Indication for ERCP is less strongly predictive of procedure related complications but is a predictor of procedure time. Emergent procedures and find more those done for bile leak take longer but are not associated with increased PEP or unplanned hospital stay. MR SMITH,1 A CHONG,3 M CHIN,1 S EDMUNDS,1 S RAFTOPOULOS,2 I YUSOFF,2 D SEGARAJASINGAM,2 C SIAH1 1Gastroenterology Department, Royal Perth Hospital, 2Sir Charles Gairdner Hospital, 3Fremantle Hospital, Western Australia Introduction: Gastric subepithelial lesions are commonly found during routine gastroscopy. The majority of these lesions are gastrointestinal stromal tumors (GISTs). While surgery is advocated for large lesions (20–30 mm+), management of small

(<20 mm)

ABT-737 mw lesions is controversial. A strategy of endoscopic ultrasound surveillance is commonly used, but data on its utility is limited. We aimed to retrospectively analyse our experience evaluating gastric subepithelial lesions and in surveillance of GISTs in Western Australia across all tertiary centers. Methods: All patients undergoing EUS for the evaluation of a gastric subepithelial lesion in Western Australia between February 2002 and May 2014 were identified from our endoscopic database. Data was collected from endoscopic and clinical databases. Data was represented as mean or median +/− range as appropriate. Results: 263 patients with gastric subepithelial lesions were identified, male 107 (41%) with a median age 58.7 years (range 21–89). EUS diagnosis was GIST 161 (62%), lipoma 37 (14%), pancreatic rest 29 (11%), duplication cyst 13 (5%), artefact from organ/vessel

indentation 14 (5%), Other 9 (3%). 126 lesions were biopsied (48%): 86 fine needle aspirations, 34 tunnel biopsies, 7 biopsies, 3 snared, with a diagnostic rate of 78%, 24%, 29%, 77% respectively. Endosonographically suspected GISTS/leiomyomas were histopathologically confirmed in 66 patients (41%). 77 of the endosonographically suspected GISTs were Non-specific serine/threonine protein kinase recommended for surveillance. Of these, 55 patients proceeded to EUS surveillance, male 27 (49%) with mean age 59.1. Mean size of lesion 14.5 mm (range 6–40 mm); <10 mm 11, 10–14 mm 21, 15–19 mm 13, ≥20 mm 10. Location of lesion: antrum 21, body 15, fundus 12, cardia 7. Lesion characteristics on first EUS: hypoechoic 51 (93%), homogenous 35 (64%) vs heterogenous 20 (36%). “High risk features” present in 11 patients (cysts/echos 10, irregular border 1); no ulceration, lymph nodes or invasion detected. 155 EUS procedures were performed with mean number of EUSs per patient 2.8 (range 2–7). Mean time of EUS follow up was 33 months, median 26 months (range 4–113 months). In this time mean change in size was −0.65 mm, median 0 (range −19 to +5 mm). Longer follow up time had no relation to change in size. When categorized by lesion size, there was no significant difference in change in size: size <10 mm, +0.32 mm; 10–19 mm, −0.56 mm; >20 mm, −2.05 mm.