Search

Traditionally, dose-finding process for oncology compound is carried out in phase I dose escalation study and is driven by safety in order to find maximum tolerated dose (MTD). However, with the recent paradigm shift from cytotoxic drugs to new generation of targeted therapies and immuno-oncology therapies, it may be difficult or unnecessary to identify the MTD because of the possible non-monotonic dose–response curves, and efficacy data should be incorporated into the dose-finding process. In this article, we have proposed efficacy-driven dose-finding designs with a safety-driven warm-up phase. Both local investigation and adaptive randomization using the framework of double-sided isotonic regression are investigated. Simulation studies are used to compare the proposed design to the original double-sided isotonic design. The results show that a safety-driven warm-up phase at the beginning can significantly improve the performance of double-sided isotonic regression, and both local investigation and adaptive randomization have good operating characteristics for finding the best dose/dose range under different tested scenarios.

The Multiple Comparison Procedure and Modelling (MCPMod) approach has been shown to be a powerful statistical tool that can significantly improve the design and analysis of dose finding studies under model uncertainty. Due to its frequentist nature, however, it is difficult to incorporate into MCPMod information from historical trials on the same drug. Recently, a Bayesian version of MCPMod has been introduced by Fleischer et al. (2022) to resolve this issue, which is particularly tailored to the situation where there is information about the placebo dose group from historical trials. In practice, information may also be available on active dose groups from early phase trials, e.g., a dose escalation trial and a preceding small Proof of Concept trial with only a placebo and a high dose. To address this issue, we introduce a Bayesian hierarchical framework capable of incorporating historical information about both placebo and active dose groups with the flexibility of modelling both prognostic and predictive between-trial heterogeneity. Our method is particularly useful in the situation where the effect sizes of two trials are different. Our goal is to reduce the necessary sample size in the dose finding trial while maintaining its target power.

Multiple comparison procedures and modeling (MCPMod) has established itself as a method for dose-finding under model uncertainty. A downside of MCPMod is that due to its frequentist nature in particular with respect to the multiple comparison part it is tough to incorporate historical information in a systematic fashion. A typical situation where such historical information is available is existing data for the placebo group from previous trials. There are multiple Bayesian concepts for integrating historical data in a systematic and even dynamic fashion like the meta-analytic prior approach. In this article, we define Bayesian MCPMod (BMCPMod) that is build upon these two aspects. BMCPMod is able to mimic the results of the classical MCPMod for non-informative priors. At the same time, it allows for the inclusion of historical data in a systematic fashion. After the definition of BMCPMod related characteristics for a Bayesian approach similar to the MCP-testing part are derived. The BMCPMod is compared to classical MCPMod/non-informative priors via simulations. Aspects of mixture priors, optimal contrast vectors, and impact of allocation ratios are discussed and an example for designing a BMPCMod trial is given.

After exploratory drug development, companies face the decision whether to initiate confirmatory trials based on limited efficacy information. This proof-of-concept decision is typically performed after a Phase II trial studying a novel treatment versus either placebo or an active comparator. The article aims to optimize the design of such a proof-of-concept trial with respect to decision making. We incorporate historical information and develop pre-specified decision criteria accounting for the uncertainty of the observed treatment effect. We optimize these criteria based on sensitivity and specificity, given the historical information. Specifically, time-to-event data are considered in a randomized 2-arm trial with additional prior information on the control treatment. The proof-of-concept criterion uses treatment effect size, rather than significance. Criteria are defined on the posterior distribution of the hazard ratio given the Phase II data and the historical control information. Event times are exponentially modeled within groups, allowing for group-specific conjugate prior-to-posterior calculation. While a non-informative prior is placed on the investigational treatment, the control prior is constructed via the meta-analytic-predictive approach. The design parameters including sample size and allocation ratio are then optimized, maximizing the probability of taking the right decision. The approach is illustrated with an example in lung cancer.

A drug is defined as highly variable if its intra-individual coefficient of variation (CV) is greater than or equal to 30%. In such a case, bioequivalence may be assessed by means of methods that take the (high) variability into account. The Scaled Average Bioequivalence (SABE) approach is such a procedure and represents the recommendations of FDA. The aim of this investigation is to compare the performance characteristics of classical group sequential designs (GSD) and fixed design settings for three-period crossover bioequivalence studies with highly variable drugs, where the SABE procedure is utilized. Monte Carlo simulations were performed to assess type I error rate, power, and average sample size for GSDs with Pocock’s and O’Brien-Fleming’s stopping rules and various timings of the interim analysis and for fixed design settings. Based on our investigated scenarios, the GSDs show comparable properties with regard to power and type I error rate as compared to the corresponding fixed designs. However, due to an advantage in average sample size, the most appealing design is Pocock’s approach with interim analysis after 50% information fraction. Due to their favorable performance characteristics, two-stage GSDs are an appealing alternative to fixed sample designs when assessing bioequivalence in highly variable drugs.