Gastric cancer possesses great histological and molecular diversity, which creates obstacles for rapid and efficient diagnoses. To overcome the limitations of the classic diagnostic procedure in gastric cancer, the authors established a deep learning system to achieve intelligent tumor differentiation grading and microsatellite instability status recognition using hematoxylin-eosin stained whole slide images from 467 patients. They used the convolutional neural network visualization to demonstrate the key pathological features learned by the system to increase system interpretability.
- Feng Su
- Jianmin Li
- Jiafu Ji
