BACKGROUND The non-small cell lung cancer TNM classification system uses only the anatomic extent of lymph node (LN) metastases to define the N category. The number of LNs resected and the ratio of positive LNs to total examined LNs are prognostic in other solid tumors. We used the Surveillance, Epidemiology and End Results database to investigate the effect of these factors on the overall survival of non-small cell lung cancer. METHODS All patients with non-small cell lung cancer in the Surveillance, Epidemiology and End Results database from 1988 through 2007 who had curative resections and had at least one LN examined were included. The prognostic value of age, race, sex, tumor size, histologic grade, number of examined LNs, and ratio of positive LNs to total examined LNs was assessed using a multivariate Cox proportional hazards model for overall survival. The number of LNs examined was categorized into four levels. The percentage of positive LNs was stratified into three levels. RESULTS Among patients with localized disease, fewer LNs examined corresponded with a worse prognosis. Prognosis improved as more LNs were examined. For patients with regional disease, the differences were significant only at the extremes. Older patients, males, and those with higher grade or larger tumors did worse. Patients with low or moderate ratios of positive to total LNs had better prognoses than those with high ratios. CONCLUSIONS More LNs resected and lower ratios of positive LNs to total examined LNs are associated with better patient survival after non-small cell lung cancer resection independent of age, sex, grade, tumor size, and stage of disease.