OBJECTIVES: While immuno-oncology (IO) treatments are established in advanced cancer stages, their recent expansion to earlier stages raises questions about implications on IO re-challenge in later stages. This study reviews existing modelling approaches and explores scenarios to assess the significance of IO re-challenge assumptions in health technology assessments (HTAs).

METHODS: A literature review of HTAs published by NICE, CADTH, and HAS was conducted focusing on IOs assessed in early cancer stages. No time restriction was applied. Information extracted included cost-effectiveness model (CEM) type, assumptions about IO re-challenge, modelling approaches for cost and efficacy in the re-challenge phase, scenarios tested and criticism by the HTA agencies. Based on these findings, a state transition CEM with four health states was developed to evaluate the impact of the different IO re-challenge assumptions and methods on the final incremental cost-effectiveness ratio (ICER).

RESULTS: Among 23 identified HTAs of IOs, 15 assessed immune checkpoint inhibitors in adjuvant breast, lung, urothelial carcinoma, renal cell carcinoma and melanoma . The assumption of IO re-challenge was accepted by most agencies and included in 67% of base case analyses or explored in 27% of scenario analyses. While all HTAs considering IO re-challenge evaluated costs, explicit modelling of increased survival due to IO use in metastatic settings was included in 73% of the models, mostly using mean/median OS data weighted by treatment distribution. IO re-challenge assumptions were always supported by clinical expert opinion. The identified approaches were tested in an exploratory CEM. It showed minor impact on the ICER (0-3% vs base case) for adjuvant settings, irrespective of the re-challenge scenarios selected, but significant differences between assumptions for locoregional settings (22-46%).

CONCLUSIONS: IO re-challenge approaches significantly impact the ICER, particularly in locoregional settings or multiple post-progression states, where IOs are considered. Enhanced clinical evidence is crucial for refining IO re-challenge modelling.