Abstract

Introduction: Magnifying narrow-band imaging (NBI) is a beneficial modality for the prediction of histology in colonoscopy. The Japan NBI Expert Team (JNET) classification has been established based on the results of online surveys of the diagnostic accuracy of histological features. However, the reliability of the JNET classification has not been examined. The aim of this study was to evaluate the degree of interobserver agreement for JNET classification among endoscopists when evaluating colonic lesions. Methods: We analyzed a total of 246 colonic lesions in this interpretation test. Each lesion image was subsequently reviewed by 3 experienced endoscopists and then classified based on 4 established JNET categories and indeterminate categories between fixed categories. The primary endpoint was the interobserver agreement rate for evaluation of endoscopic findings using the JNET classification. Results: The complete concordance (defined as 3 out of 3 observer agreement) rate was 73% (179/246), while the partial concordance (defined as 2 out of 3 observer agreement) rate was 26% (64/246). Therefore, concordance of at least 2 out of 3 was achieved in 99% of all the lesions and disagreement of the 3 endoscopists was found in only 1% (3/246). Fleiss’s kappa coefficient (κ) was 0.72 for all cases, which represents moderate interobserver agreement for observers. Discussion/Conclusion: Our study highlights the reliability of the JNET classification for colonic neoplastic lesions among experienced endoscopists and demonstrates acceptable interobserver agreement. Further studies validating the clinical applicability of JNET classification are required among endoscopists with a variety of expertise in several clinical settings.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call