Contrast Enhanced Mammography in Routine Clinical Practice: Frequency and Malignancy Rates of Enhancing Otherwise Occult Findings

The purpose of this study is to report the frequency and malignancy rates of enhancing findings seen only on subtraction images on contrast enhanced mammography in a tertiary care breast practice. Consecutive review of CEM exams from December 2015 to May 2020 identified 4% of CEM examinations containing an enhancing finding seen only on subtraction images, ie, without low energy mammographic or sonographic correlate with PPV3 of 26%. Background: In routine clinical practice, contrast-enhanced mammography (CEM) examinations identify enhancing findings seen only on subtraction images that have no low-energy mammographic or sonographic correlate. The purpose of this study is to report the frequency and malignancy rates of enhancing findings seen only on subtraction images in a tertiary care breast imaging practice. Materials and methods: Consecutive review of CEM exams from December 2015 to May 2020. Chi square tests were used to determine associations between cancer diagnosis and clinical characteristics of enhancing findings seen only on subtraction images, P <.05 indicating a statistically significant difference. Results: Four percent (100/2464) of CEM examinations identified 108 enhancing findings seen only on subtraction images. Twenty of those CEM enhancing findings were directly managed as a multifocal disease. Of those further evaluated with MR, 23% (19/78) with associated MR correlates were treated surgically as presumed multicentric or multifocal disease following multidisciplinary review. The remaining 76% (59/78) of enhancing findings were seen only on subtraction images, these included: 20% (12/59) and downgraded to benign on MR 80% (47/59) with suspicious findings which underwent MR vacuum assisted breast biopsy yielding: 26% (12/47) malignancy, 9% (4/47) high risk, and 66% (31/47) benign diagnoses. Conclusion: Enhancing findings seen on subtraction only CEM images are seen in 4% of cases in clinical practice. MR correlation can help characterize CEM findings to: (1) avoid unnecessary biopsy for benign findings, and (2) guide tissue sampling or empiric surgical planning for suspicious findings.