Related Experiment Videos
Selection designs for pilot studies based on survival
P Y Liu1, S Dahlberg, J Crowley
1Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, Washington 98104.
Abstract:
In cancer clinical trials new regimens are typically tested for antitumor activities in patients with advanced disease. The promising ones are then compared to the standard treatment in a randomized study, sometimes performed on patients with earlier-stage disease. When there are multiple promising regimens, it may not be possible to compare all of them to the control group because of the prohibitive sample size and study length requirements. We propose a design that uses the Cox regression model to select a best treatment based on survival before the randomized comparison. Sample sizes for an asymptotically correct selection probability of .90 are presented for Weibull survival distributions with parameters in a range we consider to be of practical interest. Simulations verify that the asymptotic approximations to the correct selection probabilities are quite satisfactory. Simulations also indicate that the procedure is reasonably robust to the proportional hazards assumption. In contrast to the two-stage screening design recommended by Schaid, Wieand, and Therneau (1990, Biometrika 77, 507-513), our design has the advantage of fitting naturally to a progression of cancer trials where the selection and comparison phases are carried out on different populations of patients. When the population of interest stays the same, our design can be more conservative on the average but offers the opportunity to base the comparative trial on the experience gained during the selection phase.
Insights
This study introduces a new Cox regression model design for cancer clinical trials to efficiently select the best treatment regimen before large-scale randomized comparisons. This approach optimizes resource allocation and accelerates the identification of effective cancer therapies.
Area of Science:
- Clinical Trials Methodology
- Biostatistics
- Oncology Research
Background:
- Cancer clinical trials typically test new regimens in advanced disease patients, then compare promising ones to standard treatments.
- Comparing multiple promising regimens against a control can be infeasible due to sample size and study duration constraints.
Purpose of the Study:
- To propose a novel clinical trial design utilizing the Cox regression model for selecting the optimal treatment regimen based on survival data prior to randomized comparison.
- To provide sample size calculations for achieving a .90 selection probability with Weibull survival distributions.
Main Methods:
- Application of the Cox regression model to survival data for treatment selection.
- Development of sample size guidelines for Weibull survival distributions.
- Monte Carlo simulations to verify asymptotic approximations and assess robustness.
Main Results:
- Asymptotic approximations for correct selection probabilities are satisfactory.
- The proposed design is robust to violations of the proportional hazards assumption.
- The design integrates seamlessly with sequential cancer trial phases, potentially using different patient populations.
Conclusions:
- The proposed Cox regression-based design offers an efficient alternative for selecting the best cancer treatment regimen in clinical trials.
- This method addresses limitations of traditional multi-regimen comparisons by optimizing sample size and study length.
- The design provides a flexible framework for advancing cancer treatment research.