Received: 15 May 2017 | Revised: 23 September 2017 | Accepted: 2 October 2017

DOI: 10.1111/tbj.13040

ORIGINAL ARTICLE

Incidence and outcomes of incidental breast lesions detected

on cross-sectional imaging examinations

Eniola Falomo MD1 | Roberta M. Strigel MD, MS2,3,4 | Richard Bruce MD2 |
Alejandro Munoz del Rio PhD5 | Catherine Adejumo MBBS6 | Frederick Kelcz MD, PhD2

1
Russell H. Morgan Department of
Radiology and Radiologic Science, Johns Abstract
Hopkins School of Medicine, Baltimore, The aim of this study was to determine the frequency and outcomes of incidental
MD, USA
2 breast lesions detected on nonbreast specific cross-sectional imaging examinations.
Department of Radiology, University of
Wisconsin School of Medicine and Public A retrospective review of the medical records was performed to identify all patients
Health, Madison, WI, USA
without a known history of breast cancer, who had an incidentally discovered breast
3
Department of Medical Physics, University
of Wisconsin, Madison, WI, USA lesion detected on a nonbreast imaging examination performed at our institution
4
Carbone Cancer Center, University of between September 2008 and August 2012 for this IRB-approved, HIPAA compliant
Wisconsin, Madison, WI, USA
study. Outcomes of the incidental lesions were determined by follow-up with dedi-
5
Epic Systems, Verona, WI, USA
6
cated breast imaging (mammography, breast ultrasound, and/or breast MRI) or
Emory University School of Public Health,
Atlanta, GA, USA results of biopsy, if performed. Imaging modality of detection, imaging features,
patient age, patient location at the time of the nonbreast imaging examination, type
Correspondence
Eniola Falomo, The Russell H. Morgan of follow-up, and final outcome were recorded. Rates of malignancy were also calcu-
Department of Radiology and Radiological lated, and comparison was made across the different cross-sectional imaging modali-
Science, Johns Hopkins School of Medicine,
Baltimore, MD, USA. ties. Kruskal-Wallis and Fisher’s exact tests were used to identify factors associated
Email: [email protected] with an increased rate of malignancy. Logistic regression was used to model the risk
Funding information of malignancy as a function of continuous predictors (such as patient age or lesion
Dr. Strigel reports receiving grant funding size); odds ratios and 95% confidence intervals were obtained. A total of 292
from the Radiological Society of North
America, and the Wisconsin Women’s patients with incidental breast lesions were identified, 242 of whom had incidental
Health Foundation lesions were noted on computed tomography (CT) studies, 25 on magnetic resonance
imaging (MRI), and 25 on positron emission tomography (PET). Although most of the
incidental breast lesions were detected on CT examinations, PET studies had the
highest rate of detection of incidental breast lesions per number of studies per-
formed (rate of incidental breast lesion detection on PET studies was 0.29%, com-
pared to 0.10% for CT and 0.01% for MRI). Of the 121 of 292 (41%) patients who
received dedicated breast imaging work-up at our institution, 40 of 121 (33%) under-
went biopsy and 25 of 121 (21%) had malignancy. There was a significantly increased
rate of malignancy in older patients (odds ratio: 1.05, 95% CI: 1.02-1.093; P = .006).
Additionally, patients with PET-detected incidental breast lesions had a significantly
higher rate of malignancy (55%), compared to patients with CT-detected (35%) and
MRI-detected (8%) incidental breast lesions (P = .038). The rate of malignancy upon
follow-up of incidental breast lesions detected on nonbreast imaging examinations in

Based on a presentation at the Radiological Society of North America 2013 annual meeting,
Chicago, IL, USA.

Breast J. 2018;1–6. wileyonlinelibrary.com/journal/tbj © 2018 Wiley Periodicals, Inc. | 1

2 | FALOMO ET AL.

this retrospective study was 21%, supporting the importance of emphasizing further
work-up of all incidentally detected breast lesions with dedicated breast imaging.
Additionally, we found that PET examinations had the highest rate of detection of
incidental breast lesions and the highest rate of malignancy, which suggests that PET
examinations may be more specific for predicting the likelihood of malignancy of
incidental breast lesions, compared to CT and MRI.

KEYWORDS
breast cancer, breast imaging, breast incidentalomas, incidental breast lesions

1 | INTRODUCTION studies, and 8687 (2.0%) were PET examinations. We performed a

retrospective review of the electronic medical record using keyword
The use of cross-sectional imaging examinations has become increas- search to identify all patients without a known history of breast can-
ingly common, with computed tomography (CT), magnetic resonance cer, who had a breast lesion incidentally discovered on a cross-sec-
imaging (MRI), and positron emission tomography (PET) now widely tional nonbreast imaging examination performed at our institution
used for various clinical indications.1 This increased use of radiologic between September 1, 2008 and August 31, 2012.
imaging studies sometimes results in the detection of incidental find-
ings in parts of the body that were not the intended focus of clinical
2.3 | Data collection and statistical analysis
concern.1 Cross-sectional imaging studies (particularly those evaluat-
ing the thorax and upper abdomen) often include portions of the For all incidental breast lesions, lesion size, imaging modality of detec-
breast tissue and could result in the detection of unexpected breast tion, patient age, gender, and patient location at the time of the non-
abnormalities. breast imaging examination (ie, emergency room, inpatient, or
Previous studies have looked at the incidence and outcomes of outpatient), indication for the examination, type of follow-up, and final
incidental breast lesions detected on CT examinations,2-7 those found outcome were retrospectively reviewed and recorded. Outcomes of the
on MRI examinations,8 and those seen on PET examinations.9-15 incidental breast lesions were determined by follow-up with dedicated
However, to our knowledge, there are no known published studies breast imaging (mammography, breast ultrasound, and/or breast MRI)
evaluating the outcomes of incidental breast lesions detected across and results of biopsy, if performed. All data were manually entered into
different cross-sectional imaging modalities. an Excel spreadsheet (Microsoft Corporation, Redmond, WA).
The purpose of this retrospective study was to evaluate and The data were analyzed to identify the number of patients who
compare the incidence and outcomes of breast lesions that were underwent follow-up imaging at our institution, those who under-
incidentally detected on cross-sectional imaging studies of differing went biopsy and those who were subsequently diagnosed with
modalities that were performed for nonbreast-related clinical indica- breast malignancies. The frequency of detecting incidental lesions
tions. was calculated for each imaging modality. Rates of malignancy were
also calculated, and comparison was made across the different cross-
sectional imaging modalities. Kruskal-Wallis and Fisher’s exact tests
2 | MATERIALS AND METHODS
were used to identify factors associated with an increased rate of
malignancy. Logistic regression was used to model the risk of malig-
2.1 | Study approval
nancy as a function of continuous predictors (such as patient age or
This study was approved by our institutional review board, and a lesion size); odds ratios and 95% confidence intervals were obtained.
waiver of informed patient consent was obtained. The study was A value of P < .05 was considered significant. R 3.1.0 (R Develop-
compliant with Health Insurance Portability and Accountability Act ment Core Team 2014) was used for the statistical analysis.16
regulations.

3 | RESULTS
2.2 | Lesion inclusion criteria
3.1 | Patient demographics and overview
A review of the electronic medical record was performed to identify
the number of nonbreast cross-sectional imaging studies performed A total of 292 patients with incidental breast lesions were identified.
at our institution from September 1st, 2008 to August 31st, 2012. The mean patient age was 59.6 years (range: 14-100 years). The
During this time period, a total of 429 425 cross-sectional nonbreast majority of the patients (97%) were female. Two hundred and seven
imaging examinations were performed at our institution, 253 602 (71%) were outpatients, while 49 (17%) were inpatients, and 36
(59.0%) of which were CT examinations, 167 136 (38.9%) were MRI (12%) were Emergency Department (ED) patients.
FALOMO ET AL. | 3

T A B L E 1 List of the types of cross-sectional imaging The method of incidental breast lesion detection was as follows:
examinations that incidental lesions were detected on 242 patients (83%) with CT-detected lesions, 25 patients (8.5%) with
Examination type Frequency % MRI-detected lesions, and 25 patients (8.5%) with PET-detected
CT studies lesions. A detailed listing of the specific cross-sectional imaging stud-
CT abdomen/Pelvis 27 9.2 ies on which the incidental breast lesions were discovered is pro-

CT abdomen 3 1
vided in Table 1. The lesions ranged in size from 0.4 to 7.2 cm, with
an average lesion size of 1.6 cm.
CT angiogram of the coronary arteries 2 0.7
CT angiogram of the abdomen/pelvis 2 0.7
CT angiogram of the chest/abdomen/pelvis 7 2.4 3.2 | Frequency of incidental breast lesions across
CT angiogram of the chest 30 10.3 different cross-sectional imaging modalities
CT angiogram of the head and neck 1 0.3
For 25 patients, incidental breast lesions were detected on a PET
CT angiogram run-off 1 0.3 examination, representing a frequency of 0.29% (25 of 8687). In
CT chest/abdomen/pelvis 61 20.9 comparison, the frequency of CT-detected incidental breast lesions
CT chest 96 32.9 during the study period was 0.10% (242 of 253 602), and 0.01%
CT chest and abdomen 3 1 for MRI-detected incidental breast lesions (25 of 167 136).
CT flank pain 1 0.3
CT neck 3 1
3.3 | Work-up of incidental breast lesions
CT urography 2 0.7
CT colonography 3 1 Review of the electronic medical record revealed that 121 (41%)

MR studies of the 292 patients with incidental breast lesions had subsequent
follow-up with breast imaging work-up at our institution. Of these
MRA abdomen 2 0.7
121 patients, 40 underwent breast biopsy (33%) and 21 of these
MRA chest 2 0.7
(53%) were found to have biopsy-proven malignancy (See Fig-
MRA chest and abdomen 1 0.3
ures 1-4). Of the remaining 81 patients who underwent breast
MRCP/MRI abdomen 6 2
imaging work-up but did not have breast biopsy, 74 had negative
MRI abdomen 9 3.1
or benign findings on dedicated breast imaging evaluation. Three
MRI cardiac 2 0.7
of the 81 patients had suspicious findings on dedicated breast
MRI chest 1 0.3 imaging, but refused the recommended biopsy. The remaining four
MRI/MRA abdomen 1 0.3 were presumed to have malignancy: three known lymphoma
MRI/MRA chest 1 0.3 patients with findings concerning for lymphomatous involvement
PET/CT studies 25 8.6 of the breast and one known nonsmall cell lung cancer patient
with presumed metastatic disease to the breast (See Table 2).

Patients with incidental breast

lesions (n=292)

Patients with no known Patients with breast imaging

follow-up (n=171) follow-up (n=121)

Patients without biopsy Patients who underwent

(n=81) biopsy (n=40)

Patients with Patients with Patients with biopsy-

negative or suspicious findings Patients with proven malignancy
benign findings (BI-RADS 4 or 5) presumed (n=21)
(BI-RADS 1 or 2) who refused malignancy (n=4)
(n=74) biopsy (n=3)

FIGURE 1 Flowchart depicting the patients with incidental breast lesions and their outcomes
4 | FALOMO ET AL.

Colour online, B&W in print

(A)

(B)

F I G U R E 2 Axial image from a CT Colonography examination

(performed to screen for colon cancer) in a 50-y-old woman showed
a 1.5-cm mass in the inferior aspect of the left breast (see arrow),
which was subsequently proven to represent invasive ductal
carcinoma, following diagnostic work-up and biopsy

F I G U R E 4 A, B, Axial fused image from a PET/CT (left) and CT

(right) (performed to evaluate for metastatic disease) in a 65-y-old
woman with ovarian cancer showed a 2-cm hypermetabolic mass in
the inferior aspect of the right breast (see arrow), which was
subsequently proven to represent invasive lobular carcinoma,
following diagnostic work-up and biopsy [Color figure can be viewed
at wileyonlinelibrary.com]

T A B L E 2 List of the malignancies diagnosed following the

detection of incidental breast lesions
Diagnosis N
Invasive ductal carcinoma 16 (64%)
Invasive lobular carcinoma 2 (8%)
F I G U R E 3 Axial image of a magnetic resonance angiography of Adenocarcinoma of uncertain origin 1 (4%)
the chest (performed to evaluate for pulmonary embolism) in a 28-y-
Lymphoma 5 (20%)
old woman shows an enhancing mass in the medial aspect of the
left breast (see arrow). The mass was biopsied under ultrasound Metastasis (Nonsmall cell lung cancer) 1 (4%)
guidance, with pathology revealing a fibroepithelial lesion. She Total 25
subsequently underwent surgical excision, with final pathology result
N, number of cases.
of a benign phyllodes tumor

3.4 | Rate of malignancy

also a significant association between the rate of malignancy and the
Twenty-five of the 121 patients (21%) with known follow-up were modality of incidental breast lesion detection; six of the 11 patients
found to have malignancy (21 that were biopsy-proven): 16 (64%) with (55%) with PET-detected lesions were found to have malignancy,
invasive ductal carcinoma (IDC), two (8%) with invasive lobular carci- compared to 18 of the 52 patients (35%) with CT-detected lesions
noma (ILC), five (20%) with lymphoma (two biopsy-proven cases and and 1 of the 12 patients (8%) with MRI-detected lesions (P = .038).
three presumed), one (4%) with adenocarcinoma of uncertain origin, and There was no significant association between the rate of malig-
one (4%) with presumed metastasis (from nonsmall cell lung cancer). nancy and either the lesion size (P = .17), patient gender (P = 1), or
There was a significantly increased rate of malignancy in older the patient location at the time of the initial nonbreast cross-sec-
patients (odds ratio 1.05; 95% CI: 1.02-1.093; P = .006). There was tional imaging examination (P = .91).
FALOMO ET AL. | 5

4 | DISCUSSION up of such lesions with dedicated breast imaging. Additionally, our

study showed that on a per modality basis, incidental breast lesions
In this study, we evaluated the incidence and outcomes of incidental were most frequently detected on PET examinations, and PET-
breast lesions that were detected on nonbreast cross-sectional imag- detected incidental breast lesions had the highest likelihood of
ing studies. We found that the incidence and rate of malignancy of malignancy.
incidental breast lesions in our study cohort were overall similar to
those reported in previously published studies.2-15 However, to our
knowledge, there are no previous studies comparing the incidence 5 | CONCLUSION
and outcomes of incidental breast lesions detected on different
cross-sectional imaging studies. The rate of malignancy upon follow-up of incidental breast lesions
Although most of the incidental breast lesions in our study were detected on nonbreast imaging examinations in this retrospective
detected on CT, when evaluated from the point of view of likelihood study was significant (21% of patients with known breast imaging
of detection by a specific modality, PET studies had the highest rate follow-up at our institution), which highlights the importance of
of incidental breast lesions detection. This may be due to the fact emphasizing further work-up of all incidentally detected breast
that (i) PET studies typically include the entire breast region, whereas lesions with dedicated breast imaging.
inclusion of the breast region on CT and MRI examinations is highly
variable and depends on the specific indication and type of the CT
DISCLOSURES
or MRI examination, and (ii) PET examinations always include a
radiotracer (FDG) which may accumulate in a malignancy. MRI and The University of Wisconsin receives research support from GE
CT examinations may be performed without contrast, markedly Healthcare.
decreasing the likelihood of detecting malignancy, even if the breast
anatomy is included in the imaging field-of-view.
We also found that PET-detected incidental breast lesions had CONFLICT OF INTEREST
the highest likelihood of malignancy, compared to MRI-detected and
None.
CT-detected lesions. These results suggest that PET examinations
may be more specific for predicting the likelihood of malignancy of
REFERENCES
incidental breast lesions, compared to CT and MRI. However, these
findings do not obviate the need for dedicated breast imaging 1. Berland LL, Silverman SG, Gore RM, et al. Managing incidental find-
work-up of an incidental breast lesion, regardless of the modality of ings on abdominal CT: white paper of the ACR incidental findings
committee. J Am Coll Radiol. 2010;7:754-773.
detection.
2. Lin WC, Hsu HH, Li CS, et al. Incidentally detected enhancing breast
Another factor that was found to be significantly associated with lesions on chest computed tomography. Korean J Radiol. 2011;12:44-
increased rate of malignancy was patient age, as incidental breast 51.
lesions in older patients were significantly more likely to be malig- 3. Hussain A, Gordon-Dixon A, Almusawy H, Sinha P, Desai A. The inci-
dence and outcome of incidental breast lesions detected by com-
nant, which is consistent with the known increase in the risk of
puted tomography. Ann R Coll Surg Engl. 2010;92:124-126.
breast cancer with age.17 4. Porter G, Steel J, Paisley K, Watkins R, Holgate C. Incidental breast
We found that only 41% of the patients with incidentally masses detected by computed tomography: are any imaging features
detected breast lesions had any known breast imaging follow-up to predictive of malignancy? Clin Radiol. 2009;64:529-533.
5. Moyle P, Sonoda L, Britton P, Sinnatamby R. Incidental breast lesions
work-up the incidental findings. It is unclear how many of these
detected on CT: what is their significance? Br J Radiol. 2010;83:233-
patients received follow-up at a different institution and how many 240.
simply did not get any breast imaging work-up. Given that 21% of 6. Surov A, Fiedler E, Wienke A, Holzhausen HJ, Spielmann RP, Behr-
the patients with known breast imaging follow-up at our institution mann C. Intramammary incidental findings on staging computer
tomography. Eur J Radiol. 2012;81:2174-2178.
were found to have malignancy, dedicated breast imaging work-up
7. Monzawa S, Washio T, Yasuoka R, Mitsuo M, Kadotani Y, Hanioka
of all incidental breast lesions detected on nonbreast cross-sectional
K. Incidental detection of clinically unexpected breast lesions by
imaging examinations should be emphasized. computed tomography. Acta Radiol. 2013;54:374-379.
Our study has several limitations, one of which is the single-site 8. Prabhu V, Chhor CM, Ego-Osuala IO, Xiao JM, Hindman NM,
retrospective design of the study. Another limitation is the lack of Rosenkrantz AB. Frequency and outcomes of incidental breast
lesions detected on abdominal MRI over a 7-year period. AJR Am J
tracking of patients who did not undergo dedicated breast imaging
Roentgenol. 2017;208:107-113.
work-up at our institution, which prevents us from being able to 9. Litmanovich D, Gourevich K, Israel O, Gallimidi Z. Unexpected foci
account for any patients who may have presented to a different of 18F-FDG uptake in the breast detected by PET/CT: incidence and
institution for subsequent dedicated breast imaging work-up. clinical significance. Eur J Nucl Med Mol Imaging. 2009;36:1558-
1564.
In summary, we found that although incidental breast lesions
10. Beatty JS, Williams HT, Gucwa AL, et al. The predictive value of
are relatively rare, they may represent malignancy in up to 21% of incidental PET/CT findings suspicious for breast cancer in women
cases, which supports the importance of emphasizing further work- with non-breast malignancies. Am J Surg. 2009;198:495-499.
6 | FALOMO ET AL.

11. Kang BJ, Lee JH, Yoo IR, et al. Clinical significance of incidental find- 16. R Core Team (2014). R: A Language and Environment for Statistical
ing of focal activity in the breast at 18F-FDG PET/CT. AJR Am J Computing. Vienna, Austria: R Foundation for Statistical Computing.
Roentgenol. 2011;197:341-347. http://www.R-project.org.
12. Shin KM, Kim HJ, Jung SJ, et al. Incidental breast lesions identified by 17. Howlader N, Noone AM, Krapcho M, et al. (eds). SEER Cancer Statis-
(18)F-FDG PET/CT: which clinical variables differentiate between tics Review, 1975-2014. Bethesda, MD: National Cancer Institute.
benign and malignant breast lesions? J Breast Cancer. 2015;18:73-79.
13. Benveniste AP, Marom EM, Benveniste MF, Mawlawi O, Fox PS,
Yang W. Incidental primary breast cancer detected on PET-CT.
Breast Cancer Res Treat. 2015;151:261-268. How to cite this article: Falomo E, Strigel RM, Bruce R,
14. Chung A, Schoder H, Sampson M, Morrow M, Port E. Incidental Munoz del Rio A, Adejumo C, Kelcz F. Incidence and
breast lesions identified by 18F-fluorodeoxyglucose-positron emis-
outcomes of incidental breast lesions detected on cross-
sion tomography. Ann Surg Oncol. 2010;17:2119-2125.
15. Chae EY, Cha JH, Kim HH, et al. Analysis of incidental focal hyper- sectional imaging examinations. Breast J. 2018;00:1–6.
metabolic uptake in the breast as detected by 18F-FDG PET/CT: https://doi.org/10.1111/tbj.13040
clinical significance and differential diagnosis. Acta Radiol.
2012;53:530-535.