2-hydroxyglutarate (2HG) is the first imaging biomarker that is specific for IDH-mutated gliomas, which are associated with better treatment response and longer patient survival^1,2, and therefore may have important implications for use in management of patients with brain tumors. Recent development of in vivo 1H magnetic resonance spectroscopy (MRS) at 3T may provide an effective tool for noninvasive measurement of 2HG in IDH-mutated gliomas^3-5. However, the potential impact of this imaging biomarker for patients can only be realized when the technique is widely available in clinical scanners at academic/community radiology centers. The purpose of this study is to test the reproducibility of 2-hydroxyglutarate MRS between research and clinical scans.Twenty-seven glioma patients were recruited (17 IDH mutated glioma patients confirmed with biopsy, 10 IDH mutation status unknown). Each subject underwent two MRS examinations using a published 2HG-optimized PRESS TE=97ms method at 3T³; one in a clinical scanner and the other in a research scanner on the same day. The 2HG MRS in the clinical scanner were operated solely by technologists and those in the research scanner by a research member. Two patients were scanned at 4 time points in both scanners, and the total number of clinical/research scan pairs being 35. The research scanners were equipped with an 8-channel head coil and the clinical scanner with a 32-channel head coil (both. Voxel size was 4 – 8 mL depending on tumor volume identified with T2w-FLAIR. Data acquisition parameters included TR = 2 s, sweep width = 2.5 kHz, number of sampling points = 2048, and number of averages = 128 – 512. In clinical site, total scan time was less than 10 minutes and sub-echo times of PRESS sequence, which were tuned to TE₁ = 32 ms and TE₂ = 65 ms in research scanner, were TE₁ = 34 ms and TE₂ = 63 ms. Water unsuppressed data was acquired for eddy current compensation and multi-channel combination. Spectral fitting was performed with LCModel software⁶, using basis spectra calculated incorporating the volume localizing RF and gradient pulses of PRESS using published chemical shift and J-coupling constants^7,8. Metabolite quantification was conducted with reference to water at 45 M. Representative in vivo spectra of reproducibility test scans of two glioma patients are presented in Figure 1. With the high-risk of getting biopsy from the brain stem region, tumor and IDH mutation type were unknown for the first patient shown. The estimation of 2HG, as well as glutamate and glutamine was well reproduced in the two scanners (6.1 mM vs 6.5 mM; 2.7mM vs 2.7 mM; 0.8 mM vs 0.9 mM, respectively). The second patient had an IDH 2 mutated glioma (biopsy confirmed) with reproducible 2HG estimates from the two scanners (2.0 mM vs 1.7 mM) and ignorable residuals. Of the 35 pairs of MRS data sets, 4 pairs of data had singlet linewidth larger than 0.1 ppm and thus were excluded for metabolite estimation. Excluding these 4 pairs, 31 pairs of data (shown in Fig. 2) were included for subsequent statistical analysis. 2HG and total choline (tCho) estimates from the clinical and research scans were shown in Figure 3a and 3c. The concentrations of 2HG from the two scanners were found to be very similar (3.69 ± 2.47 mM vs. 3.73 ± 2.45 mM, paired t-test p value = 0.74). Small coefficient of variation (CV) (=0.11) and large intraclass correlation coefficient (ICC) (= 0.97) of 2HG estimation further confirmed the high reproducibility of the 2HG MRS between the pairs of scans. Shown in Figure 3b, 2HG estimate differences (0.03 ± 0.61 mM) between the two scanners were less than 1 mM in all cases. Similarly, small CV (0.09) and large ICC (0.97) were obtained for tCho, which is elevated in tumors, indicating good reproducibility between clinical and research scans. The present study demonstrates that evaluation of 2HG across research and clinical scans using optimized PRESS is highly reproducible, suggestive of widespread dissemination of the 2HG MRS protocol for clinical use.