LiteOdyssey: 一种用于可解释罕见病诊断的轻量级推理AI智能体
Teaching agentic AI to generalize expert diagnostic reasoning in rare diseases
AI总结：
提出轻量级框架LiteOdyssey，通过人类-AI协作的诊断策略和公共生物医学工具增强单个推理语言模型，在罕见病诊断基准上达到最先进性能，无需微调或多智能体集成。
摘要：
大多数医疗AI系统通过扩展额外机制来改进：更多的微调数据、更多的智能体和/或更大的检索数据库。然而，在罕见病诊断中，这种扩展可能导致系统难以部署、审计和维护。我们探究是否可以通过扩展单个AI智能体的推理链来实现最先进的诊断性能：通过人类-AI协作开发的诊断策略指导它，并利用可免费获取的生物医学工具进行增强。我们引入了LiteOdyssey，一个轻量级罕见病诊断框架，通过临床遗传学工作流引导推理语言模型。该框架通过人类反馈的策略迭代（PIHF）开发，并动态访问公共生物医学工具。在两个仅提供患者临床特征的高难度基准上，LiteOdyssey取得了最先进的性能，在LIRICAL（n=370）和PhenoPacket Store（n=873）的合并1243个病例中，总体疾病Recall@1达到59.3%。这两个基准中超高罕见病（患病率低于1/1,000,000）的比例很高，分别约为45%和52.8%。在更困难的PhenoPacket子集上（其中因果疾病未在我们的稀有性映射流程中映射到Orphanet），LiteOdyssey实现了60.7%的Recall@1，而相同基线模型（GPT-5.4）不使用工具时为10.7%。这一性能是在没有微调、多智能体集成或大型病例检索数据库的情况下实现的。在开发过程中未见过的病例、真实世界罕见病患者的私人队列以及较小的开源权重模型上也观察到了增益。LiteOdyssey为罕见病AI系统指明了一条路径，使其准确、易于部署且对医生审查更透明。
英文摘要：
Rare disease diagnosis depends on expert reasoning that is scarce and difficult to transfer. Large language models rank the correct disease first in only 35.4% of benchmark cases and often rely on learned phenotype-disease associations rather than reusable diagnostic reasoning strategies. We developed liteOdyssey through Policy Iteration with Human Feedback, a process in which model failures and expert corrections are iteratively consolidated into a clinician-gated, natural-language policy executed by a language model. Across 1,243 public benchmark cases spanning 722 rare diseases, liteOdyssey ranked the correct disease first in 59.3% of cases versus 26.5% without the policy, with comparable gains in cases involving diseases excluded from policy development. The same policy transferred across model families and sizes without retraining. Adaptation of the policy to the Undiagnosed Diseases Network (UDN) improved diagnostic accuracy among 515 UDN patients, with gains confirmed by blinded physician adjudication. These results show that expert reasoning can be externalized into an inspectable and revisable natural-language policy that generalizes across rare diseases, transfers across model backbones, and adapts to a real-world patient cohort.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
