BACKGROUND: We compared survival between video-assisted thoracoscopic surgery (VATS) and thoracotomy approaches to lobectomy for non-small cell lung cancer. METHODS: Overall survival of patients who had lobectomy for any stage non-small cell lung cancer without previous chemotherapy or radiation from 1996 to 2008 was evaluated using the Kaplan-Meier method and multivariate Cox analysis. Propensity scoring was used to assess the impact of selection bias. RESULTS: Overall, 1,087 patients met inclusion criteria (610 VATS, 477 thoracotomy). Median follow-up was not significantly different between VATS and thoracotomy patients overall (53.4 versus 45.4 months, respectively; p=0.06) but was longer for thoracotomy for surviving patients (102.4 versus 67.9 months, p<0.0001). Thoracotomy patients had larger tumors (3.9±2.3 versus 2.8±1.5 cm, p<0.0001), and more often had higher stage cancers (50% [n=237] versus 71% [n=435] stage I, p<0.0001) compared with VATS patients. In multivariate analysis of all patients, thoracotomy approach (hazard ratio [HR] 1.22, p=0.01), increasing age (HR 1.02 per year, p<0.0001), pathologic stage (HR 1.45 per stage, p<0.0001), and male sex (HR 1.35, p=0.0001) predicted worse survival. In a cohort of 560 patients (311 VATS, 249 thoracotomy) who were assembled using propensity scoring and were similar in age, stage, tumor size, and sex, the operative approach did not impact survival (p=0.5), whereas increasing age (HR 1.02 per year, p=0.01), pathologic stage (HR 1.44 per stage, p<0.0001), and male sex (HR 1.29, p=0.01) predicted worse survival. CONCLUSIONS: The thoracoscopic approach to lobectomy for non-small cell lung cancer does not result in worse long-term survival compared with thoracotomy.