New BI-RADS guidelines may multiply dense-breast counts

Reviewing their group’s implementation of the fifth edition of the American College of Radiology’s BI-RADS Atlas, breast radiologists at the Medical University of South Carolina have observed considerable reader variability in determinations of which patients have dense breast tissue.

However, they also found an overall increase in the number of women deemed to have dense breasts, leading them to conclude that the new density-assessment guidelines may well be helping to identify more women who will benefit from supplemental screening.

The American Journal of Roentgenology published their findings online Aug. 11.

Abid Irshad, MD, and colleagues gathered data on mammographic breast density as reported by five radiologists, comparing 19,066 assessments using 4th-edition BI-RADS against 16,907 using the 5th edition, which was published in 2013.

They found that, as a group, the five radiologists had an overall increase of 2.3 percent in tissue they deemed dense versus nondense after implementing BI-RADS 5.

Individually, two radiologists showed increased dense assessments using the 5th edition, while three colleagues showed no change.

Drilling down into several key categories, the researchers found most of the increase turned up in the category of “heterogeneously dense.”

Specifically, they recorded an overall 5 percent decrease in fatty assessments, 2.8 percent increase in scattered densities, 2.6 percent increase in heterogeneously dense and 0.4 percent decrease in extremely dense assessments.

In their discussion, Irshad et al. comment that the new guidelines make sense to radiologists both clinically and from a legal standpoint.

“Because of the fact that a relatively small patch of dense tissue can obscure an underlying cancer, radiologists are now able to label these patients as having dense breasts,” the authors write. “This may in turn provide these patients (who can be labeled as having dense breasts only according to the 5th edition guidelines) an opportunity for additional screening tests such as ultrasound that could find mammographically occult breast cancers that would have been missed otherwise.”

Following the new guidelines could drive up healthcare costs, as additional staff and resources may be needed to see to the supplementary exams, but the payoff would be better cancer care, the authors point out.

Another consideration they take up: If their single-institution results prove broadly representative, interreader variability on density assessments could confuse patients. Group practices may need to develop “some kind of uniform consensus approach in assessing breast density under the new guidelines,” they write.

Along with their single-group study design, Irshad and team acknowledge as a limitation the fact that their study period—about six months using the BI-RADS 5 density-assessment guidelines—may not have given every radiologist sufficient time to adapt to the new guidelines.

“A larger and multiinstitutional study comparing density distribution datasets obtained from routine clinical practice using the 4th and 5th edition guidelines may give more accurate results,” they write. “Additionally, because many radiologists are still in the process of adapting to the new guidelines, the density distribution pattern using the 5th edition guidelines may still be settling at a larger scale. It will be of interest to compare the data again several years down the road.”