Probably benign findings, BIRADS category 3 (MR-Bi3), on magnetic resonance imaging of the breast (BrMR): Is repeated MRI an alternative to biopsy?

21 Background: The shift from the use of BrMR for confirmation of known abnormality to the screening mode in problematic breast is associated with higher frequency of probably benign findings. Some of them still require biopsy, with or without close BrMR follow up (f/u), some may remain on observation only. Our pilot study analyzed clinical outcomes of MR-Bi3 pts and added value of close BrMR f/u. Proportion of MR-Bi3, some after “second look ultrasound” (2ndUS) with or without biopsy could be used among other monitors of BrMR practices. Methods: We reviewed BIRAD categories assigned to all 743 BrMRs done in period of year 2007 in our specialized breast imaging practice with dedicated BrMR (Aurora) system and full complement of diagnostic and interventional modalities. Frequency of MR-Bi3 and studies resulting in MR-Bi3 or biopsies after 2ndUS were analyzed. With benefit of 3 year clinical data, the outcome of patients on such short term f/u schedule was reviewed. Results: MR-Bi3 was assigned to 40 (5.4%) of our BrMRs on the initial evaluation, all of them yielded benign findings on follow up. Additional 26.2% of our studies required 2ndUS. After 2ndUS, 40% of these pts were converted to Bi4+5, continued with image guided biopsy with cancer detected in 35% of those. Remaining 60% of 2ndUS yielded “probably benign findings”, MR-Bi3, which means that total of 21.6% of our BrMRs remained on close f/u schedule with MRI. There were no additional delayed findings of malignancy in this group. Single additional BrMR was sufficient to confirm benign findings only in 47% of cases, 24% of patients required multiple BrMRs at 6 month intervals, remaining patients were followed by other modalities. Conclusions: Utilization of MR-Bi3 category reflects readers' confidence avoiding or justifying biopsies for border line findings. Targeted use of 2ndUS helps to identify patients who benefit from biopsy rather than close f/u. Proportion of repeated Bi3 and outcome of Bi3 could become valuable QA indices for monitoring of BrMR practices. Only in small proportion of cases, the use of Bi3 was not contributory, particularly in case of repeated BrMRs for the same abnormality.