OBJECTIVES: To assess the cost-effectiveness of atezolizumab as adjuvant treatment, compared to best supportive care (BSC), following complete resection and platinum-based chemotherapy in adult patients with early-stage non-small cell lung cancer (NSCLC) at high risk of recurrence, whose tumors have programmed death-ligand 1 (PD-L1) expression ≥50%, excluding epidermal growth factor receptor-mutant or anaplastic lymphoma kinase-positive NSCLC.

METHODS: A Markov model including disease-free survival (DFS), locoregional recurrence, metastatic recurrence (1^st and 2^nd lines), and death health states was used. The model uses data from IMpower010 (atezolizumab’s pivotal trial in this therapeutic indication) for DFS and from external sources for the remaining health states. As patient reported outcomes were not collected during IMpower010, utility scores are based on the literature and on IMpower150. The use of this trial, in which atezolizumab was evaluated in metastatic NSCLC patients, allowed estimating EQ-5D-3L scores with Portuguese tariffs for the respective health states and standardizing the scores available from the literature. Portuguese-specific disease management resource use was based on an experts panel and Portuguese diagnosis-related group microdata. The main sources for unit costs were national legislation and official drug cost databases. The analysis was conducted from the National Health Service perspective, assuming a lifetime horizon and a 4% discount rate. Deterministic and probabilistic sensitivity analyses show that results are robust.