Quantification of Cerebral Glucose Concentrations via Detection of the H1 ‐α‐Glucose Peak in 1 H MRS at 7 T

• Published in: Journal of magnetic resonance imaging Vol. 59; no. 2; pp. 661 - 672
• Main Authors: Kuribayashi, Hideto, Urushibata, Yuta, Imai, Hirohiko, Ahn, Sinyeob, Seethamraju, Ravi Teja, Isa, Tadashi, Okada, Tomohisa
• Format: Journal Article
• Language: English
• Published: United States 01.02.2024
• Subjects: Brain - diagnostic imaging; Brain - metabolism; Humans; Magnetic Resonance Spectroscopy - methods; Prospective Studies; Retrospective Studies; Signal-To-Noise Ratio; Water; brain glucose concentration; magnetic resonance spectroscopy; H1-α-glucose; LCModel; 7 T; baseline correction
• ISSN: 1053-1807; 1522-2586; 1522-2586
• DOI: 10.1002/jmri.28834

Sensitive detection and quantification of cerebral glucose is desired. To quantify cerebral glucose by detecting the H1-α-glucose peak at 5.23 ppm in H magnetic resonance spectroscopy at 7 T. Prospective. Twenty-eight non-fasted healthy subjects (aged 20-28 years). Short echo time stimulated echo acquisition mode (short-TE STEAM) and semi-localized by adiabatic selective refocusing (semi-LASER) at 7 T. Single voxel spectra were obtained from the posterior cingulate cortex (27-mL) using a 32-channel head coil. The H1-α-glucose peak in the spectrum with retrospective removal of the residual water peak was fitted using LCModel with a glucose basis set of only the H1-α-glucose peak. Conventional spectral analysis was performed with a glucose basis set of a full spectral pattern of glucose, also. Fitting precision was evaluated with Cramér-Rao lower bounds (CRLBs). The repeatability of glucose quantification via the semi-LASER sequence was tested. Paired or Welch's t-test were used for normally distributed values. A P value of <0.05 was considered significant. The repeatability of measures was analyzed using coefficient of variation (CV). Removal of the residual water peak improved the flatness and stability of baselines around the H1-α-glucose peak and reduced CRLBs for fitting the H1-α-glucose peak. The semi-LASER sequence was superior to the short-TE STEAM in the higher signal-to-noise ratio of the H1-α-glucose peak (mean ± SD 7.9 ± 2.5, P < 0.001). The conventional analysis overfitted the H1-α-glucose peak. The individual CVs of glucose quantification by detecting the H1-α-glucose peak were smaller than the corresponding CRLBs. Cerebral glucose concentration is quantitated to be 1.07 mM by detecting the H1-α-glucose peak in the semi-LASER spectra. Despite requiring long scan times, detecting the H1-α-glucose peak allows true glucose quantification free from the influence of overlapping taurine and macromolecule signals. 2 TECHNICAL EFFICACY STAGE: 1.