Purpose:To investigate the impact of cerebrospinal fluid (CSF) contamination on metabolite evaluation in the superior cerebellar vermis with single-voxel 1 H-MRS in normal subjects and patients with degenerative ataxias.
Materials and Methods:Twenty-nine healthy volunteers and 38 patients with degenerative ataxias and cerebellar atrophy were examined on a 1.5 Tesla scanner. Proton spectra of a volume of interest placed in the superior vermis were acquired using a four TE PRESS technique. We calculated N-acetyl aspartate (NAA)/creatine (Cr), choline (Cho)/Cr, and NAA/Cho ratios, T 2 relaxation times and concentrations of the same metabolites using the external phantom method. Finally, concentrations were corrected taking into account the proportion of nervous tissue and CSF, that was determined as Volume Fraction (VF).
Results:In healthy subjects, a significant difference was observed between metabolite concentrations with and without correction for VF. As compared to controls, patients with ataxias showed significantly reduced NAA/Cr and NAA concentrations, while only corrected Cr concentration was significantly increased. The latter showed an inverse correlation with VF.
Conclusion:CSF contamination has a not negligible effect on the estimation of brain metabolites. The increase of Cr concentration in patients with cerebellar atrophy presumably reflects the substitutive gliosis which takes place along with loss of neurons. EVALUATION IN VIVO of the amount of the brain metabolites with proton MR spectroscopy ( 1 H-MRS) is an important clinical tool in several diseases (1). This can be accomplished using a semiquantitative approach, in which metabolite ratios are computed by assuming that a reference metabolite, usually creatine (Cr), is constant, or using a quantitative approach in which the concentration of each brain metabolite is determined with several different techniques. These include external phantoms, reference to water signal, and compartmental analysis (2).A controversial issue in clinical 1 H-MRS of the brain is the evaluation of spectra obtained from volume of interest (VOI) containing cerebrospinal fluid (CSF). In particular, it is unclear the relevance of correction for the CSF contained in the VOI. In a prior study corrections from CSF inclusion were found to decrease the coefficient of variation of spectral data acquired in the hippocampus (3).The peripheral cerebellum is a region particularly suited to investigate the impact of CSF contamination on metabolite assessment. In fact, because of the thinness of the cerebellar folia, spectra of the peripheral cerebellum contain variable amounts of nervous tissue and CSF also in normal subjects. While spectra from the cerebellar hemispheres can be of suboptimal quality due to difficult shimming caused by the nearby skull and blood flow in the venous dural sinuses, good quality spectra are consistently obtained when the VOI is placed in the superior vermis (4).To investigate the impact of CSF contamination on the metabolite evaluation we examined the c...