Mammography models predict heart attack risk from breast scans
Mammography Foundation Models for Opportunistic Prediction of Major Adverse Cardiovascular Events
Computer Vision and Pattern RecognitionMachine Learning
Summary
Heart disease is the top cause of death in women, but doctors often lack complete information to assess risk. The authors used computer models originally made to find breast cancer to look for signs of heart problems in mammogram images. These models could predict major heart problems within five years better than just using age, even without extra health data. The models noticed patterns related to calcium in breast arteries, which are linked to heart risks, without being told to look for them. This suggests mammograms might help spot women at risk for heart disease without needing extra tests.
What this means in practice
- •For radiology departments: Identify patients at risk of cardiovascular events by analyzing existing mammogram images without additional tests or clinical data.
- •For health insurers: Improve risk stratification models for women using routinely collected mammograms to guide preventative cardiovascular care cost-effectively.
Authors
Paula Feldman, Nusrat Binta Nizam, Sunwoo Kwak, Batuhan Karaman, Katerina Dodelzon, Mert Sabuncu
Abstract
Cardiovascular disease (CVD) remains the leading cause of death among women, yet cardiovascular risk assessment often relies on clinical variables that may be missing, outdated, or unavailable in routine care. Screening mammography offers an opportunity for opportunistic cardiovascular risk stratification because it is routinely acquired and contains vascular features, including breast arterial calcifications (BAC), that are associated with cardiovascular risk and events. We evaluate whether mammography specific foundation models, originally pretrained for breast cancer-related tasks, can transfer to cardiovascular risk prediction without cardiovascular specific supervision or explicit BAC annotation. We constructed a 5-year major adverse cardiovascular event (MACE) cohort of 22,497 women linked to electronic health record outcomes, including 500 events (2.22% prevalence). The foundation models achieved AUROCs of 0.823 and 0.822 substantially exceeding an age-only model (AUROC 0.765), despite using only the screening mammogram as input, with no clinical variables. Both foundation models evaluated assigned substantially higher predicted risk to patients with radiologist-documented BAC, despite BAC never being used as a training label, and showed activation patterns consistent with vascular findings. Together, these findings suggest that mammography foundation models can recover clinically relevant cardiovascular risk information directly from mammographic pixels and suggest that screening mammography may provide an opportunistic source of cardiovascular risk information to complement conventional clinical assessment without additional imaging. Code is available in https://github.com/PauFeld/MammoCVD