Title of Research Proposal: The relative efficacy and safety of apalutamide, enzalutamide, abiraterone, and darolutamide for nonmetastatic castration-resistant prostate cancer
Potential Conflicts of Interest: None
Until recently, nonmetastatic castration-resistant prostate cancer (nmCRPC) was managed with continued androgen-deprivation therapy (ADT) and surveillance for metastases. One-third of nmCRPC patients will develop metastasis within two years of diagnosis. Progress has been made since 2018, when four antiandrogen drugs were evaluated to be efficacious for nmCRPC. Namely, apalutamide, enzalutamide, abiraterone, and darolutamide.While all four drugs are promising for nmCRPC, it is unclear which one performs best, as there have been no studies directly or indirectly comparing all of them. Such a comparison is important for decision-making, especially when drug costs are brought into the picture. All four drugs are patent-protected except abiraterone, for which the patent expired in October 2018. The monthly costs are estimated to be $3,275 for the generic version of abiraterone as compared to $12,196 and $12,065 for apalutamide and enzalutamide, respectively. The monthly costs are provided here because these drugs are to be used until the development of metastases. The cost for darolutamide is yet to be determined as its new drug application has just been submitted to the FDA.
Considering the huge disparity in costs, it is critical to evaluate the relative efficacy and safety of these drugs and inform value-based decision-making. Are higher costs justified by better treatment effects or a better safety profile? Or is there actually no difference, making abiraterone is the most economic choice?
Statistical Analysis Plan
Eligibility for trials
- Study designs: RCTs or single-arm trials. Phase I trials, cross-over trials, cluster RCTs, or quasi-randomized trials will be excluded.
- Participants: included nmCRPC patients.
- Interventions: apalutamide, enzalutamide, abiraterone, or darolutamide, in combination with ADT.
- Comparators: any comparators or single-arm trials with no comparator.
- Outcomes: reported efficacy and/or safety outcomes. The outcome of interest in this study is described in the data extraction section.
- Timing and setting: median duration of follow-up more than 12 months; any setting.
- Language and year: no restrictions by language; year since 2011- the earliest approval year of the above-mentioned drugs (abiraterone approved in 2011).
Ways to handle missing data: we will contact principle investigators for missing data. And in case where missing data were not because of the reporting of trial results but because of loss-to-follow up or data collection issues, we will use multiple imputation to address missing data in covariates and outcomes.
Ways to handle differences in outcome measures: we specified our primary and secondary outcomes based on a preliminary search and review of eligible trials. We selected outcomes that are consistently reported across trials.
Ways to handle the differences in study designs: we limited eligible trials to randomized controlled trials and single-arm trials with at least 12 months of median follow-up time. As the drug dosage and regimen are established for the drug therapies of interest, we expect little variation in terms of the administration of intervention. Statistically, we will use random effects model to account for heterogeneity in study results derived from different study designs.
We will conduct network meta-analysis under a Bayesian framework using generalized linear modeling and Markov Chain Monte Carlo (MCMC) methods. The dichotomous outcomes will be determined by using an odds ratio (OR) with 95% credible intervals (CI). The time-to-event outcomes will be determined by using a hazard ratio (HR) with 95% CI. As the dosages for our alternative drug therapies are established, we expect little or no variation and will combine different dosage if necessary. We will estimate the relative treatment effect between any two drugs and use the Surface Under the Cumulative RAnking curve (SUCRA) as well as mean ranks to compare drugs.
The combination of aggregate data and individual patient data:
- We will bring aggregate trial data that are publicly available from five other studies into the Vivli research environment.
- We will then combine individual patient data from Vivli studies with aggregate data from other studies in a statistical model. We will simultaneously model study level treatment effect and individual level treatment effect by Bayesian hierarchical modeling. We will simultaneously adjust for patient level covariates and trial level covariates to estimate the treatment effect.
A Phase III, Randomised, Placebo-controlled, Double-blind Study to Assess the Efficacy and Safety of Once-daily Orally Administered ZD4054 (Zibotentan) 10 mg in Non-metastatic Hormone-resistant Prostate Cancer Patients
Study ID: NCT00626548
Sponsor ID: D4320C00015
Safety and Efficacy Study of Enzalutamide Versus Bicalutamide in Men With Prostate Cancer (STRIVE)
Data Contributor: Pfizer Inc.
Study ID: NCT01664923
Sponsor ID: NCT01664923
We will include aggregated level trial results from the following trials: 1. SPARTAN trial. NCT01946204. 2. ARN-509-001 trial. NCT01171898. 3. PROSPER trial NCT02003924. 4. IMAAGEN trial. NCT01314118. 5. ARAMIS trial. NCT02200614.
Data Contributor: I WILL BRING MY OWN
Study ID: Aggregate results data from five trials
Lin Wang, MD, PhD, Channing Paller, MD, Hwanhee Hong, PhD, Lori Rosman, MLS, Anthony De Felice, MHS, Otis Brawley, MD, G Caleb Alexander, MD, MS, Comparison of Treatments for Nonmetastatic Castration-Resistant Prostate Cancer: Matching-Adjusted Indirect Comparison and Network Meta-Analysis, JNCI: Journal of the National Cancer Institute, 2021;, djab071, DOI: 10.1093/jnci/djab071