S460 Abstracts 2370 POSTER Relative risk analysis of safety profile of lanreotide autogel/depot vs. placebo in patients with pancreatic and intestinal neuroendocrine tumours A. Phan 1 , M. Caplin 2 , M. Pavel 3 , J. Cwikła 4 , M. Raderer 5 , E. Sedl´ aˇ ckov ´ a 6 , G. Cadiot 7 , E. Wolin 8 , J. Capdevila 9 , L. Wall 10 , G. Rindi 11 , A. Langley 12 , E. Gomez-Panzani 13 , P. Ruszniewski 14 . 1 Houston Methodist Hospital, Dept of Gastrointestinal Medical Oncology, Houston TX, USA; 2 Royal Free Hospital, Dept of Gastroenterology and Hepatobiliary Medicine, London, United Kingdom; 3 Charit ´ e University Medicine, Dept of Gastroenterology and Hepatology, Berlin, Germany; 4 University of Varmia and Masuria, Department of Nuclear Medicine − Faculty of Medical Science, Olsztyn, Poland; 5 University Hospital, Clinical Division of Oncology − Dept of Medicine I, Vienna, Austria; 6 First Faculty of Medicine and General Teaching Hospital, Institute of Medical Biochemistry and Laboratory Medicine, Prague, Czech Republic; 7 Robert-Debr ´ e Hospital, Dept of Hepato-Gastroenterology and Digestive Oncology, Reims, France; 8 Markey Cancer Center, University of Kentucky, Lexington, KY, USA; 9 Vall d’Hebron University Hospital, Vall d’Hebron Institute of Oncology, Barcelona, Spain; 10 Western General Hospital, Edinburgh Cancer Centre, Edinburgh, United Kingdom; 11 Universit ` a Cattolica del Sacro Cuore, Institute of Anatomic Pathology, Rome, Italy; 12 Ipsen, Dept of Biometry, Les Ulis, France; 13 Ipsen, Dept of Global Drug Development, Les Ulis, France; 14 Beaujon Hospital, Dept of Gastroenterology-Pancreatology, Clichy, France Background: In the CLARINET study (NCT00353496), progression- free survival (PFS) of patients with metastatic pancreatic and intestinal neuroendocrine tumours (NETs) was significantly increased with lanreotide Autogel (Depot in USA) 120 mg vs placebo (hazard ratio [HR] for progressive disease [PD]/death 0.47 [95% confidence interval CI: 0.30, 0.73]). We now present detailed analysis of the safety and tolerability data from CLARINET evaluating the relative risk of adverse events. Methods: Patients with metastatic grade 1/2 (Ki-67 <10%) non-functioning pancreatic and intestinal NETs received lanreotide Autogel 120 mg (n = 101) or placebo (n = 103) for 96 weeks or until death/disease progression. The relative risks and 95% CIs of any adverse event occurring in 5% of patients in either group were calculated post hoc for lanreotide vs. placebo. Results: The overall incidence of adverse events was similar for lanreotide- and placebo-treated patients (88% vs 90%). The most common adverse events were gastrointestinal disorders, which occurred in 67% of the lanreotide-treated group vs. 63% of the placebo-treated group (RR 1.1 [0.9, 1.3]). Of these, diarrhoea was the most common individual adverse event (35% vs 35%; RR 1.0 [0.7, 1.4]). None of the adverse events occurring in 5% of either group were statistically significantly different between the lanreotide and placebo treated patients when based on the RRs although the CIs for the RRs were wide in several cases (lower limit of CIs were 1). The most frequent adverse events in either group are shown in the Table. Conclusions: No new safety signals for lanreotide were identified in the CLARINET study supporting its favourable benefit–risk profile in patients with pancreatic and intestinal NETs. Table: RRs for most frequent AEs (occurring in 10% in either group) Adverse event Lanreotide, n (%) (N = 101) Placebo, n (%) (N = 103) RR (95% CI) Gastrointestinal Diarrhea 35 (34.7) 36 (35.0) 1.0 (0.7, 1.4) Abdominal pain 24 (23.8) 17 (16.5) 1.4 (0.8, 2.5) Vomiting 19 (18.8) 9 (8.7) 2.2 (1.0, 4.5) Nausea 14 (13.9) 14 (13.6) 1.0 (0.5, 2.0) Constipation 12 (11.9) 13 (12.6) 0.9 (0.5, 2.0) Flatulence 12 (11.9) 9 (8.7) 1.4 (0.6, 3.1) Musculoskeletal Headache 16 (15.8) 11 (10.7) 1.5 (0.7, 3.0) Back pain 12 (11.9) 11 (10.7) 1.1 (0.5, 2.4) Hepatobiliary Cholelithiasis 14 (13.9) 7 (6.8) 2.0 (0.9, 4.8) Vascular Hypertension 13 (12.9) 5 (4.9) 2.7 (1.0, 7.2) General and administration site Fatigue 10 (9.9) 15 (14.6) 0.7 (0.3, 1.4) Infections and infestations Nasopharyngitis 9 (8.9) 16 (15.5) 0.6 (0.3, 1.2) For significant difference between treatment groups, lower limit of CIs would need to be >1. Conflict of interest: Ownership: Authors Langley and Gomez-Panzani are employees of Ipsen, France. 2371 POSTER Efficacy and safety of 2nd line treatment for advanced esophagogastric cancer (AEGC): A network meta-analysis E. Ter Veer 1 , N. Haj Mohammad 1 , G. Van Valkenhoef 2 , L.L. Ngai 1 , R. Mali 1 , M. Van Oijen 1 , H.W.M. Van Laarhoven 1 . 1 Acadamic Medical Centre, Medical Oncology, Amsterdam, Netherlands; 2 University Medical Centre Groningen, Epidemiology, Groningen, Netherlands Background: Previous meta-analyses have shown that chemotherapy (CT) or targeted therapy(TT) may provide survival benefit compared to best supportive care (BSC) in 2 nd Line treatment of AEGC. However, the efficacy of several single agents or combined treatment strategies have not yet been mutually compared. Therefore, we conducted a systematic review using network meta-analysis (NMA) of randomized controlled trials (RCT) to investigate the efficacy and safety of 2 nd Line therapy for AEGC. Method: Using Cochrane methodology, MEDLINE, EMBASE, CENTRAL databases, and ASCO and ESMO abstracts were searched up to March 2015 for RCTs studying CT and/or TT-based 2 nd Line therapy. Hazard Ratios (HR) were extracted for Overall Survival (OS) and event counts for grade 3−4 adverse events were compared using Risk Ratio’s. Using WinBUGS, Bayesian pair-wise and NMAs with random effects models were conducted based on intention to treat to make direct, indirect and combined comparisons between therapies and presented as HRs with 95% credibility intervals. Results: Fourteen RCTs (3166 patients) were identified and analysed in a 5-node network including BSC, single agent CT (i.e. paclitaxel, docetaxel, irinotecan), single agent TT (i.e. ramucirumab, apatinib), two-chemotherapy combination (2-CT) and ramucirumab+paclitaxel (RAM+PTX). OS was significantly prolonged with CT (combined HR = 0.64, 0.49–0.81) and TT (combined HR 0.77, 0.59–0.93) vs BSC and with RAM+PTX vs CT (direct HR = 0.82, 0.70–0.97). Compared to CT alone, 2-CT (combined HR = 0.97, 0.78–1.19) and TT (combined HR = 1.22, 0.88–1.59) showed no survival benefit. Indirect estimates showed OS benefit for RAM+PTX (indirect HR = 0.53, 0.37–0.77) and 2-CT vs BSC (indirect HR = 0.62, 0.44–0.84) but not for RAM+PTX vs TT (indirect HR = 0.67, 0.48–1.08) although a non- significant HR <0.70 was estimated, which can be considered as clinically meaningful. RAM+PTX had 80% probability to be the most efficacious treatment. For RAM+PTX vs CT a significant increase in grade 3−4 neutropenia (41% vs 19%), leukopenia (17% vs 7%), fatigue (12% vs 5%), neuropathy (8% vs 5%), abdominal pain (6% vs 3%), gastro-intestinal haemorrhage (4% vs 2%) and hypertension (15% vs 3%) were detected. Leukopenia was increased in 2-CT vs CT (19% vs 11%). Conclusion: RAM+PTX is more effective compared to CT or TT as single agents, although toxicity was increased. 2-CT has no survival benefit and was associated with increased toxicity compared to CT in the 2 nd line treatment of AEGC. No conflict of interest. 2372 POSTER Is there a role for palliative gastrectomy in asymptomatic metastatic gastric cancer? G. Musettini 1 , C. Caparello 1 , G. Pasquini 1 , C. Vivaldi 1 , M. Lencioni 2 , I. Petrini 1 , M.G. Fabrini 3 , G. Pallabazzer 4 , S. D’Imporzano 4 , B. Solito 4 , S. Santi 4 , L. Fornaro 1 , E. Vasile 1 , A. Falcone 1 . 1 U.O. Oncologia Medica 2 Universitaria, Azienda Ospedaliero-Universitaria Pisana, Pisa, Italy; 2 U.O. Oncologia Medica 1 Ospedaliera, Azienda Ospedaliero-Universitaria Pisana, Pisa, Italy; 3 U.O. Radioterapia, Azienda Ospedaliero-Universitaria Pisana, Pisa, Italy; 4 U.O. Chirurgia dell’ Esofago, Azienda Ospedaliero-Universitaria Pisana, Pisa, Italy Background: Surgery has an established role in gastric cancer for the palliation of major symptoms such as bleeding or obstruction. In other malignancies (like colorectal, kidney and breast cancer), resection of the primary tumor, even in presence of metastatic disease, appears to have a positive prognostic impact also in paucisymptomatic or asymptomatic patients and several randomized trials are ongoing Material and Methods: We retrospectively evaluated the data of patients treated from 2009 to 2014 at our Institution who fulfilled the following selection criteria: histologically confirmed gastric or gastroesophageal junction adenocarcinoma, locally advanced (unresectable) or metastatic disease and palliative treatment with at least one line of systemic chemotherapy. Primary end point of the analysis was overall survival (OS) from the start of first-line chemotherapy, estimated using the Kaplan–Meier method. Two-tailed log-rank test was used for survival comparison between groups and statistical significance was set at p < 0.05 Results: We identified 153 patients: five patients were excluded since they received palliative surgery because of symptoms referred to the primary