Full Breakdown
AI-Enhanced Mammograms Offer Early Detection of Women’s Heart Disease
8/28/2026, 12:49:35 AM
Core Development: Dual-Purpose Breast Screening
Researchers from Tel Aviv University’s Chaim Sheba Medical Center presented a machine-learning model at the 2026 European Society of Cardiology (ESC) annual congress in Munich. The model analyzes routine mammograms—already used for breast-cancer screening—to identify three common cardiovascular conditions: hypertension, coronary heart disease (ischemic heart disease) and stroke.
Background & Context
Cardiovascular disease (CVD) remains the leading cause of death among women worldwide, yet it is frequently underdiagnosed and treated later than optimal. In the United Kingdom, coronary heart disease kills twice as many women as breast cancer each year and accounts for the highest female mortality globally in 2023, according to the British Heart Foundation. Conversely, mammography is a well-established, widely accessed screening tool for women in midlife, providing an opportunity to capture cardiovascular risk without additional imaging.
Data & Statistics
- The retrospective cohort comprised 29,921 women (median age 54) who underwent 97,364 mammograms.
- Medical record linkage revealed prevalence rates of 16 % for hypertension, 2.5 % for coronary heart disease, and 2.5 % for stroke.
- Model performance, measured by area under the receiver-operating-characteristic curve (AUROC), was 0.79 for hypertension, 0.78 for coronary heart disease and 0.86 for stroke.
- Accuracy was consistent across age groups and irrespective of concurrent cancer status.
Official Statements & Responses
Dr Viana Copeland, a physician at Chaim Sheba Medical Center, emphasized that CVD “is consistently under-diagnosed and under-treated” and noted that many women only seek care after disease has advanced, whereas routine breast-cancer screening reaches a broader population. She argued that repurposing mammograms could provide a scalable, non-invasive avenue for early cardiovascular risk assessment.
Elena Arbelo, a member of the ESC communication committee, described the findings as “compelling” and highlighted the need to prove the model’s accuracy and reliability before clinical rollout.
Dr Sonya Babu-Narayan, consultant cardiologist and clinical director of the British Heart Foundation, reiterated that heart disease is the world’s biggest killer of both sexes and that the myth of it being a “men’s disease” hampers awareness and research inclusion. She welcomed the prospect that AI-augmented mammography might enable earlier detection and prevention for women.
Ongoing Work and Remaining Gaps
The research team is refining the algorithm to reduce false-positive results and expand the range of detectable cardiovascular conditions. While initial AUROC values are promising, the study’s retrospective design and reliance on existing medical records mean prospective clinical trials are still required. Experts note that validation in diverse, real-world populations will be essential before integration into national screening programs.
Future Outlook
Stakeholders identified the transition from experimental proof-of-concept to routine clinical implementation as the next critical step. Ongoing efforts will focus on improving model precision, confirming reproducibility across healthcare settings, and establishing guidelines for how cardiovascular findings from mammograms should be communicated to patients and primary-care providers.
