Assessment of the human brain metabolism in vivo using magnetic resonance spectroscopic editing techniques at 3T

Proton MR spectroscopic imaging (1H-MRSI) is a metabolic MRI technique that could be added to the brain imaging protocols for a detailed investigation of the brain metabolism. Neurodegenerative diseases and brain tumors manifest changes in glutamate/glutamine (Glu/Gln) complex (Glx) and myo-inositol (mI) peaks in 1H-MRSI. These metabolites could not be separated with routine 1H-MRSI techniques, so there is a need for MR spectral editing techniques. J-resolved point resolved spectroscopy (PRESS) and echo time (TE)-averaged PRESS are MR spectral editing techniques, which allow for separation of the Glu, Gln, mI, and glutathione (GSH) metabolites. In this thesis, Monte Carlo simulations were used to optimize the number of TEs to reduce the acquisition time of 2D J-resolved PRESS for a detailed investigation of Glu and GSH metabolites with TE-averaged PRESS and accelerate the data acquisition with compressed sensing. A Wilcoxon signed-rank test was used to detect any significant differences in metabolite to creatine (Cr) and Glu/Gln ratios and signal-to-noise ratios between multi-voxel 35 ms PRESS and optimized TE-averaged PRESS. Mean metabolite to Cr ratios were compared between full and compressed sensing accelerated optimized TE-averaged PRESS data using Bland Altman statistical test. This study indicated that optimized TE-averaged PRESS successfully and separately quantified the peaks of Glu at 2.35 ppm, and compressed sensing reconstruction could be applied for further reduction of scan time of optimized TE-averaged PRESS.