Is automated breast density measurement more accurate?

Measuring breast density with automated software might be a more precise tool for assessing breast cancer risk than visual measurements based on BI-RADS criteria, according to a new study published in the January issue of the American Journal of Roentgenology.


Researchers from Elizabeth Wende Breast Care (EWBC) found that the sensitivity of automated volumetric breast density assessment was more similar to that of mammography than visual calculations of density based on BI-RADS assessment categories (AJR, January 2017, Vol. 208:1, pp. 222-227).

The team led by Dr. Stamatia Destounis included 771 breast cancers diagnosed at EWBC between 2009 and 2012. Of these, 652 were detected on screening and 119 were interval cancers. The group assessed breast density using both BI-RADS categories and VolparaDensity software from Volpara Solutions. The software generates a measurement of breast density as Volpara Density Grades, which correspond with BI-RADS density categories.

Destounis and colleagues found that overall mammographic sensitivity was 85%. Across visual BI-RADS categories, mammographic sensitivity was 82% for fatty breasts, 90% for breasts with scattered fibroglandular tissue, 84% for heterogeneously dense breasts, and 66% for extremely dense breasts. Using VolparaDensity categories, mammographic sensitivity was 95% for fatty breasts, 89% for breasts with scattered fibroglandular tissue, 83% for heterogeneously dense breasts, and 65% for extremely dense breasts.

"For women with dense breasts, there is a large range in sensitivity that is not being captured using the visual BI-RADS system which could impact the selection of women who would most likely benefit from supplemental imaging," Destounis said in a statement released by EWBC.