Single-Voxel Proton Spectroscopy at 7 telsa: Optimization and Establishment of Protocol

Proton magnetic resonance spectroscopy (1H-MRS) offers an opportunity to probe tissue biochemistry, non-invasively and in vivo. It is a fast-evolving NMR technique with vast research applications. One of the current advances in 1H-MRS research is pulse sequence development and optimization. The choice of pulse sequence parameters in a typical 1H-MRS study determines both the number and nature of spectra recorded. For the first time, 1H-MRS was being implemented on a 7 T MRI/MRS research scanner installed at the St. Olavs Hospital in Trondheim, Norway. There was therefore the need to empirically determine optimum pulse sequence parameters in an appropriate pulse sequence, which could be used for future 1H-MRS studies at the research facility. The standard STEAM pulse sequence was compared with the sLASER pulse sequence and spectra obtained from these two pulse sequences were compared on five main objectively quantifiable indicators namely, signal-to-noise ratio (SNR), water suppression efficiency, shimming efficiency, signal intensity, and coefficient of variation (CoV, a measure of reproducibility). Studies were conducted on an in-house MRS brain phantom and three healthy volunteers (recording spectra from the anterior cingulate cortex). Using TE/TR = 40 ms/5000 ms and TE/TR = 6 ms/4500 ms for the sLASER and STEAM acquisitions, respectively (but 64 signal averages and voxel size of 8.0 mL with both pulse sequences), the spectral analyses results (using an in-house spectral analysis and quantification script deployed in MATLAB) showed that sLASER offered better SNR and shimming efficiency, while both pulse sequences had the same performance on water suppression efficiency (even though sLASER showed a tendency to do better than STEAM). STEAM was shown to be more reproducible in repeated scans in both phantom and human studies. In terms of signal intensities, significant variations (p < 0.05) were observed in the intensities of lactate and NAcetyl Aspartate (higher) and myo-inositol (lower) with sLASER compared to STEAM in phantom scans; in the human scans, NAA and taurine had higher signal intensities (p < 0.05) with sLASER than with STEAM pulse sequences. The results thus showed that, generally, at the optimum pulse sequence parameters, sLASER offers better MRS signal quality than STEAM, but may not be suitable for longitudinal studies compared to STEAM. Future studies may consider using a dedicated spectral analysis and quantification software package, increasing the sample size for the human studies, and proceeding to estimate absolute metabolite concentrations to validate the spectra quantification scheme implemented in this study. (Author)