ABSTRACT Magnetic resonance spectroscopy (MRS) enables noninvasive assessment of brain metabolites and is commonly implemented using single‐voxel spectroscopy (SVS) or magnetic resonance spectroscopic imaging (MRSI). This study directly compares the reproducibility of single‐voxel sLASER and 3D‐Concentric Ring Trajectory–based Free Induction Decay MRSI (3D‐CRT‐FID‐MRSI) at 3 T and 7 T in the same cohort. Five healthy adults were each scanned twice on a 3 T PrismaFit (Siemens; 64‐channel head coil) and a 7 T MAGNETOM 7 T Plus (Siemens; 32‐channel head coil), with sessions 5–9 days apart. To explore MRSI's capabilities for regional metabolite quantification and reproducibility assessment, three masking strategies were applied. Additionally, two spatial averaging approaches for MRSI, averaging before vs. after spectral fitting, were evaluated. Coefficients of variation (CVs) and voxel‐wise correlation analyses were used to assess intrasubject and intersession reproducibility. Results showed good‐to‐excellent reproducibility across both techniques, with SVS generally providing lower CVs at 7 T, while MRSI outperformed SVS in several metabolites at 3 T. MRSI allowed tissue‐specific analysis, with lower CVs observed in WM compared to GM, especially at 7 T. Although MRSI reproducibility was slightly reduced at 7 T likely due to longer scan times and lack of prospective motion correction at 7 T (which was available at 3 T in this study), the spatial coverage and retrospective region analysis makes it an attractive alternative to SVS for many brain regions. This study demonstrates that both sLASER and CRT‐FID‐MRSI provide reproducible metabolite measurements at 3 T and 7 T. The findings highlight MRSI's advantages for retrospective multiregional and tissue‐specific analysis, facilitating its integration into future clinical research.
Eckstein et al. (2025) studied this question.