The treatment of patients with a solitary brain metastasis has been evolving, with most centers recommending resection in patients with good performance status. To evaluate the results of resection of brain metastases from non-small-cell lung cancer, we reviewed our 16-year experience with 185 consecutive patients undergoing resection of brain metastases from 1974 to 1989, inclusive. There were 89 men and 96 women; ages ranged from 34 to 75 years (median 54). Sixty-five (35%) had synchronous and 120 (65%) metachronous brain metastases. Discounting the brain metastasis, 68 patients (37%) had stage I, 13 (7%) stage II, 62 (33%) stage IIIA, 30 (16%) stage IIIB, and 12 (6%) stage IV carcinoma. There was no significant difference in age, locoregional stage (TN), or histologic features in patients with synchronous versus metachronous lesions. The overall survival rates (n = 185) were as follows: 1 year, 55%; 2 years, 27%; 3 years, 18%; 5 years, 13%; and 10 years, 7% (median 14 months). There was no significant difference in survival between patients with synchronous and metachronous lesions. To evaluate the impact of locoregional stage and treatment of the primary site, we analyzed only those patients with synchronous brain metastases. Multivariate analysis demonstrated that locoregional stage had no significant effect on survival (p = 0.97), but complete resection of the primary disease significantly prolonged survival (p = 0.002). Therefore complete resection, and not stage, of the locoregional primary lesion is the primary determinant of survival in patients undergoing resection of brain metastases from non-small-cell lung cancer.