Targeted Screening for Alzheimer's Disease Clinical Trials Using Data-Driven Disease Progression Models

Area of Science:

• Neuroscience
• Clinical Trials
• Computational Biology

Background:

• Alzheimer's disease (AD) progression is highly heterogeneous, hindering the evaluation of disease-modifying therapies.
• Treatment effects in subgroups of responders can be masked by non-responders in clinical trials.
• Effective methods for screening potential responders are lacking.

Purpose of the Study:

• To pilot a computational screening tool for improved patient stratification in Alzheimer's disease clinical trials.
• To enhance the sensitivity to treatment effects by identifying and excluding non-responders.
• To potentially reduce clinical trial size, duration, and cost.

Main Methods:

• Utilized data-driven disease progression modeling to develop a computational screening tool.
• Retrospectively analyzed a completed double-blind clinical trial of donepezil (NCT00000173) in mild cognitive impairment.
• Identified a data-driven subgroup based on cognitive impairment severity.

Main Results:

• The computational tool demonstrated potential for improved patient stratification.
• A specific subgroup with more severe cognitive impairment exhibited a clearer treatment response to donepezil.
• This subgroup's response was more pronounced than that of the overall trial cohort.

Conclusions:

• Computational screening tools can enhance the identification of responders in Alzheimer's disease trials.
• Improved stratification may increase the likelihood of detecting therapeutic effects.
• This approach holds promise for optimizing the design and efficiency of future clinical trials.