A predictive model for longitudinal cognitive subtypes in Parkinson's disease.

Area of Science:

• Neuroscience
• Neurology
• Clinical Research

Background:

• Longitudinal cognitive changes in Parkinson's disease (PD) are highly variable.
• Predicting cognitive trajectories in early PD is essential for patient counseling and clinical trials.

Purpose of the Study:

• To identify distinct longitudinal cognitive subtypes in early Parkinson's disease.
• To develop a predictive model for cognitive deterioration in early PD patients.

Main Methods:

• Utilized data from 337 early PD patients in the Parkinson's Progression Markers Initiative (PPMI) database with a 6-year follow-up.
• Employed Latent Class Mixed Models (LCMM) to identify cognitive trajectories and constructed a nomogram using baseline clinical variables.
• Assessed cognitive function using the Montreal Cognitive Assessment (MoCA).

Main Results:

• Identified two cognitive subtypes: cognitive stable (81.9%) and cognitive deteriorating (18.1%).
• The deteriorating subtype showed poorer baseline cognition, older age, more severe motor/non-motor symptoms, higher serum NFL, and lower striatal DAT uptake.
• A nomogram predictive model using six baseline variables achieved an AUC of 0.92, demonstrating high predictive accuracy.

Conclusions:

• Two longitudinal cognitive subtypes were identified in early PD patients over 6 years.
• Eighteen percent of early PD patients fall into a cognitive deterioration subtype.
• A validated predictive model using six baseline variables can estimate the risk of cognitive deterioration in PD.