hematogenous metastases Hypogenetic lung (scimitar) syndrome,
endobronchial, 476f to lungs, 714 269, 269f
Fracture Hematoma(s) Hypothermia with circulatory arrest
sternal, 834f, 838f mediastinal, 148f for aortic surgery, 167
Fungal infections. See also specific infections pericardial, 43
lymph node enlargement in, 389f, perigraft, 171f I
389–390 primary aortic intramural, 130, 131f Image reconstruction and interpretation
tracheobronchitis/aspergillosis, 498 aortic pseudoaneurysm in, 163, 163f in pulmonary embolism, 228–229,
clinical features and natural history 229f
G of, 155–156 Immunocompromised patient
Gadolinium-enhanced 3D techniques in CT evaluation of, 156–157, 156–159f acute lung disease in, 738–748
MR angiography, 91–92, 92f MR evaluation of, 157–163, noninfectious complications of,
in aortic dissection, 145, 146f 159–163f 748
breath holding effects in, 181, pathogenesis of, 155 Inclusion graft
182–183f in penetrating atherosclerotic ulcers, for aortic aneurysm of dissection,
in intramural hematoma, 158, 162–163f 150, 153f 168–172, 169f
optimization and timing of, 92–93, Hematopoiesis Infections. See also specific infections
92–93f extramedullary, 436–437, 437–438f diaphragm as pathway for spread of,
technical aspects of, 93–96f, 94–96 Hemiazygos vein, 191f, 191–196, 197f 873, 875f
Ganglia Hemoptysis mediastinal lymph node involvement
sympathetic tumors of, 432–435, 435t bronchoscopy in, 545 in, 388–390
Ganglioneuroblastoma causes of, 543, 544f poststernotomy, 835, 836f
mediastinal, 435 CT evaluation of, 543–546, 544t sternal dehiscence and, 837f
Ganglioneuroma Hemorrhagic pleural effusion, 772, 773f. Infectious bronchiolitis
mediastinal, 435 See also Pleural effusion(s) tree-in-bud pattern in, 687, 688f
Germ cell tumors Hepatocelluar carcinoma Infectious tracheobronchitis, 497f,
mediastinal, 337t, 338f, 403 diaphragm invasion by, 877f 497–498
nonseminomatous, 341t, 341–343, Hernia(s) Inflammatory bowel disease, 503
342–343f containing fat, 399f, 403 Intercostal muscle (innermost)
Goiter diaphragmatic, 863–866, 866f CT appearance of, 785f
anterior mediastinal, 344, 345f hiatal, 423–427, 425–426f, 872 Intercostal veins, 784, 785f
middle mediastinal, 344, 345f lung Interface sign
multi-nodular, 346, 348–349f postoperative, 858f in diffuse lung disease, 683, 683f
posterior mediastinal, 344, 344f Morgagni, 863–866, 866f in peridiaphragmatic fluid localization,
Granulomatosis Hiatal hernia, 423–427, 425–426f, 872 870
Wegener, 500f, 500–501 Hibernoma, 402–403 Interlobar fissures, 772–778, 775–780f
Graves disease, 346–347 Histiocytoma accessory, 778–780f
Great cell arteritis, 186, 189f malignant fibrous, 410 major (oblique), 774, 775f
Ground-glass attenuation Histiocytosis X. See Langerhans cell histio- minor (horizontal), 774–778,
differential diagnosis of, 691t cytosis 775–777f
in diffuse lung disease, 690f, 690–691 Hodgkin lymphoma, 368, 369–371 Interlobular septa, 676f, 678–681, 680f
disease activity and, 752f, 752–755, abnormalities in, 369, 369t thickening of
754f CT in, 369t, 369–370 differential diagnosis of, 683–684,
diseases characterized by, 714–725 lymph node enlargement in, 370–371f 684f
in lung cancer, 655, 663 nodular sclerosing, 374, 375f Interposition graft
in Pneumocystitis pneumonia, 690, 690f thymic involvement in, 334–335f, for aortic aneurysm or dissection,
in sarcoidosis, 754–755f 335–336, 370–371 168–172, 169f
Index 891
Interstitial pneumonia Lung(s) on MRI, 816f

acute abscess of MRI evaluation of, 648–652, 649–653f
consolidation in, 692, 692f CT evaluation of, 793–797, 796–797f with neuroendocrine morphology,
desquamative (DIP), 720–721, empyema vs., 793–795f 625–626
721–722f MR evaluation of, 803f nodal disease in
ground glass attenuation in, 690, attenuation of, 682, 696 CT evaluation of, 641–643f, 641–644
690f biopsy of non-small cell
HRCT in, 721, 721f in diffuse lung disease PET for initial staging of, 652–656,
idiopathic, 697–701 HRCT-guided, 755–757 655–657f
nonspecific, 718–719f, 718–720 HRCT findings to obviate need for, TNM classification of, 626–637
ground-glass attenuation in, 690, 757t, 757–758 pathology of, 622–626
692f, 752f surgical, 756 pleural effusion with, 636, 812–814f,
HRCT in, 718–719f, 718–720 centrilobular region of, 681f, 812–816
Intimointimal intusseption 681–682 positive immunohistochemical results
in aortic dissection, 143f, 144–145 decreased attenuation of for, 622, 623t
Intimomedial rupture sign bronchiolar disease and, 541 positron emission tomography in eval-
in aortic dissection, 144, 144f disease of. See also specific disorders uation of, 652–658, 655–657f
Intracardiac shunts acute preinvasive lesions and, 622
assessment of in immunocompromised patient, radiofrequency ablation of, 660,
MRI in, 18 738–748 661–662f
Intralobular interstitium benign recurrent, 830f
thickening of, 685, 686t cavitation in, 574, 580–581f screening for
Intramural hematoma. See Hematoma(s), chronic infiltrative, 697t, 697–734 CT in, 660–664
primary aortic intramural CT in small cell, 625–626, 626f, 658–659
Iodine clinical utility of, 748–758 squamous cell, 472, 481–482f, 624,
concentrations of cystic 625f
in contrast for CT angiography, 88–89 differential diagnosis of, 695f staging of
Iron overload cardiomyopathy, 38 diffuse. See Diffuse lung disease current controversies in, 631–637
Ischemic heart disease, 24–34 drug-induced, 723–724, 723–724f distant metastasis in, 629–631,
focal. See Focal lung disease; Solitary 630f
J pulmonary nodule groupings in, 631–637, 632f
Joints obstructive International System for Staging
manubriosternal, 833–834f HRCT in, 675, 675f Lung Cancer and Regional Lymph
sternoclavicular, 833–834f pleural disease vs., 770, 771–772f Nodes, 626, 627t
interstitium of issues in, 636–637
K components of, 676, 676f multidisciplinary imaging in,
Kartagener syndrome, 458f postoperative herniation of, 858 637–658
small cell cancer of nodal disease and, 627–629, 628f,
L imaging evaluation in, 658–659 629t
Langerhans cell histiocytosis, 728–731, therapeutic considerations in, 658 primary tumor extent in, 627, 627t
729–731f transplantation of substaging stage 1, 632–634, 634f
HRCT in, 729–730f, 730t obliterative bronchiolitis after, synchronous satellite lesions and,
Laryngotracheobronchial papillomatosis, 541–543, 542f 634–636, 635–636f
491, 493f Lung abscess, 574, 577f superior vena caval thrombosis in, 194,
Leiomyoma, 410–411, 419–421, 421f Lung cancer, 621–669 194f
Leiomyomatosis chest wall invasion by surgical evaluation in, 659–660
esophageal, 421–422 CT evaluation of, 638f, 638–639 treatment of
Leiomyosarcoma, 52, 411 MRI evaluation of, 649–652, CT follow-up of, 647f, 647–648
Leukemia 651–653f Lymph node stations, 355, 357–358f,
mediastinal involvement in, 382–384, dominant scar, 604 360t
382–384f “early hilar,” 636 Lymph nodes
Linear opacities epidemiology of, 621–622 aorticopulmonary, 354, 355f
in diffuse lung disease, 682–685, histologic classes of intercostal, 355
683–685f histologic and cytologic differences lung cancer spread and, 643f,
Lipoblastoma in, 622, 622t 643–644
mediastinal, 402, 402f large cell, 624–625 mediastinal. See Mediastinal lymph
Lipoid pneumonia, 727–728, 728f loculated pleural fluid vs., 793, 795f nodes
Lipoma, 47–48, 48f lymph node enlargement in, 359f, 361, paravertebral, 355, 356f
mediastinal, 400, 402f 361t peribronchial, 354
pleural/chest wall, 772, 773f lymphatic spread of prepericardiac, 353
Lipomatosis patterns of, 643–644 retrocural, 355, 357f
mediastinal, 399–400, 400–401f mediastinal invasion by, 639f, 639–641 retrotracheal, 354, 354f
Liposarcoma metastatic subcarinal, 355, 355f
chest wall, 884f CT evaluation of, 644–647, 645–646f tracheobronchial, 354–355
mediastinal, 400–402, 402f to pleura, 810, 814f Lymphangioleiomyomatosis, 731–733,
Lobule(s) missed, 558, 559f, 611–613, Y–612f 732–733f
of lung computer-assisted diagnosis of, cystic air spaces in, 694, 695f
secondary, 678, 679–681f 611–612f, 611–613 HRCT findings in, 731, 732t
Lymphangioma sagittal, 301–303, 302–306f paracardiac, 353

mediastinal, 308, 403–405, 404–407f techniques for, 293, 294f posterior, 355–357f
Lymphangiomatosis, 405, 407f of neurogenic tumors, 432–434f, 435 prevascular, 352–353, 354f
Lymphangitic carcinomatosis, 685f, of penetrating atherosclerotic ulcer, in sarcoidosis, 386–388, 386–388f,
706–707, 706–707f 150, 152f 387t
Lymphoid follicular hyperplasia of thy- phase contrast, 10–12, 12f in tuberculosis, 388t, 388–389
mus, 314 pleural evaluation using, 857f, Mediastinal pseudocyst, 874f
Lymphoma, 52–53, 53f, 368–382. See also 858–860 Mediastinitis, 428–429, 428–429f
Hodgkin lymphoma of primary aortic intramural fibrosing (granulomatous), 390–392,
anterior lymph nodes in, 352, 354f hematoma, 154–159, 156–159f 390–392f, 391t
chest wall invasion by, 857f in pulmonary embolism, 232–237 MR evaluation of, 391–392f, 392
CT in, 372 fast techniques in, 235–236, 236f Mediastinum, 289–452
cutaneous T-cell, 851f future developments in, 237 chordoma, 411f, 412
lymphoblastic techniques of, 233–235, 234t, CT of
precursor T-cell, 373t, 374, 374f 234–235f indications for, 290
MR diagnosis of, 374–376, 375f in residual lymphoma after treatment, techniques for, 291–293, 292f
non-Hodgkin, 371–373 377–382, 378f, 380–381f cystic lesions of, 392–396, 393–397f
primary mediastinal, 373t, 373–374 respiratory referencing and, 8 differential diagnosis of, 396–398
thymic rebound in, 315, 316–317f of ruptured aortic aneurysm, 135, 136f fatty lesions of, 398–403, 399–402f
pleural, 827–828f, 827–829 technical parameters in, 6–8 MR evaluation of, 339–340f, 403
pleural effusions in, 827 of thymoma, 327–330, 327–330f germ cell tumors of, 336–343, 337t
primary mediastinal large-B-cell, in thyroid disease, 346–347, 347–349f lung cancer invasion of
373–374 tomographic imaging and, 8–10, 9–10f CT evaluation of, 639–641,
residual tumor after treatment of venography using 639–641f
CT evaluation of, 376–377, clinical applications of, 205–207, lymph nodes of. See Mediastinal lymph
376–377f 208f nodes
fluorine-18 fluorodeoxyglucose technical considerations in, 204–205, mass in
positron emission tomography in, 205–207f differential diagnosis on CT,
382 for ventricular analysis, 16–18, 17f, 303–308, 306–308t
MR evaluation of, 377–382, 378f, 17t in pretracheal space, 308
380–381f Malperfusion syndrome MRI of
staging classification for, 369t in aortic dissection, 141–142 indications for, 290
thymic involvement in, 334–336f, Manubriosternal joints, 833–834f techniques for, 293
335–336 Mediastinal growing teratoma syndrome, neurogenic tumors in, 430–435,
Lymphoproliferative disorders. See specific 340f, 341 431–433f
disorders Mediastinal hematoma, 148f MR evaluation of, 432–434f, 435
Mediastinal lipomatosis, 399–400, non-Hodgkin lymphoma in, 373–374
M 400–401f normal anatomy of, 293–303,
Magnetic resonance imaging (MRI) Mediastinal lymph nodes, 352–368 294–296f, 298f, 300–306f
acceleration techniques in, 14–15 abnormalities of on CT, 294–295f, 294–303, 298f,
angiography using, 13–14, 15f, 25–27 CT evaluation of, 359–360 300–301f
of aorta, 90–91 MR evaluation of, 368 on MRI, 294f, 294–303, 296f, 298f,
of heart, 12–14, 14–15f anterior, 352–353, 354f 302–303f
of aorta, 89–96, 90–91f attenuation of, 352–366, 363–364f, paracardiac, 299–301, 300–301f
of aortic arch, 296–297f 364t sagittal and coronal planes of, 301–303,
in aortic trauma, 166–167, 167f calcified, 308, 308t, 362–364, 302–306f
in aortitis, 185–186, 187–189f 363–364f, 364t in thyroid disease, 343–346, 344–345f
cardiac, 6–14 enhancing, 366, 366t, 367f tumors of, 398–403, 399–402f
cardiac acquisition sequences, 8–14 enlargement of, 354f, 356f, 359f, of bone and cartilage, 411–412
first pass perfusion, 12 360–362 of fibroblastic origin, 408–410, 409t
inversion recovery, 14, 16f in Hodgkin lymphoma, 370–371f of muscle origin, 410–411, 411t
motion compensation and, 8, 8f location of, 366–368 of sympathetic ganglia, 432–435
in specific evaluations, 16–18 in fibrosing mediastinitis, 390–392, vascular, 403–408, 404t, 404–407f
technical design for, 6–8 390–392f Melanoma
of costochondral injury and pulmonary groups of, 352–355 metastatic
contusion, 859f in infections, 388–390 pulmonary artery involvement in,
of diaphragm, 874, 876–877f in leukemia, 382–384, 382–384f 276, 276f
in evaluation of lung cancer, 648–652, location of, 352–353 Meningocele
649–653f in lympholiferative disorders, 385 thoracic, 436–437f
in fibrosing mediastinitis, 391–392f, metastases to, 385–386 Mesothelioma, 873
392 morphology of benign fibrous, 770, 771f
in focal lung disease, 560–561 CT assessment of, 354–355f, 359f, malignant, 820–823, 822–823f
image analysis, 98–99, 99f 362, 362f asbestos and, 817
of mediastinum, 294f, 294–303, 296f, necrotic, 364t, 364–366, 365–366f asbestos-related pleural fibrosis vs.,
298f, 302–303f normal 806–807
coronal, 301–303, 302–306f CT appearance of, 354–356f, contrast-enhanced CT evaluation of,
indications for, 290 356–359, 359f, 360t 820f
Index 893
diagnosis of, 817 Myofibroblastic tumor Papilloma

localized, 832 inflammatory, 409, 410f of trachea, 491, 493f
metastatic disease causing, 823–824, Myxoma, 45–47 Papillomatosis
824–826f laryngotracheobronchial, 491, 493f
MRI evaluation of, 819f N Paraganglioma
pleural, 816–824 Needle aspiration/biopsy mediastinal, 308, 435f, 435–436, 436t
round atelectasis and, 810 of solitary pulmonary nodule, 607f, Paraspinal regions
staging of, 818, 820t, 821t 607–610 abnormalities of, 429f, 429–430
three forms of, 817–818 transbronchial (TBNA), 472–484, Parathyroid adenoma, 351, 351f, 353f
Metastasis(es) 483f Parathyroid disease
hematogenous transthoracic, 607–608f, 608–609 CT evaluation of, 350
to lungs, 714 Neoplasms. See specific tumors; Tumor(s) MRI evaluation of, 351f, 351–352, 353f
mediastinal lymph node involvement cardiac Parathyroid glands
in, 385–386 benign, 45–50 ectopic, 349–350, 350t
Middle lobe syndrome malignant, 50–53f Parenchymal disease
chronic, 507f, 511 metastatic, 50 characterization of, 793–797, 796–797f
Miliary tuberculosis, 688, 689f pericardial, 43–44 and disease of pleura
Missing calcium sign Nerve sheath tumors differentiation of, 793–795f
in ruptured aortic aneurysm, 134, 135f mediastinal, 430t, 430–432, 431–432f Patent ductus arteriosus, 64, 65f
Mitral regurgitation, 57, 57f Neurenteric cyst, 394–395 Penetrating ulcer
Mitral stenosis, 56–57 Neurilemoma aortic. See Atherosclerotic ulcer, pene-
Mitral valve, 56–57 mediastinal, 430, 431 trating
Morgagni hernia, 863–866, 866f Neuroblastoma Peribronchovascular interstitium, 683, 684f
Mosaic perfusion mediastinal, 432–434, 434f Pericardial cyst, 395, 397f
in small airways disease, 696, 696f Neuroendocrine tumor Pericardial disease, 40–45
Mounier-Kuhn syndrome, 494–496 thymic, 332–333, 332–333f Pericardial effusion, 41–42
MRI. See Magnetic resonance imaging Neurofibroma Pericardial lesions
(MRI) mediastinal, 430, 431–432 congenital, 44–45, 45f
Mucoid impaction, 574, 577f Neurogenic tumors Pericardial masses, 42–44
in bronchiectasis, 513, 516f, 519–520, mediastinal, 430–435, 431–433f Pericardial recess
519–522f MR evaluation of, 432–434f, 435 superior, 299, 359f
Multidetector row helical computed to- Nitrofurantoin Pericarditis
mography pneumonia induced by, 723f, 724 acute, 42
of artery of Adamkiewicz, 130 Nodular opacities constrictive, 42
of infectious aortitis, 190, 190f in diffuse lung disease, 685–690, Pericardium, 24
of ruptured aortic aneurysm, 133f, 134 686–689f, 707–714, 708–713f Peridiaphragmatic fluid collections,
Multiplanar reconstructions (MPRs) Nodule(s) 867–871, 870–873f
of airway images, 229, 230f, 231 centrilobular Perifissural opacities, 774–778, 777f
of angiographic data differential diagnosis of, 688, 688f Perilymphatic nodules
from MRI, 96 HRCT appearance of, 687, 687f differential diagnosis of, 686, 686t
Multiplanar volume reconstructions perilymphatic Phenytoin
(MPVRs) differential diagnosis of, 686, 686t pneumonia induced by, 724, 724f
of angiographic data, 96–97, 97f in sarcoidosis, 709–710, 709–710f Phrenic nerves
Muscle(s) small inferior pulmonary ligaments and,
intercostal (innermost) with perilymphatic distribution, 778–781, 781–782f
CT appearance of, 785f 686f, 687f Pleura
sternalis, 838f solitary. See Solitary pulmonary nodule adenocarcinoma metastatic to, 810,
subcostal, 785f Non-Hodgkin lymphoma, 371–373 811f, 825f
thoracic, 843, 844–850f primary mediastinal, 373t, 373–374 asbestos-related disease of
Myasthenia gravis benign, 803–810, 804–806f
CT in, 330–331 O CT evaluation of, 806–807, 806–807f
thymic imaging in, 314, 324f, 330–331 Obstructive atelectasis. See under Atelectasis benign fibrous tumors of, 770, 771f,
Mycobacterial infection. See also Oleothorax 831–832, 832f
Tuberculosis CT assessment after, 802f CT evaluation of
atypical Osteodystrophy technique for, 769–772
bronchiectasis in, 525, 525f renal, 835f disease of
Myocardial diseases, 34–40 Osteomyelitis and parenchymal disease
Myocardial imaging sternoclavicular, 837f differentiation of, 793–795f
functional, 27–31, 28f Osteosarcomas, 52 invasion of
Myocardial stress perfusion assessment, by lung cancer, 770, 771f
28–29 P lesions of
Myocardial stress systolic function assess- Panbronchiolitis localization of, 770–772f
ment, 11f, 29–30 diffuse, 533–534, 538f lipoma of, 772, 773f
Myocardial viability assessment, 30f, Pancoast tumor lymphoma of, 827–828f, 827–829
30–31 surgical evaluation of, 659–660 malignant disease of, 810–831
Myocarditis, 40, 41f Pancreatic pseudocyst CT’s role in, 810–812
Myocardium, 24 mediastinal extension of, 437 metastatic, 810, 811f
Pleura (continued) Pneumomediastinum, 487 branching of, 219

MDCT of, 769 Pneumonia congenital anomalies of, 268–271,
MRI evaluation of, 857f, 858–860 chronic eosinophilic, 727, 727f 268–271f
surfaces of cytomegalovirus (CMV) disease of, 219–288
and adjacent chest wall, 781–784, 783f in AIDS, 744f, 744–745 gradation of severity of, 246–258
CT appearances of, 782–784, HRCT findings in, 744f, 744–745 tumor mimicking embolic disease in,
783–785f in non-AIDS immunocompromised 252–254, 253f
tumor(s) of patient, 743–744 malignancy of, 274–277, 274–277f
recurrent, 830f, 831 exogenous lipoid, 573, 576f neoplastic involvement of, 274–277,
Pleural adhesions, 787–788f interstitial. See Interstitial pneumonia 274–277f
Pleural effusion(s) lipoid, 727–728, 728f obstruction of, 277–278, 278–279f
asbestos-related, 806f Pneumocystic carinii. See Pneumocystis due to fibrosing mediastinitis,
clinical characterization of, 784–810 carinii pneumonia 277–278, 278f
CT evaluation of, 786–787f, 797–803 Pneumonitis right
exudative, 787, 799, 804–805 hypersensitivity. See Hypersensitivity congenital absence of, 268, 268f
diagnostic criteria for pneumonitis Pulmonary contusion
on CT, 799 interstitial. See Pulmonary fibrosis evaluation by MRI, 859f
on thoracentesis, 798 radiation Pulmonary edema
hemorrhagic, 772, 773f adhesive atelectasis in, 505, 508f, interlobular septal thickening in, 684f
loculated, 787f 512 Pulmonary embolism
characterisitcs of, 798–799 Pneumothorax(ces), 487 acute
free fluid vs., 786–787f, 797–798 CT evaluation of, 772, 774f CT findings in, 221f, 221–225,
lung cancer mimicking, 793, 795f recurrent 237–240, 238–241f
in lymphoma, 827 bullae in, 738, 738f evaluation of
malignant, 810 Polychondritis clinical, 219, 219t
MRI evaluation of, 858 relapsing, 499, 508f CT angiography in, 220–225
parapneumonic, 787–789, 788–789f Polymyositis/dermatomyositis, 722 biochemical indicators of, 247
CT characteristics of, 798, 798f Polyps chronic, 258–266
percutaneous catheter drainage for fibrovascular CT findings in, 259–266, 259–267f
CT and, 800 of esophagus, 422 mosaic perfusion in, 696, 696f
transudative, 784, 786f, 799 Positive bronchus sign (air bronchogram) preoperative evaluation and postop-
VATS in, 799, 803 solitary pulmonary nodule and, erative changes in, 265–267,
Pleural fibrosis, 789, 789f 586–588f 266–267f
CT evaluation of, 790–792f Positron emission tomography (PET) pulmonary hypertension with,
diffuse See FDG-positron emission tomography 263–265, 264–265f
asbestos-related, 805–806, 806f Postpneumonectomy space, 829–830f, CT in
round atelectasis, 808 829–831 scan protocols for, 225–226, 226t
Pleural plaques Pregnancy diagnosis of
asbestos-related, 803–804, 804–805f pulmonary embolism in, 231–232, computer-aided, 229–231, 232f
Pleural space 233t multiplanar reconstructions in, 229,
CT appearance of, 781–784, 783f Pretracheal space 230–231f
Pleural splenosis, 826–827f masses involving, 308 diagnostic study of
Pleural thickening Progressive systemic sclerosis prospective investigation on, 225
apical, 816f pulmonary interstitial involvement in, exclusion by D-dimer, 220, 220t
diffuse 721, 722f gradation of severity of, 246–258
asbestos-related, 805–806, 806f Prosthetic valves, 59, 60f image reconstruction and interpretation
CT definition of, 806 Proteinosis in, 228–229, 229f
mesothelioma vs., 806–807 alveolar, 684, 724–725 isolated subsegmental, 223–224, 224f
Pleurodesis sequela, 790, 791f ground-glass attenuation in, 724, massive
Pleuroparenchymal disease 725f outcome of
complex Pseudoaneurysm CT prediction of, 240–242, 241t,
CT evaluation of, 790–793, 792f aortic. See Aortic pseudoaneurysm 241–243f
Pleuroparenchymal findings Pseudobronchiectasis, 523 MR in, 232–237
on CT Pseudocoarctation, 114, 115–116f fast techniques in, 235–236, 236f
in acute pulmonary embolism, Pseudocyst future developments in, 237
237–240, 239–241f mediastinal, 874f scan protocols for, 236f, 236–237
Plombage pancreatic techniques of, 233–235, 234t,
CT assessment after, 802f mediastinal extension of, 437 234–235f
Pneumoconiosis Pseudodiaphragm sign, 871 in pregnancy, 231–232, 233t
coalworker’s, 712–713 Pseudodissection ventilation-perfusion imaging in,
Pneumocystis carinii (jeroveci) pneumonia aortic, 143–145, 144–145f 221–222
(PCP) Pseudonodule Pulmonary embolism obstruction index,
in AIDS, 740–743, 740–743f pulmonary, 840–842f 246
ground-glass attenuation in, 690, Pseudotumors, 53–54, 53–54f Pulmonary fibrosis
690f, 755 Pulmonary artery(ies), 24 drug-induced, 723–724, 723–724f
calcified lymph nodes in, 363f, 364 anatomy of, 217–219, 218f idiopathic (IPF), 698–701, 699–702f
HRCT findings in, 741t, 741–743, aneurysm of, 271–274, 272–273f ground glass attenuation in,
741–743f anomalous left, 271, 271f 698–699, 701f
Index 895
honeycombing in, 694, 694f, 700f, Rib(s) fat in, 573–574, 575–576f
733–734, 733–734f anatomy of, 835–836, 839f feeding vessel sign with, 581f,
HRCT in, 699–700, 699–701f trauma to 583–585f, 583–586
linear opacities in, 697–701 CT evaluation of, 836–843, 842f fluid in, 574, 577–578f
mediastinal lymph nodes in, 701 Right pneumonectomy syndrome, 831 growth characteristics of, 601–603f,
Pulmonary hypertension Round atelectasis, 807–810, 808–809f 601–604
chronic thromboembolic, 258–266, CT evaluation of, 808–810, 809f halo sign with, 586–589, 588f
259–260f location of, 583
main pulmonary artery findings and, S management of, 604–610
263–264, 264f Saber-sheath trachea, 496f, 496–497 future directions in, 610–613
parenchymal findings in, 264–265, Sarcoidosis, 472, 482–483f, 707–711 morphologic evaluation of, 589
265f air trapping in, 710, 710f positive bronchus sign with, 586–588f
Pulmonary infarction, 237–240, cardiac, 37, 37f positron emission tomography in eval-
239–241f CT in, 386–388, 386–388f uation of, 596–601, 598–601f
Pulmonary ligament(s), 778–781, ground-glass reversibility in, 754, 754f size of, 581–583, 603f
781–782f HRCT in, 673f, 708t, 708–710, 708–711f Spinal cord ischemia
Pulmonary nodule(s) mediastinal lymph nodes in, 386–388, during aortic surgery, 163–164
pseudonodule, 840–842f 386–388f, 387t Splenosis
solitary. See Solitary pulmonary nodule nodules in, 709–710, 709–710f pleural, 826–827f
Pulmonary regurgitation, 58, 58f pulmonary nodules in, 755, 755f Sputum cytology
Pulmonary sequestration, 117–119, Sarcoma in lung nodules, 607
119f cardiac, 50–52 Squamous cell carcinoma
Pulmonary stenosis, 58 pulmonary artery, 274, 274f of lung, 472, 481–482f, 623, 625f
Pulmonary thromboembolism. See Scan delay of trachea, 472, 474f, 488–489, 489f
Pulmonary embolism for CT angiography, 97, 227–228f Stent graft
Pulmonary valve, 57–58 Scan parameters for aortic aneurysm, 181–185, 184f
Pulmonary veins CT angiography, 88 Stent(s)
congenital anomalies of, 269–271f MR angiography of aorta, 95 endobronchial, 472, 481–482f, 486f
Pulmonary venous connection Scan protocols patency of
anomalous, 62–64, 64f for CT in pulmonary emboli, 225–226, assessment of, 31, 32f
partial anomalous, 199, 201f 226t Sternalis muscle, 838f
Pulmonary vessels for MR in pulmonary emboli, 236f, Sternoclavicular joints, 833–834f
anatomy of, 677–678, 677–678f 236–237 Sternotomy
Scar infection following, 835, 836f
R lung cancer arising in, 604 Sternum, 832–835, 833–835f
Radiation Schwannoma dehiscence of
risk of cancer due to, 675 mediastinal, 430, 431 and infection, 837f
Radiofrequency thermal ablation (RFA) Scleroderma Subcarinal space, 299, 300f
in lung cancer, 660, 661–662f dilated esophagus in, 422, 424f Subclavian artery(ies), 101, 102
Radiography Septic embolus(i), 747–748, 748f artifactual stenosis of
chest Shaded-surface rendering on MR angiography
in diffuse lung disease, 671 of angiographic data, 97–98 due to patient position, 94–95,
versus CT Sialadenoid tumors 95f
diagnostic accuracy of, 750–752, of trachea, 489–491, 490f Subcostal muscle, 785f
751f Silicosis, 711–714, 712–713f Sulfasalazine
in diffuse lung disease, 748–749 HRCT in, 712–713f, 712–714 pneumonia induced by, 724, 724f
Relapsing polychondritis, 499, 499f Small airways disease. See Bronchiolitis Superior sulcus tumor
Renal cell carcinoma Small cell cancer surgical evaluation of, 659–660
metastatic to diaphragm, 875f of lung, 625–626, 626f, 658–659 Superior vena cava
thrombus in vena cava from, 207, imaging evaluation in, 658–659 occlusion of, 853f
208f therapeutic considerations in, 658 Swyer-James syndrome. See Bronchiolitis,
Renal failure Smokers’ bronchiolitis, 720, 720f constrictive (obliterative)
chronic Soft tissues masses
vena cava occlusion in, 853–854f of axilla and chest wall, 852–853f T
Renal osteodystrophy, 835f Solitary pulmonary nodule T-cell lymphoma
Respiratory bronchiolitis interstitial lung biopsy of, 607f, 607–610 cutaneous, 851f
disease, 720, 720f calcification in, 567–573, 569–574f Tagging
Respiratory referencing characteristics of, 561–604 MRI, of heart, 10, 11f
MRI and, 8 clinical evaluation of, 561–562 Takayasu disease, 185–186, 186–188f
Reticular opacities clinical management of Teratoma, 49
in diffuse lung disease, 682–685, factors affecting, 604–605 immature, 337
683–685f, 697–707 contrast enhancement in mature
Rhabdomyoma, 49, 50f on CT, 589–593, 590–596f rupture of, 341
Rhabdomyosarcoma(s), 51, 411 on MRI, 593–596, 597f mature cystic, 337–338, 338–339f
Rheumatoid arthritis definitions of, 561 mediastinal, 337t, 337–341,
pulmonary interstitial involvement in, detection of 338–340f
723, 723f CT in, 589–593, 590–596f malignant, 340f, 341
Rhinoscleroma, 497f, 497–498 MRI in, 593–596, 597f Tetralogy of Fallot, 64–67, 66f
Thoracentesis mass in Traction bronchiectasis, 523, 524f

diagnostic criteria for exudative effusions differential diagnosis of, 313, 313t Transbronchial needle aspiration/biopsy
on, 787–789, 788–789f, 798–799 normal and thymic tumor (TBNA)
Thoracic duct differentiation of, 313, 313t CT-guided, 472–484, 483f
dilatation of, 396, 398f normal appearance of, 308–312, Transthoracic needle aspiration/biopsy
Thoracic duct cyst, 396 310–312f of solitary pulmonary nodule,
Thoracic veins, 191–192f rebound of, induced, 315, 316–318f 607–608f, 608–609
Thoracoplasty Thyroid Transversus thoracis muscle, 784, 785f
CT assessment after, 800, 801f cancer of Trauma
Thoracoscopic surgery CT in, 344–346, 346f aortic, 163–167, 165–167f
video-assisted (VATS) disease of. See Thyroid disease Tree-in-bud pattern
for diffuse lung disease assessment, normal appearance of bronchiolar disease associated with,
755f, 755–756 on MRI, 347f 533–538, 538f
Thoracoscopy Thyroid disease in infectious bronchiolitis, 687, 688f
video-assisted (VATS) CT evaluation of, 343–346, 344–346f in tuberculosis, 533, 534f, 687–688,
for diffuse lung disease assessment, mediastinal involvement by, 343–346 688f
755f, 755–756 CT evaluation of, 344–345f Tricuspid regurgitation, 59
for solitary pulmonary nodule assess- MR evaluation of, 346–347, 347–349f Tricuspid stenosis, 59
ment, 609–610 Tomographic imaging Tricuspid valve, 58–59
Thoracotomy MRI and, 8–10, 9–10f Troponin T, 247
video-assisted (VATS) Trachea Tuberculosis
in pleural effusions, 799, 803 adenoid-cystic carcinoma of, 489–491, airway involvement in
Thromboembolism 490f tree-in-bud pattern in, 687–688,
pulmonary. See Pulmonary embolism anatomy of, 460–463f, 460–470, 470f 688f
Thromboendarterectomy, 265, 266–267f diffuse narrowing of, 496–497f, centrilobular nodules in, 538f
Thrombus(i) 496–498 CT in, 388–389
intracavitary cardiac, 54, 54f diffuse widening of, 494–496 of mainstem bronchus, 494
perigraft disease of mediastinal abscess in, 429f
after aortic inclusion graft placement, classification of, 486t mediastinal lymph nodes in, 388t,
168, 170–171f diffuse, 494–503, 496–502f 388–389
superior vena cava, 198f focal, 487f, 487–488 miliary, 688, 689f
in lung cancer, 199, 202f neoplastic, 488–494, 489–490f, round atelectasis and, 808, 809f
thoracoabdominal aneurysm with, 127, 492–494f of trachea, 494, 495f
129f and mainstem bronchi Tuberculous empyema(s), 789–791f
venous abnormalities involving, 486t, postoperative CT evaluation of,
contrast venography in detection of, 486–503 802f
204, 204f neoplasms of Tumor(s). See also specific tumors
CT signs of, 202, 203–204f CT findings in, 491–494, pleural
Thymic epithelial tumors 493–494f benign fibrous, 770, 771f, 831–832,
classification of, 315–316, 317, 319t neoplasia of, 488–494 832f
Thymic neuroendocrine tumor, 332–333, papilloma of, 491, 493f tracheal, 488–494
332–333f saber-sheath, 496f, 496–497
Thymolipoma, 333f, 333–334 sialadenoid tumors of, 489–491, 490f U
Thymoma, 315–330 squamous cell carcinoma of, 472, 474f, Ulcerative colitis
cervical, 319–320, 322f 488–489, 489f airway involvement in, 503
classification of, 315–317, 320t stenosis of Ulcers
CT of, 319–325, 321–326f following intubation, 488, 489f penetrating aortic. See Atherosclerotic
cystic, 320, 322–323f trauma to, 484f, 487–488 ulcer, penetrating
invasive, 317–318, 320f, 324–325f, 325, tuberculosis of, 494, 495f Ultrasound
329, 329f Tracheal bronchus, 469, 469f endobronchial, 484–485, 484–485f
MRI appearance of, 327–330, 327–330f Tracheal diverticulum, 469, 470f
in myasthenia gravis, 314, 324f, Tracheal stenosis, 488, 489f V
330–331 Tracheitis Valvular heart disease, 54–59
noninvasive, 321–322f tuberculous, 494, 495f Varix(ces)
pleural involvement in, 826f Tracheobronchial lymph nodes, 344–345, esophageal, 422–423, 424–425f
recurrent, 329, 330f 354–355f Vascular abnormalities
staging of, 317–319 Tracheobronchitis CT findings of
Thymus infectious, 497f, 497–498 in acute pulmonary embolism, 237,
carcinoid of, 332–333f Tracheobronchitis/aspergillosis 238–239f
carcinoma of, 331f, 331–332 fungal, 498 peripheral
cyst of, 331f, 334, 334f Tracheobronchomalacia, 462, 462f in chronic pulmonary embolism,
enlargement of, 313–314 Tracheobronchomegaly, 494–496 259–263, 259–263f
in Hodgkin lymphoma, 370–371 Tracheobronchopathia osteochondroplas- Vegetations
hyperplasia of, 314–315f tica, 498f, 498–499 intracardiac, 54
lymphoma involving, 334–336f, Tracheocele, 469, 471f Veins
335–336 Tracheomalacia, 496 coronary, 23–24
Index 897
great Venography Ventricular analysis

intrathoracic, 191f, 191–199 MR MRI in, 16–18, 17f, 17t
pulmonary, 24 clinical applications of, 205–207, Ventricular septal defect, 62, 63f
thoracic, 191–193f 208f Video-assisted thoracoscopy (VATS)
Vena cava technical considerations in, 204–205, for diffuse lung disease assessment,
inferior 205–207f 755f, 755–756
thrombosis of, 198f, 198–199 Venous anomalies for solitary pulmonary nodule assess-
superior acquired, 199–204, 202–204f ment, 609–610
left-sided, 196, 197–199f Venous thrombosis Video-assisted thoracotomy (VATS)
ligation of, 202, 203–204f signs of, 202, 203–204f in pleural effusions, 799, 803
thrombosis of Ventilation/perfusion scans
in metastatic lung cancer, 199, in pulmonary embolism evaluation, W
202f 221–222 Wegener granulomatosis, 500f, 500–501
5636_Naidich_color_insert 12/7/06 6:46 PM Page 1
Color Figure 2-7B Color Figure 2-39E
Color Figure 2-99C Color Figure 2-105B

Color Figure 2-129B
Color Figure 2-105C
Color Figure 2-129A Color Figure 3-12

Color Figure 3-59B Color Figure 3-59D
Color Figure 3-66B
Color Figure 3-73F Color Figure 3-73G

I, J K, L
M, N O, P
Color Figure 5-8I-P
Color Figure 5-9B Color Figure 5-21D
Color Figure 5-23C Color Figure 5-23D

Color Figure 5-27C
Color Figure 5-24C
Color Figure 5-24D

Color Figure 5-30C
Color Figure 5-26I Color Figure 5-30D

Color Figure 5-32C

Color Figure 5-32D
Color Figure 5-33B Color Figure 5-33C

Color Figure 5-34D
Color Figure 5-41C Color Figure 5-42B

Color Figure 5-43D

Color Figure 5-44B
Color Figure 5-70H Color Figure 5-70I
Color Figure 5-77A Color Figure 5-77B

Color Figure 5-77C Color Figure 5-77D
Color Figure 6-6A,B Color Figure 6-6C,D

Color Figure 6-6E,F
Color Figure 6-40I,J
Color Figure 6-43D Color Figure 6-46D

E, F
Color Figure 7-5A

G, H
Color Figure 6-55E-H
Color Figure 7-5B

5636_Naidich_color_insert 1/3/07 11:56 AM Page 11
Color Figure 7-7A
Color Figure 7-5D
Color Figure 7-7D Color Figure 7-7E

Color Figure 7-20B
Color Figure 7-23C
Color Figure 7-34B
Color Figure 7-35B

Color Figure 7-36B
Color Figure 7-37B
Color Figure 7-38B

Color Figure 8-44C

Color Figure 8-44D
Color Figure 8-61B
Color Figure 8-64C

Color Figure 8-108C
Color Figure 9-75C Color Figure 9-75F

Color Figure 9-75I
Color Figure 9-117B

@literaturamedica@ CT Thorax

Uploaded by

Copyright:

Available Formats

@literaturamedica@ CT Thorax

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

@literaturamedica@ CT Thorax

Uploaded by

Copyright:

Available Formats

5636_Naidich_fm_ppi-xiv 12/8/06 11:07 AM Page i

Computed Tomography and

DAVID P. NAIDICH, MD W. RICHARD WEBB, MD

NESTOR L. MÜLLER, MD, PhD IOANNIS VLAHOS, MB, BS, BSC

GLENN A. KRINSKY, MD MONVADI B. SRICHAI, MD

Publisher: Lisa McAllister

© 2007 by LIPPINCOTT WILLIAMS & WILKINS, a Wolters Kluwer business

Third edition, © 1999 Lippincott-Raven Publishers

Printed in the USA

Library of Congress Cataloging-in-Publication Data

To my wife Jocelyn, my son Zachary, and to my father Harry

To the learners and teachers:

To Alison and Phillip Müller

To my wife Jo, my boys Emile and Sebastian, and to my parents

To Stacie, Emma, and Reginald

Preface to the First Edition ix

1 Heart 1 6 Focal Lung Disease 557

Preface to the First Edition

Preface to the Fourth Edition

IMAGING TECHNIQUES 2 Tricuspid Valve 58

PERICARDIAL DISEASE 40 Cardiac disease is one of the leading causes of morbidity

2 Computed Tomography and Magnetic Resonance of the Thorax

IMAGING TECHNIQUES motion. However, variables including collimation, pitch,

for oblique reconstructions without loss of spatial resolu- Systole Diastole

Advantages Pulsed radiation Spiral acquisition

4 Computed Tomography and Magnetic Resonance of the Thorax

Delay Table Delay Table Delay Table Delay

Continuous data acquisition

20% Tube current

6 Computed Tomography and Magnetic Resonance of the Thorax

8 Computed Tomography and Magnetic Resonance of the Thorax

one cardiac cycle to ensure that a complete heart cycle is

10 Computed Tomography and Magnetic Resonance of the Thorax

the achieved contrast between blood and tissue becomes a Tagging

Figure 1-10 Gradient-echo MR

12 Computed Tomography and Magnetic Resonance of the Thorax

14 Computed Tomography and Magnetic Resonance of the Thorax

16 Computed Tomography and Magnetic Resonance of the Thorax

LV Function Quantitative Assessment

End-diastole & End-systole

Figure 1-18 Volumetric acquisition of MR slices for calculation of ventricular volumes.

LV ejection fraction (%) 69  6 69  6 (57–81) 69  6 (57–81)

LV, left ventricular; RV, right ventricular; SD, standard deviation.

18 Computed Tomography and Magnetic Resonance of the Thorax

20 Computed Tomography and Magnetic Resonance of the Thorax

Figure 1-21 A–D: Volume rendered dis-

22 Computed Tomography and Magnetic Resonance of the Thorax

24 Computed Tomography and Magnetic Resonance of the Thorax

26 Computed Tomography and Magnetic Resonance of the Thorax

28 Computed Tomography and Magnetic Resonance of the Thorax

normally enhanced myocardium. After about 1 to 3 min-

Nuclear/ ECG Changes tion or fibrosis can be demonstrated. Unlike standard

30 Computed Tomography and Magnetic Resonance of the Thorax

Figure 1-30 Different MR se-

32 Computed Tomography and Magnetic Resonance of the Thorax

34 Computed Tomography and Magnetic Resonance of the Thorax

Figure 1-33 A–D: Curved multi-

Figure 1-34 CT (A, B) and MR

36 Computed Tomography and Magnetic Resonance of the Thorax

LV ejection fraction (%) 69 6 69 6 (57–81) 69 6 (57–81)

Level 21–40 41–60 60 21–40 41–60 60