Randomized phase III comparison of paclitaxel/carboplatin/etoposide versus gemcitabine/irinotecan, both followed by gefitinib, in patients (pts) with carcinoma of unknown primary site (CUP)

Abstract

4631 Background: Empiric chemotherapy for pts with CUP has resulted in modest survival improvements. Several regimens have been active in phase II trials, but no randomized comparison of newer regimens have been completed. In this multicenter randomized phase III trial, we compared the efficacy and toxicity of 2 active regimens. Methods: Previously untreated pts with CUP (adenocarcinoma, poorly differentiated adenocarcinoma, poorly differentiated carcinoma, poorly differentiated squamous carcinoma) were eligible. Pts with specific treatable syndromes were excluded. Additional eligibility: ECOG PS 0–2; controlled brain metastases; adequate organ function. Pts were randomized (1:1) to paclitaxel 200mg/m2 day 1/carboplatin AUC 6.0 day 1/etoposide 50mg/100mg alternating po days 1–10 (PCE) or gemcitabine 1000mg IV days 1, 8/irinotecan 100mg/m2 IV days 1, 8 (GI). Both regimens were repeated at 21-day intervals for 4–6 courses. Responding/stable pts then received gefitinib 250mg po qd until tumor progression. The primary endpoint was the 2-year survival rate. Initially, a total of 320 pts were planned, to allow detection of a 50% improvement in 2-year survival (from 20% to 30%). However, due to slow accrual, enrollment was stopped after 198 pts were randomized. Results: Between September 2003 and July 2008, 198 pts were randomized (PCE, 93 patients; GI 105 patients). Pt characteristics were similar in both groups. Median progression-free survival for PCE versus GI was 3.2 months versus 5.3 months, p=0.19. Median overall survivals were 7.4 months (PCE) versus 8.6 months (GI), p=0.34; 2-year survivals were 16% (PCE) and 19% (GI). Response rates were similar (PCE 19%, GI 20%). GI was less toxic, with lower rates of grade 3/4 neutropenia (11% vs. 35%; p< 0.01), febrile neutropenia (0% vs. 9%; p<0.01), thrombocytopenia (3% vs. 8%; p=.05), anemia (3% vs. 9%; p=0.05), and RBC transfusions (10% vs. 24%; p<0.01). Conclusions: The PCE and GI regimens had comparable efficacy in the treatment of CUP, while the GI regimen was better tolerated. The 2-year survival was similar (16%, 19%). Better treatments are needed for pts with CUP. [Table: see text]