• 2019-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • 2021-03
  • br This e learning program can be accessed on the


    This e-learning program can be accessed on the web. Therefore, it is possible to educate many endoscopists on the diagnosis process of M-NBI, and it is expected that the web program will be widely provided in the future. Based on the results of this study, a good learning effect on the e-learning system for depressed lesion was obtained from the AUC analysis. The most common EGC shape is the depressed type. Therefore, educating endoscopists through e-learning appears to contribute greatly to the improvement of EGC diagnosis.
    The AUC result for cancer diagnosis rates achieved with M-NBI for pathologic cancer or noncancer in our 365 subjects was favorable (.84). In our investigation, Volume -, No. - : 2019 GASTROINTESTINAL ENDOSCOPY 7
    Learning effect after e-learning training in endoscopic diagnosis Ikehara et al
    .4379-.7894; P Z .22). However, when MV and MS were taken into account for the diagnosis, the AUC increased to .88, indicating significantly improved cancer diagnostic capability. This indicates the importance of taking the 3 items into account when making a diagnosis. Meanwhile, investigation of the kappa values of DL, MV, and MS indi-cated a moderate agreement (.42) was obtained for MV, whereas only slight to fair agreement was obtained for DL and MS. Results suggest that each participant created his or her own diagnostic algorithm in the Resiniferatoxin similar to that of artificial intelligence through deep learning. Because the algorithm was different for each participant, it seems that the kappa value of the answer was lowered. An algorithm developed in a participant’s brain by learning is difficult to share. Meanwhile, accuracy is improved by e-learning, and an individual’s algorithm tends to improve as a whole. It appears that diagnosis must be complemented with diagnostic imaging using artificial intelligence to achieve further improvement in M-NBI diagnostic capability.
    The content that can be provided by e-learning on the web is limited. Furthermore, the time needed will be
    longer. At present, many studies on endoscopic diagnosis using artificial intelligence have been reported.17,18 The de-
    gree of diagnosis is obtained as a numerical value by analyzing the endoscopic image by artificial intelligence through deep learning. In addition, to give lectures in real time in daily endoscopic examination, it is important to perform endoscopic examination under the instruction of an expert endoscopist, but the number of expert endo-scopists is limited.
    An artificial intelligence diagnostic system is estab-lished through annotating by an expert endoscopist. Its diagnostic ability has been improved, and an equivalent diagnostic accuracy as that of an expert endoscopist has been obtained. The greatest advantage of the artifi-cial intelligence diagnostic system is that Domain of a protein can be installed in many endoscopic systems, and by using auxiliary artificial intelligence, novices are provided an environment similar to using an endoscope while receiving advice from an expert. Future evaluation of the educational effects of these newly developed tech-nologies is awaited.
    This study has some limitations. First, in the 40 lesions used in test, there were only 2 cases of noncancerous le-sions in elevated lesion and cancer lesions in flat type. Because these numbers of lesions were limited, the effects of e-learning on elevated and flat regions may have been underestimated. Second, a large number of participants were examined. In some cases, Fleiss’ kappa may return low values even when agreement is high.
    In conclusion, in M-NBI education for endoscopists, a good learning outcome was obtained in depressed and small lesions. In contrast, a poor learning outcome was demonstrated in elevated and flat lesions.
    This work was supported by a research grant from the Japanese Foundation for Research and Promotion of Endoscopy and the Central Research Institute for Endos-copy, Fukuoka University.
    2. Hamashima C, Okamoto M, Shabana M, et al. Sensitivity of endoscopic screening for gastric cancer by the incidence method. Int J Cancer 2013;133:653-9.
    5. Yao K, Oishi T, Matsui T, et al. Novel magnified endoscopic findings of microvascular architecture in intramucosal gastric cancer. Gastrointest Endosc 2002;56:279-84.
    6. Yao K, Anagnostopoulos GK, Ragunath K. Magnifying endoscopy for diagnosing and delineating early gastric cancer. Endoscopy 2009;41: 462-7.
    7. Ezoe Y, Muto M, Uedo N, et al. Magnifying narrowband imaging is more accurate than conventional white-light imaging in diagnosis of gastric mucosal cancer. Gastroenterology 2011;141:2017-25.
    8. Nakanishi, Doyama H, Ishikawa H, et al. Evaluation of an e-learning sys-tem for diagnosis of gastric lesions using magnifying narrow-band im-aging: a multicenter randomized controlled study. Endoscopy 2017;49: 957-67.