Radiomics-based Modelling Unveils Cerebellar Involvement in Parkinson's Disease.

Area of Science:

• Neuroimaging
• Radiomics
• Biomarker Discovery

Background:

• Emerging evidence implicates the cerebellum in Parkinson's disease (PD) pathophysiology.
• Cerebellar dysfunction is increasingly recognized in neurodegenerative disorders.

Purpose of the Study:

• To investigate the utility of cerebellar magnetic resonance imaging (MRI) radiomics for differentiating PD patients from healthy controls (HC).
• To identify key radiomic features associated with PD in the cerebellum.

Main Methods:

• Retrospective analysis of 3D-T1 MRI data from 374 participants (PPMI and in-house cohort).
• Extraction of 883 radiomic features from cerebellar gray and white matter.
• Development and evaluation of machine learning models (gray matter, white matter, combined) for PD classification.

Main Results:

• The combined gray and white matter radiomics model achieved the highest performance, with an area under the ROC curve (AUC) of 0.936 in the training set and 0.881 in the testing set.
• Cerebellar gray matter model: AUC of 0.931 (training) and 0.874 (testing).
• Cerebellar white matter model: AUC of 0.846 (training) and 0.868 (testing).
• Key features included Gray-level Dependence Matrix, Gray-level Co-occurrence Matrix, First-Order (gray matter), and Gray-level Size Zone Matrix (white matter).

Conclusions:

• Cerebellar MRI radiomics effectively differentiates PD patients from HC.
• These findings support the cerebellum's significant role in PD.
• Cerebellar radiomics shows promise as a novel imaging biomarker for Parkinson's disease.