Sa1112 Three-Dimensional Myoarchitecture of Porcine Gastro-Esophageal Junction With Diffusion Tensor Imaging Hans Gregersen, Donghua Liao, Jingbo Zhao, Steffen Ringgaard, Peter Agger, Christoffer Laustsen, Hans Stødkilde-Jørgensen Introduction: The anatomy and function of the gastro-esophageal junction (GEJ) is complex with mucosa-submucosa layers and smooth muscle layers organized into circular and longitu- dinal muscle layers. These layers continue from the esophagus into the GEJ and the stomach. Furthermore, the crura of the diaphragm and the lower esophageal sphincter are components of the GEJ. Little is known about the microstructure of the GEJ. Diffusion tensor imaging (DTI) allows mapping of the diffusion process of molecules, mainly water, in biological tissues. Water molecule diffusion patterns can therefore reveal microscopic details about tissue architecture. Its main clinical application has been in the study of neurological disorders such as acute stroke. Until now DTI has not be used much in esophagology. Aims: This study aims to reconstruct an anatomically realistic 3D model of the GEJ on the basis of DTI and more specifically to distinguish the muscle fiber transition between the esophagus, the stomach, and the diaphragmatic crura. Methods: Three 40-60kg pigs were used for obtaining the GEJ and adjacent parts of the esophagus, stomach and diaphragm. The pig was sacrificed during deep anesthesia before opening the abdominal and thoracic cavities. The intestines were removed whereas the liver and stomach was kept in place for maintaining the anatomical position of GEJ. The abdominal cavity and thoracic cavities were filled with formalin and closed. After 24h of tissue fixation, the pigs were reopened and the esophagus, upper part of stomach and intact diaphragm were dissected out and fixed with 4% formalin in an organ bath for further 2-3 days. A 7x7x7cm 3 block including the entire GEJ region and lower part of the esophagus, upper part of the stomach, and the diaphragmatic crura was prepared for Magnetic Resonance (MR) scanning. Scanning was done in a Philips Intera, 1.5T MR scanner for obtaining anatomy imaging and DTI. The wall thickness at different layers of the GEJ as well as the helical angles of the longitudinal muscle fiber, the circular muscle fiber, the gastric sling fiber, gastric oblique fibers and diaphragm crura fiber along the central axis of the GEJ were calculated on the basis of the anatomic imaging and DTI. Results: It was possible to visualize the anatomical features of GEJ using DTI. The figure shows the total muscle fibers 3D distribution along the GEJ in a traced DT image. Differences in diffusion properties between the GEJ structures such as the sling fibers were easy to recognize and detailed layer data were obtained and used as input for an ongoing development of a realistic functional model of the GEJ. Conclusions and perspectives: This study clearly demonstrated that DTI may be useful in esophagology and moved a step further by establish- ing an anatomical-functional realistic computer model of the GEJ. Sa1113 Randomized Controlled Trial of Self-Directed vs. In-Classroom Didactic Teaching of Narrow Band Imaging in Diagnosing Barrett's Esophagus Associated Neoplasia: One Year Follow Up Nikhil Banerjee, Tarun Rai, Jonathan Lattell, Prashanth Vennalaganti, Prateek Sharma, Neil Gupta Background: Narrow band imaging (NBI) has been shown to be more effective in diagnosing Barrett's esophagus (BE) associated neoplasia compared to standard endoscopic imaging. While previous studies shown that NBI can be taught to inexperienced gastroenterologists, the durability of a one time teaching session is unknown. Aim: Determine the durability of a one time teaching session of NBI in diagnosing BE associated neoplasia. Methods: This study was a continuation of a previously conducted randomized controlled trial. GI fellows with no prior experience with NBI in BE were randomized to in-classroom or self-directed didactic training. A PowerPoint (PPT) presentation (teaching tool) was developed that pro- vided (1) information about NBI, (2) NBI patterns of non-dysplastic BE (NDBE) and BE with high grade dysplasia/cancer (HGD/CA), and (3) NBI image examples of both NDBE and BE with HGD/CA. The in-classroom subjects attended a single session where an endos- copist with expertise in NBI in BE reviewed the teaching tool and answered questions. The self-directed subjects received an automated version of the teaching tool with audio. All S-229 AGA Abstracts trainees completed a test of 40 NBI images (22 NDBE, 18 HGD/CA), where they predicted histology and marked their confidence level. For this study, 1 year after completion of the initial training session, every participant was once again given a test of 40 new NBI images (22 NDBE, 18 HGD/CA) with no follow up training. Results: 17 trainees completed the 1 year follow up testing, 6 from center 1 and 11 from center 2. This included 7 first year GI fellows at the time of initial training, 7 second year GI fellows and 3 third year GI fellows. During the initial study, 9 had been randomized to in-classroom training and 8 to self- directed training. At 1 year, the overall accuracy was higher when compared to the initial post training accuracy, however the percentage of predictions made with high confidence was lower and there was no difference in the accuracy of high confidence predictions. (Table 1). At 1 year, there were significant variations in the percentage of predictions made with high confidence based on the type of training, study center, and trainee rank. (Table 2) However, there were no differences in overall accuracy or accuracy of high confidence predictions. Conclusions: In this first ever assessment of long-term follow up of teaching of NBI in diagnosing BE associated neoplasia, we found no change in the accuracy of histology prediction among GI trainees but a reduction in the percentage of predictions made with high confidence over time. There were variations in confidence based on several factors including training method, study center, and trainee rank. Moving forward, training programs on advanced imaging need to focus on creating confidence in optical diagnoses through continued exposure and training. Table 1 (comparison of initial post training vs 1 year followup) Table 2 ( 1 year followup only) comparisons made by mode of initial training, institution, and fellow rank Sa1114 Early Improvement of Liver Stiffness Measured by ARFI During Protease Inhibitor Based Triple Therapy for Chronic Hepatitis C Takuma Ajimine, Takashi Goto, Hirohide Ohnishi, Kouichi Miura, Shigetoshi Ohshima, Tomomi Shibuya, Wataru Sato, Ryo Kanata, Toshitaka Sakai BACKGROUND & AIMS Protease inhibitor based triple therapy has increased the rate of sustained virological response (SVR) in the patient of chronic hepatitis C (CHC). Acoustic radiation force impulse (ARFI) has been applied to examination of liver stiffness by measuring shear wave velocity (SWV) without resorting to liver biopsy. Liver stiffness is assumed to reflect liver inflammation and/or liver fibrosis. The aim of this study was to assess early changes in liver stiffness using ARFI after protease inhibitor based triple therapy. PATIENTS AND METHODS 21 patients (11 males and 10 females, average age was 55.8 years old) who completed the triple therapy for 24 weeks at Akita University Hospital were enrolled. 17 patients were treated with Telaprevir +Peginterferon (Peg-IFN)+Ribavirin (RBV) and 4 patients were treated with Simeprevir sodium+Peg-IFN+RBV. We compared aspartate aminotransferase (AST), alanine aminotransferase (ALT), Platelet count, serum hyaluronic acid, type IV collagen, FIB-4 index, and SWV in the liver before (baseline), at the end of treatment. FIB-4 index =[Age (Years) x AST (IU/L)]/[Platelet count (x10 9 /L) x 1/2 ALT (IU/ L)] RESULTS HCV-RNA was not detected in any patients at the end of triple therapy. Comparing baseline and the end of treatment, AST and ALT significantly improved from 53.9 IU/L to 29.2 IU/L and 66.6 IU/L to 27.1 IU/L, respectively (p<0.01). Platelet count changed from 14.9 x10 4 /μL to 13.7 x10 4 /μL (N.S.). Serum hyaluronic acid changed from 127.2 ng/mL to 115.2 ng/mL (N.S.), and type IV collagen changed from 143.2 ng/mL to 147.6 ng/mL (N.S.). FIB-4 index significantly decreased from 1.67 to 1.10 (p<0.01). Mean SWV changed from 1.46 m/s to 1.30 m/s (p<0.05), a significant decrease. CONCLUSION AST, ALT, FIB-4 index and SWV of the liver were significantly improved during protease inhibitor based triple therap. These data suggest that SWV as measured by ARFI may be useful for noninvasively evaluating the improvement of liver stiffness, which may reflect the reduction of liver inflammation and/or fibrosis, at the early stage after protease inhibitor based triple therapy for CHC. AGA Abstracts