Ordinary chest scans predict heart risk as well as dedicated calcium scans

Coronary artery calcium is one of the strongest markers of silent heart disease, but the dedicated gated scan used to measure it is a separate test many people never have. Meanwhile, large numbers of ordinary chest CT scans are performed every year for lungs, injuries and other reasons, and calcium is visible on those images too. Whether that incidental reading predicts heart events as reliably as the formal test had not been well established. If it does, risk information already sitting in medical records could be put to work in prevention.

Researchers used the Multi-Ethnic Study of Atherosclerosis, in which 2601 participants had a gated and a nongated chest CT on the same day between April 2010 and December 2011. After excluding 106 people with previous coronary or cardiovascular disease and 23 with missing data, 2472 remained; 53% were female and the group was ethnically diverse, including white, Chinese, Black and Hispanic or Latino participants. Both scan types were read at a core laboratory blinded to how each was acquired, and statistical models linked each calcium score to later coronary and cardiovascular events, comparing predictive performance head to head.

Calcium on the nongated scan carried clear prognostic weight. Compared with a score of zero, moderate calcium (101 to 299) more than doubled coronary heart disease risk, with a hazard ratio of 2.67, while severe calcium (300 or more) raised it more than fivefold, with a hazard ratio of 5.22; cardiovascular disease showed the same stepped pattern. Scores from the two scan types were almost perfectly correlated (r=0.961). On formal measures of discrimination and overall accuracy, the gated and nongated scores performed statistically similarly.

The result supports a shift already under way: the 2026 American College of Cardiology and American Heart Association dyslipidemia guidelines endorse using incidental calcium from nongated scans to stratify cardiovascular risk and guide lipid-lowering therapy. This study supplies the paired, same-day comparison that strengthens that endorsement. Open questions remain about how such scores should be reported and acted on at scale, and prediction is not the same as proof that responding to a score improves outcomes.

If you have ever had a chest CT for any reason, it may already contain a usable estimate of your coronary calcium, and this study suggests that estimate forecasts risk about as well as the dedicated test. For someone weighing a borderline decision about cholesterol treatment, that is potentially useful information to raise with a clinician. It is context, not a directive; the study measured prediction, not the benefit of any action taken in response.