Cardiovascular magnetic resonance to predict clinical outcome in chronic total coronary artery occlusion

Abstract

Abstract Background Chronic total coronary artery occlusions (CTOs) are present in approximately 20–30% of patients undergoing invasive angiography. Despite their prevalence, the optimum management strategy of CTOs remains uncertain. A potential limitation in published trials of CTO revascularisation is their failure to incorporate systematic assessment of ischaemia/viability in informing revascularisation decisions. Aim We sought to determine the prognostic utility of ischaemia/viability assessment by cardiovascular magnetic resonance (CMR) in a large, contemporaneous, real-world CTO population. Methods We retrospectively studied consecutive adult patients with≥1angiographically identified CTO who were referred for clinical CMR imaging during a consecutive 8-year period in our centre (2010–2018). Multi-parametric CMR comprised functional assessment, adenosine-stress perfusion and scar imaging. For perfusion assessment, images were analysed qualitatively with a concurrent examination of scar images. Myocardial segments were assigned to CTO or non-CTO territories according to standard criteria, taking into account coronary dominance. Significant ischaemia was defined as ≥10% and/or ≥2 contiguous myocardial segments with hibernation. Angiographic collateral flow to the CTO territory was graded using the Rentrop classification and the Collateral Connection (CC) Score. Significant CAD in non-CTO vessels was defined angiographically as ≥50% stenosis in any epicardial coronary artery/branch with diameter ≥2mm. The composite clinical endpoint comprised all-cause mortality, myocardial infarction and heart failure hospitalisation. Results From a total of 27,201 invasive angiograms performed during the study period, 389 patients were diagnosed with CTO and underwent CMR imaging (mean age 65.0±11.0 years, 84% male). CTO was present most frequently in the right coronary artery (59% of subjects, 229/389), with left circumflex (LCx) artery involvement in 29% (112/389) and left anterior descending (LAD) artery in 29% (111/389). Collaterals with CC grade ≥2 were identified in 186 subjects (48%), and Rentrop score ≥2 in 300 (77%). Significant ischaemia was present in 61% of patients, and infarction in 71% (median infarction 8.6% [interquartile range (IQR) 4.5–14.1]. With a median follow-up time of 3.30 years [IQR 0.04–8.64], 65 (17%) met the composite endpoint. On multivariate analysis, neither significant ischaemia nor infarction was associated with the composite endpoint. However, non-CTO territory ischaemia was independently predictive of adverse outcome (hazard ratio 1.93; 95% CI 1.16–3.21; p=0.0113). Conclusion CTO-territory ischaemia and infarction are not predictive of adverse clinical outcome, challenging the assertion that CTO revascularisation may be guided by ischaemia assessment. The finding that non-CTO territory ischaemia is associated with adverse cardiovascular events warrants further investigation. Kaplan-Meier curves_CTO Funding Acknowledgement Type of funding source: Foundation. Main funding source(s): British Heart Foundation