Using short-term endpoints to improve interim decision making and trial duration in two-stage phase II trials with...

Area of Science:

• Clinical Trials
• Biostatistics
• Oncology Research

Background:

• Phase II oncology trials commonly use two-stage designs with binary endpoints.
• Simon's design is popular but can lead to long trial interruptions due to follow-up times.
• Using short-term endpoints for interim decisions risks misleading results if assumptions are misspecified.

Purpose of the Study:

• To propose novel methods for interim decisions in two-stage oncology trials.
• To improve statistical power and avoid premature trial closure due to misspecified short-term endpoints.
• To leverage all available patient outcome data for more accurate trial guidance.

Main Methods:

• Developed two approaches: conditional power and Bayesian posterior predictive probability.
• Utilized both short-term and long-term endpoint assessments for interim decision-making.
• Conducted extensive simulations to evaluate method performance and power.

Main Results:

• Both proposed methods demonstrated similar performance when calibrated.
• These approaches significantly improve statistical power compared to existing methods.
• The methods are particularly beneficial in trials with slow patient recruitment.

Conclusions:

• The proposed methods offer a robust alternative for interim analyses in phase II oncology trials.
• Accurate utilization of all patient data enhances trial efficiency and reliability.
• Publicly available software facilitates the implementation of these advanced statistical techniques.