all USL(s). Every 3 mos, patients had HR chromoendoscopy with biopsies from TA (2 bx/2cm) and any USLs (1 bx/1cm). Focal RFA was applied to USL(s), if present. If all interim visit biopsies were negative for MGIN or worse, patient was released to 12 month visit.Outcome: Percent of cases with complete response of neoplasia (CR-Neo), defined as no MGIN or worse in any biopsy from the TA, at 12 months.Results: 29 patients were enrolled (14 male, mean age 59.6 yrs) with MGIN (18), HGIN(10), and ESCC(1). Mean length of USL(s) was 6.2 cm. After one ablation, 25 (86%) achieved CR-Neo. In the 4 patients with persistent neoplasia at 3 months, an additional single focal ablation achieved CR- Neo. At this time, all patients are CR-Neo (n=29, 100%). There were four strictures, all successfully dilated.Conclusion: This is the first prospective trial of RFA for early squamous neoplasia of the esophagus. In this homogeneous group of patients, we rigorously confirmed the diagnosis of eligible neoplasia (MGIN, HGIN, T1m2), excluded more advanced disease, and used HR chromoendoscopy for follow-up biopsies. We found RFA to be highly effective and well tolerated. A CR-Neo was achieved in all patients (100%), the majority of whom would have otherwise been treated by radical EMR or esophagectomy. 587e Complete Barrett’s Excision (CBE) by Endoscopic Resection (ER): A Highly Effective and Durable Curative Option for Barrett’s Segments With High Grade Dysplasia (HGD) or Early Adenocarcinoma (EAC) Alan Moss, Michael J. Bourke, Luke F. Hourigan, Robert G. Walker, Saurabh Gupta, Andrew J. Metz, Stephen J. Williams BACKGROUND AND AIMS: After focal treatment of Barrett’s HGD or EAC, patients are at risk of metachronous neoplasia unless the Barrett’s segment is completely eliminated. Surgery has significant morbidity and mortality. Radiofrequency ablation does not allow complete histological evaluation and is expensive. ER is minimally invasive, provides complete histology and possibly CBE. This prospective, two center study aims to determine the safety, efficacy and long term outcomes of attempted CBE. METHODS: CBE by 2-3 stage radical mucosectomy was attempted via upper gastrointestinal endoscopy for patients aged 75 with minimal comorbidities and Barrett’s segments 3 cm in circumferential (C) extent who were referred with biopsy demonstrated HGD or EAC. Technique was multiband mucosectomy (MBM) or Inoue cap, with Endocut Q Effect 3 electrocautery (ERBE, Germany). In circumferential disease, a 60-80% circumferential resection was performed initially. Procedures were day cases under conscious sedation with midazolam, fentanyl and propofol. Following ER, twice daily proton pump inhibitor was prescribed for 4 weeks. Repeat endoscopy was performed 6-8 weekly until CBE was achieved. CBE was defined as the endoscopic absence of Barrett’s mucosa, with biopsy confirmation at the most recent follow-up endoscopy. RESULTS: CBE was attempted in 53 patients [41 men; mean age 69; median C length 2cm, median maximal extent 2cm (but up to 5cm)]. MBM was used in 41 and cap in 12. On intention to treat analysis, the success rate was 50/53 (94%). This was achieved in a mean (and median) of 2 sessions (range 1-4). The mean (median) number of resections performed at the first session was 3 (2) with range 1-10. At subsequent sessions, the mean (and median) number of resections was 2. Two patients’ ER specimens revealed submucosally invasive adenocarcinoma, so esophagectomy was performed, and CBE could not be pursued. One patient did not complete the treatment protocol due to progression of co-morbid illness. Thus, per protocol CBE success was 100%. Mean follow-up duration was 28 months (range 3-75). 8 patients (15%) had oesophageal dilatation (median 2, range 1-8) for symptomatic stricture with complete symptom resolution. There were no perforations, and no deaths. Follow-up endoscopy demonstrated complete neo-squamous re- epithelialization of the treated segments, with no sub-squamous intestinal metaplasia on biopsies. No metachronous lesions occurred. CONCLUSIONS: In patients with Barrett’s HGD or EAC, CBE is safe and effective. The stricture rate is low with the use of Endocut electrocautery. CBE consistently achieves cure of Barrett’s esophagus with excellent long term outcomes. 587f Multi-Band Mucosectomy in Barrett’s Esophagus: A Prospective Registration of 1060 Resections in 243 Procedures Lorenza Alvarez Herrero, Roos E. Pouw, Bas L. Weusten, Jacques Bergman BACKGROUND: Multi-Band Mucosectomy (MBM) is a relative new technique for endoscopic resection (ER) in Barrett Esophagus (BE). This suck-and-cut technique uses a modified variceal band ligator which allows for up to 6 consecutive resections without prior submucosal lifting. AIM: To prospectively evaluate the safety of MBM and to evaluate its efficacy for complete endoscopic removal of predefined focal lesions in BE. METHODS: Prospective registration of all MBM procedures in BE was performed between Nov’04-Oct’09. Prior to MBM, the target area was delineated with electrocoagulation markers followed by ERs until the delineated area, including all markers, was resected. Pts were discharged after 2-4 hrs observation and followed up by telephone at 24 hrs and 2 wks. Primary endpoints were the number of acute (during procedure) and early (1 week) complications and the rate of complete endoscopic removal of the delineated focal lesion. RESULTS: 243 MBM procedures, with a total number of 1060 resections, were performed in 170 BE pts (median age 68 [IQR 61-75]; 150 male). MBM was performed for focal lesions in 113 procedures (32 en-block; 81 piecemeal (279 resections; median 3 [IQR 2-5])), for removal of BE as part of a (stepwise) radical ER protocol in 117 procedures (713 resections; median 5 [IQR 3-9]) and for escape treatment after radiofrequency ablation in 13 procedures (36 resections; median 2 [IQR 1-3]). Acute complications occurred in 7/243 procedures (2.9% [95% CI 1.4-5.8%]): all bleedings treated with standard endoscopic therapy during the procedure and graded as “mild”. Pts were observed for 1-2 days and only 1 pt underwent elective endoscopic re- inspection. Despite the absence of submucosal lifting no perforations occurred in 1060 MBM resections (0% [95% CI 0-0.4%]). Early complications consisted of 5/ 243 delayed bleedings (2.1% [95% CI 0.9%- 4.7%]), 4 occurring within 48 hrs and 1 after 6 days. All delayed bleedings were effectively managed endoscopically and graded as “moderate”. Bleedings were significantly more frequent after MBM of focal lesions. Complete resection of targeted area was achieved in 103/113 (91% [95% CI 84%- 95%]). All failures were due to scarring of the tissue after prior ulceration prohibiting suctioning of tissue into the MBM-cap (p0.001). CONCLUSION: This is the largest prospective series of MBM in BE. Despite the absence of submucosal lifting, perforations did not occur. Post-MBM bleeding occurs in approximately 2% and can generally be managed endoscopically. MBM allows for effective removal of the delineated target area in the vast majority of cases unless the area is scarred due to prior ulceration or endoscopic therapy. 587g Prospective Randomized Comparison of Cold Snare Polypectomy and Conventional Polypectomy Akira Horiuchi, Yoshiko Nakayama BACKGROUND AND AIMS: The removal of small colon polyps by colon snare transection without electrocautery effectively eliminates polyps, but the technique raises some concerns with the difficulty in retrieving the polyp and the risk of bleeding. The aim of this study is to compare cold snare polypectomy with conventional polypectomy. METHODS: Patients with small colorectal polyps up to 8 mm were randomized to two groups using either cold snare technique (Cold group) or conventional polypectomy (Conventional group). The size and location of all polyps were recorded and the retrieval rate of all collected specimens were evaluated histologically. Complications and all gastrointestinal symptoms within two weeks after each polypectomy were recorded. The principal outcome measures were complete retrieval rate of colorectal polyps and bleeding. Secondary outcome measures were abdominal symptoms within two weeks after each polypectomy or other complications.RESULTS: 64 patients: Cold group, N=32 (94 polyps), or Conventional group, N=32 (92 polyps) were randomized. The patients’ demographic characteristics and the number and size of polyps removed were similar between the two techniques. There was a significant difference in the mean procedure time between Cold group and Conventional group (18 min vs. 25 min, P=0.03). There was no significant difference in the complete retrieval rate of colorectal polyps between Cold and Conventional group (96% (90/94) vs. 97% (89/92)). No bleeding requiring hemostasis occurred in either group. There was a significant increase in abdominal symptoms after polypectomy with Conventional group compared to Cold group (25% (8/32) vs. 3% (1/32), P0.001).CONCLUSIONS: Neither difficulty in polyp retrieval nor bleeding after polypectomy was a problem with cold snare polypectomy. Cold snare technique was possibly more patient friendly as it was associated with less abdominal symptoms after polypectomy than conventional polypectomy. 587h Estimating Adenoma Detection Rate With Polyp Detection Rate: Proof of Concept Daniel T. Rodriguez Correa, Anna M. Buchner, Michael B. Wallace, Dawn L. Francis Background: Adenoma detection rate has been accepted as a quality benchmark for colonoscopy. Many practices find it difficult to determine adenoma detection rate because it requires the combination of endoscopic and histologic findings which are often in different places in the medical record. It is unknown whether there is a constant that could be applied to polyp detection rate to estimate adenoma detection rate. We evaluated the adenoma to polyp ratio amongst 37 endoscopists at two institutions and determined the variability of this ratio and the utility of creating a constant by which polyp detection rate could be converted to an estimate of adenoma detection rate.Methods: We retrospectively evaluated colonoscopies over a two year period at two different practice sites. Only board certified staff gastroenterologists performed these procedures with or without fellows and for all indications. We determined the average adenoma to polyp detection rate for all individual endoscopists and for the group as a whole. We then took the group average and used it as a constant multiplier for each individual endoscopist’s polyp detection rate. We used a student t test to Abstracts www.giejournal.org Volume 71, No. 5 : 2010 GASTROINTESTINAL ENDOSCOPY AB127