Fractional flow reserve (FFR) remains the gold standard for assessing myocardial ischemia; however, its invasive nature limits widespread clinical application. Computed tomography–derived FFR (CT-FFR) provides a noninvasive alternative, though its diagnostic consistency across different stenosis severity levels remains incompletely characterized. This study aims to evaluate the diagnostic accuracy and correlation of CT-FFR compared with invasive FFR, stratified by stenosis severity, to improve precision in ischemia evaluation.

In this prospective study, 138 patients (mean age 62.4 ± 9.7 years; 64.5% male) with suspected or confirmed coronary artery disease (CAD) and stenosis ranging from 30% to 90% in major coronary arteries (diameter ≥ 2.0 mm) underwent coronary CT angiography (CCTA), followed by CT-FFR analysis (Coronary Scope, Shenzhen Yueying Technology Co., Ltd., China). Invasive FFR (St. Jude Medical, Inc., USA) was performed within 15 days after CCTA as the reference standard. Both CT-FFR and FFR evaluations focused on a single target vessel per patient. Ischemia was defined as FFR ≤ 0.80, with the same threshold applied to CT-FFR. Stenosis severity was categorized into three groups: 30%–49%, 50%–69%, and 70%–90%. Diagnostic performance metrics—including sensitivity, specificity—and correlation coefficients (r) were analyzed.

Fifty-three patients exhibited CT-FFR values ≤ 0.80. Among the 138 evaluated lesions, two (1.4%) were located in the left main artery, 99 (71.7%) in the left anterior descending artery, 28 (20.3%) in the right coronary artery, and nine (6.5%) in the left circumflex artery. Stenosis distribution was as follows: 36 vessels had 30%–49% stenosis, 64 vessels had 50%–69% stenosis, and 38 vessels had 70%–90% stenosis. The mean CT-FFR was 0.81 ± 0.11, and the mean invasive FFR was 0.80 ± 0.15. CT-FFR demonstrated excellent diagnostic performance, with sensitivity, specificity, and accuracy of 96.2%, 97.7%, and 97.1%, respectively. A strong overall correlation was observed between CT-FFR and invasive FFR (r = 0.832, 95% CI: 0.773–0.877; p < 0.001). Subgroup analysis revealed progressively stronger correlations with increasing stenosis severity: moderate for 30%–49% stenosis (r = 0.700, 95% CI: 0.483–0.836; p < 0.0001), strong for 50%–69% stenosis (r = 0.755, 95% CI: 0.625–0.844; p < 0.0001), and nearly perfect for 70%–90% stenosis (r = 0.914, 95% CI: 0.840–0.955; p < 0.0001). The correlation in the 70%–90% stenosis group was significantly superior to that in the other groups (p < 0.05). 

CT-FFR demonstrates high diagnostic accuracy across a broad spectrum of coronary stenosis (30%–90%), with particularly outstanding performance in cases of severe stenosis (70%–90%). These findings highlight the potential of CT-FFR as a first-line noninvasive tool for guiding revascularization strategies, especially in high-grade stenosis, and support its integration into precision-oriented CAD management protocols.