Mutations in isocitrate dehydrogenase (IDH) 1 and 2 occur in over 80% of low-grade gliomas and secondary glioblastomas (1). Wild-type IDH catalyzes the conversion of isocitrate to α-ketoglutarate (α-KG); IDH1 (cytosolic) and IDH2 (mitochondrial) mutant tumors accumulate 2-hydroxyglutarate (2-HG) as a result of a neomorphic IDH activity, which additionally catalyzes the reduction of α-KG to produce 2-HG (2,3). Prior MRS studies reported the in vivo detection of 2-HG in gliomas at 3T (4,5). Recently, and benefiting from substantial gains in signal-to-noise ratio (SNR) and spectral resolution at 7T, an elevated 2-HG accumulation in IDH2 R172K (mitochondrial) compared to the IDH1 R132H (cytosolic) mutant tumors in human brains non invasively has been demonstrated (6). On the basis that tumor metabolism is known to differ from normal tissue, we propose that different subtypes of IDH mutations might be distinguished by their characteristic 2-HG and associated metabolite profiles due to the metabolic reprogramming associated with IDH mutations. Thus, the goal of this work is to determine whether different subtypes of IDH mutation lead to additional MRS–detectable alterations. Here we demonstrate this in the case of lactate since it is by-product of anaerobic metabolism. 10 glioma patients (6 IDH1 R132H, 3 IDH2 R172K and a patient without histopathologic diagnosis) and 7 healthy volunteers were scanned under IRB ethical approval at 7T using a whole body MR system (Siemens, Erlangen) with a Nova Medical 32-channel receive array head-coil. A semi-LASER sequence with VAPOR water suppression and outer volume saturation was modified and optimized for maximum 2-HG signal at TR = 5s and TE = 110 ms (5). For each patient, spectra were acquired from a volume of interest (VOI) within the tumor of 20x20x20 mm³, 128 transients. Eddy current correction, reconstruction, and zero-order phasing of array coil spectra were carried out using a reference water spectrum acquired from the same VOI. The residual water resonance was removed using Hankel-Lanczos singular value decomposition (HLSVD) time-domain selective filtering (7). After transforming the signal to the frequency domain, any baseline offset was subtracted from the spectrum. The normalization of the spectral data vector to the L2-norm was performed based on the data-points in the region [0.5, 4.2] ppm. Finally, the spectral range restricted to [0.5, 2.9] ppm was used as an input to SpectraClassifier 3.1, an automated MRS-based classifier-development system (8). Feature selection was performed with Correlation-based Feature Subset Forward Selection and the resulting features were used as an input to a Fisher Linear Discriminant Analysis (LDA). The number of spectral features selected using the correlation analysis was set to 3 (<n/3, where n is the number of cases in the smallest group).The untargeted feature extraction of in vivo spectra from controls and glioma patients resulted in a spectral pattern deviation at 2.25 and 1.35 ppm where the 2-HG and Lac peaks are located (Figure 1). In addition to the increased 2-HG signal in IDH2 compared to IDH1 mutant gliomas, an increased lactate signal was observed in IDH2 mutants. The projection space plot of these two identified features showed a distinct clustering with a complete separation between IDH1- and IDH2-mutant tumors at 7T (Figure 2). In summary, in addition to 2-HG, lactate signal was modulated by different subtypes of IDH mutations. While a larger sample size is needed to confirm these findings, this pilot study indicates that in vivo ¹H-MRS of 2-HG and associated metabolites at 7T offers the possibility to make important contributions not only in the early differential diagnosis of brain tumors, but also more importantly in assisting the study of disease progression and treatment response that cannot be obtained with other methods.