Use of real-world evidence to benchmark clinical trial results: A case study in von Hippel-Lindau (VHL) disease.

Abstract

4531 Background: Data from a natural history study can be used to benchmark results from single arm trials with comparable ascertainment of exposures and outcomes. Despite limitations of externally controlled trials, natural history study data can provide important information, especially when clinical trial results provide compelling evidence of change in the established progression of disease. Such results could include partial or complete response for patients receiving drug in a disease where spontaneous regression is not observed. The aim was to conduct a natural history study to benchmark results of a Phase 2 single-arm trial (LITESPARK [LS]-004) for a new HIF-2a inhibitor (belzutifan) using comparable RECIST-based study endpoints. After a median follow-up of 37.8 months, the objective response rate (ORR) for patients with renal cell carcinoma (RCC) in LS-004 was 64.0% (95% CI: 50.6 to 75.8). Methods: A study of existing electronic medical records was conducted for VHL patients with disease-associated RCC undergoing active surveillance at the US National Cancer Institute. VHL patients with ≥1 cm solid RCC, confirmed by independent review, were followed until death or last encounter. Additional LS-004 exclusion criteria were applied to define the natural history study cohort: ≥3.0 cm tumor and procedure performed within 60 days, prior systemic oncologic therapy, and metastasis prior the index date. The index date was the date of the earliest radiology report with measurable disease during the study window (July 2004 - June 2020). An analysis of radiographic outcomes was performed according to RECIST V1.1 for patients with ≥3 scans prior to last encounter, death, or renal therapy using an imaging charter similar to LS-004. Adjudication of best overall response (BOR) was performed if there was discordance between assessments by the two independent radiologists. Results: The natural history study population included 244 VHL patients (100% genetic confirmation; 55% male; median age: 41 years, median follow-up: 10 years). The median size of largest RCC tumor diameter at baseline was 2.5 cm (IQR: 1.9, 3.1), and 76% of patients had ≥1 surgery after the index date. Among patients with ≥3 scans (n=178/244), the confirmed real-world ORR within 5 years of follow-up was 1.8% (95% CI: 0.4%, 5.2%), all with partial response. The real-world BOR was progressive disease and stable disease for 22% and 74% of patients, respectively. Within 5 years, 80% of patients had disease progression. Conclusions: Using similar trial endpoints, this natural history study demonstrates that spontaneous regression of RCC solid tumors is unlikely for patients with VHL undergoing active surveillance in the absence of therapy, and most patients experience disease progression. These findings help contextualize the antitumor activity and efficacy of belzutifan in the treatment of patients with VHL disease-associated RCC.