Evaluation of two-stage designs of Phase 2 single-arm trials in glioblastoma: a systematic review

BACKGROUND: Due to economical and ethical reasons, the two-stage designs have been widely used for Phase 2 single-arm trials in oncology because the designs allow us to stop the trial early if the proposed treatment is likely to be ineffective. Nonetheless, none has examined the usage for published articles that had applied the two-stage designs in Phase 2 single-arm trials in brain tumor. A complete systematic review and discussions for overcoming design issues might be important to better understand why oncology trials have shown low success rates in early phase trials. METHODS: We systematically reviewed published single-arm two-stage Phase 2 trials for patients with glioblastoma and high-grade gliomas (including newly diagnosed or recurrent). We also sought to understand how these two-stage trials have been implemented and discussed potential design issues which we hope will be helpful for investigators who work with Phase 2 clinical trials in rare and high-risk cancer studies including Neuro-Oncology. The systematic review was performed based on the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA)-statement. Searches were conducted using the electronic database of PubMed, Google Scholar and ClinicalTrials.gov for potentially eligible publications from inception by two independent researchers up to May 26, 2022. The followings were key words for the literature search as index terms or free-text words: “phase II trials”, “glioblastoma”, and “two-stage design”. We extracted disease type and setting, population, therapeutic drug, primary endpoint, input parameters and sample size results from two-stage designs, and historical control reference, and study termination status. RESULTS: Among examined 29 trials, 12 trials (41%) appropriately provided key input parameters and sample size results from two-stage design implementation. Among appropriately implemented 12 trials, discouragingly only 3 trials (10%) explained the reference information of historical control rates. Most trials (90%) used Simon’s two-stage designs. Only three studies have been completed for both stages and two out of the three completed studies had shown the efficacy. CONCLUSIONS: Right implementation for two-stage design and sample size calculation, transparency of historical control and experimental rates, appropriate selection on primary endpoint, potential incorporation of adaptive designs, and utilization of Phase 0 paradigm might help overcoming the challenges on glioblastoma therapeutic trials in Phase 2 trials.