Noriko Ishigami and Takahiko Tokuda contributed equally to this work.
Version of Record online: 1 JUN 2012
Copyright © 2012 Movement Disorder Society
Volume 27, Issue 7, pages 851–857, June 2012
How to Cite
Ishigami, N., Tokuda, T., Ikegawa, M., Komori, M., Kasai, T., Kondo, T., Matsuyama, Y., Nirasawa, T., Thiele, H., Tashiro, K. and Nakagawa, M. (2012), Cerebrospinal fluid proteomic patterns discriminate Parkinson's disease and multiple system atrophy. Mov. Disord., 27: 851–857. doi: 10.1002/mds.24994
Relevant conflicts of interest/financial disclosures: Nothing to report.
Full financial disclosures and author roles may be found in the online version of this article.
- Issue online: 20 JUN 2012
- Version of Record online: 1 JUN 2012
- Manuscript Accepted: 9 MAR 2012
- Manuscript Revised: 27 FEB 2012
- Manuscript Received: 20 JUN 2011
- Parkinson's disease;
- multiple system atrophy;
- cerebrospinal fluid;
The differential diagnosis of Parkinson's disease and multiple system atrophy can be challenging, especially in the early stages of the diseases. We developed a proteomic profiling strategy for parkinsonian diseases using mass spectrometry analysis for magnetic-bead-based enrichment of cerebrospinal fluid peptides/proteins and subsequent multivariate statistical analysis. Cerebrospinal fluid was obtained from 37 patients diagnosed with Parkinson's disease, 32 patients diagnosed with multiple system atrophy, and 26 patients diagnosed with other neurological diseases as controls. The samples were from the first cohort and the second cohort. Cerebrospinal fluid peptides/proteins were purified with C8 magnetic beads, and spectra were obtained by matrix-assisted laser desorption ionization time-of-flight mass spectrometry. Principal component analysis and support vector machine methods are used to reduce dimension of the data and select features to classify diseases. Cerebrospinal fluid proteomic profiles of Parkinson's disease, multiple system atrophy, and control were differentiated from each other by principal component analysis. By building a support vector machine classifier, 3 groups were classified effectively with good cross-validation accuracy. The model accuracy was well preserved for both cases, training by the first cohort and validated by the second cohort and vice versa. Receiver operating characteristics proved that the peak of m/z 6250 was the most important to differentiate multiple system atrophy from Parkinson's disease, especially in the early stages of the disease. A proteomic pattern classification method can increase the accuracy of clinical diagnosis of Parkinson's disease and multiple system atrophy, especially in the early stages. © 2012 Movement Disorder Society