Amide proton transfer (APT) imaging has been reported to be useful for assessing malignancy or evaluating treatment efficacy. In this study, we evaluated the validity of APT signals in brain tumors by direct voxel-wise comparison with standardized uptake values (SUVs) from fluorodeoxyglucose positron emission tomography (FDG-PET) on a PET-magnetic resonance (PET/MR) system. APT imaging showed discrepancies with FDG-PET due to structural inhomogeneity, and the correlation between APT signals and SUVs was poor. The correlation was significantly improved after correcting for the apparent diffusion coefficient (ADC). APT/ADC could be a reliable metabolic marker with better correlation with SUVs.