Introduction: Myocardial infarction is associated with dangerous ventricular tachycardia (VT), a major cause of sudden cardiac death. During healing, necrotic cells are replaced by heterogeneous fibrosis, leading to tissue structural and electrical remodelling, and VT occurrence. Unfortunately, current VT diagnostic methods are invasive and often inaccurate. Owing to its excellent tissue contrast, MRI has become an ubiquitous non-invasive modality to characterize myocardial structure. This preclinical study investigates if MRI can identify markers of heterogeneous fibrosis causing VT, by employing two free-breathing MR methods: a custom multi-contrast T1* mapping technique and conventional late gadolinium enhancement (LGE), respectively.
Methods: 3D MR imaging was performed in n=7 chronically infarcted swine, at 1.4mm isotropic resolution. All pigs also underwent VT inducibility tests, mimicking catheter-based clinical protocols. For analysis, we first computed the amount of heterogeneous fibrosis replacing the chronic infarct (i.e., dense scar together with the VT substrate identified as 'grey zone' in MR images). Specifically, we used a fuzzy C-means clustering algorithm to obtain T1* maps, and the clinically accepted 5SD (standard-deviation) threshold for signal intensity for LGE image segmentation. Finally, the VT cycle length recorded for each case during the VT test was correlated with the volumes of fibrosis derived by each MR method.
Results: our quantitative analysis showed that the T1*-derived fibrosis volume correlated better with VTCL (R2~0.81) than LGE-derived fibrosis volume (R2~0.57). Additionally, the extent and patterns of histopathology-defined fibrosis matched best with those defined by T1*-mapping.
Conclusion: Our preclinical pilot study suggests that 3D T1* mapping is more robust than conventional LGE imaging in predicting scar-related malignant VT. Furthermore, since many post-infarction patients are not hemodynamically stable and cannot undergo clinical VT tests, we envision that non-invasive predictions of VTCL based on T1*-defined fibrosis may allow clinicians to evaluate larger patient populations at risk of potentially developing lethal VT.