Association between inhaled corticosteroids and anti-PD-1/PD-L1 efficacy in patients with NSCLC with concurrent COPD.

Introduction
Lung cancer remains a leading cause of cancer-related mortality worldwide, with non-small cell lung cancer (NSCLC) comprising approximately 85% of all cases.1 The clinical management of NSCLC is further complicated in patients with comorbid chronic obstructive pulmonary disease (COPD).2 Epidemiological studies have demonstrated a bidirectional relationship between these diseases: the incidence of lung cancer is elevated in individuals with COPD,3 and pre-existing COPD is present in 50–70% of patients with lung cancer at diagnosis.4 This comorbidity is frequently associated with accelerated disease progression and poor survival, thereby substantially increasing the clinical burden.5
The management of this patient population remains constrained by two principal clinical challenges. Firstly, COPD-related lung function impairment and increased disease severity significantly limit the eligibility for and tolerance of surgical resection, which remains a cornerstone of curative-intent treatment for early-stage disease.6 Secondly, although targeted therapies offer promise for advanced NSCLC with driver mutations, patients with comorbid COPD exhibit both a lower prevalence of epidermal growth factor receptor (EGFR) mutations and a diminished response to targeted agents even when such mutations are present.7 As a result, the treatment of patients with NSCLC with concurrent COPD has long posed a considerable clinical challenge.
Recent advances in immunomodulation have redirected research focus toward the COPD-associated tumor microenvironment. Characteristic immune dysregulation, including abnormalities in T cells, B cells, and regulatory T cell (Treg)/T helper (Th)17 cell responses, appears to influence both COPD pathogenesis and the efficacy of programmed death-1 (PD-1)/programmed death-ligand 1 (PD-L1) inhibitors.8,9,10 Furthermore, retrospective analyses indicate that immune checkpoint inhibitors (ICIs), particularly those targeting the PD-1/PD-L1 axis, may have enhanced activity in patients with NSCLC with comorbid COPD.11 However, the therapeutic benefit of these agents is heterogeneous,12 and concurrent COPD medications introduce potential confounders. For instance, systemic glucocorticoids (SCS), commonly administered during acute COPD exacerbations, exert immunosuppressive effects by inhibiting cytokines such as interleukin (IL)-10, interferon-γ (IFN-γ), and tumor necrosis factor-α (TNF-α).13,14 They also impair effector T cell function by suppressing IL-12 production and promoting Treg proliferation and differentiation, thereby hindering effective antitumor immunity and potentially worsening outcomes in patients with NSCLC.15 In contrast, inhaled corticosteroids (ICS) exhibit a more complex immunomodulatory profile. ICS have been shown to promote the polarization of tumor-associated macrophages (TAMs) from an M1 to an M2 phenotype,16 which may adversely affect anti-PD-1/PD-L1 efficacy.17 Furthermore, ICS suppresses key pro-inflammatory cytokines, including IL-6, IFN-γ, and TNF-α, and may enhance CD4+/CD8+ T cell infiltration, thereby potentially remodeling the local immune landscape.18,19,20
Nevertheless, existing studies on PD-1/PD-L1 inhibitor efficacy in patients with NSCLC with COPD have not systematically accounted for key confounding variables, including heterogeneity in inhalation regimens. Moreover, it remains controversial whether the ICS-mediated modulation of TAMs and effector T cells synergizes with or antagonizes checkpoint inhibition. Thus, the impact of ICS on PD-1/PD-L1 inhibitor efficacy in this patient subgroup remains unclear. To address this gap, we conducted a retrospective cohort study evaluating the effect of ICS-containing inhalation regimens on outcomes in patients with NSCLC with comorbid COPD receiving PD-1/PD-L1 inhibitor therapy. The aim is to provide a rationale for the personalized identification of patient subsets most likely to benefit from anti-PD-1/PD-L1 immunotherapy.

Results

Baseline characteristics
A total of 268 patients with NSCLC and concurrent COPD who were treated with PD-1/D-L1 inhibitors were initially screened. After applying the predefined exclusion criteria (Figure 1), 46 cases were excluded, resulting in a final analytical cohort of 222 patients prior to matching. This cohort comprised 103 patients in the ICS group and 119 in the N-ICS group. Their baseline characteristics are presented in Table 1.
The majority of the cohort (93.2%) presented with unresectable disease (stage IIIB/C or IV). A small subset (6.8%) had early-stage (I-IIIA) disease and received immunotherapy due to compromised pulmonary function that precluded surgical intervention. Comparative analysis of the pre-matched groups revealed several notable differences. Patients in the ICS group had significantly more severe COPD, as indicated by a higher proportion of Global Initiative for Chronic Obstructive Lung Disease (GOLD) stages 3–4 (73.8% vs. 33.6%, p < 0.001) and poorer lung function (forced expiratory volume in 1 s [FEV1]% predicted: 41.92 ± 15.57 vs. 56.38 ± 18.53, p < 0.001). They also had a higher prevalence of stage IIIB/C disease (51.5% vs. 32.8%, p = 0.003). In contrast, no statistically significant differences were observed between the groups in demographic characteristics (age, gender, body mass index [BMI]), key clinical features (smoking history, Eastern Cooperative Oncology Group performance status [ECOG-PS], histology), immunotherapy regimens, or baseline hematological markers, including eosinophils, lactate dehydrogenase (LDH), neutrophil-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), lymphocyte-monocyte ratio (LMR) and prognostic nutritional index (PNI) (all p > 0.05).
To account for these identified imbalances, we performed 1:1 propensity score matching based on factors from the univariate Cox analysis (Table 2). This process yielded a well-balanced matched cohort of 90 patients (45 in the ICS group and 45 in the N-ICS group). The post-matching baseline characteristics of this cohort are detailed in Table 1.

Survival outcomes
In the pre-matched cohort, patients receiving ICS-containing regimens demonstrated significantly prolonged progression-free survival (PFS) those in the N-ICS group, with median PFS of 14 months (95% confidence interval [CI]: 11.4–16.6) versus 10 months (95% CI: 8.7–11.3), respectively (unadjusted hazard ratio [HR]: 0.683, 95% CI: 0.513–0.908, p = 0.009; Figure 2A). This association still existed after propensity score matching, which balanced baseline characteristics between the groups. The ICS group maintained a superior PFS profile, with a median of 16.0 months (95% CI: 12.8–19.2), in contrast to 9.0 months (95% CI: 7.0–11.0) in the N-ICS group (unadjusted HR: 0.596, 95% CI: 0.379–0.936, p = 0.025; Figure 2B). The beneficial effect of ICS on PFS remained statistically significant after adjusting for potential confounders in a multivariate Cox proportional hazards model. After adjustment for combination with radiotherapy, pre-treatment NLR >3.36, LDH >240 U/L, and PNI <45, the adjusted HR was 0.605 (95% CI: 0.384–0.592, p = 0.030; Table 3 and Figure 2B). This finding was corroborated by a multivariate analysis of the entire pre-matched cohort, which yielded a similarly significant adjusted HR of 0.699 (95% CI: 0.504–0.970, p = 0.032; Figure 2A). No statistically significant difference in overall survival (OS) was observed between the ICS (median OS: 27 months, 95% CI: 24.0–30.0) and N-ICS group (median OS: 26 months, 95% CI: 22.3–29.7) in the unmatched cohort (unadjusted HR: 0.818, 95% CI: 0.568–1.180, p = 0.284; Figure 3). This result was confirmed in a multivariate Cox regression analysis adjusted for influential prognostic factors, which yielded an adjusted HR of 0.765 (95% CI: 0.505–1.158, p = 0.205; Figure 3).

Tumor responses
Treatment with ICS-containing regimens was associated with a significantly higher overall response rate (ORR) compared to the N-ICS group (59.2% vs. 42.9%; p = 0.016). The disease control rate (DCR) was also significantly improved in the ICS group (90.3% vs. 79.0%; p = 0.026). A comprehensive summary of tumor responses is presented in Table 4.

Immune-related adverse events
As summarized in Table 5, immune-related adverse events (irAEs) of any grade occurred in 45.9%(102/222) of the overall cohort, with the majority being grade 1–2 in severity (n = 91, 41.0%). The incidence of irAEs was comparable between the ICS and N-ICS groups (46.6% vs. 44.5%, p = 0.863). The most common irAE across both groups was fatigue (ICS group: 14.6%; N-ICS group: 11.8%, p = 0.676), followed by gastrointestinal reactions. Reactive cutaneous capillary endotheliosis (RCCEP), a known adverse event associated with camrelizumab, was observed in 5 cases receiving regimens containing this agent. Other documented irAEs included rash, checkpoint inhibitor pneumonitis (CIP), thyroid dysfunction, cardiac events, transaminitis, and leukopenia. A detailed summary of adverse events is presented in Table 5.

Subgroup analysis
To assess the prognostic value of key inflammatory and nutritional biomarkers, we performed a subgroup analysis based on pre-treatment NLR, LDH, and PNI levels.
Among patients receiving ICS-containing regimens, those with a low NLR (≤3.36; n = 28) demonstrated significantly longer PFS compared to those with a high NLR (>3.36; n = 75). The median PFS was 19.0 months (95% CI: 17.1–20.9) versus 12.0 months (95% CI: 10.3–13.7), respectively (unadjusted HR: 1.687, 95% CI: 1.045–2.725, p = 0.032). This association remained significant after multivariate Cox regression analysis, adjusting for relevant clinical covariates (adjusted HR: 1.810, 95% CI: 1.036–3.160, p = 0.037) (Figure 4A). A similar effect of NLR on PFS was observed in the non-ICS group. Patients with a low NLR (≤3.36) also experienced superior survival outcomes, with a median PFS of 15.0 months (95% CI: 9.7–20.3) compared to 9.0 months (95% CI: 7.5–10.5) in the high NLR subgroup (unadjusted HR: 1.596, 95% CI: 1.006–2.531, p = 0.047). The significance persisted after multivariate adjustment (adjusted HR: 1.589, 95% CI: 1.002–2.520, p = 0.049) (Figure 4B).
Low LDH levels (≤240 U/L) were also associated with improved PFS, irrespective of ICS use. In the ICS group, patients with low LDH achieved a median PFS of 18.0 months (95% CI: 16.0–20.0), compared to 8.0 months (95% CI: 5.9–10.1) in the high LDH (>240U/L) subgroup (unadjusted HR: 1.649, 95% CI: 1.030–2.642, p = 0.037; adjusted HR: 1.730, 95% CI: 1.046–2.860, p = 0.033; Figure 4C). Similarly, in the N-ICS group, low LDH was associated with a median PFS of 11.0 months (95% CI: 8.9–13.1) versus 6.0 months (95% CI: 3.9–8.1) in patients with high LDH (unadjusted HR: 1.718, 95% CI: 1.153–2.561, p = 0.008; adjusted HR: 1.634, 95% CI: 1.068–2.500, p = 0.024; Figure 4D).
In contrast, no significant association was observed between PNI levels and PFS in either treatment group (Figures 4E and 4F). Corresponding baseline characteristics are detailed in Tables S1–S3.
We subsequently evaluated whether the treatment benefit derived from inhaled corticosteroids was specific to the particular agent used. Of the patients receiving ICS-containing regimens, 42 were managed with fluticasone and 61 with budesonide. Comparative survival analysis demonstrated no statistically significant disparity in progression-free survival between these two subgroups (median PFS: 13.0 months, 95% CI 9.38–16.63 for fluticasone versus 16.0 months, 95% CI 13.04–18.97 for budesonide; log rank p = 0.427; Figure 5). This observation suggests that the survival advantage associated with ICS administration represents a class effect rather than being dependent on any specific corticosteroid molecule. To further validate this finding, we performed a multivariate Cox regression analysis incorporating established prognostic covariates. This adjusted model consistently demonstrated no significant survival difference between the fluticasone and budesonide subgroups (adjusted HR: 1.218, 95% CI: 0.797–1.863, p = 0.362; Figure 5 and Table 6), thereby reinforcing the conclusion that the therapeutic benefit of ICS remains consistent across different corticosteroid types.

Discussion
PD-1/PD-L1 inhibitors have demonstrated clinically meaningful benefits in patients with NSCLC and comorbid COPD.21 However, the pharmacologic interplay between COPD maintenance treatment regimens and PD-1/PD-L1 inhibitors remains poorly characterized. This retrospective cohort study provides initial clinical evidence that ICS-containing maintenance regimens could be associated with enhanced efficacy of PD-1/PD-L1 inhibitors in this dual-diagnosis population. Within the propensity score-matched cohort, ICS-containing regimens demonstrated a significant PFS advantage. Crucially, this survival benefit was consistent across different ICS subtypes, as demonstrated by the comparable PFS outcomes between patients receiving fluticasone-based versus budesonide-based regimens. In the broader unmatched cohort, the ICS group was additionally associated with a superior ORR and DCR, while maintaining a comparable burden of irAEs.
The complex interplay between COPD-related airway inflammation and the tumor immune microenvironment provides a critical biological context for our findings. Airway inflammation in COPD has been shown to enhance tumor-infiltrating lymphocyte density and upregulate PD-1/PD-L1 expression,22,23,24 potentially sensitizing tumors to PD-1/PD-L1 inhibition.25,26 Paradoxically, this inflammatory milieu may simultaneously drive the expansion of Tregs and Th17 cells,10 fostering an immunosuppressive microenvironment that can compromise response to PD-1/PD-L1 inhibition.27,28 This duality underscores the complexity of immunotherapy outcomes in this patient population. Furthermore, mechanisms shared by both lung cancer and COPD pathogenesis, including epithelial-mesenchymal transition (EMT),2 aberrant transforming growth factor β activation, and nuclear factor κB (NF-κB) signaling,29 have been verified as adverse prognostic factors for lung cancer immunotherapy.
The contrasting immunological profiles of ICS versus SCS carry profound clinical implications for patients receiving immunotherapy. Substantial evidence demonstrates that the administration of SCS during PD-1/PD-L1 inhibitor therapy correlates with inferior survival outcomes,30 likely attributable to their broad suppression of alveolar macrophage, T cell, and B-cell function, induction of M2 macrophage polarization,15 and induction of systemic metabolic disturbances.31,32 In contrast, this study indicates that the ICS-containing inhalation regimen may effectively modulate this inflammatory landscape toward a more favorable equilibrium. Unlike the broad immunosuppressive effects of SCS, ICS exerts localized immunomodulatory actions within the pulmonary environment. Mechanistically, ICS likely operates through several complementary pathways: First, by maintaining T cell homeostasis through dose-dependent modulation of CD4+ and CD8+ T cell populations while simultaneously suppressing Th2 activity33,34,35; Second, by suppressing EMT via inhibition of NF-κB signaling36,37,38; and third, by rebalancing Th17/Tregs dysfunction.39 This multifaceted immunomodulation may collectively enhance anti-tumor immunity while preserving PD-1/PD-L1 inhibitors efficacy. Consequently, when corticosteroid intervention becomes necessary for COPD exacerbation management or symptom control in this patient population, ICS represents a strategically superior alternative to systemic formulations, providing localized airway immunomodulation without compromising systemic antitumor immunity.
CIP represents a frequent and potentially life-threatening complication of PD-1/PD-L1 inhibition, occurring in 3.5–19% of treated patients.40 Patients with pre-existing COPD face a particularly elevated risk, as this condition is an established independent risk factor for developing severe (grade ≥3) pneumonitis.41 The underlying mechanisms likely involve the COPD-inflammatory milieu potentiating immune-related lung toxicity. CIP pathogenesis appears driven by dysregulated Treg/Th17 balance and elevated levels of pro-inflammatory cytokines such as IL-2 and IFN-γ within the lung tissue and tumor microenvironment.42,43 In addition, the enhanced infections, inflammatory, and immune dysregulation burden in severe COPD may predispose these patients to a generally higher incidence of irAEs.12 Against this background, our findings acquire particular clinical relevance. Despite the ICS group comprising a significantly higher proportion of patients with severe COPD, a cohort theoretically at maximal risk for pulmonary toxicity, we observed no statistically significant difference in CIP incidence between groups. This suggests that ICS may mitigate COPD-specific inflammatory drivers of pneumonitis through the localized suppression of pro-inflammatory cytokines,35 potentially reducing CIP risk without compromising anti-tumor immunity.
Based on the significant association between pre-treatment NLR, LDH, and PNI with progression-free survival in our multivariate analysis, we performed pre-specified subgroup analyses to further investigate these relationships. Our results consistently identified baseline NLR and LDH levels as reliable predictors of PD-1/PD-L1 inhibitor efficacy, independent of ICS-containing regimen use. These findings align with previous research,44,45 and suggest that systemic inflammatory status, as reflected by NLR, and tumor burden, indicated by LDH levels, represent fundamental biological determinants of PD-1/PD-L1 inhibitors response in patients with NSCLC with comorbid COPD. The prognostic significance of elevated levels of serum LDH is rooted in its role as a surrogate for aggressive tumor behavior, characterized by high proliferative activity, a switch to glycolytic metabolism, and consequent tumor necrosis.46,47 These findings indicate that the integration of these readily accessible biomarkers with ICS treatment history could enable more precise identification of patient subgroups most likely to benefit from PD-1/PD-L1 inhibitor therapy.
An interesting divergence was observed regarding the prognostic value of PNI. While the multivariate Cox model identified PNI as an independent predictor of PFS in the overall population, Kaplan-Meier analysis stratified by PNI failed to demonstrate significant survival differences between subgroups. This apparent discrepancy likely stems from the superior ability of the multivariate model to control for confounding variables. Specifically, the subgroup with PNI <45 contained a higher proportion of patients receiving combined immunotherapy-chemotherapy, as well as those with elevated pre-treatment NLR (>3.36) and PLR (>185). These imbalanced prognostic factors likely obscured the independent association between PNI and survival outcomes in the univariate analysis.
Taken together, this observational research indicates that ICS-containing inhalation regimens are associated with superior efficacy and a favorable safety profile in patients with NSCLC with comorbid COPD receiving PD-1/PD-L1 inhibitor therapy. The consistent treatment effect observed across different ICS molecules reinforces the potential clinical applicability of our findings, suggesting that the choice of a specific ICS agent may be guided by individual patient factors and local practice standards. These findings support the safety and potential efficacy of ICS in this complex patient population and highlight the need for prospective validation to establish ICS as a complementary immunomodulator in lung cancer immunotherapy.

Limitations of the study
Despite these valuable findings, this study has several limitations. First, even with propensity score matching and multivariate adjustment, the retrospective design still cannot completely eliminate unmeasured confounding factors. Details regarding ICS dosage, inhalation technique, and compliance were not uniformly available, potentially introducing classification bias. Second, the absence of overall survival differences may reflect insufficient follow-up duration or the impact of subsequent therapies. Third, the study cohort was predominantly male and exclusively comprised Asian participants, which limits the generalizability of our findings. The underrepresentation of female patients precluded robust subgroup analyses to assess the influence of sex, gender, or both on the outcomes, and the results may not be directly transferable to other racial or ethnic populations. Future studies incorporating diverse racial, ethnic, and sex-balanced cohorts are warranted to validate these findings across broader populations.

Resource availability

Lead contact
Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Hengyi Chen (hilliry@126.com).

Materials availability
This study did not generate new unique reagents.

Data and code availability
All data generated or analyzed during this study are reasonably available from the corresponding author.
This article does not report original code.
Any additional information required to reanalyze the data reported in this study is available from the corresponding author upon request.

Acknowledgments
The authors declare that financial support was received for the research, authorship, and publication of this article. This work was funded by grants from the 10.13039/501100005230Natural Science Foundation of Chongqing (Grant number: CSTB2022NSCQ-MSX0765). The graphical abstract was created with BioGDP.com.48

Author contributions
L.Z. and H.Y.C. conceptualized the study; F.W.W., H.Z., S.L., and Y.H.P. performed investigative studies; L.L.Z. and H.Y.C. performed data analysis; L.L.Z. and H.Z. interpreted the results and wrote the article; F.W.W., H.Z., S.L., Y.H.P., and H.Y.C. provided critical insights and advice and reviewed the article.

Declaration of interests
The authors declare no competing interests.

STAR★Methods

Key resources table

Experimental model and study participant details
This multicenter retrospective cohort study enrolled adult patients with histologically confirmed, inoperable NSCLC and comorbid COPD from three medical institutions in China between January 2021 and June 2023. The participating centers included the Seventh People's Hospital of Chongqing, Chongqing Hygeia Hospital, and Shapingba Hospital Affiliated with Chongqing University. The study was conducted in accordance with the principles of the Declaration of Helsinki (2013 revision) and received approval from the Medical Ethics Committee of the Seventh People's Hospital of Chongqing (Approval No.: 202402002). Due to the retrospective nature of the study and the use of anonymized clinical data, the requirement for informed consent was waived. All patient records were de-identified prior to analysis to ensure confidentiality.
Inclusion criteria comprised: (I) Histologically or cytologically confirmed inoperable primary NSCLC; (II) At least one measurable lesion as defined by Response Evaluation Criteria in Solid Tumors (RECIST) version 1.149; (III) Diagnosis of COPD meeting the 2020 GOLD criteria, confirmed by post-bronchodilator FEV1/forced vital capacity (FVC) ratio < 0.7050; (IV) Completion of at least two cycles of PD-1/PD-L1 inhibitor therapy, along with continuous COPD maintenance treatment throughout the immunotherapy period.
Exclusion criteria were: (I) Concurrent primary malignancies in other organ systems or major comorbidities with significant impact on survival (e.g., end-stage renal disease, New York Heart Association [NYHA] class III/IV heart failure); (II) Active pulmonary infection at the initiation of anti-PD-1/PD-L1 immunotherapy; (III) Insufficient clinical data or loss to follow-up.
COPD maintenance regimens were classified according to GOLD guidelines.50 The ICS group consisted of patients receiving inhaled dual or triple therapy containing corticosteroids. The N-ICS group included patients treated with bronchodilators only (long-acting β2 agonists and/or long-acting muscarinic antagonists) or other non-ICS regimens. Group assignment was verified through comprehensive prescription review and inhaler technique validation.
The study cohort comprised 222 participants of Asian descent, all of whom were of Han Chinese ethnicity, reflecting the predominant demographic composition of the enrolled medical institutions in China. The cohort was predominantly male (91.9%, 204/222), with female participants constituting 8.1% (18/222). This sex distribution mirrors the higher epidemiological prevalence of both NSCLC and COPD in the male population within the studied demographic. Given the substantial imbalance in sex representation and the limited number of female participants, subgroup analyses stratified by sex were not statistically powered to detect meaningful differences. Consequently, the influence of sex, gender, or both on the study outcomes (including progression-free survival and treatment response) could not be robustly assessed.

Method details

Data collection
Demographic and clinical characteristics, including age, gender, smoking history, and BMI, were systematically recorded. Disease-specific variables consisted of the ECOG-PS,51 TNM stage (8th edition),52 histological subtype, and key pulmonary function parameters. The profile of respiratory comorbidities was characterized by COPD disease stage and the corresponding maintenance therapeutic regimens.
Baseline hematological parameters were assessed from peripheral blood samples collected prior to the initiation of immunotherapy. These included absolute eosinophil count, NLR, PLR, LMR, PNI, and serum LDH levels.

Assessment of disease severity
The severity of airflow obstruction was graded based on post-bronchodilator FEV1, in accordance with the GOLD criteria.50 The stages were defined as follows: GOLD 1 (Mild): FEV1 ≥ 80% predicted; GOLD 2 (Moderate): 50% ≤ FEV1% < 80% predicted; GOLD 3 (Severe): 30% ≤ FEV1% < 50% predicted; and GOLD 4 (Very severe): FEV1% < 30% predicted.
Functional status was evaluated using the ECOG-PS scale,51 which is categorized as: 0 (Fully active, without restriction); 1 (Restricted in strenuous activity but ambulatory and able to carry out light work); 2 (Ambulatory and capable of all self-care but unable to perform work activities; up and about for more than 50% of the day spent out of bed); 3 (Capable of only limited self-care; confined to bed or chair for more than 50% of waking hours); 4 (Completely disabled; unable to carry out any self-care); and 5 (Dead).

Endpoints and efficacy evaluation
The primary endpoint of this study was PFS. Secondary endpoints included OS, ORR, DCR, and the incidence of irAEs.
PFS was defined as the time from the first administration of PD-1/PD-L1 inhibitor therapy to the date of radiologically confirmed disease progression according to RECIST 1.1,49 or death from any cause, whichever occurred first. Patients who had not experienced a PFS event by the data cutoff date were censored at their last radiographic assessment. OS was defined as the time from treatment initiation to death from any cause. Mortality from causes not related to lung cancer or COPD was considered a competing event and handled by censoring at the time of death. Surviving patients were censored at the last known follow-up date.
Tumor response was assessed by contrast-enhanced computed tomography or magnetic resonance imaging at approximately 8-week intervals, in accordance with RECIST 1.1.49 Responses were categorized as complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD). ORR was calculated as the proportion of patients achieving a confirmed CR or PR. DCR was defined as the proportion of patients with a best overall response of CR, PR, and SD.
All patients underwent systematic monitoring for irAEs from the initiation of ICI treatment until occurrence of disease progression, death, or the last follow-up. IrAEs were systematically identified through comprehensive review of electronic medical records, including structured oncology treatment notes, nursing assessments, radiology reports, and laboratory data. The severity of irAEs was assessed using the Common Terminology Criteria for Adverse Events, version 5.0 (CTCAE v5.0). Attribution to immunotherapy was determined by the treating oncologist during clinical care. Clinical visits were scheduled at approximately 4-week intervals until disease progression. Survival status was subsequently updated via telephone follow-up. All endpoint data were censored as of August 31, 2024.
To minimize selection and assessment bias, all cases underwent a blinded re-evaluation by two independent pulmonologists. This process included: (I) Confirmation of COPD diagnosis through reinterpretation of spirometry data; (II) verification of RECIST 1.1 compliance via independent imaging reassessment; (III) validation of the documented inhalation regimen through a pharmacy audit trail; and (IV) irAEs were verified independently by the study investigators through retrospective review, based on temporal relationship to treatment, clinical presentation, and exclusion of alternative etiologies. Any discrepancies between the two were adjudicated by a third senior oncologist to reach a final consensus.

Quantification and statistical analysis
The data analysis was performed with SPSS version 26.0 (IBM, Armonk, New York, USA).53 As established in prior studies,54,55,56 elevated pre-treatment levels of NLR (> 3.36), PLR (> 185), LDH (> 240 U/L), LMR (> 4), and reduced PNI (< 45), have been associated with inferior outcomes in patients treated with PD-1/PD-L1 inhibitors. Consistent with these findings and considering the non-normal distribution of these baseline variables, we dichotomized NLR, PLR, LMR, LDH, and PNI for subsequent analyses. Patient characteristics were summarized by group, with categorical variables expressed as frequencies and percentages, and continuous variables reported as mean ± standard deviation or median (interquartile range). To evaluate baseline characteristics between groups, categorical variables were compared using the Chi-squared (χ2) test or Fisher’s exact test, as appropriate, while the t-test was applied for continuous variables. To minimize potential confounding, we performed 1:1 propensity score matching via logistic regression with a caliper width of 0.1. We conducted univariate Cox proportional hazards regression to analysis factors potentially affecting PFS (with a significance threshold of P <0.10 for inclusion, results detailed in Table 1). These factors included COPD stage (GOLD 1 vs. GOLD 2 vs. GOLD 3 vs. GOLD 4), combination with radiotherapy, pre-treatment NLR > 3.36, pre-treatment PLR > 185, pre-treatment LDH > 240U/L, and pre-treatment PNI < 45. Following matching, between-group balance was re-assessed to evaluate the effectiveness of the adjustment.
The Kaplan-Meier method was employed to calculate PFS for both unmatched and matched cohorts, with results reported alongside their corresponding 95% CIs. The HRs for PFS with 95% CIs were evaluated with univariate Cox proportional hazards models. Subsequently, multivariate Cox regression was performed to calculate adjusted HRs for PFS with 95% CIs, controlling for potential prognostic covariates. Secondary outcomes were assessed only in the unmatched cohort, and the unadjusted and adjusted HRs with 95% CIs were derived from univariate and multivariate Cox models, respectively. Differences in ORR and incidence of irAEs between groups were compared using the χ2 test or Fisher’s exact test, as appropriate. All statistical details can be found in the figure legends, figures, and main text. P < 0.05 was considered statistically significant.

Additional resources
Not applicable. This study did not utilize any external protocols, troubleshooting forums, or additional resources beyond those described in the STAR Methods section.