BACKGROUND Patients with metastatic (stage IV) non-small-cell lung cancer usually have a poor prognosis and disease refractory to chemotherapy. Three new agents--taxol, merbarone, and piroxantrone--have shown promising antitumor treatment in vitro and in animals. Taxol is an antimicrotubular agent that interferes with mitosis during cell division. Merbarone, a conjugate of thiobarbituric acid and aniline, is a topoisomerase II inhibitor, which thus inhibits DNA synthesis and tumor growth. Piroxantrone, an anthracenedione derivative, is a DNA intercalating agent that has shown potent antitumor activity in animal studies. OBJECTIVE Our randomized phase II study was designed to evaluate the efficacy and toxicity of these agents in the treatment of stage IV metastatic non-small-cell lung cancer. METHODS Eligible patients (119) were randomly assigned to receive one of the three treatments given every 3 weeks: 250 mg/m2 taxol by a 24-hour intravenous infusion, 1000 mg/m2 merbarone by continuous intravenous infusion through a central catheter daily for 5 days, or 150 mg/m2 piroxantrone by intravenous infusion over 1 hour. Patients had received no chemotherapy. Response and toxicity were evaluated every 3 weeks. RESULTS Twenty-five patients were randomly assigned to receive taxol, 47 to receive merbarone, and 47 to receive piroxantrone. One of 44 assessable patients (2.3%) treated with piroxantrone had a complete response. Rates for partial response were 20.8% (five of 24 patients) and 5.7% (two of 35) for assessable patients treated with taxol or merbarone, respectively. One-year survival rates were 41.7%, 21.6%, and 22.6%, and median survival times were 24.1, 19.9, and 29.3 weeks for taxol, merbarone, and piroxantrone, respectively. These differences were not statistically significant, but this study was not designed to compare survival. In general, toxicity was manageable. With premedication, no anaphylaxis was observed with taxol. The most common toxic effects were leukopenia with taxol or piroxantrone treatment and thromboembolic complications with merbarone. Death directly related to treatment occurred in 4% (one patient), 11.4% (four), and 5% (two) of the assessable patients receiving taxol, merbarone, and piroxantrone, respectively. Cardiotoxicity and neurotoxicity occurred only occasionally in all three arms. CONCLUSIONS On the basis of the response rate (20.8% partial response) and 1-year survival rate (41.7%), taxol is an active agent for the treatment of metastatic non-small-cell lung cancer. Merbarone and piroxantrone are relatively inactive. CONCLUSIONS Further study of taxol is warranted. In future studies, taxol should be combined with other agents, and granulocyte colony-stimulating factor should be used to ameliorate myelosuppression.