All Times EDT
Keywords: Subgroup, Heterogeneity, Novel Framework
Potential patient heterogeneity in treatment effects challenges the clinical development of many investigational medical products. On the other hand, scientific advancement sometimes reveals that what used to be considered separate diseases may respond to the same medical product. For example, since May 2017, the US Food and Drug Administration has granted accelerated approval to three cancer treatments for tissue-agnostic indications, i.e., an indication based on a common biomarker across different tumor types rather than the location in the body where the tumor originated.
These tissue-agnostic clinical trials used tumor response rates (TRRs), well defined for all involved tumor types, as primary endpoints. Such is not the case for an investigational product intended for treatment of non-muscle invasive bladder cancer (NMIBC), discussed at an FDA advisory committee meeting in 2015. The single-arm trial used disease free survival (DFS) as the primary endpoint. However, the trial population consisted of both patients entering the trial with disease and those whose disease were resected and therefore disease-free at trial entry. For the subgroup of patients entering the trial with disease, a single-arm design with an endpoint of TRR was considered adequate to evaluate the efficacy, but DFS was not well defined for this subgroup. On the other hand, DFS was well defined for patients without disease at trial entry, but a randomized controlled design with concurrent control was thought to be needed to evaluate efficacy.
We propose a novel class of trial designs and corresponding analysis methods for similar cases. That is, cases in which there is strong a priori support that the investigational product will either be effective in two clinically distinct subgroups or in neither subgroup. In addition, the two subgroups may be investigated with different designs (e.g., single-arm vs concurrent-controlled) and different endpoints (e.g., TRR vs. DFS).