Aortic and Cardiac Structure From Routine CT Predict Cardiovascular Risk Beyond PREVENT and Coronary Calcium (JIMG May 2026)-Article: Aortic and Cardiac Structure From Routine CT Predict Cardiovascular Risk Beyond PREVENT and Coronary Calcium

Article: Aortic and Cardiac Structure From Routine CT Predict Cardiovascular Risk Beyond PREVENT and Coronary Calcium

Pdf Summary

This study asked whether routine noncontrast chest CT scans can provide better cardiovascular risk prediction than standard clinical scores. The authors used deep-learning segmentation and radiomics to measure structural features of the heart chambers and aorta from CT images.<br /><br />They first developed a model in the National Lung Screening Trial (NLST), using 13,437 CT scans to predict cardiovascular mortality. The most informative features included left ventricular volume, left atrial size/shape, and aortic surface-to-volume ratio. The model was then tested in an external Mass General Brigham cohort of 14,577 patients without prior major cardiovascular events.<br /><br />In external validation, 6.2% of patients had a major adverse cardiac event (MACE) over a median 5.7 years. The CT-based radiomics score discriminated MACE better than the PREVENT clinical risk score and performed similarly to coronary artery calcium (CAC). It also added predictive value when combined with CAC, improving overall discrimination further. Importantly, the radiomics score identified high-risk patients who were not flagged by PREVENT, including many patients missing clinical inputs needed to calculate standard risk scores.<br /><br />When patients were grouped by radiomics risk, event rates rose in a graded fashion, even within the same CAC category. The score also reclassified some patients into higher statin-eligibility groups, suggesting it could help target prevention more precisely.<br /><br />The authors conclude that structural cardiac and aortic features from routine CT scans can detect cardiovascular risk beyond current tools and may support opportunistic screening to identify patients who could benefit from intensified preventive care.

Meta Tag

Concept CT Radiomics

Concept Cardiovascular Risk

Concept Major Adverse Cardiac Events

Concept Left Ventricular Volume

Concept Left Atrial Shape

Keywords

noncontrast chest CT

cardiovascular risk prediction

deep-learning segmentation

radiomics

left ventricular volume

left atrial size

aortic surface-to-volume ratio

major adverse cardiac event

coronary artery calcium

PREVENT risk score

CT Radiomics

Cardiovascular Risk

Major Adverse Cardiac Events

Left Ventricular Volume

Left Atrial Shape