文献速递医学人工智能与信息学基础与转化研究

用于疾病诊断辅助的通用医学语言模型

A generalist medical language model for disease diagnosis assistance.
PubMed 文献 · Nat Med · 2025-01-08

实验显示该模型在八个专科的分布内(常见病)、分布外(外部验证)和长尾分布(罕见病)场景中优于其他基线大语言模型和专用模型,并通过涉及人工智能与医生对比、人工智能辅助研究和人类评估框架的临床适用性综合评估,表明该模型作为临床工作流一部分辅助医生进行疾病诊断的可行性。

要点

  • 采用自举策略的思维链方法学习医生推理诊断,并引入统一偏好对齐框架与标准临床实践对齐。
  • 在八个专科的分布内、分布外和长尾分布场景中,模型优于其他基线大语言模型和专用模型。

核心结论

  • 研究结果表明该模型作为临床工作流一部分辅助医生进行疾病诊断具有可行性。

研究设计与样本

  • 进一步微调MedFound,采用基于自举策略的思维链方法学习医生的推理诊断,并引入统一偏好对齐框架使其与标准临床实践对齐。
  • 开展涉及人工智能与医生对比、人工智能辅助研究和人类评估框架的综合评估,框架包含八个临床评估指标,涵盖病历摘要、诊断推理和风险管理等能力。

主要发现

  • 在八个专科的分布内(常见病)、分布外(外部验证)和长尾分布(罕见病)场景中,该医学大语言模型优于其他基线大语言模型和专用模型。
  • 消融研究进一步表明该医学大语言模型训练方法中关键组件的有效性。

临床怎么用

  • 原文未据此给出推荐。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

The delivery of accurate diagnoses is crucial in healthcare and represents the gateway to appropriate and timely treatment. Although recent large language models (LLMs) have demonstrated impressive capabilities in few-shot or zero-shot learning, their effectiveness in clinical diagnosis remains unproven. Here we present MedFound, a generalist medical language model with 176 billion parameters, pre-trained on a large-scale corpus derived from diverse medical text and real-world clinical records. We further fine-tuned MedFound to learn physicians' inferential diagnosis with a self-bootstrapping strategy-based chain-of-thought approach and introduced a unified preference alignment framework to align it with standard clinical practice. Extensive experiments demonstrate that our medical LLM outperforms other baseline LLMs and specialized models in in-distribution (common diseases), out-of-distribution (external validation) and long-tailed distribution (rare diseases) scenarios across eight specialties. Further ablation studies indicate the effectiveness of key components in our medical LLM training approach. We conducted a comprehensive evaluation of the clinical applicability of LLMs for diagnosis involving artificial intelligence (AI) versus physician comparison, AI-assistance study and human evaluation framework. Our proposed framework incorporates eight clinical evaluation metrics, covering capabilities such as medical record summarization, diagnostic reasoning and risk management. Our findings demonstrate the model's feasibility in assisting physicians with disease diagnosis as part of the clinical workflow.

来源:Nat Med,查看原文。版权归原作者所有。

引用

  1. Liu X, Liu H, Yang G, et al. A generalist medical language model for disease diagnosis assistance. Nat Med. 2025. PMID: 39779927. doi:10.1038/s41591-024-03416-6

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。