AI That Flags Your Heart Risk Years Early

AI can now turn a routine mammogram into a five‑year warning signal for a woman’s heart.

Story Snapshot

AI can read tiny calcium deposits in breast arteries and flag hidden heart risk years early.
Women with severe breast arterial calcification face about 2–3 times higher event risk, not 10x.
This happens with no extra scan, no extra cost, and no extra radiation—just smarter use of images.
The science is strong but still retrospective, so it has not yet reshaped official heart guidelines.

How a cancer test became an early warning for the heart

Every year, millions of women get mammograms to look for breast cancer. Radiologists focus on tumors, but those same images quietly record something else: chalk-white lines of calcium in the breast arteries. For decades, most doctors shrugged at these streaks. They were “incidental findings,” noted or ignored. Now teams at Emory and Mayo Clinic have shown that those quiet streaks are loaded with information about who will have heart attacks, strokes, and heart failure in the next few years.^[1]

Researchers trained artificial intelligence on more than one hundred twenty thousand mammograms to measure this breast arterial calcification, or BAC, in precise square millimeters.^[1] Older work graded BAC in fuzzy “mild, moderate, severe” buckets. This model turns it into a real number, like a coronary calcium score, but for the breast.^[1] That matters, because continuous scores let doctors track risk over time, see whether treatment is working, and avoid lumping very different women into the same broad category.

What the numbers really say about risk

The big question is simple: does more BAC really mean more danger? The answer from the European Heart Journal study is yes, and in a stepwise way.^[1] Women with severe BAC had around a 2.8 to 3.3 times higher risk of major heart events or death than women with none.^[10] Each tiny 1 mm² bump in BAC area added about 1% to 3% more risk across outcomes.^[10] That is not clickbait math; it is the sober, peer-reviewed signal that held up even after adjusting for standard risk scores.

This is where the messaging gap appears. In a short Mayo Clinic video, Dr. Imon Banerjee says that women with severe calcification have “10 times more risk” of heart disease within five years than women without it.^[2] The most rigorous published hazard ratios cluster closer to threefold, not tenfold.^[10] Critics will pounce on that gap, and they would be right to demand careful language. The power of this tool is strong enough without stretching the headline number beyond what the data support.^[10]

Why this matters most for younger women

Standard heart risk tools lean heavily on age, blood pressure, cholesterol, and diabetes. Those work fairly well for older men. They miss a lot of younger and middle-aged women. Studies show that BAC picked up extra risk in women under 50 who looked “low risk” by traditional models.^[2] In some cohorts, BAC flagged one in three women labeled low risk by standard scores as actually having elevated risk.^[4]

Importantly, BAC is not replacing established risk tools; it is adding to them. The artificial intelligence score stayed a strong predictor even after adjusting for major risk factors and formal atherosclerotic cardiovascular disease scores.^[4]^[8] That suggests BAC reflects artery damage and vascular aging that lab numbers alone cannot see. For women who do “everything right” yet still suffer heart events, this offers a concrete clue rather than vague talk about “bad luck” or “stress.”

What AI gets right, and what still worries the skeptics

On the technical side, the artificial intelligence model has done its homework. Against expert human tracings of BAC, it reached a correlation around 0.95 and high sensitivity and specificity for detecting calcification.^[1]^[8] It also held up across different scanner brands and breast densities.^[1] Other teams have reported similarly strong accuracy, with areas under the curve near 0.98 for identifying BAC from routine mammograms.^[8] That is a solid measurement tool, not science fiction, and it directly addresses the old complaint that BAC scoring was too subjective.

A routine mammogram may do more than detect breast cancer; it could also help identify early signs of heart disease.

Dr. Imon Banerjee, scientific director of the Arizona Advanced AI and Innovation Hub at Mayo Clinic, explains how breast arterial calcification (BAC) is captured… pic.twitter.com/inmy3feyYQ

— Mayo Clinic (@MayoClinic) June 18, 2026

The weak spot is not prediction; it is proof of benefit. Every major BAC paper so far has been retrospective, mining old images and following what happened later.^[1]^[2]^[4]^[9] None has yet shown that acting on a high BAC score—by adding statins, tightening blood pressure, or pushing lifestyle changes—actually cuts heart attacks and deaths. That is the conservative line in the sand: before turning a score into policy, run a proper trial. Until then, BAC is a powerful warning light, not a mandate for new pills.

Politics, money, and the slow grind of adoption

Guideline writers move slower than innovation, sometimes for good reason. Current American College of Cardiology and American Heart Association risk tools do not list BAC as a standard marker.^[7] Some cardiologists and radiologists worry about overreliance on artificial intelligence scores, about bias in algorithms, and about who profits if the tool is proprietary.^[17] Those concerns echo broader warnings that medical artificial intelligence, while promising, can be opaque, uneven across populations, and hard to hold accountable when it fails.^[17]

The right path runs between hype and paralysis. The evidence that artificial intelligence–measured BAC predicts risk, and does so independently, is strong.^[1]^[2]^[4]^[8]^[9] The promise is huge: no extra visit, no extra radiation, no new lab test—just reading more from the image already taken.^[2]^[11] But the field must clean up its own messaging, avoid “10x risk” shortcuts, and deliver real-world trials that show fewer women burying husbands and children because their heart risk was missed when it was still fixable.