There is a great deal of concern about metastasis of lung cancer to regional lymph nodes, due partly to the work of groups of thoracic surgeons in Japan and North America beginning in the 1970s. The classification of regional lymph node stations for lung cancer staging published by Mountain and Dresler has been widely adopted for more than ten years. Anatomic landmarks for 14 levels of intrapulmonary, hilar, and mediastinal lymph nodes stations are designated. Skip transfer and occult lymph node metastasis, confirmed by studies regarding the mode of spread of intrathoracic lymphatic metastasis, are two theoretical bases for complete mediastinal lymphadenectomy of lung cancer. However, whether or not the degree of the dissection influences prognosis, the role of systematic nodal dissection (SND) vs mediastinal lymph node sampling (MLD) in resectable non-small cell lung cancer (NSCLC) remains controversial. A systematic literature search was performed to identify relevant reports, making full use of the 'Cited by,' 'Related Records,' 'References,' and 'Author Index' functions in the PubMed and ISI Web of Science databases. This paper presents a review of the role of mediastinal lymph node distribution and methods of determining suitability for hilar and mediastinal lymphadenectomy based on the four subsets of stage IIIA-N2, balancing the cost vs effect of mediastinal lymph node dissection in resectable NSCLC, focusing on the stage migration bias in clinical trials comparing SND and MLS, recommending a reasonable node dissection sequence, improving the prospects for the perioperative anti-tumor therapy based on mediastinal lymphadenectomy, and evaluating the various preoperative staging techniques. Finally, we believe that, besides the role of complete resection and accurate staging, the complete mediastinal lymphadenectomy is the core component of the lung cancer multidisciplinary therapy, and suggest that the values of lymphadenectomy should be further assessed using decision-tree analysis based on large-scale prospective randomized trials and pooled analysis to evaluate the costs vs effects.