Effect of lesion characteristics on diagnostic performance of CT-derived fractional flow reserve: exploring the indications for application based on CT-FFR CHINA trial (Preprint)
Na Zhao,Yang Gao,Bo Xu,Weixian Yang,Lei Song,Tao Jiang,Li Xu,Hongjie Hu,Lin Li,Wenqiang Chen,Dumin Li,Feng Zhang,Lijuan Fan,Bin Liu
DOI: https://doi.org/10.2196/preprints.35006
2021-01-01
Abstract:BACKGROUND High diagnostic performance of coronary computed tomography angiography (CCTA)-derived fractional flow reserve (FFR; CT-FFR) in identifying flow-limiting stenosis has been confirmed. CT-FFR is recommended to assess the hemodynamic significance of coronary lesions. However, the optimal indications of CT-FFR relies on its ability to discriminating ischemia in situations of different types of lesions. And the effect of lesion-dependent factors on determining the diagnostic accuracy of CT-FFR have not been comprehensively evaluated yet. OBJECTIVE We aimed to investigate the effect of lesion-related factors on the diagnostic performance of CT-FFR with computational fluid dynamics algorithm, to promote the clinical application of it. METHODS This multicenter prospective clinical trial enrolled 317 patients with 30%–90% stenosis undergoing CCTA and invasive FFR from 5 centers across China. All target lesions were assigned into different lesion characteristics (target vessels, lesion location, lesion length, bifurcation lesions, and coronary calcification) subgroups. Diagnostic performance (accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under receiver operating characteristics curve (AUC)) of CT-FFR identifying ischemia were calculated and compared in all sub-groups. RESULTS Three hundred and sixty-six target vessels from 317 patients (mean age: 59.3 ± 9.6 years) were analyzed. The overall vessel-based diagnostic accuracy, sensitivity, specificity, PPV, NPV, and AUC of CT-FFR were 87.2%, 86.4%, 88.8%, 86.9%, 88.4%, and 0.90. Absence of bifurcation lesion group possessed the higher NPV of CT-FFR than presence of bifurcation lesion group (92.8% vs. 78.9%, p = 0.006). Whereas there was no statistically significant difference in diagnostic performance of CT-FFR among different target vessels, lesion location, lesion length, and coronary calcification sub-groups (all p > 0.05). CONCLUSIONS This study supported CT-FFR as a powerful noninvasive functional assessment tool for coronary lesions with different lesion characteristics, involving target vessel, lesion location, lesion length, and coronary calcification. While the diagnostic performance of CT-FFR was negatively affected by the presence of bifurcation lesions. CLINICALTRIAL https://clinicaltrials.gov; Unique identifier: NCT03692936.