Comparison of lymph node dissection and lymph node sampling for non-small cell lung cancers by video-assisted thoracoscopic surgery.

Background: Video-assisted thoracoscopic surgery (VATS) has been increasingly used in the treatment of lung cancers. But it is still unclear whether mediastinal lymph node dissection (LND) under VATS is safe and feasible. The aim of this study is to figure out whether IND by VATS is safe and feasible. Methods: Consecutive patients with primary resectable lung cancers referred for lobectomy and LND or sampling by VATS between January 2012 and December 2016 were retrospectively reviewed. Clinicopathological characteristics and perioperative results were collected for statistical analysis. Results: Seven-hundred and seventy-three VATS lobectomy patients were included in this study, 494 received LND and 279 received lymph node sampling (LNS). There were more male patients, higher pathological T and N stage in the LND group than in the LNS group. Multivariate analysis suggested that clinical N stage higher than cNO category and LND were independent risk factors for finding pN2 diseases in all lung cancers, while higher than cNO category, solid or micropapillary component, and LND were independently related to finding pN2 stage in adenocarcinomas. Propensity-score matching rendered 279 pairs of patients with no significant difference in age, gender, co-morbidity, tumor location, or T stage. Although the IND group had longer operation time (128 vs. 114 minutes, P<0.001), higher amount of postoperative drainage (920 vs. 720 mL, P<0.001), longer postoperative hospital stay (6 vs. 4 days, P<0.001) than the LNS group, no difference was observed in overall morbidity or mortality between the two groups. Conclusions: LND by VATS has acceptable perioperative results but can provide more accurate nodal staging compared with INS. IND by VATS is safe, feasible, and should be recommended in patients with tumors in clinical N stages higher cNO category or with more invasive histology.