To evaluate whether two commonly used newer platinum-based regimens offer any advantage over vinorelbine-cisplatin (reference regimen) in response rate for patients with advanced non-small-cell lung cancer (NSCLC). Chemotherapy-naive patients were randomized to receive gemcitabine 1,250 mg/m(2) days 1 and 8 plus cisplatin 75 mg/m(2) day 2 every 21 days (GC arm), or paclitaxel 225 mg/m(2) (3-hour infusion) then carboplatin (area under the concentration-time curve of 6 mg/mL x min), both on day 1 every 21 days (PCb arm), or vinorelbine 25 mg/m(2)/wk for 12 weeks then every other week plus cisplatin 100 mg/m(2) day 1 every 28 days (VC arm). Six hundred twelve patients were randomized to treatment (205 GC, 204 PCb, and 203 VC). Overall response rates for the GC (30%) and PCb (32%) arms were not significantly different from that of the VC arm (30%). There were no differences in overall survival, time to disease progression, or time to treatment failure. Median survival for the GC, PCb, and VC groups was 9.8, 9.9, and 9.5 months, respectively. Neutropenia was significantly higher on the VC arm (GC 17% or PCb 35% v VC 43% of cycles, P <.001), as was thrombocytopenia on the GC arm (GC 16% v VC 0.1% of cycles, P <.001). Alopecia and peripheral neurotoxicity were most common on the PCb arm, as was nausea/vomiting on the VC arm (P <.05). Efficacy end points were not significantly different between experimental and reference arms, although toxicities showed differences. These findings suggest that chemotherapy in NSCLC has reached a therapeutic plateau.