Core Insights - Shanghai Jiao Tong University and Xinhua Hospital have launched DeepRare, the world's first intelligent rare disease diagnostic system, addressing the global challenge of high misdiagnosis rates in rare diseases [1][2] - DeepRare utilizes a "central-body" architecture, surpassing traditional medical AI by integrating vast medical literature and clinical cases, enabling a human-like reasoning process for diagnosis [1] - Clinical data shows that DeepRare's primary diagnostic accuracy for pure phenotype diagnosis is 57.18%, a 23.79 percentage point improvement over the best international models [1] Group 1 - DeepRare's system generates a complete evidence chain for each diagnosis, addressing the trust issues associated with AI in healthcare [1] - The system's accuracy improves to 70.61% when incorporating genetic sequencing data, significantly outperforming international tools [1] - The generated inference reports have achieved a 95.40% satisfaction rate among doctors, ensuring that diagnoses are well-supported [1] Group 2 - The online diagnostic platform for DeepRare was launched on July 26, 2025, attracting over 1,000 global professional users within six months [2] - The system has been deployed and tested at Xinhua Hospital, becoming part of the quality control process for rare disease diagnosis [2] - A "10,000 clinical validation plan" has been initiated, aiming to complete thousands of clinical validations through international collaboration [2]