Advanced 2-Dimensional Quantitative Coronary Angiographic Analysis for Prediction of Fractional Flow Reserve in Intermediate Coronary Stenoses

Purpose: The widespread clinical application of coronary computed tomography angiography (CCTA) has resulted in increased referral patterns of patients with intermediate coronary stenoses to invasive coronary angiography. We evaluated the application of advanced quantitative coronary angiography (A-QCA) for predicting fractional flow reserve (FFR) in intermediate coronary lesions detected on CCTA. Materials and Methods: Fifty-six patients with 66 single intermediate coronary lesions (>= 50% to 80% stenosis) on CCTA prospectively underwent coronary angiography and FFR. A-QCA including calculation of the Poiseuille-based index defined as the ratio of lesion length to the fourth power of the minimal lumen diameter (MLD) was performed. Significant stenosis was defined as FFR <= 0.80. Results: The mean FFR was 0.86 +/- 0.09, and 18 lesions (27%) were functionally significant. FFR correlated with lesion length (R= -0.303, P = 0.013), MLD (R= 0.527, P < 0.001), diameter stenosis (R= -0.404, P = 0.001), minimum lumen area (MLA) (R= 0.530, P < 0.001), lumen stenosis (R= -0.400, P = 0.001), and Poiseuille-based index (R= -0.602, P < 0.001). The optimal cutoff values for MLD, MLA, diameter stenosis, and lumen stenosis were <= 1.3 mm, <= 1.5 mm(2), >44%, and >69%, respectively (maximum negative predictive value of 94% for MLA, maximum positive predictive value of 58% for diameter stenosis). The Poiseuille-based index was the most accurate (C statistic 0.86, sensitivity 100%, specificity 71%, positive predictive value 56%, and negative predictive value 100%) predictor of FFR <= 0.80, but showed the lowest interobserver agreement (intraclass correlation coefficient 0.37). Conclusions: A-QCA might be used to rule out significant ischemia in intermediate stenoses detected by CCTA. The diagnostic application of the Poiseuille-based angiographic index is precluded by its high interobserver variability.