Japanese narrow band imaging expert team classification of colorectal polyps: A validation study from India
摘要
Japanese narrow band imaging expert team (JNET) classification has a diagnostic accuracy above 90% in differentiating neoplastic from non-neoplastic colonic polyps as well as estimating the depth of invasion in colorectal cancer. However, its validation outside Japan is limited to expert centers and requires magnifying endoscopes.
Aims and MethodsThis study aimed at validating the JNET classification prospectively in a real-world setting in India using magnifying endoscopes with dual focus. We analyzed consecutive patients with colonic polyps detected via these endoscopes. The JNET classification was compared with histopathology, the gold standard and its diagnostic accuracy was assessed.
ResultsTotal 203 consecutive patients with colonic polyps underwent examination using a magnifying endoscope with dual focus. In real time, 331 polyps were identified and classified based on the JNET classification. Among them, 15 polyps could not be retrieved, leaving 316 polyps for histopathological comparison in the study. The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of each JNET classification type, along with their 95% confidence intervals, are as follows. For Type-1 JNET classification, the values are 78% (69–86), 97% (94–99), 92% (84–97), 92% (87–95) and 92% (88–94), respectively. Type-2 A JNET classification has corresponding values of 92% (86–96), 84% (78–89), 82% (75–88), 93% (88–97) and 88% (84–91). For Type-2B JNET classification, the values are 45% (24–68), 97% (95–99), 56% (31–78), 96% (93–98) and 93% (90–96). Lastly, Type-3 JNET classification has values of 95% (87–99), 98% (96–100), 94% (85–98), 99% (97–100) and 98% (96–99), respectively.
ConclusionsThe JNET classification has good accuracy in characterizing colonic polyps using magnifying endoscopes with dual focus. Large-volume, multicentric data is necessary to validate the findings in our study.
Graphical Abstract