Imaging-Enhanced Prediction of Upper Limb Motor Function After Stroke Using Automated Segmentation and Cluster Analysis.

Adding imaging-derived variables to clinical variables improved prediction of upper extremity motor impairment (FMA-UE) after acute ischemic stroke.

• The study retrospectively analyzed 242 patients with acute ischemic stroke who had brain MRI and motor function assessment.
• Predictive models for FMA-UE scores were built using hierarchical regression.
• The final model incorporating imaging variables reached an adjusted R² of 0.595.

Motor tract lesion volume emerged as a key predictor of FMA-UE scores.

• Imaging variables were extracted with automated software and included infarct volume, motor tract involvement, and subtype classification.
• Motor tract lesion volume was identified as a key predictor in the hierarchical regression models.

K-means cluster analysis of standardized baseline data identified three distinct patient subgroups.

• Cluster 1 consisted of patients with mild impairments and younger age.
• Cluster 2 consisted of older patients with multiple comorbidities.
• Cluster 3 consisted of patients with severe neurological deficits and large infarcts.

Functional outcomes differed significantly across the clusters, with Cluster 3 showing the greatest disability.

• In Cluster 3, 76% of patients had a follow-up modified Rankin Scale score of 4-6.
• The difference across clusters was significant (P < 0.001).

The authors conclude that explainable imaging features from early MRI and cluster-based stratification support precision rehabilitation.

• Cluster-based stratification identified clinically meaningful subgroups.
• The authors suggest this supports personalized care planning and targeted secondary prevention.
• The study notes that clinical integration of neuroimaging prognostic information remains limited.