OBJECTIVES: Economic modelling of adjuvant cancer therapy is subject to several challenges, including post-recurrence treatment options and limited overall survival (OS) data, reducing applicability of traditional oncology approaches, such as partitioned-survival models. This study aimed to identify the most appropriate modelling approach for adjuvant therapies in the UK setting.

METHODS: A targeted literature review conducted in 2021 identified all National Institute for Health and Care Excellence (NICE) health technology assessments (HTAs) appraising adjuvant treatments. Output from this review was used to inform strategy for two nivolumab HTAs, alongside clinical insights.

RESULTS: Ten HTAs were identified: disease-free survival (DFS), invasive DFS and recurrence-free survival were the most relevant endpoints; limited OS data (immature or unavailable) was reported. Markov modelling was the most common approach (8 HTAs). Where the Markov approach was used, OS could use independent sources; published literature informed post-recurrence mortality (7 HTAs) and general population mortality was assumed for recurrence-free patients (cure assumption; 4 HTAs).

Based on this review, Markov models were developed for nivolumab in the adjuvant treatment of oesophageal cancer and muscle-invasive urothelial carcinoma. Pre- and post-recurrence survival were derived from trial data and published literature, respectively. Cure was assumed for those recurrence-free at 5 years, based on smoothed hazard plots from trial data. Clinical experts validated this approach and relevant survival outcomes. Clinical benefits of nivolumab (life years and quality-adjusted life years) were predominantly accrued in the disease-free state. While outcomes varied by modelled population, the largest model drivers were DFS extrapolations and the timing of cure assumption.

CONCLUSIONS: It is important that adjuvant cancer modelling approaches are able to address relevant challenges (e.g. limited OS data). The Markov structure is most appropriate for the adjuvant setting, allowing application of published literature sources and flexible cure assumptions. Validation from external experts and sources remains essential.