大语言模型与知识图谱在体检报告解读中的应用  

Application of Large Language Models and Knowledge Graphs in Medical Examination Report Interpretation

作　　者：朱静[1] 赵艳[1] ZHU Jing;ZHAO Yan(Ruijin Hospital,Shanghai Jiaotong University School of Medicine,Shanghai 200025,China)

机构地区：[1]上海交通大学医学院附属瑞金医院信息中心,上海市200025

出　　处：《中国卫生信息管理杂志》2025年第1期32-37,共6页Chinese Journal of Health Informatics and Management

摘　　要：目的 提出一种基于大语言模型与知识图谱的体检报告智能解读系统,旨在解决传统方法效率低、个性化不足的问题。方法 体检报告智能解读系统基于结构化的医疗知识图谱,结合检索增强生成技术与大语言模型的深度协同,能够从体检报告中自动识别潜在疾病风险,并生成针对性的健康管理建议。结果 实验结果表明,本系统在准确性和体检者满意度方面显著优于传统方法,准确率提升12.8%,体检者满意度评分达4.7分(满分5分)。结论 融合知识的大语言模型不仅提升了体检报告的解读质量,还为医疗健康领域智能化发展提供了创新路径。Objective This paper proposes an intelligent interpretation for medical examination reports that integrates Large Language Models(LLMs)with knowledge graphs,aiming to address the issues of low efficiency and lack of personalization in traditional methods.Methods The intelligent physical examination report interpretation system,grounded in a structured medical knowledge graph and integrating the deep-seated collaboration of Retrieval-augmented generation technology and LLMs,is capable of automatically identifying potential disease risks from physical examination reports and generating targeted health management recommendations.Results The results indicate that the framework significantly outperforms traditional methods in terms of accuracy and user satisfaction,with an increase in accuracy of 12.8%and a user satisfaction rating of 4.7 out of 5.Conclusion The integration of knowledge with large language models not only enhances the quality of interpretation but also paves the way for innovative development in thefield of medical health intelligence.

关 键 词：检索增强生成 知识图谱 大语言模型 

分 类 号：R-39[医药卫生] R319