A mammogram is normally used to look for early signs of breast cancer, but the same image may contain information about a woman’s heart and blood vessels.
New research suggests that artificial intelligence could analyze routine mammograms and help identify women who already have common cardiovascular diseases.
The study was presented at ESC Congress 2026 by Dr. Viana Copeland of Chaim Sheba Medical Center and Tel Aviv University in Israel. The researchers tested whether an AI system could recognize patterns linked to high blood pressure, ischemic heart disease, and stroke.
Cardiovascular disease is the world’s leading cause of death in women, yet it can be missed or diagnosed later than it is in men. Symptoms in women are not always recognized quickly, and some women may not seek medical attention until disease has become more serious.
Breast cancer screening creates an interesting opportunity because many women attend mammography appointments even when they feel healthy. Screening often takes place during middle age, which is also an important period for identifying high blood pressure and other factors that can raise future heart and stroke risk.
A mammogram uses low-dose X-rays to create images of breast tissue. Although its main purpose is cancer screening, the images can also capture other physical features. Previous research has suggested that findings such as calcium deposits in breast arteries may be connected with cardiovascular health.
The new study included 29,921 women who had a total of 97,364 mammograms. Their median age was 54, meaning half were younger and half were older. Researchers also collected health information from medical records, prescriptions, procedures, and other imaging tests.
They focused on three cardiovascular conditions. About 16% of the women had hypertension, or high blood pressure, while 2.5% had ischemic heart disease, which occurs when blood flow to the heart is reduced. Another 2.5% had experienced stroke.
The researchers trained a deep-learning system to examine mammograms and identify image patterns associated with each condition. Deep learning is a form of AI that can learn complex patterns from large collections of examples rather than relying only on features manually chosen by researchers.
The model showed encouraging results. Its performance score was 0.79 for hypertension, 0.78 for ischemic heart disease, and 0.86 for stroke on a scale where 0.5 is equivalent to random guessing and 1.0 represents perfect separation between people with and without a condition.
The findings were also broadly consistent when researchers considered factors such as age and whether a woman had cancer. This suggests that the model may be detecting information in mammograms that is connected with cardiovascular disease rather than simply relying on obvious differences between patient groups.
One attraction of the approach is that it would not necessarily require women to undergo another scan. If the technology eventually proves reliable, existing mammograms might be analyzed for both breast cancer and cardiovascular information, potentially adding another layer of health screening without another imaging appointment.
However, the AI system is not ready to replace standard cardiovascular tests. A mammogram cannot currently diagnose high blood pressure, coronary artery disease, or stroke on its own, and an AI prediction could produce false alarms or miss people who actually have disease. The researchers are now working to improve accuracy and reduce these errors.
The study was presented at ESC Congress 2026, the annual congress of the European Society of Cardiology. The research team is also exploring whether mammogram images could provide useful clues about additional cardiovascular conditions.
The size of the dataset is a strength, with nearly 30,000 women and more than 97,000 mammography examinations. The idea is also practical because it builds on a screening test that is already widely used rather than asking health systems to introduce an entirely new scan.
There are important limitations, however. This was a retrospective study, meaning the researchers analyzed information that had already been collected, and the AI was trained using patients from a particular health setting.
Before clinical use, the system would need to be tested independently in women from different countries, ethnic backgrounds, ages, hospitals, and types of mammography equipment.
The most promising role may ultimately be as an early warning tool rather than a diagnostic test. If an AI system flags a woman as having possible cardiovascular risk, doctors could follow up with established assessments such as blood pressure measurements, blood tests, or heart examinations.
The study therefore offers an intriguing possibility: one day, a routine breast screening image might help doctors protect both breast and heart health.
If you care about heart health, please read studies about how drinking milk affects risks of heart disease , and herbal supplements could harm your heart rhythm.
For more information about heart health, please see recent studies about how espresso coffee affects your cholesterol level, and results showing Vitamin K2 could help reduce heart disease risk.
Source: Chaim Sheba Medical Center and Tel Aviv University.
