Efficacy and benefit of mediastinal computed tomography as a selection method for mediastinoscopy.

In 95 consecutive patients with proven or suspected bronchial carcinoma, computed tomographic evaluation of the upper mediastinum for N2 disease was performed prospectively. Patients with positive results underwent mediastinoscopy. Patients with perinodal N2 or N3 disease at mediastinoscopy were not considered candidates for operation. The mediastinum was declared negative only when intraoperative mediastinal lymph node dissection showed tumor-free nodes. Of the 95 patients, 12 had benign lesions, 14 were excluded from further evaluation because the lymph node status of the mediastinum was not proven intraoperatively, and 6 others were excluded from the final evaluation because of violation of the protocol. Twenty-two of the 75 remaining patients had a positive computed tomographic scan and underwent mediastinoscopy. Fourteen patients with positive results were considered to have inoperable disease. Fifty-three patients (70.7%) did not undergo mediastinoscopy. We performed seven probably incomplete resections, two for palliative reasons, and two thoracotomies without resection in patients with N2 disease. A policy of routine mediastinoscopy would have prevented only 5% of the thoracotomies performed in patients with lung cancer.