Evaluation of lymph nodes in various surgical approaches for clinical N0 lung cancer: a systematic review and network meta-analysis.

[BACKGROUND] Clinical N0 lung cancer is surgically treated using a variety of approaches. Nevertheless, a comparison of the effectiveness of these approaches in evaluating lymph node (LN) remains to be conducted.

[OBJECTIVE] To evaluate the upgrading rate of LN staging, the total number of resected LNs, and the number of resected LN stations in clinical N0 lung cancer among four surgical approaches.

[METHODS] A comprehensive search was conducted across the PubMed, EMBASE, MEDLINE, and CENTRAL databases, encompassing papers published between 1 January 2005, and 31 March 2025. The pooled effect size was measured using odds ratios (ORs) or standardized mean differences (SMDs), as well as 95% confidence intervals (CIs). The surface under the cumulative ranking curve (SUCRA) was utilized to calculate the cumulative probability for each approach. In order to investigate the origins of heterogeneity, meta-regression and subgroup analyses were employed.

[RESULTS] A total of 71 070 participants from 30 studies were included in the analysis. The incidence of cN0-pN1 was higher in the open thoracotomy (OPEN) group than in the robot-assisted thoracoscopic surgery (RATS) and video-assisted thoracoscopic surgery (VATS) groups [OR = 0.53 (0.32, 0.89), OR = 0.36 (0.26, 0.51), respectively]. OPEN had the highest SUCRA (97.7%), followed by RATS (59.5%) and VATS (27.3%). Regarding cN0-pN2, both the OPEN and RATS groups exhibited a higher incidence than the VATS group [OR = 0.39 (0.16, 0.91), OR = 0.20 (0.06, 0.64), respectively] with RATS demonstrating the greatest SUCRA (92.3%), followed by OPEN (56.8%) and VATS (0.9%). The video-assisted mediastinoscopic lymphadenectomy plus video-assisted thoracoscopic surgery (VM + VATS) group removed more total LNs than the VATS group [SMD = 6.66 (1.54, 11.77)], with VM + VATS having the highest SUCRA (97.7%), followed by RATS (57.8%), OPEN (44.1%), and VATS (0.4%). Heterogeneity was evident in cN0-pN2 and total LNs. Regression analysis revealed that sample size could explain some of the heterogeneity observed in cN0-pN2 [coef = 0.62 (0.04-1.26)], whereas sample size, publication year, and surgical area could explain some of the heterogeneity observed in total LNs [coef = - 21.3 (-42.03-4.23), coef = - 21.02 (-42.26-3.25), coef = - 12.54 (-31.59-3.70), respectively]. Subgroup analysis supported the previous findings.

[CONCLUSIONS] In terms of nodal upstaging incidence for cN0-pN1 and cN0-pN2, both OPEN and RATS performed well. For N1 lymph node (LN) resection, the OPEN approach is superior to the RATS and VATS approaches, whereas for N2, the RATS approach is superior to the OPEN and VATS approaches. Regarding the total number of LNs removed, the video-assisted mediastinoscopic lymphadenectomy (VM) + VATS and OPEN have obvious advantages. The publication year, sample size, and surgical area affected the incidence of cN0-pN2 and the total number of LNs removed. These findings may help thoracic surgeons to select different surgical approaches for patients with clinical N0 lung cancer and different tendencies for LN metastasis.