摘要
人工智能(AI)技术正在以前所未有的速度重塑医疗模式,尤其在实验诊断学领域,AI亟待从简单的自动化工具发展为能够统筹多源信息并辅助临床决策的“实验诊断学综合智能体”。这一智能体以ISO 15189质量管理体系为指导,整合文本、图像、时序数据等多模态信息,通过自然语言交互、知识图谱和大语言模型等技术,实现从样本采集、物流调度、检测执行、质量控制、报告解读到资源运控的全流程智能化支持。其核心能力包括多模态数据协同分析、动态学习优化、人机协同决策和全流程自动化管理,从而显著提升诊断的准确性和效率。然而,构建这一智能体仍面临多重挑战,包括数据异构性导致的“信息孤岛”问题、AI模型的“黑箱”特性带来的临床可解释性不足、现有系统兼容性差,以及数据隐私保护、算法偏见等伦理与合规风险。该文还分析了实验诊断学对AI的临床需求,强调其在提升诊断效率、优化全流程质量控制、增强报告解读能力以及推动实验室管理智能化等方面的重要作用。在技术应用层面,该文回顾了AI在形态学诊断、检测流程优化、质量控制等领域的现状,展示了深度学习、知识图谱和大语言模型等技术的演进及其在提升诊断效率和精准度方面的潜力。跨界人才匮乏、硬件成本高昂以及数据标准化缺失等问题仍制约着AI技术的普及。展望未来,汇聚多学科、多组学的实验诊断学综合智能体将成为下一代智慧实验室的终极形态。实验诊断学综合智能体将重塑实验室工作模式,推动精准医疗发展,开启智慧实验诊断的新时代。
Artificial intelligence(AI)technology is reshaping medical paradigms at an unprecedented pace,particularly in the field of laboratory medicine.Here,AI has evolved from a simple automation tool into a comprehensive agent capable of integrating multi-source information to assist in clinical decision-making.Guided by the ISO 15189 quality management system,this agent integrates multimodal information encompassing text,images and time-series data and utilizes technologies such as natural language interaction,knowledge graphs and large language model(LLM)to achieve intelligent support across the entire workflow,ranging from specimen collection,logistics scheduling,and testing execution to quality control,report interpretation,and resource operation control.Its core capabilities include multimodal data collaborative analysis,dynamic learning optimization,human-machine collaborative decision-making,and full-process automated management,thereby significantly enhancing diagnostic accuracy and efficiency.Nevertheless,developing such an agent faces several challenges,such as data heterogeneity that creates information silos,limited clinical interpretability resulting from AI′s"black-box"nature,poor interoperability with existing systems,and ethical/compliance risks pertaining to data privacy and algorithmic bias.Furthermore,this article also analyzes the clinical demands of laboratory diagnostics for AI,emphasizing its crucial role in improving diagnostic efficiency,optimizing full-process quality control,enhancing report interpretation capabilities,and promoting the intelligent management of laboratories.At the technical application level,it reviews the current status of AI in morphological diagnosis,test process optimization,quality control and other areas,demonstrating the evolution of technologies such as deep learning,knowledge graphs and LLM and their potential in enhancing diagnostic efficiency and accuracy.Critical barriers,including a shortage of cross-disciplinary talent,high hardware costs,and insufficient data standardization,remain to hinder the widespread adoption of AI technology.Looking ahead,the multi-disciplinary and multi-omics joint comprehensive agent for laboratory diagnostics will become the ultimate form of the next-generation smart laboratory.The comprehensive agent for laboratory diagnostics will reshape laboratory work modes,facilitate the development of precision medicine,and usher in a new era of smart laboratory diagnostics.
作者
中国研究型医院学会检验医学专业委员会
重庆市医师协会检验医师分会
重庆市卫生健康信息学会数智检验专业委员会
张立群
府伟灵
Professional Committee of Laboratory Medicine,Chinese Research Hospital Association;Branch of Laboratory Physicians,Chongqing Medical Doctor Association;Professional Committee of Digital and Intelligent Laboratory Medicine,Chongqing Health Information Association;ZHANG Liqun;FU Weiling(不详;Department of Laboratory Medicine,the Second Affiliated Hospital of Army Medical University,Chongqing,400037,China;Department of Laboratory Medicine,the First Affiliated Hospital of Army Medical University,Chongqing,400038,China)
出处
《检验医学与临床》
2025年第24期3313-3321,共9页
Laboratory Medicine and Clinic
基金
国家自然科学基金项目(82472384)
重庆市科卫联合医学科研重点项目(2024ZDXM012)
重庆英才创新领军人才(CQYC20220303658)
重庆市自然科学基金创新发展联合基金重点项目(CSTB2025NSCQ-WZQLHJJZDX0007)。
关键词
实验诊断学
综合智能体
人工智能
数智检验
大语言模型
laboratory diagnostics
comprehensive agent
artificial intelligence
digital and intelligent laboratory medicine
large language model
