Abstract / Summary
Background Perivascular fat attenuation index (FAI) derived from coronary computed tomography angiography (CCTA) is a novel imaging biomarker of coronary inflammation. However, its incremental value in predicting functional myocardial ischemia, defined by CT-derived fractional flow reserve (CT-FFR), beyond traditional cardiovascular risk factors and epicardial adipose tissue (EAT) volume remains unclear. Methods A total of 380 consecutive patients who underwent clinically indicated CCTA with CT-FFR evaluation were retrospectively analyzed. Functional ischemia was defined as CT-FFR ≤0.80. Patients were stratified into low and high FAI groups based on the optimal cutoff value of −81.32 Hounsfield Units (HU). Multivariate logistic regression analysis was performed to identify independent predictors of ischemia. The incremental diagnostic performance of adding FAI to a baseline clinical model was assessed using area under the receiver operating characteristic curve (AUC), continuous net reclassification improvement (NRI), integrated discrimination improvement (IDI), and decision curve analysis (DCA). Results Despite comparable baseline profiles for low-density lipoprotein cholesterol, body mass index, and glycemic status ( P > 0.05), patients in the high FAI group exhibited significantly lower CT-FFR values (0.75 ± 0.13 vs. 0.79 ± 0.12, P = 0.0017). After adjusting for age, gender, traditional cardiovascular risk factors, and EAT volume, average FAI remained a robust independent predictor of functional ischemia [Adjusted Odds Ratio [OR]: 1.056 per 1 HU increase; 95% Confidence Interval [CI]: 1.019–1.096; P = 0.004]. Notably, EAT volume did not reach statistical significance in the multivariate model ( P = 0.584). The addition of FAI to the clinical model resulted in a modest, non-significant improvement in AUC (0.731 vs. 0.707, P = 0.087), but yielded significant improvement in risk reclassification (continuous NRI: 0.339, P = 0.001; IDI: 0.024, P = 0.002). Subgroup analysis demonstrated a consistent risk trend across varying degrees of stenosis, age, and body mass index categories (OR range: 1.03–1.07). DCA confirmed superior net clinical benefit for the FAI-integrated model across a wide threshold probability range (25%–75%). Conclusions Coronary perivascular FAI is an independent imaging marker for CT-FFR-defined functional ischemia and provides incremental value beyond traditional risk factors and EAT volume. These findings support FAI as a non-invasive marker of coronary inflammation that may refine risk stratification for functionally significant coronary artery disease. However, as CT-FFR mainly reflects epicardial physiology, the association between FAI and microvascular dysfunction remains to be determined.