PURPOSE: To confirm the promising phase II results of docetaxel monotherapy, this phase III trial was conducted of chemotherapy for patients with advanced non-small-cell lung cancer (NSCLC) who had previously failed platinum-containing chemotherapy. PATIENTS AND METHODS: A total of 373 patients were randomized to receive either docetaxel 100 mg/m(2) (D100) or 75 mg/m(2) (D75) versus a control regimen of vinorelbine or ifosfamide (V/I). The three treatment groups were well-balanced for key patient characteristics. RESULTS: Overall response rates were 10.8% with D100 and 6.7% with D75, each significantly higher than the 0.8% response with V/I (P =.001 and P =.036, respectively). Patients who received docetaxel had a longer time to progression (P =.046, by log-rank test) and a greater progression-free survival at 26 weeks (P =.005, by chi(2) test). Although overall survival was not significantly different between the three groups, the 1-year survival was significantly greater with D75 than with the control treatment (32% v 19%; P =.025, by chi(2) test). Prior exposure to paclitaxel did not decrease the likelihood of response to docetaxel, nor did it impact survival. There was a trend toward greater efficacy in patients whose disease was platinum-resistant rather than platinum-refractory and in patients with performance status of 0 or 1 versus 2. Toxicity was greatest with D100, but the D75 arm was well-tolerated. CONCLUSION: This first randomized trial in this setting demonstrates that D75 every 3 weeks can offer clinically meaningful benefit to patients with advanced NSCLC whose disease has relapsed or progressed after platinum-based chemotherapy.