Diagnostic value of JNET classification in blue-laser imaging magnifying endoscopy for colorectal neoplastic lesions / 中华消化内镜杂志

ABSTRACT

Objective:To investigate the diagnostic value of Japan NBI Expert Team (JNET) classification with blue-laser imaging magnifying endoscopy (BLI-ME) for colorectal neoplastic lesions.Methods:Data of 40 colorectal neoplastic lesions in 34 patients that received BLI-ME from September 2016 to December 2018 in Peking University International Hospital were reviewed and endoscopic images from those lesions were selected. Four senior endoscopists analyzed and classified these images according to the JNET criteria and determined the possible pathologic type. The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of JNET classification under BLI-ME system were calculated with final pathologic results as the golden standard.Results:Of all the 40 lesions, there were 3 non-cancerous lesions, 20 adenoma, 10 high grade intraepithelial neoplasia/submucosal-superficial lesions and 7 submucosal-deep lesions. The accuracy of 4 endoscopists was 75.00%-87.50%, and 65.38%-95.89% for different JNET types. For each JNET type, the sensitivity was 60.71%-91.67%, specificity was 84.17%-97.73%, PPV was 63.46%-95.89%, and NPV was 88.51%-99.31%. For the 4 endoscopists, the assurance rate of surface pattern was 73.08%-100.00%, 80.77%-100.00% for deciding vascular pattern, and 67.31%-100.00% for deciding final JNET type. The Kappa value between any two endoscopists was 0.630-0.887, and the interclass correlation coefficient for all results was 0.880. Conclusion:JNET classification shows significant diagnostic value under BLI-ME system for colorectal neoplastic lesions.