Related Experiment Video
Updated: Aug 6, 2026
05:33
Introduction of an Integrated Pathology Image Management, Artificial Intelligence, and Reporting System
Published on: July 11, 2025
Deep learning-based computational pathology: Technologies, clinical applications, and future directions
Yunqiu Gao1,2,3,4,5, Teng Ma1,2, Lisha Li1,2
1College of Medicine and Biological Information Engineering, Northeastern University, Shenyang, Liaoning 110169, China.
Chinese Medical Journal
|July 22, 2026
Summary
Artificial intelligence (AI) in digital pathology faces challenges translating research to practice. This review proposes a framework linking AI methods with clinical tasks for better diagnostic and prognostic AI tools.
Area of Science:
- Digital pathology and artificial intelligence (AI)
- Deep learning in medical imaging
Background:
- Whole-slide imaging is transforming pathology, but AI integration faces hurdles.
- A gap exists between AI algorithmic advances and clinical pathology needs for diagnosis and prognosis.
Purpose of the Study:
- To propose a dual-perspective framework bridging AI methodologies and clinical pathology tasks.
- To provide a comprehensive overview of AI in pathology from 2020-2025.
- To guide the selection and design of AI solutions for specific clinical goals.
Main Methods:
- Review of deep learning architectures (CNNs, Vision Transformers, GNNs) applied to pathology.
- Analysis of AI for diagnostic classification, tissue segmentation, and prognostic prediction.
- Exploration of weakly supervised, self-supervised learning, and prediction of IHC from H&E slides.
Main Results:
- A novel algorithm-clinical task mapping framework is presented.
- Emerging trends minimize reliance on costly annotations.
- Advances enable predicting immunohistochemistry results from H&E slides.
Conclusions:
- Addressing challenges in interpretability, regulation, and generalization is crucial.
- Future AI systems should be integrated, trustworthy, and equitable.
- AI aims to augment, not replace, pathologist expertise.
