Quantification of the ischemic burden on cardiac magnetic resonance perfusion maps by perfusion threshold analysis

Abstract

Background High-resolution quantitative analysis of myocardial blood flow (MBF) by cardiac magnetic resonance (CMR) perfusion has been recently proposed and validated in preclincal studies. It aims to combine the advantages of both visual assessment and quantitative analysis in a single test. However, incorrect registration between stress and rest images can hamper myocardial perfusion reserve (MPR) calculation from high-resolution perfusion maps. The aim of this study was to test the feasibility of MPR analysis from high-resolution perfusion maps as well as to test MBF threshold analysis, a novel analysis approach capable to measure the extent of ischemic myocardium with high spatial accuracy. Methods Ten patients with single-vessel coronary artery disease (CAD) and ten control subjects were analyzed at 3 Tesla. Perfusion data were analyzed on high-resolution maps and at segmental level to simulate MPR for variable degrees of image misregistration. MBF threshold analysis was then applied to high-resolution perfusion maps. In analogy to one of the methods currently in use for the quantification of late gadolinium enhancement, pixels with abnormally reduced coronary flow reserve were defined on the basis of the average MBF in areas with visually normal perfusion. In this exploratory analysis, the ischemic myocardium was defined as pixels with MBF reduced below the 90th, 95th and 99th percentile of the average epicardial MBF of the reference segment. MBF threshold analysis defines the amount of ischemic myocardium as a percentage of the LV (global ischemic burden) or of a perfusion territory (territorial ischemic burden).