Digital Breast Tomosynthesis Imaging:

Jerry A. Thomas, MS, FAAPM, DABR, CHP, DABSNM

8:00 am Registration and Coffee

8:30 am Clinical Need for Digital Breast Tomosynthesis (DBT) Imaging

 Why is tomosynthesis needed?
 Potential benefits of DBT
 ROC studies
 Clinical studies
- Normal
- speculated lesions
- calcifications

9:30 am Technical Approach to Tomosynthesis

 How tomosynthesis images are made

- principles of operation
 2D vs. 3D imaging of the breast
- advantages and difficulties
 Machine design and differences in technical approach
 Image display
- slice vs. slab

10:20 am Informal Discussion and Break

10:30 am Digital Detector Technology

 Types of digital detectors
 How does the detector type affect the approach to DBT?
 System technical details - requirements for DBT
- target
- dose
- exposure time
- compression time

11:15 am Preparing for the Transition from 2D to 3D Imaging of the Breast

11:35 am User Interface for Quality Control

 Navigation through the user interface

 System components
- similarities and differences between 2D and 3D

12:05 noon Lunch - (on your own)

1:05 pm Quality Control Testing in Tomosynthesis

 Technolgist QC
 Medical physicist tests
 Similarities and differences between 2D and 3D

2:35 pm Quality Control of Tomosynthesis Display

 Workstation
 Printer

3:05 pm Informal Discussion and Break

3:15 pm Accreditation of DBT

 Role of the ACR and states
 Role of the FDA
 Training requirements
- Technologists, medical physicists, physicians
 Experience requirements
- Technologists, medical physicists, physicians
3:45 pm Advanced Application in Digital Imaging
 Other areas of research that may improve sensitivity and
specificity
 CAD approaches to 3D imaging
 Role of MRI in breast imaging
 Dual-energy imaging of the breast
 Breast CT
 Radioisotope imaging of the breast
- PET
- SPECT
- single detector

4:30 pm Adjourn
About the Instructor

Jerry A. Thomas is a diagnostic medical physicist at the Via Christi Regional

Medical Center (VCRMC) in Wichita, KS. Jerry holds a Master of Science degree
in Radiation Biophysics from the University of Kansas and is board certified in
Radiological Physics, Health Physics and Nuclear Medicine. He has more than 30
years’ experience in teaching, clinical service and research in many areas of
medical imaging including digital breast tomosynthesis. Jerry also has extensive
experience teaching technologists and students on medical imaging topics. MTMI
is proud to have Jerry serve as an instructor for this important new diagnostic
capability and share his knowledge with you.
 Educational Objectives

 Understand the clinical advantages offered by DBT imaging in screening

and diagnosis of breast cancer.
 Explain how DBT images are made.
 Describe the quality control tests required by the technologist and physicist.
 Understand how DBT imaging impacts accreditation.
 Describe the complementary roles of 2D and 3D breast imaging in detecting
cancer and reducing call backs.
 Understand how DBT compares with other advanced digital imaging
applications.
Copyright © 2015 by Medical Technology Management Institute, Inc.

Copying or duplicating any portion of this material is in violation

of U.S. Copyright Law and is prohibited. Duplication of this work
by any means is prohibited without the prior express written
consent of Medical Technology Management Institute, Inc.
 Clinical Need for Digital
Performance of Digital Mammography
Breast Tomosynthesis
• Annual mammography screening has
Digital Breast Tomosynthesis Imaging demonstrated the ability to significantly reduce the
mortality rate from breast cancer by 15 – 50%
• S/F mammography may have a miss rate of 20 –
14 March 2015 – Indianapolis, IN 30% for breast cancers
• Sensitivity decreases as breast density increases
Jerry A. Thomas, MS, FAAPM, DABR, CHP, DABSNM
Diagnostic Medical Physicist
Via Christi Hospitals Wichita, Wichita, KS

Why is There a Need for Tomosynthesis?

• In 2D FFDM:
Superimposed Tissue Removal using Tomo
Tissue superimposition hides
pathologies in 2D
• ACR Phantom imaged with 4 cm
Tissue superimposition mimics cadaverous breast
pathologies in 2D
• Phantom has low contrast fibers,
masses and calcifications

• Overlying breast tissue obscures

object visibility

Hologic – Proprietary and Confidential

 The Solution is Tomosynthesis
Digital Mammogram & Tomo Images
Tomosynthesis is a three-dimensional
mammographic examination that can minimize the
effects of structure overlap within the breast

PRE‐00023

Breast Tomosynthesis
Hologic Selenia Dimensions DBT
• A 3D screening modality that
preserves the very high
resolution of 2D FFDM • 2D or 3D imaging
– 2D only
• Multiple images of the breast are – 3D only
acquired at different angles
during a sweep of the x-ray tube
• Combo mode
• Allows radiologists to see around – 3D image
overlapping structures – Return to 0 degrees for 2D image
– Single compression for both images
 How Does Hologic’s Tomosynthesis Work? Potential Benefits of 3D Imaging
• Tube moves in a 15 arc
• Better imaging
– Improved lesion margin visibility
•15 low dose images are acquired
– Precise lesion localization
• 1 image at each degree
• Four second sweep – Identification and location of multi-focal cancers
• Total dose ≈ one 2D mammogram

• Images are reconstructed into 1 mm slices • Higher accuracy

– Increased breast cancer detection
•Incombo-mode imaging, 2D and
– Higher PPV for breast biopsy recommendations
3D images are taken under the same
compression, with no additional patient – Decreased workup rate for non-cancer cases
Compression Paddle
positioning required. Combo supports both
CC and MLO projections. Compressed Breast

Detector Housing • Lower recall rates

– Decreased workup rate for non-cancer cases

Overview of Reader Study

ROC Study Design – for FDA PMA
• Comparison of 2D to 2D plus 3D
• 1083 women were recruited from 5 clinical • Two retrospective Independent Reader Studies
centers • Readers were MQSA qualified
– 856 presented for screening mammography – Wide range of experience in 2D
– 227 presented for breast biopsy • Reader study enriched with:
• All subjects received 2D and 3D images of both – Cancer cases
breasts in CC and MLO positions – Recalled screening cases
• Radiation dose for a single 2D plus 3D acquisition – Benign biopsy cases
(either CC or MLO) was less than the MQSA limit • Major conclusions
for a single 2D mammogram – Improved area under ROC curve
– Reduced recall rate
 Rationale for using 2D plus 3D
ROC Curves
• Comparison of current images with prior images is standard
mammography practice and critical to perceive subtle changes Recall Rate = False Positive Fx
which may be associated with a cancer.
High rate of Cancer Detection =
High Recall Rate

• Obtaining a 2D exam along with the 3D exam will allow direct Low rate of Cancer Detection =
comparison of current 2D images with prior 2D images Low Recall Rate

• Segmental and clustered calcifications are more easily and

quickly appreciated with 2D because they can traverse multiple
slices in 3D.

• By minimizing structure overlap, 3D optimally demonstrates

masses and architectural distortion
Hologic – Proprietary and Confidential

How to compare systems with ROC Reader Study #1

• Select a False Positive Value

• Compare True positive Fx for both • Pooled ROC Curves for

systems 12 Readers
• S/F mammo has approximately a
10% false positive rate • Significant increase in
performance
• System 1 True Positive Fx = 55% 60% increased to 78%
• System 2 True Positive Fx = 80%

Hologic – Proprietary and Confidential Hologic – Proprietary and Confidential

 Reader Study 1 & Reader Study 2
Reader Study 2
Pooled ROC Curves
• Pooled ROC Curves for • Pooled ROC Curves for 2
15 Readers Reader Studies

• Identical Results from 2 • Almost complete overlap

independent reader studies between the two studies

• Significant increase in • Significant increase in

performance performance
60% increased to 76% 60% increased to 76%

Hologic – Proprietary and Confidential Hologic – Proprietary and Confidential

Pooled ROC Results

With three points to illustrate trade off between cancer
detection and recall rate changes Reader Study 2: Pooled ROC Curves

Cancer Detection Recall Rate

Reader Study 2 added a
Increase Reduction 3D MLO view
16% 0%
8% ‐25% Increased sensitivity and
0% ‐40%
decreased recall rate

2D plus 3D
2D
AUC diff 0.072, p‐value 0.0001

Hologic – Proprietary and Confidential Hologic – Proprietary and Confidential

 What are the benefits of Combo-mode *

Performance in Dense Breasts • The ROC analysis demonstrated that 2D plus 3D is

Tomo improved ROC performance in fatty breasts
In dense breasts, ROC performance increased 3X that of fatty superior to 2D alone
Conclusion: Tomo useful in fatty breasts, more useful in dense breasts • The ROC results showed that for a given sensitivity,
the recall rate should be lower using tomosynthesis
• The ROC results showed that at a given recall rate,
Fatty – BIRADS
Dense – BIRADS
sensitivity should be higher using tomosynthesis
density 1 or 2
density 3 or 4 • The ROC analysis demonstrated that the
performance of all participating radiologists
improved, regardless of experience
None of these statements could be said for the transition
from Analog to Digital Mammography…
*The Hologic Selenia Dimensions clinical studies presented to the FDA as part of Hologic’s PMA
Hologic – Proprietary and Confidential submission that compared Hologic’s Selenia Dimensions combo‐mode to Hologic 2D FFDM PRE‐00023

Clinical Image Review

Why is tomosynthesis going to revolutionize breast imaging?

Images and data courtesy of:

Better Visualization Example 1
• Hôpital Privé d’Antony, Paris France
• Massachusetts General Hospital, Boston MA USA
• Netherlands Cancer Institute –
Antoni Van Leeuwenhoek Hospital, Amsterdam Holland
• Centre de Radiologie et d’Echographie du Docteur Joussier, Paris France
• Dartmouth Hitchcock Medical Center, Lebanon NH USA
• Magee Women’s Hospital, Pittsburgh PA USA

Slides courtesy of:

• Andy Smith, Ph.D. Hologic, Inc

February 2011
PRE‐00023
 LCC LCC

A suspicious area in a 2D Mammography Image

A 2D Mammography Image

2D 2D
Hologic – Proprietary and Confidential Hologic – Proprietary and Confidential

The 2D Mammography Image next to one slice of a 3D Image Set

Better Visualization Example 2

2D 3D

The Difference is Clear

February 2011
Hologic – Proprietary and Confidential
 2D Mammography Image RMLO The 2D Mammography Image next to one slice of a 3D Image Set
suspicious area
2 – 3 x Magnification

2D 3D
2D
The Difference is Clear
Hologic – Proprietary and Confidential Hologic – Proprietary and Confidential

A 2D Mammography Image with

a suspicious area next to a 3D
RCC RCC
image set

Recall Reduction
Superimposed Tissue Examples

2D 3D 3D Click image to
Hologic – Proprietary and Confidential
run 3D in cine
 Slice 30
26
22
14
18
Recall Reduction – Superimposed Tissue (Case 2)

Stepping thru the image

set, shows that the
suspicious area is nothing
more than normal breast
structures overlapping

Click image to
run 3D in cine
Hologic – Proprietary and Confidential Hologic – Proprietary and Confidential

Invasive Ductal Carcinoma (IDC) IDC - Region of Interest

2D 3D
Click image to Click image to
Hologic – Proprietary and Confidential
run 3D in cine Hologic – Proprietary and Confidential
run 3D in cine
 Tubulolobular Adenocarcinoma Region of Interest

Click image to 2D 3D
run 3D in cine
Click image to
run 3D in cine
Hologic – Proprietary and Confidential Hologic – Proprietary and Confidential

Screening vs DBT in a
57 y/o woman Micropapillary type ductal
carcinoma in situ in 65 y/o woman

Adjacent ductal extension

39
Movie 3
Park J M et al. Radiographics 2007;27:S231-S240 Movie 1 Park J M et al. Radiographics 2007;27:S231-S240
 Metastasis from endometriod
carcinoma in 59 y/o woman

DBT
Artifacts
3 primary masses Mass not seen
(see arrows) on FFDM

Park J M et al. Radiographics 2007;27:S231-S240 Movie 4

Artifacts due to large Calcification

Large Calcification Artifact
Artifact from large
calcification
(white arrow)

On basis of US
appearance Mass
Dx as Cyst

44
Park J M et al. Radiographics 2007;27:S231-S240 Movie 7 Wu, T. et al. Med Phys 33 (7), Jul 2006
 Small Calcification Artifact Needle Artifact

45 46
Wu, T. et al. Med Phys 33 (7), Jul 2006 Wu, T. et al. Med Phys 33 (7), Jul 2006

Questions? 

Talk 2
 Technical Approach to The technology behind
Tomosynthesis tomosynthesis
• DBT
Digital Breast Tomosynthesis Imaging
– Complex imaging technique
• Uses computer algorithms and digital images to
create a 3D image
14 March 2015 – Indianapolis, IN • Underlying technologies
– Digital detectors
Jerry A. Thomas, MS, FAAPM, DABR, CHP, DABSNM – X-ray unit
Diagnostic Medical Physicist
Via Christi Hospitals Wichita, Wichita, KS
– Reconstruction algorithms
– Image Display 2

Types of Detectors X-ray Unit

• Film-screen • Tube motion
• CCD – Linear
• Computed Radiography (CR) – Circular
– Elliptic
• FFDM
– aSi • Arc of tube travel
– CsI – Wide angle – thinner slice

• Solid state
3 4
 X-ray tube travel Ends of Tube Travel
+ 7.5 degrees - 7.5 degrees
Motion
 Continuous
 Step-and-shoot

Angle of Travel
 15°-- ± 7.5°
 40° -- ± 20°

# of Projections
 15 to 25

5
Park J M et al. Radiographics 2007;27:S231-S240 Park J M et al. Radiographics 2007;27:S231-S240

Schematic of Tomosynthesis Projection of Objects in Breast

7 8
Park J M et al. Radiographics 2007;27:S231-S240 Park J M et al. Radiographics 2007;27:S231-S240
 Basic Principle of Basic Principle of
Slice Recon Projection 1 Slice Recon Shift Right

Shift each projection left or Shift each projection left or

right then add to get the plane right then add to get the plane

Projection 2 No Shift

Projection 3 Shift Left

Add shifted images to

get final slice image
9 10

Basic Principle of
Slice Recon Shift Left

Shift each projection left or

right then add to get the plane
Shift and Add
No Shift

Shift Right

Add shifted images to

get final slice image
11 12
Park J M et al. Radiographics 2007;27:S231-S240
 Tomosynthesis Acquisition
Sweep angle
Parameters
• Sweep Angle • The total arc about the center of the detector
• Sweep Direction • In DBT sweep angle ranges from 15° to 50°
• Patient Barrier-Object Distance – Also called scan angle
15° to 50°

• Number of Projections and Projection

Density
• Total Radiation Dose

13 14

DBT Scan Angles and Projections DBT Detector Motions

• Literature reports scan angles from 15° to 50° • Motions
– As scan angle increases depth resolution and – Stationary
slice thickness – Tilting
• Breast tissue further from detector
– Have increased magnification
• 11 – 15 projections
– Early research DBT systems used up to 21
projections • Grids have not been developed for use with
DBT
15
 Sweep Direction DBT Scanning motions
• The direction of x-ray tube movement • X-ray tube
relative to the object or body part of interest – Step-and-shoot
during a sweep. • Takes longer
– Can change by patient positioning • Increased motion artifacts
• E.g. CC vs. MLO views
– Continuous short-burst exposures
• Exposure times need to be short enough to avoid
focal spot motion
• Focal spot motion can cause image blurring
18

Number of Projections and

Patient Barrier-Object Distance Projection Density
• Number of projections = number of x-ray
• The minimum distance between the surface exposures acquired during a single sweep.
of the patient barrier and the object of
• Acquire at constant angle intervals.
interest
– Early DBT systems used 15, 21 and 40 projections.
• Can not be freely varied by the user
– Current Hologic system has 15 projections
– Non-FDA approved systems have varied numbers of
projections
• Projection density = # projections/sweep angle
 Total Radiation Dose Potential Artifacts
• Cumulative sum of the doses for all • Blurring – Ripple
projections • Ghost Artifact – Distortion
• Poor Spatial Resolution
• Image Noise
• Metallic Artifact

22

Blurring - Ripple
Ghost Artifact - Distortion
• Blurring occurs along the sweep direction
– Results from imaging structures that exist out-side • Caused by high-contrast structures outside
the section plane the section plane whose long axis is parallel
– Most prominent for high-contrast objects that are to the sweep direction
perpendicular to the sweep direction – Called “parasitic streak” in conventional linear
tomography
• Ripple
– Due to limited number of projections
– Caused by a high-contrast source far outside imaging
plane
 Poor Spatial Resolution DBT Reconstruction Algorithms
• In-section resolution • Filtered back projection (FBP)
– Depends on the resolution of the detector • Maximum likelihood (ML)
– Independent of acquisition parameters • Simultaneous algebraic reconstruction
technique (SART)
• Depth resolution • Gaussian frequency blending (GFB)
– Affected by sweep angle
• Voting strategy
• Section thickness decreases and depth resolution
increases as sweep angle increases.
26

Filtered back projection (FBP) Maximum likelihood (ML)

• Widely used in CT • An iterative reconstruction method
– Process involves – Has shown better results in some studies than FBP
• Generating a sinogram – Requires more computer memory and time
• Back-projecting the data
• A time intensive reconstructive method
• Filtering the data to compensate for blurring

27 28
 Simultaneous algebraic
Gaussian frequency blending (GFB)
reconstruction technique (SART)
• A refined version of the iterative method of • Combines matrix inversion tomosynthesis
algebraic reconstruction. with FBP
• More effective than FBP at making tissue • Improved reconstruction of micro
conspicuous calcifications
• Better Calcification enhancement
• Better mass margin enhancement
• FBP superior in noise reduction and CNR
29 30

Voting strategy Image Display

• Used to reduce artifacts • Single Slice
• Identifies projection that would introduce • Slab Recon
artifacts into a voxel – Arithmetic
• Rejects the slice when reconstructing the – Geometric
voxel. – Cubic
– Other?
• MIP

31
 MIP of slices showing MIP of slices showing speculated
Calcifications lesions

Tomosynthesis in Breast Imaging -

Better Sensitivity
Potential Uses
• ACR Phantom imaged with 4 cm
• Detection / screening - especially women with cadaverous breast
dense breasts
• Phantom has low contrast fibers,
• Characterization / problem solving masses, and calcifications
• Multi-focal cancers • Overlying breast tissue obscures
object visibility
• Potential for reduced compression

Reference: Andy Smith, “Overview of Breast

Tomosynthesis”, Hologic
 Better Sensitivity Lower Dose
Digital Mammogram 1X dose Tomosynthesis 1X dose Digital Mammogram 4X dose Tomosynthesis 0.5X dose

Slice at plane of phantom insert

Slice at plane of phantom insert
Tomosynthesis shows improved low contrast Tomosynthesis shows improved low contrast visibility
visibility over digital mammography over FFDM, even at much lower dose

Digital Breast Tomosynthesis (DBT)

— Visualization
A DBT reconstruction
- 30-80 slices parallel to the
detector plane
- 1mm slice thickness
- 100 µm in-plane pixel size

Visualization software functions

- Paging through DBT slices
- Window level
- Zoom in / zoom out
- Field of view magnifier

The knowledge for interpreting

conventional mammography is
valid for DBT, no learning curve!

D. Albagli, et. Al., SPIE 2005

Courtesy of Tao Wu, Ph. D and Massachusetts General Hospital

 Lag Tomosynthesis Status
Definition:
• Hologic tomosynthesis FDA approved
Residual signal measured on subsequent images
• Tomosynthesis considered a new modality by FDA
• Impact of high Lag:
• New modalities require 8 hours of training prior to
1. Reduces gain of detector doing unit surveys.
2. Can cause artifacts due to variation in gain
………...(due to temperature or dose level) • Other Companies working on DBT
3. Increases minimum time between exposures • GE
………..(impacts tomosynthesis, dual energy subtraction) • Siemens
• Controlled by Time Constant of pixel readout • Philips (Sectra)
• Planmed
•Giotto

Hologic Tomosynthesis Unit GE Tomosynthesis

• 15 projections over 15 degrees continuous arc – 3.7 ms scan
• Collect 9 projections over 25 degrees
• Rule of thumb is one projection per degree
• Use step and shoot exposures
• Reduces artifacts from calcification
• Reconstruct 0.5 to 1.0 mm slices
• Provides better visualization of spiculations and masses
• Bin them into 1cm slabs (overlap 0.5 cm)
• During projections, detector moves 5 degrees about CR
• Goal to do screening with MLO view only
• Angulation corrected in reconstruction
– Use dose equivalent to CC or MLO
• Using back projection reconstruction because it’s faster
• Preparing to submit for FDA approval
• Most important are reconstruction filters
• Pixels binned to 140 microns/recon to 95
• Slice thickness of 1mm 44
43
 Siemens Tomosynthesis Philips (Sectra) Tomosynthesis
• Collect 25 projections over 50 degrees One single scan with continuous read-out of the detector to obtain 3D data
Each detector line will obtain data from a different angle
• Use step and shoot exposures
• 85 um aSe detector • Photon Counting Detector
• Reconstruct to 1 mm slices
• Move axis of rotation pivot
point under detector for 3D

• Image is captured in a single

scan / projection

45
46

Giotto DBT Projections

Questions? 

47
 Break Time

10 minute break

Talk 3
Digital Detector Technology Acknowledgement
• Original talk content developed by
Digital Breast Tomosynthesis Imaging
Dr. J. Ed Barns, Executive Director MTMI

14 March 2015 – Indianapolis, IN

Jerry A. Thomas, MS, FAAPM, DABR, CHP, DABSNM

Diagnostic Medical Physicist
Via Christi Hospitals Wichita, Wichita, KS

Outline Detectors used in Breast Imaging

• Screen Film
• Review of detector technologies
• Computed Radiography
• FDA approved – Single sided plate
• Non-FDA approved – “Dual” sided plate
• CCD (charge coupled device)
• Advanced applications of FFDM
• FFDM
– aSi – CsI
– aSi – aSe
• Solid-state Photon Counting
 Detector Technology for FDA Centre for Evidence-based
Approved Systems Purchasing
Sensor Readout Company • Part of the National Health Service in the UK
Flat Panels • Evaluate technical performance of radiological equipment.
CsI Phosphor TFT Array GE • Reports on Hologic, Fuji, Sectra and Comparison of CR
a-Se plates TFT Array Hologic, Siemens venders available online

CR Cassette/Reader • Formerly called KCARE (Kings Centre for the Assessment

of Radiological Equipment)
PSP plates Laser scanned Fuji, Carestream
• Website: www.pasa.nhs.uk/PASAweb/
Slot Scanning
Silicon microstrip Photon Counting Sectra

FCM
Fuji Mammo CR
• Pixel size 50 micron
• Use dual sided reading to DQE
• Imaging Plate has thicker phosphor layer
• Results in 40-50% increase in NEQ (DQE)
• Post processing enhances image
(e.g. filtering to recognize and enhance calcifications)
 FUJI FCR 5000: MTF FUJI FCR 5000: DQE

From Centre for Evidence-based Purchasing Report 04094, October 2004 From Centre for Evidence-based Purchasing Report 04094, October 2004

FUJI FCR 5000: DQE vs Exposure Fuji System

Components

Breast Imaging Review Station

ClearView-CSm ClearView-1m Flash Plus IIPm

From Centre for Evidence-based Purchasing Report 04094, October 2004
 PEM PEM (Pattern Enhancement for Mammography)
Non-Calcification Edge

 Effect of PEM Edge

Detection
Increases visibility of Processing
microcalcifications through the use of  Selective edge

Combination
Processing
enhancement of
selective edge-enhancement microcalcifications

Calcification
Position
Detection Calcification
Original Processing
Image Edge
Information

Detection of Center
of Calcification Image

Location of Microcalcifications by PEM process Other CR Mammography

Venders
• Agfa
• Carestream (Kodak)
• Konica
• Canon
 Carestream Health (Kodak)
Agfa CR Mammography
Mammography CR
• Single sided plate, flexible • Single sided plate
• BaSrFBrI:Eu phosphor • Using thicker phospor (increases DQE, reduces
resolution)
• Smaller grain powdered phosphor
• Phosphor grain size uniform
• 50 micron pixel • Using a sealant
• Scan slower than conventional CR • Smoothing to reduce noise and make
calcifications more conspicuous
• Wider light collection guide
• 48.5 micron laser with slower scan time
• Thinner plate to increase resolution

Konica CR Mammography
CR Mammography Systems
• Use CsBr plate
• Columnar crystal structure which collimates the for AGFA, FUJI, KODAK, KONICA
light
• Scan with 43.75 u laser “ All measurements were equivalent to
or exceeded the performance of a typical
• Scan slower than regular plate
modern film/screen system”
• Use smoothing algorithm to reduce noise
• Motor accuracy is limiting parameter for resolution

From Centre for Evidence-based Purchasing Report 06047, October 2006

 CR – Usefulness in DBT FFDM Detectors
• Limited resolution Indirect Detector Direct Detector
– aSi – CsI – aSi – aSe
• Limited DQE
– 3 sources of noise – 2 sources of noise
– Low lag – Moderately low lag
– Little ghosting – Little ghosting
• Slow read-out
– Fast read-out – Fast read-out

NOT SATISFACTORY FOR DBT IMAGING

The Senographe Flat Panel Detector Scintillator & Reflector

Contact Leads Amorphous
for Read-Out Contact Fingers Silicon Array
Electronics X-Ray Photons Reflector

300 to 400 µ CsI

Glass Light photons
Substrate
Scintillator
guided via CsI
(CsI)
needle
structure
Graphite Cover
Scintillator
Electronics / Seal

aSi Array Glass

Substrate

10 needle width

 GE FFDM
Detector Electronics & Packaging
preamp preamp preamp preamp

Photo D
Mammo Panel with Bonded Scan & Data Modules
TFT Data Data
Modules Modules
Scintillator
Drivers

Photo D/TFT

preamp preamp preamp preamp

GE FFDM Product Line Impact of image processing

Unprocessed Thickness Premium
Image Equalization View
RAW TE PV

2000D Seno DS (19x23) Seno DS Interventional Seno Essential (24x31)

9M$ - 30units
33%
100%

Seno Advantage 2 with Premium View 1st intention

Premium View improves conspicuity and productivity
 Fine View
GE Senographe Essential, LFOV
Without Fine View
With Fine View
Generation-3 (Introduced 2004)
• FOV = 30.7 x 24 cm
Fine View
applied to PV
image • Matrix = 3070 x 2400
improves
sharpness and
conspicuity
• Pixel size 100 microns
• 4 times in dynamic range
• CsI thickness increased
• DQE to 70%
• Lag <1%

Data line

Gate Line

FET Negative HV bias

Photodiode

IP CP

Cstorage

D.Albagli, et. Al., SPIE 2005

 D.Albagli, et. Al., SPIE 2005

D.Albagli, et. Al., SPIE 2005

a-Selenium Detector Operation

a-Se Detector Diagram
X-rays

HV
Biasing
Electrode
Scan
Electric field
Line blocking
layer

selenium
Data blocking
Line layer
Selenium
Structure
Pixel thickness
TFT Collection
Active Matrix Electrode
Array
Assuming perfect material, no trapping of charge in selenium facilitates rapid imaging
 Principle of Blocking Layers High Efficiency of a-Se
Metal
electrode
n-layer •Selenium provides almost 100% charge collection efficiency due to
electrostatic field effects in the photoconductor
•Almost complete absorption of the x-rays is achieved within ~200µm layers

selenium
•Selenium allows high fill factor for small pixel size required in
mammography

p-layer
TFT
Substrate

N-layer: one way charge valve. Conducts electrons but traps holes
P-layer: one way charge valve. Conducts holes but traps electrons

Hologic a-Se Detector

-HV top bias electrode
•Hologic reports ghosting of <1%
a-Se
after 3 minutes
+
-
trapped • Detector reconditioned for 50 sec.
image
charge
with bias reversal and light flashes to
- - - - - - - - - - - - - - - -
Blocking layer
- - - - - -
reduce residual trapped charges.

Pixel electrodes
 Data line
Hologic DQE Curves
Gate Line

FET Negative HV bias

a-Selenium Plate

Ise Cse

Cstorage

Z. Jing, Hologic

a-Se Flat Panel Detector MTF of a-Se detector

Detective Quantum Efficiency vs. Exposure
Lazzari,et.al SPIE 2003
 Hologic “Next Generation” Detector Amorphous-Selenium Detector

• Faster electronics-full erasure at 3-4 fps

• Reduced lag- no ghosting
• Higher DQE at low exposures
• Improved blocking layers
• Actively cooled/ Copper heat sink
Detector images to chest wall edge
• Detector reliability improved
• Expect mean time between failures to double

a-Selenium Detector

• Hologic product: 24x29cm

The Product
* 70 micron pixel size
• Anrad Detector
* 85 micron pixel size
• OEM by:
• Siemens (Inspiration) -- Novation used Hologic
• Agfa (OEM entire unit from Siemens – in Europe)
• Planmed
•Giotti
 Siemens Mammomat Novation DR

Status of Siemens Novation Siemens Mammomat Inspiration

Novation Units
• Have stopped selling new Novation units
• Will continue servicing units and replacing detectors
• Selling refurbished Novation units
• Detectors are original design Hologic detectors Future Advanced
Applications

Developed new unit called Inspiration -25° +25°

• New unit uses Anrad detector
Stereotactic
• Received FDA Approval in Feb 2011
• Capable of tomosynthesis
52
 Siemens Digital
Positioning Small Breasts on Large Field Detectors
Mammography Unit
• Mammomat Inspiration - Anrad detector Hologic & GE 30

- Capable of taking images every 25 seconds

- Planning relative dose index to control dose creep
24
• Small fields are centered on detector
- Requires alternative positioning
30
• Provide Tungsten target option Siemens
- In addition to Mo target
- Tungsten spectrum increases DQE 24
- Allows lower dose with same image quality
– Small focal spots of 0.1 and 0.3 mm

Philips FFDM Unit

Giotto FFDM Unit

• Purchased Sectra in Sep 2011

• -- Not known if Philips will continue • Using Anrad
development of “own” system or will move detector
forward with the Sectra unit as their FFDM
 Planmed Nuance FFDM Unit

PHOTON COUNTING

Sectra MicroDose Mammography Sectra Photon Counting Mammography

Mammography stand
• Five programmable stops for individual projection
settings Ergonomic design – Adjustment buttons in
several places
• Free positioning
• Field of View - 24 x 26 cm
• Time to display image < 20 seconds
• Scan time 4-15 seconds
Detector Technology - Photon Counting Sectra Silicon Detector Geometry
• Dead volume layer on top
• 20 parallel silicon microstrips X-rays
• Detector several cm long
• 50 micron pixels
• Pre-collimator defines pixel in scan • Detection rate 5 x 106 x-rays
direction /sec/pixel
• Sample data every 50 microns of scan Dead area
motion • Slit-scanning, 5 sec. scan time
• Tungsten target tube operated at about • 3 mm deep – 95% absorption
30kVp
to 20 kev
• 5 second scan time – varies with breast Active Detector 8 degrees
size Volume • Counts individual photons
-digitized upon detection
Dead area
-suppresses noise

Detector Technology - Photon Counting Detector Technology - Photon Counting

Detector Technology - Photon Counting Detector Technology - Photon Counting

Detector Technology - Photon Counting To Count Photons – What does it mean?

Energy Result:
X-ray photon
- Number of X-rays

Threshold 1 - Pulse height give the

color for each X-ray
Threshold 2
Threshold 3 - No electronic noise
(regardless of pixel
size and dose!)
Electronic noise
Time
1 2 3
 Dose from Sectra FFDM Unit SECTRA FFDM Unit
Minimal Dose Standard Dose
Configuration Configuration Detector Array Scan Direction
mrads (1) 43 85 Direction

kVp 29 32 Peak DQE 62.9 % 60.6%

mAs 20 23.5 (detector)
Processing clinical clinical
50% MTF 6.2 lp/mm 3.3 lp/mm

(1) Mean glandular dose for TORMAX phantom with W/Al

From Centre for Evidence-based Purchasing Report 06045, August 2006 From Centre for Evidence-based Purchasing Report 06045, August 2006

Sectra FFDM Unit: MTF Sectra FFDM unit: DQE

From Centre for Evidence-based Purchasing Report 06045 August 2006 From Centre for Evidence-based Purchasing Report 06045 August 2006
 Sectra DQE Dose Dependence
Other Detectors – not discussed
• Xcounter
– Gas filled Krypton Detector Not FDA Approved
– Have developed a 3D prototype

• Fuji Aspire HD
– Dual aSe detector
– Readout – optical switching

• DASALA
– CMOS detector coupled to CsI
– 39 micro pixel Not FDA Approved
– Partnering with Giotto
From Centre for Evidence-based Purchasing Report 06045 August 2006 – 30 frames/second readout

Acknowledgements
Summary Thanks for input from:
• GE Healthcare
• Properties of a detector useful for DBT • Fujifilm Medical Systems
– Fast readout • Hologic/LORAD
• Planmed
– Low noise
• Sectra Imtec AB
– Low lag • Siemens Medical Solutions
– High DQE • XCounter
• Tao Wu, Ph.D.
– Niquist frequency: 5 lp/mm or greater • Ed Barns, PhD
• Mats Danielsson, Ph.D.
 Questions? 

Talk 4
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

ASSESSMENT OF System Description and

User Interface
TOMOSYNTHESIS
SYSTEMS AND USER Selenia® Dimensions® Digital Breast Tomosynthesis Systems

INTERFACE
1 2
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SSYYSSTTEEMM DDEESSCCRRI PI PTTI O

I ONN AANNDD UUSSEERR I N
I NTTEERRFFAACCEE

Agenda
System Description
• Go over the features of the Selenia
Dimensions DBT system
• Walk through the Selenia Dimensions user
interface

3 4
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

System Configurations Image Acquisition Modes

2

A user may purchase a

Selenia Dimensions
system for both FFDM
A Selenia Dimensions and tomosynthesis
FFDM system may be imaging
upgraded in the field via Conv & Tomo
software to add
tomosynthesis imaging under same compression
(Combo)
Today, a user may 1 3
already own a Selenia
Dimensions FFDM
system
Conventional Only Tomosynthesis Only
5 6
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

AWS Configurations The Differences

Selenia Dimensions Selenia Dimensions
with AWS-5000 with AWS-8000
• Standard Acquisition Workstation • Premium Acquisition Workstation
• 2MP freestanding color display • 3MP medical grade DICOM display on
(upgrades available) articulating arm
• Operator interfaces – mouse, • Operator interfaces – touch screen
keyboard, control panel, trackball, touch pad,
– Optional barcode scanner automated barcode scanner
• Manual login • Biometric login
• 4 standard compression paddles • 8 standard compression paddles
• Optional diagnostic paddles kit • MPPS included in DICOM Services
• Optional MPPS license • Bi-directional communications included
With AWS-8000 With AWS-5000
(premium version) (standard version) • Optional bi-directional communications
license

7 8
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SSYYSSTTEEMM DDEESSCCRRI PI PTTI O
I ONN AANNDD UUSSEERR I N
I NTTEERRFFAACCEE SYSTEM DESCRIPTION AND USER INTERFACE

X-Ray Generation
System Components

X-Ray Tube • Tungsten (W) Anode

• LFS: 0.3 mm; SMF: 0.1 mm

• Conv: 50 μm Rh; 50 μm Ag
X-Ray Filtration • Tomo: 0.7 mm Al

• 200 mA max for LFS

X-Ray • 50 mA max for SFS
Generator • Max mA varies with kVp

9 10
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

Digital Detector
Detector Field-Of-View
• Electronic cropping to
– 18 cm x 24 cm or
– 24 cm x 29 cm
• Smaller FOV dynamically moves to left,
TFT based amorphous Retracting HTC Grid center and right side of the full detector
selenium detector (aSe) • Automatic grid control • Detector cropping depends on position of
• Similar technology to • Grid retracts for
Selenia detector (70 μm • tomo the attached shifting paddle
pixel pitch) • Conventional mags
• Optimized for fast image • Grid in place for
acquisition (4 fps) conventional contact
• Active temperature control imaging
11 12
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

AEC Position Selection Automatic Collimation

• X-ray collimation
• Automatic Position automatically adjusts
– AEC sensitive area is a 5.0 cm x 14 cm area based on
– Always centered to the field of view (shifts with the paddle) – Inserted compression paddle
– Uses two 1 cm x 1 cm sensing areas – Compression paddle position
• Two sensing areas select the two densest part of the breast • Automatic sizes
• Final target exposure is computed by balancing these two – 18 cm x 24 cm
areas – 18 cm x 29 cm (for DBT)
• Manual Position – 24 cm x 29 cm
– Seven positions available to the technologist • Other available sizes
– Always centered to the field of view (shifts with the paddle) – 15 cm x15 cm
– Typically used during implant imaging or to avoid – 10 cm x10 cm
triggering on foreign objects such as pacemakers or wires – 7.5 cm x 8.0 cm

13 14
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

Compression Paddles Compression Paddles

Automatically shifts to Position over-ride

next view in pre- function bypasses
configured workflow steps automation
Standard Compression Mode FAST Compression Mode

15 16
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

Paddle Installation, Removal Screening Paddles

17 18
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

Magnification Platform C-Arm Controls

19 20
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

System Power Controls System Label Location

21 22
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SSYYSSTTEEMM DDEESSCCRRI PI PTTI O

I ONN AANNDD UUSSEERR I N
I NTTEERRFFAACCEE

System Label Location User Interface

User Login

23 24
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

physicist

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SSYYSSTTEEMM DDEESSCCRRI PI PTTI O
I ONN AANNDD UUSSEERR I N
I NTTEERRFFAACCEE

User Interface
Quality Control

34
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SSYYSSSYSTEM
TTEEMM DDEEDESCRIPTION
SSCCRRI PI PTTI O
I ONN AAND
ANNDD UUSSEERRINTERFACE
USER IN
I NTTEERRFFAACCEE SYSTEM DESCRIPTION AND USER INTERFACE

User Interface
Image Acquisition

47
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE

49 50
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SSYYSSTTEEMM DDEESSCCRRI PI PTTI O
I ONN AANNDD UUSSEERR I N
I NTTEERRFFAACCEE

User Interface
Other Tools

58
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SSYYSSTTEEMM DDEESSCCRRI PI PTTI O
I ONN AANNDD UUSSEERR I N
I NTTEERRFFAACCEE

User Interface
Test Patterns and QC Review

70
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 SYSTEM DESCRIPTION AND USER INTERFACE SYSTEM DESCRIPTION AND USER INTERFACE

Thank you!

78
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011

SYSTEM DESCRIPTION AND USER INTERFACE

Lunch TIME

Break for 1 hour

LUNCH is on your own

79
© 2011 Hologic, Inc. All rights reserved. PRE00027 rev 03/2011
 QUALITY CONTROL PROCEDURES

Quality Control
Quality Control Procedures Procedures
14 March 2015 – Indianapolis, IN
Selenia® Dimensions® Digital Breast Tomosynthesis Systems

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

Selenia Dimensions: Selenia Dimensions Systems:

Image Acquisition Modes
3 (1) Selenia Dimensions 2D
FFDM System

Conventional Only

(2) Selenia Dimensions DBT

Combo: Tomo + Conv System
under the same
compression
1 2 Combo: Tomo + Conv
under the same
1 compression
3

Conventional Only Tomosynthesis Only Conventional Only Tomosynthesis Only

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

Selenia Dimensions: Selenia Dimensions: Techniques

X-Ray Generation
Conventional 2D Imaging Tomosynthesis Imaging

• a-Se detector, 24×29 cm area • a-Se detector, 24×29 cm area

• 70 μm pixel size • 140 μm pixel size
X-Ray Tube • Tungsten (W) Anode
• Rh and Ag filters • Al filter
• LFS: 0.3 mm; SMF: 0.1 mm
• HTC grid in contact mode; • No anti-scatter grid
No grid in magnification mode • Moving tube, 15°sweep
• Conv: 50 μm Rh; 50 μm Ag
X-Ray Filtration • Tomo: 0.7 mm Al • Moving detector
• 15 projections
• 200 mA max for LFS • 3-4 seconds acquisition
X-Ray • 50 mA max for SFS
Generator • Reconstruction
• Max mA varies with kVp
- ~90 μm pixel size
- 1 mm slice spacing

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

Comparison of Hologic FFDM Systems

Selenia (Moly) Selenia(Tungsten) Selenia Dimensions

2D Tomosynthesis
Detector a-Se, 24x29 cm a-Se, 24x29 cm a-Se, 24x29 cm

Pixel 70 μm 70 μm 70 μm 140 μm

Anode Moly Tungsten Tungsten Quality Control (QC) Tests for

Focal spot 0.3 mm LFS
0.1 mm SFS
0.3 mm LFS
0.1 mm SFS
0.3 mm LFS
0.1 mm SFS
0.3 mm LFS
Medical Physicist
Filters Mo and Rh Rh and Ag Rh and Ag Al

Filter 30 μm 58 μm 50 μm 700 μm
thickness
mA (Max) 100 100 200

SID (cm) 66 66 70

Grid Contact mode: HTC Contact mode: HTC Contact mode: HTC Contact mode: no
Mag mode: no grid Mag mode: no grid Mag: no grid grid

Dose 1.60 mGy 1.20 mGy 1.20 mGy 1.45 mGy

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES

Selenia Dimensions QC Manual

The QC manual covers:

- Selenia Dimensions 2D FFDM system
- Selenia Dimensions DBT system

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

MP Tests for Dimensions 2D

MP Tests with Tomosynthesis Option
 QUALITY CONTROL PROCEDURES

Technologist Tests – Dimensions 2D

Quality Control (QC) Tests for

Technologist

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES

Alternative Standard 9
Technologist Tests –
Tomosynthesis Option

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

Selenia Dimensions QC Tests (Medical Physicist) QC Program Prerequisites

1. Mammographic Unit Assembly Evaluation • Hologic Selenia Dimensions User Manual
2. Collimation Assessment • American College of Radiology: Mammography Quality Control
3. Artifact Evaluation Manual, 1999 (USBN 1-55903-142-5)
4. kVp Accuracy and Reproducibility
5. Beam Quality Assessment — Half-Value Layer Measurement
6. Evaluation of System Resolution
7. Automatic Exposure Control (AEC) Function Performance
8. Breast Entrance Exposure, AEC Reproducibility, and Glandular
Average Dose
9. Radiation Output Rate
10. Phantom Image Quality Evaluation
11. Signal-To-Noise and Contrast-To-Noise Measurements
12. Diagnostic Review Workstation Quality Control
13. Detector Ghosting (Troubleshooting Use Only)
Printer and Acquisition Workstation Display
© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

Technologist Tests – Dimensions 2D

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

1. Mammographic Unit Assembly Evaluation

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

1. Mammographic Unit Assembly Evaluation 1. Mammographic Unit Assembly Evaluation

• Free standing unit is mechanically stable
• All moving parts move smoothly, without obstruction to motion
• All locks and detents work properly
• Image receptor is free from vibrations
• Compressed breast thickness scale is accurate to ±5 mm and
reproducible to ±2 mm.
• Patient or operator is not exposed to sharp or rough edges, or other
hazards
• Operator technique charts are posted
• Operator protected during exposure by adequate x-ray shielding
• All indicator lights work properly
• Auto-decompression can be overridden to maintain compression (and
status displayed)
• Manual emergence compression release can be activated in the
case of power failure

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

2. Collimation Assessment
Equipment
• X-ray recording media (e.g. screen-film cassettes, CR
plate, self-developing film)
• Screening paddles
2. Collimation Assessment

• Four small attenuators (i.e. coins of one size)

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

2. Collimation Assessment 2. Collimation Assessment

(1) X-Ray Field to Light Field (1) X-Ray Field to Light Field
• The total misalignment on opposite sides ≤ 2% of the SID • Measured once using the 24 × 29 cm screening paddle
• The misalignment between x-ray filed and light field, measured
(2) X-Ray Field to Image Receptor using this paddle, can be applied to other paddles
• The x-ray field must NOT extend by more than 2% of the SID at
any of the four sides of the image receptor
e.g. X-Ray Field to Light Field test using self-developing film
• The radiation field must extend beyond the edge of the digital
Unexposed area
image receptor on the chest wall side
Exposed area (color unchanged)
(3) Compression Paddle to Image Receptor (color changed)
• The anterior edge of the compression paddle must be aligned
just beyond the chest wall edge of the image receptor
• The anterior edge of the compression paddle must NOT extend
beyond the chest wall edge of the image receptor by more than
1% of the SID
Source to Image Receptor Distance (SID) = 70 cm Before x-ray exposure After x-ray exposure
© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

2. Collimation Assessment 2. Collimation Assessment

(2) X-Ray Field to Image Receptor (2) X-Ray Field to Image Receptor

• 2D
o 24 × 29 cm screening paddle
o 18 × 24 cm screening paddle, LEFT position
o 18 × 24 cm screening paddle, CENTER position
o 18 × 24 cm screening paddle, RIGHT position

• Tomosynthesis
o 18 × 24 cm screening paddle, CENTER position, using
“Zero-Degree Tomo” 18 × 24 cm, Left 18 × 24 cm, Center 18 × 24 cm, Right

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

2. Collimation Assessment 2. Collimation Assessment

(2) X-Ray Field to Image Receptor (3) Compression Paddle to Image Receptor
• Paddles to be tested:
o 24 × 29 cm screening paddle
o 18 × 24 cm screening paddle, CENTER position
o Small Breast screening paddle (when available)

Note:
• Put 4.0 cm BR-12 or PMMA under
the paddle
• Apply 10-12 lb compression force

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

2. Collimation Assessment
(1) X-Ray Field to Light Field
• 2D
o 24 × 29 cm screening paddle
(2) X-Ray Field to Image Receptor
• 2D
o 24 × 29 cm screening paddle
o 18 × 24 cm screening paddle, LEFT position
3. Artifact Evaluation
o 18 × 24 cm screening paddle, CENTER position
o 18 × 24 cm screening paddle, RIGHT position
• Tomosynthesis
o 18 × 24 cm screening paddle, CENTER position, “Zero-Degree Tomo”
(3) Compression Paddle to Image Receptor
• 2D
o 24 × 29 cm screening paddle
o 18 × 24 cm screening paddle, CENTER position
o
© 2011 Hologic, Inc. All rights reserved.
Small Breast screening paddle (when available) PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

3. Artifact Evaluation 3. Artifact Evaluation

(1) DICOM Printer Artifact Evaluation
• Send an “electronic Flat Field” image to DICOM printer

(2) System Artifact Evaluation

• Flat field images to be acquired and reviewed
o 2D contact mode (LFS) with Rh and Ag filters
o 2D magnification mode (SFS) with Rh and Ag filters
o DBT projection (LFS) with Al filter

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

3. Artifact Evaluation
“Actual Pixel” button to
get into 1:1 display mode
(1 detector pixel to 1
monitor pixel)

4. kVp Accuracy and Reproducibility

Image acquisition

Image review
© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

4. kVp Accuracy and Reproducibility 4. kVp Accuracy and Reproducibility

Performance Criteria
• Perform the test following the 1999 ACR Mammography
Quality Control Manual. • kVp must be accurate within ± 5% at:
o the lowest clinical kVp that can be
measured by the kVp test device
• Protect the detector by covering the detector with a layer of
o the 28 kVp
≥0.5 mm lead or lead equivalent
o the highest available clinical kVp

• If non-invasive kVp meter is used to measure kVp, it must • At 28 kVp, the coefficient of variation
be calibrated to the target-filter combination and the kVp of the kVp must be ≤ 0.02
range used.

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

5. Beam Quality Assessment — Half-Value

Layer Measurement
• Perform the test following the 1999 ACR Mammography
Quality Control Manual

5. Beam Quality Assessment — Half-Value • Protect the detector by covering the detector with a layer of
Layer Measurement ≥0.5 mm lead or lead equivalent

• When measuring the HVL under tomosynthesis,

o use the “Zero-Degree Tomo” view
o the compression paddle should be lower than 24 cm for
the exposure to be allowed

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

5. Beam Quality Assessment — Half-Value

Layer Measurement
• For kVp range of less than 50,
HVL > kVp/100 + 0.03 (mm Al)

6. Evaluation of System Resolution

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

6. Evaluation of System Resolution 6. Evaluation of System Resolution

Line pair phantom

• Rotated by 45º
• Placed on 4 cm flat field phantom
• Under compression paddle “Actual Pixel” button
• Apply 15-20 lb force
• Large focal spot

Image Acquisition
• One 2D acquisition
• One Tomosynthesis acquisition

“Magnification” button

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

6. Evaluation of System Resolution

Performance Criteria:

• 2D resolution >7 lp/mm at 45°

7. Automatic Exposure Control (AEC)

• Tomosynthesis resolution >3 lp/mm at 45°
Function Performance

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

7. Automatic Exposure Control (AEC) 7. Automatic Exposure Control (AEC)

Function Performance Function Performance
AEC modes
• AEC positions
• Auto-Filter o Auto-AEC
- filter, kVp, mAs automatically determined o Seven manual positions (marked on compression
paddle)
• Auto-kV
- filter manually selected • AEC Method
- kVp and mAs automatically determined o Filter and kVp are determined by breast thickness (in
Auto-Filter and Auto-kV)
• Auto-Time o A pre-exposure is made to determine mAs of the main
exposure
- filter and kVp manual selected
o The main exposure is calculated so that the exposure
- mAs automatically determined index (EI) at a thickness is within a pre-defined range
© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

7. Automatic Exposure Control (AEC) 7. Automatic Exposure Control (AEC)

Function Performance Function Performance
(1) For 2D and Tomosynthesis, check AEC performance (1) Performance at different thickness and imaging modes
- at different THICKNESSES (2, 4, 6 and 8 cm) and • Acrylic or BR-12 blocks
- in different IMAGING MODES (contact mode and • AEC position 2
magnification mode)
• Compression thickness must be precise
(2) For 2D, check AEC performance at different exposure • For 2D
COMPENSATION STEPS - Contact mode (LFS, grid): 2, 4, 6, 8 cm
- Mag mode (SFS, no grid): 4 cm
• For Tomosynthesis
- Contact mode (LFS, no grid): 2, 4, 6, 8 cm

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

7. Automatic Exposure Control (AEC) 7. Automatic Exposure Control (AEC)

Function Performance Function Performance
4.0 cm displayed
compression thickness

Magnification stand
© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

7. Automatic Exposure Control (AEC) CNR Correction Factors, 2D

Function Performance

(1) Performance at different thickness and imaging modes

• Thickness-dependent Exposure Index (EI) LFS
Corrected Pixel Value =
(Measured Mean Pixel Value – 50) / CNR_Factor

SFS

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

7. Automatic Exposure Control (AEC) 7. Automatic Exposure Control (AEC)

Function Performance Function Performance
(1) Performance at different thickness and imaging modes
• Thickness-dependent Exposure Index (1) Performance at different thickness and imaging modes
• Thickness-dependent Exposure Index (EI)
Corrected Pixel Value =
(Measured Mean Pixel Value – 50) / CNR_Factor

© 2011 Hologic, Inc. All rights reserved. Example: Data sheet for 2D AEC test
PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

7. Automatic Exposure Control (AEC)

CNR Correction Factors, Tomo Function Performance
(1) Performance at different thickness and imaging modes
• Thickness-dependent Exposure Index

LFS

Example: Data sheet for tomosynthesis AEC test

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

7. Automatic Exposure Control (AEC) 7. Automatic Exposure Control (AEC)

Function Performance Function Performance
(2) Performance at different compensation steps
(2) Performance at different compensation steps
• Take the average of the exposure index
• Only for 2D (Tomo does not have compensation
values of 3 images measured at Step 0,
steps)
V_Step0 = AVE (EI_Step0 – 50)
• Acrylic or BR-12 blocks
• At each compensation step, calculate
• AEC position 2
Ratio = (EI – 50) / V_Step0 Step -3 0.56 – 0.66
• 4 cm only (compression thickness must be precise)
• The ratio for each step must Step -2 0.66 – 0.78
• For 2D, contact mode Step -1 0.78 – 0.92
stay within the range
- Make 3 exposures at Step 0 Step +1 1.06 – 1.24
- Make 1 exposure at Steps: -3, -2, -1, 1, 2, 3, 4 Step +2 1.22 – 1.43
Step +3 1.40 – 1.64
Step +4 1.61 – 1.89
© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

7. Automatic Exposure Control (AEC)

Function Performance
(2) Performance at different compensation steps

8. Breast Entrance Exposure, AEC

Reproducibility, and Glandular Average Dose

Example: Data sheet for 2D AEC compensation step test

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

8. Breast Entrance Exposure, AEC 8. Breast Entrance Exposure, AEC

Reproducibility, and Glandular Average Dose Reproducibility, and Glandular Average Dose
• Image Acquisition
o Use ACR phantom
o Use “ACR Phantom View” provided by the system
o Set up the dosimeter as described in the 1999 ACR
Mammography Quality Control Manual
o Make 4 acquisitions in 2D mode
o Make 4 acquisitions in tomosynthesis mode

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

8. Breast Entrance Exposure, AEC 8. Breast Entrance Exposure, AEC

Reproducibility, and Glandular Average Dose Reproducibility, and Glandular Average Dose
• AEC Reproducibility
o 2D: calculate the mean, standard deviation (STD) and
the coefficient of variation (CV) for mAs and exposures
o Tomo: calculate the mean, standard deviation (STD) and
the coefficient of variation (CV) for mAs and exposures
o Performance criteria: CV < 0.05

• Average Glandular Dose (AGD)

o 2D: calculate the AGD based on dose conversion factors
provided in the QC manual
o Tomo: calculate the AGD based on dose conversion
factors provided in the QC manual
o Performance criteria:
- AGD < 3.0 mGy (2D or Tomo)
© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 - AGD < 3.0 mGy (2D + Tomo)
© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

9. Radiation Output Rate

• Protect the detector by covering the detector with a layer of
≥0.5 mm lead or lead equivalent

9. Radiation Output Rate

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

9. Radiation Output Rate

• Acquisition Technique
At 28 kVp, the tube current is 160 mA

10. Phantom Image Quality Evaluation

• Performance Criteria
For W/Rh, at 4.5 cm above the breast support
Output rate > 230 mR/s (or 2.0 mGy Air Kerma/s)

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

10. Phantom Image Quality Evaluation 10. Phantom Image Quality Evaluation – 2D
• Image Acquisition • Click the “Actual Pixel” button to
o Use ACR phantom bring the image to full resolution
o Use “ACR Phantom View” provided by the system • Score the phantom following
o Acquire 2D and tomosynthesis phantom images using AEC 1999 ACR Mammography Quality
Control Manual
• 2D Phantom Score:
- 5 fibers, 4 specs, 4 masses
- Due to phantom variation, a
score of (4.5, 4.0, 3.5) is
acceptable if SNR and high
contrast resolution tests pass

“Actual Pixel” button

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

10. Phantom Image Quality Evaluation – Tomo

• Click the “Actual Pixel” button
to bring the image to full
resolution
• Scroll to the reconstruction
slice in which the phantom
elements are in focus 11. Signal-To-Noise and Contrast-To-Noise
• Score the phantom following Measurements
1999 ACR Mammography
Quality Control Manual
• 3D Phantom Score:
4 fibers, 3 specs, 3 masses

“Actual Pixel” button

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

11. Signal-To-Noise and Contrast-To-Noise 11. Signal-To-Noise and Contrast-To-Noise

Measurements Measurements
• Image Acquisition
o Use ACR phantom
o Use “ACR Phantom View” provided by the system
o Acquire 2D phantom image using AEC

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

11. Signal-To-Noise and Contrast-To-Noise

Measurements

• SNR and CNR Calculation

SNR = (Mean_background – DC_offset) / Std_background
CNR = (Mean_background – Mean_disk) / Std_background

• Performance Criteria
SNR ≥ 40
CNR within ±15% of the reference value set by physicist

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

12. Diagnostic Review Workstation Quality Control

• Hologic QC only to be used if workstation does not
have its own QC protocol
• The SecurView workstation comes with its own QC
manual
12. Diagnostic Review Workstation Quality • Tests to be performed under Hologic modality QC
Control – All displays
• While level measurement
• DICOM GSDF compliance
– CRT displays, only
• Black level measurement
• White level uniformity
• Keep in mind that performance depends on the type of display
used. Check with manufacturer specs of each display for
performance requirements.

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

12. Diagnostic Review Workstation Quality Control

• Most review
workstation offer
QC software
13. Detector Ghosting
• Follow their QC
(Troubleshooting Use Only)
procedures as to
how to perform the
QC

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

13. Detector Ghosting (Troubleshooting Use Only) 13. Detector Ghosting (Troubleshooting Use Only)
• Create Ghosting Effect
1. Create Ghosting Effect

2. Measure Ghosting Factor

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

13. Detector Ghosting (Troubleshooting Use Only) 13. Detector Ghosting (Troubleshooting Use Only)
• Create Ghosting Effect • Measure Ghosting Factor

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

13. Detector Ghosting (Troubleshooting Use Only) 13. Detector Ghosting (Troubleshooting Use Only)
• Measure Ghosting Factor • Measure Ghosting Factor

3 2 1

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011

QUALITY CONTROL PROCEDURES QUALITY CONTROL PROCEDURES

13. Detector Ghosting (Troubleshooting Use Only)

• Measure Ghosting Factor

• | Ghosting factor | ≤ 0.3

ROI1 in background
without ghosting Ghosting Factor = Thank you!
meanregion2 - meanregion3
ROI2 in Al filter
without ghosting meanregion1 - meanregion2

ROI3 in Al filter
with ghosting

© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011 © 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 QUALITY CONTROL PROCEDURES

10 minute break
Cokes & Soft Drinks
are on back table
Talk 6
© 2011 Hologic, Inc. All rights reserved. PRE00028 rev 03/2011
 Quality Control of
Softcopy Display
Tomosynthesis Displays
• Monitor design/types
Digital Breast Tomosynthesis Imaging
• Flat Panels
• Monitor performance characteristics
• DICOM Part 14 Grayscale Standard Display
14 March 2015 – Indianapolis, IN Function
• Flat Panel Calibration
Jerry A. Thomas, MS, FAAPM, DABR, CHP, DABSNM
Diagnostic Medical Physicist
Via Christi Hospitals Wichita, Wichita, KS

Medical Display Choices

Flat Panels
1280 x 1024 2048 x 1536 (3 MP) 1200 x 1600 (2 MP) 1200 x 1600 2048 x 2560 (5 MP)
1024 x 1280 (1 MP)
55 fL / 55 fL / 35 fL / 75 fL / 235 fL / 175 fL /
200 cd/m2 200 cd/m2 120 cd/m2 255 cd/m2 800 cd/m2 600 cd/m2

3840 x 2400 (9.2 MP) 1600 x 1200 1600 x 1200

1600 x 1200 1280 x 1024
80 fL / 120 fL / 235 fL /
35 fL / 35 fL / 800 cd/m2
270 cd/m2 400 cd/m2
120 cd/m2 120 cd/m2

3
 Each Sub-Pixel is Independent
Liquid Crystal (LC) Light Valve and controlled by a TFT
Not a Liquid … and not a Crystal

• LC slurry has orientation between crystals Polarizer

–called the director relationship

• Light follows the crystal path Liquid Crystal

–when orientation is parallel on major axis V
• Light is cut-off
–when orientation is disturbed by a voltage field (classic core) Polarizer

• Each Pixel has three sub pixels (R,G,B)

LC molecules normally twist light Electric field causes LC molecules to
90° letting light pass through the re-align, blocking light – Off pixel
front polarizer – On pixel

AM-LCD Structure
Graphic boards
Backlight

Polarizing Filter • Dome 5MP

Glass Substrate – 2048 x 2560 x 8 bit
Transparent
Electrode
• Dome 2MP
Alignment Layer • BarcoMed 5MP and 2MP
Liquid Crystals – 2048 x 2560 x 8 or 10 (e.g. 256 or 1024 SOG)
G
B Color Filters • Matrox and nVidia
B R – Commercial Graphics Controllers – for CRT
G B Glass Substrate
R G and Flat Panels
R Polarizer Filter

8
Monitor Performance Characteristics Contrast Ratio
• Luminance Uniformity • Anti-reflective coating • Depends on
– Fixed at time of – Maximum luminance
– Will add color to monitor
manufacture
– Should be < 18% (center to • Glass color – Glass Transmission
edges) – Ambient Light
– Best monitors – 8% – 12% • Monitor gamma
– White areas of the
• Contrast Ratio ~ 2.2 image
– Lmax/Lmin • Focal Spot Size (CRT)
• Phosphor Type (CRT)
– P45 • Input impedance
– P104 – 50 or 75 ohms

9 Recommended Contrast ratio for Mammo:

10
at least 400:1

Phosphors - CRT
Glass Envelope Construction - CRT
• P45
– One component phosphor
– Low aging
• 27% total transmission
– Short decay time after 2 ms • Ambient Light
– No color variation over
time
attenuation (93%)
• Anti-reflective/anti-
• P104 static coating
– Dual component phosphor
– High efficiency • Same refractive index
– Can be run at luminance throughout
levels > 100 fL

11 12
11 12
Resolution vs Luminance - CRT Monitor Lifetime – CRT and FP
• Can last up to
50,000 hr

• Phosphor and
glass decrease

• Cut-off control

• Control of
Gamma function

13 14
14

Long Term Reliability & Conformance of Displays - FP

The Complete “Hands-Off” Conformance Test Monitor stability over time
Controlling the
White Level

Hands-off
Conf Test
Sampling
Points

Regulating the
Gamma
Function

Stabilizing the
Black Level

15 16
 Grayscale Standard Display GSDF
Function
• A methodology for measuring the • Neither
characteristic curve of a display system – A performance standard
– For conformance testing – Image display standard
– For altering display system to match GSDF
• Does not define
• Key elements of GSDF – Luminance
– Target: 10% of the monitor area – Luminance range
– Optical density range
– Surround: 20% of Lmax
– Presentation of picture element values

GSDF DICOM Part 14 Factors that affect visual perception

• What does it provide • Room lighting

– Ambient light
– Standardized basis for video display calibration
– Specular reflection
– Table of the GSDF as a function of the Just-
Noticeable Difference Index • Lmax and Lmin

• Icons on screen

• Veiling glare
Monitor Gamma vs Perceptual Monitor Perceptual Linearization
Linearization and
JND Determination
• Background is set
• Line Pair target in center 10% of viewing area
• Increase target until it is just visible
• Increase background to target level
– Repeat previous two steps
• Example follows:
 Grayscale Standard Display Function Grayscale Standard Display Function
DICOM Part 14 DICOM Part 14
Perceptually Linearized Response Curve
white 1000

100
Different change
in absolute luminance 10
Luminance

1
0 200 400 600 800 1000
GSDF defines
.1
transformation of
Digital Driving
.01
Levels (DDL) JND Index
to Luminance.

Same number of Just Noticeable Differences == Same perceived contrast

black
Courtesy of Brad Hemminger, UNC Chapel Hill, NC SPIE Vol. 2707-58, 1996
Just Noticeable Differences (No of JNDs)

What are the pieces needed to

give an image “Pizzazz”?

Optimized Softcopy Display: • Monitor Calibration

Putting the Pieces Together • Acquisition Device – The Processed Image

• Image Data Optimized for Display
• The Displayed Image
 The First Piece:
Monitor Calibration
• Necessary for standardization

• Corrects the natural display gamma to a FLAT PANEL CALIBRATION

perceptually linearized model

• Requires frequent conformance testing

34

Flat Panel Calibration The First Piece:

Setting White and Black Levels Monitor Calibration
• Select Lmax • Setting the display black and white levels
– Generally between 300 and 375 cd/m2
– Some newer displays maybe 600 cd/m2 • Determining the natural gamma of the display
• Select Contrast Ratio
• System AUTOMATICALY calibrates using
external sensor and/or Puck

35
 The First Piece:
Monitor Calibration
• Setting the display black and white levels

• Determining the natural gamma of the display

• Application of a Look-Up Table (LUT)

Application of an LUT
Display Natural Gamma Gamma Correction LUT Conformance Validation

• Manual Conformance Testing

+ • DICOM Part 14 Grayscale Standard Display Function
• Monitor Luminance Uniformity
Perceptually
Linearized Display
 Analysis of Measured Data

Conformance Validation
Automated Conformance Testing
• Manual Conformance Testing
• Barco
• DICOM Part 14 Grayscale Standard Display Function – Medi-Cal
• Monitor Luminance Uniformity – Medi-Cal QA Web

• Automated Conformance Testing • Double Black

• On Local Workstation – X-cal (local)
• Remotely – X-cal (remote)
 The Second Piece: Processed Image
The Processed Image Breast Image 1 algorithms
• Processing
unknown (proprietary)
Image Acquisition Processed Image

• Is not suitable for softcopy

interpretation
• Can not be optimized with
linear windowing
Processing Processing

Detector Gain, Offset Image Data Enhancement

and Flatfield Corrections

Raw Image

The Third Piece: Non-Linear Windowing

Optimization for Display

• Produces:
•An image with diagnostic content (pizzazz)

•An image that looks familiar to the radiologist

• Accomplished through non-linear windowing

 Optimized Image
How Do the Pieces Fit Together?
Breast Image 2
Processed Image Optimized Image

•The processed image data is passed through a non-

linear window

How Do the Pieces Fit Together?

Processed Image Optimized Image
• The processed image data is passed through a non-
linear window

• The non-linearly windowed data is applied to a gamma-

corrected (perceptually linearized) display system
 How Do the Pieces Fit Together?
Processed Image Optimized Image
• The processed image data is passed through a non-
linear window

• The windowed data is applied to a gamma-corrected

(perceptually linearized) display system

• The result is a composite function that:

• is optimized for viewing
• can be “log luminance linear” (i.e. look like film)

Processed Image Optimized Image

 Example – sigmoid PLUT
Presentation LUT (PLUT)
• Added into the display
– Alters the GSDF calibration
– Is Modality Dependent
• Mammo
• Chest
• Extremity

Sigmoid
Acquisition image Presentation
Dicom GSDF
LUT

Processed Image Optimization for Display

Example – sigmoid PLUT Image Acquisition Processed Image Processed Data Non-Linear Window

+
• But if the GSDF transformation
resulted in making big skips near
Processing Processing
+
black...
Perceptually Linearized
Display
Raw Image

Monitor Calibration
Image Presentation

Display Gamma Gamma Correction LUT

• Then when the PLUT is applied the Result of

dark parts look too flat or cartoon insufficient bit +
depth after
GSDF
 Flat Panel Calibration Uncorrected LCD Display
Medical Grade vs Home computer
• “Out of the box” Vendor Supplied LUT
• DICOM Part 14 calibration using 256 sampling
points
– Get less than 256 display values
Areas of lost
– Have Zero JND Shifts between some DLs or inaccurate
information up
• Optimized DICOM Part 14 calibration using to 160 gray
scales lost
“pseudo shades of gray”
– Get TRUE 256 display values
– No Zero JND shifts

Natural Response of LC pixel to DDL Discontinuities you don’t see

between 100-150 DDL
Higher Contrast
Area

LCD displays have discontinuities in

Areas of lost their response that cause calibration
or inaccurate errors if 255 steps are used for
information calibration.

Factory calibration is the key to

accuracy.

An un-calibrated display lags behind

the DICOM curve.

Ambient conditions reduce contrast

and masking of vital information.

Non-medical grade LCD’s add in

lamp instability as a variable.
 Linear LUT – “out of the box” Linear LUT – showing zero JND shifts
Red < 1 jnd Blue > 3 jnd

65 66

Video LUT with 256 Calibration values 256 Cal values - Showing JND Shifts

67 68
Optigrayscale – TRUE 256 JNDs Optigrayscale 256 JNDs Annotated

69 70

Differing #’s of Calibration Points

Linear “256” cal values True 256 cal values Display
Quality Control

72
71
 Conformance Testing
QC/QA of Digital Mammography and
Brightness Uniformity QC
Workstation
Image quality
Luminance uniformity
Ambient light
 Control of Specular reflections
DICOM Part 14 – Gray Scale
Performance Standard
 Measure at System Level and Application Level
 QC using SMPTE Pattern

73 74

Barco QC Program
• QC Tests for the Radiologic Technologist
– Display Screen Cleanliness – Daily
– Viewing Conditions Check for the Display
• Daily when clinical image review is planned
– Image Quality Check – Weekly

• QC Tests for the Medical Physicist

– Visual Check – Annually
– Ambient Light – Annually -- < 20 lux
75
 Technologist Testing Monitor Cleaning
• Check for fingerprints, dirt, scratches
Test Nio 5 MP Coronis 5 MP • Brush with nylon brush to remove dirt/dust
Display Screen cleaning Daily Daily • Soft cloth ONLY
Viewing Conditions Daily Daily – Old Tee Shirt
Image Quality Monthly Weekly – Camera – Microfiber Cloth
• NO WINDEX!!!
– Removes protective anti-glare coating on monitor

• If Dirty and/or Greasy

– Clean with mild detergent solution
1 drop dishwashing liquid to 8 oz Warm Water
77

Workstation Viewing Conditions Coronis 5MP QC Testing

• Room light levels
– Keep lights off or low -- < 5 lux
» 1 to 2 lux preferred

• Equipment positioning
– View boxes and Softcopy displays should be parallel to
one another … not at an angle or back to back

80
Coronis 5MP QC Testing Nio 5MP QC Testing

81 82

Nio 5MP QC Testing Image Display Test Patterns

83 84
Pattern Evaluation Frequency
Do you remember this?
Physician
- Daily prior to each reading session
Technologist
- Daily on acquisition workstation
- Weekly on Dx monitors in work area
Medical Physicist
- Monthly – Department QC review
- Annually – detailed eval/comparison with prior years data

85 86

Or this? If so -- It’s Howdy Doody Time !!

Buffalo Bob Howdy Doody

THE STAR of
the SHOW

First Children’s
program to run 5
days a week

87 88
 SMPTE Test Pattern Modified SMPTE Test Pattern
Evaluation
0 / 5% Black
Center Targets
95/100% White
For measuring
Density Step Ramp Lmax and Lmin
(0 to 100%)
Veiling Glare will
Increase black level
High Contrast Bars
measurement
Center
Corners Pattern Should NOT
be used for this
Low Contrast Bars measurement
Center
Corners

Black bar on White

White bar on Black

Letters

90

AAPM TG18 QC Pattern Briggs Test Pattern

• Multiple inserts embeded in a mid- • 8 surrounds
value background
– Gridlines – From Lmax
– Luminance Patches – To Lmin
– Line Pair Patterns
– Cx patches 100, 75, 50, 25% of Lmax
– Contrast Detail “Quality Control” • 17 test objects
– Vertical Bars
– From 9 x 9 square
– White and Black Bars
– Horizontal Area for Flat Panel – To 4x4 pixel square
evaluation (1 on and 1 off)
– Border around outside

91 92
92
AAPM Briggs Test Pattern Circle
• Measures:
– Geometrical distortion
– Linearity
– Warping

– Ability to see 1% contrast

93 94
94
END

Printer Overview
• High resolution Laser Printers
– Kodak 8600 – No longer manufactured Still used
WIDELY
Hard Copy Device QC –
–
–
Kodak 8610 – No longer manufactured
Kodak 8900
Kodak 6800
– Agfa LR5200 – No longer manufactured or supported
– AGFA DS4500M
– Fuji DryPix 7000
– Fuji DryPix 5000
– Fuji DryPix 4000
– Fuji FM-DP-L
– Konica-Minolta DryPro 793
– Quality Control
• Acceptance Testing
• NEMA “standard” XR 23-2006
 Printers Validated by FFDM Common Mammo Laser
Manufacturers Characteristics
VALIDATED Printers
• AGFA 4500M
In service No longer
Manufactured
• Specifications
• • AGFA 5500 – Pixel matrix: 5025 x 6200
• • AGFA 5503 • KODAK 8600
• • AGFA AXYS
• KODAK 8610 – Pixel size: 38.75 to 50 um – newer system down to 25 um
• FUJI FM-DP L
• • KODAK 8900 – 12 bit (4096 gray levels) – newer systems 14 bit
• • KODAK 6800 No Longer Supported by
• • CARESTREAM DryView 5850 Printer Manufacturer
– Resolution: 508 to 655 dpi – newer system up to 1016 dpi
• • KONICA 793 • AGFA LR5200 – Image area: 19.5 x 24 cm up to 35 x 43 cm
• • KONICA 873
• • FUJI DRYPIX 4000 (see note 1)
– Throughput: 40 to 200 films/hr
• • FUJI DRYPIX 5000 (see note 1) – Dmax: Film dependent – normally < 3.6 – now ~ 4.0
• • FUJI DRYPIX 7000 (see note 1)
• • SONY UP-DF750
– Dmin: Film dependent – normally ~ .25
97
98

Kodak Dryview 8610

Kodak Dryview AGFR LR
• Specifications
8600/8610 Fuji Drypix 7000
5200 Fuji Drypix 5000 – Pixel matrix: 5025 x 6200
– Pixel size: 38.75 um
– 12 bit (4096 gray levels)
– Resolution: 655 dpi
– Image area: 19.5 x 24 cm
– Throughput: 40 films/hr
– Dmax: Film dependent
AGFA Drystar Konica DryPro 793
Kodak Dryview
Kodak Dryview 4500M Fuji Drypix FM-DP L – Dmin: Film dependent
6800
8900
100
 Kodak Dryview 8900 Kodak DriView 6800
• Specifications • Specifications
– Pixel size: depends on film – Pixel size: 39 micron
size (5 different available) – 14 bit (16,384 gray levels)
– 14 bit (16,384 gray levels) – Resolution: 650 dpi
– Resolution: 650 dpi – Image area: All film sizes
– Image area: All film sizes – Throughput: 200 films/hr
– Throughput: 200 films/hr – Dmax: 3.6 – 4.0 for mammo
– Dmax: 3.6 – 4.0 for mammo – Dmin: Film dependent
– Dmin: Film dependent

101 102

Agfa Scopix LR5200P Agfa DRYSTAR 4500M

• Specifications
• Specifications
– Pixel matrix: Depends on
– Pixel matrix: 4,776 x 5,944 film size
– Pixel size: 40 um (high – Spot size: 50 um
resolution mode)
– 12 bit (4096 gray levels)
– 12 bit (4096 gray levels)
– Resolution: 508 ppi
– Resolution: 630 ppi
– Film sizes: 8 x 10, 10 x 12
– Film sizes: 14 x 17, 14 x 14,
– Throughput:
14 x 11, 8 x 10
• 80 8 x 10 films/hr
– Throughput: 60 films/hr • 60 10 x 12 films/hr
– Dmax: 3.5 – Dmax: > 3.5
– Dmin: <= 0.25 103
– Dmin: <= 0.25 104
 Fuji DryPix 7000 Fuji DryPix 5000
• Specifications • Specifications
– Pixel matrix: – Pixel matrix:
– Spot size: 100/50 um – Spot size: 100/50 um
– 14 bit (16384 gray levels) – 14 bit (16384 gray levels)
– Resolution: 508 ppi – Resolution: 508 ppi
– Film sizes: 8 x 10, 10 x 12 – Film sizes: 8 x 10, 10 x 12
– Throughput: – Throughput:
• 200 8 x 10 films/hr • 200 8 x 10 films/hr
• 240 10 x 12 films/hr • 180 10 x 12 films/hr
– Dmax: > 3.6 – Dmax: > 3.6
– Dmin: <= 0.25 – Dmin: <= 0.25
105 106

Fuji DryPix FM-DP L

Fuji DryPix 4000
• Specifications
• Specifications
– Pixel matrix: Depends on film
– Pixel matrix: size
– Spot size: 100/50 um – Spot size: 50/100 um
– 14 bit (16384 gray levels) – 12 bit (4096 gray levels)
– Resolution: 508 ppi – Resolution: 508 ppi
– Film sizes: 8 x 10, 10 x 12 – Film sizes: 8 x 10, 10 x 14, 14 x 17
– Throughput: – Throughput:
• 200 8 x 10 films/hr
• 130 8 x 10 films/hr
• 160 10 x 12 films/hr
• 60 10 x 12 films/hr
– Dmax: > 3.6 – Dmax: > 3.6
– Dmin: <= 0.25 – Dmin: <= 0.25 108
 Konica-Minolta DryPro 793
Automatic Calibration
• Specifications
– Pixel matrix: up to 8079 x 9725 • All systems have automatic calibration
– Spot size: 43.75/25 um
capability
– 14 bit (16,384 gray levels)
– Resolution: 580/1016 ppi
– Self calibration sets Dmax and Dmin to
predetermined values
– Film sizes: 8 x 10, 10 x 14, 11x14,
14 x 14, 14 x 17 – All can vary slope of H&D curve
– Throughput:
• 120 8 x 10 films/hr
• 150 10 x 12 films/hr
– Dmax: 4.0
– Dmin: <= 0.25
109

Kodak 8600 / 8610 Calibration Film

AGFA Calibration Film
 Daily QC Printer QC Step Wedge - 8610
• Use same as “wet”
• Step Wedge film step wedge
• Manufacturer’s Pattern – B+F = step 1
• SMPTE Pattern – DD = step 12 – step 9
– MD = step 11
• ACR Phantom
• Plot results and
monitor trends

114

Daily QC and AT – Agfa 4500M

SMPTE in LP QC

– Medium Density – (5th step

down from max white)
– Use 60% step

116
 SMPTE in LP QC SMPTE in LP QC
• Medium Density – (5th step
– Medium Density – (5th step down from max white)
down from max white) – Density Difference
– Density Difference • (5th step minus 2nd step)
• (5th step minus 2nd step)
60% - 90% – Lower Density – (1st step
above max black) (10% step)

– NOTE: Some SMPTE patterns

• white = 0%
• others white = 100%
117

QC Program Differences Alternate QC Measurements

5% in 0/5% Patch
• ACR guidance for MRI differ 95% in 95/100% Patch

– 5% patch in 0/5% SMPTE Optical Control

– 95% patch in 95/100% Patch Density Limit
– Measure OD in 0, 10%, 40%, and 90% 0 2.45 ± 0.10
• Control limits: use +/- 0.10 for dry lasers 10% 2.10 ± 0.10
40% 1.15 ± 0.10
90% 0.30 ± 0.10
• Lacks measure of contrast
From ACR MRI 2004 MRI QC Manual p 55
 Carestream Health 6800 QC AAPM TG18 QC Image –
Mammo Test Image internal to printer

Acceptance Testing Printer Performance Test Pattern

• Verify continuity of lines
• Evaluate the following AT From Kodak 6800 Printer • No duplication of arc or other
– Print area on film lines
– Size of print area • Black and white squares
– Aspect ratio of print area should not be surrounded by
– Line uniformity opposite color line
– Density uniformity • Uniform OD on bar
– Test patterns: • 21 step from Max to min OD
• GE printer pattern
• HOLOGIC Printer pattern
• Circle Pattern
 Circle Test Pattern Thank You
• Measures
– Diagonals
• 1% contrast Questions?
– Geometry
– X – Y distortions

• Lacks Grey Scale Time for a BREAK

step Wedge 10 minutes

Talk 7
126
 Accreditation of Digital Accreditation of
Breast Tomosynthesis FFDM SYSTEMS
Digital Breast Tomosynthesis Imaging
•FDA Requirements – 2D FFDM

•FDA Requirements – 3D DBT

14 March 2015 – Indianapolis, In

Jerry A. Thomas, MS, FAAPM, DABR, CHP, DABSNM

Diagnostic Medical Physicist
Via Christi Hospitals Wichita, Wichita, KS

FDA and Digital Mammography FDA Regulations

• Provide a list of all personnel who began working
in (i.e., interpreting, performing, and surveying)
the FFDM modality prior to April 28, 1999 as
well as a list of personnel who began or will begin
working in the modality after April 28, 1999.

– All personnel must have 8 hours of training in the

modality prior to working with/using the modality

3 4
 FDA Regulations FDA Regulations
• Provide a satisfactory FFDM equipment • Follow the manufacturer's guidelines for
evaluation (including an evaluation of the quality assurance and quality control
Soft Copy Display system) performed by a tests as described by the manufacturer's
qualified medical physicist within 6 manual.
months prior to the date of your application.

5 6

Growth of FFDM Imaging FDA Facility / FFDM Statistics

• As of 01 August 2006
• Started SLOW
• Picked up in 2006 – Certified facilities, as of October 1, 2005: 8,881
– Total certified facilities: 8,829
• # of units in US has rapidly expanded – Total accredited units 13,556
– Certified facilities with FFDM units: 1,130 (13%)
• TODAY – FFDM is “Standard of Care” in
– Accredited FFDM units: 1,604 (12%)
Mammography
• Total annual mammography procedures reported, as of August 2006: 33,875,835

7 8
 FDA Facility / FFDM Statistics FDA Facility / FFDM Statistics
• As of 01 August 2007 • As of 01 August 2008
– Certified facilities, as of October 1, 2006: 8,832 – Certified facilities, as of October 1, 2007: 8,837
– Total certified facilities: 8,825 – Total certified facilities: 8,835
– Total accredited units 13,402 – Total accredited units 13,465
– Certified facilities with FFDM units: 2,090 (24%) – Certified facilities with FFDM units: 3,515 (40%)
– Accredited FFDM units: 3,077 (23%) – Accredited FFDM units: 5,356 (40%)

• Total annual mammography procedures reported, as of August 2007: 34,808,145 • Total annual mammography procedures reported, as of August 2008: 36,154,201

9 10

FDA Facility / FFDM Statistics FDA Facility / FFDM Statistics

• As of 01 August 2009 • As of 01 August 2010
– Certified facilities, as of October 1, 2008: 8,814 – Certified facilities, as of October 1, 2010: 8,713
– Total certified facilities: 8,711 – Total certified facilities: 8,651
– Total accredited units 12,844 – Total accredited units 12,432
– Certified facilities with FFDM units: 4,743 (54%) – Certified facilities with FFDM units: 5,826 (67%)
– Accredited FFDM units: 7,087 (55%) – Accredited FFDM units: 8,514 (68%)

• Total annual mammography procedures reported, as of August 2009: 36,983,579 • Total annual mammography procedures reported, as of August 2010: 38,823,849

11 12
 FDA Facility / FFDM Statistics FDA Facility / FFDM Statistics
• As of 01 August 2011 • As of 01 August 2012
– Certified facilities, as of October 1, 2010: 8,652 – Certified facilities, as of October 1, 2011: 8,620
– Total certified facilities: 8,620 – Total certified facilities: 8,626
– Total accredited units 12,270 – Total accredited units 12,414
– Certified facilities with FFDM units: 6,766 (78%) – Certified facilities with FFDM units: 7,388 (86%)
– Accredited FFDM units: 9,760 (79%) – Accredited FFDM units: 10,786 (87%)

• Total annual mammography procedures reported, as of August 2011: 38,844,222 • Total annual mammography procedures reported, as of August 2012: 39,351,056

13 14

FDA Facility / FFDM Statistics FDA Facility / FFDM Statistics

• As of 01 August 2013 • As of 01 June 2014
– Certified facilities, as of October 1, 2012: 8,654 – Certified facilities, as of October 1, 2013: 8,691
– Total certified facilities: 8,658 – Total certified facilities: 8,714
– Total accredited units 12,833 – Total accredited units 13,523
– Certified facilities with FFDM units: 7,858 (91%) – Certified facilities with FFDM units: 8,154 (94%)
– Accredited FFDM units: 11,792 (92%) – Accredited FFDM units: 12,790 (95%)

• Total annual mammography procedures reported, as of August 2013: 38,233,396 • Total annual mammography procedures reported, as of 01 Feb 13: 38,747,608

15 16
 FDA Approved DBT & FFDM Systems
FDA Facility / FFDM Statistics 19 Systems
approved by
13 Systems
FDA since
approved in
• Changes from 1 Aug 2006 to 1 Jun 2014 2000
2011

– Certified facilities, as of October 1, 2013: 8,881 8,691

8,6 190 (2.1%)
– Total certified facilities: 8,829 8,712 174 (2.0%) 1st NEW technology
– Total accredited units 13,556 13,234 1,040 (7.6%)
– Certified facilities with FFDM units: 1,130 (13%) 8,154 (94%)
– Accredited FFDM units: 1,604 (12%) 12,790 (95%)

• Total annual mammography procedures reported, as of Aug: 33,875,835 38,747,608

4,859,544 (14.3%)

How longiswill
FFDM the it take for DBTOF
STANDARD to CARE
becomefor
the 5 Systems No
17
STANDARD of Imaging
Mammographic CARE? longer
manufactured

FFDM/DBT Machines by FDA Approved Accreditation Bodies

Manufacturer - American College of Radiology
- State of Arkansas
Approved
Currently
Manufactured
LOTS of CHOICES - State of Iowa
- State of Texas
iCRco 1 1
Siemens
• Why so many choices? • Only approved to accredit FFDM
5 3
Philips 2 2 • None are approved to accredit DBT
Fuji 5 5

Konica Minolta 1 1 • Many HARD Decisions

Agfa 1 1
• What to BUY?
Giotto
• When to BUY?
1 1
GE 4 3
Planmed 2 2
Hologic 6 5
• What is the Difference?
Hologic DBT
• Image Quality?
1 1
Carestream 1 1
Fischer 1 0 • Acquisition Speed? Table updated
31 26
• Detector Technology? 20 Apr 2013
20
 FDA Approved Accreditation Bodies FDA Approved Accreditation Bodies
- American College of Radiology - American College of Radiology
- State of Arkansas - State of Arkansas
- State of Iowa - State of Iowa
- State of Texas - State of Texas

• Only approved to accredit FFDM • Only approved to accredit FFDM

• None are approved to accredit DBT • None are approved to accredit DBT

• SIA and SAR – not approved as • 31 FDA Approved Systems

accreditation body for all FFDM • 25 FFDM
• 5 no longer manufactured
• 1 DBT
- Must go to FDA or ACR for
approval of those machines that • 11 different manufacturers
the State Accreditation bodies
are not approved to accredit.

Table updated Table updated

20 Apr 2013 20 Apr 2013
21 22

FFDM Under MQSA

FDA Certification Extension
• Current: ACR or State AB has a separate FDA approval
for each FFDM system to be accredited • For FFDM systems for which there is not an approved
Accreditation Body FDA will provide a Certification
Extension of Accredited Sites
• Initial FDA Equipment Approval: If no approved AB
• System is exempt from Accreditation requirement
then MUST get FDA Certification Extension
• Facility must submit a request to FDA – in Writing – to
obtain Certification Extension
• DBT: MUST be accredited for 2D imaging
– Sectra FFDM system currently is in this category for those
facilities with accreditation from ACR or State of Iowa
– DBT component of Hologic 2D Selenia falls into this category

23 24
 FDA Certification Extension FDA Certification Extension
• For FFDM systems for which there is not an approved
Accreditation Body FDA will provide a Certification
• Processing time for Certification Extension: Extension of Accredited Sites
– 14 days
• System is exempt from Accreditation requirement
• FDA Contact: Go to FDA MQSAWeb Site
• Facility must submit a request to FDA – in Writing – to
obtain Certification Extension

– Sectra FFDM system currently is in this category for those

facilities with accreditation from ACR or State of Iowa
– DBT component of Hologic 2D Selenia falls into this category

25 26

FDA Certification Extension - DBT FDA Certification Extension - DBT

• Requirements 10 required elements for Certification Extension to include DBT
1. Facility Status Information
– Must have an accredited Hologic Selenia Dimensions 2D 2. Hologic Selenia Dimensions DBT Unit Identification
– Personnel 3. Hologic Selenia Dimensions DBT Digital Image Receptor Identification (if
• Physicians, Medical Physicists and Technologists interchangeable)
4. Identification of Printer for Hard Copy Interpretation (mandatory even for facilities
• must have 8 hrs of DBT Training performing only soft copy interpretation)
– NOTE: Digital Mammography training can not be used for 5. Final Interpretation Review Monitor Identification (if soft copy display is available)
DBT training 6. Phantom Identification
– Equipment 7. Personnel Qualifications
• Must have MEE prior to use Pre 11 Feb 2011 – Qualified under MQSA and reading DBT
Post 11 Feb 2011 –Qualified under MQSA AND have 8 hrs of training in DBT
Note: FFDM training can not be substituted for DBT
– QC Procedures 8. Report of Mammography Equipment Evaluation (MEE) (must have been conducted in
• Follow Manufacturer’s QC program accordance with 900.12(e)(10) within the 6 months prior to the request for use a
9. Hologic Selenia Dimensions DBT Manufacturer’s Quality Control Program
27
10. Signature 28
 1. Facility Status Information 2. Hologic Selenia Dimensions DBT Unit ID
a. Facility Name and FDA Facility ID Number a. Machine Manufacturer
b. FDA Certificate Expiration Date b. Machine Model
c. Current Accreditation Body for the Hologic Selenia Dimensions 2D unit c. Year of Manufacture
d. Accreditation Expiration Date d. Serial Number
e. Facility Contact Person for Hologic Selenia Dimensions DBT unit
f. Contact Person’s Title
g. Contact Person’s Telephone, Fax, E-mail
h. Facility Address
i. Facility Owner

29 30

3. Hologic Selenia Dimensions DBT Digital 4. Identification of Printer for Hard Copy
Image Receptor Identification Interpretation
a. Receptor Manufacturer (mandatory even for facilities performing only soft copy
b. Receptor Model interpretation)
c. Year of Manufacture
d. Serial Number (if applicable)
a. Printer Manufacturer
b. Printer Model
c. Year of Manufacture
d. Serial Number

31 32
 5. Final Interpretation Review Monitor
6. Phantom Identification
Identification
(if soft copy display is available) a. Phantom Manufacturer
b. Phantom Model
a. Monitor Manufacturer
b. Monitor Model
c. Year of Manufacture
d. Serial Number

33 34

7. Personnel Qualifications – Pre 11 Feb 2011 7. Personnel Qualifications – Pre 11 Feb 2011
• Interpreting Physicians • Radiologic Technologists
– PERSONNEL QUALIFICATIONS: INTERPRETING PHYSICIANS WHO – PERSONNEL QUALIFICATIONS: RADIOLOGIC TECHNOLOGISTS
ARE QUALIFIED TO INTERPRET DBT MAMMOGRAMS WHO ARE QUALIFIED TO PERFORM DBT MAMMOGRAMS

– (1) meet all the requirements of CFR 900.12(a)(1) "Mammography Quality – (1) meet all the requirements of 21 CFR 900.12(a)(2) "Mammography Quality
Standards; Final Rule" that became effective on April 28, 1999 AND Standards; Final Rule" that became effective on April 28, 1999 ; AND

– (2) began interpreting DBT mammograms prior to February 11, 2011 – (2) began performing DBT mammography examinations prior to February 11,
2011.

35 36
 7. Personnel Qualifications – Pre 11 Feb 2011 7. Personnel Qualifications – Post 11 Feb 2011
• Medical Physicists • Interpreting Physicians
– PERSONNEL QUALIFICATIONS: MEDICAL PHYSICISTS WHO ARE – PERSONNEL QUALIFICATIONS: INTERPRETING PHYSICIANS WHO
QUALIFIED TO PERFORM DBT SURVEYS ARE QUALIFIED TO INTERPRET DBT MAMMOGRAMS

– (1) meet all the requirements of 21 CFR 900.12(a)(3) "Mammography Quality – (1) meet all the requirements of 21 CFR 900.12(a)(1) "Mammography Quality
Standards; Final Rule" that became effective on April 28, 1999 ; AND Standards; Final Rule" that became effective on April 28, 1999 *;

– (2) began performing equipment evaluations and/or surveys of DBT – (2) began interpreting DBT mammograms after February 11, 2011; AND
mammography units prior to February 11, 2011.
– (3) have 8 hours of initial training in DBT Mammography*.
• Note: Full Field Digital Mammography training can not be used as a substitute for DBT
training.

37 38

7. Personnel Qualifications – Post 11 Feb 2011 7. Personnel Qualifications – Post 11 Feb 2011
• Radiologic Technologists • Medical Physicists
– PERSONNEL QUALIFICATIONS: RADIOLOGIC TECHNOLOGISTS – PERSONNEL QUALIFICATIONS: MEDICAL PHYSICISTS WHO ARE
WHO ARE QUALIFIED TO PERFORM DBT MAMMOGRAMS QUALIFIED TO PERFORM DBT SURVEYS

– (1) meet all the requirements of 21 CFR 900.12(a)(2) "Mammography Quality – (1) meet all the requirements of 21 CFR 900.12(a)(3) "Mammography Quality
Standards; Final Rule" that became effective on April 28, 1999 ; Standards; Final Rule" that became effective on April 28, 1999 ;

– (2) began performing DBT mammography examinations after February 11, – (2) began performing equipment evaluations and/or surveys of DBT
2011; AND mammography units after February 11, 2011; AND

– (3) have 8 hours of initial training in DBT Mammography. – (3) have 8 hours of initial training in DBT Mammography.
• Note: Full Field Digital Mammography training can not be used as a substitute for DBT • Note: Full Field Digital Mammography training can not be used as a substitute for DBT
training. training.
39 40
 8. Report of Mammography Equipment 8. Report of Mammography Equipment
Evaluation Evaluation
Report of Mammography Equipment Evaluation (MEE) (must have been conducted in b. The results of quality control tests as required under the following sections of the MQSA
accordance with 900.12(e)(10) within the 6 months prior to the request for use final regulations 21CFR 900.12(e):
approval)
(4)(iii) Compression Device Performance (5)(vii) X-Ray Field/Light field/Image
a. Statement that equipment performance, as required under the following sections of the receptor/Compression paddle alignment
(5)(i) Automatic Exposure Control
MQSA final regulation 21 CFR 900.12(b), is met:
Performance (5)(ix) System Artifacts
(1) Prohibited Equipment (6) Magnification (if applicable to the DBT system used) (5)(x) Radiation Output
(2) Specifically Designed for (7) Focal Spot Selection (5)(ii) Kilovoltage Peak Accuracy and (5)(xi) Decompression
Mammography Reproducibility (or alternative standards allowed for these
(8) Compression
(5)(iii) Focal Spot Condition (Resolution) requirements)
(3) Motion of Tube-Image Receptor (9) Technique Factor Selection and Display
Assembly (5)(iv) Beam Quality and Half-Value Layer (6) Quality Control Tests – Other
(10) Automatic Exposure Control Modalities
(4)(iii) Removable Grid (if applicable to the (5)(v) Breast Entrance Air Kerma and AEC
(14) Lighting (Facilities must perform all DBT
DBT system used) Reproducibility (if applicable to the
(if hard copy display is used for image DBT system used) manufacturer recommended quality control
(5) Beam Limitation and Light Fields evaluation) tests including the medical physicist’s tests
(5)(vi) Dosimetry for Soft Copy Display system)
(15) Film Masking Devices
(if hard copy display is used for image
evaluation) 41 42

8. Report of Mammography Equipment 9. Hologic Selenia Dimensions DBT Quality

Evaluation Control Program
c. The results of the phantom image quality tests, including a sample image a. Name of the Quality Control Manual
b. Year published
d. If any of the requirements in 8 a, b, or c are not met, submit documentation of c. Revision number, if not the original
successful corrective action d. Printing number, if not the original

e. If any of the requirements in 8 a or b are not performed, explain why the

requirement is not applicable

f. Date of the MEE

g. Name and address of the physicist(s) who performed the MEE

43 44
 10. Signature
Lead Interpreting Physician Attestation to Staff Personnel Qualifications

To the best of my knowledge and my belief, the information provided in this

Questions? 
document is true and correct. I understand that FDA may request additional
information to substantiate the statements made in the document. I understand
that knowingly providing false information in a matter within the jurisdiction
of an agency of the United States could result in criminal liability, punishable
by up to $10,000 fine and imprisonment of up to five years, or civil liability
under MQSA, or both.

Signature (Lead Interpreting Physician) __________________________ Talk 8

45
 Advanced Applications in
Advanced
Digital Imaging Applications in
Digital Breast Tomosynthesis Imaging Digital Imaging

14 March 2015 – Indianapolis, IN

Jerry A. Thomas, MS, FAAPM, DABR, CHP, DABSNM

Diagnostic Medical Physicist
Via Christi Hospitals Wichita, Wichita, KS

Digital Detector Advanced Applications

Keys to Advanced Applications
• Computer-Aided Detection and/or Dx
• Tomosynthesis

Fast Imaging •
•
Tele-Radiology
Dual-Energy Subtraction
• Digital Subtraction Angiography

Low Noise •
•
Low-Dose Fluoroscopic Positioning
Stereo / Computer Aided Localization
• Multi-modality Imaging
• Temporal Subtraction
• Image Pasting
• ...
 Computer-Aided Detection CAD Goals
Applications
• Automated detection
• Automated classification • Improve diagnosis
Current ImageCheckerTM • Standardize interpretation
Performance
• Flagged 85% of • Facilitate interpretation
overlooked cancers • Reduce reading load
• Flagged 83% of
diagnosed cancers • Improve positive predictive value
Increase the sensitivity for • Facilitate Image “triage” in Telemammography
cancer detection when
combined with the radiologist • Educational tool

M Calcification
a
s
Cluster Detection
s

D
e
t
e
c
t
i
o
n
 Calcification
Artificial Neural Network
Original image Cluster
Classification
Classification Mammogram Enhancement
Malignant

PURPOSE: Automated enhancement of contrast and/or

mammograms texture
74% chance
for cancer Goal: To improve display of clinically relevant
structures

Use portions of CAD algorithms to identify internal

image content for selection/generation of a
Automatic segmentation
individualized display look-up table

TC & Dynamic Range Management

Film/screen or Raw Unprocessed Image

Display Full Dynamic Range

Thickness Compensation

Improve contrast in glandular tissue while

preserving visualization of the whole breast
 Original

Findings Based Filtering

Enhanced

Improve visibility of findings.

Future Applications of CAD Future Applications of CAD

Telemammography and Remote Diagnosis Automatic sorting of patient Images

Purpose: To provide “immediate” remote expert Purpose: To sort patient images into “normal”
clinical interpretation / clinical management and “possibly abnormal/abnormal” categories
advice and guidance Goal: Reduce reading time required for normal
Goal: Real-time transmission and interpretation images so Radiologist can focus on abnormal or
of Dx quality mammograms from a remote site to potentially abnormal images.
a central “expert” site.
 Tele-
mammography/
Consultation “Coming Attractions”
• New X-ray tube Technology
• Contrast Enhanced DM
• Dual Energy Contrast Enhanced DM
• Spectral Imaging vs. Dual Energy
• Multi-modality Imaging

18

Carbon Nanotube X-ray Prototype Digital Subtraction Angiography

• Cancer must recruit vessels (angiogenesis)
to grow larger than 1 mm in size
• DSA provides an opportunity to detect early
stage cancers by looking for increased
vascularity
• Determination of cancer extent
• DSA may also be used for monitor treatment
of a tumor - especially treatment using the
new anti-angiogenesis drugs
• Two images may be sufficient

19
 DSA Leg DSA Leg
No Contrast Subtracted
Image

DSA Line Phantom Contrast Enhanced Digital

Digital Detector
Mammography- CEDM

• GE Senobright
• FDA approved for sale in the US
Neuro Angiography System
• Change filter and algorithms only
• No hardware change required
• Haven’t found uptake kinetics clinically helpful
• Dose 20% higher-1.5 mGy/view

Only filled 250 um line

 Temporal CEDM Temporal CEDM
Temporal CEDM CEDM
• 3-7 images at high kVp (45-49) Case image
• Dose/image typically 5x lower mask
62 year-old
compared to standard MX image

• Mask image before injection 5

30 60 180 300
Physical examination
0

Iodine injection
Non palpable lesion
Images courtesy of Dr Diekmann Technique :
Charité – Berlin, Germany Mo/Cu, 45kV,
100mAs Mammography
- 1 stellate opacity
- 1 round opacity

Conventional Mammograms

t = 0s t = 60s t = 120s t = 180s T 60 – T 0 T 120 – T 0 T 180 – T 0

Images courtesy of Dr Dromain,
Institut Gustave Roussy – Villejuif, France 26
25

Dual Energy CEDM Dual Energy CEDM

Dual-Energy CEDM CEDM Conventional CEDM
• 1 image at low kVp, 1 image at high kVp (45-49) Case 4 image Mammograms image
• low kVp image just before high kVp image Physical examination
low kVp high kVp Normal
image image

Mammography
0 120 120 +  Time (s)
Small mass
Iodine injection only visible on CC view
• Low confidence of
presence
Low Energy Spectrum Rh/Rh 28 kVp High Energy Spectrum Rh/Cu 44 kVp
• Global
classification
BIRADS 3

Ultrasonography RMLO 2 min RCC 4 min

normal
Conventional CEDM image (performed by referring physician)
Mammogram
Iodine Iodine
K-edge K-edge

Images courtesy of Dr Dromain, 27 Images courtesy of Dr Dromain, 28

Institut Gustave Roussy – Villejuif, France Institut Gustave Roussy – Villejuif, France
 Spectral imaging an alternative to
Spectral Imaging vs Dual-Energy MR?
Spectral imaging: Photon-counting:
dual exposures single exposure
Spectra after breast Spectrum after breast
14
18
12 16
Fluence (10 6phot/cm 2)

Fluence(10 6ph./cm 2)
10 14

12
8
Rh 10
6 8 Iodine
4 Cu 6 K-edge
4
2
2

0 0
0 10 20 30 40 10 15 20 25 30 35 40 45
Energy (keV) Energy (keV)

2nd
Non-overlapping spectra  better tissue cancellation threshold
Concentrated around K-edge  improved I-contrast
30
29

Low Dose Fluoroscopic or Ultrasound / X-ray Data Fusion

Radiographic Positioning Sagittal US

• Fluoroscopy - up to several frames per

second using radiographic detectors
– Biopsy and needle positioning

• Low dose radiographic exposures

– Positioning for any radiographic exam

• Digital Dose may be adjusted to clinical need

Mammogram Axial US Coronal US

 Thank you for your attention
Don’t let the Sharks BITE!!

My contact info: [email protected]

34

Just the Sail Fish FINAL COURSE BUSINESS

• Please complete Course Critique Sheets
• Please complete Attendee Profile

• Course Certificates - on the Table in back of room

• DRIVE CAREFULLY on your way HOME

THANK YOU FOR ATTENDING!!!
WE ENJOYED SEEING YOU AGAIN!
HAVE A WONDERFUL DAY!!