Abstract PURPOSE Magnetic resonance spectroscopy (MRS) is a quantitative technique based on the chemical shift of protons that allows the metabolic profile of tissues to be analyzed and characterized. Standard glial tumor analysis comprises Creatine (Cr), N-acetil aspartato (NAA) and Choline (Cho) metabolites, usually altered in brain pathologies (low grade tumor -LGT-, high-grade cancer -HGT-, metastasis -MTT-, gliosis -GLS-). The aim of this study is to analyze the feasibility of MRS technique in brain glial tumor diagnosis. METHODS The prospective study, approved by the Institutional Review Board, included 68 randomized patients (39 female and 29 male, mean age 47 y, with confirmed histopathology). Data was acquired at a Essenza Siemens 1.5T scanner, single voxel (SV) at the lesion (IT) and at healthy tissue (HT) in the contralateral hemisphere. A multivoxel (CS) sample was obtained from a ROI containing HT and IT. Postprocessing was done at Syngo workstation, statistical analyses were performed using SPSS v25.0 and p-values < 0.05 were considered statistically significant. RESULTS The group analysis (by pathology) for HT demonstrates homogeneity. For pathological ROIs (SV, long_TE), rNAA and rCho indexes (r stands for Cr ratio) had a p=0.003 and p=0.048 respectively; however, aCho (NAA ratio) had a p=0.11. For pathological regions, rNAA coefficient for LGT and GLS: p=0.008, LGT and MTT: p=0.02, HGT and MTT: p=0.008 and GLS with MTT: p=0.019. Meanwhile, for the rCho index, LGT and HGT: p=0.005, LGT and MTT: p=0.008, GLS and MTT: p=0.002. Short-TE SV demonstrates inhomogeneity between HT and only the aCho ratio allows to distinguish LGT and MTT (p=0.01) and GLS and MTT (p=0.008). Parametric map (from CS) of both rCho and aCho visually describes pathological regions (83% and 72%). CONCLUSIONS Long_TE-SV allows to characterize brain tumor lesions and CS provides a spatial metabolic description of the IT localization.