Abstract
BACKGROUND: The aims of this study were to correlate residual mammographic microcalcifications after neoadjuvant chemotherapy (NAC) with pathological results and to compare the accuracy of mammography (MG) and magnetic resonance imaging (MRI) in predicting the size of residual tumors. METHODS: The imaging findings and pathological results for 29 patients with residual microcalcifications after NAC were reviewed. We compared the agreement of the measured extent of residual microcalcifications based on MG and residual enhancement based on MRI with the residual tumor size based on pathology. RESULTS: At final pathology, residual microcalcifications were malignant in 55.2% of cases and benign in 44.8% of cases. In 36% of non-pCR cases, the remaining microcalcifications were benign. Compared with the measurements of residual tumor obtained from pathology, MG showed poor agreement, and MRI showed moderate agreement, for the entire group (concordance correlation coefficient [CCC] = 0.196 vs. 0.566). Regarding the receptor status, the agreement of measurements obtained by MG was superior to that obtained by MRI (CCC = 0.5629, 0.5472 vs. 0.4496, 0.4279) for ER(+) and HER2(-) tumors. In ER(-) tumors, the measurements obtained by MG showed the lowest agreement with the pathological tumor size, which had the highest agreement with those obtained by MRI (CCC = - 0.0162 vs. 0.8584). CONCLUSIONS: Residual mammographic microcalcifications after NAC did not correlate with malignancy in 44.8% of cases. Residual microcalcifications on MG were poorly correlated with pathological tumor size, and MRI might be more reliable for predicting residual tumor size after NAC. Tumor receptor status affected the accuracy of both MG and MRI for predicting residual tumor size after NAC. TRIAL REGISTRATION: CRIS, KCT0002281 ; registered 6 April 2015, retrospectively registered.