1145 A Large Scale Multi-Center Study of Long-Term Outcomes After Resection for Submucosal Invasive Colorectal Cancer (Colon vs. Rectum) Hiroaki Ikematsu* 1 , Yusuke Yoda 1 , Yutaka Saito 2 , Yuichiro Yamaguchi 3 , Kinichi Hotta 4,3 , Nozomu Kobayashi 5 , Takahiro Fujii 6 , Takahisa Matsuda 2 , Madoka Takao 3 , Tomoaki Shinohara 4 , Kazuhiro Kaneko 1 1 Gastroenterology & Gastrointestinal Oncology, National Cancer Center Hospital East, Kashiwa, Japan; 2 Endoscopy Division, National Cancer Center Hospital, Tokyo, Japan; 3 Division of Endoscopy, Shizuoka Cancer Center, Shizuoka, Japan; 4 Gastroenterology, Saku Eentral Hospital, Saku, Japan; 5 Diagnostic Imaging, Tochigi Cancer Center, Utsunomiya, Japan; 6 TF Clinic, Tokyo, Japan Background: In Japanese treatment guideline of submucosal invasive colorectal cancer (SM-CRC), endoscopic resection (ER) alone has been recently identified as adequate for low-risk SM-CRC patients (pts) and subsequent curative surgical resection (SR) is recommended for high-risk pts. In DDW 2011, we reported long-term outcomes after endoscopic resection for SM-CRC pts. (Gastrointest Endosc. 2011; 73: AB296-7) However, the difference of long-term outcomes in pts with colonic and rectal SM-CRC is still unknown. Aim: To evaluate retrospectively the difference of long-term outcomes in patients following treatment for colonic and rectal SM-CRC. Patients and Methods: From January 2000 to December 2007, SM-CRC pts receiving ER alone, ER+SR, and SR among 6 institutions were analyzed. The SM-CRC lesions with 1) histologically complete resection, 2) well or moderate differentiated adenocarcinoma, 3) absence of vascular invasion, and 4) depth of submucosal invasion 1000m, were defined as low-risk group for lymph node metastases, and the remaining lesions were defined as high-risk group. Pts were classified into three groups according to pathological features and following treatments, ER alone with low-risk, ER alone with high-risk, and ER+SR or SR with high-risk. We analyzed the clinical outcomes, such as 5-year disease free survival (DFS) and recurrence-rate (RR) based on our database. Results: A total of 760 pts (colon 549 pts and rectum 211) were evaluated in this study, with median follow-up period of 50.4 months. In the 760 pts, low-risk SM-CRCs were found in 122 pts (colon 106 pts and rectum 16), and high-risk SM-CRCs were found in 638 pts (colon 443 pts and rectum 195). The 122 pts of low-risk SM-CRCs received ER alone. Of 638 pts with high risk SM-CRCs, 105 pts (colon 68 pts and rectum 37) received ER alone and 533 pts were received ER+SR or SR. The RR of low-risk SM-CRCs in the colon and rectum were 0% (0/106) and 6.3% (1/16, p=0.13), respectively. Distant metastasis was found in only 1 pts with recurrence. The RR and DFS of colonic and rectal high-risk SM-CRCs resected by ER alone were 1.5% (1/68) and 16.2% (6/37, p=0.007), and 98% and 74% (p=0.004), respectively. Of the 7 recurred pts, 6 pts (colon 1 pt and rectum 5) had local metastasis and 1 pt had distant metastasis. The RR and DFS of colonic and rectal high-risk SM-CRCs treated by ER+SR or SR were 1.9% (7/443) and 2.5% (4/195, p=0.73), and 97% and 97% (p=0.71), respectively. Of 11 pts that recurred, 2 pts (colon 1 pt and rectum 1) were local metastasis and 9 pts (colon 6 pts and rectum 3) were distant metastasis. Conclusion; Risk of the local recurrence of the high-risk rectal SM-CRCs resected by ER alone was significantly higher than that of colon. In the rectal cancer pts with high-risk pathological feature, subsequent surgery could be strongly recommended. 1169 Endoscopic Hemostasis Is Rarely Used for Severe Hematochezia: Population-Based Data From a Large Consortium of Diverse Endoscopy Practices in the United States Osnat Ron-Tal Fisher 2 , Ian M. Gralnek* 1 , Glenn M. Eisen 3 , Jennifer L. Holub 4 , Jeffrey L. Williams 4 1 Gastroenterology, Technion-Israel Institute of Technology & Rambam Medical Center, Haifa, Israel; 2 Technion-Israel School of Medicine, Haifa, Israel; 3 Gastroenterology, Oregon Health Sciences and University, Portland, OR; 4 CORI, Oregon Health Sciences and University, Portland, OR Background: Acute lower GI bleeding (LGIB), presenting as severe hematochezia, is associated with increased morbidity & mortality, especially in older patients and those with co-morbidities. As compared to acute upper GI bleeding, there are very limited data describing endoscopic hemostasis used in such patients. Aims/Methods: We used the CORI endoscopic database to describe and compare patients with severe hematochezia who received endoscopic hemostasis with those who did not. CORI has been demonstrated to be a valid reflection of community endoscopic practice. To better select for severe hematochezia, we identified all patients ( 18 years), from 1/02 to 12/08, who underwent in-patient colonoscopy for the lone indication hematochezia and had no endoscopic diagnosis of hemorrhoids. We further characterized this cohort by demographics, co-morbidities, practice setting, endoscopic diagnosis, extent of colonoscopy exam, endoscopic hemostasis type, need for repeat endoscopy, and adverse events (AE). Results: We identified 3,151 persons who underwent in-patient colonoscopy for hematochezia and had no endoscopic diagnosis of hemorrhoids. Endoscopic hemostasis was performed in only 144 patients (4.6%); 3,007 (95.4%) received no hemostasis. In both cohorts, the majority were male (64.6% and 58.0%), White (83.3% and 71.0%), with mean ages of 70.9 and 69.3 yrs, respectively. Most underwent colonoscopy at a community hospital (67.4% and 69.7%) and had ASA Score I / II (48.7% and 57.8%). In the hemostasis group, endoscopic findings* included: diverticulosis (68.1%), polyp (37.5%), AVMs (32.6%), mucosal abnormality/colitis (20.1%), solitary ulcer (8.3%), tumor (6.3%) and multiple polyps (37.5%). Hemostasis therapy included: injection (32.6%), bipolar coag (30.6%), APC (29.2%), clips (11.1%), heater probe (4.2%), other therapy (2.8%), and band ligation (2.1%)**. The two cohorts differed significantly with regard to race (p=0.02) and the hemostasis group was more likely to have more co-morbidities per ASA score (p=0.04), and have their colonoscopy exam reach the cecum (95.8% vs. 87.7%, p=0.003). Endoscopic diagnosis was significantly more likely to be AVMs or solitary ulcer, p0.0001 and p=0.0001 respectively. AEs were rare, with bleeding in 3 (2.1%). Conclusions: In patients with severe hematochezia undergoing in- patient colonoscopy, it was rare for patients to receive endoscopic hemostasis in a community practice setting. Those who did receive hemostasis were more often White males with higher ASA scores. These are novel population-based data, largely from community practice and appear to contrast published data from tertiary care referral centers.*more than one endoscopic diagnosis was allowed**more than one endoscopic therapy may have been performed 1170 Intraoperative Colonoscopy Control for Colorectal-Anastomotic Leakage: A New Solution for an Old Problem. a Pilot Study Francesca Parmeggiani* 1 , Nicola De’ Angelis 1 , Maria Clotilde Carra 1 , Barbara Bizzarri 1 , Fabiola Fornaroli 1 , Alessandro Gnocchi 1 , Carmen Madia 1 , Giorgio Nervi 1 , Raffaele Dalla Valle, 2 Gian L. De’Angelis 1 1 Unit of Gastroenterolgy and Operative Endoscopy, University of Parma, Parma, Italy; 2 Department of Emergency Surgery, University of Parma, Parma, Italy Background: The study aimed at evaluating the use of intraoperative colonoscopy control (IOCC) to assess anastomotic leakages, bleeding or early ischemias in colo-rectal anastomosis creation following left colectomy (LC) or anterior rectal resection (ARR) performed in laparotomy or laparoscopy. Materials and Methods: The study included 47 consecutive patients (26 men, 21 women, mean age 68.5 years) seen in our department between 01.01.2010 and 23.11.2011, presenting colonrectal cancer in 28 cases, and diverticular disease in the 19 cases. All patients received LC or ARR performed by the same surgeon. Patients were randomly allocated in two groups: a study group of 27 patients (5 laparoscopic LC and 16 laparotomic LC; 1 laparoscopic ARR and 5 laparotomic ARR), which systematically underwent after anastomosis creation to both intaoperative pneumatic test and IOCC; and a control group of 20 patients (13 laparotomic LC; 7 laparotomic ARR), which underwent intraoperative pnuematic test only. All endoscopes used had incoroprated narrow band imaging. Results: In the study group, the IOCC (average duration 3.6 min) revealed that 2 anastomoses needed to be recreated due to partial dehiscence, and mucosal ischemia extended to 1/3 of the anastomotic circumference. IOCC detected leakage not recognized by the classical pneumatic test in 1 case, leading to the discovery of a iatrogenic fissure in the rectal rear wall few cm from the Learning Curve for Detection of Nonpolypoid Neoplasms over Time Among Four Endoscopists, Forecasted to 3000 cases. A log-linear best fit line (dashed line) is superimposed on the scatter plot for the All Endos- copists data. R2=0.84. Abstracts www.giejournal.org Volume 75, No. 4S : 2012 GASTROINTESTINAL ENDOSCOPY AB178