Tu1482 High Definition White Light Endoscopy and I-SCAN for in-Vivo Characterisation of Small Colonic Polyps: No Need to Push the Button

Abstract

a conventional image, the lesion was each observed by NBI magnification. The diagnosis of NBI magnification was based on Sano’s classification. Also irregular findings of microvessels including “caliber change of microvessel” (CCM), “long irregular vessel” (LIV) and “decline of microvessel density” (DMD) were analyzed. When type V pit pattern was suspected, the lesion was observed after crystal-violet dye-stain. The diagnosis of chromoendoscopic magnification was based on Kudo’s classification. Both capillary pattern and pit pattern were compared with final histological diagnosis. This study was performed retrospectively and consecutively. Histological diagnosis on resected specimen was determined according to Vienna classification. Submucosal massively invasive cancer (SM-M) was defined as an invasive depth greater than 1000 m. Results: As to histological diagnoses, there were 380 intramucosal cancers, 60 submucosal slightly invasive cancers, and 81 cancers with submucosal massively invasion. The values of the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) and accuracy of type IIIB for the diagnosis of SM-M were 70.4%, 98.6%, 90.5%, 94.8% and 94.3%, respectively. Those of VI highly irregular and VN were 81.5%, 98.4%, 90.4%, 96.7% and 95.8%, respectively. Type IIIB was observed in 19 (73.1%) of 26 protruded type SM-M, 8 (40.0%) of 20 flat type SM-M and 30 (85.7%) of 35 depressed type SM-M. The values of specificity of CCM, LIV and DMD for the diagnosis of SM-M were 99.1%, 99.8% and 99.5%, respectively. CCM was observed in 14 (51.9%) of 26 protruded type SM-M. And DMD were observed in 23 (67.6%) of 35 depressed type SM-M. Conclusion: NBI magnification would have clinical advantage to diagnose the invasion depth of early colorectal cancers. Submucosal massively invasive cancers had characteristic irregular findings that depended on the morphological type. These irregular findings had high specificity for the diagnosis of submucosal massively invasive cancers. Tu1482 High Definition White Light Endoscopy and I-SCAN for in-Vivo Characterisation of Small Colonic Polyps: No Need to Push the Button Peter J. Basford*, Gaius R. Longcroft-Wheaton, Pradeep Bhandari Gastroenterology, Portsmouth Hospitals NHS Trust, Portsmouth, United Kingdom Introduction: Standard definition white light endoscopy is inadequate for in-vivo characterisation of small colonic polyps. Chromoendoscopy and electronic imaging techniques are shown to improve accuracy. The ASGE has identified prediction of polyp surveillance intervals and negative predictive value for adenomatous histology of diminutive recto-sigmoid polyps as key targets for new technologies. High definition white light endoscopy incorporating charge coupled device chips with resolution in excess of 1 million pixels are now available but there is little data on their use. Aims&Methods: We aimed to examine the in-vivo characterisation accuracy of high definition white light endoscopy (HDWL) plus a novel electronic imaging modality i-Scan (Pentax, Japan). Patients undergoing colonoscopy through the UK Bowel Cancer Screening Programme were prospectively recruited. All colonoscopies were performed by a single expert endoscopist with extensive experience in in-vivo diagnosis. Procedures were performed with Pentax EC-3890Li 1.2 Megapixel HD colonoscopes and EPKi processor. An initial classification & validation exercise was carried out to determine the optimum i-Scan settings for in-vivo diagnosis, and to develop a novel in-vivo diagnosis assessment tool. All polyps 10mm in size were assessed sequentially with HDWL and i-Scan. Optical magnification was not used. Predicted histology (non-neoplastic, adenoma, cancer) was recorded for both modalities and compared to the final histopathological diagnosis as reported by an expert gastrointestinal pathologist. Predictions were rated as high or low confidence assessments. Results were analysed for sensitivity and specificity for neoplasia, overall accuracy, and negative predictive value for rectosigmoid polyps 5mm as recommended by the ASGE PIVI statement. Results: 84 patients were recruited, in whom 209 polyps 10mm were included. Mean polyp diameter was 4.3mm, median 4mm. 134 polyps were neoplastic and 75 non-neoplastic. There were no significant differences in sensitivity (95.5% vs 97.0%) and specificity (89.3% vs 90.7%) for neoplasia and overall diagnostic accuracy (93.3% vs 94.7%) between HDWL and i-Scan. Negative predictive value for adenomatous histology of rectosigmoid polyps 5mm was 100% with both modalities. Polyp surveillance intervals using in-vivo assessment of diminutive polyps were correct in 95% and 97% of patients with HDWL and i-Scan respectively. Conclusion: 1) Excellent in vivo diagnostic accuracy, in excess of 90% can be achieved with HDWL alone.2) No significant gains in accuracy over HDWL were noted with i-Scan when used with a 1.2Megapixel HD colonoscope therefore, there is no need to push the button or spray dye to improve accuracy. 3) Both HDWL and i-Scan fulfill the ASGE criteria for ‘resect and discard’ and ‘do not resect’ strategies for diminutive polyps