Xiaobei Luo1, Jiahao Wang2, Zelong Han1, Yang Yu3, Zhenyu Chen1, Feiyang Huang4, Yumeng Xu5, Jianqun Cai1, Qiang Zhang1, Weiguang Qiao1, Inn Chuan Ng6, Robby T Tan7, Side Liu1, Hanry Yu8. 1. Department of Gastroenterology, Nanfang Hospital, Southern Medical University, Guangzhou, China. 2. Mechanobiology Institute, National University of Singapore, Singapore; Institute of Bioengineering and Nanotechnology, Agency for Science, Technology and Research (A∗STAR), Singapore. 3. Mechanobiology Institute, National University of Singapore, Singapore; Institute of Bioengineering and Nanotechnology, Agency for Science, Technology and Research (A∗STAR), Singapore; CAMP, Singapore-MIT Alliance for Research and Technology, Singapore. 4. Institute of Bioengineering and Nanotechnology, Agency for Science, Technology and Research (A∗STAR), Singapore. 5. Mechanobiology Institute, National University of Singapore, Singapore. 6. Department of Physiology, The Institute for Digital Medicine (WisDM), Yong Loo Lin School of Medicine, Singapore. 7. Yale-NUS College, Singapore; Department of Electrical and Computer Engineering, National University of Singapore, Singapore. 8. Department of Gastroenterology, Nanfang Hospital, Southern Medical University, Guangzhou, China; Mechanobiology Institute, National University of Singapore, Singapore; Institute of Bioengineering and Nanotechnology, Agency for Science, Technology and Research (A∗STAR), Singapore; CAMP, Singapore-MIT Alliance for Research and Technology, Singapore; Department of Physiology, The Institute for Digital Medicine (WisDM), Yong Loo Lin School of Medicine, Singapore.
Abstract
BACKGROUND AND AIMS: Endoscopic submucosal dissection (ESD) and EMR are applied in treating superficial colorectal neoplasms but are contraindicated by deeply invasive colorectal cancer (CRC). The invasion depth of neoplasms can be examined by an automated artificial intelligence (AI) system to determine the applicability of ESD and EMR. METHODS: A deep convolutional neural network with a tumor localization branch to guide invasion depth classification was constructed on the GoogLeNet architecture. The model was trained using 7734 nonmagnified white-light colonoscopy (WLC) images supplemented by image augmentation from 657 lesions labeled with histopathologic analysis of invasion depth. An independent testing dataset consisting of 1634 WLC images from 156 lesions was used to validate the model. RESULTS: For predicting noninvasive and superficially invasive neoplasms, the model achieved an overall accuracy of 91.1% (95% confidence interval [CI], 89.6%-92.4%), with 91.2% sensitivity (95% CI, 88.8%-93.3%) and 91.0% specificity (95% CI, 89.0%-92.7%) at an optimal cutoff of .41 and the area under the receiver operating characteristic (AUROC) curve of .970 (95% CI, .962-.978). Inclusion of the advanced CRC data significantly increased the sensitivity in differentiating superficial neoplasms from deeply invasive early CRC to 65.3% (95% CI, 61.9%-68.8%) with an AUROC curve of .729 (95% CI, .699-.759), similar to experienced endoscopists (.691; 95% CI, .624-.758). CONCLUSIONS: We have developed an AI-enhanced attention-guided WLC system that differentiates noninvasive or superficially submucosal invasive neoplasms from deeply invasive CRC with high accuracy, sensitivity, and specificity.
BACKGROUND AND AIMS: Endoscopic submucosal dissection (ESD) and EMR are applied in treating superficial colorectal neoplasms but are contraindicated by deeply invasive colorectal cancer (CRC). The invasion depth of neoplasms can be examined by an automated artificial intelligence (AI) system to determine the applicability of ESD and EMR. METHODS: A deep convolutional neural network with a tumor localization branch to guide invasion depth classification was constructed on the GoogLeNet architecture. The model was trained using 7734 nonmagnified white-light colonoscopy (WLC) images supplemented by image augmentation from 657 lesions labeled with histopathologic analysis of invasion depth. An independent testing dataset consisting of 1634 WLC images from 156 lesions was used to validate the model. RESULTS: For predicting noninvasive and superficially invasive neoplasms, the model achieved an overall accuracy of 91.1% (95% confidence interval [CI], 89.6%-92.4%), with 91.2% sensitivity (95% CI, 88.8%-93.3%) and 91.0% specificity (95% CI, 89.0%-92.7%) at an optimal cutoff of .41 and the area under the receiver operating characteristic (AUROC) curve of .970 (95% CI, .962-.978). Inclusion of the advanced CRC data significantly increased the sensitivity in differentiating superficial neoplasms from deeply invasive early CRC to 65.3% (95% CI, 61.9%-68.8%) with an AUROC curve of .729 (95% CI, .699-.759), similar to experienced endoscopists (.691; 95% CI, .624-.758). CONCLUSIONS: We have developed an AI-enhanced attention-guided WLC system that differentiates noninvasive or superficially submucosal invasive neoplasms from deeply invasive CRC with high accuracy, sensitivity, and specificity.
Authors: Quirine E W van der Zander; Mirjam C M van der Ende-van Loon; Janneke M M Janssen; Bjorn Winkens; Fons van der Sommen; Ad A M Masclee; Erik J Schoon Journal: Sci Rep Date: 2022-10-06 Impact factor: 4.996