Abstract
Background
Lung cancer remains the leading cause of cancer-related mortality worldwide. While immune checkpoint inhibitors (ICIs) have improved survival outcomes for some patients, their efficacy and adverse effects vary significantly. Thus, developing accurate and practical prognostic tools is essential to optimize treatment decision-making.
Methods
This retrospective study analyzed 436 lung cancer patients treated with ICIs, who were randomly divided into training (70%) and validation (30%) cohorts. Independent prognostic factors for overall survival (OS) and progression-free survival (PFS) were identified using LASSO regression and multivariate Cox regression. Nomograms were constructed based on clinical and blood biomarkers. Model performance was assessed using the concordance index (C-index), ROC curve, calibration curve, and decision curve analysis (DCA). Kaplan–Meier analysis validated patient stratification.
Results
The key independent predictive factors for OS and PFS included neutrophil-to-lymphocyte ratio (NLR), previous surgery, liver metastasis, clinical stage, treatment lines, and treatment response evaluation. The nomograms achieved C-index values of 0.709 (OS) and 0.730 (PFS) in the training cohort, with validation C-indexes of 0.655 (OS) and 0.694 (PFS). The ROC curves demonstrated good predictive accuracy for 12-, 24-, and 36-month outcomes. High-risk patients exhibited significantly shorter median OS and PFS (P < 0.001).
Conclusion
The nomograms developed in this study, integrating clinical and blood biomarkers, provide a cost-effective, simple, and accurate tool for predicting the prognosis of lung cancer patients receiving ICIs treatment, to facilitate personalized clinical decision-making.
Keywords: Lung cancer, Immune checkpoint inhibitors, Prognostic biomarkers, Nomogram
Background
Lung cancer, a major malignancy, poses a significant threat to human health and is responsible for the highest rates of illness and death worldwide [1]. In 2022, the Global Cancer Epidemiology Statistics (GLOBOCAN) estimated approximately 20 million new cases of lung cancer worldwide, resulting in about 9.7 million deaths [1]. According to the National Cancer Center, approximately 1.06 million new lung cancer cases and 733,300 deaths were reported in China, establishing it as the country's most deadly malignancy [2]. Immune checkpoint inhibitors (ICIs) targeting PD-(L)1 have emerged as a cornerstone therapy for patients lacking other targeted options [3–5]. However, response rates vary significantly, with only a subset of patients deriving long-term benefits [5–7]. Identifying reliable predictive biomarkers is essential to distinguish responders from non-responders and guide treatment decisions.
In lung cancer, molecular indicators such as tumor PD-L1 expression, tumor mutational burden (TMB), microsatellite instability (MSI), and the presence of tumor-infiltrating immune cells are utilized to assess susceptibility to PD-(L)1 inhibitors [8–10]. These markers collectively inform the immune landscape of a tumor and help predict therapeutic responses. However, these markers inadequately capture the full variability in outcomes, likely due to the complex nature of the immune response to cancer [11, 12].
The tumor microenvironment (TME) plays a pivotal role in modulating immune responses to ICIs. Components of the TME, including immune cell infiltration, cytokine profiles, and stromal composition, critically influence immunotherapy efficacy. For instance, immunosuppressive elements such as regulatory T cells (Tregs) and myeloid-derived suppressor cells (MDSCs) within the TME can attenuate anti-tumor immunity, while high CD8 + T cell infiltration correlates with improved outcomes [13–15]. Recent studies emphasize the dynamic interplay between TME heterogeneity and ICI responsiveness. For example, spatial transcriptomic analyses reveal that immune-excluded or desert TME phenotypes are associated with resistance to PD-1 inhibitors [16]. Additionally, metabolic reprogramming within the TME, such as lactate accumulation from aerobic glycolysis, may impair T cell function and reduce therapeutic efficacy [17]. These findings underscore the need to integrate TME characteristics into prognostic models to refine patient stratification.
In clinical practice, PD-L1 expression remains the most widely used biomarker for predicting the efficacy of immunotherapy. However, studies have shown only a weak correlation between PD-L1 expression levels in tumor biopsies and the overall response rate (ORR) to ICIs [18]. In lung cancer, evaluating factors such as smoking history, TMB, MSI, high expression of CTLA4, low expression of CX3CL1, and CD8 + T cell infiltration in the tumor microenvironment (TME) may provide more robust predictive capabilities for responses to anti-PD-1/PD-L1 therapies compared to histopathological PD-L1 quantification [19–21]. However, these markers have yet to be developed into a clinically robust and practical biomarker signature.
Nomograms, graphical tools integrating multiple predictors, offer a practical solution for personalized prognosis estimation [22]. This study aims to develop and validate a nomogram combining clinical and biomarker data to predict outcomes in lung cancer patients prior to ICI initiation.
Patients and methods
Patients
A total of 436 lung cancer patients treated with ICIs at Nanjing Jinling Hospital (January 2019–February 2024) were retrospectively analyzed. Patients were randomly allocated into training (70%) and validation (30%) cohorts.
Inclusion and exclusion criteria
The inclusion criteria encompassed the following: (I) Patients aged ≥ 18 years with histologically confirmed lung cancer subtypes, including adenocarcinoma, squamous cell carcinoma, or small cell carcinoma; (II) Documented metastatic subtypes (e.g., liver, bone, or brain metastases) via imaging (CT/MRI) or biopsy; (III) Received ≥ 2 cycles of immune checkpoint inhibitors (ICIs). Patients with prior systemic therapies (chemotherapy, targeted therapy) were eligible if they met this criterion. Revised exclusion criteria were as follows: (I) Poor general condition (ECOG PS ≥ 3) or comorbidities (e.g., immune deficiencies) precluding tolerance to immunotherapy or chemotherapy; (II) Incomplete baseline biomarker data (NLR, PLR, CEA) or missing follow-up records; (III) Missing data > 5% (e.g., CA125 in 8% of cases), which were addressed via multiple imputation by chained equations (MICE) under the missing-at-random assumption.
Treatment regimens
The treatment regimens for ICIs included: 293 patients (67.2%) received combined chemotherapy; 57 patients (13.1%) were treated with both chemotherapy and anti-angiogenesis therapies (AATs); 54 patients (12.4%) received only AATs; and 32 patients (7.3%) underwent immune monotherapy.
Data collection
Baseline characteristics collected from patients included age, sex, smoking history, metastatic sites at baseline, presence of ascites, prior treatments (surgery, radiotherapy, chemotherapy, targeted therapy, and anti-angiogenic therapy), site of disease progression, histological subtypes (adenocarcinoma, squamous cell carcinoma, small cell carcinoma, and others), gene mutations, cancer staging, types of ICIs administered, immunotherapy combination regimens, treatment lines, efficacy assessments, and immunotherapy. Additional factors such as treatment adverse reaction scores, ECOG PS, neutrophil/lymphocyte ratio (NLR), platelet/lymphocyte ratio (PLR), hemoglobin levels (HGB), carcinoembryonic antigen (CEA), carbohydrate antigen 125 (CA125) and 199 (CA199), white blood cell count (WBC), time to progression, date of death.
Evaluation of curative effect
Imaging evaluations (CT/MRI) were performed at baseline, after every two treatment cycles, and at suspected progression. Scans were independently reviewed by two radiologists blinded to clinical outcomes, with discrepancies resolved by a third reviewer. Responses were categorized per RECIST 1.1: complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD) [23]. Target lesions were measured in axial planes, and non-target lesions were assessed qualitatively. Preliminary therapy response evaluations were conducted after two treatment cycles or as medically necessary. The primary objectives of the study were to determine overall survival (OS) and progression-free survival (PFS). OS was defined as the duration from the initiation of ICIs treatment to the last follow-up or death, whereas PFS was defined as the time from the initiation of ICIs treatment until disease progression. Patient safety was evaluated, with the severity of adverse events categorized according to the National Cancer Institute's Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0. Informed consent was obtained from all participants in accordance with the ethical guidelines outlined in the Declaration of Helsinki (2013 revision), ensuring that all aspects of the study respected participant autonomy and safety.
Biomarker detection
Peripheral blood samples were collected pre-treatment. NLR and PLR used the Mindray BC-6800Plus automated hematology analyzer (Mindray Corporation, Shenzhen, China), which was calibrated, controlled, and maintained according to the manufacturer's recommendations. CEA, CA125, and CA199 levels were quantified using Maglumi X8 fully automated chemiluminescence immunoassay analyzer (New Industry Biotech, Shenzhen, China). All assays followed manufacturer protocols, with inter- and intra-assay coefficients of variation < 10%.
Statistical analysis
Statistical analyses were conducted using R software (RRID:SCR_000432, version 4.1.3). Baseline characteristics were analyzed using Chi-square or Fisher's exact tests for categorical variables (reported as frequencies/proportions) and nonparametric tests (e.g., Mann–Whitney U) for non-normally distributed continuous variables (reported as medians with interquartile ranges). For variables with > 5% missing data (e.g., CA125 in 8% of cases), multiple imputation by chained equations (MICE) with predictive mean matching was applied, generating five imputed datasets; pooled results were derived using Rubin's rules. Prognostic factors were identified via LASSO regression, with the optimal penalty parameter (λ) determined through tenfold cross-validation. Variables with non-zero coefficients were retained, and those statistically significant (P < 0.05) in multivariate Cox regression were incorporated into the nomogram, presented as hazard ratios (HRs) with 95% confidence intervals (95%CIs). Multivariable models adjusted for confounders, including comorbidities (e.g., chronic obstructive pulmonary disease, diabetes), ECOG performance status, and combination therapies (chemotherapy/anti-angiogenic agents). Propensity score weighting (inverse probability of treatment weighting) was employed to balance baseline characteristics between treatment lines [24]. Model performance was evaluated using Harrell's concordance index (C-index), time-dependent ROC curves (AUC for 12-, 24-, and 36-month survival), and decision curve analysis (DCA) to assess clinical utility. Patients were stratified into low-risk and high-risk groups based on nomogram-derived thresholds. Survival differences were analyzed via Kaplan–Meier curves and log-rank tests, with statistical significance set at P < 0.05.
Ethical statement
The authors are responsible for all aspects of this work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study adhered to the ethical principles set forth in the Declaration of Helsinki (2013 revision) and received approval from the Ethics Committee of Nanjing Jinling Hospital (No. DZQH-KYLL-23-06). All participants provided informed consent.
Result
Patient clinical characteristics
The study included a total of 436 patients, as depicted in the flowchart presented in Fig. 1. These patients were assigned to either the training set (n = 306) and the validation set (n = 130). The age distribution revealed that 314 patients were younger than 70 years, while 122 patients were aged 70 years or older.The histological breakdown of the patient cohort showed that 209 patients (47.9%) had lung adenocarcinoma, 171 (39.2%) had squamous cell carcinoma, 38 (8.7%) had small cell lung cancer, and 18 (4.1%) had various other neoplasms. Prior treatments were as follows: 108 patients had surgery, 80 had radiation therapy, 215 had chemotherapy, 66 had targeted therapy, and 85 had antivascular therapy. Fluid accumulation prior to baseline assessment was observed in 30 patients (6.9%). PD-1 monoclonal antibody usage was: 137 patients received camrelizumab (31.4%), 23 received pembrolizumab (5.3%), 89 received sintilimab (20.4%), 144 received tislelizumab (33.0%), and 43 received other treatments (9.9%). Regarding treatment lines, 264 patients (60.6%) received first-line checkpoint inhibitors, 104 (23.9%) received second-line treatments, and 68 (15.6%) received third-line treatments. Comparative analyses indicated no significant statistical variances between the training and validation sets, confirming their comparability (P > 0.05).
Fig. 1.
Flow chart of the study
Selection of prognostic factors
LASSO Cox regression analysis was employed to evaluate recorded clinical measures, including clinicopathological characteristics and biomarkers, with the aim of identifying significant prognostic factors, as detailed in Table 1. Significant correlations with OS were observed for smoking history, staging, surgery, radiotherapy, liver metastases, bone metastases, other metastases, ECOG PS, types of ICIs, immunotherapy regimens, treatment lines, efficacy assessment, AE grade, WBC, NLR, and CEA, as shown in Fig. 2a-b. Significant correlations with PFS were found for ascites, lung metastases, liver metastases, bone metastases, lymph node metastases, other metastases, treatment lines, and efficacy assessment, as detailed in Fig. 2c-d. Prospective prognostic factors for OS and PFS were identified by incorporating clinical indicators from LASSO Cox regression into multivariate Cox proportional hazards regression analysis. Multivariate Cox regression analysis was conducted to identify independent predictors of OS and PFS. Patients with an NLR < 5.06, prior surgery, clinical stage II-III, fewer than two lines of therapy, SD and PR efficacy ratings, mild adverse events (grade 1/2), and no liver metastases experienced improved survival benefits. Similarly, patients with fewer than two lines of therapy, SD and PR efficacy assessments, and no initial ascites, lung, bone, lymph node, or other metastases demonstrated superior survival advantages, as shown in Table 3.
Table 1.
Baseline characteristics
| Variables | Total (n = 436) | Training set (n = 306) | Validation set (n = 130) | p |
|---|---|---|---|---|
| Age | 0.255 | |||
| ≤ 70 | 314 (72.02) | 215 (70.26) | 99 (76.15) | |
| > 70 | 122 (27.98) | 91 (29.74) | 31 (23.85) | |
| CEA(ug/L) | 0.541 | |||
| ≤ 9.8 | 150 (34.40) | 102 (33.33) | 48 (36.92) | |
| > 9.8 | 286 (65.60) | 204 (66.67) | 82 (63.08) | |
| CA125(u/ml) | 0.925 | |||
| ≤ 35 | 91 (20.87) | 63 (20.59) | 28 (21.54) | |
| > 35 | 345 (79.13) | 243 (79.41) | 102 (78.46) | |
| CA199(u/ml) | 0.999 | |||
| ≤ 37 | 204 (46.79) | 143 (46.73) | 61 (46.92) | |
| > 37 | 232 (53.21) | 163 (53.27) | 69 (53.08) | |
| WBC((× 109/L) | 0.999 | |||
| ≥ 3.5 | 405 (92.89) | 284 (92.81) | 121 (93.08) | |
| < 3.5 | 31 (7.11) | 22 (7.19) | 9 (6.92) | |
| HGB(g/L) | 0.541 | |||
| ≥ 115 | 286 (65.60) | 204 (66.67) | 82 (63.08) | |
| < 115 | 150 (34.40) | 102 (33.33) | 48 (36.92) | |
| NEU_LYM | 0.900 | |||
| ≤ 5.06 | 292 (66.97) | 206 (67.32) | 86 (66.15) | |
| > 5.06 | 144 (33.03) | 100 (32.68) | 44 (33.85) | |
| PLT_LYM | 0.999 | |||
| ≤ 205.99 | 241 (55.28) | 169 (55.23) | 72 (55.38) | |
| > 205.99 | 195 (44.72) | 137 (44.77) | 58 (44.62) | |
| Types of ICIs | 0.378 | |||
| Carrilizumab | 137 (31.42) | 96 (31.37) | 41 (31.54) | |
| Pembrolizumab | 23 (5.28) | 15 (4.90) | 8 (6.15) | |
| Sindellizumab | 89 (20.41) | 66 (21.57) | 23 (17.69) | |
| Tirellizumab | 144 (33.03) | 104 (33.99) | 40 (30.77) | |
| Other | 43 (9.86) | 25 (8.17) | 18 (13.85) | |
| Gender | 0.739 | |||
| Male | 353 (80.96) | 246 (80.39) | 107 (82.31) | |
| Female | 83 (19.04) | 60 (19.61) | 23 (17.69) | |
| Smoking history | 0.663 | |||
| No | 206 (47.25) | 142 (46.41) | 64 (49.23) | |
| Yes | 230 (52.75) | 164 (53.59) | 66 (50.77) | |
| Surgery | 0.753 | |||
| No | 328 (75.23) | 232 (75.82) | 96 (73.85) | |
| Yes | 108 (24.77) | 74 (24.18) | 34 (26.15) | |
| Radiotherapy(baseline) | 0.999 | |||
| No | 356 (81.65) | 250 (81.70) | 106 (81.54) | |
| Yes | 80 (18.35) | 56 (18.30) | 24 (18.46) | |
| Chemotherapy | 0.737 | |||
| No | 221 (50.69) | 153 (50.00) | 68 (52.31) | |
| Yes | 215 (49.31) | 153 (50.00) | 62 (47.69) | |
| Targeted therapy | 0.595 | |||
| No | 370 (84.86) | 262 (85.62) | 108 (83.08) | |
| Yes | 66 (15.14) | 44 (14.38) | 22 (16.92) | |
| Antiangiogenic therapy | 0.310 | |||
| No | 351 (80.50) | 242 (79.08) | 109 (83.85) | |
| Yes | 85 (19.50) | 64 (20.92) | 21 (16.15) | |
| Histological subtype | 0.232 | |||
| Adenocarcinoma | 209 (47.94) | 156 (50.98) | 53 (40.77) | |
| Squamous cell carcinoma | 171 (39.22) | 115 (37.58) | 56 (43.08) | |
| Small cell lung cancer | 38 (8.71) | 24 (7.84) | 14 (10.77) | |
| Other | 18 (4.13) | 11 (3.59) | 7 (5.38) | |
| Stage | 0.866 | |||
| II-III | 110 (25.23) | 76 (24.84) | 34 (26.15) | |
| IV | 326 (74.77) | 230 (75.16) | 96 (73.85) | |
| Genetic mutation | 0.413 | |||
| No | 94 (21.56) | 71 (23.20) | 23 (17.69) | |
| Yes | 58 (13.30) | 41 (13.40) | 17 (13.08) | |
| Unknow | 284 (65.14) | 194 (63.40) | 90 (69.23) | |
| Treatment Plan | 0.451 | |||
| ICIs + chemotherapy | 293 (67.20) | 202 (66.01) | 91 (70.00) | |
| ICIs + chemotherapy + AATs | 57 (13.07) | 39 (12.75) | 18 (13.85) | |
| ICIs + AATs | 54 (12.39) | 43 (14.05) | 11 (8.46) | |
| ICIs | 32 (7.34) | 22 (7.19) | 10 (7.69) | |
| Lines | 0.858 | |||
| 1 | 264 (60.55) | 186 (60.78) | 78 (60.00) | |
| 2 | 104 (23.85) | 71 (23.20) | 33 (25.38) | |
| > 2 | 68 (15.60) | 49 (16.01) | 19 (14.62) | |
| ECOG PS | 0.438 | |||
| 0 ~ 1 | 379 (86.93) | 263 (85.95) | 116 (89.23) | |
| > 2 | 57 (13.07) | 43 (14.05) | 14 (10.77) | |
| Efficacy assessment | 0.502 | |||
| PD | 116 (26.61) | 82 (26.80) | 34 (26.15) | |
| SD | 261 (59.86) | 179 (58.50) | 82 (63.08) | |
| PR | 59 (13.53) | 45 (14.71) | 14 (10.77) | |
| AE grade | 0.863 | |||
| 0 | 209 (47.94) | 147 (48.04) | 62 (47.69) | |
| 1 ~ 2 | 182 (41.74) | 126 (41.18) | 56 (43.08) | |
| 3 + | 45 (10.32) | 33 (10.78) | 12 (9.23) | |
| Ascites | 0.818 | |||
| No | 406 (93.12) | 286 (93.46) | 120 (92.31) | |
| Yes | 30 (6.88) | 20 (6.54) | 10 (7.69) | |
| Lung metastases | 0.784 | |||
| No | 344 (78.90) | 243 (79.41) | 101 (77.69) | |
| Yes | 92 (21.10) | 63 (20.59) | 29 (22.31) | |
| Liver metastases | 0.076 | |||
| No | 408 (93.58) | 291 (95.10) | 117 (90.00) | |
| Yes | 28 (6.42) | 15 (4.90) | 13 (10.00) | |
| Bone metastases | 0.236 | |||
| No | 389 (89.22) | 269 (87.91) | 120 (92.31) | |
| Yes | 47 (10.78) | 37 (12.09) | 10 (7.69) | |
| Lymph node metastases | 0.999 | |||
| No | 380 (87.16) | 267 (87.25) | 113 (86.92) | |
| Yes | 56 (12.84) | 39 (12.75) | 17 (13.08) | |
| Other metastases | 0.831 | |||
| No | 373 (85.55) | 263 (85.95) | 110 (84.62) | |
| Yes | 63 (14.45) | 43 (14.05) | 20 (15.38) | |
| Radiotherapy | 0.563 | |||
| No | 302 (69.27) | 215 (70.26) | 87 (66.92) | |
| Yes | 134 (30.73) | 91 (29.74) | 43 (33.08) |
WBC White blood cell count, CEA Carcinoembryonic antigen, NEU_LYM Neutrophil/lymphocyte ratio, PLT_LYM Platelet/lymphocyte ratio, HGB Hemoglobin, CA125 Carbohydrate antigen 125, CA199 Carbohydrate antigen 199, AATs Anti-angiogenesis therapies, ECOG PS Eastern Cooperative Oncology Group Performance Status, AE Adverse event, PD Progressive disease, PR Partial response, SD Stable disease
Fig. 2.
Factor selection and linear map development using clinicopathological characteristics and peripheral blood markers. a OS-LASSO coefficient profiles of candidate variables. c PFS-LASSO coefficient profiles of candidate variables. b and d, The LASSO study utilized ten-fold cross-validation to identify the ideal tuning parameter lamb λ. The non-zero coefficients established via ten-fold cross-validation are represented by the dashed and vertical lines. The initial vertical line corresponds to the smallest measured error. The second vertical line shows that the cross-validation error is within a standard error minimum
Table 3.
PFS multivariate cox hazards analysis
| Variables | HR | 95%CI | P | Wald | B | SE |
|---|---|---|---|---|---|---|
| Lines | ||||||
| 1 | Ref | |||||
| 2 | 0.967 | 0.694–1.347 | 0.842 | 0.039 | −0.034 | 0.169 |
| > 2 | 1.864 | 1.282–2.712 | 0.001 | 10.612 | 0.623 | 0.191 |
| Efficacy assessment | ||||||
| PD | Ref | |||||
| SD | 0.368 | 0.267–0.508 | < 0.001 | 36.838 | −0.999 | 0.165 |
| PR | 0.380 | 0.233–0.620 | < 0.001 | 14.979 | −0.968 | 0.250 |
| Ascites | ||||||
| No | Ref | |||||
| Yes | 2.200 | 1.368–3.537 | 0.001 | 10.576 | 0.788 | 0.242 |
| Lung metastases | ||||||
| No | Ref | |||||
| Yes | 1.737 | 1.262–2.390 | 0.001 | 11.484 | 0.552 | 0.163 |
| Liver metastases | ||||||
| No | Ref | |||||
| Yes | 1.260 | 0.757–2.098 | 0.374 | 0.790 | 0.231 | 0.260 |
| Bone metastases | ||||||
| No | Ref | |||||
| Yes | 1.718 | 1.099–2.688 | 0.018 | 5.631 | 0.541 | 0.228 |
| Lymph node metastases | ||||||
| No | Ref | |||||
| Yes | 1.587 | 1.089–2.313 | 0.016 | 5.786 | 0.462 | 0.192 |
| Other metastases | ||||||
| No | Ref | |||||
| Yes | 1.493 | 1.064–2.093 | 0.020 | 5.388 | 0.400 | 0.173 |
PD Progressive disease, PR Partial response, SD Stable disease
Build OS and PFS nomogram
Based on the multivariate Cox regression findings (Tables 2 and 3), OS Prediction nomogram 1 was developed using NLR, surgery, staging, treatment lines, efficacy assessment, AE grade, and liver metastases. PFS Prediction nomogram 2 was constructed with variables such as ascites, lung, bone, lymph node, other metastases, treatment lines, and efficacy assessments. Patients were randomly allocated to the training set or the validation set in a 7:3 ratio. The training set was critical for constructing the prediction model, while the validation set was used for internal validation. Figure 3 shows nomograms forecasting OS and PFS at 12, 24, and 36 months.
Table 2.
OS multivariate cox hazards analysis
| Variables | HR | 95%CI | P | Wald | B | SE |
|---|---|---|---|---|---|---|
| CEA | ||||||
| ≤ 9.8 | Ref | |||||
| > 9.8 | 0.704 | 0.487–1.018 | 0.062 | 3.486 | −0.351 | 0.188 |
| WBC | ||||||
| ≥ 3.5 | Ref | |||||
| < 3.5 | 1.326 | 0.706–2.493 | 0.38 | 0.769 | 0.282 | 0.322 |
| NLR | ||||||
| ≤ 5.06 | Ref | |||||
| > 5.06 | 1.538 | 1.045–2.263 | 0.029 | 4.772 | 0.430 | 0.197 |
| Types of ICIs | ||||||
| Camrelizumab | Ref | |||||
| Pembrolizumab | 1.756 | 0.789–3.909 | 0.168 | 1.905 | 0.563 | 0.408 |
| Sintilimab | 0.654 | 0.400–1.067 | 0.089 | 2.887 | −0.425 | 0.250 |
| Tislelizumab | 1.236 | 0.785–1.947 | 0.360 | 0.837 | 0.212 | 0.232 |
| Other | 1.344 | 0.701–2.578 | 0.373 | 0.792 | 0.296 | 0.332 |
| Smoking history | ||||||
| No | Ref | |||||
| Yes | 1.377 | 0.965–1.964 | 0.078 | 3.115 | 0.320 | 0.181 |
| Surgery | ||||||
| No | Ref | |||||
| Yes | 0.493 | 0.312–0.780 | 0.003 | 9.131 | −0.707 | 0.234 |
| Radiotherapy(baseline) | ||||||
| No | Ref | |||||
| Yes | 1.388 | 0.855–2.254 | 0.185 | 1.755 | 0.328 | 0.247 |
| Staging | ||||||
| II-III | Ref | |||||
| IV | 1.781 | 1.124–2.822 | 0.014 | 6.030 | 0.577 | 0.235 |
| Treatment Plan | ||||||
| ICIs + chemotherapy | Ref | |||||
| ICIs + chemotherapy + AATs | 1.643 | 0.976–2.767 | 0.062 | 3.486 | 0.497 | 0.266 |
| ICIs + AATs | 0.746 | 0.439–1.269 | 0.280 | 1.166 | −0.292 | 0.271 |
| ICIs | 0.477 | 0.215–1.060 | 0.069 | 3.300 | −0.740 | 0.407 |
| Lines | ||||||
| 1 | Ref | |||||
| 2 | 1.234 | 0.751–2.028 | 0.407 | 0.687 | 0.210 | 0.254 |
| > 2 | 2.634 | 1.531–4.533 | < 0.001 | 12.229 | 0.968 | 0.277 |
| ECOG PS | ||||||
| 0 ~ 1 | Ref | |||||
| > 2 | 1.404 | 0.869–2.271 | 0.166 | 1.920 | 0.340 | 0.245 |
| Efficacy assessment | ||||||
| PD | Ref | |||||
| SD | 0.568 | 0.387–0.833 | 0.004 | 8.381 | −0.566 | 0.195 |
| PR | 0.213 | 0.104–0.435 | < 0.001 | 18.008 | −1.547 | 0.365 |
| AE grade | ||||||
| 0 | Ref | |||||
| 1 ~ 2 | 0.681 | 0.468–0.990 | 0.044 | 4.042 | −0.384 | 0.191 |
| 3 + | 1.204 | 0.673–2.153 | 0.532 | 0.391 | 0.186 | 0.297 |
| Liver metastases | ||||||
| No | Ref | |||||
| Yes | 2.514 | 1.217–5.195 | 0.013 | 6.199 | 0.922 | 0.370 |
| Bone metastases | ||||||
| No | Ref | |||||
| Yes | 1.509 | 0.913–2.495 | 0.109 | 2.573 | 0.411 | 0.257 |
| Other metastases | ||||||
| No | Ref | |||||
| Yes | 0.673 | 0.403–1.122 | 0.129 | 2.309 | −0.396 | 0.261 |
WBC White blood cell count, CEA Carcinoembryonic antigen, NEU_LYM Neutrophil/lymphocyte ratio, AATs Anti-angiogenesis therapies, ECOG PS Eastern Cooperative Oncology Group Performance Status, AE Adverse event, PD Progressive disease, PR Partial response, SD Stable disease
Fig. 3.
Nomograms for predicting 12-, 24-, and 36-month OS (a) and PFS (b) in lung cancer patients. NLR: Neutrophil to lymphocyte count ratio
Validation of predictive models
In the training set, the C-index values were 0.709 (95% CI: 0.666–0.752) for OS and 0.730 (95% CI: 0.698–0.763) for PFS. In the validation set, the C-index values were 0.655 (95% CI: 0.588–0.722) for OS and 0.694 (95% CI: 0.637–0.751) for PFS. The effectiveness of the risk prediction models for OS and PFS was assessed using the area under the receiver operating characteristic (ROC) curve (AUC). The ROC curves from the training set show AUC values for OS and PFS at 12, 24, and 36 months as follows: for OS, 0.745 (95% CI: 0.680–0.809), 0.733 (95% CI: 0.664–0.801), 0.731 (95% CI: 0.641–0.821); for PFS, 0.801 (95% CI: 0.748–0.853), 0.815 (95% CI: 0.749–0.880), 0.848 (95% CI: 0.781–0.915). In the validation set, AUC values for OS at 12, 24, and 36 months were 0.676 (95% CI: 0.564–0.787), 0.720 (95% CI: 0.613–0.827), 0.707 (95% CI: 0.569–0.845); for PFS, they were 0.789 (95% CI: 0.704–0.873), 0.882 (95% CI: 0.824–0.941), 0.837 (95% CI: 0.752–0.992), as shown in Fig. 4. Calibration curves from both the training and validation sets show a high level of concordance between the predicted and actual OS values, as demonstrated by nomogram 1. Similarly, Nomogram 2's predicted PFS values closely align with the observed values, as shown in Fig. 5. These results confirm the robust performance of our prediction models.
Fig. 4.
The time-dependent ROC curves of nomogram predicting OS and PFS. a and b, 12-month, 24-month, and 36-month OS and PFS in the validation set; c and d, 12-month, 24-month, and 36-month OS and PFS in the training set. ROC, receiver operating characteristic; OS, Overall Survival; PFS, Progress Free Survival
Fig. 5.
Calibration curves used to predict OS and PFS. a and b, 12-month, 24-month, and 36-month OS and PFS in the validation set; c and (d), 12-month, 24-month, and 36-month OS and PFS in the training set. OS, Overall Survival; PFS, Progress Free Survival
Clinical application of nomogram
Decision Curve Analysis (DCA) was used to evaluate the clinical utility of the nomograms for OS and PFS, showing favorable outcomes in both the training and validation sets (Figs. 6 and 7). Using risk scores and cutoff values from the nomograms, patients were categorized into high-risk (OS ≥ 161.580, PFS ≥ 9.780) and low-risk (OS < 161.580, PFS < 9.780) groups. In the high-risk group, median overall survival (mOS) was 16.3 months and median progression-free survival (mPFS) was 5.5 months, with 95% confidence intervals of 0.428–0.579 and 0.438–0.560, respectively. For the low-risk group, the mOS was 45.9 months and mPFS was 35.1 months, with 95% confidence intervals of 0.400–0.611 and 0.427–0.624, respectively. The Kaplan–Meier curves validate that the risk prediction model effectively forecasts both OS and PFS (Fig. 8a and b). There was a significant correlation between the model's low-risk prediction and extended OS, as well as prolonged PFS, both with P-values < 0.001.
Fig. 6.
DCA of nomogram. Analysis of decision curves for 12-, 24-, and 36-month OS and PFS in the validation set. Blue line: All patients died. Green line: No patient died. Red line: nomogram model. DCA: Decision curve analysis
Fig. 7.
DCA of nomogram. Analysis of decision curves for 12-, 24-, and 36-month OS and PFS in the training set. Blue line: All patients died. Green line: No patient died. Red line: nomogram model. DCA: Decision curve analysis
Fig. 8.
Kaplan–Meier survival analysis. Patients were divided into high and low risk groups based on risk scores and optimal cutoff values derived from nomograms for OS and PFS. a Kaplan–Meier curve of OS. b Kaplan–Meier curve of PFS
Discuss
Accurately predicting tumor prognosis following treatment is essential for optimizing patient management and individualizing therapy. This study identified independent risk factors for OS and PFS at 12, 24, and 36 months in lung cancer patients using our training set. We developed and validated two novel nomograms to predict these outcomes at the corresponding time points. Nomogram 1 incorporates several factors, including the NLR, surgical history, clinical staging, treatment lines, efficacy assessments, immune-related adverse events (irAEs), and the presence of liver metastases. Nomogram 2 includes treatment lines, efficacy assessments, ascites, and the presence of lung, bone, lymph node, and other metastatic sites. Both nomograms demonstrated superior differentiability, achieving C-index values for OS and PFS of 0.709 and 0.730 in the training cohort, and 0.655 and 0.694 in the validation cohort. Subsequent calibration curves and decision curve analyses confirmed the predictive accuracy and clinical utility of both nomograms, with predictions closely aligning with observed outcomes.
We identified eight key predictors for lung cancer prognosis: NLR, surgical history, cancer stage, ascites, baseline metastatic sites, treatment lines, efficacy assessments, and immune-associated adverse events. The relationship between inflammation and cancer is multifaceted; NLR, as a blood-based inflammation marker, has shown effectiveness in prognostic evaluation across various cancer types [25]. Specifically, studies indicate that NLR serves as a reliable prognostic biomarker for lung cancer patients undergoing immunotherapy [26]. Our findings are consistent with a large review by Chen et al. [27], which revealed that patients with non-small cell lung cancer (NSCLC) exhibiting reduced NLR levels at weeks 6 and 12 following immunotherapy had longer PFS and OS compared to those with elevated NLR (P < 0.001). This emphasizes the viability of using baseline NLR as a biomarker for lung cancer prognosis, reinforcing the effectiveness of our risk prediction model. Importantly, NLR data can be readily obtained from routine blood tests, offering a cost-effective and convenient prognostic tool, although it is essential to account for potential confounding factors related to comorbidities and treatments [26, 28, 29].
Common metastatic sites for lung cancer, including bones, brain, lungs, and liver, significantly influence patient prognosis [30, 31]. Patients presenting with metastases at baseline experienced poorer outcomes. A systematic meta-analysis revealed that patients with brain, liver, or bone metastases exhibited reduced OS when treated with PD-1 inhibitors. Similarly, brain and liver metastases were associated with higher progression rates under PD-1 inhibitor treatment. Patients with multiple metastatic sites showed worse PFS than those with a single site, with no significant difference in OS observed. Multiple organ metastases are frequently correlated with poor physical condition and cancer-related cachexia. The metastatic site independently predicts survival outcomes in advanced NSCLC patients treated with PD-1 inhibitors [32]. Research consistently shows that patients with metastases at baseline experience poorer outcomes. A systematic meta-analysis corroborated that patients with brain, liver, or bone metastases exhibit reduced overall survival when treated with PD-1 inhibitors. Furthermore, multiple metastatic sites correlate with worse progression-free survival compared to a single site, though no significant differences in OS have been observed. Malignant pleural effusion was found to be an independent prognostic risk factor for PFS in our study; patients without pre-existing fluid accumulation fared better than those with it. Previous studies have identified malignant pleural effusion as a negative predictor of PFS due in part to its association with tumor-induced immunosuppression [30, 33]. Kawachi et al. proposed that vascular endothelial growth factor (VEGF) mediates pleural effusion's adverse effects by increasing vascular permeability, which can impair response to treatment. Elevated VEGF levels in patients with malignant pleural effusion have been associated with poor responses to pembrolizumab monotherapy, indicating a critical area for further investigation [30, 34, 35].
We observed that treatment lines and clinical stage were independent predictors for both PFS and OS. Notably, patients receiving immune checkpoint inhibitors (ICIs) as first-line therapy showed significant benefits. Comparative analyses revealed that pembrolizumab demonstrated a 5-year survival rate of 32% as a first-line treatment in advanced NSCLC patients with a PD-L1 expression of ≥ 50% [36]. Randomized trials have also indicated the advantage of ICIs over standard chemotherapy in second-line settings [37]. Interestingly, our findings indicated that patients treated with ICIs in the second-line setting experienced improved PFS compared to those receiving first-line treatment. This phenomenon may be attributed to the initial use of chemotherapy, which can modify the tumor microenvironment favorably, enhancing subsequent immunotherapy responses. Prior studies have shown that preconditioning with chemotherapy alters the cellular composition of the tumor microenvironment, enhancing the efficacy of immune checkpoint inhibitors [38]. Analyzing drug resistance mechanisms and optimizing immunotherapy [39] could improve outcomes in second-line treatments. Clinicians may choose a more appropriate immunotherapy or combination regimen for second-line treatment, based on patient response to first-line therapy and resistance mechanisms. Individualized therapeutic optimization can enhance the efficacy and PFS of second-line immunotherapies.
With the increasing availability and efficacy of ICI therapies, concerns regarding treatment-specific toxicities, particularly immune-related adverse events, are growing [40]. These events are often interpreted as indicators of enhanced immune responses and improved anti-tumor efficacy [41]. Consequently, numerous studies have explored whether these events could serve as indicators of ICI efficacy in cancer treatment. Emerging data suggest that irAEs may correlate with increased efficacy of ICIs in advanced malignancies such as NSCLC, indicating that patients who discontinue therapy due to irAEs might still experience therapeutic benefits [42–44]. Interestingly, our study found that patients experiencing grade 1–2 irAEs had longer OS compared to those without such events or those with severe irAEs. This observation aligns with findings from large clinical trials indicating that lower-grade irAEs are associated with better treatment outcomes, suggesting that irAEs may reflect effective immune activation and significant treatment responses [45–48].
The nomograms provide a practical tool for stratifying ICI-treated patients into low- and high-risk groups, guiding decisions on therapy escalation (e.g., combination regimens) or palliative care. For example, high-risk patients (OS < 16 months) may benefit from early integration of anti-angiogenic agents or clinical trials targeting TME immunosuppression. However, challenges include biomarker accessibility in resource-limited settings and the need for real-time NLR/CEA monitoring. While the nomogram's calculations are straightforward (summing weighted scores), integration into electronic health records requires user-friendly interfaces to minimize computational burden.
As a real-world study, our findings are subject to inherent limitations. Conducted retrospectively at a single institution, the results may have limited generalizability. The small sample size and the non-randomized nature of participant selection may introduce biases. To validate these findings, large-scale, multicenter clinical trials involving diverse patient populations are warranted. Additionally, the relatively short follow-up period necessitates ongoing monitoring and extended evaluations. Our study's approach, which assessed lung cancer patients prior to ICI treatment while considering clinical, pathological, and laboratory characteristics, generates valuable data that could assist in predicting lung cancer prognosis after immunotherapy.
Conclusion
In conclusion, this study successfully identifies and validates critical prognostic factors for lung cancer treatment outcomes, culminating in the development of two effective nomograms. These tools hold significant promise for enhancing individualized treatment strategies and improving patient prognostication in clinical settings. Further research and validation in larger cohorts are required to ensure their broader applicability and optimization in the management of lung cancer.
Acknowledgements
We would like to express our sincere gratitude to all individuals and organizations that contributed to this research. We also acknowledge the financial support provided by Jiangsu Provincial Medical Key Discipline(Laboratory) (N0. ZDXYS202208); Postdoctoral Fellowship Fund of Jinling Hospital Affiliated to Nanjing University Medical School (No.96727); Project supported by Nanjing Medical Science and technology development (No. ZKX22026). These funding made it possible for us to conduct this research and complete this manuscript.
Abbreviations
- PFS
Progression-free survival
- OS
Overall survival
- C-index
Concordance index
- ROC
Receiver operating characteristic
- DCA
Decision curve analyses
- ECOG PS
Eastern Cooperative Oncology Group Performance Status
- WBC
White blood cell count
- CEA
Carcinoembryonic antigen
- NLR
Neutrophil-to-lymphocyte ratio
- PLR
Platelet-to-lymphocyte ratio
- HGB
Hemoglobin
- CA125
Carbohydrate antigen 125
- CA199
Carbohydrate antigen 199
- AATs
Anti-angiogenesis therapies
- ICIs
Immune checkpoint inhibitors
- PD-1
Anti-programmed death-1
- PD-L1
Anti-programmed death ligand-1
- CTLA4
Anti-cytotoxic T lymphocyte antigen 4
- PD
Progressive disease
- PR
Partial response
- SD
Stable disease
Authors’ contributions
Conceptualization: CXY, CC and YM. Data curation: CC. Formal analysis: QJ. Funding acquisition: CC, QJ and YM. Investigation: LTW, CA and HN. Methodology: OYT, ZF and YY. Software: OYT. Validation: CC. Visualization: ZF. Writing - original draft: OYT. Writing - review & editing: CXY, CC and YM. All authors reviewed the manuscript.
Funding
The study was supported by Jiangsu Provincial Medical Key Discipline(Laboratory) (N0. ZDXYS202208); Postdoctoral Fellowship Fund of Jinling Hospital Affiliated to Nanjing University Medical School (No.96727); Project supported by Nanjing Medical Science and technology development (No. ZKX22026).
Data availability
The data that support the findings of this study are not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
All procedures performed in this study involving human participants were in accordance with the guidelines of the Ethics Committee of Nanjing University School of Medicine affiliated Jinling Hospital (No. DZQH-KYLL-23-06) and with the 1964 Helsinki Declaration. Patients treated in our hospital will be informed of the potential research use of personal information. Only patients who agreed with and signed the consent were included in our study.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Ting Ouyang, Fei Zhang and Yan Yang contributed equally to this work.
Contributor Information
Xiaoyuan Chu, Email: chuxiaoyuan000@163.com.
Chao Chen, Email: njjloncologycc@163.com.
Mi Yang, Email: yangmi@nju.edu.cn.
References
- 1.Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Glo bal cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024;74(3):229–63. [DOI] [PubMed] [Google Scholar]
- 2.Zheng RS, Chen R, Han BF, Wang SM, Li L, Sun KX, et al. Cancer incidence and mortality in China, 2022. Zhonghua Zhong Liu Za Zhi. 2024;46(3):221–31. [DOI] [PubMed] [Google Scholar]
- 3.de Castro G Jr, Kudaba I, Wu YL, Lopes G, Kowalski DM, Turna HZ, et al. Five-year outcomes with pembrolizumab versus chemotherapy as first-line therapy in patients with non-small-cell lung cancer and programmed death ligand-1 tumor proportion score ≥ 1% in the KEYNOTE-042 study. J Clin Oncol. 2023;41(11):1986–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Rudin CM, Awad MM, Navarro A, Gottfried M, Peters S, Csőszi T, et al. Pembrolizumab or placebo plus etoposide and platinum as first-line therapy for extensive-stage small-cell lung cancer: randomized, double-blind, phase III KEYNOTE-604 study. J Clin Oncol. 2020;38(21):2369–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Salas-Benito D, Pérez-Gracia JL, Ponz-Sarvisé M, Rodriguez-Ruiz ME, Martínez-Forero I, Castañón E, et al. Paradigms on immunotherapy combinations with chemotherapy. Cancer Discov. 2021;11(6):1353–67. [DOI] [PubMed] [Google Scholar]
- 6.Jin R, Liu C, Zheng S, Wang X, Feng X, Li H, et al. Molecular heterogeneity of anti-PD-1/PD-L1 immunotherapy efficacy is correlated with tumor immune microenvironment in East Asian patients with non-small cell lung cancer. Cancer Biol Med. 2020;17(3):768–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Sun L, Bleiberg B, Hwang WT, Marmarelis ME, Langer CJ, Singh A, et al. Association between duration of immunotherapy and overall survival in advanced non-small cell lung cancer. JAMA Oncol. 2023;9(8):1075–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Gainor JF, Rizvi H, Jimenez Aguilar E, Skoulidis F, Yeap BY, Naidoo J, et al. Clinical activity of programmed cell death 1 (PD-1) blockade in never, light, and heavy smokers with non-small-cell lung cancer and PD-L1 expression ≥50. Ann Oncol. 2020;31(3):404–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kim ES, Velcheti V, Mekhail T, Yun C, Shagan SM, Hu S, et al. Blood-based tumor mutational burden as a biomarker for atezolizumab in non-small cell lung cancer: the phase 2 B-F1RST trial. Nat Med. 2022;28(5):939–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tostes K, Siqueira AP, Reis RM, Leal LF, Arantes L. Biomarkers for immune checkpoint inhibitor response in NSCLC: current developments and applicability. Int J Mol Sci. 2023. 10.3390/ijms241511887. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Negrao MV, Skoulidis F, Montesion M, Schulze K, Bara I, Shen V, et al. Oncogene-specific differences in tumor mutational burden, PD-L1 expression, and outcomes from immunotherapy in non-small cell lung cancer. J Immunother Cancer. 2021. 10.1136/jitc-2021-002891. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Johnson ML, Cho BC, Luft A, Alatorre-Alexander J, Geater SL, Laktionov K, et al. Durvalumab with or without tremelimumab in combination with chemotherapy as first-line therapy for metastatic non-small-cell lung cancer: the phase III POSEIDON study. J Clin Oncol. 2023;41(6):1213–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Liu M, Wei F, Wang J, Yu W, Shen M, Liu T, et al. Myeloid-derived suppressor cells regulate the immunosuppressive functions of PD-1(-)PD-L1(+) Bregs through PD-L1/PI3K/AKT/NF-κB axis in breast cancer. Cell Death Dis. 2021;12(5):465. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Huang J, Cai X, Ng C, Luo Y, Chen Q, Wang Z, et al. A phase 2 study of tislelizumab as neoadjuvant treatment of cisplatin-ineligible high-risk upper tract urothelial carcinoma. J Urol. 2025;213(6):739–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zhang M, Cui J, Chen H, Cheng Y, Chen Q, Zong F, et al. Increased SOAT2 expression in aged regulatory T cells is associated with altered cholesterol metabolism and reduced anti-tumor immunity. Nat Commun. 2025;16(1):630. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Liu X, Xi X, Xu S, Chu H, Hu P, Li D, et al. Targeting T cell exhaustion: emerging strategies in non-small cell lung cancer. Front Immunol. 2024;15:1507501. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Xu S, Lu Z. Exploring FNDC4 as a biomarker for prognosis and immunotherapy response in lung adenocarcinoma. Asian journal of surgery 2024. [DOI] [PubMed]
- 18.Teng F, Meng X, Kong L, Yu J. Progress and challenges of predictive biomarkers of anti PD-1/PD-L1 immunotherapy: a systematic review. Cancer Lett. 2018;414:166–73. [DOI] [PubMed] [Google Scholar]
- 19.Herbst RS, Soria JC, Kowanetz M, Fine GD, Hamid O, Gordon MS, et al. Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients. Nature. 2014;515(7528):563–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Champiat S, Ferté C, Lebel-Binay S, Eggermont A, Soria JC. Exomics and immunogenics: bridging mutational load and immune checkpoints efficacy. Oncoimmunology. 2014;3(1):e27817. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Taube JM, Klein A, Brahmer JR, Xu H, Pan X, Kim JH, et al. Association of PD-1, PD-1 ligands, and other features of the tumor immune microenvironment with response to anti-PD-1 therapy. Clin Cancer Res. 2014;20(19):5064–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Balachandran VP, Gonen M, Smith JJ, DeMatteo RP. Nomograms in oncology: more than meets the eye. Lancet Oncol. 2015;16(4):e173-180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer. 2009;45(2):228–47. [DOI] [PubMed] [Google Scholar]
- 24.Qiu C, Wang W, Xu S, Li Y, Zhu J, Zhang Y, et al. Construction and validation of a hypoxia-related gene signature to predict the prognosis of breast cancer. BMC Cancer. 2024;24(1):402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hainaut P, Plymoth A. Targeting the hallmarks of cancer: towards a rational approach to next-generation cancer therapy. Curr Opin Oncol. 2013;25(1):50–1. [DOI] [PubMed] [Google Scholar]
- 26.Balkwill F, Mantovani A. Inflammation and cancer: back to Virchow? Lancet. 2001;357(9255):539–45. [DOI] [PubMed] [Google Scholar]
- 27.Zhang N, Jiang J, Tang S, Sun G. Predictive value of neutrophil-lymphocyte ratio and platelet-lymphocyte ratio in non-small cell lung cancer patients treated with immune checkpoint inhibitors: a meta-analysis. Int Immunopharmacol. 2020;85:106677. [DOI] [PubMed] [Google Scholar]
- 28.Xie X, Liu J, Yang H, Chen H, Zhou S, Lin H, et al. Prognostic value of baseline neutrophil-to-lymphocyte ratio in outcome of immune checkpoint inhibitors. Cancer Invest. 2019;37(6):265–74. [DOI] [PubMed] [Google Scholar]
- 29.Jin J, Yang L, Liu D, Li W. Association of the neutrophil to lymphocyte ratio and clinical outcomes in patients with lung cancer receiving immunotherapy: a meta-analysis. BMJ Open. 2020;10(6): e035031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Kawachi H, Tamiya M, Tamiya A, Ishii S, Hirano K, Matsumoto H, et al. Association between metastatic sites and first-line pembrolizumab treatment outcome for advanced non-small cell lung cancer with high PD-L1 expression: a retrospective multicenter cohort study. Invest New Drugs. 2020;38(1):211–8. [DOI] [PubMed] [Google Scholar]
- 31.Adachi Y, Tamiya A, Taniguchi Y, Enomoto T, Azuma K, Kouno S, et al. Predictive factors for progression-free survival in non-small cell lung cancer patients receiving nivolumab based on performance status. Cancer Med. 2020;9(4):1383–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Grove CS, Lee YC. Vascular endothelial growth factor: the key mediator in pleural effusion formation. Curr Opin Pulm Med. 2002;8(4):294–301. [DOI] [PubMed] [Google Scholar]
- 33.Hegde PS, Wallin JJ, Mancao C. Predictive markers of anti-VEGF and emerging role of angiogenesis inhibitors as immunotherapeutics. Semin Cancer Biol. 2018;52(Pt 2):117–24. [DOI] [PubMed] [Google Scholar]
- 34.Khan KA, Kerbel RS. Improving immunotherapy outcomes with anti-angiogenic treatments and vice versa. Nat Rev Clin Oncol. 2018;15(5):310–24. [DOI] [PubMed] [Google Scholar]
- 35.Kang DH, Chung C, Kim JO, Jung SS, Park HS, Park DI, et al. Pleural or pericardial metastasis: a significant factor affecting efficacy and adverse events in lung cancer patients treated with PD-1/PD-L1 inhibitors. Thorac Cancer. 2018;9(11):1500–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Reck M, Rodríguez-Abreu D, Robinson AG, Hui R, Csőszi T, Fülöp A, et al. Five-year outcomes with pembrolizumab versus chemotherapy for metastatic non-small-cell lung cancer with PD-L1 tumor proportion score ≥ 50. J Clin Oncol. 2021;39(21):2339–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Borghaei H, Gettinger S, Vokes EE, Chow LQM, Burgio MA, de Castro CJ, et al. Five-year outcomes from the randomized, phase III trials CheckMate 017 and 057: nivolumab versus docetaxel in previously treated non-small-cell lung cancer. J Clin Oncol. 2021;39(7):723–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.An M, Mehta A, Min BH, Heo YJ, Wright SJ, Parikh M, et al. Early immune remodeling steers clinical response to first-line chemoimmunotherapy in advanced gastric cancer. Cancer Discov. 2024;14(5):766–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Nagasaki J, Ishino T, Togashi Y. Mechanisms of resistance to immune checkpoint inhibitors. Cancer Sci. 2022;113(10):3303–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Lee DJ, Lee HJ Jr, Farmer JR, Reynolds KL. Mechanisms driving immune-related adverse events in cancer patients treated with immune checkpoint inhibitors. Curr Cardiol Rep. 2021;23(8):98. [DOI] [PubMed] [Google Scholar]
- 41.Ghiringhelli F, Apetoh L. The interplay between the immune system and chemotherapy: emerging methods for optimizing therapy. Expert Rev Clin Immunol. 2014;10(1):19–30. [DOI] [PubMed] [Google Scholar]
- 42.Akamatsu H, Murakami E, Oyanagi J, Shibaki R, Kaki T, Takase E, et al. Immune-related adverse events by immune checkpoint inhibitors significantly predict durable efficacy even in responders with advanced non-small cell lung cancer. Oncologist. 2020;25(4):e679–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Hussaini S, Chehade R, Boldt RG, Raphael J, Blanchette P, Maleki Vareki S, et al. Association between immune-related side effects and efficacy and benefit of immune checkpoint inhibitors - a systematic review and meta-analysis. Cancer Treat Rev. 2021;92:102134. [DOI] [PubMed] [Google Scholar]
- 44.Remon J, Reguart N, Auclin E, Besse B. Immune-related adverse events and outcomes in patients with advanced non-small cell lung cancer: a predictive marker of efficacy? J Thorac Oncol. 2019;14(6):963–7. [DOI] [PubMed] [Google Scholar]
- 45.Socinski MA, Jotte RM, Cappuzzo F, Nishio M, Mok TSK, Reck M, et al. Association of immune-related adverse events with efficacy of atezolizumab in patients with non-small cell lung cancer: pooled analyses of the phase 3 IMpower130, IMpower132, and IMpower150 randomized clinical trials. JAMA Oncol. 2023;9(4):527–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Teraoka S, Fujimoto D, Morimoto T, Kawachi H, Ito M, Sato Y, et al. Early immune-related adverse events and association with outcome in advanced non-small cell lung cancer patients treated with Nivolumab: a prospective cohort study. J Thorac Oncol. 2017;12(12):1798–805. [DOI] [PubMed] [Google Scholar]
- 47.Fehrenbacher L, Spira A, Ballinger M, Kowanetz M, Vansteenkiste J, Mazieres J, et al. Atezolizumab versus docetaxel for patients with previously treated non-small-cell lung cancer (POPLAR): a multicentre, open-label, phase 2 randomised controlled trial. Lancet. 2016;387(10030):1837–46. [DOI] [PubMed] [Google Scholar]
- 48.Noguchi S, Suminaga K, Kaki T, Kawachi H, Fukao A, Terashita S, et al. Correlation of immune-related adverse events and effects of pembrolizumab monotherapy in patients with non-small cell lung cancer. Lung Cancer (Auckl). 2020;11:53–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request.