Diagnostic value of susceptibility-based imaging and apparent diffusion coefficient for differentiating multiple system atrophy-Parkinsonian type from Parkinson's disease
Related Experiment Video
Updated: Oct 31, 2026
Diffusion Tensor Magnetic Resonance Imaging in the Analysis of Neurodegenerative Diseases
Published on: July 28, 2013
Byeongpil Moon1, So Yeon Won2, Jonghyo Youn1
1Department of Electrical and Computer Engineering, Seoul National University, Seoul, Republic of Korea.
Objective:
To characterize the diagnostic performance of susceptibility-weighted imaging (SWI), χ-separation-derived paramagnetic susceptibility (χpara), and apparent diffusion coefficient (ADC) for differentiating multiple system atrophy-Parkinsonian type (MSA-P) from Parkinson's disease (PD).
Materials And Methods:
This retrospective case-control study included 27 drug-naïve patients with MSA-P and 27 age- and sex-matched drug-naïve patients with PD. Clinical diagnosis by board-certified neurologists served as the reference standard. χpara maps were generated from multi-echo gradient-recalled echo images using χ-separation, and ADC maps were derived from diffusion-weighted imaging. Two blinded radiologists qualitatively assessed SWI and ADC images using prespecified putaminal criteria. Quantitative analyses included voxel-wise group comparison and receiver operating characteristic (ROC) analysis using mean region-of-interest (ROI) values. Candidate logistic-regression models were assessed using leave-one-out cross-validation.
Results:
Qualitative assessment yielded sensitivity (85.2%), specificity (100%) and accuracy (92.6%) for SWI; the corresponding ADC assessments were 70.4%, 77.8% and 74.1%, respectively. Voxel-wise χpara analysis revealed a rim-shaped susceptibility increase in the posterior putamen, whereas ADC differences were localized to cerebellar white matter, including the middle cerebellar peduncle (MCP). In ROI-based classification, putaminal χpara yielded a sensitivity of 92.6% and ADC of the MCP showed a specificity of 96.3%. An exploratory model combining putaminal χpara and ADC in the MCP yielded a leave-one-out cross-validation AUC of 0.927.
Conclusion:
By evaluating susceptibility-based imaging and ADC in the same cohort, this study characterized their respective regional imaging features and diagnostic performance for differentiating MSA-P from PD. χpara revealed an observed rim-shaped susceptibility pattern in the posterior putamen, supporting further evaluation of this pattern as a candidate imaging finding in MSA-P.
