Determination of biomarkers for mild cognitive impairment in parkinson's disease using magnetic resonance spectroscopic imaging

Parkinson's disease (PD) patients could be categorized as PD with cognitively normal (PD-CN), PD with mild cognitive impairment (PD-MCI), and PD with dementia (PDD). There is a need for finding noninvasive biomarkers for the early diagnosis of PD-MCI. Proton magnetic resonance spectroscopic imaging (1H-MRSI) is a non-invasive MR technique that provides spectroscopic information about metabolic activity of the brain. 19 patients with PD-MCI and 21 patients with PD-CN were included in this study and neuropsychological tests were performed. Multi-voxel 1H-MRSI data were acquired in all patients. An MRSI data analysis tool was developed to create 1H MR spectroscopic peak parameter maps out of raw MRSI data and overlay them onto reference T2-weighted MR images. FMRIB Software Library (FSL) tool was used to register metabolite maps overlaid onto T2-weighted MR images to an MNI152 brain atlas. A Mann-Whitney rank-sum test was applied to compare the differences of metabolic parameters and neuropsychological test scores between PD-MCI and PD-CN. A Friedman test was used to analyze the MR spectroscopic metabolite ratio variations in different brain regions of PD-MCI and PD-CN. Spearman rank correlation coefficient was used to find correlations of neuropsychological test scores and MRS metabolite ratios. There were no significant differences in MRS metabolite ratios in different brain regions of PD-MCI and PD-CN after accounting for multiple comparisons. However, frontal lobe and cerebral white matter showed trends for metabolic differences. Neuropsychological test scores were correlated with several spectroscopic parameters. The results of this study might enable a definition of a biomarker for PD-MCI diagnosis in the future, when combined with possible other MR based biomarkers.