NL
DE
In the Netherlands, thousands of mammograms are interpreted each day by two radiologists as part of breast cancer screening. A new study from Radboudumc has shown that the sequence in which mammograms are reviewed could influence radiologists’ accuracy and efficiency in detecting abnormalities.
Traditionally, mammograms are reviewed in the order they are taken, resulting in a random mix of images. This study, conducted by Jessie Gommers and Sarah Verboom, and led by Ioannis Sechopoulos from the Medical Imaging department and Mireille Broeders from the IQ Health department, investigated whether ordering mammograms based on the underlying characteristics of the mammograms, such as breast density, rather than randomly, could enhance radiologists’ detection capabilities. Based on visual adaptation theory, which posits that our visual system adjusts to repeated exposure to certain features, researchers hypothesized that radiologists might become better at identifying cancerous abnormalities when they adjust to the visual characteristics of the breast tissue.
To explore this, 13 experienced Dutch radiologists examined 150 screening mammography exams, half of which showed breast cancer while the other half were normal or benign. Three sequencing methods were tested: random ordering, increasing volumetric breast density (from fatty to dense tissue), and an AI-generated sequence using a self-supervised learning (SSL) algorithm to group similar mammograms.
Results point to improved detection with density-based ordering
The study, published in Radiology on October 8, 2024, found that radiologists reviewing mammograms in order of increasing density showed better accuracy and faster reading times. Radiologists also perceived the images as more dense after starting with less dense ones, supporting the adaptation hypothesis. The SSL ordering, however, did not yield the same improvements.
This research suggests that reordering mammograms by breast density could be an effective, low-cost adjustment in screening programs, potentially leading to earlier and more accurate breast cancer detection. Jessie Gommers: "It’s a simple change that could lead to better outcomes for women undergoing breast cancer screening. Sometimes, it’s the small tweaks that can make a difference."
This research is part of Radboudumc Research program: Breast cancer
About the publication
Gommers JJJ, Verboom SD, Duvivier KM, van Rooden JK, van Raamt AF, Houwers JB, Naafs DB, Duijm LEM, Abbey CK, Webster MA, Broeders MJM, Sechopoulos I. Enhancing Radiologist Reading Performance by Ordering Screening Mammograms Based on Characteristics That Promote Visual Adaptation. Radiology. 2024 Oct;313(1):e240237. doi: 10.1148/radiol.240237. PMID: 39377678.
