Background Cardiac magnetic resonance imaging (CMRI) can accurately determine infarct size. Prior studies using indirect methods to assess infarct size have shown that patients with larger myocardial infarctions (MI) have a worse prognosis than those with a smaller MI. This study assessed the prognostic significance of infarct size by CMRI. Methods Cine and contrast MRI were performed in patients with coronary artery disease (CAD) undergoing routine cardiac evaluation. Results 100 patients (mean age 66±11 years, 87% male, diabetes 23%, hypertension 49%, prior MI 62%, mean ejection fraction (EF) 34±13%) underwent CMRI. Mean follow-up was 25±18 months after MRI, during which time 15 patients died. Cox regression was used to estimate risk of death associated with traditional risk factors, heart failure symptoms, EF, angiographic severity of CAD, and extent of infarct size. Evidence of MI based on CMRI was present in 91% of patients. The only two significant univariate predictors of death (all-cause) were evidence of myocardial infarction greater than 15% of left ventricular (LV) mass, and extent of LV dysfunction based on EF (p≤0.05). On multivariate analysis, the presence of MI ≥15% of LV mass was the single best independent predictor of death (p=0.01) with an adjusted relative risk of 9.9 (95% CI 1.6-63), figure 1. Conclusions The extent of myocardial infarction determined by CMRI is an independent predictor of death in patients with CAD.