Recent pooled analyses suggested that combining chemotherapy with immune checkpoint inhibitors (ICIs) improved survival versus ICI monotherapy for advanced non–small-cell lung cancer (NSCLC) with PD-L1 TPS ≥ 50%. Whether that apparent benefit reflects an incremental effect of chemotherapy or differences between specific ICIs (agent heterogeneity) has been uncertain.
The authors performed an agent-stratified reassessment using the randomized phase 3 trials included in the prior meta-analysis to distinguish chemotherapy effect from ICI heterogeneity.
To reassess whether adding chemotherapy to ICI monotherapy improves overall survival (OS) and progression-free survival (PFS) in advanced NSCLC with PD-L1 TPS ≥ 50%, using agent-stratified comparisons anchored to chemotherapy.
The analysis used the 24 phase 3 randomized clinical trials that were included in the original meta-analysis (search date reported as August 3, 2025). The authors extracted published hazard ratios (HRs) for OS and PFS from each trial in that set.
Two complementary analytic approaches were specified:
Within-agent comparisons: Trials where the same ICI agent was used in both the chemoimmunotherapy arm and the ICI monotherapy arm. This allows direct agent-level comparisons to isolate chemotherapy contribution when the ICI is held constant.
Across-agent comparisons: Comparisons in which different ICIs appear in the chemoimmunotherapy and monotherapy strategies. These comparisons reflect the approach used in typical meta-analyses that do not stratify by agent and therefore can be influenced by ICI heterogeneity.
For within-agent comparisons the authors used a two-stage random-effects meta-analysis. Stage 1 pooled ICI-specific HRs versus chemotherapy for the chemoimmunotherapy and monotherapy strategies and calculated the ratio of HRs (RHR = HRchemoimmuno / HRmono). An RHR < 1 favors chemoimmunotherapy. Stage 2 pooled the ICI-specific RHRs.
For across-agent comparisons the pooled HRs by treatment strategy were used to derive the RHR. Endpoints were OS and PFS. Heterogeneity was reported with I2 and statistical significance with P values, as available in the source.
Within-agent comparisons included 4 ICIs across 13 trials with a combined sample size of N = 3,252. The pooled RHR for OS was 0.94 (95% CI, 0.78–1.13; P = .48; I2 = 0.0%), indicating no statistically significant OS advantage for adding chemotherapy to the same ICI. For PFS the pooled RHR was 0.85 (95% CI, 0.68–1.06; P = .14; I2 = 0.0%), similarly showing no statistically significant benefit.
These within-agent analyses, which hold the ICI agent constant across comparison arms, did not demonstrate a clear advantage from chemoimmunotherapy for either survival endpoint in patients with PD-L1 TPS ≥ 50% disease.
Across-agent comparisons involved 7 different ICIs across 11 trials (N = 2,231). In this framework, the pooled RHR favored chemoimmunotherapy for both endpoints: OS RHR 0.68 (95% CI, 0.50–0.92; P = .01) and PFS RHR 0.46 (95% CI, 0.37–0.58; P < .001).
These results reproduce the pattern seen in the original meta-analysis but do not account for agent-level differences between ICIs used in the treatment strategies.
A sensitivity analysis limited to trials evaluating regimens recommended by the National Comprehensive Cancer Network produced a pooled RHR for OS of 1.02 (95% CI, 0.81–1.28; P = .87). This estimate is consistent with no OS benefit from adding chemotherapy to ICI monotherapy when restricting to guideline-recommended regimens.
The discordance between the within-agent and across-agent analyses suggests that the observed benefit of chemoimmunotherapy in conventional cross-trial meta-analysis is likely driven by ICI heterogeneity rather than a consistent, additive chemotherapy effect in PD-L1 TPS ≥ 50% disease.
Accordingly, these findings support continued consideration of ICI monotherapy as a standard first-line option for patients with advanced NSCLC and PD-L1 TPS ≥ 50%, and they underscore the importance of agent-level stratification when comparing across trials that use different ICIs.
The study used aggregate trial data (published HRs and 95% CIs) extracted from the 24 phase 3 trials included in the original meta-analysis. Details beyond the reported HRs and confidence intervals (for example individual patient data or trial-level covariates) were not reported as available in the source. The authors note that all data produced are available upon reasonable request to the authors.
The source reports no competing interests and acknowledges funding from the Natural Science Foundation of Jilin Province (grant YDZJ202401204ZYTS). Ethical statements indicate that the analysis used published aggregate data.
In agent-stratified, within-agent analyses, adding chemotherapy to ICI monotherapy did not improve OS or PFS for patients with PD-L1 TPS ≥ 50% advanced NSCLC. The survival benefit reported in prior pooled analyses appears to be driven by differences between ICIs when trials are compared across agents. These results recommend caution in interpreting across-trial meta-analyses that do not account for ICI agent heterogeneity and support ICI monotherapy as an appropriate first-line approach in this biomarker-defined population.