Docs’ eyes with BI-RADS slightly better than software at separating breast cancer patients from control subjects

Following imaging with full-field digital mammography, radiologists’ visual inspection paired with clinical BI-RADS breast-density scores has bested both fully automated and semiautomated computer-assisted methods at distinguishing between 125 breast-cancer patients and 274 control subjects.

In the study, conducted at Stanford and published online Sept. 5 in Radiology, the margin of the present victory was tiny. But it was a win for visual assessment nonetheless, leading Abra M. Jeffers, PhD, MPhil, and colleagues to conclude that BI-RADS ought to be considered an appropriate measure of mammographic density for estimating patients’ risk of breast cancer and guiding treatment decisions.

The team used Cumulus software to assess percentage of density and dense area semiautomatically and Volpara software to assess volumetric percentage of density and dense volume automatically. They extracted BI-RADS classifications of breast density from mammography reports.

The area under the receiver operating characteristic curve (AUC) for mammographic percentage of density was 0.68, 0.66, and 0.64 for BI-RADS, the semiautomated method and the fully automated method, respectively.

“[T]hus, BI-RADS was as accurate as computer-assisted methods for discrimination of patients with breast cancer from control subjects,” Jeffers et al. write.

The researchers further found that all density measures were positively associated with breast cancer risk.

“Clinical assessment with BI-RADS allowed the best discrimination of patients from control subjects, followed by Cumulus and Volpara measurements, although the AUC differences were small and not statistically significant,” they write.

The authors acknowledge several study limitations. These included the small sample size, which limited the ability to detect subtle differences in discrimination among the density assessment methods, and the use of a single reader for clinical BI-RADS density assessment. (This is the standard clinical practice in the U.S.).

“[O]ur results are interesting in that, if visual assessment is the standard, as is now legislated in more than half of all states in the United States, then Cumulus and Volpara do not appear to provide much additional value for prediction of risk beyond that of routine clinical BI-RADS assessment,” the authors write.

Additional studies would be helpful, they add, to determine whether there may be small incremental benefits of current computerized methods.

“Beyond differences in risk prediction, automated methods may provide benefits of reducing interobserver variability and improving reproducibility of breast density assessments over time,” Jeffers and colleagues write. “Future development of automated methods to quantify breast density and other risk features could improve risk prediction with full-field digital mammography images.”