Deep Learning Models for Histopathological Classification of Gastric and Colonic Epithelial Tumours

TLDR

Histopathological classification of gastric and colonic epithelial tumours is a routine diagnostic task, and AI‑based computational pathology could alleviate pathologists’ workloads, especially in resource‑limited regions. The study trains convolutional neural networks and recurrent neural networks on stomach and colon biopsy whole‑slide images to classify gastric and colonic epithelial tumours. The CNNs and RNNs were trained to classify whole‑slide images into adenocarcinoma, adenoma, and non‑neoplastic categories. On three independent test sets, the models achieved AUCs of 0.97–0.99 for gastric adenocarcinoma and adenoma and 0.96–0.99 for colonic adenocarcinoma and adenoma, demonstrating strong generalisation and promising potential for clinical deployment.

Abstract

Abstract Histopathological classification of gastric and colonic epithelial tumours is one of the routine pathological diagnosis tasks for pathologists. Computational pathology techniques based on Artificial intelligence (AI) would be of high benefit in easing the ever increasing workloads on pathologists, especially in regions that have shortages in access to pathological diagnosis services. In this study, we trained convolutional neural networks (CNNs) and recurrent neural networks (RNNs) on biopsy histopathology whole-slide images (WSIs) of stomach and colon. The models were trained to classify WSI into adenocarcinoma, adenoma, and non-neoplastic. We evaluated our models on three independent test sets each, achieving area under the curves (AUCs) up to 0.97 and 0.99 for gastric adenocarcinoma and adenoma, respectively, and 0.96 and 0.99 for colonic adenocarcinoma and adenoma respectively. The results demonstrate the generalisation ability of our models and the high promising potential of deployment in a practical histopathological diagnostic workflow system.