Parkinson's Disease Levodopa-Induced Dyskinesia Classification via Cerebellar Radiomics

Data from the Parkinson's Progression Markers Initiative (PPMI) database, comprising 211 patients (69 with Levodopa-Induced Dyskinesia and 142 without). Author Yi Chen published this feasibility study on figshare in April 2026, which used 3D T1-weighted MRI images to extract radiomic features from cerebellar gray and white matter for diagnostic modeling.

Use Cases

• Building a binary classifier for Levodopa-Induced Dyskinesia (LID) based on cerebellar structural radiomics.
• Investigating feature importance for LID diagnosis using SHAP model visualization methods.
• Benchmarking radiomics model performance (e.g., XGBoost) on neuroimaging data from a specific clinical cohort.

Strengths

• Includes data from 211 patients sourced from the established PPMI database.
• Model reported an AUC of 0.962 on the training set and 0.849 on the testing set.
• Focuses on a specific, clinically relevant problem (LID diagnosis) with a novel cerebellar radiomics approach.

Limitations

• Row count and column-level documentation are unknown, limiting suitability assessment.
• The dataset is small (287.4 KB), suggesting it contains summary or supplementary results rather than raw image data.
• Data freshness should be verified; the last update date is in the future (2026-04-22).

Provenance

Source

Parkinson's Progression Markers Initiative (PPMI) database

Collection Method

Retrospective collection of 3D T1WI MRI data, segmented using Fastsurfer, with radiomic features extracted via Python.

Time Range

null

Freshness

Last updated 2026-04-22 05:46:25; freshness should be verified.

Geography

null