Comparison of two programs for quantification of 1 H MR spectra of rat brain using a vascular dementia model

Abstract

The quantification of in vivo 1 H magnetic resonance (MR) spectra measured from the rat brains provides important information about the brain metabolite concentrations and can help to understand the role of the metabolites under normal and pathological conditions. The purpose of this study was to compare the most frequently used algorithms for quantification of 1 H spectra: LCModel (Linear Combination of Model spectra) and QUEST (QUantitation based on QUantum ESTimation) from jMRUI software (Java based Magnetic Resonance User Interface). The comparison was done on a rat model of vascular dementia (VD). The MR spectra were measured on 4.7T spectrometer with ultra-short echo time by sequence SPECIAL. For these types of spectra the contribution from the macromolecules and lipids is large. Our analysis revealed that all values determined by QUEST, except for one value, were lower in comparison to values obtained by LCModel. The minimal differences were found in N-acetylaspartate/(phospho) creatine (−0.3 %) and maximal in inositol in both control and VD rats. This underestimation of a metabolite concentration in QUEST may be caused by an overestimation of baseline.Although our study found the different values of metabolite concentrations by these two methods, the quantified metabolite changes in pathological brain were comparable in both analyses.