Fractional flow reserve derived from computed tomography and angiography: measuring agreement and discrepancy

Background In addition to conventional wire-derived fractional flow reserve (FFR), image-based FFR techniques, including CT-derived FFR and angiography-derived FFR, are increasingly used for ischemia assessment in chronic coronary syndromes (CCS). While both methods have shown high concordance with invasive FFR, direct comparisons remain limited. This study aimed to evaluate the agreement between CT-derived FFR and angio-derived FFR at different measurement locations and to identify factors contributing to discrepancies. Methods We retrospectively analyzed 77 coronary arteries in 29 patients (mean age: 70.4±9.6 years, 72% male) who underwent CT-derived FFR evaluation between April 2024 and January 2025. Patients were included if they had at least one vessel with CT-derived FFR ≤ 0.80 at the far distal segment and subsequently underwent coronary angiography with angio-derived FFR assessment. CT-derived FFR values were obtained at two locations: (1) distal to the lesion (within 20 mm) and (2) at the far distal coronary segment, with corresponding angio-derived FFR values extracted. Bland-Altman analysis assessed measurement agreement, and t-tests compared measurement variability. Additionally, plaque burden was analyzed using optical coherence tomography or intravascular ultrasound in the culprit lesions and CT imaging software in the others. The association between calcified plaque and FFR discrepancy was statistically evaluated. Results Bland-Altman analysis showed a mean difference of 0.02 for both locations. However, the standard deviation in distal to the lesion was lower compared to far distal measurements (0.05 vs. 0.08), resulting in narrower limits of agreement (-0.08 to 0.12 vs. -0.15 to 0.19). Measurement variability was significantly lower at distal to the lesion (t = 2.15, p = 0.037). Furthermore, FFR discrepancies >0.05 were more frequent in the lesions with calcified plaque than those without calcified plaque (68% [26/38] vs. 46% [18/39], p = 0.066). Notably, lesions with calcified plaques subtending an arc ≤180° exhibited a significantly higher rate of discrepancy compared to the other lesions (61% [23/38] vs. 21% [8/39], p < 0.001). Conclusion CT-derived FFR measurements taken distal to the lesion demonstrate smaller errors and higher agreement with angio-derived FFR compared to far distal measurements. Additionally, the presence of calcified plaques, particularly those with an arc ≤180°, significantly contributes to measurement discrepancies. These findings highlight the importance of precise image-based FFR assessment in CCS evaluation.