Photo credit: www.sciencedaily.com
Recent research unveiled at the American College of Cardiology’s Annual Scientific Session (ACC.25) indicates that mammograms, when combined with artificial intelligence (AI) models, could provide insights beyond cancer detection. This study emphasizes the potential of mammograms to assess calcium deposits in breast tissue arteries, which may serve as a marker for cardiovascular health.
The Centers for Disease Control and Prevention (CDC) advises that women aged middle-aged and older should undergo mammograms—X-ray imaging of the breast—every one to two years for breast cancer screening. With approximately 40 million mammograms administered annually in the U.S., the findings from this study are particularly relevant. While radiologists can observe breast artery calcifications in mammographic images, they typically do not quantify these findings or report them to patients or healthcare providers. The new research utilized a novel AI image analysis technique to automatically evaluate breast arterial calcification, translating these results into a cardiovascular risk score.
“This presents a unique opportunity for women to receive cancer screening alongside cardiovascular risk assessment during their mammograms,” stated Theo Dapamede, MD, PhD, who is leading the study as a postdoctoral fellow at Emory University. “Our findings indicate that breast arterial calcification is a strong predictor of cardiovascular disease, particularly for those under 60. Early identification of these patients allows for timely referrals to cardiology for further evaluation.”
Cardiovascular disease is the predominant cause of mortality in the U.S.; however, it is often underdiagnosed in women, partially due to a lack of awareness. Researchers believe that the integration of AI-based enhancements in mammogram screenings could enable the earlier identification of cardiovascular issues among women, leveraging routine tests that many already participate in.
Calcium accumulation in blood vessels is indicative of cardiovascular damage and can signify early-stage heart disease or age-related decline. Prior investigations have linked arterial calcium deposits to a 51% increased risk of experiencing heart disease or stroke in women.
In this study, the researchers employed a deep-learning AI model trained to identify calcified vessels in mammograms, where they appear as bright pixels in X-ray images, and to assess the risk of future cardiovascular events based on electronic health record data. The segmentation method utilized marks a departure from earlier AI models aimed at analyzing breast arterial calcifications, with this approach benefiting from a substantial dataset involving over 56,000 patients who underwent mammography at Emory Healthcare between 2013 and 2020 and had at least five years of follow-up health records.
“The advancements in deep learning and AI now permit the extraction of more comprehensive information from images, facilitating opportunistic screening,” Dapamede noted.
The study’s results indicated that the AI model effectively categorized patients’ cardiovascular risk levels as low, moderate, or severe based on their mammograms. After evaluating the likelihood of mortality from any cause, or experiencing an acute heart attack, stroke, or heart failure within two to five years, the model found that risk escalated with increasing levels of breast arterial calcification in two of the three age groups studied—particularly in women under 60 and those aged 60 to 80—while showing no significant correlation in women over 80. This indicates that the tool may be exceptionally useful in identifying heart disease risk in younger women, enabling them to seek early interventions.
The findings further revealed that women exhibiting the highest levels of breast arterial calcification (above 40 mm²) had markedly lower rates of event-free survival over five years compared to those with minimal calcification (below 10 mm²). Specifically, only 86.4% of women with significant calcification survived for five years, contrasted with 95.3% of those with minimal calcification. This data suggests that patients with severe breast arterial calcification face approximately 2.8 times more risk of mortality within five years than their counterparts with little to no calcification.
The AI model was collaboratively developed by Emory Healthcare and Mayo Clinic and is not yet available for clinical use. Researchers expressed optimism that, following external validation and potential approval from the U.S. Food and Drug Administration, the tool could become commercially accessible for integration into routine mammography practices. Additionally, the team plans to investigate the applicability of similar AI models for evaluations of other conditions, such as peripheral artery disease and kidney disease, that may also be identifiable through mammograms.
Source
www.sciencedaily.com