autogel

Abstracts 2330 POSTER Value of neutrophil-lymphocyte ratio in predicting treatment response and disease free survival in patients with gastric cancer D. Romero1 , C. Ruiz2 , H. Martinez2 , F. Castro3 , R. Garcia2 , L. Basave4 , T. Duran3 , V. Sarmiento4 , P. Rosciano4 . 1 INCAN, Oncology, Distrito Federal, Mexico; 2 National Cancer Institute, Medical Oncology, Mexico, Mexico; 3 National Cancer Institute, Radio Oncology, Mexico, Mexico; 4 National Cancer Institute, Surgial Oncology, Mexico, Mexico Background: Several studies have shown that the neutrophil to lymphocyte ratio (NLR) in peripheral blood is a prognostic factor of various cancers. Some biomarkers have been associated with response and prognosis in gastric cancer (GC) identify depends specimens on biopsies and limits their use in clinical practice.Pheripheral blood tests at the time of diagnosis can reflect evolution within the tumor. We evaluated whether the NLR would predict response in patients with GC with treatment neoadyuvant or palliative. Methods: We retrospectively analyzed 190 patients with GC (2011– 2013) in National Cancer Institute Mexico. 73 patients were treated with chemotherapy neoadjuvant, 30 with chemoradiotherapy neoadjuvant and 117 patients palliative chemotherapy. NLR were calculated from complete blood counts in laboratory test at diagnosis, cut of values for the NLR were >2.5 and <2.5 using median values reported in previous studies. Results: Characteristics of Patients: male 107, female 83, mean age 54 (20−81).The median number of cycles of chemotherapy before of asseses response by tomography was 4. The response rate was partial 20.5%, stable disease 36.3% and progression 43.2%. The median NLR: 2.9 (1.1– 15.2).84 patients were detected with NLR <2.5 and 106 patients with NLR greater 2.5. Low NLR group patients had a better disease control (partial response and stable disease) in 68% vs. 47.1% than high NLR (p = 0.028). Progression of disease was reported in 32% (n = 27) with low NLR and 52.9% (n = 56) of group high NLR (P = 0.002). Patients with neoadjuvant treatment 50 underwent surgery 30 with NLR ratio lesser 2.5 and 20 high. There are more advanced disease in patients greater 2.5 NLR than <2.5 80 vs 56 (p = 0.036). Response pathological complete after surgery in 2 patients with a lower value and 1 patient with 2.5. Overall survival in low NLR is longer than in group patients NLR >2.5, 11.3 vs. 9.1 months (p = 0.49). Conclusions: These data give evidence that baseline NLR could be predictive marker of response and prognosis in patients with GC undergoing treatment. Limitations to our analysis are a small number of patients, short follow-up and will need to be stratified according to patient characteristics and treatment. No conflict of interest. 2331 POSTER Prognostic value of neutrophil/lymphocyte ratio in intestinal and pancreatic neuroendocrine tumors: exploratory analysis of data from the CLARINET trial of lanreotide depot/autogel T. Grenader1 , P. Ruszniewski2 , M. Pavel3 , J. Cwikła4 , A. Phan5 , M. Raderer6 , E. Sedlaˇ ´ ckova´ 7 , G. Cadiot8 , E. Wolin9 , J. Capdevila10 , L. Wall11 , G. Rindi12 , A. Lang13 , E. Gomez-Panzani14 , M. Caplin1 . 1 Royal Free Hospital, Dept. of Gastroenterology and Hepatobiliary Medicine, London, United Kingdom; 2 Beaujon Hospital, Dept. of GastroenterologyPancreatology, Clichy, France; 3 Charite´ 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 The Methodist Hospital, Dept. of Gastrointestinal Medical Oncology, Houston, TX, USA; 6 University Hospital, Clinical Division of Oncology, Dept of Medicine I, Vienna, Austria; 7 First Faculty of Medicine and General Teaching Hospital, Institute of Medical Biochemistry and Laboratory Medicine, Prague, Czech Republic; 8 Robert-Debre´ Hospita, Dept of Hepato-Gastorenterology and Digestive Oncology, Reims, France; 9 University of Kentucky, Markey Cancer Center, Lexington, KY, USA; 10 Vall d’Hebron University Hospital, Vall d’Hebron Institute of Oncology, Barcelona, Spain; 11 Western General Hospital, Edinburgh Cancer Centre, Edinburgh, United Kingdom; 12 Universita` Cattolica del Sacro Cuore, Institute of Anatomic Pathology, Rome, Italy; 13 Ipsen, Dept. of Biometry, Les Ulis, France; 14 Ipsen, Dept. of Global Drug Developmen, Les Ulis, France Background: The CLARINET trial demonstrated longer progression free survival (PFS) among patients with inoperable/metastatic intestinal and pancreatic NETs of grade 1 or 2 (Ki67 <10%) treated with somatostatin analog lanreotide as compared to patients treated with placebo. High NLR has been reported to be prognostic of poor outcomes in majority of solid tumors. The aim of the current analysis was to evaluate the prognostic value of NLR in intestinal and pancreatic NETs. Methods: A post-hoc exploratory analysis of CLARINET data was performed on all patients with available absolute neutrophil count (ANC) and absolute lymphocyte count (ALC) data. Progression-free survival (PFS)

S443 curves were generated for subgroups based on NLR value using the Kaplan–Meier method and raw survival rates were computed at 24 months. P values and hazard ratios for prognostic effects were obtained using the Cox proportional hazards model. Results: ANC and ALC were available in 201 patients: 100 patients received lanreotide and 101 patients received placebo. Baseline characteristics were balanced between the two treatment arms. The 24-month raw PFS rates across subgroups based on tertiles of the NLR distribution, irrespective of treatment, were comparable (37.3%, 38.8% and 38.8% with n = 67 per group). Likewise the 24 month PFS rate was 38.1% in patients with NLR 4 (n = 176) and 38.8% in patients with NLR >4 (n = 25). These findings were corroborated by the Cox model where NLR was not found to be prognostic (p > 0.6 for NLR effect irrespective of the NLR variable included). Additionally the therapeutic effect of lanreotide was independent of NLR value (p > 0.1). Conclusions: Contrary to most solid tumors, NLR seems to have no prognostic value in advanced intestinal and pancreatic NETs of grade 1−2 (Ki67 <10%). This finding, although based on relatively few patients with NLR >4 (n = 25 [vs. n = 176 for NLR 4]), might be explained by lack of inflammation affecting the tumor microenvironment in this cohort and the relatively slow rate of disease progression in the CLARINET trial. Tal Grenader was supported by an ESMO clinical Research Fellowship Grant. No conflict of interest.

2332 POSTER A simple prognostic scoring system for patients treated with sorafenib for advanced hepatocellular Carcinoma: The SAP score J. Edeline1 , J.F. Blanc2 , J. King3 , M. Chatterjee4 , J. Mathurin2 , S. Le ` Sourd1 , D. Palmer4 , T. Meyer5 . 1 Centre Eugene Marquis, Oncologie, ˆ ´ Bordeaux, Hepatology, Rennes, France; 2 CHU Hopital Saint-Andre, 3 Bordeaux, France; Royal Free Hospital, Oncology, London, United Kingdom; 4 Cancer Research UK Centre, University of Liverpool, Oncology, Liverpool, United Kingdom; 5 University College London, Oncology, London, United Kingdom Background: No prognostic classification is used for patients treated with sorafenib for advanced Hepatocellular Carcinoma (HCC). We aimed to create a simple score able to stratify patients according to their prognosis and to select patients with a low chance of benefit from sorafenib. Methods: We retrospectively analyzed data from patients treated with sorafenib for advanced HCC from 2 centers in France and 2 in the United Kingdom. Data from 2 centers were used as a training set, and data from the 2 other centers as a validation set. Continuous variables were dichotomized according to easy-to-use thresholds. Variables significantly associated with Overall Survival (OS) in the training set were used to build a new score, named the SAP (Sorafenib Advanced HCC Prognosis) score. The score was tested in the validation set, then compared with other prognostication systems in the overall cohort. OS was defined as the time from start of sorafenib to death, and was analyzed with a Cox regression model, the Kaplan–Meier method and a log-rank test. Results: The training and validation sets comprised 356 and 232 patients. Characteristics of the 2 sets were similar, with Barcelona Clinic for Liver Cancer (BCLC) class A in 1.1%, B in 16.8%, C in 80.2% and D in 1.9%. Child Pugh class was A in 81.5%, B in 16.9% and C in 1.6%. Extrahepatic spread was present in 34.0%, portal vein thrombosis in 40.9%, and performance status >0 in 49.3%. AFP was >400ng/mL in 40.0%, tumor size >7cm in 42.2%, bilirubin >17 mmol/L in 46.5%, and albumin <36g/dL in 44.3%. In the training set, variables independently associated with OS in multivariate model were: performance status >0, AFP >400ng/mL, tumor size >7cm, bilirubin >17 mmol/L and albumin <36 g/dL. The SAP score was built giving one point to each abnormal variable, and 3 classes were constructed: SAP A if 0−1 point, SAP B if 2−3 points, SAP C if 4−5 points. The SAP score was significantly associated with OS, with median OS of 14.7 (95% CI: 10.8–18.5), 7.3 (6.4−8.2) and 2.5 months (1.8−3.2) for patients with SAP A, B and C, respectively (p < 0.001). In the validation set, median OS was 18.0 (10.1–25.9), 9.1 (6.5–11.6) and 4.0 months (3.2−4.9), for SAP A, B and C, respectively (p < 0.001). Results of the worst class of different prognostication systems in the overall set are reported in the Table. Score

% of patients in worst class

Median OS (95% CI)

Cumulative survival at 6 months

Cumulative survival at 12 months

SAP C HAP D Albi 3 BCLC D

16.6% 23.9% 6.8% 1.8%

3.1 3.7 3.1 2.2

23.7% 28.1% 27.6% 11.4%

11.8% 15.0% 19.1% 0%

(2.4−3.7) (3.1−4.5) (1.7−4.5) (1.1−3.4)