The utility of angiography-derived fractional flow reserve in predicting graft occlusions prior to bypass surgery

Abstract Background Occurrence of graft occlusion following coronary artery bypass graft (CABG) surgery has been linked to the presence of native coronary artery competitive flow (1). Despite its routine use in guiding revascularization, coronary angiography lacks sensitivity for assessing coronary stenosis significance. Data regarding the use of pressure-derived fractional flow reserve (FFR) to assess the hemodynamic significance of coronary stenoses prior to CABG are inconclusive (2) and the penetration of FFR in clinical routine is limited for various reasons (3). Different systems have been developed to predict FFR based on coronary angiographic images, eliminating pressure wires or hyperemic agents (4). Preliminary data demonstrated a high concordance between off-site angiography-derived FFR and pressure wire–based FFR (5-7). However, the use of this technique prior to CABG is not well studied. Purpose This study aims to correlate pre-CABG angiography-derived FFR with graft occlusion. Secondary endpoint is the correlation between the presence of at least one occluded bypass graft and overall mortality. Methods Retrospective study of patients undergoing CABG in whom a follow-up angiogram has been performed. Angiography-based FFR was calculated with a dedicated software (vFFR, CAAS 8.2 Workstation, Pie Medical Imaging), obtaining each major native coronary vessel before CABG. Post-CABG angiograms were conducted at least 6 weeks post-surgery to exclude graft-failures due to technical factors and thrombosis. Results From 2005 and 2014, 204 patients underwent pre-CABG angiography, with post-CABG angiograms at a median of 3.85 years [IQR 1.33-8.80] post-surgery. In total, 125 patients had to be excluded due to inadequate angiogram quality for vFFR assessment (outdated imaging system, improper vessel angles, vessel overlap, insufficient contrast). Consequently, 79 patients were suitable for pre-CABG vFFR analysis, including 132 grafted vessels (93 arterial and 39 saphenous grafts). The median patient age was 68 years [IQR 58-72] and mean time for clinical follow-up was 9.9±3.6 years. Vessels analyzed using vFFR were left anterior descending (LAD) (n=60, 45.5%), left circumflex (LCX) (n=39, 29.5%), ramus intermedius (RIM) (n=8, 6.1%) and right coronary artery (RCA) (n=25, 18.9%). Mean pre-CABG vFFR was 0.59±0.15. To predict graft occlusion, a vFFR cut-off value of 0.745 with an area under the curve (AUC) of 0.847 (95%CI 0.752–0.942) had a specificity of 0.722