BACKGROUND: There is no consensus on the optimal treatment for patients with advanced non-small cell lung cancer and stable disease after cisplatin-based chemotherapy. The objective of the trial was to evaluate a switch to a different dual-agent chemotherapy. METHODS: Patients with stage IV non-small cell lung cancer and stable disease after two cycles of cisplatin (P) and gemcitabine (G) (P day1 (d(1)): 75 mg/m(2), G: 1250 mg/m(2) d(1) and d(8) every 3 weeks) were randomized to receive either two further cycles of PG (arm A) or paclitaxel (100 mg/m(2) d(1), d(8), d(15)) plus gemcitabine (1250 mg/m(2) d(1) and d(8), every 4 weeks) (arm B). RESULTS: Two-hundred-twenty-eight patients were enrolled between October 2003 and August 2006. After two cycles of PG, 98 patients (43%) had stable disease; 87 were randomized: 45 to arm A and 42 to arm B. The objective response rates were 15.6% (6.5-29.4) and 21.4% (10.3-36.8) in arms A and B. Overall survival after randomization was 9.6 months (7.0-13.8) in arm A and 9.3 months (7.4-13.3) in arm B. Adverse events were similar in the two arms for hematological and non hematological toxicities. CONCLUSIONS: Sequential first-line chemotherapy in these patients is feasible with no difference in response rates. These results do not warrant a phase III trial.