Preliminary application of in vivo cardiac diffusion weighted MRI in chronic myocardial infarction porcine model

Abstract

Background Cardiac diffusion-weighted MRI is a non-contrast technique that has the potential to identify changes in tissue microstructure in acute myocardial infarction (MI) in humans and rats [1,2]. The trace apparent diffusion coefficient (trADC) was found to be significantly increased in the infarcted region relative to remote regions. The increased trADC is attributed to an increase in extracellular space following cell death, where restriction in water diffusion is less. The aim of this study was to perform in vivo cardiac diffusionweighted MRI in a chronic MI porcine model to see if this increased trADC chronically persists and correlate it with scar tissue delineated by late gadolinium enhancement (LGE) imaging. Methods Four chronic MI (8 weeks post MI induced by completely occluded proximal LAD) mini pigs were scanned on a 3T MR system (Siemens Verio). Before contrast was administered, M1M2 motion compensated [3] diffusionprepared 2D segmented TSE was performed to derive trADC maps (TR/TE = 3RR/8.4 ms, a = 180°, 2.1 × 2.1 ×8m m 3 , 3 slices, diffusion TEprep = 105 ms, b = 400 s/ mm 2 ,G diff = 43 mT/m, three orthogonal DW directions). Diffusion preparation was applied in the most quiescent phase identified by CINE imaging (35 phases). Contrast LGE GRE imaging was performed roughly 15 min post contrast injection (TR/TE/TI = 326/1.47/300 ms, a =2 0°, 1.3 ×1 .3 ×8m m 3 , 12 slices). The lateral wall of the most basal slice (furthest from the LAD occlusion) was designated as the remote region. The region of infarction was determined by elevated trADC and signal intensity in LGE (both required μinfarct >μremote +5 sremote). For each pig, the mean trADC’s calculated for the infarct and remote regions. The infarct area determined by elevated trADC was compared to the LGE infarct area. Differences in means were tested for significance with a two tailed paired t-test. Cardiac phase mismatch between trADC maps and LGE images was corrected using a nonlinear B-spline registration, whose transform was derived from two phases of CINE closest to the trADC map and LGE. Results M1M2 diffusion-prepared TSE scans resulted in mean infarct trADC values of 2.4 ± 0.2 × 10 -3 mm 2 /s and mean remote trADC values of 1.4 ± 0.4 × 10 -3 mm 2 /s (p < 0.05). The location of elevated trADC infarct region agreed well with the LGE infarct area. The mean trADC and mean LGE infarct area was 0.9 ± 0.1 cm 2 and 0.8 ± 0.1 cm 2 , respectively (p = 1). In the infarct regions, wall motion was akinetic. Conclusions In this preliminary study, infarct regions delineated by elevated trADC correlated well with infarct regions defined by protocol matched LGE in area size and location. The roughly 50% increase in trADC values found in the infarct region relative to the remote region was statistically significant. Further histological studies and larger sample size are needed to confirm the origin of the increase in trADC and its relationship with infarcted tissue.