End-diastolic wall thickness (EDWT) and thickness of the residual non-contrast-enhanced myocardial rim have been suggested as markers for the assessment of myocardial viability by cardiovascular magnetic resonance (CMR) imaging. This study compared these parameters as derived from contrast-enhanced CMR images for the prediction of myocardial viability as determined by fluorine-18 deoxyglucose positron emission tomography (FDG-PET). Twenty-two patients with ischemic cardiomyopathy (ejection fraction 31 +/- 11%) were investigated. For contrast-enhanced CMR imaging, a standard inversion-recovery sequence was used. FDG-PET was performed using a hyperinsulinemic-euglycemic clamp. Data were analyzed with a 17-segment model. Of 146 severely dysfunctional segments, 112 were assessed as viable and 34 as nonviable by nuclear imaging. Using receiver-operator characteristic analysis, areas under the curve were 0.95 for unenhanced myocardial rim (95% confidence interval 0.92 to 0.98) and 0.86 for EDWT (95% confidence interval 0.80 to 0.93, p <0.001 vs unenhanced myocardial rim) for the prediction of viability as assessed by FDG-PET. Cutoffs of 5.4 mm for EDWT and 3.0 mm for unenhanced myocardial rim were found to optimally differentiate viability by FDG-PET. In 25 segments with divergent results, 94% of segments with an EDWT < or =5.4 mm and an unenhanced myocardial rim >3.0 mm were scored as viable by FDG-PET, whereas 57% of segments with an EDWT >5.4 mm and an unenhanced myocardial rim < or =3.0 mm were scored nonviable with the reference technique. In conclusion, unenhanced myocardial rim is superior to EDWT for the prediction of myocardial viability as determined by FDG-PET and may be clinically useful for assessment of myocardial viability in patients with ischemic cardiomyopathy and regional wall thinning.