HD WLE guided biopsy confirmed diagnosis of GITB in 28.12% (P, 0.04). The two sets of biopsies together confirmed diagnosis of GITB in 53.1%. The area under curve for NBI-M plus HDWLE, NBI-M alone and HDWLE alone were 0.77(0.63-0.87) 0.73 (0.60-0.85) and 0.64 (0.50-0.77) respectively. Patients were started on anti-tubercular therapy for nine months. Four patients underwent surgery for intestinal obstruction while on antitubercular therapy. Twenty eight completed full therapy and improved. Conclusion: NBI-M guided biopsy improved the yield of histology for diagnosis of GITB. Endoscopic and computed tomography abdomen findings of 32 patients with GITB High definition white light endoscopic image and corresponding narrow band imaging image of a patient with colonic tuberculosis Sa2042 MEASUREMENT OF PANCREATIC SIZE IN A POPULATION-BASED COHORT: COMPARISON OF ULTRASOUND AND MAGNETIC RESONANCE IMAGING Ali Aghdassi, Birgit Schauer, David Duscha, Till Ittermann, Tilman Pickartz, Jan Hartmann, Christoph Budde, Marie Kromrey, Henry Völzke, Jens Kühn, Markus M. Lerch, Peter Simon Background: Transabdominal ultrasound (US) und magnetic resonance imaging (MRI) are two important imaging methods for abdominal organs. Despite its broad availability ultrasonography has limitations as it is user-dependent, and accessability is often impaired by meteorism and obesity. An exact measurement of the size of the pancreas helps for correct diagnosis of various pancreatic disorders. Aims: The intention of this study was to determine comparability of these two imaging methods in a population-based cohort without known pancreatic abnormality or disease. Methods: In the prospective Study of Health in Pomerania (SHIP) in Germany, 342 participants without any known pancreatic disorder underwent whole-body MRI at 1.5 T and transabdominal ultrasound on the same day. Sizes of the pancreatic head, body and tail were measured. MR images were analysed by a single examiner under supervision of an experienced radiologist, and ultrasound examiners were trained for sonography of the abdomen. Agreement between ultrasound and MR measure- ments were assessed by Bland-Altman Plots. Distribution of ultrasound measurements per examiner was visualized by boxplots. A multivariate regression model was applied using the independent variables age, sex and body mass index (BMI). Results: Abdominal ultrasound underestimated the size of each part of the pancreas, however with a high level of inconsis- tency between these two methods. Mean difference was 0.39 cm, 0.18 cm and 0.54 cm for the head, body and tail, respectively. There was a high interobserver variability in ultrasound with intraclass coefficients for US examiner of 0.25 (95% CI: 0.11 - 0.47) for the head, 0.150 (95% CI: 0.054 - 0.36) for the body and 0.30 (95% CI: 0.13 - 0.54) for the tail. After excluding the random examiner effect, mean difference decreased to 0.31 (11.8% of MRI), 0.12 (9.3%) and 0.45 cm (24.4%) for the head, body and tail, respectively. Multivariate analysis showed that pancreatic size of all three parts increased with BMI in both genders whereas pancreatic head and tail size decreased with age, which was more marked in females. Conclusions: Concordance of pancreatic size measurements is poor between US and MRI. Furthermore, pancreatic size measured by MRI varied with age, gender and BMI. These limitations should be considered when evaluating morphologic features for pathologic condi- tions or setting limits of normal size. Adjustments for BMI and age may also be warranted. Sa2043 PERORAL CHOLANGIOPANCREATOSCOPY WITH SPYGLASS® SYSTEM: A PORTUGUESE EXPERIENCE Pedro Pereira, Armando Peixoto, Patrícia Andrade, Guilherme Macedo Background Since the introduction of the direct view of the SpyGlass® System (Boston Scientific Corp., Natick, MA, USA) to cholangiopancreatoscopy, it has been shown in small series to be an effective therapeutic and diagnostic tool, and it is reasonable to think that this technique will become universally adopted for the evaluation and treatment of complex biliary and pancreatic tract diseases. Aims and Methods Our aim is to present a retrospective S-429 AGA Abstracts analysis of patients undergoing cholangiopancreatoscopy with SpyGlass® system between since its availability in a european referral center in 2015. Biliary and pancreatic strictures were considered indeterminate if it remained without diagnosis after conventional ERCP with negative or inconclusive cytology and/or biopsy. Lithiasis was considered complex after two unsuccessful attempts to remove the stones with ERCP. Cholangioscopy and pancreatoscopy were considered a technically successful if was possible to progress to the lesion or stone(s), with successful biopsies in case of undetermined strictures or success in early fragmentation or removal of calculi in cases of complex lithiasis. Results Sixteen procedures were performed, including 14 cholangioscopies and 2 pancreatoscopies. Indica- tions included indeterminate biliary or pancreatic strictures (n=13), difficult biliary stones (n=2) and pancreatic stones (n=1). The technical success rate was 93.75%. In 85% of cases with indeterminate strictures biopsies were done using the SpyBite® system. Histological evaluation revealed benign findings in five of the thirteen strictures and malignancy in one case. In other five patients findings were considered inconclusive, resulting in a 55% accuracy. Comparing the malignant diagnostic impression given by endoscopic features with histologi- cal findings, there was an 100% accuracy. Therapeutic success in complex lithiasis differed according to its origin. In both cases of lithiasis of the biliary tree it was possible to remove the stones after electro-hidraulic lithotripsy. By marked angulation it was not possible to treat the patient with pancreatic main duct stones. There were no complications during the procedure. In the post-procedure monitoring three patients developed acute pancreatitis (18.8%) and two patients developed acute cholangitis (12.5%) in both situations without severity criteria. Cholangiopancreatoscopy with the SpyGlass® system allowed a subsequent change in approach in 69% of cases. Conclusions By its specificity, SpyGlass® system is mainly used in the diagnosis of indeterminate biliary strictures and in the treatment of difficult-to-remove stones. The short experience in pancreatoscopy relates to the limitations that SpyGlass® still presents in exploring this organ, where the rate of diagnostic and technical failures are still considerable. Sa2044 EUS-GUIDED ENTERO-ENTEROSTOMY USING LUMEN-APPOSING METAL STENTS FOR THE MANAGEMENT OF AFFERENT LOOP SYNDROME: A MULTICENTER EXPERIENCE Olaya Isabella Brewer Gutierrez, Saowonee Ngamruengphong, Rastislav Kunda, Ali Siddiqui, Markus Dollhopf, Jose Nieto, Yen-I Chen, Majidah A. Bukhari, Omid Sanaei, Shayan S. Irani, Marcia I. Canto, Vikesh Singh, Mouen A. Khashab Background: Afferent loop syndrome (ALS) is defined as obstruction of the afferent limb after a gastrojejunostomy reconstruction. This condition generally requires surgical management. Endoscopic-assisted luminal stenting (EALS) across the obstruction can provide symptoms relief; however, access to the level of obstruction is not always feasible and placed stents are typically of small caliber. Recently, EUS guided entero-enterostomy (EUS-EE) using lumen-apposing metal stents (LAMS) for trans-enteric drainage of ALS has been described. Potential advantages of LAMS and transmural drainage include decrease in the risk of stent migration and improved drainage due to wider stent diameter. Nevertheless, the efficacy and safety of EUS-EE in patients with ALS is not well known. Aims: To assess the technical success, clinical success and safety of EUS-EE in patients with ALS. Methods: This is a multicenter, retrospective study at 4 US and 2 European tertiary centers. All consecutive adult patients referred for management of ALS by EUS-EE between 04/2012 and 08/2016 were included. Symptoms of ALS included abdominal pain, vomiting, jaundice and cholangitis. The primary outcome was clinical success, defined as resolution of symptoms after the procedure. Secondary outcomes included technical success, procedural time, symptoms recurrence and adverse events (severity graded according to the ASGE lexicon). Results: A total of 18 patients (11 males (61 %), mean age 64.22 ± 1.03 yrs.) were included. Of these, 13 patients had pancreaticoduodenectomy, 2 partial gastrectomy with Roux-en-Y reconstruction, 1 Roux-en-Y hepaticojejunostomy, 1 partial gastrectomy with Billroth II, and 1 unspecified Roux-en-Y reconstruction. Six patients presented with jaundice, 5 with vomiting, 4 with abdominal pain, and 3 with cholangitis. A total of 11 patients had prior episode(s) of ALS. One patients had previous luminal stenting within the obstructed afferent loop,7 percutaneous transhepatic biliary drainage and 3 more than one treatment. Technical success was achieved in all (100%) patients with a mean procedure time of 29.67±25.48 min (Table I). A total of 3 (16.7%) adverse events occurred (all abdominal pain, conservatively managed), 2 of which were graded as mild and 1 as moderate per the ASGE lexicon. Clinical success was achieved in 16 (88.9%) patients. Length of hospital stay was 7.22±7.51 days. During a median follow-up duration after EUS-GE of 120.5 days (range 10-449 days), 2 patients experienced symptoms recurrence: one patient treated with placement of a second LAMS and one with venting PEG (Table II). Conclusion: EUS-EE is a feasible and safe technique for the management of ALS. Comparative studies with endoscopic luminal stenting are needed to determine the optimal endoscopic technique for the management of this challenging disease. AGA Abstracts