A team of Chinese researchers has developed the nation’s first versatile artificial intelligence (AI) model proficient in analyzing a wide array of pathological images. This groundbreaking model, named PathOrchestra, can examine more than 20 human organs, including the lungs, breast, and liver, marking a significant advancement in AI-assisted disease diagnosis.
Developed collaboratively by researchers from the Air Force Medical University (AFMU), Tsinghua University, and SenseTime, PathOrchestra represents a transformative shift from specialized cancer analysis to a comprehensive approach capable of addressing multiple cancers simultaneously. Leveraging China’s largest domestic dataset of nearly 300,000 whole-slide digital pathology images—amounting to an impressive 300 terabytes of data—the team harnessed self-supervised learning to enable the model to ‘cross-learn’ and analyze diverse organs.
The complexity and diversity inherent in pathological images have historically posed significant challenges for AI applications. Wang Zhe, a professor from the Basic Medical Science Academy under the AFMU, referred to this complexity as the ‘jewel in the crown’ of image processing. Despite these challenges, PathOrchestra has achieved an accuracy rate exceeding 95 percent in nearly 50 clinical tasks, including lymphoma subtype diagnosis and bladder cancer screening, according to an AFMU news release on Tuesday.
The model’s ability to perform tasks such as pan-cancer classification, lesion identification and detection, multi-cancer subtype differentiation, and biomarker assessment demonstrates its potential to substantially reduce the workload of pathologists. ‘This advancement notably increases the efficiency of reviewing medical images,’ the researchers stated.
PathOrchestra stands as a testament to the rapid and vibrant growth of AI in China. Among the more than 1,300 AI large language models (LLMs) globally, 36 percent are from China, the second-largest proportion after the United States, according to a white paper on the global digital economy released recently by the China Academy of Information and Communications Technology at the Global Digital Economy Conference 2024.
Reference(s):
Chinese team develops AI model capable of reading cancer images
cgtn.com
