center. Result: A total of 97 patients, with a total of 105 lesions were enrolled and randomly assigned to undergo ESD using either SH or NS. Self-completion rate was higher in the SH group (85% (45/53)) than in the NS group (75% (39/52)), although there was no statistically significant difference in the two groups (p=0.203). Subclass analysis on lesions with resected specimen diameter of less than 40mm (n=82) showed that the use of SH (97% (38/39) improved self-completion rate compared with the use of NS (81% (35/43), (p=0.031). The median operation time was 47 min in the SH group and 54 min in the NS group (p=0.083). The use of SH reduced the median number of submucosal injections (2 vs. 4, p0.001), and the median volume of injections (13ml vs. 40ml, p=0.001) compared with the use of NS. No severe adverse events were observed in either group. Limitation: Singe center study. Couclusion: For the ESD by supervised residents, we could not find statistical difference in self-completion rate between SH and NS groups, although the self- completion rate was higher in the SH group for the lesions with resected specimen diameter of less than 40mm (p=0.031). This study suggests that a larger multicenter study is needed to establish the efficacy of SH. 314b Effect of Routine Mucosal Closure After Endoscopic Submucosal Dissection for Gastric Neoplasms: A Randomized Controlled Trial Bo-in Lee, Byung Wook Kim, Hwang Choi, Jeong-Seon Ji, Sun-Mee Hwang, Hiun Suk Chae, Kyu-Yong Choi Background: A larger mucosal defect usually leaves after endoscopic submucosal dissection (ESD) for gastric neoplasms, and this may be associated with ESD- associated complications such as bleeding. Aim: This study was performed to determine whether routine closure for mucosal defect after ESD can enhance mucosal healing and reduce ESD-associated complications. Subject & Method: ESD was performed for 42 gastric neoplasms (18 adenomas, 22 intramucosal cancers, 1 subtle submucosal invasive cancer, 1 MALT-lymphoma) of 42 patients. Mucosal defect was left for 21 control subjects while routine closure of mucosal defect was tried with a two-channel endoscope, hemoclips, and a detachable snare for 21 subjects of closure group (Fig. 1). Both groups received follow-up endoscopy next day and endoscopic treatment was performed for lesions with active bleeding or visible vessel. All subjects received proton pump inhibitors for 6 weeks and 2nd follow-up endoscopy was done after 8 weeks. Results: There was no significant difference in age, sex, history of anti-platelet agents, prevalence of H. pylori infection, area of resected specimen, or area of mucosal defect between control and closure group. In closure group, mucosal closure could be performed completely in 11 patients (52%) and incompletely in 8 (38%). There was significant difference in elapsed time per area of mucosal defect (0.8 vs. 1.1 min/cm2, P=0.013), hospital stay (4.0 vs. 3.0, P=0.000), presence of lesions with active bleeding or visible vessel after 1 day (38.1% vs. 4.8%, P=0.020), and open ulcer after 8 weeks (53% vs. 6%, P=0.007) between control and closure group. No difference was observed in ESD-associated complications or decrease in hemoglobin concentration. Conclusions: Routine mucosal closure after gastric ESD reduces lesions with risk of bleeding and enhances mucosal healing. Larger-scale trial is necessary to determine whether mucosal closure can really reduce ESD-associated complications. 314c Risk Factors and Clinical Outcomes of Gastric Perforation Induced by Endoscopic Submucosal Dissection (ESD) for Gastric Neoplasms Toshitatsu Takao, Kohei Takizawa, Masaki Tanaka, Naomi Kakushima, Hisatomo Ikehara, Yuichiro Yamaguchi, Hiroyuki Matsubayashi, Hiroyuki Ono Background & Aim: ESD has been widely performed as a treatment for gastric neoplasms. The aim of this study was to clarify the risk factors for gastric perforation induced by ESD and to assess the clinical outcomes after perforation.Patients & Methods: From September 2002 to July 2008, 1591 gastric neoplasms in 1371 patients were treated by ESD using IT knife in our institution. Lesion characteristics, treatment after perforation and clinical outcomes were collected from the medical records. Results: The overall rate of gastric perforation in this study was 5% (74 lesions in 73 patients). Univariate analysis showed that tumor location (the upper and the middle third [7%], especially greater curvature in the upper third [12%]), specimen size (more than median size, 46mm [7%]), histological ulceration (present within the lesion [9%]), histological depth (submucosal or deeper invasion [9%]) and operation time (longer than median operation time, 38 minutes [8%]) were significantly related to the risk of perforation. When perforation occurred, immediate closure with endoclips was performed in all the cases except for one case. Naso- gastric tube was inserted in all the patients. Oral intake was prohibited for a median of 3 days. Median hospitalization period of the patients with perforation was 6 days. All the cases except for one case did not require emergent surgical rescues. 37 patients among 73 patients with perforation had non-curative endoscopic treatment. 14 patients among 37 patients underwent additional surgery in our institution. No peritoneal dissemination were found during surgery and cytological examinations of ascites were negative in all the 14 cases. Among 73 patients, 64 patients were followed-up at our institution. 62 patients including the case of emergent surgery were alive and had no recurrence for median follow-up periods of 364 days (range 6-1868 days). Two patients died during follow-up. The first patient refused further surgery after non-curative ESD and died from peritonitis carcinomatosa. The second patient died from unrelated causes to gastric neoplasm. Conclusions: The risk factors for perforation induced by ESD were tumor location, specimen size, histological ulceration, histological depth and operation time. Perforation was unlikely to lead to peritoneal dissemination in patients who had curative resection. 314d The Utility of a New Jumbo Biopsy Forceps for Tissue Acquisition of Gastric Subepithelial Masses Srinadh Komanduri, Shriram Jakate Introduction: Subepithelial masses (GES) of the stomach have provided a diagnostic dilemma for gastroenterologists. Conventional biopsies using standard forceps generally do not procure diagnostic tissue. As a result, the majority of these lesions are referred for Endoscopic Ultrasound (EUS) evaluation for imaging and tissue diagnosis. Given the majority of (GES) are benign; an alternative means of tissue diagnosis would preclude many unnecessary EUS exams. We prospectively evaluated jumbo tunnel biopsies for the evaluation of GES referred for EUS. Methods: We prospectively evaluated 72 consecutive patients referred to our institution from 2007-2008, for EUS for gastric GES found on endoscopy. 32 patients were excluded for: extralumenal masses (13), extrinsic compression by adjacent organs (10), vascular anomaly (1), and inability to identify a lesion (8). 40 patients with previous nondiagnostic biopsies qualified for the study and underwent EUS examination. FNA was performed when technically feasible as was jumbo tunnel biopsies (Radial Jaw 4 ™, Boston Scientific Corp). Primary outcomes included: technical success, acquisition of diagnostic tissue (including immunohistochemistry when relevant), and complications.Results: All 40 patients had sampling with the jumbo forceps using tunnel biopsy technique. 23/40 (58%) patients had lesions amenable to FNA (mean # passes =3). FNA was not performed in 17 patients due to size (1cm) or anatomical difficulty (cardia location). 13/23 (56%) lesions that underwent FNA were diagnostic while 34/40 (85%) patients who underwent jumbo biopsy sampling were diagnostic. Of the 10 patients who had non-diagnostic FNA, jumbo biopsies were diagnostic in 8/10 (80%), while only 1/6 non-diagnostic jumbo biopsies was diagnostic by EUS-FNA. Although not powered to compare modalities, there is a trend towards statistical significance with jumbo biopsy sampling vs. EUS- FNA. Most importantly, only 14/34 (41%) lesions were lesions that required further follow up or resection (e.g. GIST, Carcinoid). The remainder of the lesions required no further surveillance (e.g. Leiomyoma, pancreatic rest, lipoma). There was a single complication after jumbo biopsy and FNA (bleeding), which did not require any intervention.Conclusions: Our study shows the efficacy and safety of the radial jaw 4 jumbo biopsy forceps in obtaining diagnostic tissue in GES. As the majority of GES are benign, utilization of jumbo forceps would be cost-saving by limiting unnecessary referral for EUS and endoscopic surveillance. Utilization of tunneled jumbo biopsies is safe and effective and should be considered when a GES is encountered during endoscopy. Abstracts www.giejournal.org Volume 71, No. 5 : 2010 GASTROINTESTINAL ENDOSCOPY AB105