Feasibility of virtual fractional flow reserve derived from coronary angiography and its correlation with invasive functional assessment

Abstract

Abstract Introduction Invasive functional assessment (iFA) of coronary artery disease (CAD) needs expensive devices, has potential procedure-related complications and is still underutilized. Virtual Fractional Flow Reserve (vFFR) derived from invasive coronary angiography (ICA) has the potential to overcome these limitations. Purpose To investigate the feasibility of vFFR analysis and its correlation with iFA (iFR, RFR or FFR). Methods Retrospective analysis of consecutive patients (pts) who underwent iFA in a tertiary center between 2019 and 2020. vFFR was calculated using a dedicated software (CAAS Workstation 8.4) based on standard non-hyperaemic coronary angiograms acquired in ≥2 different projections, by operators blinded to iFA results. Diagnostic performance and accuracy of vFFR were evaluated. vFFR was considered positive when <0.80. FFR <0.8 and iFR/RFR <0.90 were classified as positive according to current clinical standards. Results Out of 113 coronary arteries of 102 pts, vFFR was successfully analysed in 106 (94%). Reasons for vFFR analysis failure were: vessel projection overlap (48%), <2 angiographic projections (28%) and table movement while acquisition (24%). From 106 coronary arteries of 95 pts with analysable vFFR (78% male, mean age 67.8±9.7 years), 90 (85%) showed agreement with the respective iFA result. The vFFR predicted which lesions were physiologically significant and which were not with accuracy, sensitivity, specificity, positive and negative predictive values of 73%, 73%, 83%, 53%, and 92% respectively. The mean difference between vFFR and iFA were −0.0484±0.096 and Pearson's correlation coefficient was 0.533 (p<0.001). The ROC area under the curve was 0.839 (0.751–0.928, p<0.001). Conclusion FFR were feasible in 94% of cases analysed retrospectively. As compared to gold-standard iFA, vFFR had an overall moderate accuracy in detecting ischemia-producing lesions and a negative predictive value >90%. vFFR has the potential to substantially simplify physiological coronary lesion assessment and thus improve its current uptake. Funding Acknowledgement Type of funding sources: None. Bland-Altman plot between vFFR and IFA