TY - JOUR T1 - S85 British Thoracic Oncology Group Trial, BTOG2: Randomised phase III clinical trial of gemcitabine combined with cisplatin 50 mg/m<sup>2</sup> (GC50) vs cisplatin 80 mg/m<sup>2</sup> (GC80) vs carboplatin AUC 6 (GCb6) in advanced NSCLC JF - Thorax JO - Thorax SP - A41 LP - A41 DO - 10.1136/thoraxjnl-2011-201054b.85 VL - 66 IS - Suppl 4 AU - D Ferry AU - L J Billingham AU - H W Jarrett AU - D Dunlop AU - J Thompson AU - M Kumar AU - G Skailes AU - M Nicolson AU - R Shah AU - P Leonard AU - A Chetiyawardana AU - P Wells AU - C Lewanski AU - P Woll AU - B Crosse AU - M Hill AU - S Pirrie AU - K J O'Byrne Y1 - 2011/12/01 UR - http://thorax.bmj.com/content/66/Suppl_4/A41.1.abstract N2 - Background Platins are considered key drugs in treating advanced NSCLC. Carboplatin has been reported as inferior to cisplatin in meta-analyses while the optimal dose of cisplatin is unclear.Methods Eligibility was by histologically proven NSCLC, PS 0–2, stage IIIB/IV disease and a GFR of &gt;60 ml/min, using the Wright equation. Treatment was gemcitabine (1250 mg/m2) combined with cisplatin 50 mg/m2, cisplatin 80 mg/m2 or carboplatin AUC 6, for up to four cycles. Carboplatin dosing was calculated using the Calvert equation. At the time of analysis 1223 deaths had been reported, allowing analysis according to the statistical plan.Results 1363 patients were randomised between April 2005 and November 2009. Trial arms were well balanced for PS, stage and age. Median age was 63, 32% were PS0, 60% PS1 and 8% PS2. The median delivered dose intensities (planned=100%) for platinum were GC50 99%, GC80 96% and GCb6 87%, for gemcitabine were 95%, 88% and 80% respectively. During treatment the proportion of patients experiencing at least one grade 3/4 AE were; GC50 27%, GC80 41% and GCb6 57%. At analysis 140 patients were alive and median follow-up was 21 months. Response rates were significantly different between arms; GC50 23%, GC80 33% and GCb6 28% (p=0.01). Median survival was: GC50 8.3 months, GC80 9.5 months and GCb6 10.0 months, with the GC50 arm statistically identified as differing from the other two. For subsequent primary comparisons of non-inferiority of GC50 v GC80 (HR=1.11) and GCb6 v GC80 (HR=0.96), the 95% CI for the cisplatin dose comparison (0.96, 1.27) did not exclude the pre-defined inferiority region of HR&gt;1.2 whereas the 95% CI for the GCb6 v GC80 comparison (0.84, 1.10) fell well below this inferiority region.Conclusion In advanced NSCLC the dose of cisplatin is important with GC50 giving the poorest outcome in terms of overall survival and response rate. GCb6 is not inferior to GC80 thus, in combination with gemcitabine, and in relation to survival time, carboplatin is clinically equivalent to that of cisplatin but other factors, such as quality of life, may influence treatment choice. ER -