Endoscopic detection and differentiation of esophageal lesions using a deep neural network

• Published in: Gastrointestinal endoscopy Vol. 91; no. 2; pp. 301 - 309.e1
• Main Authors: Ohmori, Masayasu, Ishihara, Ryu, Aoyama, Kazuharu, Nakagawa, Kentaro, Iwagami, Hiroyoshi, Matsuura, Noriko, Shichijo, Satoki, Yamamoto, Katsumi, Nagaike, Koji, Nakahara, Masanori, Inoue, Takuya, Aoi, Kenji, Okada, Hiroyuki, Tada, Tomohiro
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.02.2020
• Subjects: Adult; Aged; Aged, 80 and over; Deep Learning; Esophageal Diseases - diagnostic imaging; Esophageal Diseases - pathology; Esophageal Neoplasms - diagnostic imaging; Esophageal Neoplasms - pathology; Esophageal Squamous Cell Carcinoma - diagnostic imaging; Esophageal Squamous Cell Carcinoma - pathology; Esophagus - pathology; Female; Humans; Image Processing, Computer-Assisted - methods; Male; Middle Aged; Narrow Band Imaging - methods; Neoplasm Invasiveness; Neural Networks, Computer; Observer Variation; Optical Imaging - methods; Precancerous Conditions - diagnostic imaging; Precancerous Conditions - pathology; Reproducibility of Results; Sensitivity and Specificity; CNN; SCC; NBI; WLI; AI; ME; BLI
• ISSN: 0016-5107; 1097-6779
• DOI: 10.1016/j.gie.2019.09.034

Diagnosing esophageal squamous cell carcinoma (SCC) depends on individual physician expertise and may be subject to interobserver variability. Therefore, we developed a computerized image-analysis system to detect and differentiate esophageal SCC. A total of 9591 nonmagnified endoscopy (non-ME) and 7844 ME images of pathologically confirmed superficial esophageal SCCs and 1692 non-ME and 3435 ME images from noncancerous lesions or normal esophagus were used as training image data. Validation was performed using 255 non-ME white-light images, 268 non-ME narrow-band images/blue-laser images, and 204 ME narrow-band images/blue-laser images from 135 patients. The same validation test data were diagnosed by 15 board-certified specialists (experienced endoscopists). Regarding diagnosis by non-ME with narrow-band imaging/blue-laser imaging, the sensitivity, specificity, and accuracy were 100%, 63%, and 77%, respectively, for the artificial intelligence (AI) system and 92%, 69%, and 78%, respectively, for the experienced endoscopists. Regarding diagnosis by non-ME with white-light imaging, the sensitivity, specificity, and accuracy were 90%, 76%, and 81%, respectively, for the AI system and 87%, 67%, and 75%, respectively, for the experienced endoscopists. Regarding diagnosis by ME, the sensitivity, specificity, and accuracy were 98%, 56%, and 77%, respectively, for the AI system and 83%, 70%, and 76%, respectively, for the experienced endoscopists. There was no significant difference in the diagnostic performance between the AI system and the experienced endoscopists. Our AI system showed high sensitivity for detecting SCC by non-ME and high accuracy for differentiating SCC from noncancerous lesions by ME.