External control arms can validate treatment efficacy in single-arm studies

To address uncertainty related to belzutifan’s real world efficacy for treatment of von Hippel-Lindau (VHL) disease associated renal cell carcinoma (RCC), after United States Food and Drug Administration (FDA) approval based on a small single arm study, Linehan et al. employed a novel strategy of using an existing external cohort as a rigorously matched external control arm to provide further validation.1 Although there are only approximately 7000 patients with VHL disease in the United States, this disease can carry a lifetime risk of renal cell carcinoma of approximately 70%, with tumors that can be bilateral and multifocal, and which can require multiple surgical and focal therapies that can negatively impact a patient’s quality of life.2,3 For many years, these tumors had only limited options available for systemic therapy. This changed with the regulatory approval of belzutifan (a hypoxia-inducible factor 2 alpha, HIF-2α, inhibitor), based on results from the LITESPARK-004 (LS-004, MK-6482-004) trial.4 This is a single-arm, phase 2 study of belzutifan in patients with VHL disease-associated RCC (VHL-RCC) and other malignancies. It included 61 patients with VHL disease-associated localized RCC, with the initial trial manuscript noting a 49% objective response rate (ORR) (95% CI = 36% to 62%), a 49% stable disease rate, and a remarkable 96% progression free survival rate at 24 months (95% CI = 87% to 99%).4 Further follow-up data are available, with the most recent update providing results after a median follow-up of 61.8 months. At this point, the ORR was 70% (95% CI = 57% to 81%), including an 11% complete response rate and a 59% partial response rate. Among patients with response, 76.2% remained in response at 48 months.5,6 Additionally, belzutifan treatment was reasonably well tolerated, and no treatment-associated death has been observed in the LS-004 trial.6 Due to the critical unmet need for this rare disease, the FDA approved belzutifan for the treatment of VHL disease-associated cancers, including RCC, in August 2021.7 Despite the impressive positive results from LITESPARK-004, efficacy demonstrated by a small single-arm study in patients with a biologically indolent disease may be subject to cofounders. This warranted further validation of belzutifan’s effectiveness, and the authors used a rigorous statistical approach to compare outcomes of VHL-RCC treated with belzutifan (LS-004 trial) to the natural history of VHL-RCC in the real-world cohort of patients included in a prior VHL-RCC natural history study.8 From this 244-patient observational cohort, the authors sought to optimize validity by using the key eligibility criteria for LS-004, employing propensity score weighting for significant prognostic factors, and including those with adequate imaging availability for objective response rate comparison. Using this methodology, there were large differences in objective response rate (63.9% vs 1.5%) and median time to surgery (not reached vs 51.3 months) between the 2 arms, supporting prior conclusions that the favorable outcomes seen in the LS-004 study are attributable to drug efficacy rather than the indolent natural history of VHL disease-associated RCC. These promising results are an important addition to the body of evidence surrounding the use of belzutifan treatment in this rare disease. They are presented in the context of known difficulties with developing effective treatments for other rare diseases due to the inherent challenge of recruiting for randomized controlled studies from limited populations of patients. These results demonstrate how a rigorously established and matched external control arm may be the best methodology feasible to demonstrate treatment efficacy in the short term. Use of external control arms may also be preferable when compared to using real-world evidence generation, which can take years after regulatory approval and be associated with great cost, and which is therefore difficult to pursue. Indeed, there has been an increased use of external control arm comparison for evaluating the efficacy and safety of treatments for rare malignancies in recent years. While 1 review noted only 23 such studies that have been published since 1996, with varying methods for their design, it also noted an increasing prevalence of these studies in recent years.9 Previously, the US FDA has commented on the use of these external control arms to establish the natural history of diseases and for other comparative analyses, and provided guidance on how to best incorporate these external control arms into robust investigations of drug efficacy.10 The FDA does, however, note the difficulty of developing appropriate external control arms for fair comparison, with some obstacles including non-comparable endpoints, lack of detail regarding relevant biomarkers, missing information potentially relevant to prognosis, and differences in granularity of data when compared to respective intervention arms.10 These concerns are similarly shared by European regulatory agencies, including issues related to limited data, risk of bias, and statistical methodology.11 Looking forward, it could eventually become commonplace for regulatory agencies to require the use of external control arms when considering drug approval applications that are based on single-arm, small cohort trials of treatments for rare diseases. At the same time, the increasing digitization of health records and availability of real-world datasets via commercial and shared platforms will likely continue to support the feasibility of this strategy and the associated data collection. From the use of these external control arms to provide further validation of results derived from trials with limited scope, there is the opportunity to provide increasing confidence in the efficacy of treatments for diseases that are rare and difficult to study, especially in application to real world patients. Eric James Miller (Conceptualization, Data curation, Formal analysis, Methodology, Project administration, Supervision, Validation, Writing—original draft, Writing—review & editing), Saad Atiq (Formal analysis, Writing—review & editing), Jonathan F. Anker (Formal analysis, Writing—review & editing), and Che-Kai Tsao (Conceptualization, Formal analysis, Writing—review & editing) No funding was used for this editorial. E.M., S.A., J.A., and C.T. have no relevant disclosures. No new data were generated or analyzed for this editorial.