Negative predictive value and specificity of contrast enhanced spectral mammography in 309 women with suspicious findings at breast ultrasound

Abstract

Background: There are ongoing efforts to discover readily available noninvasive techniques to accurately exclude the presence of cancer when a suspicious breast lesion is identified. In order to be useful, such a technique must also accurately exclude a sizeable number of women who would have been otherwise biopsied. Therefore, such a test should have a high negative predictive value in excluding malignancy and a high enough specificity in predicting which women will turn out to have a benign biopsy result. The goal of this study was to investigate the negative predictive value and specificity of contrast enhanced spectral mammography (CESM) in women with a suspicious finding on diagnostic mammography and ultrasound. Methods: With IRB approval, CESM reports from two academic breast centers were retrospectively evaluated. Four hundred and twenty-four women were identified, and 309 women met inclusion criteria. All women had a suspicious finding on diagnostic mammography and ultrasound and had CESM prior to a same day scheduled ultrasound guided core needle biopsy. Results: Two hundred/309 (65%) women had histologically proven malignancy and 109/309 (35%) women had benign results. One hundred and ninety-seven/200 cancers enhanced at CESM [sensitivity 98.5%; 95% confidence interval (CI): 95.7–99.7%]. Fifty-eight/109 women with benign results did not enhance and were true negatives (specificity 53.2%; 95% CI: 43.4–62.8%). Of the 248 women with CESM lesion enhancement, 197/248 were cancer (positive predictive value =79.4%; 95% CI: 76.0–82.5%). Of the 61 women without lesion enhancement on CESM, 58/61 were benign (negative predictive value =95.1%; 95% CI: 86.1–98.4%). The 3 false negatives consisted of 2 sub centimeter Nottingham grade 1/3 estrogen receptor (ER) positive HER2/neu negative invasive mammary cancers of no special type, and one case of ductal carcinoma in situ (DCIS) in a papilloma which was 3 mm in size at excision. Conclusions: CESM was accurate in excluding malignancy in over half of the women that had otherwise suspicious findings at diagnostic mammography and ultrasound. 12 fibrocystic changes, 5 of 8 papillomas, 1 of 8 stromal fibrosis, and 4 of 7 sclerosing adenosis results.