| 摘要: |
| ［摘要］ 人工智能（AI）作为引领新一轮科技革命的核心技术，已深度融入临床医学的各个环节。自20世纪50年代诞生以来，AI经历了从符号逻辑到深层神经网络的跨越式发展，目前已在疾病诊断、方案制订、药物研发及医院管理中展现出广阔的应用前景。国际监管方面，美国食品药品监督管理局（FDA）已建立并持续更新AI-Enabled Medical Device List，并对相关软件和器械实施风险分级与全生命周期监管。国内方面，中国在积极推进“AI+医疗卫生”行动、促进新技术临床转化的同时，监管部门亦同步加强制度建设，在医疗器械注册审评、AI医疗软件风险管理、数据安全与隐私保护等方面逐步完善规范要求，形成覆盖技术准入、应用实施与风险防控的监管体系，为AI技术在临床实践中的规范落地提供制度支撑。在上述国际与国内监管框架的共同推动下，AI技术正逐步应用于临床诊疗与公共卫生管理，在提升基层医疗服务能力、肿瘤早筛及流行病学管理等方面展现出重要应用前景。该文综述了AI在临床医学中的最新应用进展，涵盖计算机视觉与自然语言处理等技术基础，系统分析了其在实际应用中面临的数据隐私、算法透明度及法律责任界定等挑战，并强调AI医疗应用应坚持“以人为本”和“人类主导”原则，通过多中心验证与协同监管，推动精准医学的可持续发展。 |
| 关键词: 人工智能 医学影像 临床决策支持 多模态大模型 精准医学 伦理挑战 |
| DOI：10.3969/j.issn.1674-3806.2026.01.01 |
| 分类号:R 4；TP 18 |
| 基金项目: |
|
| Advances in the application of artificial intelligence in clinical medicine |
|
SHEN Weixi, ZHANG Jian
|
|
Shenzhen Tianyou Medical Research Institute, Shenzhen 518110, China
|
| Abstract: |
| ［Abstract］ Artificial intelligence(AI), as the core technology leading a new round of scientific and technological revolution, has been deeply integrated into every aspect of clinical medicine. Since its inception in the 1950s, AI has experienced a great-leap-forward development from symbolic logic to deep neural network. At present, it has shown broad application prospects in disease diagnosis, formulation of treatment plan, drug research and development, and hospital management. Regarding international regulation, the US Food and Drug Administration(FDA) has established and continuously updated the AI-Enabled Medical Device List, and applied a risk-based, total product life-cycle regulatory framework to the related softwares and medical devices. In China, the government is actively promoting the “AI+Healthcare” initiative to facilitate the clinical transformation of new technologies. Meanwhile, regulatory authorities are also simultaneously strengthening institutional building. They are gradually improving regulatory requirements in areas such as medical device registration review, risk management of AI medical software, data security and privacy protection, and establishing a regulatory system covering technology access, application implementation and risk prevention and control to provide institutional support for the standardized implementation of AI technology in clinical practice. Under the joint impetus of the above-mentioned international and domestic regulatory frameworks, AI technology is gradually being applied to clinical diagnosis and treatment as well as public health management, demonstrating significant application prospects in enhancing the capacity of primary medical services, early screening of tumors, and epidemiological management. This paper reviews the latest application progress of AI in clinical medicine, covering technical foundations such as computer vision and natural language processing. It systematically analyzes the challenges AI faces in practical applications, including data privacy, algorithmic transparency, and definition of legal liability. It also emphasizes that the application of AI in medical care should adhere to the principles of “people-oriented” and “human-led”, and the sustainable development of precision medicine should be promoted through multi-center validation and collaborative supervision. |
| Key words: Artificial intelligence(AI) Medical imaging Clinical decision support Multi-modal large models Precision medicine Ethical challenge |
