AI Spots Cardiovascular Risk in Mammograms

AI detected signs linked to hypertension, heart disease and prior stroke in routine mammograms, an ESC Congress study found.

2 Sep 2026

Routine mammograms may reveal more than signs of breast cancer. At ESC Congress 2026 in Munich, researchers presented an AI model that detected cardiovascular signals in standard mammography images, raising the prospect of identifying heart risk without another scan.

The retrospective study drew on 97,364 mammograms from 29,921 women with a median age of 54. Using those images alone, the deep-learning model identified markers associated with hypertension, ischemic heart disease and previous stroke.

Researchers from Chaim Sheba Medical Center and Tel Aviv University developed the tool with a persistent screening gap in mind. Many women undergo regular breast imaging in midlife without receiving dedicated cardiovascular screening, even as that period can be important for identifying emerging risk.

For health systems, the appeal lies partly in what the approach does not require. Because the analysis uses mammograms already being taken, it could potentially add cardiovascular information without new imaging hardware or a separate scan, although the findings presented at the congress do not by themselves establish how the model would perform in routine clinical practice.

The approach could also broaden the reach of preventive care if further research supports its use. Existing mammography workflows already bring large numbers of women into contact with healthcare services, creating an opportunity to flag patients who may benefit from cardiovascular assessment.

Presented at the ESC Congress in Munich, held from August 28 to 31, the research remains an early step rather than a new screening standard. Still, it points to a wider possibility for medical AI: extracting clinically useful information from images that healthcare systems already collect.