Fast virtual functional assessment of intermediate coronary lesions using routine angiographic data and blood flow simulation in humans: comparison with pressure wire - fractional flow reserve
Michail I Papafaklis, Takashi Muramatsu, Yuki Ishibashi, Lampros S Lakkas, Shimpei Nakatani, Christos V Bourantas, Jurgen Ligthart, Yoshinobu Onuma, Mauro Echavarria-Pinto, Georgia Tsirka, Anna Kotsia, Dimitrios N Nikas, Owen Mogabgab, Robert-Jan van Geuns, Katerina K Naka, Dimitrios I Fotiadis, Emmanouil S Brilakis, Héctor M Garcia-Garcia, Javier Escaned, Felix Zijlstra, Lampros K Michalis, Patrick W Serruys, Michail I Papafaklis, Takashi Muramatsu, Yuki Ishibashi, Lampros S Lakkas, Shimpei Nakatani, Christos V Bourantas, Jurgen Ligthart, Yoshinobu Onuma, Mauro Echavarria-Pinto, Georgia Tsirka, Anna Kotsia, Dimitrios N Nikas, Owen Mogabgab, Robert-Jan van Geuns, Katerina K Naka, Dimitrios I Fotiadis, Emmanouil S Brilakis, Héctor M Garcia-Garcia, Javier Escaned, Felix Zijlstra, Lampros K Michalis, Patrick W Serruys
Abstract
Aims: To develop a simplified approach of virtual functional assessment of coronary stenosis from routine angiographic data and test it against fractional flow reserve using a pressure wire (wire-FFR).
Methods and results: Three-dimensional quantitative coronary angiography (3D-QCA) was performed in 139 vessels (120 patients) with intermediate lesions assessed by wire-FFR (reference standard: ≤0.80). The 3D-QCA models were processed with computational fluid dynamics (CFD) to calculate the lesion-specific pressure gradient (ΔP) and construct the ΔP-flow curve, from which the virtual functional assessment index (vFAI) was derived. The discriminatory power of vFAI for ischaemia- producing lesions was high (area under the receiver operator characteristic curve [AUC]: 92% [95% CI: 86-96%]). Diagnostic accuracy, sensitivity and specificity for the optimal vFAI cut-point (≤0.82) were 88%, 90% and 86%, respectively. Virtual-FAI demonstrated superior discrimination against 3D-QCA-derived % area stenosis (AUC: 78% [95% CI: 70- 84%]; p<0.0001 compared to vFAI). There was a close correlation (r=0.78, p<0.0001) and agreement of vFAI compared to wire-FFR (mean difference: -0.0039±0.085, p=0.59).
Conclusions: We developed a fast and simple CFD-powered virtual haemodynamic assessment model using only routine angiography and without requiring any invasive physiology measurements/hyperaemia induction. Virtual-FAI showed a high diagnostic performance and incremental value to QCA for predicting wire-FFR; this "less invasive" approach could have important implications for patient management and cost.
Source: PubMed