Changes in Breast Density Reporting Patterns of Radiologists After Publication of the 5th Edition BI-RADS Guidelines: A Single Institution Experience.

Abstract

The objective of our study was to determine the impact of 5th edition BI-RADS breast density assessment guidelines on density reporting patterns in our clinical practice. PenRad reporting system was used to collect mammographic breast density data reported by five radiologists: 16,907 density assignments using 5th edition BI-RADS guidelines were compared with 19,066 density assessments using 4th edition guidelines. Changes in the density assessment pattern were noted between the 4th and 5th edition guidelines, and agreement in density distribution was compared using the intraclass correlation coefficient. A chi-square analysis was conducted for each reader to examine the change in the proportion of dense versus nondense assignments and on each category type to examine specific changes in proportion of density assignments from the 4th to the 5th edition. All reported p values are two-sided, and statistical significance was considered at the p < 0.001 threshold. Using the 5th edition, there was an overall 5.0% decrease in fatty assessments (p < 0.001), 2.8% increase in scattered densities (p < 0.001), 2.6% increase in heterogeneously dense (p < 0.001), and 0.4% decrease in extremely dense assessments (p = 0.15). Comparing the dense with nondense categories, there was a 2.3% overall increase in the dense assessments (p < 0.001) using 5th edition guidelines, mainly in the heterogeneously dense category. Two radiologists showed increased dense assessments (p < 0.001) using the 5th edition, and three radiologists showed no change (p = 0.39, 0.67, and 0.76). There was an overall increase in the dense assessments using the 5th edition, but individual radiologists in our clinical practice showed a variable adaptation to new guidelines.