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

Area of Science:

• Cardiology
• Radiology
• Medical Imaging Analysis

Background:

• Accurate cardiovascular disease (CVD) risk assessment is crucial for prevention, but current risk scores lack precision.
• Routine imaging, such as computed tomography (CT), offers potential for opportunistic CVD risk prediction.

Purpose of the Study:

• To evaluate if CT-derived cardiac and aortic structural features can predict major adverse cardiac events (MACE) beyond established risk assessment tools.
• To assess the added value of radiomics in CVD risk stratification.

Main Methods:

• A least absolute shrinkage and selection operator (LASSO) model was developed using radiomics features from 13,437 lung cancer screening CT scans to predict cardiovascular mortality.
• The radiomics score was compared against the PREVENT tool and coronary artery calcium (CAC) score in an independent cohort.
• Discrimination was assessed using Harrel's C-index, and MACE rates were analyzed in high-risk groups defined by the PREVENT or radiomics scores.

Main Results:

• The radiomics score demonstrated superior discrimination for MACE compared to the PREVENT score (C-index 0.66 vs 0.61).
• The radiomics score was complementary to CAC, improving overall MACE prediction (combined C-index 0.69 vs 0.66 for CAC alone).
• Individuals identified as high-risk by the radiomics score (but not PREVENT) exhibited a 3.6-fold higher MACE incidence; key predictive features included aortic surface-to-volume ratio and cardiac chamber dimensions.

Conclusions:

• CT-derived cardiac and aortic radiomics effectively identify high-risk individuals missed by conventional clinical scores.
• This imaging-based approach enhances risk stratification, particularly when combined with CAC scoring.
• Intensified primary prevention strategies may be beneficial for high-risk patients identified through radiomics analysis.