Prioritizing mammogram reads based on breast density improves radiologist performance

Prioritizing mammogram reads based on breast density appears to improve radiologists’ performance, according to new research published Tuesday in RSNA’s Radiology [1]. 

Amid staffing shortages and rising imaging volumes, practices are seeking ways to increase efficiency. Dutch scientists recently tested the effect of rearranging radiologist worklists to start with the easier-to-read, low-density mammograms, with physicians building their way up to the more difficult scans with denser tissue. 

Following the intervention, radiologists executed their duties more quickly and produced better results. 

“In conclusion, the findings of our study indicate that ordering mammography examinations from low to high volumetric breast density (VBD) improved screening performance while reducing reading and fixation times,” Ioannis Sechopoulos, PhD, a professor of advanced X-ray imaging methods at Radboud University Medical Center in the Netherlands, and co-authors wrote Oct. 8. “Although larger studies are needed to validate these findings, our results suggest that ordering screening mammograms for reading by increasing VBD may be feasible.”

The retrospective study utilized mammograms originally obtained between 2016 and 2019. Thirteen radiologists interpreted screening examinations (each consisting of four mammograms) in three distinct orders. These included random ordering, by increasing breast density, and grouped based on those that appeared similar. Researchers also used a tracker to follow radiologists’ eye movements during interpretation, counting the number of “fixations.” The study sample included images from 150 women, with half containing cancer while the other half did not. 

Reader-averaged area under the receiver operating characteristic curve was greater when radiologists interpreted images ordered by breast density versus randomly (0.93 vs. 0.92). Reading times dropped by about 3.6 seconds (median, 24.3 vs. 27.9 seconds), while fixation counts (brief pauses on parts of the images) dropped by five (median, 47 vs. 52). Fixation duration in the malignant-lesion regions also dropped by about 0.9 seconds (median, 3.7 vs. 4.6 seconds). 

“Put simply, the readers were faster, more accurate, and more efficient when screening mammography examinations were ordered by breast density,” Lars J. Grimm, MD, an associate professor in the Breast Imaging Division at Duke University, wrote in a corresponding editorial [1]. “In contrast, there were no performance benefits for the AI-organized approach versus random shuffling. In fact, the reading time with the AI approach was significantly longer than that with random shuffling (median, 28.4 vs. 27.9 seconds). The authors thus demonstrated that deliberate ordering could benefit the interpretation of screening mammograms but that not all ordering approaches are better than the current random shuffle standard.”

Read more about the results, including potential study limitations, at the links below.