Artificial intelligence-enabled cancer diagnosis, type classification and mutation prediction

Data-driven computational oncology promises to leverage the power of big data and machine learning techniques to transform and revolutionise clinical diagnosis and decision making to improve cancer treatment and healthcare. In our preliminary studies, we have developed a deep learning-based framework termed HEAL to streamline the whole-slide image (WSI) analysis. This project aims to develop an AI system for gastric cancer diagnosis, molecular typing classification, and prognosis from histopathological images and benchmark the performance of such models based on different deep learning architectures. We will address these questions using high-performance GPU supercomputers using >1.6 million tile-level images. This AI system will be valuable for gastric cancer diagnosis and risk assessment and help identify pathologically and clinically relevant image patches and regions of interest from the WSIs.