Effect of selective lymph node dissection based on patterns of lobe-specific lymph node metastases on patient outcome in patients with resectable non-small cell lung cancer: A large-scale retrospective cohort study applying a propensity score

Objective: Lobectomy with systematic complete mediastinal lymph node dissection is standard surgical treatment for localized non-small cell lung cancer. However, selective mediastinal lymph node dissection based on lobe-specific metastases (selective dissection) has often been performed. This study was designed to evaluate the validity of the selective lymph node dissection. Methods: From 1995 through 2003, 625 patients in our hospital had surgery for complete mediastinal lymph node dissection and 147 for selective dissection. We evaluated whether selective dissection adversely affected overall survival. To minimize possible biases due to confounding by treatment indication, we performed a retrospective cohort analysis by applying a propensity score. The propensity score was calculated by logistic regression based on 15 factors available that were potentially associated with treatment indication. Patients were divided into 4 groups according to quartile, and comparison between selective dissection and complete mediastinal lymph node dissection was made using propensity score quartile-stratified Cox proportional hazard models. Results: Comparison of baseline characteristics between patients having selective dissection and patients having complete mediastinal lymph node dissection according to propensity score quartile supported comparability of the 2 groups. The 5-year overall survival rates were 76.0% for selective dissection versus 71.9% for complete mediastinal lymph node dissection. The 5-year survival probabilities stratified by propensity score quartile consistently showed no marked difference. In multivariate models, there was no significant difference between the 2 groups (hazard ratio = 1.17, P = .500) as also seen in the analysis without propensity score (hazard ratio 1.06; 95% confidence interval, 0.68-1.64; P = .810). Therefore, selective dissection showed no significant impact on poor survival compared with complete mediastinal lymph node dissection. Conclusions: Selective lymph node dissection did not worsen the survival of patients with non-small cell lung cancer. (J Thorac Cardiovasc Surg 2010;139:1001-6)