W-28: Two Independent Phase 3 Studies Did Not Meet Pre-specified Endpoints, What's Next?
Poster Presenter
Gajanan Bhat
Vice President
Spectrum Pharmacueticals United States
Objectives
The gold standard for a drug regulatory approval is meeting statistical criteria for pre-specified study endpoints. We attempt to address potential questions that arises when the primary endpoint (s) fail as to what should be done next by using a case of a Phase 3 program.
Method
We attempted to systematically evaluate the data and attempt to answer some of the possible questions on why studies failed and how these studies can be salvaged or lessons learned. We used some of the criteria used by Pocock, et al (NEJM, 2016) to model our introspection.
Results
The Sponsor has completed two randomized controlled trials of single dose of an intravesical therapy (IVT) or placebo immediately after transurethral resection of the bladder tumor (TURBT) with the objective of reducing recurrence rate of bladder tumor from IVT over placebo. A 12% decrease in recurrence rate in IVT over placebo was assumed a priori to power each Phase 3 study. However, when complete, both studies did not individually meet statistical significance due to the observed reduction in recurrence rates were lower than assumed a priori (6.6% and 6.7%). These studies were reviewed by Oncology Advisory Committee and was rejected by the FDA for failure to demonstrate statistical significance.
This poster will delineate multidimensional arguments, clinical as well as statistical to key questions including: was there a potential clinical benefit to patients; was trial underpowered; appropriateness of the definition of primary endpoint or if methods evolved over time; was population appropriate; was treatment regimen appropriate; was there study compliance issue or missing data; was a claim of non-inferiority relevant; do subgroup findings elicit positive signal; is secondary outcome correlated or does it elicit positive effect; was there deficiency in methodology; does any positive external evidence exist; is there a strong biologic rationale favoring treatment. A detailed statistical analysis backed by clinical rationale will be presented in the poster.
Conclusion
The post-study analysis provides a strong argument for the consideration of regulatory approval in the patient population with modification in design parameters. The above example provides an insight into a potential divulgence between scientific merit and regulatory approval criteria in the interpretation of a failed study.
