Tu1436 Learning Effect of Magnifying Narrow Band Imaging on Endoscopic Diagnosis of Gastric Intestinal Metaplasia: A Pilot Study of Asian-Pacific International Working Group

Noriya Uedo, Enders K. Ng, Deng-Chyang Wu, Han‐Mo Chiu, Khek Yu Ho, Chong Vui Heng, Rajvinder Singh, Rupert W. Leong, Yoshinori Morita, Atsushi Imagawa, Udom Kachintorn, Joseph J.�y. SungGastrointestinal 2011 · DOI

Abstract

Diagnosis of intestinal metaplasia (IM) is an important step to identify subjects with increased risk of developing gastric cancer. However, ordinary endoscopy has a low sensitivity in detecting these potentially pre-malignant changes. To investigate whether training on the use of narrow band imaging (NBI) with magnifying endoscopy improve diagnostic accuracy for gastric IM. Observational study. Setting: An international working group in the Asian-Pacific region. Patients: With written informed consent, 80 consecutive asymptomatic or mildly dyspeptic patients presenting for screening esophago-gastro-duodenoscopy were enrolled in a Japanese hospital. Endoscopic images of the antrum were recorded under white light imaging (WLI), narrow band imaging (NBI), and magnifying NBI. Biopsy specimens were taken from the lessor and the greater curvature of the antrum (3 cm from the pylorus) to provide histological reference standard for IM. Interventions: Representative images of the each endoscopic method in the former 40 patients (a total of 120 images) were randomly ordered and presented to 23 endoscopists from the Asian-Pacific region as a pre-test. A lecture about endoscopic diagnosis of gastric IM by magnifying NBI was then delivered by an experienced endoscopist. A post-test was conducted with another set of randomly ordered 120 images in the latter 40 patients. Diagnostic criteria of each method were, slightly elevated or flat whitish areas for WLI, whitish patchy areas for NBI and ridge or villous surface structure with finding of the light blue crest for magnifying NBI. Main outcome measurements: Changes of diagnostic ability of each method after learning magnifying NBI finding of IM. Sensitivity of all methods, WLI, NBI and magnifying NBI, was significantly improved in the post-test compared to the pre-test (0.52 to 0.65, p=0.007; 0.66 to 0.75, p=0.060; 0.67 to 0.81, p=0.004, respectively). Inter-observer agreement (κ value) among the participant endoscopists was low and did not change. Limitation: Quasi-experimental study using recorded images. Learning magnifying NBI could improve diagnostic ability of IM and it warrants validation in clinical settings. Inter-observer variability did not ameliorated much with the training lecture, thus case discussion may be needed to increase agreement of endoscopic diagnosis.Tabled 1hanges of diagnostic ability after training lecture of magnifying NBIMethodTestSensitivitySpecificityAccuracyAgreement (κ value)WLIPre-test0.520.720.620.21Post-test0.65⁎p<0.05 for pre- vs. post-test0.660.660.23NBIPre-test0.660.610.640.26Post-test0.75⁎p<0.05 for pre- vs. post-test0.630.680.30Magnifying NBIPre-test0.670.530.600.16Post-test0.81⁎p<0.05 for pre- vs. post-test0.500.630.15 p<0.05 for pre- vs. post-test Open table in a new tab
BibTeX
@inproceedings{uedo2011tu,
  author = {Noriya Uedo and Enders K. Ng and Deng-Chyang Wu and Han‐Mo Chiu and Khek Yu Ho and Chong Vui Heng and Rajvinder Singh and Rupert W. Leong and Yoshinori Morita and Atsushi Imagawa and Udom Kachintorn and Joseph J.�Y. Sung},
  title = {{Tu1436 Learning Effect of Magnifying Narrow Band Imaging on Endoscopic Diagnosis of Gastric Intestinal Metaplasia: A Pilot Study of Asian-Pacific International Working Group}},
  booktitle = {Gastrointestinal},
  year = {2011},
  doi = {10.1016/j.gie.2011.03.928},
}

Loading related content...

Figure