Abstract
PURPOSE: This study aimed to investigate the application of the number of positive lymph nodes (PLNs) in tumor, node, metastasis (TNM) staging system of non-small cell lung cancer (NSCLC) patients. Patients and Methods. We screened a total of 15820 patients with resected NSCLC between 2004 and 2015 from SEER database. The X-tile model was used to determine the cutoff values of the number of PLNs. Overall survival (OS) curves were plotted using the Kaplan-Meier method, and the differences among the individual groups were defined using the log-rank test. Cox regression model was used to perform univariate and multivariate analyses and to assess the association between the number of PLNs and OS. RESULTS: In this study, using the X-tile model, we screened three different cutoff values, including nN0, nN1-3, and nN4-. Survival curves demonstrated that our defined nN stage had a significant predictive value for OS (P < 0.001). In the univariate and multivariate Cox analyses, the result showed that nN stage was a significant prognostic factor of OS for NSCLC patients (P < 0.001). Subsequently, we classified the patients into five subgroups based on the combination of pN and nN stages, including pN0 + nN0, pN1 + nN1-3, pN2 + nN1-3, pN1 + nN4-, and pN2 + nN4-. Moreover, survival curves revealed significant differences among these five groups (P < 0.001). CONCLUSION: A combination of pathological LNs (pN) and the number of LN (nN) involvement in NSCLC patients had a better prognostic value than the current TNM staging system based on only pN stage.