Backgrounds: Cardiac magnetic resonance (CMR) T1 mapping and the extracellular volume (ECV) have been developed to quantitative analysis of diffusely abnormal myocardial fibrosis (MF). However, dual-energy CT (DECT) has a potential for calculation of ECV. The aim of this study is to evaluate the feasibility and accuracy of DECT technique in determining the ECV in patients with heart failure, with 3T CMR as the reference. Methods: Thirty-five patients with various reasons of heart failure were enrolled in this study. Both DECT and CMR exams were completed within 24 h. ECVs were calculated, and the relationship between DECT-ECV, CMR-ECV, and other heart function parameters, including left ventricular end systolic and diastolic volume, cardiac output and ejection fraction (LVESV, LVEDV, CO, LVEF), Brain natriuretic peptide (BNP) was determined. All participants gave informed consent, and the study was approved by the institutional review board. Results: The median ECVs on DECT and CMR were 33% (95% CI: 32%-36%) and 30% (95% CI: 30%-32%), respectively. A good correlation between myocardial ECV at DECT and that at CMR (r=0.945, P < 0.001) was observed. Bland-Altman analysis between DECT and CMR showed a small bias (2.6%), with 95% limits of agreement of -0.4% and 5.6%. Interobserver agreement for ECV at DECT was excellent (ICC = 0.907). Both ECVs, for DECT and CMR, were inversely associated with LVEF and CO. Conclusion: DECT-based ECV could be an alternative non-invasive imaging tool for myocardial tissue characterization. However, overestimation of the extent of diffuse MF is observed with use of DECT. (C) 2018 Elsevier B.V. All rights reserved.