physiologic feedback of fullness and less hunger. It is done with the Incisionless Operating Platform (IOP) that uses four instruments and has shown to produce durable tissue remodeling at 12 months. We present our initial experience in the safety with the POSE procedure in 32 patients included in a long term program of follow and change of alimentary and living habits and preliminary two and four months follow-up data in weight loss. Patients and Methods: Indications have been obesity grade 1 and II, and Grades III that refused surgical approach. We have performed 32 operations in 32 patients (11 men, 21 women) mean BMI was 38,26 (sd 3,65 )and age 42.32 (sd 9.67 ). Written consent was obtained and the study passed through the ethics committee of the hospital. Preoperative, intra-operative and follow up data was obtained prospectively every two weeks initially and monthly afterwards. The aim of the operation was to put eight plications in fundus and three in the anterior wall of the body-antrum area. Results: Eleven plications were performed in all the patients. Mean time of the operation was 54 minutes. All the patients were discharged at 12-18 hours of the operation. Mild sore throat for 24 hours in half of the patients, mild abdominal pain lasting less than 24 hours in 18 and low grade fever at 72 hours controlled with oral antibiotics in one and 12 hours vomiting in two, were the only complications observed. Oral intake with liquids was started at 12 hours of the operation. Mean weight loss has been of 12,07 Kg (sd 3,96 ) at 2nd month ( 23 patients) and 13,09 (sd 5.92) at 4rth month (16 patients) and % EBMIL at two months of 33,37 (sd 14,12 ) and 45,81 (sd 19.43) at four months. %EWL of 27.56 (sd 9.71) and 30.76 (sd 14,83 ) at second and fourth month . All the patients referred loss of hanger and early satiety. Conclusions: The POSE procedure seems to be a safe intervention without significant adverse effects to date. Initial results in weight loss are encouraging although long term follow-up has to be made to know its real utility and place in the obese patient always associated to life style modifications. Sa1613 The Role of Trimodal Endoscopy in Detection and Characterization of Gastric Lesions Roman O. Kuvaev* 1 , Sergey V. Kashin 1 , Igor O. Ivanikov 2 , Alexander S. Nadezhin 3 , Alexey A. Gvozdev 1 1 Endoscopy, Yaroslavl Regional Cancer Hospital, Yaroslavl, Russian Federation; 2 Gastroenterology, Central Clinical Hospital of the Russian President Administration, Moscow, Russian Federation; 3 Pathology, Yaroslavl Regional Cancer Hospital, Yaroslavl, Russian Federation Introduction: Trimodal endoscopy (TME) is a combination of autofluorescence endoscopy (AFI) with high-magnification endoscopy (HME)and narrow-band imaging (NBI). AFI produces real-time pseudocolor images from detection of autofluorescence emitted by endogenous fluorophores in the mucosa. Gastric lesions typically appear purple or green in the AFI images due to their morphology (elevated, or depressed lesions respectively). This novel endoscopic technique is a powerful tool for detecting early cancer. However, the value of TME for detecting and differentiatinggastric lesions is still under research. Aims and Methods: The aim was to determine the efficacy of AFI as the first step in TME by analyzing AFI-patterns ofgastric lesions utilizing HME and NBI. This study comprised 44 lesions in 32 patients (pts). Initially the first endoscopist investigated all pts by WLE combined with chromoendoscopy (CE) with indigocarmine. Afterwards the second endoscopist examined these pts by TME. We defined any lesion that was different from the surrounding mucosa in color as AFI-positive. There were two types of AFI-positive lesions: purple lesions in green background (P/G) and green lesions in purple background (G/P). AFI- positive lesions were analyzed by employing HME and NBI. HME findings based on the two microanatomies: microvascular architecture (MV) and microsurface structure (MS). Irregular (IMVP) MV pattern with irregular (IMSP) or absence (AMSP) MS pattern is used as the criterion of neoplasia. Biopsies were taken from all lesions for histological assessment. Results: From 41 AFI-positive lesions there were 33 (80.49%) P/G-pattern and 8 (19.51%) G/P-pattern. There were 3 AFI-negative lesions (2 neoplastic, 1 nonneoplastic) detected with only WLE with CE. P/G-pattern included 19 (57.58%)nonneoplastic (chronic gastritis, intestinal metaplasia, hyperplasia) and14 (42.42%)neoplastic (LGD, HGD,adenocarcinoma, ring-cell cancer)lesions (n.s.). G/P-pattern included 3 (37.5%) nonneoplastic and 5 (62.5%) neoplastic lesions (n.s.). The sensitivity AFI in detection of neoplastic lesions was 90.48% comparable with WLE with CE (95.23%). From analysis by HME in all AFI-positive lesions IMSP/AMSP with IMVP were found in 11 cases (10 neoplastic and 1 nonneopastic lesions), the specificity was 95.45%. Conclusion: Due to high sensitivity AFI can be used as “red-flag” technique to detect suspected lesions for further detailed inspection with high specificity methods of HME and NBI. This combination known as TME may increasedetection rate of high-grade dysplasia/early cancer. Sa1614 Introducing BACE - Balloon-Assisted Confocal Endomicroscopy: the First Endoscopic Technique to Perform Confocal Imaging Within the Small Bowel Helmut Neumann* 1 , Michael Vieth 2 , Lucia C. Fry 3 , Klaus Monkemuller 3 , Markus F. Neurath 1 1 Department of Medicine 1, University of Erlangen-Nuremberg, Erlangen, Germany; 2 Institute of Pathology, Klinikum Bayreuth, Bayreuth, Germany; 3 Department of Internal Medicine, Gastroenterology and Infectious Diseases, Marienhospital Bottrop, Bottrop, Germany Background: Confocal laser endomicroscopy (CLE) is an established endoscopic technique to obtain optical biopsies during ongoing endoscopy. Moreover, CLE harbours the potential to perform molecular imaging within the gastrointestinal tract. Various studies have shown the potential of CLE for diagnosis and characterization of gastrointestinal diseases with high overall accuracy rates and excellent agreement to histopathology. Nevertheless, according to the length of available CLE-systems there is currently no possibility to perform confocal imaging within the small bowel. Aims: To develop a new endoscopic technique to perform confocal imaging within the small bowel and to evaluate this technique in the clinical setting. Material and Methods: A standard CLE- colonoscope with a working length of 150 cm was used in combination with the balloon-guided endoscopy system. The disposable was mounted on the tip of the CLE-endoscope with its front balloon sheltered with its rear balloon. Advancement of the endoscope and overtube was accomplished using the push- and-pull method as previously described. CLE was performed after intravenous injection of fluorescein in a standardized fashion. Results: A new endoscopic technique to perform CLE within the small bowel was developed. In a preliminary series of 10 cases a mean ileal advancement depth of 45 cm (Range 7-70cm) with an average ileoscopy time of 14.5 minutes (range 8-25 minutes) was averaged. Confocal imaging within the small bowel was technically easily feasible and revealed typical architectural details including intestinal vili and microvessels within the lamina propria. Confocal image quality and agreement to corresponding histological analysis was almost perfect. No complications were recorded. Conclusions: BACE - Balloon-assisted confocal endomicroscopy- was developed as a new endoscopic technique that readily allows to perform CLE within the deep small bowel. Therefore, this technique has the enormous potential to increase our understanding of small bowel pathologies in the near future. Sa1615 Prospective Randomized Trial of Magnetic Endoscopic Submucosal Dissection for Large Early Gastric Neoplasms (EGN) Chang Beom Ryu* 1 , Moon Sung Lee 1 , Jun Yong Bae 3 , Jae Hee Lim 1 , Hee Yong Yoo 1 , Hyung Su Ahn 1 , Bong Min Ko 1 , Su Jin Hong 1 , Hee Kyung Kim 2 1 Department of Internal Medicine, Soon Chun Hyang University School of Medicine, Bucheon, Republic of Korea; 2 Departments of Pathology, Soon Chun Hyang University School of Medicine, Bucheon, Republic of Korea; 3 Departments of Internal Medicine, Goodmorning hospital, Pyeongtaek, Republic of Korea Background: Endoscopic submucosal dissection (ESD) has been one of the standard treatment for early gastric neoplasms (eGN) in Korea. However, the procedure is time consuming. A simple ESD using a magnet allows the use of two hands to operate. Compared to the former conventional ESD usually practiced by one hand the procedure becomes more effective. Aim: The purpose of prospective comparative clinical trial was to evaluate the feasibility of magnetic ESD (MAG-ESD) for large eGN in human beings. Method: From Jan 2010 to Sep 2011, total 58 patients (MAG-ESD: 33, conventional ESD group: 25) CLE images of a lg adenoma (left) and normal colonic crypts (right), in which crypt boundaries have been automatically detected and over- printed. Abstracts www.giejournal.org Volume 75, No. 4S : 2012 GASTROINTESTINAL ENDOSCOPY AB221