AI Enhanced Pathology: Improving Metastasis Detection Accuracy with Deep Learning in Gastric Cancer

# min read

  • Digital Health
  • Artificial Intelligence
  • Davies Award
  • APAC

Chang Gung Memorial Hospital (CGMH) in Linkou developed an AI-enhanced pathology system to improve the detection of lymph node metastasis in gastric cancer. Using a deep learning model called the Enhanced Streaming Convolutional Neural Network (ESCNN), the system analyzes digital pathology slides to identify even small cancerous lesions that are often missed during manual review.

The AI-assisted workflow has significantly increased diagnostic accuracy, improving micrometastasis detection from 82% to 96% and reducing review time by more than 30%. It also minimizes interobserver variability, ensuring consistent and timely cancer staging that directly supports more effective treatment planning.

The project contributed to CGMH’s recognition as a 2025 HIMSS Davies Award winner.