Abstract Implantable cardioverter defibrillator (ICD) for primary prevention of mortality is an approved therapy in nonischemic cardiomyopathy (NICM) patients with ejection fraction (EF) ≤35%. However, the DANISH ICD trial enrolled 1,116 patients with NICM and EF ≤35% randomized to ICD (n=556) or no ICD (n=560), demonstrated lack of significant mortality reduction in ICD patients when compared to non-ICD patients (hazard ratio=0.87; p=0.28), but significant reduction in the risk of sudden cardiac death (hazard ratio=0.50; p=0.005). Data from prior clinical trials leading to ICD indications in NICM showed a nonsignificant but clinically meaningful reduction in mortality, and significant reductions in sudden cardiac death (DEFINITE: 458 patients; HR=0.65; p=0.08 and a subgroup analysis of the SCD-HeFT: 792 patients; HR=0.73; p=0.06). The DANISH patients were on optimal pharmacological therapy and 58% had CRT devices. Their 3-year mortality in non-ICD arm was around 10% vs. around 20% in DEFINITE and SCD-HeFT trials, which were conducted 10 years earlier, a reflection of different patient populations and different underlying therapies. Since arrhythmogenic myocardial substrate is characterized by ventricular arrhythmias detected in long-term ECG recordings, we analyzed data from MADIT-CRT trial in 416 NICM patients in NYHA class II with QRS ≥130 ms who were randomized to CRT-D and who had a 24- hour Holter ECG recording prior to device implantation. NSVT on preimplantation Holter (observed in 194 [47%] patients) was associated with hazard ratio of 3.08 (p<0.001), enlarged LVESVi >86 ml/m2 was associated with hazard ratio of 2.81 (p<0.001), and Non-LBBB QRS morphology (observed in 44 (11%) patients) with hazard ratio of 3.33 (p<0.001) in a multivariate model predicting fast VT≥188 bpm or VF. This risk prediction model could guide therapeutic decisions regarding implantation of CRT-D, CRT-P, or ICD devices in NICM heart failure patients with QRS≥120 ms. In specific aim 1 of this proposal, we propose to validated this MADIT-CRT risk stratification model in 400 NICM with wide QRS≥120ms. With an increasing interest in observations that late gadolinium enhancement assessed in cardiac MRI is associated with increased risk of sudden cardiac death in NICM patients, we also aim to: 2) determine whether abnormal cardiac magnetic resonance (CMR) imaging with late gadolinium enhancement will be predictive for fast VT/VF and will further improve risk stratification in NICM with QRS≥120 ms, and 3) to evaluate cost- effectiveness of the proposed risk stratification approach leading to utilization of less costly devices (CRTP vs. CERT-D).