Effect of surgery on survival in patients with stage III N2 small cell lung cancer: propensity score matching analysis and nomogram development and validation.

ABSTRACT

BACKGROUND: The standard treatment of stage III N2 small cell lung cancer (SCLC) is concurrent chemoradiation, and surgery is not recommended. This study was aimed to evaluate whether surgery has survival benefits in patients with stage III N2 SCLC and investigate the factors influencing survival of surgery. METHODS: Patients diagnosed with stage T1-4N2M0 SCLC from 2004 to 2015 were selected from the Surveillance Epidemiology End Results database. Propensity score matching (PSM) was used to balance confounders between patients who underwent surgery and those treated with radiation and/or chemotherapy. We compared overall survival (OS) of the two groups using Kaplan-Meier curves and a Cox proportional hazard model. We also identified prognostic factors in patients with surgical resection, and a nomogram was developed and validated for predicting postoperative OS. RESULTS: -A total of 5576 patients were included in the analysis; of these, 211 patients underwent surgery. PSM balanced the differences between the two groups. The median OS was longer in the surgery group than in the non-surgery group (20 vs. 15 months; p = 0.0024). Surgery was an independent prognostic factor for longer OS in the multivariate Cox regression analysis, and subgroup analysis revealed a higher survival rate in T1 stage patients treated with surgery (hazard ratio = 0.565, 95% confidence interval 0.401-0.798; p = 0.001). In patients who underwent surgery, four prognostic factors, including age, T stage, number of positive lymph nodes, and radiation, were selected into nomogram development for predicting postoperative OS. C-index, decision curve analyses, integrated discrimination improvement, and time-dependent receiver operating characteristics showed better performance in nomogram than in the tumor-node-metastasis staging system. Calibration plots demonstrated good consistency between nomogram predicted survival and actual observed survival. The patients were stratified into three different risk groups by prognostic scores and Kaplan-Meier curves showed significant difference between these groups. CONCLUSIONS: These results indicate that surgery can prolong survival in patients with operable stage III N2 SCLC, particularly those with T1 disease. A nomogram that includes age, T stage, number of positive lymph nodes, and radiation can be used to predict their long-term postoperative survival.