This free educational material is provided by

333 N. Commercial Street, Suite 100, Neenah, WI 54956

Breast Imaging
Donald L. Renfrew, MD

Breast cancer is the most frequent non-skin detect possible breast cancer. It then discusses
cancer diagnosis in women, with an estimated diagnostic studies.
192,370 new cases in 20091. Knowing what
diagnostic imaging tests are available, which test to Screening studies
order when, and what to do with the results Screening studies are usually chosen for a
presents a challenge to the primary care practitioner. combination of factors including relatively low cost
This chapter reviews three key concepts regarding and high sensitivity: the screening test should pick
breast imaging: up as much disease as possible, with the idea that
subsequent studies will provide more specificity
1. There are certain relatively widely accepted regarding the diagnosis.
rules about how to screen asymptomatic
Screening mammography
women, and how to image symptomatic Mammography remains the king of breast
women. imaging (Figure 1). It has been shown in multiple
2. Mammography is the mainstay of diagnosis, trials to reduce mortality in the screened population
frequently supplemented by ultrasound, by about 30%2. It’s the best screening test we have.
with MR playing a minor role. That being said, it has problems as a screening test:
3. Careful follow-up and handoff of the patient it is relatively insensitive, it involves ionizing
is critical for optimal patient care. radiation, it is at least somewhat painful for most
women, and it can be inconvenient. It also results in
a fair number of false positives, causing a lot of
RULES TO GUIDE BREAST IMAGING needless worry on the part of patients and driving
up the costs of medical care. If we had some
There are a few relatively widely accepted rules alternate method of early diagnosis – for example, a
regarding breast imaging that are helpful to know serum test for tumor markers – this would be a great
when ordering imaging studies. Breast imaging advantage. This may happen, but it hasn’t yet, so
studies may be divided into screening and we continue to do mammography.
diagnostic exams, and the rules differ for these two General recommendations are that women have
categories of exams. This chapter first covers screening commencing at age 40, and continue as
screening studies, done on asymptomatic patients to long as life expectancy is at least ten years 3. For

For additional free educational material regarding symptoms and imaging, please visit www.symptombasedradiology.com
Page 118 Breast Imaging

patients who have had a mother, sister, recommendation to return for an annual screening
grandmother, or aunt diagnosed at a young age mammography in one year, if the study is normal;
(prior to 40) with breast cancer, it is generally or 2) a recommendation for additional imaging
accepted that screening should begin at an age studies if the screening study is abnormal (see
earlier than 40. One commonly used rule is to start below). Usually, the additional imaging study is
screening at 5 years prior to the age of diagnosis of either additional mammography, with, for example,
cancer in the relative. spot compression or magnification views, or
Note that a screening mammography report will ultrasound evaluation. It is uncommon to proceed
usually contain one of two recommendations: 1) a directly to biopsy on the basis of a screening study.

Figure 1. Normal digital screening mammogram, mediolateral oblique (MLO) views. Modern digital mammography
technique shows exquisite detail of breast tissue allowing screening for malignancy. Note the inclusion of the pectoralis
muscle along the posterior margin of the study. Screening mammography usually includes both bilateral mediolateral
oblique views (shown) and craniocaudal views (not shown).
Chapter 9 Breast Imaging Page 119

Screening MRI compared to mammography, there are multiple

MR is more sensitive than mammography in the organizations, including the American Cancer
detection of breast cancer. The generally accepted Society, that advocate screening MRI for patients
sensitivity for MRI is over 90%, but it will miss small with a 20 – 25% lifetime risk of breast cancer 5.
cancers or areas of DCIS4. There are two major Patients will generally fall into this high risk
problems with breast MR, however: 1) specificity is category if they have a breast cancer gene (BRCA1
only in the 50-70% range secondary to false positives or BRCA2), or if they have close relatives with breast
from fibroadenomas and other benign lesions, and cancer. There are several online calculators which
2) cost. The false positives necessitate either biopsy will allow precise determination of cancer risk, for
or follow-up MR, both of which are also costly. example at: http://www.cancer.gov/bcrisktool/.
However, because of the increased sensitivity of MR

Figure 2. Abnormal screening mammogram, prompting recall of the patient for a diagnostic mammogram with additional
views showing normal tissue. A. Screening mammogram from 7-23-07 is normal. B. The patient’s left craniocaudal view
from 7-24-08 shows an apparent developing mass in the inner aspect (arrow). C. Spot compression study shows no
discrete mass but normal, although dense, breast tissue. D. Follow-up mammogram study of 8-3-09 is normal.
Page 120 Breast Imaging

Figure 3. Abnormal screening mammogram, prompting recall of the patient for a diagnostic mammogram with additional
views prompting biopsy. A. Screening cradiocaudal mammogram shows a small, dense cluster of calcifications (arrow).
The patient was recalled for a diagnostic mammogram. B. Spot magnification craniocaudal mammogram better shows
these calcifications (arrow), which demonstrate variable size. Stereotactic needle biopsy was performed, and the pathology
interpretation was an involuted fibroadenoma and focal ductal hyperplasia without atypia.

Screening ultrasound Diagnostic Studies

Ultrasound is presently not routinely used as a Screening mammography is done on asymptomatic
screening study, although the modality is patients with no known imaging abnormality.
undergoing evaluation as an adjunct (or possible Diagnostic mammography is performed when there is
replacement) to mammography, particularly in either an abnormality on a screening examination
patients with dense breasts6 7. (also known as a callback) or the patient has
symptoms. Ultrasound and MR may also be used as
Chapter 9 Breast Imaging Page 121

diagnostic studies, and again this usually occurs the results of that additional evaluation are clearly
either because of an abnormal screening benign (Figure 2), then the patient returns to a
examination or patient symptoms. yearly screening schedule. If the results of the
additional evaluation are not clearly benign, it may
Abnormal screening studies resulting in diagnostic be necessary to proceed with biopsy (Figure 3).
studies Ordering of studies and the decision to proceed with
Nowadays, most radiology departments handle biopsy should generally follow the radiologist’s
callbacks internally, with the department notifying recommendations.
the patient that additional evaluation is necessary. If

Figure 4. Infiltrating ductal carcinoma in a 39 year old woman with a breast mass found at breast self examination.
A. Right mediolateral oblique (MLO) diagnostic mammogram is normal. B. Left MLO diagnostic mammogram
demonstrates a large, dense mass (arrow).
Page 122 Breast Imaging

Figure 5. Infiltrating ductal carcinoma in a in a 75 year old woman with a palpable lesion found at clinical breast
examination. A. Craniocaudal mammogram spot compression views (following initial full field exam) shows a subtle lesion
of the right breast by the radio-opaque marker. B. Breast ultrasound demonstrates a hypodense, shadow-casting, irregular
lesion (arrows) worrisome for malignancy. Biopsy revealed infiltrating ductal carcinoma.

Breast lump or focal pain, age > 35 the patient’s breast at the time of the physical
Generally speaking, lumps and focal pain should examination, prior to sending the patient for
be worked up in a similar fashion. Lumps found at imaging. This way, the technologist will know
clinical breast examination (CBE) or breast self where to place the radiographic marker. The
examination (BSE) are both evaluated using the mammogram should include both breasts if the
same algorithm, although lumps found at CBE are asymptomatic breast has not undergone
more likely to be malignant than those found at BSE3. mammography in the past year.
For patients over the age of 35 with a lump or If the mammogram fails to show, or does not
focal pain, mammography should be performed first adequately characterize, an explanatory lesion at the
(Figure 4), with ultrasound to follow if necessary location of the palpable abnormality or focal pain,
(Figure 5)8. The mammogram should be scheduled the patient will typically proceed to ultrasound
as a “diagnostic” (not a “screening”) study, and the (Figure 5). The ultrasound study is done because
technologist will typically put a radiographic ultrasound will demonstrate some malignant lesions
marker at the location of the palpable lump or area that escape detection on mammography, and
of maximum pain. If the palpable abnormality is ultrasound may better demonstrate some lesions
subtle on clinical exam, particularly if the patient which are poorly demonstrated on mammograms.
cannot feel the abnormality herself, it is best to mark
Chapter 9 Breast Imaging Page 123

discharge with a small, specially designed blunt

catheter and injecting contrast material into the duct
in a retrograde fashion.

Figure 6. Ruptured epidermal inclusion cyst in a 23 year

old woman with a palpable abnormality. The ultrasound
exam demonstrates a hypoechoic lesion. Since the
abnormality did not represent a simple cyst or a
prominent but normal ridge of breast tissue, it was
surgically excised, and the pathologic diagnosis was a
ruptured epidermal inclusion cyst.
Figure 7. Intraductal papilloma in a 55 year old woman
Breast lump or focal pain, age < 35 with bloody single pore nipple discharge. A standard
For patients under the age of 35, ultrasound mammogram (not shown) failed to demonstrate any
should be performed first, followed by cause of discharge. The catheter tip is at the nipple, and
mammography if necessary8 (Figure 6). These contrast material fills the dilated duct which has a filling
women have denser breasts and a lower pretest defect (arrow), found at pathology to represent a benign
probability of having a malignancy with a higher intraductal papilloma.
likelihood that the palpable lesion is a cyst or benign
but prominent ridge of breast tissue. Therefore, it What NOT to image
makes sense to perform ultrasound first, followed For patients with diffuse pain, or with bilateral,
by mammography if the ultrasound provides no multipore discharge, no imaging beyond standard,
explanation but there is still a strong suspicion of a age-appropriate screening mammography is useful.
lesion.

Breast discharge MAMMOGRAPHY IS THE PRIMARY

Multipore, blood-negative, expressed-only IMAGING MODALITY, SUPPLEMENTED
discharge is best categorized as benign physiologic BY ULTRASOUND, WITH A SMALL ROLE
discharge, and is not worrisome for malignancy.
FOR MRI
Such discharge may require medical evaluation and
medical work-up9. As noted above, mammography is the screening
Unilateral, single pore discharge, particularly if modality of choice, and is the most frequently used
bloody, is worrisome and needs further evaluation10. diagnostic modality as well. Since mammography is
The first imaging step in evaluation is usually the primary method of breast evaluation in both the
ultrasound, particularly in patients under 30, to screening and diagnostic roles, how do radiology
detect dilated ducts and focal masses. This may be departments know that they are doing a good job?
followed by mammography, and if these tests do not
provide a definitive answer, then a ductogram (also Mammography quality assurance
known as a “galactogram”) may provide a diagnosis Mammography quality assurance has evolved
(Figure 7). The ductogram is performed by through the years in part because of work done by
cannulating the nipple pore that shows the
Page 124 Breast Imaging

the American College of Radiology (ACR), and in

part because of legislation known as the Screening mammography metrics
Mammography Quality Standards Act (MQSA). The BI-RADS categories allow relatively easy
evaluation of large amounts of data. The United
ACR Lexicon and BI-RADS States Department of Health and Human Services
In response to complaints about the variability of has created benchmarks or metrics for community
mammography reports, the American College of radiologists which may be calculated with the use of
Radiology developed a lexicon of mammography these categories14. Of these metrics, the most useful
terms11. As it turns out, this lexicon has not been are probably recall rate, biopsy rate, biopsy yield,
universally adopted although the ACR publishes an and cancer detection rate*. Note that these metrics
excellent handbook illustrating these terms12. At the can be calculated from the BI-RADS codes given to
same time they developed the lexicon, the ACR also the screening studies and follow-up on those specific
developed the Breast Imaging Reporting and Data studies where biopsy was recommended (which
System or BI-RADS (Table 1). BI-RADS is a quality should represent about 1% of the screening exam
assessment tool, but it is also directly clinically results). Also note that the cancer prevalence rate is
relevant, because it forces the radiologist to reduce different in those women undergoing screening
the mammogram result to a single number which mammography for the very first time than the
determines the next step in patient care. As noted cancer incidence in patients undergoing annual
by Sistrom and Langlotz writing on the topic of screening. While the general recall rate is set at 10%,
improving radiology reporting “One of the greatest the recall rate is also different between women
benefits of the entire BI-RADS initiative arises from undergoing their first study (where 10% is a
the mandated forced choice between clinically reasonable figure) versus women undergoing
diagnostic categories”.13 Each BI-RADS category is repeated screening (where 3% or 4% is probably
linked to a specific next step, making management more reasonable15). However, the prevalence data
unambiguous (Table 1) (exams for first-time screening mammograms) and
incidence data (exams with prior studies for
BI-RADS Description Next Step comparison) are often pooled in evaluating
Category mammography quality assurance. An example for
0 Incomplete Return for additional data in one small community hospital is presented
assessment imaging or obtain in Table 2. Note that in the “Analog” column, the
prior comparison data represents a two year period and demonstrates
studies adequate performance with respect to the recall rate,
1 Negative Return for routine biopsy yield, and cancer detection rate. The biopsy
screening rate is higher than the benchmark (1.6% versus 1.0%),
2 Benign Return for routine but given that the biopsy yield is still significantly
findings screening above the benchmark, this is acceptable.
3 Probably Return for initial
benign short term follow-up
findings (usually 6 months)
4 Suspicious Biopsy should be
abnormality considered
5 Highly Appropriate action
suggestive of should be taken
malignancy
6 Known Appropriate action *
Other benchmarks or metrics include: sensitivity of at least 85%, prevalent
malignancy should be taken cancer detection rate of 0.6 – 1.0%, incident cancer rate of 0.2 – 0.4%, less
Table 1. BI-RADS categories with descriptions and than 25% with positive lymph node metastases at the time of diagnosis,
mean tumor size of less than 1.5 cm, at least 30% DCIS or invasive cancer <
resulting actions. 1 cm; at least 50% stage 0 or 1 cancer.
Chapter 9 Breast Imaging Page 125

Figure 8. Superior detail with digital mammography. Film-screen (A.) versus digital (B.) normal screening mammogram.
Note the superior visualization of both the central, dense parenchymal tissue, and also the peripheral, predominantly fatty
breast tissue, with digital mammography.
demonstrated a significant reduction of the call-back
Digital mammography rate for digital mammography versus film
Digital mammography uses different technology mammography in return patients (2.4 % versus
than analog mammography, and provides greater 3.6%)15, without a decrease in the rate of malignancy
detail, particularly in the superficial tissues and in detection. This reduction in call-back rate is
dense breasts (Figure 8). Image data is collected, important, since women being recalled for
stored, and displayed electronically rather than with additional views may experience significant,
film. Digital mammography shows greater ongoing anxiety17. When comparing the data at the
sensitivity for detection of cancer in women with same small community hospital (Table 2 again), note
dense breasts, as seen in women under the age of 50 that following implementation of digital
or women who are premenopausal and mammography, there was a decrease in the recall
perimenopausal . In addition, Sala et al
16 rate (in this table, both initial and return recall rates
Page 126 Breast Imaging

are pooled), similar to the Sala et al study, while the Ultrasound is used frequently and MR is used
biopsy rate (the percentage of screening patients occasionally for problem solving
eventually undergoing biopsy) fell, while the biopsy Ultrasound is used to distinguish normal
yield (the likelihood that a given biopsy tissue and cysts from solid masses. Ultrasound
demonstrated cancer) increased. The cancer can be used to evaluate palpable lesions, focal
detection rate showed a statistically insignificant, tender spots, or lesions seen on mammography
small decrease. or MRI requiring further work-up. Lesions seen
Wherever mammography is done, these metrics on ultrasound may be placed into one of four
should be available. If, as is often the case, there is basic categories, two of which typically require
more than one available location for mammography no further evaluation or work-up. If a normal
service, these metrics provide a handy way to ridge of breast tissue or a cyst explain the
compare the locations. abnormality, then no further evaluation is
Metric Benchmark Analog Digital necessary (Figure 9). If a solid lesion is identified,
(5742) (6128) this typically requires biopsy, although some
Recall Rate <10% 6.3% 4.6% solid lesions are relatively typical of benign
Biopsy Rate <1% 1.6% 1.1% lesions such as fibroadenomas (Figure 10),
Biopsy Yield >25% 30.4% 40.6%
Cancer 0.2 – 0.5% 0.49% 0.46%
Detection Rate
Table 2. Mammography data from Door County
Memorial Hospital, Sturgeon Bay, WI. Rates are for
screening mammograms performed in a community
hospital, with historical comparison between Analog and
Digital examinations.

Figure 9. Cyst in a 47 year old woman with an abnormal screening mammogram, with ultrasound demonstrating a benign
cyst at the location of the new lesion. A. Craniocaudal screening mammogram (cropped) shows a circumscribed
hypodense lesion of the inferior right breast (arrow). B. Ultrasound (right) demonstrates a simple cyst at the location of
the lesion (arrow), and no further work-up required.
Chapter 9 Breast Imaging Page 127

while others are quite suspicious for malignancy MRI shows malignancy as a mass or enhancing
(Figure 5). Some women would rather have even tissue
benign appearing solid lesions removed rather than In addition to its role as a screening tool in
followed, whereas other women would rather avoid patients with a high risk of breast malignancy, MR
biopsy. Malignant appearing solid lesions should may be used to evaluate the ipsilateral breast for
certainly undergo biopsy. mammographically occult disease, the contralateral
breast in a patient with known malignancy (Figure
11), and, on occasion, to better characterize a lesion
seen on mammography or ultrasound18.

Figure 10. Fibroadenoma in a 48 year old woman with an abnormal screening mammogram, with ultrasound
demonstrating a solid, benign appearing lesion at the location of the abnormality. A. Screening mammogram shows a
circumscribed isodense mass (arrow) in the right breast. B. Breast ultrasound (with a different magnification) shows an
oblong, sharply marginated, isodense solid mass without shadowing (arrow), characteristic of a fibroadenoma. The patient
wanted the lesion removed despite its benign appearance, and pathology confirmed a fibroadenoma.
Page 128 Breast Imaging

Figure 11. Infiltrating ductal carcinoma in a 41 year old woman with MRI demonstrating additional disease not detected at
initial surgery. A. Right craniocaudal screening mammogram shows a mass in the lateral breast (arrow). B. US of the
breast confirms a malignant appearing mass (arrow). C. Contrast enhanced MR examination of the breasts done following
excision of an infiltrating ductal carcinoma demonstrates the operative site (arrow). Abnormal tissue extends from the
biopsy site to the nipple (double arrow). Imaging directed biopsy of this region demonstrated multifocal high-grade DCIS
beyond the margins of the initial surgery.

CAREFUL HANDOFFS ENSURE of missed reports “falling through the cracks”. With
the development of BI-RADS, the responsibility to
THE BEST PATIENT CARE
notify the patients to return for additional views or
ultrasound examination largely shifted from the
Careful handoffs from practitioner to practitioner referring physician to the radiology department.
prevent the tragic mistakes that can happen because Many of these same departments also schedule and
Chapter 9 Breast Imaging Page 129

perform ultrasound-directed biopsy or stereotactic

biopsy, whereas at other locations biopsies are
performed by surgeons. Local referral patterns, as
well as preference for ultrasound directed biopsy,
core needle biopsy with mammographic guidance,
and biopsy using needle localization techniques
vary with locations as well as patient
circumstances18. Regardless of the local distribution
of duties, it is imperative that all involved
physicians know the pathway the patient is taking.
In the unfortunate event of a bad outcome, all
parties will likely be held liable, so it is good to have
redundancy built into the system in those instances
when a patient is sent to biopsy. There are various
mechanisms to achieve this, such as keeping a list of
patients you know are going to biopsy and setting
up automated forwarding of pathology results to
you from the laboratory. Making sure the patient
knows who to call, and that she should call someone,
if she does not hear about her results, adds an
additional layer of security. Do not assume the
patient will call if she hears nothing: there are
patients who, hoping for the best, will assume that
“no news is good news”.

SUMMARY

Breast imaging usually follows several widely

accepted rules about when and how to screen
patients, and when and how to image the breast
symptoms of a palpable mass or focal breast pain.
Mammography remains the mainstay of diagnosis,
frequently supplemented by ultrasound with MR
typically playing a minor role. Careful follow-up
and handoff of the patient are critical for optimal
patient care.
Page 130 Breast Imaging

15 Sala M et al. Implementation of digital

REFERENCES mammography in a population-based breast cancer
screening program: effect of screening round on
1 American Cancer Society “Cancer Facts & Figures recall rate and cancer detection. Radiology 2009;
2009” available at 252:31-39.
http://www.cancer.org/downloads/STT/500809web. 16 Pisano ED et al. Diagnostic Performance of Digital

pdf versus Film Mammography for Breast-Cancer

2 Nystrom L et al. Long-term effects of Screening NEJM 2005; 353:1773-1783.
mammography: updated review of the Swedish 17 Brewer NT et al. Systematic review: the long-term

Randomized trials. Lancet 2002; 359:909-919. effects of false-positive mammograms. Ann Intern
3 Fletcher SW. Screening for breast cancer. Med 2007; 146:502-510
UpToDate, accessed 7/27/09. 18 Esserman LJ, Joe BN. Diagnostic evaluation and

4 Macura KJ et al. Patterns of enhancement on breast initial staging work-up of women with suspected
MR images: interpretation and imaging pitfalls. breast cancer. UpToDate, accessed 7/28/09.
RadioGraphics 2006; 26:1719-1734
5 Saslow D et al. American Cancer Society

guidelines for breast screening with MRI as an

adjunct to mammography. CA Cancer J Clin 2007;
57:75-89
6 Fajardo LL. Question and answer. AJR 2003;

181:1715
7 Berg WA. Rationale for a trial of screening breast

ultrasound: American College of Radiology Imaging

Network (ACRIN) 6666. AJR 2003; 180:1225-1227.
8 Fletcher SW, Barton MB. Primary care evaluation

of breast lumps. UpToDate, accessed 7/29/09.

9 Golshan M, Iglehart D. Nipple discharge.

UpToDate, accessed 7/30/09.

10 Leis HP. Management of nipple discharge. World

J. Surg 1989; 13:736-42.

11 D’Orsi CJ. The American College of Radiology

mammography lexicon: an initial attempt to

standardize terminology. AJR 1996; 166:779-780.
12 Breast Imaging Reporting and Data System (BI-

RADS) Atlas. 4th Edition. American College of

Radiology, Reston, VA 2003.
13 Sistrom CL, Langlotz CP. A framework for

improving radiology reporting. J Am Coll Radiol

2005; 2:159-167.
14 Bassett LW et al. Quality determinants of

mammography. Rockville (MD): Agency for Health

Care Policy and Research, Public Health Service. US
Department of Health and Human Services; 1994.
Clinical Practice guideline #13: AHCPR Publication
#95-0632.