7089 Background: Trials testing chemotherapy for elderly pts with SCLC are scanty. Gemcitabine (Gem) is active and well tolerated. The G-STEP program looked for optimal 2-drug combination of G with either vinorelbine (Vin), etoposide (Eto), cisplatin (Cis), or carboplatin (Car). Methods: Extended stage SCLC pts, aged >70 years, PS 0–2, were eligible. Dose of G was 1,000 mg/m2, dd 1&8, every 3 weeks in all the four combinations. As safety data in the elderly were already available for Gem+Vin (Vin dose 25 mg/m2, dd 1&8), a two-stage minimax flexible design was applied with response as end-point: ≥13 responses/30 pts required at 1st stage, with p0=0.40, p1=0.60, α and β=0.10. For Gem+Car, Gem+Cis, Gem+Eto a phase 1–2 Bayesian design to select optimal dose was applied (Thall & Russell, Biometrics, 1998), with 3 possible outcomes for each patient: “active” (response and no unacceptable toxicity [UT]), “inactive” (no response, no UT), or “toxic” (UT independently of response). A response rate [RR] ≥60% and a rate of UT ≤25% were acceptable. Dose levels to explore were: Car: AUC 3.5 / 4 / 4.5, d1; Cis: 50 / 60 / 70 mg/m2, d1; Eto: 60 / 70 / 80 mg/m2, dd 1,2,3. Results: From May 2000 to September 2005, 78 eligible pts were enrolled; median age was 74 years (42% of pts older than 75yrs). In the whole group, median progression-free survival and overall survival were 20.3 weeks (95% CI 17.6 - 24.1) and 33.7 weeks (95% CI 23.7 - 41.6), respectively. Study of Gem+Vin was closed for inactivity after 1st stage with 11 responses / 30 pts (RR 36.7%, 95% exact CI: 19.9–56.1). Gem+Eto arm was closed after 10 pts (5 inactive and 5 toxic) because too high probability (0.994) of inactivity. Gem+Cis arm was closed after 12 pts (2 active, 5 inactive, 5 toxic) because too high probability (0.988) of inactivity. With Gem+Car (December 2005: 2 pts ongoing, 24 analysed: 12 active, 6, inactive and 6 toxic), RR in the 20 pts receiving Car at AUC 3.5 was 60% (95% exact CI: 36.1–80.9). Conclusions: The combination of gemcitabine and carboplatin seems promising for future trials in elderly patients with extended SCLC. The G-STEP program was supported by GIOGER. No significant financial relationships to disclose.
Read full abstract