Lung cancer is a leading cause of cancer-related deaths globally, with non-small cell lung cancer (NSCLC) comprising the majority of cases. Surgical resection and lymph node dissection are standard treatments for resectable NSCLC. Pathological N stage (pN), determined by lymph node sampling after surgery, is a crucial prognostic factor, influencing decisions regarding adjuvant therapies. However, the optimal number of lymph nodes to examine remains debated, with guidelines varying across organizations. The number of examined lymph nodes has been linked to improved prognosis, but surgical quality can affect its accuracy in representing pN stage. Lymph node ratio (LNR), the ratio of metastatic to total harvested lymph nodes, has proven prognostic value in other cancers. This study aimed to compare the prognostic value of LNR and pN stage in NSCLC patients undergoing surgery and to propose a revised pN (r-pN) classification combining both.

Existing literature highlights the importance of pN staging in NSCLC prognosis, with survival decreasing with increasing pN stage. However, the heterogeneity within each pN stage necessitates a more precise classification system. Studies have shown a correlation between the number of examined lymph nodes and survival, but the optimal number remains unclear, varying from 6 to 16 across different guidelines. Surgical quality introduces variability in the number of examined lymph nodes, thus influencing the reliability of pN staging. Previous research on LNR in various cancers indicates its potential as a prognostic biomarker. This study builds on this existing knowledge to explore the utility of LNR in the context of NSCLC surgery.

This retrospective cohort study utilized data from the SEER databases (2004-2015) on NSCLC patients who underwent surgery as initial treatment. Inclusion criteria included histologically confirmed adenocarcinoma or squamous cell carcinoma, age ≥18 years, lymph node examination performed, definite TNM staging (7th edition AJCC), non-M1 and non-N0 stages, and no prior chemotherapy or radiotherapy. Patient characteristics (age, sex, race, primary site, tumor grade, histology, T stage, N stage, number of resected and positive lymph nodes) were extracted. LNR was calculated as the ratio of metastatic to total examined lymph nodes. X-tile software identified optimal LNR cut-off values, categorizing patients into LNR1 (LNR < 0.19), LNR2 (0.19 ≤ LNR ≤ 0.73), and LNR3 (LNR > 0.73). Cancer-specific survival (CSS) and overall survival (OS) were the primary and secondary endpoints, respectively. Kaplan-Meier analysis and log-rank tests compared survival across LNR and pN stages. A revised pN (r-pN) classification was created by stratifying pN categories based on LNR stages. Harrell's C-index and Akaike Information Criterion (AIC) assessed the predictive performance of pN, LNR, and r-pN stages in multivariable proportional hazards models. Statistical significance was set at P<0.05.

The study included 7,792 NSCLC patients. X-tile analysis determined the optimal LNR cut-off values as 0.19 and 0.73. Median CSS for LNR1, LNR2, and LNR3 groups were 71, 41, and 17 months, respectively. Similarly, median OS for these groups were 50, 35, and 16 months. LNR2 and LNR3 showed significantly worse CSS and OS compared to LNR1 (P<0.001 for all comparisons). Median CSS for pN1, pN2, and pN3 were 62, 33, and 14 months, respectively, while median OS was 46, 29, and 13 months. pN2 and pN3 also showed significantly poorer survival than pN1 (P<0.001 for all comparisons). In multivariable analysis, LNR stages demonstrated better penalized goodness-of-fit (lower AIC) and discriminatory ability (higher Harrell's C-index) than pN stages in predicting both CSS and OS. The r-pN classification exhibited the best predictive performance, with the lowest AIC and highest C-index for both CSS and OS. The r-pN staging effectively categorized patients into low, medium, and high-risk groups with significantly different survival outcomes (P<0.001).

This study confirms the prognostic importance of LNR in NSCLC patients undergoing surgery. The superior predictive performance of LNR compared to pN stage highlights the limitations of solely relying on the presence and location of positive lymph nodes for prognostication. LNR incorporates the extent of nodal involvement relative to the total number of examined nodes, offering a more nuanced assessment of disease burden and thus a more accurate prediction of treatment outcomes. The improved performance of the r-pN classification underscores the benefits of integrating both LNR and pN information for more precise risk stratification. This could lead to better-tailored adjuvant treatment decisions, potentially optimizing treatment strategies and improving patient outcomes. The findings support the use of LNR as a valuable tool for improving the accuracy and precision of NSCLC prognosis and treatment planning.

This study demonstrated that LNR is a superior prognostic factor compared to pN stage for predicting CSS and OS in NSCLC patients undergoing surgery. The r-pN classification, integrating both LNR and pN, showed the best predictive ability. These findings suggest that incorporating LNR into clinical practice could lead to improved risk stratification and more individualized treatment strategies for NSCLC patients. Further research could investigate the clinical utility of r-pN staging in guiding adjuvant therapy selection and assessing its impact on long-term outcomes.

This study was a retrospective cohort study, susceptible to inherent limitations such as selection bias and potential confounding factors not fully accounted for. The use of SEER database data may limit the generalizability to populations outside the United States. Furthermore, the study did not account for the specific types of adjuvant therapies administered, which could further refine risk stratification and treatment optimization. Further prospective studies are warranted to validate these findings and assess the clinical impact of LNR integration into routine clinical practice.