- Email: [email protected]
Utility of routine use of breast ultrasound following contrast-enhanced spectral mammography
Clinical Radiology 73 (2018) 908.e11e908.e16
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Utility of routine use of breast ultrasound following contrast-enhanced spectral mammography E. Klang a, b, A. Krosser b, *, M.M. Amitai a, b, V. Sorin b, O. Halshtok Neiman a, b, A. Shalmon a, b, M. Gotlieb a, b, M. Sklair-Levy a, b a b
Department of Diagnostic Imaging, Sheba Medical Center, Ramat Gan, Israel Tel Aviv University, Sackler Faculty of Medicine, Tel Aviv, Israel
article in formation Article history: Received 27 December 2017 Received in revised form 12 May 2018 Accepted 24 May 2018
AIM: To evaluate whether breast ultrasound (US) is routinely indicated following contrastenhanced spectral mammography (CESM). MATERIALS AND METHODS: Consecutive screening and diagnostic CESM examinations with concurrent breast US were collected retrospectively (May 2012 to February 2016). Radiologists assigned a separate Breast Imaging-Reporting and Data System (BIRADS) score for CESM and for US. BIRADS scores were grouped into three categories: normal/benign appearing (BIRADS 1, 2); probably benign, short-term follow-up (BIRADS 3); or suspicious appearing (BIRADS 0, 4, 5). Patients with a suspicious-appearing lesion in either US or CESM underwent biopsy. The associations between malignant pathology with either suspicious-appearing CESM or suspicious-appearing US were calculated. The sensitivities and specificities of CESM and US were analysed. RESULTS: Eighty-seven lesions were biopsied, 37 (43%) biopsies were malignant and 50 (57%) were benign. Although suspicious-appearing CESM was associated with malignant biopsies (p<0.0001), suspicious-appearing US was not (p¼0.985). Among 37 malignant biopsies, CESM had a sensitivity of 97% (36/37 lesions), compared to 92% (34/37 lesions) with US. None of the malignant biopsies were normal/benign appearing with CESM. One case of follow-up CESM was suspicious-appearing at US and proved to be malignant on biopsy. The specificity of CESM was 40%, which was significantly higher than US at 8%. CONCLUSION: When CESM is suspicious appearing, subsequent US and biopsy is appropriate. With a CESM BIRADS 3, correlation with US is suggested. If the CESM is benign appearing, the routine use of US is questionable, as it may lead to unnecessary benign biopsies. Ó 2018 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction
* Guarantor and correspondent: A. Krosser, The Chaim Sheba Medical Center, Derech Sheba 2, Ramat Gan, Israel. Tel.: þ972 58 6314570; fax: þ972 3 5357315. E-mail address: [email protected] (A. Krosser).
Breast cancer is the most commonly diagnosed cancer among women and ranks second among all cancers in female mortality.1 Early detection improves survival, making early identification of breast cancer imperative for
https://doi.org/10.1016/j.crad.2018.05.031 0009-9260/Ó 2018 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
908.e12
E. Klang et al. / Clinical Radiology 73 (2018) 908.e11e908.e16
lengthening survival time.2 Currently, mammography is the only approved breast cancer screening test, and has been shown to decrease mortality in breast cancer patients.3e5 Despite mammography being the reference standard for diagnosing breast cancer, factors such as breast density have been shown to lower its sensitivity.6 Mammography with supplementary ultrasound (US) is recommended in women with dense breast parenchyma, as concurrent use has been shown to detect more malignant lesions in this population.7,8 Both mammography and US can be used separately for breast cancer detection as well. Concomitant US and mammography does result in an increased false-positive (FP) rate.9 US has high sensitivity for finding lesions in the breast, but low specificity for malignancy, leading to an increased number of benign biopsies.10,11 Contrast-enhanced spectral mammography (CESM) is a relatively novel technique for breast cancer assessment, approved for use by the US Food and Drug Administration (FDA) in 2011. CESM provides anatomical and functional imaging of breast tissue, combining the standard twodimensional (2D) digital mammography performed following the intravenous injection of an iodine-based contrast agent. Contrast material accumulates in metabolically active tissue with increased vascular supply, and is therefore, useful in identifying cancerous lesions.12,13 Unlike conventional mammography, CESM is mostly unaffected by dense breast tissue, and is preferred over unenhanced mammography in women with dense breast tissue.14 CESM has been shown to have a slightly higher sensitivity than conventional mammography, and the increase in sensitivity is amplified in denser breast tissue.15 Although US is indicated following mammography in women with dense breasts, very little is reported about the utility of breast US in addition to CESM. The aim of this study was to evaluate whether breast US is routinely indicated following CESM.
Materials and methods US and CESM technique An institutional review board approval was granted for this retrospective study. This review was a retrospective study with analysis of interpretations given at the time of the radiological studies. On searching the institutional database, 953 patients who underwent CESM between May 2012 and February 2016 were identified. Subsequent breast US examination was conducted for 1,669 breasts. The average age of the patients was 51.89 years. Previous medical history and additional data for patients included in this study are shown in Tables 1 and 2. The standard of care is to perform US after mammography in women with dense breast tissue. As CESM is a new technique, an attempt was wanted to ensure we do not miss findings in cases of a normal CESM, US was performed on the majority of high-risk patients regardless of their breast density. US examinations were always bilateral and of the whole-breast and were performed with a hand-held device.
Table 1 Previous medical history for patients in this study. History
n (%)
Known familial predisposition (breast, ovarian cancer) Known breast tumour (right breast) Known breast tumour (left breast) Known other tumour/metastasis S/P right lumpectomy S/P left lumpectomy S/P right mastectomy S/P left mastectomy Breast density BIRADSa BIRADS 0 BIRADS 1 BIRADS 2 BIRADS 3 BIRADS 4 NS
207 (21.7%) 37 (3.9%) 35 (3.7%) 13 (1.4%) 97 (10.2%) 105 (11%) 17 (1.8%) 11 (1.2%) 9 (1%) 7 (0.7%) 104 (11%) 776 (82.3%) 47 (5%) 10
BIRADS, Breast Imaging-Reporting and Data System; NS, not specified; s/p, status post. a The 10 NS were excluded when calculating percentages. Table 2 Indications for mammography. Indication
n (%)
Screening Mastitis Palpable lump right Palpable lump left Breast tenderness right Breast tenderness left Nipple discharge right/retraction Nipple discharge left/retraction Pre-op FNL MG known nodule follow-up Search for primary breast cancer
725 (76.1%) 4 (0.4%) 61 (6.4%) 83 (8.7%) 14 (1.5%) 29 (3%) 8 (0.8%) 5 (0.5%) 12 (1.3%) 18 (1.9%) 8 (0.8%)
FNL, fine-needle localisation; MG, mammography.
Although the standard of care in the time frame of the study was US following CESM examinations, a minority of the women did not undergo US for the following reasons: patients with recent US conducted in the previous 3 months; women who preferred not to undergo the recommended US; and patients with prior mastectomies only underwent unilateral US. All CESM studies were performed on a digital mammography system (Senographe Essential, GE Healthcare; Chalfont St Giles, UK) upgraded to enable the acquisition of dual-energy exposures. Low-energy exposures were obtained at 27e31 kVp with the use of molybdenum and rhodium targets and filters. High-energy exposures were acquired at 45e50 kVp using a molybdenum target with an aluminium and copper filter. Using an imageprocessing software algorithm, the two exposures were subtracted, generating two images: one low-energy image, providing maximum soft-tissue contrast, and one subtracted image displaying areas of contrast enhancement only. A dose of 1.5 ml/kg body weight non-ionic contrast agent (Iopamiro 370, Bracco S.p.A, Milano, Italy) was intravenously injected in the antecubital fossa, using an automated power injector (Medrad Mark V ProVis; Bayer HealthCare) at a flow rate of 3 ml/s, followed by a saline flush.
E. Klang et al. / Clinical Radiology 73 (2018) 908.e11e908.e16
All US examinations were conducted using the Sieman’s Acuson S2000 ultrasound system with a linear transducer 14-5Mhz or 18-6Mhz (Sieman’s Medical Solutions USA, Mountain View, CA, USA). After each examination, a Breast Imaging-Reporting and Data System (BIRADS) score was determined separately for US and CESM in each patient. BIRADS scores were separately grouped into three categories: normal/benign appearance (BIRADS 1, 2), probably benign, short-term follow-up (BIRADS 3), and suspicious appearance (BIRADS 0, 4, 5). For CESM interpretation, four radiologists each with 2e4 years of experience reading CESM assigned BIRADS scores. For US, all evaluations were performed by the same four radiologists each with 15e20 years of US reading experience, who subsequently assigned BIRADS scores. The radiologists performing US evaluations were not blinded to the CESM findings. Patients with a suspicious-appearing lesion in either US or CESM underwent biopsy of the breast. All biopsies were US-guided. Biopsies that were suspicious-appearing with CESM were US-guided even if the lesion did not appear suspicious with US. Imaging findings for lesions proven histologically to be breast cancer were considered true positive (TP) with BIRADS 0,4,5, and false negative (FN) with BIRADS 1,2 or BIRADS 3. Imaging findings for the lesions found to be histologically benign were considered true negative (TN) with BIRADS 1, 2 or BIRADS 3, and FP with BIRADS 0, 4, 5. Both BIRADS 3 and BIRADS 1, 2 (follow-up and normal/ benign appearance, respectively) were considered to be negative imaging findings because neither result warrants an immediate biopsy of the breast. For statistical analysis, the association between suspicious-appearing CESM (BIRADS 0, 4, 5) and suspiciousappearing US to malignant biopsy were calculated (Chisquare test), and odds ratios (ORs) were evaluated. The accuracies, sensitivities, and specificities for CESM and US were also analysed, and confidence intervals were calculated. McNemar’s test was used to evaluate the statistical difference between the accuracies of CESM and US.
908.e13
Results Of the 953 patients examined in this study, US and CESM were conducted for 1,669 total breasts. The demographics and clinical data of the study cohort are presented in Table 1. Indications for mammography are shown in Table 2. Some patients had more than one indication. Biopsies were performed for 87 lesions (5.2%), in which either US or CESM classified the findings as a suspicious-appearing lesion. Histological analysis of the biopsies showed that 37 lesions (43%) were malignant and 50 (57%) were benign. Histology of malignant lesions showed 24/37 (64.9%) cases of infiltrating ductal carcinoma (IDC), 6/37 (16.2%) cases of ductal carcinoma in situ (DCIS) and 7/37 (18.9%) cases of combined IDC and DCIS. Graphs showing a breakdown of CESM and US BIRADS scores for malignant biopsies can be seen in Fig 1a and for benign biopsies in Fig 1b. Among biopsies that were histologically determined to be malignant, CESM classified 36 (97%) lesions as suspicious appearing; this includes three patients who had a concurrent normal/benign appearing US. US classified 34 (92%) malignant lesions as suspicious appearing, with no normal/benign appearing CESM findings among these patients (both tests TP; see Fig 2). Only one US examination classified a malignant lesion as suspicious appearing with a subsequent follow-up (BIRADS 3) CESM (TP US with FN CESM; Fig 3). A more detailed BIRADS score breakdown among malignant lesions can be seen in Table 3. In the biopsies that were histologically determined to be benign, 20 (40%) images were normal/benign appearing or follow-up with CESM while suspicious appearing with US (TN CESM with FP US; Fig 4). US found that four (8%) benign lesions were normal/benign appearing or follow-up with a suspicious-appearing CESM (TN US with FP CESM). Both CESM and US were suspicious appearing in 26 (52%) lesions (both tests FP). A more detailed BIRADS score breakdown among benign lesions can be seen in Table 4. While suspicious-appearing CESM was associated with malignant biopsies (p<0.0001, OR¼24.0), suspiciousappearing US was not (p¼0.985, OR¼0.9). Table 5 shows the accuracies, sensitivities, specificities, and OR for CESM
Figure 1 Percentage of malignant lesions (a) and benign lesions (b) in each group of BIRADS categories. Comparison between CESM and US. Of the 87 lesions biopsied, 37 were histologically identified as breast cancer and 50 were found to be benign lesions. (a) Among malignant lesions, suspicious BIRADS findings are indicative of TP findings and normal/benign appearing or follow-up indicate FN. (b) Among benign lesions, normal/benign appearing or follow-up are considered TN and suspicious BIRADS findings are FP findings.
908.e14
E. Klang et al. / Clinical Radiology 73 (2018) 908.e11e908.e16
Figure 2 Mammography (a), CESM (b), and US (c) of a 58-year-old woman who underwent biopsy of the left breast, and has a histologically confirmed carcinoma. The carcinoma can be observed with CESM and US (BIRADS 4, 5).
and US based on the BIRADS categories and malignancy. CESM showed significantly higher accuracy than US (64.4% versus 44.7%, p¼0.003; Table 6). A graph showing the sensitivity and specificity for CESM and US can be seen in Electronic Supplementary Material Fig S1. For calculations both normal/benign appearing (BIRADS 1, 2) and follow-up (BIRADS 3) cases were considered TN for benign lesions and FN for malignant lesions as neither finding warrants an immediate biopsy.
Discussion Although CESM has been shown to be useful in screening for breast cancer, little is reported about concurrent US following CESM. In the present study, the role of US following CESM was examined among 953 patients with 87 biopsied lesions. Of the 87 lesions, 43% were malignant and 57% were benign. Sensitivity, specificity, and ORs for diagnosis of malignant versus benign biopsied lesions were calculated for both imaging methods. The sensitivity of CESM was 97%, higher than the 92% sensitivity of US. Specificity of CESM was 40%, significantly higher than 8% of US. The OR for malignant biopsies was 24 for CESM, compared to 0.9 for US. A study by Luczynska et al. comparing CESM and US found that CESM better detected breast cancer than breast US, with a CESM sensitivity of 100% compared to US
Table 3 Breast Imaging-Reporting and Data System score breakdown among patients with malignant biopsies.
Figure 3 CESM (a) and US (b) of a 66-year-old woman. This is the one case of malignancy where US detected BIRADS 4 and CESM BIRADS 3. Biopsy confirmed carcinoma.
Malignant group
US 1,2
US 3
US 4,5
CESM 1,2 CESM 3 CESM 0,4,5
3
0
0 1 33
US, ultrasound; CESM, contrast-enhanced spectral mammography.
E. Klang et al. / Clinical Radiology 73 (2018) 908.e11e908.e16
908.e15
Figure 4 CESM (a) and US (b) of a 55-year-old woman who underwent biopsy of the left breast. Initial CESM was BIRADS 2, but US was BIRADS 4 so ultrasound-guided biopsy of the left breast was performed (4c). Biopsy was benign.
Table 4 Breast Imaging-Reporting and Data System score breakdown among patients with benign biopsies. Benign group
US 1,2
US 3
US 4,5
CESM 1,2 CESM 3 CESM 0,4,5
2
2
13 7 26
Table 6 Table describing the accuracy of CESM and US.
CESM US
TP
FP
TN
FN
36 34
30 46
20 4
1 3
US, ultrasound; CESM, contrast-enhanced spectral mammography; TP, true positive; FP, false positive; TN, true negative; FN, false negative.
US, ultrasound; CESM, contrast-enhanced spectral mammography.
Table 5 Sensitivities, specificities, and odds ratios for CESM and US. Sensitivity % (95% CI) CESM 97.3% (85.8 e99.3) US 91.9% (78.1 e98.3)
Specificity % (95% CI)
Association with malignant biopsy (p-value)
40% (26.4 0.0001 e54.8) 8% (2.2e19.2) 0.985
US, ultrasound; CESM, contrast-enhanced spectral mammography.
Odds ratio 24 0.9
sensitivity of 97%; however, their results showed a less pronounced difference in specificity than the present results, with CESM specificity of 27%.16 The present results, which yielded a significant difference in specificity for malignant biopsies between the two tests, do compare to other publications that show variability between US and CESM, and confirm a higher rate of FP diagnoses using US.9,10 These findings suggest that US is not indicated for CESM with a BIRADS 1, 2. By using CESM alone 40% of the benign biopsies, which were due to FP US, could have potentially been spared. US may even be misleading in the
908.e16
E. Klang et al. / Clinical Radiology 73 (2018) 908.e11e908.e16
case of malignancy, as seen in 8% of malignant cases, where CESM findings were suspicious (BIRADS 4, 5), but US findings were not suspicious (BIRADS 1, 2, 3). The routine use of additional US examinations leads to both an increased cost, burdening the health system, and an increased rate of FP diagnoses.17 An increase in the cost of screening and an increase in FP diagnoses may deter patients from screening. The present study is limited by its small cohort of biopsied lesions. Some subgroups, such as BIRADS 3, contained a small number of patients, making it difficult to make conclusions. In addition, due to the small number of cases, the confidence intervals of the calculated sensitivities and specificities are wide. Another limitation of the study is its retrospective nature and the radiologists who assigned US BIRADS scores having access to the CESM report. Further prospective studies with a larger cohort would be beneficial. In conclusion, for CESM with a suspicious-appearing lesion (BIRADS 0, 4, 5), subsequent US and biopsy of the lesion is appropriate. When CESM shows follow-up (BIRADS 3), correlation with US is recommended and biopsy based on the result of the US. If the CESM is normal/benign appearing (BIRADS 1, 2), routine use of US is questionable as it may lead to unnecessary biopsies and unnecessary costs.
Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.crad.2018.05.031.
References 1. Torre LA, Bray F, Siegel RL, et al. Global cancer statistics, 2012. CA Cancer J Clin 2015;65(2):87e108. 2. Youlden DR, Cramb SM, Dunn NA, et al. The descriptive epidemiology of female breast cancer: an international comparison of screening, incidence, survival and mortality. Cancer Epidemiol 2012;36(3):237e48.
3. Tabar L, Vitak B, Chen HH, et al. Beyond randomized controlled trials: organized mammographic screening substantially reduces breast carcinoma mortality. Cancer 2001;91(9):1724e31. 4. Madsen EL, Berg WA, Mendelson EB, et al, Investigators for AP. Anthropomorphic breast phantoms for qualification of investigators for ACRIN Protocol 6666. Radiology 2006;239(3):869e74. 5. Kerlikowske K, Grady D, Rubin SM, et al. Efficacy of screening mammography. A meta-analysis. JAMA 1995;273(2):149e54. 6. Hersh MR, Marla R. Imaging the dense breast. Appl Radiol 2004; 33(1):22e7. 7. Okello J, Kisembo H, Bugeza S, et al. Breast cancer detection using sonography in women with mammographically dense breasts. BMC Med Imaging 2014;14:41. 8. Ohuchi N, Suzuki A, Sobue T, et al. Sensitivity and specificity of mammography and adjunctive ultrasonography to screen for breast cancer in the Japan Strategic Anti-cancer Randomized Trial (J-START): a randomised controlled trial. Lancet 2016;387(10016):341e8. 9. Kuhl CK, Schrading S, Leutner CC, et al. Mammography, breast ultrasound, and magnetic resonance imaging for surveillance of women at high familial risk for breast cancer. J Clin Oncol 2005;23(33):8469e76. 10. Berg WA, Bandos AI, Mendelson EB, et al. Ultrasound as the primary screening test for breast cancer: analysis from ACRIN 6666. J Natl Cancer Inst 2016;108(4). 11. Berg WA, Blume JD, Cormack JB, et al. Combined screening with ultrasound and mammography vs mammography alone in women at elevated risk of breast cancer. JAMA 2008;299(18):2151e63. 12. Fallenberg EM, Schmitzberger FF, Amer H, et al. Contrast-enhanced spectral mammography vs. mammography and MRI d clinical performance in a multi-reader evaluation. Eur Radiol 2017;27(7):2752e64. 13. Thibault F, Balleyguier C, Tardivon A, et al. Contrast enhanced spectral mammography: better than MRI? Eur J Radiol 2012;81(Suppl. 1):S162e4. 14. Smith A. The principles of contrast mammography. Hologic white paper WP-00084e001. 2014 https://pdfs.semanticscholar.org/56ba/d03f5f62 c390d8f770c5c2d961b769b9f583.pdf. 15. Diekmann F, Freyer M, Diekmann S, et al. Evaluation of contrastenhanced digital mammography. Eur J Radiol 2011;78(1):112e21. 16. Luczynska E, Heinze S, Adamczyk A, et al. Comparison of the mammography, contrast-enhanced spectral mammography and ultrasonography in a group of 116 patients. Anticancer Res 2016; 36(8):4359e66. 17. Fayanju OM, Kraenzle S, Drake BF, et al. Perceived barriers to mammography among underserved women in a breast health center outreach program. Am J Surg 2014;208(3):425e34.
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Utility of routine use of breast ultrasound following contrast-enhanced spectral mammography E. Klang a, b, A. Krosser b, *, M.M. Amitai a, b, V. Sorin b, O. Halshtok Neiman a, b, A. Shalmon a, b, M. Gotlieb a, b, M. Sklair-Levy a, b a b
Department of Diagnostic Imaging, Sheba Medical Center, Ramat Gan, Israel Tel Aviv University, Sackler Faculty of Medicine, Tel Aviv, Israel
article in formation Article history: Received 27 December 2017 Received in revised form 12 May 2018 Accepted 24 May 2018
AIM: To evaluate whether breast ultrasound (US) is routinely indicated following contrastenhanced spectral mammography (CESM). MATERIALS AND METHODS: Consecutive screening and diagnostic CESM examinations with concurrent breast US were collected retrospectively (May 2012 to February 2016). Radiologists assigned a separate Breast Imaging-Reporting and Data System (BIRADS) score for CESM and for US. BIRADS scores were grouped into three categories: normal/benign appearing (BIRADS 1, 2); probably benign, short-term follow-up (BIRADS 3); or suspicious appearing (BIRADS 0, 4, 5). Patients with a suspicious-appearing lesion in either US or CESM underwent biopsy. The associations between malignant pathology with either suspicious-appearing CESM or suspicious-appearing US were calculated. The sensitivities and specificities of CESM and US were analysed. RESULTS: Eighty-seven lesions were biopsied, 37 (43%) biopsies were malignant and 50 (57%) were benign. Although suspicious-appearing CESM was associated with malignant biopsies (p<0.0001), suspicious-appearing US was not (p¼0.985). Among 37 malignant biopsies, CESM had a sensitivity of 97% (36/37 lesions), compared to 92% (34/37 lesions) with US. None of the malignant biopsies were normal/benign appearing with CESM. One case of follow-up CESM was suspicious-appearing at US and proved to be malignant on biopsy. The specificity of CESM was 40%, which was significantly higher than US at 8%. CONCLUSION: When CESM is suspicious appearing, subsequent US and biopsy is appropriate. With a CESM BIRADS 3, correlation with US is suggested. If the CESM is benign appearing, the routine use of US is questionable, as it may lead to unnecessary benign biopsies. Ó 2018 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction
* Guarantor and correspondent: A. Krosser, The Chaim Sheba Medical Center, Derech Sheba 2, Ramat Gan, Israel. Tel.: þ972 58 6314570; fax: þ972 3 5357315. E-mail address: [email protected] (A. Krosser).
Breast cancer is the most commonly diagnosed cancer among women and ranks second among all cancers in female mortality.1 Early detection improves survival, making early identification of breast cancer imperative for
https://doi.org/10.1016/j.crad.2018.05.031 0009-9260/Ó 2018 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
908.e12
E. Klang et al. / Clinical Radiology 73 (2018) 908.e11e908.e16
lengthening survival time.2 Currently, mammography is the only approved breast cancer screening test, and has been shown to decrease mortality in breast cancer patients.3e5 Despite mammography being the reference standard for diagnosing breast cancer, factors such as breast density have been shown to lower its sensitivity.6 Mammography with supplementary ultrasound (US) is recommended in women with dense breast parenchyma, as concurrent use has been shown to detect more malignant lesions in this population.7,8 Both mammography and US can be used separately for breast cancer detection as well. Concomitant US and mammography does result in an increased false-positive (FP) rate.9 US has high sensitivity for finding lesions in the breast, but low specificity for malignancy, leading to an increased number of benign biopsies.10,11 Contrast-enhanced spectral mammography (CESM) is a relatively novel technique for breast cancer assessment, approved for use by the US Food and Drug Administration (FDA) in 2011. CESM provides anatomical and functional imaging of breast tissue, combining the standard twodimensional (2D) digital mammography performed following the intravenous injection of an iodine-based contrast agent. Contrast material accumulates in metabolically active tissue with increased vascular supply, and is therefore, useful in identifying cancerous lesions.12,13 Unlike conventional mammography, CESM is mostly unaffected by dense breast tissue, and is preferred over unenhanced mammography in women with dense breast tissue.14 CESM has been shown to have a slightly higher sensitivity than conventional mammography, and the increase in sensitivity is amplified in denser breast tissue.15 Although US is indicated following mammography in women with dense breasts, very little is reported about the utility of breast US in addition to CESM. The aim of this study was to evaluate whether breast US is routinely indicated following CESM.
Materials and methods US and CESM technique An institutional review board approval was granted for this retrospective study. This review was a retrospective study with analysis of interpretations given at the time of the radiological studies. On searching the institutional database, 953 patients who underwent CESM between May 2012 and February 2016 were identified. Subsequent breast US examination was conducted for 1,669 breasts. The average age of the patients was 51.89 years. Previous medical history and additional data for patients included in this study are shown in Tables 1 and 2. The standard of care is to perform US after mammography in women with dense breast tissue. As CESM is a new technique, an attempt was wanted to ensure we do not miss findings in cases of a normal CESM, US was performed on the majority of high-risk patients regardless of their breast density. US examinations were always bilateral and of the whole-breast and were performed with a hand-held device.
Table 1 Previous medical history for patients in this study. History
n (%)
Known familial predisposition (breast, ovarian cancer) Known breast tumour (right breast) Known breast tumour (left breast) Known other tumour/metastasis S/P right lumpectomy S/P left lumpectomy S/P right mastectomy S/P left mastectomy Breast density BIRADSa BIRADS 0 BIRADS 1 BIRADS 2 BIRADS 3 BIRADS 4 NS
207 (21.7%) 37 (3.9%) 35 (3.7%) 13 (1.4%) 97 (10.2%) 105 (11%) 17 (1.8%) 11 (1.2%) 9 (1%) 7 (0.7%) 104 (11%) 776 (82.3%) 47 (5%) 10
BIRADS, Breast Imaging-Reporting and Data System; NS, not specified; s/p, status post. a The 10 NS were excluded when calculating percentages. Table 2 Indications for mammography. Indication
n (%)
Screening Mastitis Palpable lump right Palpable lump left Breast tenderness right Breast tenderness left Nipple discharge right/retraction Nipple discharge left/retraction Pre-op FNL MG known nodule follow-up Search for primary breast cancer
725 (76.1%) 4 (0.4%) 61 (6.4%) 83 (8.7%) 14 (1.5%) 29 (3%) 8 (0.8%) 5 (0.5%) 12 (1.3%) 18 (1.9%) 8 (0.8%)
FNL, fine-needle localisation; MG, mammography.
Although the standard of care in the time frame of the study was US following CESM examinations, a minority of the women did not undergo US for the following reasons: patients with recent US conducted in the previous 3 months; women who preferred not to undergo the recommended US; and patients with prior mastectomies only underwent unilateral US. All CESM studies were performed on a digital mammography system (Senographe Essential, GE Healthcare; Chalfont St Giles, UK) upgraded to enable the acquisition of dual-energy exposures. Low-energy exposures were obtained at 27e31 kVp with the use of molybdenum and rhodium targets and filters. High-energy exposures were acquired at 45e50 kVp using a molybdenum target with an aluminium and copper filter. Using an imageprocessing software algorithm, the two exposures were subtracted, generating two images: one low-energy image, providing maximum soft-tissue contrast, and one subtracted image displaying areas of contrast enhancement only. A dose of 1.5 ml/kg body weight non-ionic contrast agent (Iopamiro 370, Bracco S.p.A, Milano, Italy) was intravenously injected in the antecubital fossa, using an automated power injector (Medrad Mark V ProVis; Bayer HealthCare) at a flow rate of 3 ml/s, followed by a saline flush.
E. Klang et al. / Clinical Radiology 73 (2018) 908.e11e908.e16
All US examinations were conducted using the Sieman’s Acuson S2000 ultrasound system with a linear transducer 14-5Mhz or 18-6Mhz (Sieman’s Medical Solutions USA, Mountain View, CA, USA). After each examination, a Breast Imaging-Reporting and Data System (BIRADS) score was determined separately for US and CESM in each patient. BIRADS scores were separately grouped into three categories: normal/benign appearance (BIRADS 1, 2), probably benign, short-term follow-up (BIRADS 3), and suspicious appearance (BIRADS 0, 4, 5). For CESM interpretation, four radiologists each with 2e4 years of experience reading CESM assigned BIRADS scores. For US, all evaluations were performed by the same four radiologists each with 15e20 years of US reading experience, who subsequently assigned BIRADS scores. The radiologists performing US evaluations were not blinded to the CESM findings. Patients with a suspicious-appearing lesion in either US or CESM underwent biopsy of the breast. All biopsies were US-guided. Biopsies that were suspicious-appearing with CESM were US-guided even if the lesion did not appear suspicious with US. Imaging findings for lesions proven histologically to be breast cancer were considered true positive (TP) with BIRADS 0,4,5, and false negative (FN) with BIRADS 1,2 or BIRADS 3. Imaging findings for the lesions found to be histologically benign were considered true negative (TN) with BIRADS 1, 2 or BIRADS 3, and FP with BIRADS 0, 4, 5. Both BIRADS 3 and BIRADS 1, 2 (follow-up and normal/ benign appearance, respectively) were considered to be negative imaging findings because neither result warrants an immediate biopsy of the breast. For statistical analysis, the association between suspicious-appearing CESM (BIRADS 0, 4, 5) and suspiciousappearing US to malignant biopsy were calculated (Chisquare test), and odds ratios (ORs) were evaluated. The accuracies, sensitivities, and specificities for CESM and US were also analysed, and confidence intervals were calculated. McNemar’s test was used to evaluate the statistical difference between the accuracies of CESM and US.
908.e13
Results Of the 953 patients examined in this study, US and CESM were conducted for 1,669 total breasts. The demographics and clinical data of the study cohort are presented in Table 1. Indications for mammography are shown in Table 2. Some patients had more than one indication. Biopsies were performed for 87 lesions (5.2%), in which either US or CESM classified the findings as a suspicious-appearing lesion. Histological analysis of the biopsies showed that 37 lesions (43%) were malignant and 50 (57%) were benign. Histology of malignant lesions showed 24/37 (64.9%) cases of infiltrating ductal carcinoma (IDC), 6/37 (16.2%) cases of ductal carcinoma in situ (DCIS) and 7/37 (18.9%) cases of combined IDC and DCIS. Graphs showing a breakdown of CESM and US BIRADS scores for malignant biopsies can be seen in Fig 1a and for benign biopsies in Fig 1b. Among biopsies that were histologically determined to be malignant, CESM classified 36 (97%) lesions as suspicious appearing; this includes three patients who had a concurrent normal/benign appearing US. US classified 34 (92%) malignant lesions as suspicious appearing, with no normal/benign appearing CESM findings among these patients (both tests TP; see Fig 2). Only one US examination classified a malignant lesion as suspicious appearing with a subsequent follow-up (BIRADS 3) CESM (TP US with FN CESM; Fig 3). A more detailed BIRADS score breakdown among malignant lesions can be seen in Table 3. In the biopsies that were histologically determined to be benign, 20 (40%) images were normal/benign appearing or follow-up with CESM while suspicious appearing with US (TN CESM with FP US; Fig 4). US found that four (8%) benign lesions were normal/benign appearing or follow-up with a suspicious-appearing CESM (TN US with FP CESM). Both CESM and US were suspicious appearing in 26 (52%) lesions (both tests FP). A more detailed BIRADS score breakdown among benign lesions can be seen in Table 4. While suspicious-appearing CESM was associated with malignant biopsies (p<0.0001, OR¼24.0), suspiciousappearing US was not (p¼0.985, OR¼0.9). Table 5 shows the accuracies, sensitivities, specificities, and OR for CESM
Figure 1 Percentage of malignant lesions (a) and benign lesions (b) in each group of BIRADS categories. Comparison between CESM and US. Of the 87 lesions biopsied, 37 were histologically identified as breast cancer and 50 were found to be benign lesions. (a) Among malignant lesions, suspicious BIRADS findings are indicative of TP findings and normal/benign appearing or follow-up indicate FN. (b) Among benign lesions, normal/benign appearing or follow-up are considered TN and suspicious BIRADS findings are FP findings.
908.e14
E. Klang et al. / Clinical Radiology 73 (2018) 908.e11e908.e16
Figure 2 Mammography (a), CESM (b), and US (c) of a 58-year-old woman who underwent biopsy of the left breast, and has a histologically confirmed carcinoma. The carcinoma can be observed with CESM and US (BIRADS 4, 5).
and US based on the BIRADS categories and malignancy. CESM showed significantly higher accuracy than US (64.4% versus 44.7%, p¼0.003; Table 6). A graph showing the sensitivity and specificity for CESM and US can be seen in Electronic Supplementary Material Fig S1. For calculations both normal/benign appearing (BIRADS 1, 2) and follow-up (BIRADS 3) cases were considered TN for benign lesions and FN for malignant lesions as neither finding warrants an immediate biopsy.
Discussion Although CESM has been shown to be useful in screening for breast cancer, little is reported about concurrent US following CESM. In the present study, the role of US following CESM was examined among 953 patients with 87 biopsied lesions. Of the 87 lesions, 43% were malignant and 57% were benign. Sensitivity, specificity, and ORs for diagnosis of malignant versus benign biopsied lesions were calculated for both imaging methods. The sensitivity of CESM was 97%, higher than the 92% sensitivity of US. Specificity of CESM was 40%, significantly higher than 8% of US. The OR for malignant biopsies was 24 for CESM, compared to 0.9 for US. A study by Luczynska et al. comparing CESM and US found that CESM better detected breast cancer than breast US, with a CESM sensitivity of 100% compared to US
Table 3 Breast Imaging-Reporting and Data System score breakdown among patients with malignant biopsies.
Figure 3 CESM (a) and US (b) of a 66-year-old woman. This is the one case of malignancy where US detected BIRADS 4 and CESM BIRADS 3. Biopsy confirmed carcinoma.
Malignant group
US 1,2
US 3
US 4,5
CESM 1,2 CESM 3 CESM 0,4,5
3
0
0 1 33
US, ultrasound; CESM, contrast-enhanced spectral mammography.
E. Klang et al. / Clinical Radiology 73 (2018) 908.e11e908.e16
908.e15
Figure 4 CESM (a) and US (b) of a 55-year-old woman who underwent biopsy of the left breast. Initial CESM was BIRADS 2, but US was BIRADS 4 so ultrasound-guided biopsy of the left breast was performed (4c). Biopsy was benign.
Table 4 Breast Imaging-Reporting and Data System score breakdown among patients with benign biopsies. Benign group
US 1,2
US 3
US 4,5
CESM 1,2 CESM 3 CESM 0,4,5
2
2
13 7 26
Table 6 Table describing the accuracy of CESM and US.
CESM US
TP
FP
TN
FN
36 34
30 46
20 4
1 3
US, ultrasound; CESM, contrast-enhanced spectral mammography; TP, true positive; FP, false positive; TN, true negative; FN, false negative.
US, ultrasound; CESM, contrast-enhanced spectral mammography.
Table 5 Sensitivities, specificities, and odds ratios for CESM and US. Sensitivity % (95% CI) CESM 97.3% (85.8 e99.3) US 91.9% (78.1 e98.3)
Specificity % (95% CI)
Association with malignant biopsy (p-value)
40% (26.4 0.0001 e54.8) 8% (2.2e19.2) 0.985
US, ultrasound; CESM, contrast-enhanced spectral mammography.
Odds ratio 24 0.9
sensitivity of 97%; however, their results showed a less pronounced difference in specificity than the present results, with CESM specificity of 27%.16 The present results, which yielded a significant difference in specificity for malignant biopsies between the two tests, do compare to other publications that show variability between US and CESM, and confirm a higher rate of FP diagnoses using US.9,10 These findings suggest that US is not indicated for CESM with a BIRADS 1, 2. By using CESM alone 40% of the benign biopsies, which were due to FP US, could have potentially been spared. US may even be misleading in the
908.e16
E. Klang et al. / Clinical Radiology 73 (2018) 908.e11e908.e16
case of malignancy, as seen in 8% of malignant cases, where CESM findings were suspicious (BIRADS 4, 5), but US findings were not suspicious (BIRADS 1, 2, 3). The routine use of additional US examinations leads to both an increased cost, burdening the health system, and an increased rate of FP diagnoses.17 An increase in the cost of screening and an increase in FP diagnoses may deter patients from screening. The present study is limited by its small cohort of biopsied lesions. Some subgroups, such as BIRADS 3, contained a small number of patients, making it difficult to make conclusions. In addition, due to the small number of cases, the confidence intervals of the calculated sensitivities and specificities are wide. Another limitation of the study is its retrospective nature and the radiologists who assigned US BIRADS scores having access to the CESM report. Further prospective studies with a larger cohort would be beneficial. In conclusion, for CESM with a suspicious-appearing lesion (BIRADS 0, 4, 5), subsequent US and biopsy of the lesion is appropriate. When CESM shows follow-up (BIRADS 3), correlation with US is recommended and biopsy based on the result of the US. If the CESM is normal/benign appearing (BIRADS 1, 2), routine use of US is questionable as it may lead to unnecessary biopsies and unnecessary costs.
Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.crad.2018.05.031.
References 1. Torre LA, Bray F, Siegel RL, et al. Global cancer statistics, 2012. CA Cancer J Clin 2015;65(2):87e108. 2. Youlden DR, Cramb SM, Dunn NA, et al. The descriptive epidemiology of female breast cancer: an international comparison of screening, incidence, survival and mortality. Cancer Epidemiol 2012;36(3):237e48.
3. Tabar L, Vitak B, Chen HH, et al. Beyond randomized controlled trials: organized mammographic screening substantially reduces breast carcinoma mortality. Cancer 2001;91(9):1724e31. 4. Madsen EL, Berg WA, Mendelson EB, et al, Investigators for AP. Anthropomorphic breast phantoms for qualification of investigators for ACRIN Protocol 6666. Radiology 2006;239(3):869e74. 5. Kerlikowske K, Grady D, Rubin SM, et al. Efficacy of screening mammography. A meta-analysis. JAMA 1995;273(2):149e54. 6. Hersh MR, Marla R. Imaging the dense breast. Appl Radiol 2004; 33(1):22e7. 7. Okello J, Kisembo H, Bugeza S, et al. Breast cancer detection using sonography in women with mammographically dense breasts. BMC Med Imaging 2014;14:41. 8. Ohuchi N, Suzuki A, Sobue T, et al. Sensitivity and specificity of mammography and adjunctive ultrasonography to screen for breast cancer in the Japan Strategic Anti-cancer Randomized Trial (J-START): a randomised controlled trial. Lancet 2016;387(10016):341e8. 9. Kuhl CK, Schrading S, Leutner CC, et al. Mammography, breast ultrasound, and magnetic resonance imaging for surveillance of women at high familial risk for breast cancer. J Clin Oncol 2005;23(33):8469e76. 10. Berg WA, Bandos AI, Mendelson EB, et al. Ultrasound as the primary screening test for breast cancer: analysis from ACRIN 6666. J Natl Cancer Inst 2016;108(4). 11. Berg WA, Blume JD, Cormack JB, et al. Combined screening with ultrasound and mammography vs mammography alone in women at elevated risk of breast cancer. JAMA 2008;299(18):2151e63. 12. Fallenberg EM, Schmitzberger FF, Amer H, et al. Contrast-enhanced spectral mammography vs. mammography and MRI d clinical performance in a multi-reader evaluation. Eur Radiol 2017;27(7):2752e64. 13. Thibault F, Balleyguier C, Tardivon A, et al. Contrast enhanced spectral mammography: better than MRI? Eur J Radiol 2012;81(Suppl. 1):S162e4. 14. Smith A. The principles of contrast mammography. Hologic white paper WP-00084e001. 2014 https://pdfs.semanticscholar.org/56ba/d03f5f62 c390d8f770c5c2d961b769b9f583.pdf. 15. Diekmann F, Freyer M, Diekmann S, et al. Evaluation of contrastenhanced digital mammography. Eur J Radiol 2011;78(1):112e21. 16. Luczynska E, Heinze S, Adamczyk A, et al. Comparison of the mammography, contrast-enhanced spectral mammography and ultrasonography in a group of 116 patients. Anticancer Res 2016; 36(8):4359e66. 17. Fayanju OM, Kraenzle S, Drake BF, et al. Perceived barriers to mammography among underserved women in a breast health center outreach program. Am J Surg 2014;208(3):425e34.