Forecasting individual progression trajectories in Alzheimer's disease

Area of Science:

• Neuroscience
• Biostatistics
• Medical Imaging

Background:

• Accurate prediction of Alzheimer's disease (AD) progression is vital for evaluating secondary prevention strategies.
• Forecasting AD trajectory is challenging due to variable rates of decline in cognitive and imaging biomarkers across individuals.

Purpose of the Study:

• To evaluate AD Course Map, a statistical model designed to predict the progression of neuropsychological assessments and imaging biomarkers in early-stage Alzheimer's disease.
• To assess the model's accuracy in identifying patients likely to progress in clinical trials.

Main Methods:

• Utilized AD Course Map, a statistical model, to predict disease progression using patient medical and radiological data.
• Tested the model on a large dataset (>96,000 cases, >4,600 patients) across four continents.
• Assessed the model's accuracy in selecting participants with progressing clinical endpoints for hypothetical trials.

Main Results:

• Enriching trial populations with predicted progressors reduced required sample sizes by 38%–50%.
• The model demonstrated no bias related to sex or geographic location and was robust to missing data.
• Optimal performance was observed at the earliest disease stages, including asymptomatic individuals at risk of AD.

Conclusions:

• AD Course Map is a robust and accurate tool for predicting Alzheimer's disease progression.
• The model significantly enhances the efficiency of clinical trials for AD prevention by reducing sample size requirements.
• Its suitability for early-stage disease makes it highly valuable for prevention research.