CAC density derived from chest CT using automated AI quantification was independently and inversely associated with MACE, providing additional prognostic value when added to CAC volume.
CAC volume was independently associated with significantly increased risk of MACE after adjusting for clinical covariates.
CAC density demonstrated a significant, independent, inverse association with MACE after adjusting for CAC volume and clinical covariates.
Ten-year restricted mean survival times (RMSTs) differed across four combined CAC volume-density groups.
Differences in time to MACE were predominantly driven by CAC volume rather than CAC density when examined within volume strata.
The study analyzed 1,109 asymptomatic adults with detectable CAC identified from a national health screening chest CT program.
AI software was used to automatically quantify CAC Agatston score and volume from non-ECG-gated chest CT scans.
This research suggests that artificial intelligence can extract meaningful heart disease risk information from routine chest CT scans that were not specifically designed for cardiac imaging. Normally, measuring calcium buildup in the coronary arteries (a strong predictor of heart attack and stroke) requires a special heart scan timed to the heartbeat. This study showed that AI software can measure both the amount (volume) and the hardness (density) of these calcium deposits from standard chest scans taken during health check-ups, and that both measurements carry useful risk information. Among more than 1,100 adults followed for nearly 8 years, those with more calcium had substantially higher rates of major cardiovascular events like heart attacks. Importantly, the density of the calcium also mattered independently: people with denser (harder) calcium deposits had lower risk than those with the same amount of softer calcium. The group with high calcium volume and low density had the worst outcomes (average 7.68 years free from events over 10 years), while the group with low volume and high density had the best outcomes (9.45 years free from events). This research suggests that the millions of chest CT scans already performed annually for lung cancer screening or other health checks could be 'opportunistically' analyzed by AI to provide heart disease risk information at no additional cost or radiation exposure to the patient. Adding calcium density to the commonly used calcium volume measurement may help doctors better identify who is at highest risk and who might benefit most from preventive treatments.
Kim N, Kim Y, Lee J, Suh Y. (2026). Opportunistic Assessment of Coronary Artery Calcium Volume and Density From Non-Electrocardiogram-Gated Chest CT Using Artificial Intelligence: Prognostic Implications in a Screening Cohort.. Korean journal of radiology. https://doi.org/10.3348/kjr.2025.1614