Comparison of false positive rates for screening breast MRI in high risk women when studies are done stacked versus alternating.

Abstract

1512 Background: Screening breast MRI added to mammography increases screening sensitivity for high risk women. However, false positive rates are high for MRI and the optimal screening schedule is unclear. In this study we compare rates of false positive MRI when studies were performed on a stacked or alternating schedule. Methods: We reviewed charts for women at increased risk for breast cancer who had screening breast MRI between 2004 - 2012 at the University of Vermont. Eligible women had at least 1 MRI and 1 mammogram performed within one year. Charts were abstracted for clinical, radiological, and biopsy data. Screening was considered stacked if both studies were performed within 90 days and alternating if studies were 4-8 months apart. False positive was defined as MRI result of BI-RADS 3-4-5-0 with additional negative imaging within 12 months or benign biopsy. Results: 143 women had screening which met inclusion criteria; 45 per stacked schedule, 52 alternating, 40 mixed and 6 neither. Women in this study had similar characteristics with respect to age, ethnicity, menopausal status and indications for MRI (i.e. family history, BRCA mutation, biopsy history and prior chest irradiation). 371 MRIs were reviewed (165 stacked and 206 alternating). The overall false positive rate was higher in the stacked group vs. alternating [30(18.2%) vs. 21(10.2%), p=0.0264]. Using only BI-RADS 4-5-0 as a positive result that difference was lost. There were significantly more BI-RADS category 3 interpretations in the stacked vs. alternating MRIs [16(9.7%) vs. 6(2.9%), p=0.006]. The rate of BI-RADS category 4-5-0 was not different between the two groups [16(9.7%) vs. 17(8.3%), p= 0.6272]. A similar number of biopsies were performed in both groups Conclusions: MRI added to mammography for women at increased risk for breast cancer was associated with higher rates of false positive interpretations when studies were done on a stacked compared to alternating schedule. In this study the greater number of BI-RADS 3 interpretations with a stacked schedule accounted for this difference. Further studies are needed to identify the optimal screening schedule when adding MRI to mammography.