P254 Virtual chromoendoscopy with Narrow Band Imaging (NBI) for the classification of polypoid and non-polypoid lesions in ulcerative colitis: A prospective study using different endoscopic classifications

A. Cassinotti1, P. Fociani2, E. Beretta1, P. Duca3, M. Nebuloni2, S. Ardizzone1

1ASST Fatebenefratelli Sacco, Gastroenterology Unit, Milan, Italy, 2ASST Fatebenefratelli Sacco, Pathology Unit, Milan, Italy, 3University of Milan, Chair of Statistics, Milan, Italy


In ulcerative colitis (UC), the diagnostic accuracy of virtual chromoendoscopy with non-magnified Narrow Band Imaging (NBI) was disappointing in the detection and differential diagnosis of neoplastic lesions in most trials performed according to unclear diagnostic criteria. The Kudo and NICE criteria are validated in the general population, but controversial in UC. Recently, a modified Kudo classification specific for the inflammatory setting of UC has been developed using virtual chromoendoscopy with Fuji Intelligence Color Enhancement (FICE). The performance of NBI according to these three classifications for optical diagnosis of neoplastic and non-neoplastic lesions in UC has never been analysed.


All visible, polypoid and non-polypoid, lesions found during consecutive surveillance colonoscopies with non-magnified NBI (Exera II CV180) for long-standing UC (>8 years from disease onset) were included in a prospective, single-centre study, and classified as suspected or not suspected for neoplasia according to the NICE, Kudo and Kudo-modified criteria. The primary endpoint was to compare the sensitivity (SE), specificity (SP), negative (-LR) and positive (+LR) likelihood ratios of the three classifications, using histology as the reference standard.


394 lesions (mean size 6 mm, range 2–40 mm; type 0-Is 83%) from 84 patients (males 62%; mean age 57 y, range 30–79 y; mean disease duration 20 y) were found: 21 (5%) were neoplastic, 49 (12%) hyperplastic and 324 (82%) inflammatory lesions. The diagnostic accuracy of the NICE, Kudo and modified-Kudo classifications were, respectively: SE 76% vs. 71% vs. 86% (p = not significant); SP 55% vs. 69% vs. 79% (p < 0.05 Kudo-mod vs. both Kudo and NICE); +LR 1.69 vs. 2.34 vs. 4.15 (p < 0.05 Kudo-mod vs. both Kudo and NICE); -LR 0.43 vs. 0.41 vs. 0.18 (p = not significant).


The diagnostic accuracy of NBI for the differentiation of neoplastic and non-neoplastic lesions in UC is low if used with the conventional classifications validated in the general population, but it is significantly better with the modified Kudo classification specific for UC. The role of NBI in UC surveillance for neoplasia should be re-evaluated according to more accurate diagnostic criteria.