Identification of Infarct Border Zone Using Late Gadolinium Enhanced MRI in Rats

Abstract

BackgroundFollowing acute myocardial infarction (MI), cardiac myocyte death and the formation of a fibrotic scar in the left ventricular free wall (LVFW) alters the biomechanical behavior of the myocardium resulting in changes to organ‐level function. Despite significant advances in understanding LVFW remodeling in the setting of MI, the extent of tissue remodeling in the regions where infarct tissue transitions to healthy tissue, commonly called the border zone (BZ), is not well understood. Additionally, the effect of alterations in biomechanical behavior of the BZ on LV pump function remains understudied. We hypothesize that image‐based cardiac modeling can effectively address this gap. To this end, our objective was to develop an automated method to identify the infarct and BZ regions in rats using ex‐vivo late gadolinium‐enhanced magnetic resonance images (LGE‐MRI) and make use of these delineations to study regional structural remodeling in MI.MethodsAnterobasal MI was induced in Wistar‐Kyoto rats (n=2) by ligating the left anterior descending artery. Prior to sacrifice, subjects in the infarct group were given gadolinium contrast agent for MRI enhancement. Animals were sacrificed at 4 weeks post‐infarct and cardiac MRI scans were performed at an isotropic 100 µm resolution. The mean value of the voxel signal‐intensity (SI) of the posterobasal region of the heart was calculated and is hereafter referred to as the baseline SI (BSI). The image voxels were then segmented in to three regions by using the following SIs as threshold value cutoffs: i) SI of greater than 125% of BSI was considered infarcted myocardium, ii) SI between 115% and 125% of BSI was considered BZ, iii) SI of less than 115% BSI was considered remote myocardium. The geometry and dimensions of the infarct region were confirmed with measurements from the excised heart of the animal.ResultsIncreased SI was noted in the anterobasal region of the infarcted hearts (Fig 1), indicating myocyte necrosis and scar formation. Mean SI of the infarct region defined via this method was 1.6‐fold larger compared to that of the remote region. The infarct region resulting from the automated identification method closely matched the gemometry and dimensions of the infarct scar region of the excised hearts.ConclusionsThe extent of myocardial remodeling in LVFW during the post‐infarct healing process determines the contractile behavior in the LVFW and influences long‐term outcomes. Automated identification of infarct, BZ, and remote myocardium regions using LGE‐MRI is a step towards image‐based characterization of the heterogenous remodeling in the LVFW. Additionally, our algorithm enables the simple definition of a BZ of varied size and could allow us to better understand the influence of distinct regional remodeling events on organ‐level function as MI proceeds from early‐stage scarring to late‐stage remodeling.