A coronary artery calcium (CAC) scan added only a modest amount of predictive accuracy to a standard cardiovascular risk calculator across a diverse cohort of more than 6,000 adults followed for a decade — but for people whose estimated risk sits in the ambiguous middle, it separated them sharply. Among borderline-risk adults, those with a high calcium score had roughly seven times the 10-year event rate of those whose score was zero.

The study, Predictive Utility of Coronary Artery Calcium Score Added to the PREVENT Atherosclerotic Cardiovascular Disease Equations, was published in JAMA on August 26, 2026 (DOI 10.1001/jama.2026.13233) by a Northwestern University Feinberg School of Medicine team with senior author Dr. Nilay Shah, alongside Xiaoning Huang, Lucia Petito, Norrina Allen, Philip Greenland and Sadiya Khan.

What is a calcium scan?

A CAC scan is a fast, low-dose CT of the chest performed without contrast. It measures calcified plaque in the coronary arteries and returns an Agatston score: zero means no detectable calcification, and higher numbers indicate more atherosclerotic burden. It is inexpensive relative to most cardiac imaging, and it has become popular — frequently marketed direct-to-consumer — precisely because the output feels concrete in a way that a percentage risk estimate does not.

What is PREVENT?

PREVENT is the American Heart Association’s current cardiovascular risk model. It estimates 10- and 30-year risk from routinely available inputs — age, sex, blood pressure, cholesterol — plus kidney and metabolic measures, and it removed race as an input variable. It replaced the older Pooled Cohort Equations, which had a well-documented tendency to overestimate risk in contemporary populations.

That matters here, because a better baseline model raises the bar for any add-on test. If the calculator is already good, an additional marker has less left to explain.

The findings

The researchers used the Multi-Ethnic Study of Atherosclerosis (MESA), following 6,098 adults aged 45–79 at baseline for 10 years, during which 6% had a heart attack or stroke.

Adding CAC to PREVENT moved the model’s discrimination — the C-statistic, a measure of how well it separates people who go on to have events from those who do not — from 0.73 to 0.75. That is a real but small improvement, and on its own it would be an argument against routine scanning.

The important result is not the average. It is where the information landed. For adults in the borderline or intermediate band — a 3–9% 10-year risk under PREVENT — the calcium score meaningfully reclassified people in both directions, and the spread was large: high scorers carried roughly seven times the event rate of those at zero.

Why a small average gain can still be clinically useful

This is a general property of risk prediction that is easy to misread. A C-statistic is computed across an entire population, most of whom are not near any decision threshold. For someone whose estimated risk is 1%, or 25%, a calcium score rarely changes what happens next: the first is not going to be started on preventive therapy on the strength of a scan, and the second generally already has an indication for treatment.

The people for whom a test can change a decision are those sitting close to the threshold. Under current cholesterol guidelines, coronary calcium already has a defined role as a tiebreaker for patients in the borderline-to-intermediate range, where the case for a statin is genuinely uncertain. What this study does is put a number on how much that tiebreaker is worth — substantial in that band, marginal outside it.

“Our findings suggest that not everyone necessarily needs or would benefit from a coronary artery calcium scan for the purpose of predicting risk of heart attack and stroke,” Shah said. For borderline patients, he noted, “knowing their calcium score can help determine whether their risk is actually lower or higher than initially estimated, which can guide treatment decisions.”

The cost of scanning everyone

Scanning low-risk people is not free of consequences. It carries a small radiation dose, out-of-pocket cost, and the possibility of incidental findings in the chest that generate follow-up imaging, specialist visits and anxiety with no demonstrated benefit. Applied broadly to a population unlikely to have events, those harms accumulate while the useful information does not.

Limitations worth knowing

MESA participants were 45–79 at baseline, so the study says nothing about adults under 45 — a group in which calcium scanning is increasingly marketed. The cohort also had limited representation of South Asian and Filipino adults, populations with elevated atherosclerotic risk where the added value of CAC is a live question.

And one distinction is essential: this is a study of prediction, not of outcomes. It shows how much a calcium score sharpens a risk estimate. It does not show that CAC-guided care leads to fewer heart attacks and strokes — that would require a randomized trial of scanning strategies, which is a different and harder question. A score of zero is also reassuring rather than protective: non-calcified soft plaque does not show up on the scan.

Research news, not medical advice. Whether a calcium scan is appropriate for any individual is a decision for that person and their clinician.