Optimal predictive probability designs for randomized biomarker-guided oncology trials

TL;DR

This paper introduces three innovative randomized designs for early phase biomarker-guided oncology trials, utilizing optimal predictive probability methods to improve efficiency and decision-making in subpopulation assessments.

Contribution

It proposes novel randomized trial designs that incorporate optimal predictive probability monitoring for multiple biomarker subgroups, enhancing early phase trial efficiency.

Findings

Designs effectively monitor multiple subgroups for futility

Simulation shows improved decision-making efficiency

Potential to better inform phase III trial decisions

Abstract

Efforts to develop biomarker-targeted anti-cancer therapies have progressed rapidly in recent years. Six antibodies acting on programmed death ligand 1 or programmed death 1 pathways were approved in 75 cancer indications between 2015 and 2021. With efforts to expedite regulatory reviews of promising therapies, several targeted cancer therapies have been granted accelerated approval on the basis of single-arm phase II trials. And yet, in the absence of randomization, patient outcomes may not have been studied under standard of care chemotherapies for emerging biomarker subpopulations prior to the submission of an accelerated approval application. Historical control rates used in single arm studies often arise as population averages, lacking specificity to the targeted subgroup. For example, a recent phase III trial of atezolizumab in patients with metastatic urothelial carcinoma found a 21.6% response rate to standard of care chemotherapy in the biomarker subgbroup of interest, much higher than the historical control rate of 10% that had been used to declare success in the preceding phase II trial. Innovations in design methodology are needed to enable efficient implementation of randomized trials for agents that target biomarker subpopulations. This article proposes three randomized designs for early phase biomarker-guided oncology clinical trials. Each design utilizes the optimal efficiency predictive probability method to monitor multiple biomarker subpopulations for futility. A simulation study motivated by the results reported in the atezolizumab trial is used to evaluate the operating characteristics of the various designs. Our findings suggest that efficient statistical design can be conducted with randomization and futility stopping to effectively acquire more evidence pertaining to comparative efficacy before deciding to conduct a phase III confirmatory trial.