Effectiveness and safety of combination immunotherapy with or without ipilimumab according to PD-L1 expression in patients with non-small cell lung cancer: a multi-center retrospective cohort study

Background: The standard first-line treatment for patients with driver mutation-negative non-small cell lung cancer (NSCLC) is chemotherapy and immunotherapy, including chemotherapy plus anti-programmed death-1 (PD-1)/programmed cell death ligand 1 (PD-L1) antibody (C + PD-1/PD-L1) or anti-PD-1 antibody (nivolumab) and anti-cytotoxic T-lymphocyte antigen-4 antibody (ipilimumab) (C + NI). Subgroup analyses of phase III trials and network meta-analyses suggest that treatment effectiveness varies according to the PD-L1 tumor proportion score (TPS), with the potential benefit of ipilimumab-containing regimens in PD-L1-negative patients. However, real-world evidence across PD-L1 subgroups, particularly in patients with PD-L1 TPS between 1–24%, remains limited. This study compared the real-world effectiveness of C + PD-1/PD-L1 and C + NI across PD-L1 TPS subgroups and evaluated their safety in the overall population. Methods: Consecutive patients with advanced or metastatic NSCLC who received C + PD-1/PD-L1 or C + NI from March 2019 to April 2022 at five institutions were included retrospectively. Patients with driver mutations or those who received second-line or later treatments were excluded. In the C + PD-1/PD-L1 group, platinum doublets were administered for up to four cycles, with pemetrexed maintenance permitted when the initial regimen included pemetrexed, whereas in the C + NI group, platinum doublets were limited to a maximum of two cycles with no maintenance chemotherapy. All chemotherapeutic agents were administered at standard doses according to the approved prescription information. Medical records were reviewed for baseline characteristics, treatment exposure, PD-L1 TPS, and grade ≥3 immune-related adverse events. Patients were routinely followed-up according to institutional practice, and survival outcomes were assessed through regular outpatient visits and medical record reviews. Results: A total of 212 patients received C + PD-1/PD-L1 and 55 patients received C + NI. The proportion of patients with a low PD-L1 TPS was significantly higher in the C + NI group than in the C + PD-1/PD-L1 cohort. Overall survival (OS) in the overall population was not significantly different [median OS: 21.1 months vs. not reached, HR: 1.06; 95% confidence interval (CI): 0.68–1.67, P=0.79], and OS in the PD-L1 TPS <1% or 1–24% group did not significantly differ between the two groups (PD-L1 TPS: <1%: median OS: 21.5 months vs. not reached, HR: 1.14; 95% CI: 0.56–2.30, P=0.72, PD-L1 TPS: 1–24%: median OS: 22.4 vs. 17.8 months, HR: 0.76; 95% CI: 0.34–1.70, P=0.50). Multivariate Cox regression analysis showed that poor performance status was an independent prognostic factor for the overall population. In the subgroups with PD-L1 TPS <1% or 1–24%, no variables were identified as independent prognostic factors. The incidence of grade ≥3 immune-related adverse events was significantly more frequent in the C + NI group. Conclusions: C + NI did not demonstrate superior effectiveness compared with C + PD-1/PD-L1 treatment across the PD-L1 TPS subgroups and was associated with increased toxicity, suggesting no clear clinical advantage of prioritizing C + NI in routine practice. C + NI should be reserved for selected cases in which the potential benefits outweigh the risks.