Survival in patients with locally advanced (stage III Mo) and metastatic (Ml) non-small-cell lung cancer (NSCLC) is short. Phase II studies have reported objective responses ranging from 20% to 60% using cisplatin-based chemotherapeutic regimens, yet few have shown improvement in median survival. In our phase II pilot studies with cisplatin (CDDP) and etoposide (VP-16), we observed a 26% response rate; with CDDP, VP-16, and mitomycin-C, a 38% response rate was obtained in advanced NSCLC patients. A total of 156 consecutive patients with locally advanced and metastatic NSCLC were randomized to one of three treatment arms to determine whether the chemotherapy protocols had any effect on response rate and median survival in a large, randomized study. Arm 1 consisted of CDDP (120 mg/m2 x 3 weeks); arm 2, of CDDP (120 mg/m2) and VP-16 (100 mg/m2 given i.v. on days 1-3), repeated every 3 weeks; and arm 3, of CDDP (120 mg/m2) and VP-16 (100 mg/m2 on days 1-3) given every 3 weeks, plus mitomycin C (10 mg/m2 on days 1, 21, and 42, then every 6 weeks, for a maximal dose of 100 mg). After 71 patients had been enrolled in the study, we stopped accrual in the CDDP arm due to a lack of response [1 complete response (CR) in 24 patients; 4%] and continued enrollment in the two combination-chemotherapy arms. In the CDDP/VP-16 arm a 30% response rate [1 CR, 18 partial responses (PRs)] was obtained, and in the CDDP/VP-16 mitomycin C arm a 26% response rate (4 CRs, 11 PRs) was seen among a total of 150 evaluable patients. Responses were observed in 31% of patients with favorable performance status (PS) (ECOG 0-1) vs 14% in patients with a poor PS (ECOG 2-3). Of patients with locally advanced disease (III Mo), 17 (33%) obtained an objective response, compared with 20 patients (20%) with metastatic disease. Median survival was 18 weeks in the CDDP arm, 35 weeks in the CDDP/VP-16 arm, and 37 weeks in the CDDP/VP-16/mitomycin C arm. The median survival in the multimodal chemotherapy arms was significantly greater than that obtained with CDDP alone. Toxicity was predominantly myelosuppression in the mitomycin C-containing arm (27%, wtto grade 3-4). Our study shows that combination chemotherapy using CDDP/VP-16 is active and safe in the treatment of advanced NSCLC patients with a good performance status. The addition of mitomycin C did not improve the therapeutic response.