Risk stratification for sudden cardiac death (SCD) in patients with non-ischemic cardiomyopathy (NICM) remains suboptimal. Although current guidelines rely on severe left ventricular systolic dysfunction (left ventricular ejection fraction (LVEF) \< 35%) as key predictor of arrhythmic risk and clinical indication of prophylactic implantable cardioverter defibrillator (ICD), this approach seems inadequate, since registries report that only a minority of NICM ICD carriers experience an appropriate ICD shock during follow-up, whereas out-of-hospital cardiac arrests (OHCA) occur in patients with LVEF\>35% in up to 80% of cases. Moreover, pivotal primary prevention trials (DANISH trial, long-term outcome of the SCD-HeFT trial) failed to demonstrate a net mortality benefit of ICD in patients with NICM. As for most structural heart diseases (SHD), scar-related reentry has been addressed as the pathophysiological mechanism of ventricular arrhythmias (VAs) in patients with NICM, with fibrotic tissue being the substrate of this reentry. Late gadolinium enhancement cardiac magnetic resonance (LGE-CMR) is the gold standard for the non-invasive visualization and characterization of the myocardial fibrosis and according to retrospective studies is detected in nearly 30% of patients with NICM. In latest years, several studies and subsequent metanalyses have explored the correlation between CMR-detected LGE and occurrence of VAs, showing that presence, extent, location (septal vs lateral) and patten (focal vs multifocal vs ring-like) of non-ischemic fibrosis help in stratifying arrhythmic risk. Nonetheless, scar heterogeneity (that is, inherent composition of dense scars vs border zone (BZ), presence of strands of viable myocardium within the scar) has been indicated as a potential novel predictor of VAs. In a recent prospective multicenter registry on patients with class I indication for cardiac resynchronization therapy (CRT) (\>60% with NICM), not only scar mass, but even border zone (BZ) mass and presence of BZ channels were identified as independent predictors for VT occurrence in NICM patients. This BZ mass and BZ channels can be automatically identified using a commercially available, post-processing imaging platform named ADAS 3D LV (ADAS3D Medical SL, Barcelona, Spain), with FDA 510(k) Clearance and CE Mark approval. Thus, CMR-derived BZ mass might be used as an automatically reproducible criterium to reclassify those patients with NICM at highest risk for developing VAs/SCD in a relatively short period of at least 2 years. In the present cohort study, the investigators sought to: i) evaluate the usefulness of CMR-derived BZ mass measurement and identification of heterogeneous tissue channels (HTC) (among other scar characteristics derived from image post-processing) to predict the occurrence of VT events in an international, retrospective, multicenter, unselected series of patients with NICM without previous arrhythmia evidence (main study); ii) subsequently validate these predictors of VT occurrence in a prospectively-collected multicenter cohort study (substudy 1); iii) retrospectively evaluate in the subset of patients with \> 1 LGE-CMR performed as part of standard clinical practice if any change in BZ mass and HTC presence occurs over time and if this correlates with occurrence of VAs (substudy 2).