How to improve statistical power in a trial with SCA2 patients using natural history data

Area of Science:

• Clinical Trials Methodology
• Neurodegenerative Diseases
• Statistical Modeling

Background:

• Randomized controlled trials (RCTs) are standard but face ethical and logistical challenges in rare neurodegenerative diseases due to small patient populations.
• Limited sample sizes in rare disease trials can reduce statistical power, making it difficult to detect treatment effects.

Purpose of the Study:

• To enhance statistical power in clinical trials for rare neurodegenerative diseases by applying advanced statistical methodologies.
• To evaluate novel methods, including prediction-powered inference for clinical trials (PPCT), prognostic covariate adjustment, and Hybrid Augmented Inverse Probability Weighting (H-AIPW), using the ATRIL study data for spinocerebellar ataxia type 2 (SCA2).

Main Methods:

• Trained a progression model using Disease Course Mapping on natural history data to predict disease progression.
• Applied PPCT, prognostic covariate adjustment, and H-AIPW to the ATRIL trial data, comparing their performance against classical variance.
• Utilized external patient data from US and European cohorts (EUROSCA, SPATAX, CRC-SCA) to train the progression model.

Main Results:

• Disease Course Mapping accurately forecasted prognostic scores, with correlations around 0.15 for predicted one-year SARA score progression.
• PPCT, prognostic covariate adjustment, and H-AIPW methods demonstrated lower variance in treatment effect estimation (0.342-0.348) compared to classical methods (0.407).
• PPCT could reduce the required sample size by 14.5% (6 patients), achieving the same variance with 39 patients instead of 45.

Conclusions:

• Advanced statistical methods like PPCT, prognostic covariate adjustment, and H-AIPW can significantly improve statistical power or reduce sample size requirements in rare neurodegenerative disease trials.
• Leveraging external natural history data through these methods enhances the efficiency of clinical trials, crucial for diseases with limited patient populations.
• While the ATRIL trial's conclusion on riluzole efficacy remained unchanged, these methods offer a pathway to more efficient trial design for future studies.