AI system analyzes years of 3D mammograms to forecast breast cancer risk
Researchers at NYU Langone have created an artificial-intelligence model that examines a woman's series of 3D mammograms to estimate her chance of developing breast cancer within five years.
A study published in the American Journal of Roentgenology describes an AI algorithm developed by NYU Langone Health and its Perlmutter Cancer Center to predict a woman's breast-cancer risk up to five years ahead. Using a dataset of more than 313,000 3D mammograms from over 161,000 women, the system tracks longitudinal changes in breast tissue rather than a single snapshot. Specialists note that the approach may identify women whose risk would otherwise go unnoticed, potentially prompting earlier or additional screening such as MRI.
The technology can be applied to existing mammograms, avoiding extra radiation exposure. However, the researchers stress that the risk score is not a diagnosis and the model is not yet FDA-authorized; broader testing with diverse imaging systems is required before clinical use.
Why it matters
Early, personalized risk assessment could improve breast-cancer screening and outcomes for many women.
In this story
