Coronary Computed Tomographic Angiography-Derived Radiomics Combing CT-Fractional Flow Reserve for Detecting...

Area of Science:

• Cardiovascular Imaging
• Radiology
• Medical Diagnostics

Background:

• Coronary artery disease (CAD) diagnosis relies on invasive procedures, which carry risks.
• Non-invasive imaging modalities like CCTA are crucial for early detection and risk stratification.
• CT-derived fractional flow reserve (CT-FFR) has emerged as a promising non-invasive tool, but its diagnostic performance can be further enhanced.

Purpose of the Study:

• To develop and validate a novel hybrid diagnostic model for hemodynamically significant CAD.
• To combine CT angiography (CCTA)-derived myocardial radiomics signatures with CT-derived fractional flow reserve (CT-FFR).
• To investigate the diagnostic accuracy of this hybrid model compared to existing methods.

Main Methods:

• Prospective recruitment of stable angina patients undergoing CCTA and invasive coronary angiography.
• Extraction of radiomics features from left ventricular myocardium across three perfusion territories.
• Feature selection using minimum redundancy maximum relevance and model construction with logistic regression.
• Generation of CT-FFR values and evaluation of radiomics, CT-FFR, and hybrid models using AUC-ROC, sensitivity, and specificity.

Main Results:

• A hybrid model integrating CCTA radiomics and CT-FFR achieved a superior Area Under the Receiver Operating Characteristic Curve (AUC-ROC) of 0.928.
• The hybrid model demonstrated high sensitivity (83.3%) and specificity (88.4%) for detecting significant CAD.
• This combined approach significantly outperformed radiomics or CT-FFR alone (P=0.040).

Conclusions:

• CCTA-derived myocardial radiomics signatures offer incremental value to CT-FFR in diagnosing significant CAD.
• The hybrid model shows potential for improving vessel-specific ischemia detection and non-invasive diagnostic accuracy.
• This combined CT-FFR and radiomics approach is a feasible strategy for enhancing CAD assessment.