A Prospective Validation and Observer Performance Study of a Deep Learning Algorithm for Pathologic Diagnosis of Gastric Tumors in Endoscopic Biopsies

• Published in: Clinical cancer research Vol. 27; no. 3; pp. 719 - 728
• Main Authors: Park, Jeonghyuk, Jang, Bo Gun, Kim, Yeong Won, Park, Hyunho, Kim, Baek-Hui, Kim, Myeung Ju, Ko, Hyungsuk, Gwak, Jae Moon, Lee, Eun Ji, Chung, Yul Ri, Kim, Kyungdoc, Myung, Jae Kyung, Park, Jeong Hwan, Choi, Dong Youl, Jung, Chang Won, Park, Bong-Hee, Jung, Kyu-Hwan, Kim, Dong-Il
• Format: Journal Article
• Language: English
• Published: United States 01.02.2021
• Subjects: Adult; Aged; Aged, 80 and over; Biopsy - statistics & numerical data; Deep Learning; Feasibility Studies; Female; Gastric Mucosa - diagnostic imaging; Gastric Mucosa - pathology; Gastroscopy - statistics & numerical data; Humans; Image Interpretation, Computer-Assisted - methods; Image Interpretation, Computer-Assisted - statistics & numerical data; Male; Middle Aged; Observer Variation; Pathologists - statistics & numerical data; Prospective Studies; Retrospective Studies; Sensitivity and Specificity; Stomach Neoplasms - diagnosis; Stomach Neoplasms - pathology
• ISSN: 1078-0432; 1557-3265
• DOI: 10.1158/1078-0432.ccr-20-3159

Gastric cancer remains the leading cause of cancer-related deaths in Northeast Asia. Population-based endoscopic screenings in the region have yielded successful results in early detection of gastric tumors. Endoscopic screening rates are continuously increasing, and there is a need for an automatic computerized diagnostic system to reduce the diagnostic burden. In this study, we developed an algorithm to classify gastric epithelial tumors automatically and assessed its performance in a large series of gastric biopsies and its benefits as an assistance tool. Using 2,434 whole-slide images, we developed an algorithm based on convolutional neural networks to classify a gastric biopsy image into one of three categories: negative for dysplasia (NFD), tubular adenoma, or carcinoma. The performance of the algorithm was evaluated by using 7,440 biopsy specimens collected prospectively. The impact of algorithm-assisted diagnosis was assessed by six pathologists using 150 gastric biopsy cases. Diagnostic performance evaluated by the AUROC curve in the prospective study was 0.9790 for two-tier classification: negative (NFD) versus positive (all cases except NFD). When limited to epithelial tumors, the sensitivity and specificity were 1.000 and 0.9749. Algorithm-assisted digital image viewer (DV) resulted in 47% reduction in review time per image compared with DV only and 58% decrease to microscopy. Our algorithm has demonstrated high accuracy in classifying epithelial tumors and its benefits as an assistance tool, which can serve as a potential screening aid system in diagnosing gastric biopsy specimens.