Reviewing screening breast MRI data from six regional Breast Cancer Surveillance Consortium (BCSC) registries representing 49 facilities, researchers have found breast rads performing quite impressively in their respective communities: Their interpretive accuracy from 2005 to 2013 met most BI-RADS benchmarks that are based on expert practice in clinical trials.

The radiologists collectively fell a little short with positive predictive values and MR imaging specificity, approaching but not quite getting to the benchmark of 85 percent to 90 percent. This suggests that “a continued focus on reducing false-positive results while maintaining sensitivity and cancer detection remains important in ongoing quality improvement efforts,” write lead author Janie Lee, MD, MSc, of the University of Washington in Seattle and colleagues in a study report published online June 5 in Radiology.

The data Lee and team analyzed came from 8,387 BCSC screening MR exams carried out on 5,343 women.

Key findings included:

• The practitioners’ positive predictive value for biopsy recommendation came in at 19 percent, topping the BI-RADS benchmark of 15 percent.
• Sensitivity was 81 percent (benchmark = >80 percent).
• Specificity was 83 percent (benchmark = 85 to 90 percent).

The overall cancer detection rate showed room for improvement, as BI-RADS sets the bar at 20 to 30 cancers per 1,000 exams and the BCSC study group detected 17 per 1,000.

The study itself will contribute to ongoing quality improvement, as some of its results can provide clinical practice-based values for measures currently listed in the BI-RADS manual as “To Be Determined.” These measures include percentage of stage 0 or 1 breast cancers (study finding: 87 percent) and median size of invasive breast cancers (10 mm).

The regional registries supplying the data were in Chicago, Western Washington, New Hampshire, North Carolina, the San Francisco Bay Area and Vermont.

The authors note that screening breast MRI is recommended as an adjunct to screening mammography for many women at high risk.

“[T]he interpretive performance of screening breast MR imaging in U.S. community practice meets or approaches current BI-RADS benchmarks,” Lee et al. conclude. “Individual practices can compare their performance with BCSC performance as well as BI-RADS benchmarks to better understand their performance. Clinical practice performance data can inform and supplement ongoing benchmark development.”