Diagnostic accuracy of vessel fractional flow reserve compared to conventional diagnostic tools in patients with coronary artery stenosis

Abstract Background Wire-based fractional flow reserve (FFR) is currently the golden standard in a physiological assessment for intermediate coronary artery stenosis (CAS). However, the clinical advantage of FFR has been limited by its invasiveness. In the previous study with a small number of participants (334 patients), three-dimensional quantitative coronary angiography (3D-QCA)-based vessel FFR (vFFR) has shown non-inferiority to wire-based FFR in diagnostic accuracy. There has been no reports regarding the relationship between vFFR and other conventional physiological assessment tools. Methods The present study was a retrospective registry designed to evaluate the diagnostic accuracy of vFFR compared to the conventional reference standard (wire-based FFR ≤0.80). Between January 2019 and February 2022, patients with stable or unstable angina, and non-ST-elevation acute coronary syndrome who had undergone physiological assessments [wire-based FFR, instantaneous wave-free ratio (iFR), resting full-cycle ratio (RFR), and/or coronary computed tomography angiography-derived FFR (FFRCT)] before percutaneous coronary intervention were included for the present analysis. The exclusion criteria were ST-elevation myocardial infarction, previous coronary artery bypass grafting, cardiogenic shock, and adenosine intolerance. In the present study, we investigate the relationship between vFFR and conventional tools such as wire-based FFR, iFR, RFR, and/or FFRCT in patients with intermediate CAS, including multi-vessel disease, severe coronary calcium, in-stent restenosis, and hemodialysis. Results The study included 722 patients (mean age, 70±10 years; male, 80%) who underwent vFFR in 698 patients (97%), wire-based FFR in 711 (98%), iFR in 523 (72%), RFR in 109 (15%), and FFRCT in 48 (7%). Most patients presented stable angina (93%). In the present study, multi-vessel disease was identified only 20% of all. A total of 1108 target vessels were LAD in 549 (50%), LCx in 287 (26%), and RCA in 272 (24%), respectively. Bifurcation lesions and in-stent restenosis were present in 12% and 13%, respectively. The subjects were diagnosed several comorbidities, such as hypertension (77%), dyslipidemia (57%), diabetes mellitus (40%), and hemodialysis (12%). Overall, vFFR showed good correlations with wire-based FFR (r=0.70; p<0.001), IFR (r=0.57; p<0.001), and RFR (r=0.69; p<0.001). However, the correlation between vFFR and FFRCT was weak (r=0.36; p<0.001). Furthermore, vFFR had a good diagnostic accuracy (sensitivity, specificity, negative predictive value and positive predictive value; 81%, 88%, 82% and 87%, respectively) to identify the lesions with wire-based FFR≤0.80 (AUC, 0.87; 95% CI, 0.84-0.90; p<0.001). Conclusion In this large number cohort, the 3D-QCA based vFFR has a good correlation with wire-based FFR, and a high diagnostic accuracy to detect the lesion with FFR ≤ 0.80 even in patients with severe coronary calcification, in-stent restenosis, and hemodialysis.Scatter plotsBland-Altman plots