Attention must be paid to high abnormal-interpretation rates in modern screening mammography

Most radiologists performing digital screening mammography in U.S. community practice are meeting or exceeding most of the American College of Radiology’s performance recommendations. However, almost half have higher abnormal interpretation rates (AIRs) than established benchmarks and ACR’s recommendations on that metric—and that’s a serious problem calling for a concerted effort to solve.  

The findings and conclusion come from a study published online Dec. 5 in Radiology.

Constance Lehman, MD, PhD, of Harvard and colleagues describe their work measuring the performance of digital screening mammography as interpreted by 359 radiologists across 95 facilities in six registries of the Breast Cancer Surveillance Consortium (BCSC).

The assessment looked at nearly 1.7 million digital screening mammograms of 792,808 women between 2007 and 2013.

The researchers calculated performance measures using the fifth edition of ACR’s BI-RADS atlas, then compared these calculations with benchmarks published by the BCSC, the National Mammography Database and other sources.  

They found that an encouraging 92.1 percent of radiologists in community practice achieved recommended cancer detection rates, and some 97.1 percent achieved recommended ranges for sensitivity.

Further, compared with prior performance reports of screening mammography in the BCSC (1996 to 2008), the overall sensitivity of screening mammography has increased from 78.7 percent to 86.9 percent.

Now the bad news. Lehman and team found that only 59 percent of the radiologists achieved recommended AIRs, and only 63 percent achieved recommended levels of specificity.

Also, the mean AIR in the study was 11.6 percent—higher than that recorded in 2005 and 2008 BCSC reports (10.9 percent and 10 percent, respectively) and higher than the 10 percent rate recently reported by the National Mammography Database.

“This is particularly concerning, given that recall rates have continually failed to meet the recommendations of the ACR and other expert panels going back to the initial BCSC report in 2005, despite calls for attention to this matter,” the authors write in their discussion.

They suggest that increasing access to tomosynthesis imaging for screening could help reduce recall rates.

They also suggest the implementation of programs to support second reviews of mammograms recalled by radiologists who are known to “overcall” mammograms.

Such second reviews of the recalls, the authors point out, could be performed by radiologists with documented high performance for both recall and cancer detection rates.

The resource investment would be manageable for most practices, they state, as second reads would only be required of the mammograms read by radiologists with poor specificity.

Lehman et al. point to previous studies showing the “potential positive impacts on our patients and healthcare expenditures if all radiologists were to meet minimally acceptable standards of performance.”

But getting there will take resolve and, probably, some level of discomfort.

“[A]chieving this end will likely require remedial or restrictive action to be taken regarding subpar performers,” the authors write. “Whether we are ready to take this next step in quality assurance and cost containment in screening mammography warrants careful consideration.”