NYU AI Predicts 5-Year Breast Cancer Risk
Analysis based on 6 articles · First reported Sep 10, 2026 · Last updated Sep 10, 2026
The study highlights growing adoption of AI in medical imaging, potentially boosting demand for AI-assisted diagnostic tools and benefiting equipment makers such as Hologic. However, the near-term market impact is limited as the tool remains in research stages and requires further validation across other health centers and manufacturers.
Researchers at NYU Langone Health and its Perlmutter Cancer Center developed a deep-learning tool, NYU-DRP, that analyzes a woman's past and recent annual 3D mammograms (longitudinal digital breast tomosynthesis) to predict five-year breast cancer risk. The study, published in the American Journal of Roentgenology, found NYU-DRP correctly ranked higher-risk women 72 percent of the time, versus 70 percent for single 3D mammography and 68 percent for AI-assisted 2D testing. Against the widely used Tyrer-Cuzick risk assessment, NYU-DRP predicted five-year risk correctly 67 percent of the time versus 56 percent. The model was built from 313,531 yearly 3D mammograms from 161,165 women tested at NYU Langone hospitals between 2016 and 2020. The study also found breast density alone did not correspond to predicted risk. All testing used breast imaging equipment manufactured by Hologic. Funding came from the United States — National Science Foundation, the United States — National Institutes of Health, the Milstein Pilot Project Fund, the Shifrin-Myers Breast Cancer Discovery Fund, and the Manhasset Women's Coalition Against Breast Cancer. Lead investigator Yanqi Xu and senior investigator Yiqiu Shen, along with co-investigators Laura Heacock, Jungkyu Park, Felicia Pasadyn, Lei Stanley Qi, Alana Lewin, Krzysztof Geras, Linda Moy, and Freya Schnabel, contributed to the research. The team plans to share and cross-check the tool with other academic health centers and different 3D mammogram manufacturers.
Set up alerts, explore entity relationships, search across thousands of events, and build custom intelligence feeds.
Open Dashboard