New AI Tool Predicts Breast Cancer Risk Using 3D Mammograms

NYU Langone Health researchers developed an AI model called NYU-DRP that analyzes longitudinal 3D mammograms over multiple years to predict a woman's five-year breast cancer risk with 72% accuracy.

The study, published on August 12, used 313,531 annual mammograms from 161,165 women, outperforming the traditional Tyrer-Cuzick risk assessment model which achieved only 56% accuracy in comparative testing.

Led by Dr. Yanqi Xu and Dr. Yiqiu Shen, the team plans to validate the Hologic-trained tool across other academic centers to help doctors customize screening schedules for an estimated 382,640 US cases in 2026.