By Billy Candra 
A groundbreaking study, set to be unveiled at the American College of Cardiology’s Annual Scientific Session (ACC.25), reveals that mammograms, traditionally vital for breast cancer screening, can now, with the aid of sophisticated artificial intelligence (AI) models, provide crucial insights into a woman’s cardiovascular health. The research underscores the potential for these routine cancer screening tools to quantify calcium buildup in breast tissue arteries – a significant indicator of cardiovascular disease (CVD) risk – thereby transforming them into dual-purpose diagnostic instruments.
The U.S. Centers for Disease Control and Prevention (CDC) recommends that women in middle age and beyond undergo regular mammographic screening for breast cancer, typically every one to two years. Annually, approximately 40 million mammograms are performed across the United States, generating a vast repository of imaging data. While breast artery calcifications (BAC) are frequently visible on these X-ray images, radiologists have not historically quantified or routinely reported this information to patients or their referring clinicians. This omission is largely due to the primary focus on breast cancer detection, the lack of standardized quantification tools, and the absence of clear clinical guidelines for reporting BAC. The new study introduces an innovative AI image analysis technique, previously unapplied to mammograms in this specific manner, demonstrating how AI can effectively bridge this gap. By automatically analyzing breast arterial calcification, the AI model translates these findings into a tangible cardiovascular risk score, opening an unprecedented avenue for opportunistic screening.
The Silent Threat: Cardiovascular Disease in Women
Heart disease remains the leading cause of death for women in the United States, accounting for approximately one in every five female deaths. Globally, CVD is responsible for an estimated 8.6 million deaths in women each year. Despite its prevalence, cardiovascular disease is often underdiagnosed and undertreated in women, partly due to atypical symptom presentation compared to men, historical biases in medical research, and a lagging awareness among both the public and some healthcare providers regarding women’s specific risk factors. This disparity often leads to delayed diagnosis and intervention, contributing to poorer outcomes. The economic burden of CVD is also substantial, with direct medical costs and lost productivity totaling hundreds of billions of dollars annually in the U.S. alone.
"We see an unparalleled opportunity for women to receive not only their essential cancer screening but also an additional cardiovascular screen from their existing mammograms," stated Theo Dapamede, MD, PhD, a postdoctoral fellow at Emory University in Atlanta and the study’s lead author. Dr. Dapamede emphasized the predictive power of BAC, noting, "Our study showed that breast arterial calcification is a good predictor for cardiovascular disease, especially in patients younger than age 60. If we are able to screen and identify these patients early, we can refer them to a cardiologist for further risk assessment, potentially altering the trajectory of their health." The researchers highlighted that AI-enabled mammogram screening tools could significantly enhance the identification of women displaying early signs of cardiovascular disease, leveraging a screening test many women already routinely undergo.
Understanding Breast Arterial Calcification (BAC)
A buildup of calcium in blood vessels, known as calcification, is a recognized sign of cardiovascular damage. It is closely associated with the progression of atherosclerosis – the hardening and narrowing of arteries due to plaque accumulation – and is an indicator of early-stage heart disease or the natural process of aging. Previous epidemiological studies have consistently demonstrated a strong correlation between the presence of calcium buildup in the arteries and an elevated risk of cardiovascular events. Specifically, women identified with breast artery calcifications have been shown to face a 51% higher risk of developing heart disease and experiencing a stroke compared to those without such calcifications. This substantial increase underscores BAC’s potential as a powerful, yet historically underutilized, biomarker. It is crucial to differentiate breast arterial calcification from mammographic microcalcifications, which are tiny specks of calcium found within breast tissue that can be an early sign of breast cancer. BAC refers specifically to calcium deposits within the walls of the arteries themselves, indicating systemic vascular health rather than breast malignancy.
The AI Breakthrough: A New Era in Diagnostic Imaging
To develop this innovative screening tool, the research team employed a deep-learning AI model, a sophisticated form of machine learning capable of identifying complex patterns in vast datasets. The model was rigorously trained to "segment" calcified vessels within mammogram images. In the context of medical imaging, segmentation involves delineating the precise boundaries of specific structures – in this case, the bright pixels on X-rays that represent calcium deposits in the arteries. This segmentation approach is a critical differentiator, setting this model apart from previous AI models developed for analyzing breast artery calcifications, which often relied on less precise detection methods. By segmenting the calcifications, the AI can more accurately quantify their extent and distribution.
The robustness of this AI model is further strengthened by its reliance on an exceptionally large and comprehensive dataset for both training and testing. The study incorporated images and electronic health records (EHR) from over 56,000 patients who underwent mammograms at Emory Healthcare between 2013 and 2020. Crucially, each patient in the dataset had at least five years of follow-up EHR data, allowing researchers to correlate BAC findings with actual cardiovascular outcomes over an extended period. This longitudinal data is essential for validating the predictive power of any new biomarker.
Dr. Dapamede emphasized the broader implications of this technological leap: "Advances in deep learning and AI have made it much more feasible to extract and use more information from images to inform opportunistic screening." This statement highlights a paradigm shift in medical imaging, moving beyond the primary diagnostic purpose of an image to glean additional, often hidden, health insights.
Study Findings: Precision in Risk Stratification
The overall findings of the study demonstrated that the new AI model performed exceptionally well in characterizing patients’ cardiovascular risk, stratifying individuals into low, moderate, or severe categories based solely on their mammogram images. The model was designed to calculate the risk of dying from any cause or suffering an acute heart attack, stroke, or heart failure at two-year and five-year intervals.
A key finding was that the rate of these serious cardiovascular events significantly increased with higher levels of breast arterial calcification in two crucial age categories: women younger than 60 and those aged 60-80. Notably, this association was not observed in women over 80, suggesting that for the oldest demographic, other comorbidities and general aging processes may overshadow the predictive power of BAC. This age-specific insight makes the AI tool particularly well-suited for providing an early warning of heart disease risk in younger women, a demographic that stands to benefit most from timely interventions. Early identification allows for proactive lifestyle modifications, targeted medical management, and closer monitoring, potentially preventing or delaying the onset of severe cardiovascular events.
Further illustrating the model’s predictive accuracy, the results revealed a stark difference in event-free survival rates. Women with the highest level of breast arterial calcification (quantified as above 40 mm²) exhibited a significantly lower five-year rate of event-free survival compared to those with the lowest level of calcification (below 10 mm²). For instance, only 86.4% of women with the highest BAC survived for five years without a major cardiovascular event, in contrast to 95.3% of those with the lowest level of calcification. This translates to an approximately 2.8 times higher risk of death within five years for patients presenting with severe breast arterial calcification compared to those with little to no BAC. These figures underscore the clinical significance of BAC as a powerful, independent predictor of adverse cardiovascular outcomes.
Clinical Implications and the Future of Preventative Cardiology
The integration of this AI-powered mammography analysis into routine clinical practice holds profound implications for preventative cardiology and women’s health. Currently, standard cardiovascular risk assessment tools, such as the Framingham Risk Score or the ASCVD Risk Estimator, rely on factors like age, cholesterol levels, blood pressure, smoking status, and diabetes. While effective, these tools can sometimes underestimate risk in women or miss early signs of subclinical disease. The addition of an objectively quantified BAC score from a routine mammogram could provide a powerful, complementary piece of information, leading to more precise risk stratification.
Once implemented, a woman identified with moderate to severe BAC could be flagged for earlier referral to a cardiologist. This proactive approach would enable a more comprehensive cardiovascular workup, including discussions about lifestyle modifications (diet, exercise, smoking cessation), potential initiation of preventive medications (e.g., statins), and closer monitoring for developing symptoms. This "opportunistic screening" model leverages existing healthcare infrastructure, minimizing the need for additional appointments or specialized tests, thereby reducing barriers to early detection.
The collaboration between Emory Healthcare and Mayo Clinic in developing this AI model highlights the growing trend of interdisciplinary research in medicine. Radiologists, cardiologists, data scientists, and AI engineers are working together to unlock new diagnostic capabilities from existing imaging modalities.
Regulatory Pathway and Broader Horizons
Despite its promising findings, the AI model developed by Emory Healthcare and Mayo Clinic is not yet available for widespread clinical use. The path to commercial availability involves several critical steps. First, the model must undergo rigorous external validation in diverse patient populations and healthcare settings to ensure its generalizability and robustness beyond the initial study cohort. This independent validation is crucial for establishing the tool’s reliability and reproducibility. Following successful validation, the AI model will need to secure approval from the U.S. Food and Drug Administration (FDA). The FDA’s regulatory process for AI-powered medical devices is evolving but typically requires demonstration of safety, effectiveness, and clinical utility.
If the AI model successfully navigates these regulatory hurdles and gains FDA approval, researchers anticipate that the tool could be made commercially available. This would allow other healthcare systems to seamlessly incorporate the AI analysis into their routine mammogram processing and follow-up care protocols.
Beyond its immediate application to cardiovascular risk, the researchers are already exploring the broader potential of similar AI models. They plan to investigate how other biomarkers for conditions such as peripheral artery disease (PAD) and kidney disease might be extracted from mammograms or other routinely performed imaging studies. This vision points towards a future where medical images, once interpreted for a single primary purpose, become rich data sources for a multitude of health insights, all powered by advanced AI.
Challenges and Ethical Considerations
While the promise of AI-enhanced mammograms is immense, the path to widespread adoption is not without challenges. Ensuring equitable access to this technology across different socioeconomic groups and geographic regions will be critical. There is also a need for comprehensive training for healthcare professionals – radiologists, cardiologists, and primary care physicians – on how to interpret and act upon these new AI-generated risk scores.
Ethical considerations, such as data privacy and the potential for algorithmic bias, must also be carefully addressed. As large datasets of patient health information are used to train AI models, robust security measures and ethical guidelines are paramount to protect patient confidentiality. Furthermore, AI models must be continuously monitored and refined to ensure they perform accurately and equitably across diverse patient demographics, avoiding any perpetuation or amplification of existing health disparities. Over-diagnosis or "over-medicalization" is another potential concern, where patients might be unnecessarily alarmed or undergo further tests based on findings that may not always translate to immediate clinical urgency. Careful guideline development will be essential to manage these scenarios effectively.
Conclusion
The study presented at ACC.25 represents a significant leap forward in preventative medicine and the application of artificial intelligence in healthcare. By transforming routine mammograms into powerful tools for cardiovascular risk assessment, especially in younger women, this research offers a compelling opportunity to identify and intervene in heart disease earlier than ever before. As this AI model moves through validation and regulatory approval, it holds the potential to profoundly impact women’s health, redefine the scope of diagnostic imaging, and usher in a new era of proactive, personalized cardiovascular care. The integration of AI promises not just more efficient healthcare, but a future where hidden health risks are routinely uncovered, allowing for timely interventions that save lives and improve quality of life.
