Bayesian local exchangeability design for phase II basket trials

Abstract

We propose an information borrowing strategy for the design and monitoring of phase II basket trials based on the local multisource exchangeability assumption between baskets (disease types). In our proposed local‐MEM framework, information borrowing is only allowed to occur locally, that is, among baskets with similar response rate and the amount of information borrowing is determined by the level of similarity in response rate, whereas baskets not considered similar are not allowed to share information. We construct a two‐stage design for phase II basket trials using the proposed strategy. The proposed method is compared to competing Bayesian methods and Simon's two‐stage design in a variety of simulation scenarios. We demonstrate the proposed method is able to maintain the family‐wise type I error rate at a reasonable level and has desirable basket‐wise power compared to Simon's two‐stage design. In addition, our method is computationally efficient compared to existing Bayesian methods in that the posterior profiles of interest can be derived explicitly without the need for sampling algorithms. R scripts to implement the proposed method are available at https://github.com/yilinyl/Bayesian-localMEM.

A Bayesian platform trial design with hybrid control based on multisource exchangeability modelling

“…Liu et al take a test-and-pool approach to limit information sharing to a similarity-defined local scope (local-MEM) while prohibiting borrowing between data sources without sufficient evidence of exchangeability. 12 Kotalik et al 13 combined MEM-based dynamic borrowing with a Pocock-type efficacy boundary to achieve additional sample size reductions for randomized, controlled trials.…”

“…Built on previous work, 11,12,13 we propose a Bayesian group sequential design with hybrid control constructed using a modified MEM approach. The hybrid control consists of both internal and external data from different historical control sources.…”

“…Moreover, it has been noted that the selection of model weights can have a substantial impact on the performance of an MEM method. 12,13,14 By establishing a connection between MEM and power prior, we can estimate the prior and posterior effective sample size for the hybrid control conveniently, which allows us to propose a novel strategy to determine suitable prior weights for MEMs. Last but not least, compared to competing Bayesian approaches, 4,6,7,15 our method is easy to implement because all posterior quantities of interest can be derived in close form without the use of Markov Chain Monte Carlo (MCMC).…”

“…The multisource exchangeability model (MEM) method proposed by Kaizer et al 11 provides a powerful approach for integrating data from different sources by averaging models representing all possible configurations of the external data sources and combining them with the internal data collected during the current trial. Liu et al take a test‐and‐pool approach to limit information sharing to a similarity‐defined local scope (local‐MEM) while prohibiting borrowing between data sources without sufficient evidence of exchangeability 12 . Kotalik et al 13 combined MEM‐based dynamic borrowing with a Pocock‐type efficacy boundary to achieve additional sample size reductions for randomized, controlled trials.…”

“…While consensus among external controls will give us more confidence about the state of belief in the historical response rate, disagreement between external controls will indicate potential differences in patient populations suggesting extra caution should be taken when relying on external data at hand. Moreover, it has been noted that the selection of model weights can have a substantial impact on the performance of an MEM method 12,13,14 . By establishing a connection between MEM and power prior, we can estimate the prior and posterior effective sample size for the hybrid control conveniently, which allows us to propose a novel strategy to determine suitable prior weights for MEMs.…”