医学成像和放射治疗中的AI治理审查框架  

Artificial intelligence governance review framework for medical imaging and radiotherapy

作　　者：范文文 范晓娟 白桦 张红梅 王洁 Fan Wenwen;Fan Xiaojuan;Bai Hua;Zhang Hongmei;Wang Jie(Department of Radiology,National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital,Chinese Academy of Medical Sciences and Peking Union Medical College,Beijing 100021,China;College of Public Administration,Shandong Agricultural University,Taian 271018,China;State Key Laboratory of Molecular Oncology,AMS Key Laboratory of Translational Research on Lung Cancer,Department of Medical Oncology,National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital,Chinese Academy of Medical Sciences and Peking Union Medical College,Beijing 100021,China)

机构地区：[1]国家癌症中心/国家肿瘤临床医学研究中心/中国医学科学院北京协和医学院肿瘤医院影像诊断科,北京100021 [2]山东农业大学公共管理学院/泰山法治研究院,山东泰安271000 [3]国家癌症中心/国家肿瘤临床医学研究中心/中国医学科学院北京协和医学院肿瘤医院分子肿瘤学国家重点实验室,AMS肺癌转化研究重点实验室肿瘤内科,北京100021

出　　处：《中华临床医师杂志(电子版)》2024年第11期973-979,共7页Chinese Journal of Clinicians(Electronic Edition)

摘　　要：医学成像是最早采用人工智能(AI)技术的学科之一。目前,AI/ML(机器学习)模型已经开始用于辅助影像诊断与治疗。然而,基于AI驱动的广泛临床应用,仍需要在准确可信任的技术应用与数据隐私保护为代表的法律伦理中寻找一种创新基础。在医学成像与放射领域,AI治理框架构建需求呈指数级增长。本文基于实用性基本伦理原则,同时考虑到AI模型的长短期验证与评估程序,提出一个实用治理框架,旨在引导医学成像与放射领域的AI应用。Medical imaging is one of the earliest disciplines to adopt artificial intelligence(AI)technology.Currently,AI/ML(machine learning)models are being used to assist in imaging diagnosis and treatment.However,the widespread clinical application of AI necessitates finding an innovative foundation that balances reliable,trustworthy technology with legal and ethical considerations such as data privacy protection.The demand for constructing AI governance frameworks in the field of medical imaging and radiology is growing exponentially.This paper proposes a practical governance framework based on fundamental ethical principles,considering both short-term and long-term validation and evaluation procedures of AI models,with an aim to guide AI applications in the field of medical imaging and radiology.

关 键 词：人工智能 影像诊断 数据黑箱 隐私权 治理框架 

分 类 号：R-05[医药卫生]