Drug Development.

Area of Science:

• Neuroscience
• Clinical Trials
• Biostatistics

Background:

• Alzheimer's disease (AD) impacts cognitive and functional domains, necessitating sensitive clinical trial assessments.
• Current composite scores can be noisy, limiting statistical power and increasing patient burden.
• Optimizing item selection and weighting in assessments is crucial for improving trial efficiency, particularly in early-phase studies.

Purpose of the Study:

• To introduce and apply a novel method for optimizing composite scores in Alzheimer's disease (AD) clinical trials.
• To enhance statistical power and reduce patient burden by refining endpoint selection.
• To leverage financial portfolio optimization techniques for clinical trial design.

Main Methods:

• Utilized the 'efficient frontier' method from finance to optimize composite scores.
• Applied the method to five common early AD assessments: ADAS-Cog13, MMSE, CDR, FAQ, and ADCS-ADL.
• Integrated digital twin predictions of item-level progression for cohort-specific optimization.

Main Results:

• Demonstrated significant improvements in statistical power for early AD populations, e.g., an 18% gain for ADAS-Cog13.
• Identified novel composite scores by combining assessments, highlighting cognitive domains, functional impairment, or both.
• The efficient frontier approach yielded globally optimal composite scores maximizing control arm progression for a given variance.

Conclusions:

• The efficient frontier method offers a novel, powerful tool for clinical trial endpoint selection, especially in early-phase AD studies.
• This approach yields interpretable results and can be applied to individual or multiple assessments.
• Optimized composite scores improve sensitivity to treatment effects, reduce patient burden, and aid trial decision-making.