Ying-Chun Jheng1,2,3,4, Yen-Po Wang1,2,5,4, Hung-En Lin1,2,4, Kuang-Yi Sung1,2,4, Yuan-Chia Chu6,4, Huann-Sheng Wang1,7,4, Jeng-Kai Jiang7,4, Ming-Chih Hou1,2,4, Fa-Yauh Lee2,4, Ching-Liang Lu8,9,10,11. 1. Endoscopy Center for Diagnosis and Treatment, Taipei Veterans General Hospital, Taipei, Taiwan. 2. Division of Gastroenterology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan. 3. Department of Medical Research, Taipei Veterans General Hospital, Taipei, Taiwan. 4. Faculty of Medicine, National Yang-Ming University School of Medicine, Taipei, Taiwan. 5. Institute of Brain Science, National Yang-Ming University School of Medicine, Taipei, Taiwan. 6. Information Management Office, Taipei Veterans General Hospital, Taipei, Taiwan. 7. Division of Colon and Rectum Surgery, Department of Surgery, Taipei Veterans General Hospital, Taipei, Taiwan. 8. Endoscopy Center for Diagnosis and Treatment, Taipei Veterans General Hospital, Taipei, Taiwan. cllu@ym.edu.tw. 9. Division of Gastroenterology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan. cllu@ym.edu.tw. 10. Institute of Brain Science, National Yang-Ming University School of Medicine, Taipei, Taiwan. cllu@ym.edu.tw. 11. Faculty of Medicine, National Yang-Ming University School of Medicine, Taipei, Taiwan. cllu@ym.edu.tw.
Abstract
OBJECTIVES: Computer-aided diagnosis (CAD)-based artificial intelligence (AI) has been shown to be highly accurate for detecting and characterizing colon polyps. However, the application of AI to identify normal colon landmarks and differentiate multiple colon diseases has not yet been established. We aimed to develop a convolutional neural network (CNN)-based algorithm (GUTAID) to recognize different colon lesions and anatomical landmarks. METHODS: Colonoscopic images were obtained to train and validate the AI classifiers. An independent dataset was collected for verification. The architecture of GUTAID contains two major sub-models: the Normal, Polyp, Diverticulum, Cecum and CAncer (NPDCCA) and Narrow-Band Imaging for Adenomatous/Hyperplastic polyps (NBI-AH) models. The development of GUTAID was based on the 16-layer Visual Geometry Group (VGG16) architecture and implemented on Google Cloud Platform. RESULTS: In total, 7838 colonoscopy images were used for developing and validating the AI model. An additional 1273 images were independently applied to verify the GUTAID. The accuracy for GUTAID in detecting various colon lesions/landmarks is 93.3% for polyps, 93.9% for diverticula, 91.7% for cecum, 97.5% for cancer, and 83.5% for adenomatous/hyperplastic polyps. CONCLUSIONS: A CNN-based algorithm (GUTAID) to identify colonic abnormalities and landmarks was successfully established with high accuracy. This GUTAID system can further characterize polyps for optical diagnosis. We demonstrated that AI classification methodology is feasible to identify multiple and different colon diseases.
OBJECTIVES: Computer-aided diagnosis (CAD)-based artificial intelligence (AI) has been shown to be highly accurate for detecting and characterizing colon polyps. However, the application of AI to identify normal colon landmarks and differentiate multiple colon diseases has not yet been established. We aimed to develop a convolutional neural network (CNN)-based algorithm (GUTAID) to recognize different colon lesions and anatomical landmarks. METHODS: Colonoscopic images were obtained to train and validate the AI classifiers. An independent dataset was collected for verification. The architecture of GUTAID contains two major sub-models: the Normal, Polyp, Diverticulum, Cecum and CAncer (NPDCCA) and Narrow-Band Imaging for Adenomatous/Hyperplastic polyps (NBI-AH) models. The development of GUTAID was based on the 16-layer Visual Geometry Group (VGG16) architecture and implemented on Google Cloud Platform. RESULTS: In total, 7838 colonoscopy images were used for developing and validating the AI model. An additional 1273 images were independently applied to verify the GUTAID. The accuracy for GUTAID in detecting various colon lesions/landmarks is 93.3% for polyps, 93.9% for diverticula, 91.7% for cecum, 97.5% for cancer, and 83.5% for adenomatous/hyperplastic polyps. CONCLUSIONS: A CNN-based algorithm (GUTAID) to identify colonic abnormalities and landmarks was successfully established with high accuracy. This GUTAID system can further characterize polyps for optical diagnosis. We demonstrated that AI classification methodology is feasible to identify multiple and different colon diseases.
Authors: Luigi Della Corte; Antonio Mercorio; Pierluigi Giampaolino; Salvatore Giovanni Vitale; Giuseppe Vizzielli; Giuseppe Bifulco; Giada Lavitola Journal: Updates Surg Date: 2021-11-05