Second‐Look Ultrasound for MRI ‐Detected Lesions: MRI Features, Detection Rate, and Sonographic Findings

BACKGROUND: Second-look US emerged as a feasible tool to evaluate MRI-detected lesions, although data regarding the accuracy and optimal lesion selection for second-look US is limited. Therefore, we evaluated the second-look US detection rate of breast lesions detected initially on MRI by using MRI features and compared second-look US findings across benign and malignant lesions. MATERIAL AND METHODS: A total of 110 MRI-detected lesions subsequently evaluated with the second-look US were retrospectively reviewed. The performance of MRI features in the prediction of second-look US detection was evaluated with ROC analyses. RESULTS: The positive and negative predictive values of second-look US for malignancy detection were 28.8% and 96.1%, respectively. In multivariate analyses, the MRI BI-RADS category (OR: 11.896, p < 0.001) and lesion type (OR: 9.470, p = 0.001) were significantly associated with second-look US detection. Combined scoring with these factors could predict second-look US detection [AUC: 0.712, 95% CI: 0.617-0.807 p < 0.001]. Furthermore, lesions with irregular shapes and margins were more often malignant in the analysis of US findings (p < 0.001). CONCLUSIONS: The MRI BI-RADS category and lesion type could predict detection in the second-look US for MRI-detected lesions. Besides, second-look US could give additional information to distinguish between benign and malignant masses.