Reproducibility of hippocampal single-Voxel proton MR spectroscopy and chemical shift imaging.

Abstract

We investigated between- and within-acquisition reproducibility of hippocampal metabolite ratios obtained using automated proton MR spectroscopy. We examined 30 healthy adults with a 1.5-T scanner four times on 3 days using single-voxel spectroscopy over the left hippocampus, chemical shift imaging over the left hippocampus, and chemical shift imaging over the bilateral hippocampi. Metabolite ratios were derived from the integral values of three major peaks: N:-acetylaspartate (NAA), choline-containing compounds (Cho), and creatine plus phosphocreatine (Cr). The random-effects model of one-way analysis of variance was used to evaluate the reproducibility in terms of coefficient of variation; the mixed-effects model was used to compare the results of different hippocampal regions and spectroscopic techniques. Most coefficients of variation for the NAA/(Cho+Cr) ratio were less than 20%. All the coefficients of variation for the posterior hippocampus (15-25%) were less than those for the anterior hippocampus (20-44%). The posterior hippocampal NAA/(Cho+Cr) ratio of unilateral chemical shift imaging had the lowest coefficient of variation (<16%). Single-voxel spectroscopy and unilateral chemical shift imaging had similar coefficients of variation for the anterior hippocampal NAA/(Cho+Cr) ratios (17-20%). There was a significant difference in metabolite ratios measured in different hippocampal regions (p<0.01) and in those acquired with different spectroscopic techniques (p<0.001). The NAA/(Cho+Cr) ratio is the most reproducible parameter for hippocampal MR spectroscopy on a 1.5-T scanner. Regional variation and technical differences in metabolite ratios must be considered when interpreting proton spectra of the hippocampus.