patients (91.7%) had a prior history of dysplasia but not detected during the FB (Table). Conclusion: Results from this multi-center, prospective, randomized trial demonstrate an increased detection of patients with HGD/Ca using Wide Area Transepithelial Sampling in conjunction with 4-quadrant biopsies compared to bi- opsies alone. Use of WATS can improve surveillance programs in patients with BE. Table Additional HGD/CA diagnosed by WATS 23 Prior history of dysplasia, n(%) 21 (91.7) HGD/Cancer, n 14 LGD, n 8 No dysplasia, n 2 Follow-up history No histology available, n (%) 14 (61) Ablation, n 11 No follow-up, n 3 Histology available, n (%) 9 (39) HGD/Cancer, n 4 LGD, n 3 No dysplasia, n 2 61 Absence of Suspicious Findings on Volumetric Laser Endomicroscopy Strongly Predicts Histopathologic Complete Remission of Dysplasia and Intestinal Metaplasia in Patients After Visual Eradication of Barrett’s Esophagus Brooks D. Cash* 2 , Virendra Joshi 9 , Herbert C. Wolfsen 6 , Vani J. Konda 8 , Steven R. DeMeester 3 , Eric K. Ganguly 5 , Michael B. Wallace 6 , Robert Hawes 7 , Douglas K. Pleskow 4 , Michael S. Smith 1 1 Medicine/Gastroenterology, Lewis Katz School of Medicine at Temple University, Philadelphia, PA; 2 University of Southern Alabama Medical Center, Mobile, AL; 3 Department of Surgery, University of Southern California, Los Angeles, CA; 4 Medicine/Gastroenterology, Beth Israel, Boston, MA; 5 Medicine/Gastroenterology, University of Vermont Medical Center, Burlington, VT; 6 Medicine/Gastroenterology, Mayo Clinic, Jacksonville, FL; 7 Medicine/Gastroenterology, Florida Hospital, Orlando, FL; 8 Medicine/Gastroenterology, University of Chicago, Ch, IL; 9 Medicine/ Gastroenterology, Oschner Medical Center, New Orleans, LA Background: Volumetric laser endomicroscopy (VLE) uses next generation optical coherence tomography to provide real-time imaging of human tissue microstruc- ture (depth 3 mm, resolution 7 microns). Most VLE research has focused on its utility in the management of Barrett’s esophagus (BE), primarily to facilitate targeted tissue sampling during surveillance of non-treated disease. Our aim was to assess the degree of correlation between a VLE negative for suspicious findings and histo- pathologic confirmation of complete remission of intestinal metaplasia (CR-IM) and dysplasia (CR-D). Methods: Interim analysis was performed on a US-based multi- center registry that prospectively collected de-identified demographic, endoscopic and histologic data from patients undergoing VLE during upper endoscopy. All procedures were performed with the NvisionVLEÔ Imaging System (NinePoint Medical, Bedford, MA). This study evaluated patients seen for surveillance following prior BE treatment. These treatments could include tissue resection, such as endoscopic mucosal resection (EMR), and/or field ablation techniques including radiofrequency ablation (RFA) or cryotherapy (Cryo). Results: 152 procedures from 130 patients were included. The average age of patients was 67 years and 82% were male. In 71 procedures, white light endoscopy (WLE) was suspicious for BE or dysplasia, including 15 with focally suspicious regions. VLE showed areas of concern in 47 (66%) of the procedures suspicious for BE or dysplasia on WLE and in 12 (80%) of the procedures with focal lesions. In the remaining 81 procedures with normal WLE findings, 29 (36%) had suspicious areas on VLE and targeted biopsies showed non-dysplastic BE (NDBE) in 4 (14%) and low grade dysplasia (LGD) in 1 (3%). In patients with normal WLE and VLE, random biopsies were performed in 51 of 52 procedures and in 4 procedures (8%) abnormal histology was found: NDBE in 3 and LGD in 1. The negative predictive value for metaplasia and dysplasia was 84% and 97% respectively. Conclusion: VLE identified suspicious lesions for residual or recurrent BE or dysplasia in 36% of post-ablation cases with normal WLE findings, and targeted biopsies showed IM or dysplasia in 14% of these procedures. A negative finding on WLE and VLE correctly predicted CR-D in 98% of cases and CR-IM in 92% of cases. Improvements in image interpretation and the ability to target areas for biopsy will likely lead to increased detection of abnormal areas in patients with otherwise negative findings. As this is a new technology, operator error may be responsible for missed VLE lesions; therefore performance characteristics should increase with time. As a strong predictor of both CR-D and CR-IM in patients who have a negative WLE, VLE may improve post-treatment surveillance by decreasing the number of biopsies required to correctly stage residual or recurrent disease. Patient Demographics Included procedures (patients) 152 (130 patients) Mean age 67 years (range 43-88) Male 106 (82%) Prior treatments 152 procedures RFA only 63 (40%) RFA and EMR 26 (17%) EMR Only 18 (11%) Cryo and RFA and EMR 11 (7%) Cryo and RFA 11 (7%) Cryo only 6 (4%) Other 12 (8%) Prior highest grade of pathology 124 patients with available results Invasive Adenocarcinoma 10 (8%) Intramucosal Carcinoma 18 (14%) High-Grade Dysplasia 44 (34%) Low-Grade Dysplasia 26 (20%) Indefinite for Dysplasia 5 (4%) Non-Dysplastic BE 17 (13%) Other 4 (6%) 114 Upper Gastrointestinal Haemorrhage With Glasgow-Blatchford Score of 3 or Less Can Be Safely Managed in Outpatient Setting: Results of a Prospective Validating Study Arrhchanah Balachandran, Vinh-An Phan, Tsai Ow, Leong Lee, Mark Schoeman, Nam Q. Nguyen* Gastroenterology and Hepatology, Royal Adelaide Hospital, Adelaide, South Australia, Australia Glasgow-Blatchford score (GBS) is the most widely validated scoring tool in pre- dicting the risk of adverse clinical outcomes following an upper gastrointestinal haemorrhage (UGIH). Our recent study observed that no patients with a GBS 3 required endoscopic management, experienced rebleeding or death secondary to UGIH (Bryant et al. GIE 2013). Aims: To prospectively assess the occurrence of an UGIH-related adverse event prior to endoscopy, and the need for therapeutic intervention in patients with a GBS 3 who will be managed in outpatient setting. Methods: This was a prospective, risk-stratified management study. Patients with a GBS 3 were discharged and managed with a non-urgent endoscopy within 5 working days. Patients with a GBS >3 were admitted and managed with an urgent endoscopy within 24 hours. The accuracy of the GBS was evaluated using receiver operating characteristic (ROC) analyses. The sensitivity and specificities were also calculated for the GBS cut-offs of zero to three in their ability to predict a low-risk UGIH. Results: 171 patients were recruited to the study over 7 months, and 16% (28/171) patients had GBS3. No patient with a GBS 3 experi- enced an adverse outcome or required hospital re-admission prior to their non-ur- gent endoscopy. On endoscopy, only one patient has a non-bleeding gastric angiodysplasia that required endoscopic therapy. In contrast, 118 (82.5%) patients with a GBS >3 (nZ143) required some form of therapeutic intervention with 7 patients experienced UGIH-related readmission and 2 patients experienced UGIH- related mortality. ROC analyses demonstrated an area under the curve of 0.89 (95% confidence interval, 0.83-0.94) in predicting the need for therapeutic intervention. A GBS 3 yielded a sensitivity of 99.2% and a specificity of 51.9% in its ability to predict a low-risk bleed. Conclusions: Patients with UGIB with GBS of 3 or less can be safely managed in outpatient setting, which can significantly reduce resources related to inpatient management as well as need for emergency endoscopy for UGIB. 115 Hemospray Versus the Combined Conventional Technique for Endoscopic Treatment of Bleeding Peptic Ulcers: A Pilot Randomized Controlled Study Andrew Kwek*, Tiing Leong Ang, Peng Lan Jeannie Ong, Yi-Lyn Jessica Tan, Daphne Ang, Ngai Moh Law, Prem Harichander Thurairajah, Kwong Ming Fock Changi General Hospital, Singapore, Singapore Introduction: Endoscopic hemostasis with local adrenaline injection, thermal coag- ulation or mechanical coaptation with clips are standard first-line techniques for bleeding peptic ulcers. Although adoption of dual endotherapy has improved suc- cess rates, recurrent bleeding still occurs in up to 25% of patients. Preliminary studies on a new nanopowder hemostatic agent TC-325 (Hemospray, Cook Endos- copy) show its safety and effectiveness in the treatment of arterial upper GI bleeding. However, there are no randomized controlled trials comparing Hemos- pray with the combined conventional technique (CCT); saline adrenaline injection www.giejournal.org Volume 83, No. 5S : 2016 GASTROINTESTINAL ENDOSCOPY AB123 Abstracts