Diagnostic Potential of Multimodal MRI Markers in Atypical Parkinsonian Disorders

Article type: Research Article

Authors: Seki, Morinobu^{a; b; *} | Seppi, Klaus^{a; b} | Mueller, Christoph^a | Potrusil, Thomas^{a; b} | Goebel, Georg^c | Reiter, Eva^a | Nocker, Michael^a | Kremser, Christian^{b; d} | Wildauer, Matthias^{b; d} | Schocke, Michael^d | Gizewski, Elke R.^{b; d} | Wenning, Gregor K.^a | Poewe, Werner^a | Scherfler, Christoph^{a; b}

Affiliations: [a] Department of Neurology, Medical University of Innsbruck, Innsbruck, Austria | [b] Neuroimaging Research Core Facility, Medical University of Innsbruck, Innsbruck, Austria | [c] Department of Medical Statistics, Informatics and Health Economics, Medical University of Innsbruck, Innsbruck, Austria | [d] Department of Neuroradiology, Medical University of Innsbruck, Innsbruck, Austria

Correspondence: [*] Correspondence to: Morinobu Seki, MD, PhD, Department of Neurology, Medical University of Innsbruck, Anichstrasse 35, A-6020 Innsbruck, Austria. Tel.: +43 512 504 23850; Fax: +43 512 504 23852; E-mail: mseki-neuro@umin.ac.jp.

Abstract: Background:The diagnostic potential of multimodal MRI approaches to discriminate among progressive supranuclear palsy (PSP), Parkinson variant of multiple system atrophy (MSA-P) and Parkinson’s disease (PD) has not been well investigated. Objective:To identify disease-specific neurodegenerative patterns and evaluate the diagnostic accuracy of dedicated MRI, iron concentration (R2*), microstructural integrity (mean diffusivity; MD and fractional anisotropy; FA) as well as volumes were analyzed in patients with PSP, MSA-P and PD. Methods:3T MRI of 18 PSP and 16 MSA-P patients were compared with 16 PD patients matched for age and disease duration as well as 21 healthy controls. Statistical parametric mapping (SPM) was applied to objectively identify focal MRI changes throughout the whole-brain. Following dimensionality reduction of significant and multiple comparison-corrected SPM clusters through principal component analysis (PCA), stepwise receiver-operating characteristic curve analysis (ROC) was applied to determine the diagnostic potential of multimodal MRI parameters. Results:PCA revealed two components involving multiple regions identified from SPM analysis. The first component was primarily composed of the mean MD value of the thalamus and the mean MD and FA values of the dentatorubrothalamic tract and the corpus callosum. The second component mainly consisted of mean MD and FA values of the middle cerebellar peduncle. ROC analysis showed 92% of PSP patients were differentiated correctly from MSA-P and PD and 80% of MSA-P patients could be distinguished from PD. Conclusion:Multimodal MRI improved the detection of disease-specific neurodegenerative patterns in PSP and MSA-P and highlights its potential to improve the diagnostic accuracy of atypical parkinsonian disorders.

Keywords: Multimodal MRI, principal component analysis, R2*, diffusion tensor imaging, progressive supranuclear palsy, multiple system atrophy

DOI: 10.3233/JPD-181568

Journal: Journal of Parkinson's Disease, vol. 9, no. 4, pp. 681-691, 2019