Phase II study of sorafenib with gemcitabine and erlotinib (GES) in first-line advanced pancreatic cancer.
266 Background: Addition of erlotinib (E) to gemcitabine (G) results in improved OS for patients (pts) with advanced pancreatic cancer. Many pancreatic tumors have constitutively activated ras/raf pathways and overexpress VEGF. Sorafenib (S), a multitargeted tyrosine kinase inhibitor, including VEGR 1-3, PDGFR-α and β and the RAF/MEK/ERK pathway, when combined with G and E may synergize, resulting in a more complete blockade of the signal transduction cascade in pancreatic cancer growth and progression, and improved outcome. Methods: Pts with previously untreated, histologically confirmed, unresectable pancreatic adenocarcinoma, PS 0-1, and adequate organ function were eligible and received G 1,000 mg/m2 over 30 min qw x 3 every 4w. E 150 mg PO daily and S 400 mg PO bid were given continuously. CT scans were performed every 2 cycles (8w). Primary endpoint included PFS at 4 mos and secondary endpoints included safety and tolerability of the novel combination, RR, and OS. Results: 45 pts enrolled, 44 evaluable for toxicity(1 not treated), and 30 evaluable for response. Median age was 64 (45-84), 32 males (71%), 43 (96%) had metastatic disease, PS was 0 in 26 (58%). Median number of cycles was 2 (0-10). Grade (gr) 3 toxicity included: thrombocytopenia 5; diarrhea 4; vomiting 4; HFS, hyperbilirubinemia, hyperglycemia and SOB each 3 pts; 1 bowel perforation and 1 epistaxis. Gr 4 toxicity included 1 each: bowel perforation, GI bleed, transaminitis, hyperglycemia and sepsis with hypotension. 2 patients only required dose reduction of S for HFS. There were 2 PR (7%), 13 SD (43%), overall RR of 7% and DCR of 50%. Median TTP 111 days (95% CI = 53-175) and median OS 195 days (144-290). PFS at 16 weeks was 49%, just reaching statistical significance. Conclusions: The combination of G and E plus S in the treatment of advanced pancreatic cancer is a well tolerated regimen without significant increased toxicity as compared to G alone, except for very manageable cutaneous reactions. While the primary endpoint met our pre-determined criteria (compared with NCIC PA.3), this study does not suggest a major benefit for addition of S to G-E when compared to published data. Supported in part by grants from Bayer Healthcare Pharmaceuticals/Onyx and OSI. No significant financial relationships to disclose.