Safety and efficacy of video-assisted versus conventional lung resection for lung cancer

Objective: We sought to evaluate the use of video-assisted thoracoscopy among patients with lung cancer and its safety and effectiveness relative to conventional resection. Methods: A cohort study (1994-2002) was conducted by using the Surveillance, Epidemiology, and End-Results Medicare database. Video-assisted thoracoscopy and conventional resection were hypothesized to be equivalent in terms of risks of death. Equivalency was defined by a confidence interval of 0.72 to 1.28 for the odds of 30-day death and 0.89 to 1.11 for the hazard of death, corresponding to a difference of no more than 1% for 30-day mortality and 5% for 5-year survival, respectively. Results: Among 12,958 patients who underwent segmentectomy or lobectomy (mean age, 74 +/- 5 years), 6% underwent video-assisted thoracoscopy. The use of video-assisted thoracoscopy increased from 1% to 9% between 1994 and 2002. Compared with those who underwent conventional resection, patients who underwent video-assisted thoracoscopy more frequently had smaller tumors (P < .001) and stage I disease (P = .03), underwent lymphadenectomy (P < .001), and were cared for by higher-volume surgeons (P < .001) and at higher-volume hospitals (P < .001). After adjusting for differences in patient, cancer, management, and provider characteristics, the odds of early death were not significantly different between patients undergoing video-assisted thoracoscopy and those undergoing conventional resection, although equivalency was not demonstrated ( adjusted odds ratio, 0.93; 95% confidence interval, 0.57-1.50). The hazard of death was equivalent for video-assisted thoracoscopy and conventional resection (adjusted hazard ratio, 0.99; 95% confidence interval, 0.90-1.08). Conclusions: Video-assisted thoracoscopy was uncommonly used to manage lung cancer, although its use has increased over time. Video-assisted thoracoscopy and conventional resection were equivalent in terms of long-term survival.