Linking microstructural remodeling detected by cardiac diffusion tensor imaging to electrophysiological vulnerability and ventricular fibrillation wave break in chronic myocardial infarction

Background Non-invasive identification of functionally critical sites of ventricular arrhythmias remains a major challenge in clinical practice. Whether microstructural remodeling after myocardial infarction detected by cardiac diffusion tensor imaging (cDTI) can non-invasively predict electrophysiological (EP) alterations and localise sites of wave break at ventricular fibrillation (VF) initiation is unknown. Methods In a chronic infarct porcine model (5 weeks post LAD balloon-occlusion) in-vivo global epicardial mapping was performed using a 256-electrode sock to record unipolar activation time (AT), repolarisation time (RT) and estimate activation-recovery intervals (ARI) during right ventricular pacing including short-coupled S2. AT, RT and ARI dispersion was quantified as the standard deviation (SD) of the metrics across each electrode and its neighbours. VF was induced with programmed ventricular stimulation, and sites of wave break identified by phase mapping (Hilbert transform, topological charge technique). Ex-vivo 3D cDTI models (b-values 100 s/mm2 & 1250 s/mm2, 30 directions) were acquired using MRI-markers to enable co-registration with 3D electrode locations. Established cDTI metrics (Mean Diffusivity (MD), Fractional Anisotropy (FA), Mode, Helix Angle (HA), Transverse Angle (TA)) were quantified and correlated with EP metrics. Regression models assessed the predictive value of DTI metrics for each EP metric. Results 8 female pigs (73±5.2kg) with transmural anterior infarct in MRI were studied. Compared with remote normal myocardium, scar regions exhibited prolonged ARI duration (234.8±34.4 vs 226.2±34.6ms, p=0.014) and increased ARI dispersion (15.5ms [23.3] vs 6.8ms [7.1], p<0.001). cDTI detected significantly elevated MD, reflecting increased diffusivity and microstructure disruption (0.75±0.22 vs 0.48±0.09 x 10-³mm²/s, p<0.001), reduced FA, indicating disarray (0.22±0.05 vs 0.32±0.07, unitless) and less negative epicardial HA (-11.7±29.0° vs -19.2±22.9°, p<0.001) within scar. Dynamic ARI-dispersion during short-coupled pacing correlated with MD (CC +0.46, p<0.001) and FA (CC -0.37, p<0.001). cDTI-metrics predicted ARI-duration (adj.R²=0.127) and ARI-dispersion (adj.R²=0.276), both p<0.001. MD was the strongest positive predictor for ARI and ARI-dispersion, FA was a significant negative predictor for ARI-dispersion (all p<0.001). Sites of wave break at VF onset were characterised by elevated MD (0.63±0.17x 10-³mm²/s) and reduced FA (0.27±0.07) compared to normal myocardium, but 'near-normal’ epicardial HA (-17°±19°). Conclusion cDTI-derived microstructure alterations correlate with increased ARI-dispersion during pacing and may support identification of sites with increased susceptibility to wave break initiating VF. This is the first study to provide evidence for cDTI as a non-invasive tool to characterise arrhythmogenic microstructural patterns after infarction and predict regions of electrophysiological instability.