Objective: To evaluate proof-of-concept performance of two artificial intelligence (AI) multivariable models predicting (1) motor disease progression and (2) treatment-related side effects over 12 months in Parkinson’s disease (PD), and to assess the feasibility and usability of the supporting digital health applications.

Background: PD is clinically heterogeneous, with variable motor progression and treatment response. Existing prognostic models, often based on cross-sectional data, insufficiently capture longitudinal dynamics and medication effects. Combining demographic, clinical, genetic and real-world digital biomarkers (dBMs) may enhance individualized prediction and decision-making. This study is part of AI-PROGNOSIS, a European research project (funded by the European Union under Grant Agreement No. 101080581) that aims to advance PD diagnosis and care through novel predictive AI-based models.

Method: AI-PMP is a 12-month, multicentre, prospective European study including 100 patients with PD (disease duration 5–10 years; Hoehn and Yahr 2–3). Baseline and 12-month visits include comprehensive clinical assessment and polygenic profiling. During follow-up, dBMs are collected via a smartwatch and mobile application.

The primary endpoint is the performance of the model in predicting motor progression at 12 months, defined as either an increase of >3 points in MDS-UPDRS part III (ON) or a >10% increase in levodopa equivalent daily dose. Secondary and exploratory endpoints include model performance in predicting motor complications, selected disease severity milestones (cognitive dysfunction, hallucinations, freezing, dysphagia), and treatment-related side effects. Performance will be assessed using discrimination, calibration and agreement metrics, and compared with clinician predictions. Feasibility and usability will be evaluated through adherence measures and System Usability Scale scores.

Results: The study is currently ongoing in France and Spain, with expansion to Germany and the UK planned throughout 2026. The study protocol and progress to date will be presented.

Conclusion: AI-PMP will prospectively evaluate proof-of-concept performance of AI-driven prediction of PD progression and treatment response using multimodal and real-world data, helping to pave the way for more personalized disease management.

« Back to 2026 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/ai-based-progression-and-medication-response-prediction-in-parkinsons-disease-the-ai-pmp-study-protocol/