Background: This retrospective study investigated whether mediastinal lymphadenectomy compliant with the National Comprehensive Cancer Network (NCCN) criteria will improve the oncological outcomes of clinical early-stage lung cancer. Methods: From 2003-2010, 712 consecutive cases of clinical N0/1 were included for retrospective analysis, including 152 confirmed cases of pN2 and 560 of pN0-1 disease following surgery. Group A was defined as the cases fulfilling NCCN lymphadenectomy criteria (≥ three stations of N2 nodes dissection) and group B included all other cases. The groups were stratified according to pN status and the outcomes were assessed. Results: Five-year overall survival (OS) and 5-year disease-free survival (DFS) were significantly different between group A versus B [72%±2% vs. 63%±4% (OS), P=0.014; 58.0%±2% vs. 49%±4% (DFS), P=0.038] in the whole cohort. After stratification by pN status, this difference was remained in pN2 subgroup [50%±5% vs. 25%±9% (OS), P=0.006; 31.0%±4% vs. 13%±7% (DFS), P=0.014], but not in pN0-1 subgroups. Cox regression analysis showed that performing a lymphadenectomy fulfilling NCCN criteria was a significant prognostic factor for OS either in the whole cohort [P=0.003, hazard ratio (HR): 0.598, 95% confidence interval (CI): 0.425-0.841] or in patients of pN2 status (P=0.038, HR: 0.559, 95% CI: 0.323-0.968). Cases with ≥4 N2 stations dissected did not achieve better survival benefit compared to those harvesting 3 stations in cN0/1-pN2 group (P=0.152). Conclusions: Mediastinal lymphadenectomy fulfilling NCCN criteria appears to improve the survival of unexpected N2 group (cN0/1-pN2) among early-stage lung cancer patients. More extended N2 node dissection may not further improve the outcome in this group.