DC3 Diagnostic Case Challenges
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/codiag-public/dc3
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是由临床专家手工整理并解决的31个具有挑战性的诊断案例，其中包括医生生成的观察结果和确认的诊断。此外，该数据集还包含了针对2018年美国国家医学图书馆PubMed数据库快照的密集相关性判断，这些判断识别出与目标诊断相关的论文。规模上，数据集包含了31个复杂的诊断案例。其任务是对诊断决策支持及评估诊断决策支持系统。
This dataset comprises 31 challenging diagnostic cases manually curated and resolved by clinical experts, including physician-generated clinical observations and confirmed diagnoses. Furthermore, it contains dense relevance judgments based on the 2018 snapshot of the PubMed database from the U.S. National Library of Medicine, which identify academic papers relevant to the target diagnoses. In terms of scale, the dataset includes 31 complex diagnostic cases. The primary purpose of this dataset is to support and evaluate diagnostic decision support systems.
提供机构：
Massachusetts General Hospital (MGH)
