We report our combined experience on video-assisted thoracoscopic (VAT) anatomic lung resections from two major hospitals in Hong Kong over a 17-month period. From August 1993 to December 1994, 82 cases of major lung resections were attempted using the VATS approach, of which 60 were successfully completed (55 lobectomies, 2 bilobectomies, 2 pneumonectomies, and 1 segmentectomy). Of these 60 cases, there were 43 men and 17 women with a mean age of 66 years (range, 37 to 85 years). The final pathologies were 52 primary lung cancers (37 adenocarcinoma, 11 squamous cell carcinoma, 2 bronchoalveolar carcinoma, 1 adenosquamous carcinoma, and 1 undifferentiated carcinoma); 1 pulmonary metastasis (from nasopharyngeal carcinoma); and 7 benign lesions (3 tuberculosis, 1 bronchiectasis, 1 sclerosing hemangioma, 2 organizing pneumonia). There was one postoperative death (mortality rate, 1.8%). Complications include persistent air leak over 10 days (four), wound infection (two), supraventricular tachycardia (three), and recurrence of tumor over the utility thoracotomy scar (one). All the patients were followed up from 8 weeks to 19 months (mean, 10 months). The mean duration of chest drainage was 5.4 days (range, 2 to 25 days). The mean hospital stay was 7.2 days (range, 4 to 35 days). The average postoperative parenteral narcotic (meperidine hydrochloride [Pethidine]) requirement by patient-controlled analgesia was 275 mg (range, 75 to 800 mg). This compared favorably with an age- and sex-matched historic group of patients who underwent posterolateral thoracotomy and had a hospital stay of 10.4 days (statistically non-significant) and narcotic requirement of 950 mg (statistically significant by paired t test). We conclude that VAT anatomic lung resection is technically feasible. However, there are some specific complications associated with major lung resection through minimal access. Refinement of our present technique and attention to details are important to improve our results.