Survival benefits in non-small cell lung cancer during the immune checkpoint inhibitor era: integrating lymph node burden for prognostic precision

Introduction

Lung cancer remains a major global health challenge, with its high incidence and mortality rates posing a significant burden on healthcare systems worldwide. According to Global Cancer Observatory (GLOBOCAN) 2022, lung cancer continues to be the most commonly diagnosed cancer globally, with approximately 2.5 million new cases, representing 12.4% of all cancers, and accounting for 18.7% of cancer-related deaths, underscoring its substantial global burden (1). Among the different subtypes of lung cancer, non-small-cell lung cancer (NSCLC) is the most prevalent, comprising 80% to 85% of all lung cancers (2) and is responsible for the majority of cancer-related deaths worldwide.

The advent of the immune checkpoint inhibitors (ICIs) has fundamentally transformed cancer treatment paradigms, particularly in NSCLC. In the Phase III trial CheckMate 017, nivolumab demonstrated a median overall survival (OS) of 9.2 months compared to 6.0 months with docetaxel, reducing the risk of death by 41% (3). Based on these findings, the U.S. Food and Drug Administration (FDA) approved nivolumab in March 2015 for the treatment of advanced squamous NSCLC following platinum-based chemotherapy failure. Since then, nivolumab has become the first approved immunotherapy for squamous cell lung cancer (4), marking the beginning of an era characterized by the widespread introduction of ICIs in NSCLC treatment. Subsequently, multiple ICIs and targeted therapies—including representative agents such as nivolumab, pembrolizumab, osimertinib, and alectinib (5)—have been approved and widely adopted in clinical practice, collectively expanding the therapeutic options for advanced NSCLC in the post-ICI era.

Several studies have highlighted the survival benefits with ICI treatment in metastatic NSCLC (6,7). For instance, the RATIONALE 304 trial demonstrated that tislelizumab plus chemotherapy provided significantly better survival outcomes compared to platinum-based chemotherapy in the first-line treatment of metastatic NSCLC [median progression-free survival (PFS): 9.7 vs. 7.6 months] (6). Similarly, the KEYNOTE-407 trial confirmed the efficacy of pembrolizumab combined with platinum-based chemotherapy as a first-line treatment for metastatic NSCLC (7). However, the strict inclusion and exclusion criteria of randomized controlled trials (RCTs) often limit their generalizability to broader patient populations. Additionally, the highly controlled environments in which RCTs are conducted may fail to fully account for the complexities and variability inherent in real-world clinical practice. Therefore, analysis based on real-world data is particularly important, as it not only supplements existing findings but also provides more practical guidance for treatment decisions in routine clinical care (8). The mechanisms underlying the success of ICIs are multifaceted, with one notable mechanism lying in its impact on lymphatic structures (9-11). Several studies have highlighted the close link between the effectiveness of immunotherapy and its ability to induce nodal downstaging and clear metastatic lymph nodes (12,13). However, prior to 2024, the lymph node N classification of the Tumor-Node-Metastasis (TNM) staging system was based solely on the anatomical location of lymph node metastasis without accounting for the extent of metastatic burden (14). In the ICI-introduction era, these areas warrant further investigation to better understand and optimize treatment strategies. To more comprehensively and stably assess the true extent of lymph node involvement, the concept of ‘lymph node metastatic burden’ has emerged and aims to more precisely reflect the degree and density of tumor dissemination through more quantitative metrics, such as the number of positive lymph nodes (NPLN), lymph node ratio (LNR), and log odds of positive lymph nodes (LODDS). These indicators, particularly LNR and LODDS, have shown potential as superior prognostic assessment tools in various malignancies (15,16). Besides, the ninth edition of the TNM staging system, updated in 2024, introduced a new distinction within N2, categorizing it into single-station (N2a) and multi-station (N2b) involvement, demonstrating significant prognostic differences and independent predictive value (14,17), which highlights the critical role of lymph node burden in precision medicine. In the era of precision medicine for lung cancer (15), the evolving understanding of metastatic burden in lymph nodes highlights the need for deeper insights into its role in treatment and prognosis.

Therefore, the study utilizes large-scale real-world data from the Surveillance, Epidemiology, and End Results (SEER) database to evaluate the impact of ICI introduction on survival outcomes in patients with advanced NSCLC and explores the mediating role of key prognostic indicators, including the NPLN, LNR, and LODDS, in the survival benefits observed following ICI introduction. Furthermore, we assess the prognostic value of these indicators in early-stage NSCLC to optimize risk stratification and refine the TNM staging system. We present this article in accordance with the STROBE reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-447/rc).

Methods

Data source

The SEER database (https://seer.cancer.gov/), developed by the U.S. National Cancer Institute, is a population-based comprehensive resource focused on collecting and organizing epidemiological and outcome data related to cancer. The data used in this study were extracted and analyzed using the SEER*Stat software (version 8.4.4.0) provided by the National Cancer Institute, encompassing cancer cases from 17 registries and representing approximately 30% of the U.S. population. Notably, although the SEER database is updated annually, data were only available and validated through 2021 at the time of data extraction (November 2023 release). The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Patient selection

Our study identified NSCLC cases through the International Classification of Diseases for Oncology, 3rd Edition (ICD-O-3) morphology codes, including the following histologic codes: 8010, 8012, 8013, 8020, 8046, 8050–8052, 8070–8078, 8140, 8141, 8143, 8147, 8250–8255, 8260, 8310, 8430, 8480, 8481, 8490, 8560, and 8570–8575. Staging information was provided by the American Joint Committee on Cancer (AJCC). Lung tumor cases with site codes C34.0–C34.9 were first extracted from the SEER database. Advanced NSCLC patients diagnosed microscopically between 2010 and 2021, aged ≥18 years, and with follow-up records that included clearly documented date of diagnosis, survival duration in months, and vital status were selected for inclusion in the study. The exclusion criteria were: (I) unknown survival time or survival less than one month; (II) missing data on race, marital status, gender, annual household income, geographical area, grade, T stage, or N stage; (III) absence of information on examined lymph nodes or positive lymph nodes; (IV) conflicting data, including but not limited to cases with zero examined lymph nodes but nonzero positive nodes (NPLN >0), discrepancies between N stage and NPLN count (e.g., N0 with NPLN >0), and inconsistencies between survival time and vital status (e.g., recorded death with zero months of follow-up), etc.

Additionally, for the validation cohort, we selected early-stage non-metastatic N1 and N2 NSCLC patients with >0 positive lymph nodes. The inclusion and exclusion criteria for the validation cohort were consistent with those described above (Figure S1).

Lymph node metastatic burden

In our study, NPLN, LODDS, and LNR were defined as indicators of lymph node metastatic burden. LNR was calculated as the ratio of NPLN to the total number of dissected lymph nodes (NDLN), through the formula: LNR = NPLN / NDLN. LODDS was calculated by the formula: LODDS = log((NPLN + 0.50) / (NDLN − NPLN + 0.50)). The constant 0.50 was added to both the numerator and denominator to avoid infinite values (16). The metrics provide a multidimensional quantification of lymph node burden, capturing different aspects of lymph node involvement to enhance the assessment of metastatic features and prognostic evaluation.

Study outcomes

The SEER cause of death classification system was used to accurately record and confirm the cause of death for each patient. The primary outcomes of this study were OS and cancer-specific survival (CSS). OS was calculated as the time interval beginning at the date of diagnosis and ending at either the date of death from any cause or the date of last available follow-up. CSS was defined as the time span from diagnosis to death specifically attributed to NSCLC. These endpoints provide critical metrics for evaluating patient prognosis and treatment efficacy.

Statistical analysis

Considering the 2015 FDA approval of nivolumab as the first ICI for NSCLC, which marked the beginning of a new therapeutic era, patients were divided into two cohorts. Those diagnosed between 2010 and 2014 were assigned to the pre-ICI era cohort, and those diagnosed between 2015 and 2021 were assigned to the ICI-introduction era cohort, reflecting the distinct periods before and after the widespread clinical availability and integration of ICIs into NSCLC treatment. A complete list of FDA-approved therapies during the study period has been summarized in Table S1 to provide context for this era-based classification. Continuous variables were expressed as mean ± standard deviation, while categorical variables were expressed as percentages. Group differences were assessed using Student’s t-test, Mann-Whitney U test, or chi-squared test. Univariable and multivariable Cox proportional hazard models [reporting hazard ratios (HR) with 95% confidence intervals (CI)] were used to evaluate the association between immunotherapy and all-cause mortality as well as cause-specific mortality. Cumulative survival rates were estimated using the Kaplan-Meier method and compared between groups with the log-rank test. Conditional survival (CS) analysis assessed time-specific survival rates, defined as the probability of surviving an additional years given survival for years, using the formula: CS(y|x) = S(x + y) / S(x), where S(x + y) and S(x) represent survival probabilities at specific time points (18).

Due to the retrospective and observational nature of the SEER dataset, confounding factors were addressed by implementing 1:1 propensity score matching (PSM) with a caliper width of 0.008 for variables found to be significant in the Cox regression analysis of OS. This approach aimed to balance patient characteristics across cohorts (Figure S2). To evaluate the effectiveness of the PSM, we calculated the c-statistic (area under the receiver operating characteristic curve), which was 0.593, indicating limited discriminative ability in distinguishing between treatment groups. Despite this, the standardized mean differences (SMDs) for all covariates after matching were below 0.1, suggesting adequate covariate balance was achieved.

Stratified analyses were conducted to explore the association and interaction between immunotherapy and mortality across different subpopulations. Causal mediation analysis was performed to examine the mediating roles of NPLN, LNR, and LODDS in the relationship between immunotherapy and mortality.

To interpret the contribution of each variable to mortality risk, the SHapley Additive exPlanations (SHAP) framework was applied. A gradient boosting model (e.g., XGBoost or LightGBM) was trained using the same covariates as those included in the Cox regression model to predict survival risk, with the endpoint defined as a binary outcome at a specified follow-up time or a transformed survival outcome. SHAP values were then computed to estimate the marginal effect of each variable on the model’s output. Positive SHAP values indicated increased predicted mortality risk, whereas negative values reflected a protective association.

Restricted cubic splines (RCS) were used to visualize the dose-response relationships between LNR or LODDS and mortality risk, and the maximal selected rank method was applied to identify inflection points for key variables and survival outcomes (19).

All statistical analyses were performed using R software (Version 4.2.1, The R Foundation; http://www.R-project.org) and EmpowerStats software (Version 5.0, X&Y Solutions, Inc., Boston, MA, USA; http://www.empowerstats.com). A P value of less than 0.05 was considered statistically significant.

Results

Baseline characteristics of the study population

A total of 5,587 patients with advanced NSCLC from the SEER database were included in this study (Figure S1), of whom 48.20% were female and 80.62% were White, reflecting recent epidemiological trends in NSCLC, particularly the rising incidence of adenocarcinoma among non-smoking women (20,21). Among them, 2,948 patients were in the pre-ICI era cohort, while 2,639 patients were in the ICI-introduction era cohort. Patients in the ICI-introduction era cohort had higher annual incomes, a greater likelihood of living in metropolitan counties, and higher divorce rates (Table 1).

However, significant differences were observed in baseline characteristics between the ICI-introduction era and pre-ICI era cohorts, including age distribution, gender, race, T stage, and N stage. These baseline differences were likely driven by real-world changes in diagnostic practices, staging criteria, and data quality improvements. Therefore, one-to-one PSM was employed. In the post-PSM dataset of 4,496 cases, no significant differences were observed between the two groups in baseline demographic data, staging, or treatment information (Table 1).

Survival outcomes

In advanced NSCLC patients with a median survival time of 12 months, 4,459 deaths were recorded. The relationship between ICI introduction and the risk of death in advanced NSCLC patients was determined using Cox regression analysis. In the unadjusted model, compared to the pre-ICI era cohort, the introduction of ICIs was associated with a 19% reduction in all-cause mortality (HR =0.81, 95% CI: 0.76–0.86, Table 2) and a 23% reduction in cancer-specific mortality (HR =0.77, 95% CI: 0.72–0.83, Table 2). After multivariate adjustment, compared to the pre-ICI era cohort, the introduction of ICIs remained associated with a 20% reduction in the risk of all-cause mortality (HR =0.80, 95% CI: 0.75–0.85, Table 2) and a 24% reduction in the risk of cancer-specific mortality (HR =0.76, 95% CI: 0.71–0.81, Table 2).

Kaplan-Meier curves were also used to compare survival outcomes between the ICI-introduction era and pre-ICI era cohorts in advanced NSCLC patients (Figure 1). Compared with patients diagnosed in the pre-ICI era, those diagnosed during the ICI-introduction era showed significantly improved OS and CSS survival times. Similar survival trends were observed in the post-PSM dataset (log-rank test P<0.001).

Figure 1 Comparison of Kaplan-Meier survival curves for OS and CSS between the pre-ICI era cohort and the ICI-introduction era cohort. (A) Kaplan-Meier curve for OS before PSM; (B) Kaplan-Meier curve for OS after PSM; (C) Kaplan-Meier curve for CSS before PSM; (D) Kaplan-Meier curve for CSS after PSM. CSS, cancer-specific survival; ICI, immune checkpoint inhibitor; OS, overall survival; PSM, propensity score matching.

CS analysis

To further explore the survival benefits observed in the ICI-introduction era cohort, we conducted CS analysis. The analysis demonstrated that, in both the pre-ICI era cohort and the ICI-introduction era cohort, the conditional probabilities of OS and CSS increased with each additional year of survival for advanced NSCLC patients.

In the pre-ICI era cohort, the 5-year conditional OS and CSS rates were 17%/22% at diagnosis, and progressively improved to 33%/39%, 50%/56%, 69%/74%, and 84%/86% after surviving 1, 2, 3, and 4 years, respectively. In the ICI-introduction era cohort, the corresponding 5-year conditional OS and CSS survival rates were 23%/30% at diagnosis, and 39%/48%, 56%/65%, 71%/77%, and 86%/89% after surviving 1, 2, 3, and 4 years, respectively, consistently higher than those in the pre-ICI era cohort (Figure 2).

Figure 2 Conditional survival analysis for the pre-ICI era cohort and the ICI-introduction era cohort. (A) OS in the pre-ICI era cohort; (B) CSS in the pre-ICI era cohort; (C) OS in the ICI-introduction era cohort; (D) CSS in the ICI-introduction era cohort. CSS, cancer-specific survival; ICI, immune checkpoint inhibitor; OS, overall survival.

Subgroup analysis

Subgroup analyses assessed the impact of ICI introduction on the survival of patients with different stratifications of advanced NSCLC. The findings revealed that ICI introduction conferred varying degrees of OS and CSS benefits across subgroups defined by age, gender, ethnicity, income level, geographical region, chemotherapy, radiation therapy, T stage, and N stage (Figure 3). Additionally, patients under 60 years of age, those receiving chemotherapy or radiotherapy, and those with higher N stages were more likely to derive OS or CSS benefits following ICI introduction (although some subgroup interaction did not reach statistical significance), which suggested potential therapeutic differences.

Figure 3 Subgroup analysis. (A) Subgroup analysis of all-cause mortality risk; (B) subgroup analysis of cancer-specific mortality risk. CI, confidence interval; ICI, immune checkpoint inhibitor.

Mediating role of the lymph node metastatic burden

Next, to better understand how the introduction of ICIs improved survival outcomes, we examined whether reductions in lymph node metastatic burden (measured by LNR and LODDS) mediated part of the survival benefit.

Data analysis revealed that patients diagnosed in the pre-ICI era cohort faced significantly higher all-cause mortality (HR =1.25, 95% CI: 1.18–1.33, P<0.001) and cancer-specific mortality (HR =1.32, 95% CI: 1.23–1.41, P<0.001). After adjusting for LNR or LODDS, the HR for all-cause and cancer-specific mortality for patients diagnosed in the pre-ICI era cohort were 1.24 (95% CI: 1.16–1.32, P<0.001)/1.30 (95% CI: 1.22–1.39, P<0.001) and 1.24 (95% CI: 1.17–1.32, P<0.001)/1.31 (95% CI: 1.22–1.40, P<0.001), respectively. Mediation analysis indicated that LNR and LODDS mediated 10.7% and 9.0% of the difference in all-cause mortality risk between the two era cohorts. For the difference in cancer-specific mortality risk, the proportions explained by LNR and LODDS were 9.0% and 7.6%, respectively (Table S2). Notably, adjusting for NPLN did not significantly reduce the all-cause and cancer-specific mortality risk for patients diagnosed in the pre-ICI era cohort compared to the ICI-introduction era cohort (P=0.33 and 0.39), suggesting that NPLN may not play a significant mediating role in the survival benefits observed following ICI introduction.

SHAP analysis further confirmed the prognostic impact of lymph node metastatic burden indicators. As illustrated in the SHAP plot (Figure S3), higher values of LODDS and LNR were associated with increased mortality risk (positive SHAP values), while NPLN showed limited and mixed influence. Cox regression analysis demonstrated that each 1% increase in LNR or each unit increase in LODDS was associated with a 87% (HR =1.87, 95% CI: 1.69–2.06, P<0.001)/19% (HR =1.19, 95% CI: 1.16–1.22, P<0.001) increase in all-cause mortality risk and a 94% (HR =1.94, 95% CI: 1.74–2.15, P<0.001)/20% (HR =1.20, 95% CI: 1.17–1.23, P<0.001) increase in CSS mortality risk among advanced NSCLC patients (Table S3).

Overall, these findings suggest that a portion of the survival improvement for advanced NSCLC patients diagnosed in the ICI-introduction era cohort can be explained by reduced lymph node metastatic burden.

The prognostic value of LNR and LODDS in NSCLC patients

To further assess whether the lymph node metastatic burden indicators (LNR and LODDS) identified through mediation analysis could provide precise prognostic stratification in early-stage NSCLC, we constructed an additional validation cohort comprising early-stage (M0-stage) NSCLC patients with at least one positive lymph node from the SEER database (Table S4). In this validation cohort, after multivariable adjustment, each unit increase in LNR or LODDS was associated with a 125% (HR =2.25, 95% CI: 2.10–2.40, P<0.001) and 22% (HR =1.22, 95% CI: 1.20–1.24, P<0.001) increase in the risk of all-cause mortality, respectively, and a 146% (HR =2.46, 95% CI: 2.28–2.65, P<0.001) and 25% (HR =1.25, 95% CI: 1.23–1.28, P<0.001) increase in the risk of cancer-specific mortality, respectively (Table S5).

RCS analysis further confirmed significant nonlinear relationships between LNR and LODDS and the risk of OS and CSS mortality in NSCLC patients (Figure S4). Using the maximum selection rank statistics method, significant cutoff values for LNR and LODDS were identified as 31% and −0.75, respectively (Figure S5). Based on these cutoffs, patients were categorized into high and low lymph node metastatic burden groups. Cox regression analysis revealed that high lymph node metastatic burden significantly increased the risk of all-cause and cancer-specific mortality (Table S5). Moreover, significant differences in mortality risk were observed with each unit increase in LNR above and below the cutoff values (interaction P<0.001, Figure S4).

Finally, we further investigated the prognostic value of combining anatomical staging with lymph node metastatic burden in NSCLC patients. Kaplan-Meier survival curves revealed that patients with N1 and low LNR burden had the best prognosis, followed by N2 with low LNR burden, N1 with high LNR burden, and finally N2 with high LNR burden (Figure S6). Cox regression analysis further supported these findings: after adjusting for multiple factors, compared to N1 with low LNR burden, the risk of all-cause and cancer-specific mortality increased by 55%/61% for N1 with high LNR burden, 26%/31% for N2 with low LNR burden, and 98%/121% for N2 with high LNR burden (Table S6). Similar trends were observed when lymph node burden was calculated using LODDS in combination with anatomical staging. These results further support the critical role of LNR and LODDS in prognostic assessment for NSCLC patients.

Discussion

Following the introduction of ICIs in 2015, OS and CSS among patients with advanced NSCLC showed significant improvement compared to the pre-ICI era. Further mediation analysis demonstrated that lymph node metastatic burden (LNR and LODDS) played an important mediating role in the survival benefits. Moreover, LNR and LODDS were also significantly associated with mortality risk in early-stage NSCLC patients and, when combined with anatomical staging, provided additional prognostic insights.

Over the past decade, ICIs, particularly PD-1/PD-L1 inhibitors, have demonstrated remarkable efficacy in locally advanced and metastatic NSCLC, providing varying degrees of survival benefits in both first-line and second-line treatments for metastatic NSCLC (22,23). For example, in second-line treatment, the CheckMate 017 (24), KEYNOTE-010 (25), and OAK studies (26) showed that nivolumab, pembrolizumab, and atezolizumab, respectively, significantly improved OS compared to docetaxel. Additionally, studies such as KEYNOTE-407 (27), KEYNOTE-189 (28), KEYNOTE-042 (29), CheckMate 227 (30), and CheckMate 9LA (31) have also provided evidence that first-line immunotherapy can deliver sustained benefits with manageable safety profiles for patients with advanced NSCLC. Consistent with previous studies, our findings based on real-world data also demonstrated that the introduction of ICIs significantly improves survival outcomes in advanced NSCLC patients, reducing the risk of all-cause mortality and cancer-specific mortality by approximately 20% and 24%, respectively. Beyond statistical significance, the clinical relevance of our observed survival benefit (adjusted HR for all-cause mortality: 0.80; 95% CI: 0.75–0.85) warrants discussion. Applying the Minimal Clinically Important Difference (MCID) framework for HRs, recently proposed by Horita et al. (32), an HR of 0.83 (Cohen’s d=0.2) is suggested as a small MCID, and 0.76 (d=0.3) as a medium MCID for critical outcomes like all-cause mortality. Our observed HR of 0.80, representing a 20% relative reduction in the hazard of death, thus surpasses the threshold for a small MCID and approaches that for a medium MCID. This suggests the survival benefit in our large real-world advanced NSCLC cohort is likely clinically important, in addition to being statistically significant.

Traditional OS does not reflect how prognosis changes over time (33). In most cancer treatments, patients’ survival probabilities increase significantly with prolonged survival, especially in those with advanced disease (34-36). For such patients, it is essential to predict long-term prognosis and estimate dynamic survival probabilities. CS is a dynamic method of assessing survival risk over time, providing more favorable prognostic estimates compared to the initial prognosis at the time of diagnosis (33). In our results, the subsequent survival probabilities (CS rates) in both cohorts gradually increased after surviving 1, 2, 3, and 4 years. Furthermore, we observed that patients in the ICI-introduction era cohort demonstrated higher survival probabilities at all time points compared to the pre-ICI era cohort, which suggests that the treatment paradigm characteristic of the ICI-introduction era cohort, notably the integration of ICIs, not only significantly reduces mortality risk in the early stages but also provides sustained survival advantages during long-term follow-up. The phenomenon may be closely related to the persistent immune surveillance mechanism induced by ICIs (37). By activating and maintaining the immune system’s tumor-suppressing effects, ICIs enable patients who respond well in the early stages to maintain a lower risk of mortality over time (37,38). The trend has also been observed in several immunotherapy trials, including the CheckMate 017 and CheckMate 057 studies, where patients treated with ICIs demonstrated sustained OS benefits at five years (30,31).

An increasing number of studies and real-world data have shown that lymph node status significantly influences survival and treatment response in advanced NSCLC patients. Inadequate control of lymph node metastases may lead to persistent or new metastatic spread (39,40). Through mediation analysis, we found that reductions in lymph node metastatic burden—reflected by LNR and LODDS—mediated a portion of the OS and CSS benefits observed following the introduction of ICIs. By contrast, the absolute NPLN did not exhibit a similar mediating effect. These findings suggest that rate-based metrics, which also account for negative lymph nodes, may more sensitively capture the true depth of tumor invasion and partly explain how treatment advancements characteristic of the ICI-introduction era cohort, with ICIs as a key component, confer survival advantages in advanced NSCLC.

Recently, the IASLC introduced refined analyses of N2 subcategories (N2a: single mediastinal or subcarinal lymph node station; N2b: multiple mediastinal lymph node stations) in its ninth edition of lung cancer staging proposals, significantly enhancing the granularity of the staging system (14,17). The advancement continues the efforts of the International Association for the Study of Lung Cancer (IASLC)’s seventh and eighth editions by incorporating the anatomical concept of ‘single/multi-station metastasis’ into the staging system for the first time in the ninth edition. However, IASLC remains cautious about incorporating quantitative lymph node metrics due to concerns about data completeness, sampling standardization, and compatibility with the existing staging framework (14). Therefore, it is of critical importance to develop a standardized and robust quantification approach. Our results demonstrate that rate-based quantification methods, such as LNR and LODDS, provide a more comprehensive reflection of the biological significance of lymph node metastasis. Compared to the simple count of positive nodes, rate-based metrics serve as standardized measures with greater consistency and stability, offering more accurate prognostic information. Furthermore, LNR and LODDS have demonstrated potential superiority over the N classification within the traditional TNM staging system in several other malignancies, including colorectal cancer, gallbladder cancer, and esophageal cancer (41-44). Our results showed that by combining anatomical location with thresholds of rate-based metrics, we found that patients in N2 stage with low-burden may have better prognoses than those in N1 stage with high-burden. The difference cannot be explained solely by anatomical location, further underscoring the critical role of rate-based metrics in prognostic evaluation. Additionally, this rate-based standardization concept holds potential for application in imaging modalities. For example, imaging-based assessment of the proportion of positive lymph node regions to total regions, such as regional lymph node density, could enable monitoring of lymph node burden and serve as an important tool for precise prognostic evaluation (45). In conclusion, our findings offer important clinical and research implications. The demonstrated prognostic value of LNR and LODDS in both early and advanced NSCLC can enhance current TNM staging for more precise risk stratification, potentially guiding patient management. Future research should validate these metrics prospectively and explore their potential predictive value for response to treatments characteristic of the ICI-introduction era, and investigate the biological mechanisms behind their mediating role in the observed era-dependent survival improvements.

We must acknowledge several limitations of this study. First, as the study is based on observational data from the SEER database, it is subject to potential selection bias, confounding factors, and measurement errors. Second, while the SEER database provides comprehensive data, it lacks certain critical patient information, such as whether patients received immunotherapy and details related to the latest ninth edition of the TNM staging system, which could offer deeper insights. Additionally, the database does not include important patient characteristics that may influence outcomes, such as performance status, smoking history, genetic mutation status, PD-L1 expression, and the number or type of irradiated sites. And, the SEER database lacks detailed comorbidity data, preventing adjustment for potential differences in comorbidity burden between cohorts. Third, the SEER database provides limited prognostic data, covering only OS and CSS, while excluding other important measures of treatment efficacy, such as objective response rate (ORR), disease control rate (DCR), PFS, and quality of life. Besides, due to a lack of explicit variables for ICIs in SEER, we used year of diagnosis (post-2015, which we define as our ‘ICI-introduction era’) as a real-world proxy for ICIs availability, which may lead to misclassification bias and inadvertently overlook concurrent advancements and impacts of targeted therapies approved during the same period, potentially confounding the attribution of observed survival improvements exclusively to the introduction of ICIs. Finally, while PSM analysis helps balance baseline differences between groups, it may not fully account for unmeasured confounding variables, potentially impacting the accuracy of the findings. And, the c-statistic of the PSM was 0.593, which is below the commonly accepted threshold (0.7–0.9), indicating modest ability to separate treatment groups. This may reflect limited observed differences in baseline characteristics and suggests potential residual confounding due to unmeasured variables.

Conclusions

The study, based on large-scale real-world data from the SEER database, evaluated survival outcomes in advanced NSCLC patients diagnosed in an era following the introduction of ICIs. The analysis suggested that the introduction of ICIs expanded patients’ treatment options and contributed to improved OS and CSS to some extent, with lymph node metastatic burden rate indicators (LNR and LODDS) identified as critical mediators of the observed survival benefits. Compared to simple numerical metrics, these rate-based metrics more accurately reflect the depth of tumor invasion and metastatic burden. Furthermore, LNR and LODDS were closely associated with mortality risk in early-stage NSCLC patients and, when combined with anatomical staging, provided more comprehensive prognostic insights. These findings underscore the potential of rate-based indicators in precision medicine, offering valuable guidance for optimizing TNM staging and individualized treatment strategies.

Acknowledgments

We are grateful to the US participants of SEER database who generously contributed their data for this study. In addition, thanks to the EmpowerStats software (Version 5.0, X&Y Solutions, Inc., Boston, MA; http://www.empowerstats.com) for providing support for the main analysis in this study.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-447/rc

Peer Review File: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-447/prf

Funding: This study was supported by the National Natural Science Foundation of China (No. 824B2055), the Ningbo Top Medical and Health Research Program (No. 2022030208), Foundation of QiangJi of Naval Medical Center (No. 22M3501), and Shanghai Hospital Development Center (No. SHDC22024308).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-447/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Hendriks LEL, Remon J, Faivre-Finn C, et al. Non-small-cell lung cancer. Nat Rev Dis Primers 2024;10:71. [Crossref] [PubMed]
2. Sorin M, Prosty C, Ghaleb L, et al. Neoadjuvant Chemoimmunotherapy for NSCLC: A Systematic Review and Meta-Analysis. JAMA Oncol 2024;10:621-33. [Crossref] [PubMed]
3. Brahmer J, Reckamp KL, Baas P, et al. Nivolumab versus Docetaxel in Advanced Squamous-Cell Non-Small-Cell Lung Cancer. N Engl J Med 2015;373:123-35. [Crossref] [PubMed]
4. Wang Y, Kondrat K, Adhikari J, et al. Survival trends among patients with metastatic non-small cell lung cancer before and after the approval of immunotherapy in the United States: A Surveillance, Epidemiology, and End Results database-based study. Cancer 2025;131:e35476. [Crossref] [PubMed]
5. Herrera-Juárez M, Serrano-Gómez C, Bote-de-Cabo H, et al. Targeted therapy for lung cancer: Beyond EGFR and ALK. Cancer 2023;129:1803-20. [Crossref] [PubMed]
6. Lu S, Wang J, Yu Y, et al. Tislelizumab Plus Chemotherapy as First-Line Treatment for Locally Advanced or Metastatic Nonsquamous NSCLC (RATIONALE 304): A Randomized Phase 3 Trial. J Thorac Oncol 2021;16:1512-22. [Crossref] [PubMed]
7. Paz-Ares L, Luft A, Vicente D, et al. Pembrolizumab plus Chemotherapy for Squamous Non-Small-Cell Lung Cancer. N Engl J Med 2018;379:2040-51. [Crossref] [PubMed]
8. Mc Cord KA, Ewald H, Agarwal A, et al. Treatment effects in randomised trials using routinely collected data for outcome assessment versus traditional trials: meta-research study. BMJ 2021;372: [Crossref] [PubMed]
9. Rahim MK, Okholm TLH, Jones KB, et al. Dynamic CD8(+) T cell responses to cancer immunotherapy in human regional lymph nodes are disrupted in metastatic lymph nodes. Cell 2023;186:1127-1143.e18. [Crossref] [PubMed]
10. Saddawi-Konefka R, Schokrpur S, Gutkind JS. Let it be: Preserving tumor-draining lymph nodes in the era of immuno-oncology. Cancer Cell 2024;42:930-3. [Crossref] [PubMed]
11. Deng H, Zhou J, Chen H, et al. Impact of lymphadenectomy extent on immunotherapy efficacy in postresectional recurred non-small cell lung cancer: a multi-institutional retrospective cohort study. Int J Surg 2024;110:238-52. [Crossref] [PubMed]
12. Provencio M, Nadal E, Insa A, et al. Neoadjuvant chemotherapy and nivolumab in resectable non-small-cell lung cancer (NADIM): an open-label, multicentre, single-arm, phase 2 trial. Lancet Oncol 2020;21:1413-22. [Crossref] [PubMed]
13. Zhai WY, Zhao ZR, Chen S, et al. Response of primary tumor and lymph node in non-small cell lung cancer after neoadjuvant immunotherapy: a pooled analysis. J Immunother Cancer 2022;10:e005160. [Crossref] [PubMed]
14. Huang J, Osarogiagbon RU, Giroux DJ, et al. The International Association for the Study of Lung Cancer Staging Project for Lung Cancer: Proposals for the Revision of the N Descriptors in the Forthcoming Ninth Edition of the TNM Classification for Lung Cancer. J Thorac Oncol 2024;19:766-85.
15. Liu S, Wang Z, Ge Y, et al. Prognostic significance of an innovative staging system based on the logarithmic odds of positive lymph nodes for resectable gastroesophageal cancer after neoadjuvant chemoradiation: a population-based study with external validation of data. J Transl Med 2024;22:801. [Crossref] [PubMed]
16. Deng W, Xu T, Wang Y, et al. Log odds of positive lymph nodes may predict survival benefit in patients with node-positive non-small cell lung cancer. Lung Cancer 2018;122:60-6. [Crossref] [PubMed]
17. Rami-Porta R, Nishimura KK, Giroux DJ, et al. The International Association for the Study of Lung Cancer Lung Cancer Staging Project: Proposals for Revision of the TNM Stage Groups in the Forthcoming (Ninth) Edition of the TNM Classification for Lung Cancer. J Thorac Oncol 2024;19:1007-27. [Crossref] [PubMed]
18. Zabor EC, Gonen M, Chapman PB, et al. Dynamic prognostication using conditional survival estimates. Cancer 2013;119:3589-92. [Crossref] [PubMed]
19. Hothorn T, Lausen BJRN. Maximally selected rank statistics in R. R News 2002;2:3-5.
20. Siegel RL, Kratzer TB, Giaquinto AN, et al. Cancer statistics, 2025. CA Cancer J Clin 2025;75:10-45. [Crossref] [PubMed]
21. Tolwin Y, Gillis R, Peled N. Gender and lung cancer-SEER-based analysis. Ann Epidemiol 2020;46:14-9. [Crossref] [PubMed]
22. Cramer-van der Welle CM, Verschueren MV, Tonn M, et al. Real-world outcomes versus clinical trial results of immunotherapy in stage IV non-small cell lung cancer (NSCLC) in the Netherlands. Sci Rep 2021;11:6306. [Crossref] [PubMed]
23. Voruganti T, Soulos PR, Mamtani R, et al. Association Between Age and Survival Trends in Advanced Non-Small Cell Lung Cancer After Adoption of Immunotherapy. JAMA Oncol 2023;9:334-41. [Crossref] [PubMed]
24. Borghaei H, Gettinger S, Vokes EE, 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;39:723-33. [Crossref] [PubMed]
25. Herbst RS, Garon EB, Kim DW, et al. Five Year Survival Update From KEYNOTE-010: Pembrolizumab Versus Docetaxel for Previously Treated, Programmed Death-Ligand 1-Positive Advanced NSCLC. J Thorac Oncol 2021;16:1718-32. [Crossref] [PubMed]
26. Rittmeyer A, Barlesi F, Waterkamp D, et al. Atezolizumab versus docetaxel in patients with previously treated non-small-cell lung cancer (OAK): a phase 3, open-label, multicentre randomised controlled trial. Lancet 2017;389:255-65. [Crossref] [PubMed]
27. Paz-Ares L, Vicente D, Tafreshi A, et al. A Randomized, Placebo-Controlled Trial of Pembrolizumab Plus Chemotherapy in Patients With Metastatic Squamous NSCLC: Protocol-Specified Final Analysis of KEYNOTE-407. J Thorac Oncol 2020;15:1657-69. [Crossref] [PubMed]
28. Rodríguez-Abreu D, Powell SF, Hochmair MJ, et al. Pemetrexed plus platinum with or without pembrolizumab in patients with previously untreated metastatic nonsquamous NSCLC: protocol-specified final analysis from KEYNOTE-189. Ann Oncol 2021;32:881-95. [Crossref] [PubMed]
29. Mok TSK, Wu YL, Kudaba I, et al. Pembrolizumab versus chemotherapy for previously untreated, PD-L1-expressing, locally advanced or metastatic non-small-cell lung cancer (KEYNOTE-042): a randomised, open-label, controlled, phase 3 trial. Lancet 2019;393:1819-30. [Crossref] [PubMed]
30. Reck M, Schenker M, Lee KH, et al. Nivolumab plus ipilimumab versus chemotherapy as first-line treatment in advanced non-small-cell lung cancer with high tumour mutational burden: patient-reported outcomes results from the randomised, open-label, phase III CheckMate 227 trial. Eur J Cancer 2019;116:137-47. [Crossref] [PubMed]
31. Paz-Ares L, Ciuleanu TE, Cobo M, et al. First-line nivolumab plus ipilimumab combined with two cycles of chemotherapy in patients with non-small-cell lung cancer (CheckMate 9LA): an international, randomised, open-label, phase 3 trial. Lancet Oncol 2021;22:198-211. [Crossref] [PubMed]
32. Horita N, Yamamoto S, Mizuki Y, et al. Minimal Clinically Important Difference (MCID) of Effect Sizes other than Mean Difference. Journal of Clinical Question 2024;1:116-27.
33. Hieke S, Kleber M, König C, et al. Conditional Survival: A Useful Concept to Provide Information on How Prognosis Evolves over Time. Clin Cancer Res 2015;21:1530-6. [Crossref] [PubMed]
34. Qaderi SM, Dickman PW, de Wilt JHW, et al. Conditional Survival and Cure of Patients With Colon or Rectal Cancer: A Population-Based Study. J Natl Compr Canc Netw 2020;18:1230-7. [Crossref] [PubMed]
35. Hagens ERC, Feenstra ML, Eshuis WJ, et al. Conditional survival after neoadjuvant chemoradiotherapy and surgery for oesophageal cancer. Br J Surg 2020;107:1053-61. [Crossref] [PubMed]
36. Shi Y, Zheng Y, Zhang H, et al. Dynamic estimates of survival in oncocytic cell carcinoma of the thyroid. Discov Oncol 2023;14:217. [Crossref] [PubMed]
37. Friedlaender A, Liu SV, Addeo A. Tracking the tail. J Immunother Cancer 2020;8:e000971. [Crossref] [PubMed]
38. Huang J, Xiong L, Tang S, et al. Balancing Tumor Immunotherapy and Immune-Related Adverse Events: Unveiling the Key Regulators. Int J Mol Sci 2024;25:10919. [Crossref] [PubMed]
39. Zhu J, Wang W, Xu S, et al. Evaluation of the Effect of Lymph Node Status on the Survival of Non-Small Cell Lung Cancer Patients With Brain Metastases: Applications of a Novel Grade Prognostic Assessment Score Model Involving N Stage. Front Oncol 2020;10:563700. [Crossref] [PubMed]
40. Chen D, Mao Y, Wen J, et al. Numbers and Stations: Impact of Examined Lymph Node on Precise Staging and Survival of Radiologically Pure-Solid NSCLC: A Multi-Institutional Study. JTO Clin Res Rep 2020;1:100035. [Crossref] [PubMed]
41. Huang B, Ni M, Chen C, et al. LODDS is superior to lymph node ratio for the prognosis of node-positive rectal cancer patients treated with preoperative radiotherapy. Tumori 2017;103:87-92. [Crossref] [PubMed]
42. Zhou YY, Du XJ, Zhang CH, et al. Comparison of three lymph node staging schemes for predicting the outcome in patients with small bowel adenocarcinoma: A population-based cohort and international multicentre cohort study. EBioMedicine 2019;41:276-85. [Crossref] [PubMed]
43. Li T, Yang Y, Wu W, et al. Prognostic implications of ENE and LODDS in relation to lymph node-positive colorectal cancer location. Transl Oncol 2021;14:101190. [Crossref] [PubMed]
44. Ye G, Chen Z, Wang L, et al. Log odds of positive lymph nodes predicts survival in patients treated with neoadjuvant therapy followed by esophagectomy. J Surg Oncol 2020;121:1074-83. [Crossref] [PubMed]
45. Ma J, Zhao R, Wu YL, et al. Regional lymph node density-based nomogram predicts prognosis in nasopharyngeal carcinoma patients without distant metastases. Cancer Imaging 2023;23:123. [Crossref] [PubMed]