A phase III randomised trial comparing sequential to standard chemotherapy in advanced non-small cell lung cancer (NSCLC)

7012 Background: Cisplatin-based chemotherapy and taxans are effective treatments for advanced NSCLC. We performed a phase III randomised trial to determine if the sequential administration of cisplatin-based chemotherapy followed by paclitaxel is superior to a cisplatin-based standard chemotherapy, with the use of paclitaxel as salvage treatment. Methods: Untreated advanced NSCLC with adequate PS, hematological, hepatic, cardiac and renal functions were treated by 3 courses of GIP (cisplatin 50 mg/m2, ifosfamide 3 g/m2, gemcitabine 1 g/m2). Patients with a non-progressing tumour were randomised between 3 further courses of GIP or 3 courses of paclitaxel (225 mg/m2). To detect an increase in the survival rates (primary endpoint) from 20% in the GIP arm to 35% in the sequential arm, with 80% probability using a two-sided logrank test with a significance level of 5%, we needed to observe 178 events, requiring to randomise 123 patients in each arm. Results: From January 2000 to February 2004, 485 patients received 3 courses of induction GIP of which 140 were randomised in the GIP arm and 141 in the paclitaxel arm. Median survival times were 14.1 (95% CI: 12.0–16.3) and 16.4 (95% CI: 14.0–18.8) months for the paclitaxel and the GIP arms, respectively (p = 0.17). When treatment comparison was adjusted for the two independent prognostic factors (sex and haemoglobin) revealed by a Cox multivariate analysis, the observed HR was 0.81 (95% CI: 0.63–1.09) in favour of the GIP arm (p = 0.10). There were more grades III/IV thrombopenia with GIP (p< 0.01) and more alopecia with paclitaxel (p = 0.04). Conclusion: Sequential cisplatin-based chemotherapy by paclitaxel does not result in better survival than standard chemotherapy, with the use of paclitaxel as salvage treatment. No significant financial relationships to disclose.