Feasibility of Randomized Controlled Trials for Cancer Drugs Approved by the Food and Drug Administration Based on Single-Arm Studies

Abstract

Background The United States (US) Food and Drug Administration (FDA) introduced an Accelerated Approval (AA) pathway to expedite patient access to new drugs. AA accepts less rigorous trial designs, including single arm studies (SAS), owing to perceived lack of feasibility of timely RCTs. Methods We designed hypothetical RCTs with endpoints of objective response rate (ORR), progression free survival (PFS), and overall survival (OS) for FDA approvals based on SAS for solid tumors during 2010–2019. Existing standards of care served as controls. RCTs were designed to detect a difference with power of 0.80, α-error of 5% (two-sided), and 1:1 randomization. Accrual duration was estimated based on participation by < 5% of eligible patients derived from cancer-specific incidence and mortality rates in the US. Results Thirty-one of 172 (18.0%) approvals during the study period were based on SAS. Median sample size was 104 (range = 23–411), and 77.4% were AA. All studies reported ORR, 55% reported duration of response, 19.4% reported PFS, and 22.5% reported OS. Median sample size needed to conduct RCTs with endpoints of ORR, PFS and OS were 206, 130, and 396 respectively. It would have been theoretically possible to conduct RCTs within duration comparable to that required by SAS for 84.6%, 94.1%, and 80.0% of approvals with endpoints of ORR, PFS and OS respectively. Conclusion An overwhelming majority of FDA approvals based on SAS should be feasible as RCTs within a reasonable time-frame. Given the collateral harms to patients and to scientific rigor, drug approval based on SAS should only be permitted under exceptional circumstances.

Basket Trials: Review of Current Practice and Innovations for Future Trials

“…24 In a few instances, evidence from basket trials has led to regulatory approval, as discussed above. 25 Table 1 summarizes several pivotal basket trials implemented in the oncology setting. Enrolling a minimum of 54 patients and maximum of 498, most studies were designed to evaluate objective response rate as the primary end point.…”

Assessing patient burden and benefit: A decade of cabozantinib clinical trials

“…In our sample of 54 cabozantinib trials, 31 trials were single arm and 39 trials were non‐randomized. The reliance on non‐randomized designs for investigating cabozantinib is concerning, as these study designs are likely to report ORRs that may be exaggerated versus a randomized study of the same intervention 24,25 . Rittberg et al suggest that single arm, non‐randomized trials are more susceptible to bias and potentially have decreased statistical power compared to randomized trials 25 .…”

“…The reliance on non‐randomized designs for investigating cabozantinib is concerning, as these study designs are likely to report ORRs that may be exaggerated versus a randomized study of the same intervention 24,25 . Rittberg et al suggest that single arm, non‐randomized trials are more susceptible to bias and potentially have decreased statistical power compared to randomized trials 25 . The decision to conduct a single arm or non‐randomized trial in place of more robust randomized design for cabozantinib may be economically driven, as single arm and non‐randomized trials are cheaper to conduct, require smaller sample sizes, 26 and are regularly used for FDA approval 27,28 .…”

“…24,25 Rittberg et al suggest that single arm, non-randomized trials are more susceptible to bias and potentially have decreased statistical power compared to randomized trials. 25 The decision to conduct a single arm or non-randomized trial in place of more robust randomized design for cabozantinib may be economically driven, as single arm and non-randomized trials are cheaper to conduct, require smaller sample sizes, 26 and are regularly used for FDA approval. 27,28 Prior work has also found that drugs approved based on response rate often have an antecedent approval with as high or higher response rate.…”

Common misconceptions of randomized controlled trials in oncology

“…Finally, an analysis by Rittberg et al demonstrates that is feasible to enrol participants in randomized trials for drugs currently approved by a single arm study. 6 The researchers showed that RCTs could have been conducted within an acceptable time frame using the same primary endpoints, preserving time to market through the accelerated approval pathway without compromising evidence quality. Such a method would prevent drugs reaching the market prematurely.…”