BCSC Blog

The Radiologist Learning Curve for 3D Mammography

BCSC study finds radiologists quickly improve screening performance with 3D mammography

Posted by Diana Miglioretti, PhD at 10:46 AM on Mar 4, 2019

Share:


A BCSC study found radiologists quickly learn to read 3D mammography more accurately than they read standard 2D mammograms. Published in Radiology, the study found that radiologists who interpret traditional mammograms, which are two-dimensional, required little start up time for transitioning to reading digital breast tomosynthesis (DBT), often called 3D mammography, with improved screening accuracy.

Using prospectively collected data from 104 radiologists from 53 facilities in five U.S. states, the study found radiologists recalled women for additional testing at a lower rate on DBT (9.4%) than they did on 2D mammography (10.4%), without sacrificing cancer detection (DBT=4.6/1000, digital=4.0/1000). These improvements were seen regardless of the patient’s breast density. In fact, improvements in recall rate were greater for women with non-dense breasts (odds ratios=0.68-0.76) compared to dense breasts (odds ratios=0.86–0.90).

The study is the largest of its kind and represents a broad range of radiology centers and providers. It is novel because it tracked radiologists’ performance over time as they transitioned from digital mammography to DBT. In an accompanying editorial by Dr. Regina Hooley of Yale School of Medicine in New Haven, CT noted: "The results of this study are important because the authors showed that the well-established benefits of DBT are present among subspecialist breast imagers and nonspecialist radiologists throughout the United States, and that a significant reduction in screening recall rates can be seen in the immediate DBT adoption period."

Full Text Citation and Link: Miglioretti DL, Abraham L, Lee CI, Buist DSM, Herschorn SD, Sprague BL, Henderson LM, Tosteson ANA, Kerlikowske K; Breast Cancer Surveillance Consortium. Digital Breast Tomosynthesis: Radiologist Learning Curve. Radiology. 2019 Feb 26:182305. doi: 10.1148/radiol.2019182305. [Link to Article]

Press release: https://health.ucdavis.edu/publish/news/newsroom/13593/?WT.rss_f=UC%20Davis%20Health%20news&WT.rss_ev=a&WT.rss_a=Article_13593

Study highlighted in the following media sources:


Posted by: Diana Miglioretti, PhD