Advancing Coronary Risk Assessment Through Combined Radiomic, Mechanical, and Hemodynamic Analysis.

Area of Science:

• Cardiovascular Imaging
• Biomedical Engineering
• Computational Fluid Dynamics

Background:

• Detecting vulnerable coronary plaques is critical for preventing cardiovascular events.
• Current methods often analyze vulnerability markers in isolation, limiting comprehensive assessment.
• An integrated approach is needed to synergistically evaluate multiple plaque characteristics.

Purpose of the Study:

• To introduce and evaluate an integrated analysis of radiomic, mechanical, and hemodynamic factors for coronary plaque vulnerability.
• To explore the synergistic contribution of these markers in predicting plaque vulnerability.
• To develop a machine learning pipeline for classifying coronary arteries and patients based on integrated features.

Main Methods:

• Analysis of 161 plaques from 39 patients using coronary computed tomography angiography (CCTA).
• Extraction of CCTA radiomic features, computation of mechanical markers via finite element simulations, and derivation of hemodynamic markers from computational fluid dynamics.
• Development of a machine learning pipeline for classification using individual and combined features.

Main Results:

• Radiomics showed high sensitivity (1.00) but lower specificity (0.69).
• Mechanics (0.94) and hemodynamics (0.97) demonstrated higher specificity but lower sensitivity (0.86).
• Integration of at least two marker types improved performance to 1.00 sensitivity and 0.97 specificity.

Conclusions:

• A multi-level integrative approach significantly enhances coronary plaque assessment.
• Hemodynamics showed superior performance over mechanics and radiomics individually.
• Combining radiomic, mechanical, and hemodynamic data adds substantial value for improved plaque vulnerability detection.