Analysis of phase II methodologies for single-arm clinical trials with multiple endpoints in rare cancers: An example...

Area of Science:

• Clinical trial design
• Biostatistics
• Rare disease research

Background:

• Conducting clinical trials for rare diseases or rare subpopulations presents significant challenges in patient identification, recruitment, and timely treatment.
• Treatments for rare diseases are often repurposed from other areas after initial Phase I testing.
• Frequent interim analyses are crucial in Phase II trials to promptly stop for futility or toxicity, and for efficacy to accelerate progress.

Purpose of the Study:

• To review and compare frequentist and Bayesian statistical methods for adaptive trial designs.
• To evaluate these methods in the context of incorporating two binary endpoints and frequent interim analyses.
• To utilize the Eurosarc Trial of Linsitinib in advanced Ewing Sarcoma (LINES) as a case study for comparison.

Main Methods:

• Review of frequentist and Bayesian statistical approaches for clinical trial design.
• Adaptation of methods to handle dual binary endpoints and multiple interim analyses.
• Application and comparison of methods using data from the LINES trial.

Main Results:

• Both frequentist and Bayesian methods can be adapted for frequent interim analyses with binary endpoints.
• In a single trial with non-informative priors, the Bayesian approach offers no significant advantage over frequentist methods.
• Bayesian designs demonstrate potential for improved efficiency by incorporating prior information from previous studies.

Conclusions:

• Frequentist and Bayesian methods offer viable strategies for adaptive rare disease clinical trials.
• Bayesian approaches provide greater design flexibility and can enhance efficiency when prior data is available.
• Adaptive designs are essential for ethical and efficient rare disease research, enabling prompt decision-making based on interim results.