Comparison of Diagnostic Accuracy of CT Derived Fractional Flow Reserve for Predicting Revascularization

CT-derived fractional-flow-reserve (FFRCT) is recommended for determining the indication of revascularization. Consensus on optimal fractional-flow-reserve (FFR) measurement methods and lesions is lacking. To investigate the diagnostic accuracy of FFRCT measurement methods for predicting revascularization and factors affecting FFRCT measurement methods. A total of 340 patients (382 vessels) who underwent CT angiography and showed ≥50% diameter stenosis (DS) on invasive coronary angiography were evaluated. Enrolled vessels were classified into four groups according to lesion length (focal, <40 mm; diffuse, ≥40 mm) and DS (moderate, 50%-69%; severe, 70%-99%). FFRCT was measured at the distal end of the vessel (vessel-distal FFRCT), distal to the stenotic lesion (lesion-specific FFRCT), and trans-lesion FFRCT (ΔFFRCT) and compared with DS. Vessel morphology and left ventricular mass were assessed. In focal lesions, the predictive ability for revascularization of ΔFFRCT (AUC = 0.71) was comparable to that of DS (AUC = 0.76, p = 0.166) and higher than that of lesion-specific FFRCT (AUC = 0.67, p < 0.001) and vessel-distal FFRCT (AUC = 0.63, p < 0.001). Multivariable analysis showed that lesion length was the strongest predictor of ΔFFRCT (β-coefficient = 0.42, p = 0.006). In diffuse lesions, the predictive ability for revascularization of ΔFFRCT (AUC = 0.73) was comparable to that of DS (AUC = 0.77, p = 0.667), lesion-specific FFRCT (AUC = 0.72, p = 0.653), and vessel-distal FFRCT (AUC = 0.69, p = 0.242). Multivariable analysis showed that lesion length was the strongest predictor of ΔFFRCT (β-coefficient = 0.99, p < 0.001). The predictive performance of ΔFFRCT for revascularization was similar to DS in focal and diffuse lesions. ΔFFRCT was influenced by lesion length.