AI + mammography could simultaneously screen for breast cancer and identify women with increased CV risk.A retrospective analysis presented at ESC 2026 demonstrates the ability of artificial intelligence (AI) to detect cardiovascular diseases (CVDs) from routine mammograms.
The three major CVDs detected were hypertension, ischaemic heart disease (IHD), and cerebrovascular accident (CVA), noted Dr Viana Copeland from the Heart Institute, Sheba Medical Center, Tel-Hashomer, Ramat Gan, Israel, at ESC 2026.
The deep learning-based algorithm yielded AUROCs* of 0.79 for hypertension, 0.78 for IHD, and 0.86 for CVA. “[Given] that these predictions were generated from images required solely for breast cancer screening, these results are encouraging and demonstrate that mammograms contain latent CV information,” Copeland explained.
“The curves illustrate the consistent performance across the three disease categories. Interestingly, the prediction was strongest for previous CVA, [suggesting] that chronic vascular disease leaves imaging signatures that can be captured by deep learning,” she continued.
The mediolateral oblique mammographic views consistently exhibited the highest predictive performance, with AUROCs of 0.80, 0.80, and 0.88 for hypertension, IHD, and CVA, respectively. These findings suggest that certain mammographic views may capture more CV information than others, Copeland noted.
Strong discrimination was achieved even without dedicated CV imaging. The findings were robust across multiple sensitivity analyses, with consistent results across age groups and cancer status.
Underdiagnosed, undertreated
“CVD remains the leading cause of death in women; however, it is consistently underdiagnosed and undertreated. A common finding in our medical centre and around the world is that when women seek medical help, their CVD is already advanced. On the other hand, many women attend routine breast cancer screening, even when they have not sought care for CV symptoms,” Copeland said in the ESC press release.
Copeland and colleagues sought to evaluate whether AI could help mammography serve an additional purpose in early CVD detection to enable preventive strategies. They evaluated data from 29,921 women (median age 54 years) who underwent 97,364 mammography examinations between 2011 and 2025. Of these, 5,485 women (18 percent) had breast cancer. The median follow-up was 7.3 years.
The prevalence of hypertension was 16 percent, while that of IHD and CVA were both 2.5 percent. [Copeland, V, et al, ESC 2026]
Dual-purpose platform
“[Taken together,] the findings suggest that routine mammography could evolve into a scalable, dual-purpose screening platform without requiring any additional imaging, radiation exposure, or patient appointments,” said Copeland.
“AI could simultaneously screen for breast cancer and identify women with increased CV risk. Such an approach could facilitate earlier opportunistic CV assessment and preventative interventions in a population that is already participating in routine screening programmes,” she continued.
Beyond breast cancer detection
In the ESC press release, Assoc Prof Elena Arbelo from the University of Barcelona, Spain, and a member of the ESC Communication Committee, shared her insights into the results. “As a cardiologist and a woman, I find this concept compelling: One day, a mammogram may do more than look for breast cancer. It may also offer a window into CV health. [This] matters because CVD in women is still too often recognized late.”
“It is great to see innovative AI studies … aiming to address unmet needs. The challenge now is to establish accuracy and reliability—to move from experimentation to clinical implementation,” Arbelo added.