能看罕见病的AI系统！中国科研团队研发

Core Insights - A team led by professors from Shanghai Jiao Tong University has developed an AI system named DeepRare, which is the world's first "traceable reasoning" intelligent doctor for rare diseases, published in the journal Nature [1][3] Group 1: System Features - DeepRare provides a complete "evidence chain" with each diagnosis, allowing for greater trust from medical professionals compared to previous AI systems that only provided conclusions without justification [3] - The system simulates human reasoning by proposing hypotheses, validating them with evidence, and self-correcting if necessary, leading to a more thorough diagnostic process [3] Group 2: Performance Metrics - The system achieved an initial diagnostic accuracy of 57.18% based solely on clinical symptoms, which is an improvement of nearly 24 percentage points over the previous best international model [3] - When genetic data is included, the diagnostic accuracy can exceed 70% [3] Group 3: Implementation and Adoption - Since its launch in July of the previous year, over 600 medical institutions worldwide have registered to use the online diagnostic platform, including major hospitals in China and top laboratories in Europe and the U.S. [3] - The system has been deployed internally at Shanghai Xinhua Hospital and will assist doctors in diagnosing rare disease patients [3] Group 4: Future Plans - The research team plans to initiate a global alliance for rare disease diagnosis and treatment, aiming to validate the system with 20,000 real cases in the next six months [4]