OBJECTIVE To describe an initial 3-year experience with video-assisted thoracic surgical procedures (VATS) at Mayo Clinic Rochester. METHODS We review the cumulative data on 771 VATS performed between June 1, 1991, and May 31, 1994, and assess the applications for this technique. METHODS The indications for VATS, our techniques used, and the associated mortality and morbidity are summarized. In addition, the frequency of conversion of VATS to open procedures and the reasons for choosing this strategy are discussed. RESULTS The 771 study patients (401 male and 370 female patients) had a median age of 62 years (range, 7 to 96). For all VATS. we used one-lung general anesthesia, without carbon dioxide insufflation. Indications for performing VATS were a pulmonary nodule in 333 patients, pleural effusion in 208, pulmonary infiltrate in 117, pneumothorax in 51, mediastinal mass in 22, pleural mass in 17, air leak in 13, and other in 10. The procedure was a wedge excision in 352 patients, examination of the pleural cavity in 128, pleural biopsy in 86, talc pleurodesis in 85, wedge excision and mechanical pleurodesis in 46, decortication in 27, excision of a mediastinal mass in 12, sympathectomy in 4, and other in 16. The rate of conversion of VATS to thoracotomy was 33.1% and did not change throughout the period of the study. The most common reasons for conversion were to complete a resection of a malignant lesion or to remove a deep nodule. The overall operative mortality was 1.9%. Complications occurred in 43 patients (8.3%) who underwent VATS without conversion to an open procedure and included prolonged air leak in 14, respiratory failure in 8, pneumothorax in 6, and atrial fibrillation in 5. The median hospitalization was 5 days (range, 1 to 104). CONCLUSIONS VATS is safe and useful for selected thoracic conditions. We favor conversion to thoracotomy when curative resection of a malignant lesion is intended.