Harm-Benefit Balance of Immune Checkpoint Inhibitors in Non–Small Cell Lung Cancer

Importance The benefits and harms of immune checkpoint inhibitor (ICI) therapy for lung cancer vary across groups, including those typically underrepresented in randomized clinical trials. Objective To quantify the harms and benefits of ICI-containing regimens in individuals with non–small cell lung cancer and assess heterogeneity across priority subgroups. Design, Setting, and Participants This retrospective cohort study conducted in 2024 used 2013 to 2019 Surveillance, Epidemiology, and End Results (SEER) Medicare data of individuals 66 years or older with non–small cell lung cancer who were exposed to any ICI. Exposures ICI + chemotherapy, single ICI (reference group). Main Outcomes Severe immune-related adverse events (irAE; harm) and mortality (when delayed mortality was the benefit). Severe irAEs were defined using validated diagnosis and medication codes. Mortality was ascertained from Medicare data. Hazard ratios (HRs) were estimated and 95% CIs were stratified by whether an ICI was used as the first or second or later systemic anticancer treatment (SACT) and in subgroups defined by preexisting autoimmune disease, sex, and age. The harm-benefit tradeoff was described as excess severe irAEs per year of life gained in which the gain in survival time was assessed using restricted mean survival time. Results Of 17 681 Medicare beneficiaries, 8797 (49.5%) were female, and the mean (SD) age was 74 (6.0) years. Compared with a single ICI (14 249 [80.6%]), individuals treated with ICI + chemotherapy (3432 [19.4%]) had an elevated risk of severe irAE in the first SACT setting (hazard ratio [HR], 1.18; 95% CI, 1.06-1.30) but not in the second or later SACT setting (HR, 1.04; 95% CI, 0.92-1.19); there was a decreased risk of mortality in the first SACT setting (HR, 0.66; 95% CI, 0.62-0.72) but not in the second or later SACT setting (HR, 0.94; 95% CI, 0.68-1.03). In the first SACT setting, ICI + chemotherapy delayed mortality more among patients with (vs without) autoimmune disease at baseline. For each 1 year of life gained, the risk of severe irAEs was 0.31 (95% CI, 0.09-0.53) and the tradeoff was also statistically significant in men and patients without autoimmune disease. Conclusions The results of this cohort study suggest that given both treatment-related harms and benefits, ICI + chemotherapy use in the first SACT setting requires informed decision-making; the potential benefits of ICI + chemotherapy vs single ICI in high-risk subgroups is encouraging.