Sixty patients with inoperable non-small-cell lung cancer (NSCLC) were entered into a phase II study that tested the combination of cisplatin (80 mg/m2, day, etoposide intravenously (IV) (100 mg, days 1 and etoposide orally (200 mg/m2, days 3 and 5). The regimen was repeated every 28 days for six courses, after which patients were allowed to receive additional treatment at the discretion of their physician. Overall objective response rate in 51 evaluable patients was 69% (95% confidence interval: range, 56% to 81%), with 16% sustaining complete remission (CR), 53% partial remission (PR), 17% stable disease (SD), and 14% progressive disease (PD). CR was pathologically confirmed by bronchoscopy and biopsy. One patient with a clinical PR underwent surgery and was shown to have a pathologic CR. Median survival of all evaluable patients was 52 weeks, greater than 75 weeks for CR patients, 52 weeks for PR patients, 42 weeks for SD patients, and 13 weeks for PD patients. Eleven patients (21.5%) developed CNS metastases, which resulted in the deaths of ten. Survival was significantly correlated with extent of disease, performance status, and albumin level, but not with histology or weight loss. Tumor response was significantly correlated only with histology (squamous-cell and large-cell undifferentiated carcinoma greater than adenocarcinoma). Side effects were nausea, vomiting, anorexia, alopecia, bone marrow suppression, and nephrotoxicity. One patient died from leukopenia and sepsis. Pharmacokinetic studies in ten patients showed the continuous presence of etoposide in plasma for six days at a level of at least 220 to 480 ng/mL. In order to investigate whether this very effective combination of cisplatin and etoposide can prolong survival in NSCLC, it will be tested as preoperative chemotherapy in a randomized trial in operable patients with T1N1 and T2N0-1 disease.