Pembrolizumab vs cemiplimab for the treatment of advanced non-small cell lung cancer with PD-L1 expression levels of at least 50%: A network meta-analysis and cost-effectiveness analysis

BACKGROUND: Pembrolizumab and cemiplimab have been approved as treatment for advanced non-small-cell lung cancer (NSCLC) with high programmed death ligand-1 (PD-L1) expression. This study aimed to evaluate the cost-effectiveness of pembrolizumab compared with that of cemiplimab in the treatment of advanced NSCLC with high PD-L1 expression from a societal perspective in the United States. MATERIALS AND METHODS: Cost-effectiveness analysis integration of the network meta-analysis framework was performed using data from the EMPOWER-Lung 1, KEYNOTE 024, and KEYNOTE 042 phase 3 randomized clinical trials. A network meta-analysis including 2289 patients was constructed, and the Markov and partitioned survival (PS) models were used to assess the cost-effectiveness of pembrolizumab compared with that of cemiplimab for the treatment of high PD-L1 expression (≥50% of tumor cells). The time horizon was 10 years. The main outcomes were overall costs, incremental cost-effectiveness ratios (ICERs), quality-adjusted life-years (QALYs), life-years, incremental net health benefits (INHB), and incremental net monetary benefits (INMB). The robustness of the model was verified using one-way and probabilistic sensitivity analyses, and subgroup analyses were conducted. RESULTS: Treatment of advanced NSCLC with high PD-L1 expression with pembrolizumab achieved 0.093 QALYs and was associated with an incremental cost of $10,657 compared with cemiplimab, yielding an ICER of $114,246/QALY. The ICER in the PS model was similar to that in the Markov model, with a difference of $3,093/QALY. At a willingness-to-pay (WTP) threshold of $100,000/QALY, INHB, and INMB of pembrolizumab were -0.013 QALYs and -$1,329, respectively, and the probability of cemiplimab was 51% when compared with pembrolizumab. When the WTP threshold increased to $150,000/QALY, the INHB and INMB of pembrolizumab were 0.022 QALYs and $3,335, respectively, and the probability of pembrolizumab was 51.85%. One-way sensitivity analysis indicated that the models were sensitive to pembrolizumab and cemiplimab costs. Subgroup analysis revealed that treatment with pembrolizumab was related to a higher INHB in several subgroups, including patients with brain metastases at baseline. CONCLUSION: Our findings suggest that the WTP threshold should be considered when choosing between cemiplimab and pembrolizumab to treat advanced NSCLC with high PD-L1 expression. Reducing the cost of pembrolizumab may lead to valuable outcomes.