AUTHOR=Ma Mingjun , Li Zhen , Yu Tao , Liu Guanqun , Ji Rui , Li Guangchao , Guo Zhuang , Wang Limei , Qi Qingqing , Yang Xiaoxiao , Qu Junyan , Wang Xiao , Zuo Xiuli , Ren Hongliang , Li Yanqing TITLE=Application of deep learning in the real-time diagnosis of gastric lesion based on magnifying optical enhancement videos JOURNAL=Frontiers in Oncology VOLUME=Volume 12 - 2022 YEAR=2022 URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.945904 DOI=10.3389/fonc.2022.945904 ISSN=2234-943X ABSTRACT=Abstract Background and aim: Magnifying image-enhanced endoscopy was demonstrated to have higher diagnostic accuracy than white-light endoscopy. However, differentiating early gastric cancers from benign lesions is difficult. We aimed to determine whether the computer-aided model for diagnosis of gastric lesion can be applied to videos rather than still images. Methods: A total of 719 magnifying optical enhancement images of early gastric cancers , 1490 optical enhancement images of the benign gastric lesions, and 1514 images of background mucosa were retrospectively collected to train and develop a computer-aided diagnostic model. Subsequently, 101 video segments and 671 independent images were used for validation, and error frames were labeled to retrain the model. Finally, a total of 117 unaltered full-length videos were utilized to test the model and compared with those diagnostic results made by independent endoscopists. Results: Except for atrophy combined with intestinal metaplasia and low-grade neoplasia, the diagnostic accuracy was 0.90 (85/94). The sensitivity, specificity, PLR, NLR, and overall accuracy of the model to distinguish early gastric cancer from noncancerous lesion were 0.91 (48/53), 0.78 (50/64), 4.14, 0.12, and 0.84 (98/117), respectively. No significant difference was observed in the overall diagnostic accuracy between the computer-aided model and experts. High kappa values were found between the model and experts, which mean kappa value were 0.63. Conclusions: The performance of the computer-aided model for diagnosis of early gastric cancer is comparable to that of experts. Magnifying optical enhancement model alone may not be able to deal with all lesions in the stomach, especially when near focus on severe atrophy with intestinal metaplasia. These results warrant further validation in prospective studies with more patients. ClinicalTrials.gov registration was obtained (identifier number: NCT04563416).