Validation of a Prediction Model From Quantitative Coronary Angiography to Detect Ischaemic Lesions as Evaluated by Invasive Fractional Flow Reserve

Abstract

Physician visual assessment (PVA) in invasive coronary angiography (ICA) is clinically used to determine stenosis severity and guide coronary intervention. However, PVA provides limited information regarding the haemodynamic significance of stenosis. This prospective study aimed to develop a model combining visual diameter stenosis (DSPVA) and quantitative coronary angiography (QCA)-derived parameters to diagnose ischaemic lesions using invasive fractional flow reserve (FFR) with pharmacologically induced maximal hyperaemia as the gold standard. A total of 103 patients (148 lesions) who underwent ICA and FFR measurement were included in the study. Quantitative coronary angiography was used to evaluate various parameters, including anatomical parameters such as lesion length (LL), minimal lumen diameter (MLD), and minimal lumen area, along with haemodynamic parameters like LL/MLD4 and stenotic flow reserve (SFR). Plaque area, a characteristic parameter of plaque, was also assessed. Lesion-specific ischaemia was defined as invasive FFR ≤0.8. The LL/MLD4 (r= -0.66, p<0.001) and SFR (r=0.66, p<0.001) exhibited inverse and positive correlations, respectively, with invasive FFR. In the multivariable logistic regression analysis, LL/MLD4 (≥10.6 mm-3 vs <10.6 mm-3; Odds ratio [OR] 10.59, 95% confidence interval [CI] 3.94-28.50; p<0.001) and SFR (≤2.85 vs >2.85; OR 4.38, 95% CI 1.63-11.79; p=0.004) were identified as the optimal dichotomised predictors for discriminating ischaemia. The area under the curve (AUC) was 0.77 using DSPVA ≥70% as a single predictor. Adding LL/MLD4 ≥10.6 mm-3 and SFR ≤2.85 into the model significantly increased the AUC to 0.87 (p<0.001). Incorporating QCA-derived haemodynamic parameters provided significant incremental value in the model's discriminatory capability for ischaemic lesions compared with visual diameter assessment alone.