Postoperative Radiotherapy (PORT) Improves the Survival of Patients with N2 Non-Small Cell Lung Cancer (NSCLC) with Regional Nodes Examined (RNE) ≤16 - A Real World Study Using Surveillance, Epidemiology, and End Results (SEER) Database
Y. Bao,Y. Men,J. Kang,X. Sun,M. Zhao,X. Yang,S. Sun,M. Yuan,Z. Ma,Z. Hui
DOI: https://doi.org/10.1016/j.ijrobp.2020.07.1236
2020-01-01
Abstract:PORT is widely used in patients with N2 NSCLC, but which subgroup population can benefit from PORT is still controversial. Our study is to evaluate the relationship between the RNE and the effect of PORT on survival of patients with N2 NSCLC from SEER database. Data from the SEER database were filtrated by the inclusion criteria of patients with NSCLC diagnosed at 2010-2016, with N2 disease (AJCC 7th) and treated with lobectomy or pneumonectomy. We excluded patients with unclear basic information (such as sex, histology or cause of death), multiple primary malignant tumors and M1 disease. All data were analyzed using the R version 3.6.2 software. And propensity-score matched (PSM) analysis was used to match the baseline characteristics between PORT group and non-PORT group. Kaplan-Meier method was used to estimate overall survival (OS) and cancer specific survival (CSS). Univariable and multivariable Cox proportional hazards models were adopted to estimate hazard ratios (HR) of predictors of survival. The HR of PORT for patients with different RNE was calculated to determine the best cut-off value of RNE. 2904 patients were included. The baseline characteristics were as follows: age≤70 (62.9%), white (80.5%), female (51.2%), grade 1-2 (53.0%), adenocarcinoma (67.1%), T2 (51.1%), lobectomy (89.9%), chemotherapy (74.9%), median number of positive regional nodes (3), median number of RNE (12). There were 1188 patients (40.9%) with PORT. RNE≤16 was chosen as the cut-off value and divided the patients into two cohorts. PSM was then conducted between the PORT group and the non-PORT group of each cohort. In RNE≤16 cohort (1398 patients), the median OS of the PORT and non-PORT groups were 48 months vs 40 months and 5-year OS rates were 42.93% vs 39.09%, respectively (P = 0.013). The median CSS were 59 months vs 44 months and 5-year CSS rates were 48.82% vs 43.15%, respectively (P = 0.008). Univariable analysis showed that PORT significantly improved OS (HR = 0.815, P = 0.013) and CSS (HR = 0.794, P = 0.008). Multivariable analysis confirmed that PORT was the significant predictor of OS (HR = 0.816, P = 0.014) and CSS (HR = 0.799, P = 0.011). In RNE>16 cohort (550 patients), The median OS of the PORT and non-PORT groups was 44 months vs 64 months and 5-year OS rates were 42.75% vs 50.69%, respectively (P = 0.210). The median CSS of the two groups were 47 months and not reached, the 5-year CSS rates were 46.17% vs 54.10%, respectively (P = 0.207). Neither univariable nor multivariable analysis showed that PORT was an individual predictive factor. For N2 NSCLC patients with RNE≤16, PORT can significantly improve the OS and CSS. However, for those with RNE>16, PORT may not improve survival, but even have the tendency of damage. These results need to be clarified by prospective randomized clinical trials.