A specialized reasoning large language model for accelerating rare disease diagnosis: a randomized AI physician assistance trial

📄 arXiv: 2606.24510v1 📥 PDF

作者: Haichao Chen, Songchi Zhou, Zhengyun Zhao, Shikai Hu, Xianghong Jin, Hongwei Ji, Li He, Shuli Li, Yiming Qin, Xin Tan, Runfeng Shi, Yih Chung Tham, Jiaye Zhu, Ye Li, Ye Jin, Longhao Cao, Dawei Li, Honghan Wu, Hongqiu Gu, Guanqiao Li, Tudor Groza, Chunying Li, Dian Zeng, Weihong Yu, Gareth Baynam, Saumya Shekhar Jamuar, Min Shen, Shuyang Zhang, Bin Sheng, Sheng Yu, Tien Yin Wong

分类: cs.AI, cs.CL

发布日期: 2026-06-23

备注: 36 pages, 5 figures

💡 一句话要点

提出RaDaR以加速罕见疾病诊断

🎯 匹配领域: 支柱九：具身大模型 (Embodied Foundation Models)

关键词: 罕见疾病 大型语言模型 推理增强 临床诊断 合成数据 开源模型 医疗AI 诊断准确性

📋 核心要点

  1. 现有的罕见疾病诊断方法面临专业知识稀缺和数据不足的挑战，导致诊断延误。
  2. RaDaR模型通过结合真实病例和合成数据进行推理增强训练，旨在提高罕见疾病的诊断效率。
  3. 实验结果显示，RaDaR在多个基准测试中表现优异，并在随机试验中显著提升了医生的诊断准确率。

📝 摘要（中文）

罕见疾病影响全球数百万人的健康，然而由于专业临床知识的稀缺，及时诊断仍然是一个重大公共卫生挑战。尽管大型语言模型（LLMs）在支持罕见疾病诊断方面展现出潜力，但现有模型在临床可部署性、临床证据的有限性和训练数据的稀缺性方面受到限制。本文提出了RaDaR（罕见疾病导航器），一个开源、紧凑的推理型大型语言模型（32B参数），专用于罕见疾病诊断。RaDaR在49,170个公开可用的自由文本病例和104,666个合成病例上进行了推理增强训练，表现出在评估的开源模型中最强的性能。在随机医生辅助试验中，RaDaR的辅助使医生的罕见疾病诊断准确率提高了21.44个百分点。

🔬 方法详解

问题定义：本文旨在解决罕见疾病诊断中的专业知识稀缺和训练数据不足的问题，现有方法在临床应用中面临较大挑战。

核心思路：RaDaR通过结合49,170个真实病例和104,666个合成病例进行推理增强训练，旨在提升模型的临床可用性和准确性。

技术框架：RaDaR的整体架构包括数据收集、模型训练、推理过程和验证阶段，采用了多种数据源以增强模型的泛化能力。

关键创新：RaDaR的主要创新在于其推理增强训练方法，利用合成数据和真实病例的结合，显著提升了模型在长尾罕见疾病上的表现。

关键设计：模型采用32B参数设置，训练过程中使用了特定的损失函数和网络结构，以优化推理能力和诊断准确性。通过对合成数据的消融实验，验证了表型锚定叙述对模型训练的重要性。

📊 实验亮点

RaDaR在多个公共基准测试中表现出色，超越了671B参数的DeepSeek-R1模型。在随机医生辅助试验中，RaDaR的使用使医生的罕见疾病诊断准确率提高了21.44个百分点，显示出其在实际应用中的显著效果。

🎯 应用场景

RaDaR模型在医疗领域的潜在应用广泛，尤其是在罕见疾病的早期诊断中。其开源特性和高效的推理能力使得医疗机构能够在数据稀缺的情况下，快速提高诊断准确性，从而改善患者的治疗效果和生活质量。未来，RaDaR有望与临床系统集成，成为医生的智能助手。

📄 摘要（原文）

Rare diseases affect millions of individuals worldwide, yet timely diagnosis remains a major public health challenge due to scarcity of specialized clinical expertise. While large language models (LLMs) show promise to support rare disease diagnosis, current models are constrained by insufficient clinical deployability, limited clinically grounded evidence, and scarcity of training data. Here we present RaDaR (Rare Disease navigatoR), an open-source, compact reasoning LLM (32B parameters) for rare disease diagnosis. RaDaR was trained with 49,170 publicly available free-text cases and 104,666 synthetic cases with reasoning-enhanced training. RaDaR showed the strongest performance among evaluated open-source models, including the 671B DeepSeek-R1, across public benchmarks and four external validation centers. In a retrospective cohort, RaDaR prioritized the final diagnosis before documented clinical suspicion in 61.06 percent of cases, corresponding to a potential lead time of 1.87 months and 50.18 percent of the within-center interval. In a randomized physician-assistance trial, RaDaR assistance improved physicians' rare-disease diagnostic accuracy by 21.44 percentage points compared with internet search alone. Synthetic-data ablations suggested that phenotype-anchored narratives provide useful training signal for long-tail rare diseases, with a monotonic scaling trend within the tested data range. Together, RaDaR and its development and validation framework provide a deployable rare-disease reasoning model and a reproducible development framework for diagnostic AI under data scarcity.