Chinese Team Develops AI-Powered System for Illness Diagnosis

Shanghai: A Chinese research team has devised a groundbreaking AI-powered system for diagnosing rare illnesses, dubbed DeepRare, which broke a new record in terms of diagnosis. This evidence-based system has been developed by a team from the Schools of Medicine and Artificial Intelligence at Shanghai Jiao Tong University. According to Oman News Agency, having launched its diagnosis portal last July, the system recorded over 1,000 professional users from more than 600 medical and research facilities worldwide. The test data showed that when relying solely on patients' clinical phenotype information without genetic data, the system in question achieved an accuracy of 57.18% in initial phenotypic diagnosis, representing an increase of nearly 24 percentage points compared to the previous global model. When genetic data were integrated, diagnostic accuracy exceeded 70%. DeepRare provides immediate access to a massive repository of knowledge derived from medical literature and real clinical case data. Regarding its diagnostic reasoning mechanism, the system leans on an iterative cycle that includes hypothesis generation, verification, and self-review to assess diagnostic evidence and correct logical gaps. Sun Kun, one of the corresponding authors of the study, said the research team is preparing to launch a global AI alliance for the diagnosis and treatment of rare diseases, noting that the team intends to complete real-world validation of 20,000 rare disease cases over the next six months.