Table. Comparision of pCLE accuracies of three observers

MeasureObserver

1Observer

2Observer

3

(%) 95% CI (%) 95% CI (%) 95% CIAccuracy 69% 58.9-77.9 89% 81.2-94.4 80% 70.8-87.3Sensitivity 89% 50.7-99.4 89% 50.7-99.4 89% 50.7-99.4Specificity 65% 50.8-77.7 89% 75.9-95.2 79% 65-89Positive predictive

value31% 15.1-51.9 57% 29.7-81.2 42% 21.1-66.1

Negative predictivevalue

97% 83.4-99.9 98% 87.3-100 98% 86-100

Su1555Comparison of High Definition With Standard White LightEndoscopy for Detection of Dysplastic Lesions DuringSurveillance Colonoscopy in Patients With InflammatoryBowel DiseaseVenkataraman Subramanian1, Vidyasagar Ramappa2, Emmanouil Telakis2,Jayan Mannath2, Krish Ragunath2, Chris J. Hawkey1

1Nottingham Digestive Diseases Centre, University of Nottingham,Nottingham, United Kingdom; 2Digestive Diseases Directorate,Nottingham University Hospitals, Nottingham, United KingdomIntroduction: Dysplasia in colonic inflammatory bowel disease (IBD) is oftenmultifocal and flat, making it easy for significant lesions to be overlooked.Dye spraying the mucosal surface is believed to enhance visualization ofsubtle mucosal abnormalities, but is cumbersome and messy and has pooruptake among endoscopists. High definition (HD) colonoscopy improvesadenoma detection rates by improving the ability to detect subtle mucosalchanges and is as good as chromoendoscopy in polyp detection. The utilityof high definition colonoscopy in dysplasia detection in patients with IBD hasnot been reported so far. We aimed to compare the yield of dysplasticlesions detected by standard definition white light endoscopy (SD) with highdefinition endoscopy (HD). Methods: Details of consecutive patients withlong standing (�7years) colonic IBD who underwent surveillancecolonoscopy at Nottingham University Hospitals between September 2008and February 2010 were extracted from the endoscopy database. Details ofdiagnosis, duration of disease and outcomes of the colonoscopy were thencollected from the electronic patient records and patient notes. Thecolonoscopies were done at 2 sites, of which one had only HD systems andthe other SD. SPSS v17 was used for the data analysis. Results: 360colonoscopies were done on 353 patients. There were 162 colonoscopies(102 UC and 60 CD) in the SD group and 208 colonoscopies (146 UC and 62CD) in the HD group. The groups were well matched for mean age ofpatients, duration of disease, gender and number of biopsies taken. Table 1gives information on the number and characterization of dysplastic lesionsdetected Conclusion: HD colonoscopy is superior to SD colonoscopy in targeteddetection of dysplastic lesions during surveillance colonoscopy of patients withcolonic IBD in routine clinical practice. HD colonoscopy could facilitateendoscopic resection in these patients. Randomized controlled studies arerequired to confirm these findings.

Table. Lesions detected in each of the groups

StandardDefinition(N�162)

HighDefinition(N�208) p value

No of lesions 15 30 NSNo of patinets with dysplasia 12 23 NSNo of patients with HGD/Cancer 1/2 2/5 NSNo. with visible targeted lesions 6 22 �0.05no with flat lesions 2 10 �0.05

NS: not significant

Su1556A Blinded Comparison of the Efficacy and Safety of Hot BiopsyForceps Electrocauterisation and Snare Based Techniques forDiminutive Colonic PolypectomyAndrew J. Metz1, Michael J. Bourke1, Alan Moss1, Duncan Mcleod2,Kayla Tran2, Craig Godfrey3, Abhilash P. Chandra1

1Gastroenterology, Westmead, Sydney, NSW, Australia; 2Pathology,Westmead, Sydney, NSW, Australia; 3Animal Care, Westmead, Sydney,NSW, AustraliaBackground: Although linked with perforation, serositis, delayed bleeding, andincomplete resection, Hot Biopsy Forceps Electrocauterisation (HBF) is still

widely used for diminutive colonic polypectomy. There has been no scientificcomparison of safety or efficacy between HBF and conventional snarepolypectomy, or it’s simple modifications such as Suction PseudopolypTechnique (SPT). Methods: Small, flat artificial lesions were resected in vivo inrandom order from 10 pig colons by one endoscopist. Technique wasstandardised. Diathermy setting was coagulating current 25watts. HBF: the tissuewas avulsed after 1-2 seconds of current caused blanching of the artificialpedicle. SPT: A pseudopolyp was created by suction of the target tissue, whichwas snared. Euthanasia and colectomy was performed at 72 hours followed byblinded histological assessment of the specimens and polypectomy ulcers.Thermal injury depth was quantified. Colonic wall layers were numericallyassigned: Mucosa (M)�0, Submucosa (SM)�1, Partial Muscularis Propria(MP)�2, Full thickness MP�3, Serosa�4 Results: One pig died with a presumedperi-procedural arrhythmia and was excluded. All polypectomies took less than30 seconds. Eighty-two resections were analysed (41 HBF, 41 SPT). SPTspecimen size was mean 79mm2. All contained complete mucosal resection andmean SM layer of 1.72mm depth. HBF specimens contained mucosa only. Onaverage 62% of the specimen was ablated and uninterpretable (range 5-100%).Mucosal necrosis adjacent to resection sites varied widely, between 2 and 9mm(mean 5.7mm). There was visible mucosa under the HBF ulcer in 14%.Conclusions: Despite standardised HBF technique, there is a wide range oflateral mucosal thermal injury as well as residual target musosa. HBF results in asignificantly greater depth of tissue injury, with a high proportion of transmuralnecrosis. In comparison to snare based polypectomy, HBF is potentially bothineffective and hazardous.

Depth of thermal injury

SPT N�41 HBF N�41 P value

Partial MP necrosis 1 (2%) 14 (34%) p�0.001Full thickness MP necrosis 1 (2%) 9 (22%) p�0.014Full thickness MP inflammation 5 (12%) 13 (32%) p�0.06Histological serositis 4 (10%) 13 (32%) p�0.027

Numerical mean depth of thermal damage

SPT HBF P value

Ulcer 0.95 0.85 p�0.45Necrosis 1.02 1.56 p�0.01Inflammation 1.95 2.49 p�0.045

Su1557Grasp and Snare Polypectomy of Large Flat Right ColonicPolyps in a Community SettingDaniel G. Luba, Maydeen M. Ogara, James A. DisarioGI, Monterey Bay GI Research Institute, Monterey, CAIntroduction: Colonoscopy is effective in reducing the incidence of left-sidedcolon cancer, however it may not be as effective in reducing right-sided cancers.Flat, right, colonic polyps are increasingly being recognized as precursors tothese cancers. The prevalence of flat, right, colonic polyps is not well defined,and they are difficult to detect and remove in their entirety. A novel technique isto use a double-channel colonoscope. Saline is injected to raise the polyp on amucosal bleb, a snare passed through one channel and a forceps through theother, the forceps are passed out through the opened snare, the polyp is graspedwith the forceps and pulled slightly up, the snare is advanced around themargins of the polyp and closed while applying monopolar energy to resect thepolyp. This may facilitate total removal of flat polyps and provide more completepathological specimens. Aim: Determine the prevalence of large, flat, rightcolonic polyps in colonoscopy patients; the outcomes of the grasp and snarepolypectomy technique, and the histopathology of these polyps in a communitysetting.Methods: IRB-approved review of all colonoscopies and prospective datacollection of grasp and snare polypectomies performed by 2 experiencedendoscopists between January 1 and November 30, 2010 in a communityambulatory endoscopy center was performed. Double channel colonoscopes(ECF3870, CF3870TLK, CF3890TLK, Pentax Montvale,NJ) All flat polyps greaterthan 1.2 cm in their largest dimension were removed using the grasp and snaretechnique. Cecal intubation time, polypectomy time, polyp histopathology, andcomplications were recorded. Results: 1208 colonoscopies (739DL, 469JD) wereperformed and 73 polyps meeting morphological criteria were found in 48(4%)patients. Of the qualifying patients, there were 20(42%) women and 28(58%)men with a mean age of 54(49-86) years. The mean colonoscopy duration was38(12-89) minutes, and grasp and snare polypectomy time was 11(2-62) minutes.A mean of 6(1-30) ml of saline was injected and 30W of pulsed energy wasused. Polyp histology included: 17(23%) hyperplastic, 37(51%) tubular adenomas,9(12%) tubulovillous adenomas, 9(12%) sessile serrated adenomas, and 1(1%)normal histology. There was no high-grade dysplasia or cancer. Thirteen patientshave had follow up colonoscopy and 1 who had a 3 cm hyperplasic polypresected had residual hyperplastic tissue at the site. Perforation occurred in3(6%) patients with a 2 cm cecal, 1.5 cm cecal, and 2 cm ascending colon polyp.

Abstracts

www.giejournal.org Volume 73, No. 4S : 2011 GASTROINTESTINAL ENDOSCOPY AB303