We aimed to analyze the accuracy of video-assisted mediastinoscopic lymphadenectomy (VAMLA) as a tool for preoperative staging and the impact of the technique on survival in patients with non-small cell lung cancer (NSCLC) undergoing pulmonary resection.
Between May 2006 and December 2010, 433 patients underwent pulmonary resection for NSCLC, 89 (21%) had VAMLA before resection and 344 (79%) had standard mediastinoscopy. The patients who had negative VAMLA/mediastinoscopy results underwent anatomic pulmonary resection and systematic lymph node dissection.
The median and mean numbers of resected lymph node stations were 5 and 4.9 in the VAMLA group and 4 and 4.2 in the mediastinoscopy group (P = .9). The mean number of lymph nodes per biopsy specimen using standard mediastinoscopy was 10.1, whereas it was 30.4 using VAMLA (P < .001). VAMLA unveiled N2 or N3 disease in 30 (33.7%) and in 6 (6.7%) of patients, respectively. The negative predictive value, sensitivity, false-negative value, and accuracy of VAMLA were statistically higher in the VAMLA groups compared with those of standard mediastinoscopy. The 5-year survival was 90% for VAMLA patients and 66% for mediastinoscopy patients (P = .01). By multivariable analysis, VAMLA was associated with better survival (odds ratio, 1.34; 95% confidence interval, 1.1-3.2; P = .02).