A drawback of visual assessment for late gadolinium enhancement (LGE) on cardiac magnetic resonance (CMR) is the subjectivity and reproducibility of the results. The aim of this study was to investigate the relationship between left ventricular (LV) reverse remodeling in response to optimal pharmacotherapies and the definite or discrepant mid-wall LGE with visual assessment in patients with dilated cardiomyopathy (DCM). A total of 65 patients who had been hospitalized with newly diagnosed DCM and had undergone CMR, were enrolled. When the visual assessment of the presence or absence of mid-wall LGE was confirmed by the two observers, patients were classified into either the positive- (n = 20) or negative-LGE (n = 29) groups. If there was discordance between the diagnoses of the two observers, patients were classified into the discrepant-LGE (n = 16) group. LV reverse remodeling was defined as an increase in LV ejection fraction by at least 10% concomitant with a decrease in the LV end-diastolic dimension by at least 10%. Among the three groups, the frequency of early LV reverse remodeling within a 1-year follow-up was significantly different (p = 0.0068). The frequency of LV reverse remodeling within a 1-year follow-up was 59, 31, and 15%, and over 2 years was 83, 62, and 40%, in patients with negative-, discrepant-LGE, and positive-LGE, respectively. The survival rate for composite end-points of cardiovascular mortality, sustained ventricular tachycardia, appropriate cardioverter-defibrillator discharge, or rehospitalization for decompensated heart failure was lower in positive-LGEs than in negative-LGEs (p =0.0011), whereas, there were no significant differences between both negative- and discrepant-LGEs, and discrepant- and positive-LGEs. This study showed that the discordance for LGE visual assessment occupied an intermediate position between positive and negative for LGE in LV reverse remodeling in patients with DCM.