OBJECTIVES: To estimate the incremental cost-effectiveness ratio (ICER) of alternating two cycles each of fluorouracil/oxaliplatin chemotherapy (FLOX) and immune checkpoint inhibitor therapy (nivolumab) as first-line treatment for the major subgroup of metastatic colorectal cancer patients with microsatellite stability/mismatch repair proficient disease selected via next-generation sequencing (NGS) compared with the standard of care (FLOX alone) and no biomarker strategy.

METHODS: Individual patient-level data (IPD) from the METIMMOX trial (NCT03388190) including EQ-5D-5L surveys were combined with Norwegian cost data and extrapolated to a lifetime horizon via partitioned survival models. We selected survival models via AIC/BIC and best practices, adopted a healthcare perspective, and estimated costs and effects of drugs, tests, second-line treatments, and end-of-life care. For the post-hoc NGS-based strategy, we used the median tumor mutational burden of >8.0 mutations/megabase as cut-off. Costs were expressed in 2023 Euros and outcomes were discounted by 4% per annum.

RESULTS: At diagnosis, patients were median age 64.5 years; 46% were female; and 58% had fully active performance status. 54% of tumors were left-sided; 72% were RAS- or BRAF-mutated; median number of metastatic sites was two; liver was affected in 83%. Weibull models provided the best IPD fit. Patients accumulated 1.74 (FLOX alone) to 2.01 (NGS strategy) life years. Adding nivolumab for all resulted in an incremental 0.12 QALYs and selecting via NGS 0.20 QALYs at an incremental cost of €75,057 and €16,626, respectively. The treating-all strategy was dominated by the NGS-based strategy which yielded an ICER of €83,376/QALYs.

CONCLUSIONS: Pending confirmatory prospective studies, NGS is an economically attractive testing strategy to select metastatic colorectal cancer patients for immunotherapy in a population-based setting.