Video-assisted mediastinoscopy compared with conventional mediastinoscopy: are we doing better?

Conventional mediastinoscopy (CM) is recently being replaced by video-assisted mediastinoscopy (VAM), with potentially better yield and better safety profile for VAM.

All 645 mediastinoscopies (505 CM, 140 VAM) performed between May 2004 and May 2008 were reviewed. Numbers of stations biopsied, total number of lymph nodes dissected, pathology results, and complications were recorded. Patients were divided into two groups: staging for lung cancer group (n = 500) and diagnostic group (n = 145). The staging group was further analyzed, using 304 patients who eventually underwent thoracotomy to evaluate accuracy and negative predictive value of mediastinoscopy, comparing between the two methods (233 CM, 71 VAM).

Average age was 65 years (range, 26 to 91), and 382 were male. There was no mortality. Eight complications (1.2%) occurred, more in the VAM group (3.8%) than in the CM group (0.8%; p = 0.04). The total number of dissected nodes was higher in the VAM group than in the CM group (7.0 +/- 3.2 versus 5.0 +/- 2.8, p < 0.001), and so was the number of stations sampled (3.6 versus 2.6, p < 0.01). Sensitivity was higher for VAM (95% versus 92.2%, p = not significant), and so was the negative predictive value (98.6% versus 95.7%, p = not significant). Most false negative biopsies (8 of 11, 73 %) occurred in station 7.

Both methods are safe. More lymph nodes and stations were evaluated by VAM, with trend toward higher negative predictive value. The higher rate of minor complications seen with VAM might be related to a more aggressive and thorough dissection.