Brain Atrophy Does Not Predict Clinical Progression in Progressive Supranuclear Palsy

Area of Science:

• Neuroscience
• Clinical Neurology
• Medical Imaging

Background:

• Progressive supranuclear palsy (PSP) clinical trials typically use clinical progression rate as the primary endpoint.
• Predicting individual disease trajectory is crucial for effective clinical trial design and patient management.

Purpose of the Study:

• To investigate if baseline clinical severity and regional brain atrophy can predict clinical progression in PSP-Richardson's syndrome (PSP-RS).
• To assess the utility of machine learning models in predicting individual patient progression rates.

Main Methods:

• A longitudinal multicohort study included 309 PSP-RS patients from clinical trial placebo arms and the Describe PSP cohort.
• Associations between baseline clinical data, volumetric MRI, and 1-year PSP rating scale (PSPRS) change were analyzed.
• Machine learning models were employed to predict individual clinical trajectories.

Main Results:

• PSP-RS patients exhibited a mean PSPRS score increase of 10.3 points per year.
• Frontal lobe volume demonstrated the strongest association with subsequent clinical progression (β: -0.34, P < 0.001).
• Machine learning models failed to accurately predict individual progression rates (R² < 0.15).

Conclusions:

• Neither baseline clinical severity nor brain atrophy reliably predicted individual clinical progression in PSP-RS.
• MRI-based stratification of patients is likely unnecessary for future PSP clinical trials.
• Findings may inform trial design and patient selection strategies.