Related Experiment Video
Updated: Sep 14, 2025
05:33
Introduction of an Integrated Pathology Image Management, Artificial Intelligence, and Reporting System
Published on: July 11, 2025
255
An end-to-end multifunctional AI platform for intraoperative diagnosis
Xueyi Zheng1, Ke Zheng1, Yongqin Wen2
1Department of Pathology, State Key Laboratory of Oncology in South China, Guangdong Provincial Clinical Research Center for Cancer, Sun Yat-sen University Cancer Center, Guangzhou, China.
NPJ Digital Medicine
|July 20, 2025
Summary
This study introduces GAS, an AI platform enhancing intraoperative frozen section quality. The system improves diagnostic confidence for pathologists, offering real-time histological insights for surgical decisions.
Area of Science:
- Pathology
- Artificial Intelligence
- Medical Imaging
Background:
- Intraoperative frozen section diagnosis is crucial for surgical guidance.
- Suboptimal section quality presents diagnostic challenges for pathologists.
- Existing methods lack real-time histological insights.
Purpose of the Study:
- To develop a comprehensive AI platform (GAS) for enhancing frozen section quality.
- To improve the diagnostic accuracy and confidence of pathologists during intraoperative procedures.
- To integrate end-to-end AI solutions into clinical workflows.
Main Methods:
- Developed GAS, a platform with Generation, Assessment, and Support modules, using over 6700 whole slide images.
- Employed a GAN-driven multimodal network for image enhancement guided by FFPE-style text descriptions.
- Utilized pathological foundation models for quality control and fine-tuned assessment models.
Main Results:
- The Generation module effectively enhanced frozen section quality across various organs.
- The Assessment module demonstrated significant microstructural quality improvements in generated images.
- A prospective study validated that GAS boosted pathologists' diagnostic confidence.
Conclusions:
- The GAS platform shows significant clinical utility in intraoperative diagnosis.
- GAS establishes a new paradigm for integrating AI into surgical pathology workflows.
- AI-driven enhancement of frozen sections can overcome current diagnostic limitations.
