Severe breast arterial calcifications on a routine mammogram mean your risk of a heart attack or stroke jumps by fourfold to eightfold — and a new AI can spot that risk without a second scan.
That finding comes from a study in the European Heart Journal led by radiologist Hari Trivedi at Emory University, who trained a model to measure something most mammography AI ignores: the bright railroad-track stripes of calcium lining breast artery walls. These breast arterial calcifications (BAC) don't signal cancer, but they do signal stiffening vessels elsewhere in the body — a direct line to cardiovascular trouble.
The Railroad-Track Marker
Breast cancer leaves tiny, clustered calcium specks. BAC looks completely different — two parallel bright lines zigzagging through tissue. Trivedi's team built an AI that quantifies the amount of BAC, not just its presence. That quantification is what makes it predictive: even small increases in BAC correlated with slightly elevated risk, and the severe end of the scale multiplied event rates by 4–8× compared to women with none.
More than 40 million mammograms are performed in the U.S. every year. Trivedi’s model, he says, “can be run on every single mammogram without any additional work.” No extra radiation, no extra appointment. The data is already on the machine.
From 120,000 Mammograms to a Risk Score
The study drew from two large cohorts: 74,124 patients at Emory Healthcare and 49,638 at Mayo Clinic sites across Arizona, Florida, Minnesota, and the upper Midwest. That's over 120,000 mammograms feeding validation. The team found that severe BAC predicted not just any cardiovascular event, but specifically heart attack and stroke — the hard endpoints that matter.
Existing FDA-cleared AI tools detect BAC, but Trivedi's model goes a step further: it quantifies the total burden. That granularity lets radiologists — or even automated reports — stratify risk instead of just flagging a binary “present/absent.”
Why a 'Bonus Finding' Matters
Mount Sinai radiologist Laurie Margolies, who was not involved in the study, calls this “a bonus finding. You come looking for breast cancer, but you also get this information. It’s no extra radiation; it’s no extra time. It’s not another appointment.” Her colleague cardiologist Mary Ann McLaughlin adds, “A lot of people could really benefit from this.”
Margolies’ own team at Mount Sinai is now studying whether patients actually follow up on cardiovascular care when BAC is included in mammogram results — and whether outcomes improve. That's the real test: detection without action is just data.
What makes this approach so compelling is the age range. Mammograms are recommended every two years starting at age 40. Trivedi notes: “If you have BAC and are under 50, you are at a higher risk of a cardiovascular event within the next 10 years.” Catching that signal a decade earlier than traditional heart screening could change the trajectory for millions of women.
Source: How breast cancer screening can predict heart disease risk
Domain: scientificamerican.com
Comments load interactively on the live page.