Is a one-year follow-up an efficient method for better management of MRI BI-RADS(®) 3 lesions?

Abstract Objectives To evaluate the necessity of a prolonged follow-up after one year for lesions incidentally diagnosed as Breast Imaging Reporting and Data System (BI-RADS) ® 3 on magnetic resonance imaging (MRI). Methods Institutional review board approved this study (CRC_june 2014_ project 11). We retrospectively analysed 218 consecutive MRI BI-RADS ® 3 lesions in 110 women followed-up at 6, 12, 18 and/or 24 months from 2011 through to 2015. We excluded MRI scans of BRCA mutation carriers, in staging before treatment and in patients undergoing therapy for breast cancer. Second-look ultrasound and mammography were normal. We analysed 43 masses, 46 foci and 129 non-mass enhancements. In the follow-up, the image was scored as BI-RADS ® 2 if it was completely resolved, BI-RADS ® 3 if it was stable, and BI-RADS ® 4 if a suspicious change occurred. Results Forty-four lesions (20%) completely resolved during the follow-up. Six were scored as BI-RADS ® 4: five at 6 months (benign) and one at 12 months (8 mm invasive ductal carcinoma). One hundred and sixty eight stable lesions were assigned BI-RADS ® 2 at 18 or 24 months. The predictive positive value of cancer for BI-RADS ® 3 lesions is 0.5% (95% CI [0.08–2.55]) in our study. Conclusions Our results suggest that a one year follow-up instead of two years is sufficient for BI-RADS 3 lesions, in case of using rigorous inclusion criteria. Given the very low PPV of the BI-RADS ® 3 lesions incidentally found on MRI, further studies with very large inclusions are necessary to establish new guidelines for MRI follow-up.