Diagnostic performance of CT-derived fractional flow reserve across coronary lesion morphologies: a prospective...

CT-derived fractional flow reserve (CT-FFR) enables noninvasive physiological assessment from coronary CT angiography. This study aimed to evaluate its diagnostic performance across six prespecified anatomical and calcification-related factors to clarify lesion-specific clinical applications. In this prospective multicenter trial, 317 patients with 366 target vessels underwent coronary CT angiography and invasive fractional flow reserve (FFR) within 7 days. CT-FFR performance was evaluated according to target vessel, lesion location, lesion length, bifurcation involvement, target-lesion calcification, and per-patient coronary calcium burden. These factors were selected a priori to address distinct anatomical or pathophysiological hypotheses. Subgroup analyses were exploratory, and nominal P values are reported. Overall vessel-level accuracy, sensitivity, specificity, positive predictive value, negative predictive value (NPV), and area under the receiver operating characteristic curve were 87.2%, 87.0%, 87.3%, 85.5%, 88.7%, and 0.90, respectively. Bifurcation lesions had a lower observed NPV than non-bifurcation lesions (80.4% vs. 92.0%; nominal P = 0.026). The corresponding AUCs were 0.86 and 0.93 (patient-cluster bootstrap P = 0.079), and the median absolute CT-FFR/FFR deviation was 0.050 in both groups (P = 0.870). No statistically significant heterogeneity was detected across target vessels, lesion locations, lesion lengths, or calcification strata. CT-FFR demonstrated reliable diagnostic performance across most evaluated coronary lesion morphologies. However, the lower NPV observed in bifurcation lesions suggests that negative CT-FFR findings in this setting should be interpreted with greater caution.Trial registration: ClinicalTrials.gov NCT03692936.