Five commercially available Angio-FFR and Angio-IMR methods showed modest (AUC 0.73-0.75) and poor (AUC 0.49-0.69) diagnostic accuracy, respectively, compared to pressure-wire standards.
Do Angiography-derived FFR and IMR methods accurately assess epicardial and microvascular physiology compared to pressure-wire reference standards in coronary vessels?
Current commercially available Angio-FFR and Angio-IMR methods demonstrate modest to poor diagnostic accuracy compared to pressure-wire standards, particularly in vessels with small minimum lumen diameters.
Tasa de eventos absoluta: 0% vs 0%
Abstract Background Beginning with automated luminal contour detection and quantitative coronary angiography (QCA), computational fluid dynamics (CFD) has enabled the derivation of Angiography-derived Fractional Flow Reserve (Angio-FFR) and Angiography-derived Index of Microcirculatory Resistance (Angio-IMR) directly from angiographic images. This approach offers integration into routine PCI workflows, supported by growing evidence and its endorsement by ESC. Nevertheless, diagnostic accuracy has varied considerably among software and trials (Figure A/B), and the reproducibility of real-world results remains a concern. Methods and Results In our prospective corelab study of 390 vessels analysed independently across five commercially available Angio-FFR methods, diagnostic performance was modest compared with pressure-wire FFR (PW-FFR), with area under the curve (AUC) values ranging from 0.73 to 0.75. (Figure A) To explore this further, we examined the relationship between QCA-derived geometric parameters and the diagnostic error of vFFR relative to PW-FFR. Minimum lumen diameter (MLD) showed the strongest association with diagnostic deviation, with smaller MLDs linked to lower accuracy and a marked improvement in accuracy once MLD exceeded 1.8 mm. (Figure C) This pattern is consistent with our earlier phantom analysis (Figure D), where limited contour-detection accuracy in small lumens was compounded by the squared function used to derive area from MLD, which is then squared again within the Lance–Gould formulation. This mathematical amplification likely explains the observed accuracy degradation in small vessel sizes. Parallel to the development of Angio-FFR, the field is moving toward wire-free assessment of the coronary microcirculation. Pressure-wire-based IMR (PW-IMR) and continuous thermodilution, derived absolute hyperaemic microvascular resistance (Rmicro), are validated methods but remain underused due to technical complexity. To address this, several Angio-IMR models have been proposed, attempting to estimate microvascular resistance from angiographic flow and pressure surrogates. However, in our comparative validation against PW-IMR and Rmicro, all five Angio-IMR algorithms demonstrated poor diagnostic accuracy (AUC 0.53–0.58 for PW-IMR ≥ 25, AUC 0.49-0.69 for Rmicro ≥475 WU), underscoring the current limitations of these wire-free surrogates in detecting microvascular dysfunction. (Figure E/F) Conclusions A clear need exists for improved contour-detection algorithms, standardisation of CFD assumptions, and comprehensive validation across diverse vessel morphologies and disease patterns. Future integration of Angio-FFR and Angio-IMR may enable a full physiological assessment, "epicardial and microvascular" from a single angiographic acquisition. Until then, balanced interpretation and careful awareness of technical constraints remain essential to ensure that physiological assessment remains accurate, reproducible, and clinically meaningful.For image description, please refer to the figure legend and surrounding text.
Oshima et al. (Sun,) reported a other. Five commercially available Angio-FFR and Angio-IMR methods showed modest (AUC 0.73-0.75) and poor (AUC 0.49-0.69) diagnostic accuracy, respectively, compared to pressure-wire standards.