综述
ENGLISH ABSTRACT
ChatGPT在眼科领域潜在应用的研究进展
吴世楠
谭钢 [综述]
邵毅 [综述]
作者及单位信息
·
DOI: 10.3760/cma.j.cn115989-20230919-00101
Advancement of ChatGPT in ophthalmology
Wu Shinan
Tan gang
Shao Yi
Authors Info & Affiliations
Wu Shinan
Eye Institute of Xiamen University, Xiamen 361005, China
Tan gang
Department of Ophthalmology, The First Affiliated Hospital of University of South China, Hengyang 421000, China
Shao Yi
Department of Ophthalmology, Shanghai General Hospital, Shanghai Jiao Tong University School of Medicine, National Clinical Research Center for Eye Diseases, Shanghai 200080, China
·
DOI: 10.3760/cma.j.cn115989-20230919-00101
147
45
0
0
1
0
PDF下载
APP内阅读
摘要

在通用人工智能(AGI)初期,ChatGPT大语言模型的涌现彰显了其在医疗保健领域中改善患者护理、拓展医疗获取和优化临床决策过程的潜力。然而,将它们融入医疗保健系统中仍需深思熟虑,以防范潜在风险,如错误的医疗建议、侵犯患者隐私、创造虚假的参考文献和图像,以及医学教育中学生过度依赖AGI等问题。必须实施适当的监督和管理,以应对这些潜在风险,确保AGI技术在眼科领域中的安全有效应用。通过不断改进其不足之处,充分利用AGI来提供眼科疾病科普及背景知识、优化医疗知识总结和改善医疗保健流程,最终可使整个社会获益。

眼科;深度学习;大语言模型;人工智能;ChatGPT
ABSTRACT

In the early stage of artificial general intelligence (AGI), the emergence of large language models like ChatGPT highlights their potential in improving patient care, expanding healthcare access, and optimizing clinical decision-making processes in the healthcare sector.However, their integration into healthcare systems requires careful consideration to mitigate potential risks, such as incorrect medical advice, infringement of patient privacy, creation of false references and images, and the overreliance of students on AGI in medical education.It is crucial to implement appropriate supervision and regulation to address these potential risks and ensure the safe and effective application of AGI technology in the field of ophthalmology.By continuously improving its limitations, AGI can be utilized to enhance the dissemination of disease knowledge and background information for ophthalmology patients, optimize the summarization of medical knowledge and improve healthcare processes, ultimately benefiting society as a whole.

Ophthalmology;Deep learning;Large language models;Artificial intelligence;ChatGPT
Tan Gang, Email: mocdef.labiamtoh99gnagnaT;
Shao Yi, Email: mocdef.3ab6199eebeerf
引用本文

吴世楠,谭钢,邵毅. ChatGPT在眼科领域潜在应用的研究进展[J]. 中华实验眼科杂志,2025,43(02):181-185.

DOI:10.3760/cma.j.cn115989-20230919-00101

PERMISSIONS

Request permissions for this article from CCC.

评价本文
*以上评分为匿名评价
ChatGPT是由OpenAI开发的一种人工智能(artificial intelligence,AI)大语言模型(large language model,LLM)。该模型基于Transformer架构,最初由Vaswani等 [ 1 ]在2017年提出。ChatGPT核心原理是基于深度学习和自然语言处理技术,其基本思想是通过大规模语料库的预训练,使模型能够理解并学到语言的语法、语义及上下文关系。在Transformer架构的基础上,通过在大规模文本语料库上进行预训练,学习语言的一般规律和知识。在针对特定任务或应用领域时,模型可以进行微调以适应具体需求。最终,通过生成式任务创作文本,能够准确回应用户输入,记忆并理解上下文信息 [ 2 , 3 ]。ChatGPT等LLM正在不断向通用人工智能(artificial general intelligence,AGI)的方向发展 [ 4 ]
高级生成式AGI具有改善患者护理和简化医疗保健流程的潜力 [ 5 , 6 ]。例如,ChatGPT可以辅助医师和护理人员进行诊断、治疗选择和个性化护理计划设计 [ 7 , 8 , 9 , 10 ]。此外,ChatGPT有望执行医疗领域的各种任务,如生成患者摘要、医疗报告,通过提供基于症状和病史的相关信息进行诊断,并通过分析广泛的医学文献为医学研究提供摘要性综述。ChatGPT的高级语言生成能力有望协助医疗保健专业人员生成精确且全面的手术记录。已有研究将GPT-4应用于生成白内障手术记录,包括患者信息、手术日期、手术过程细节、人工晶状体细节、术后护理和术后并发症部分。GPT-4正确地提到了白内障手术的关键步骤,如切口、曲线囊破裂、白内障超声乳化、晶状体植入和伤口闭合过程 [ 11 ]。此外,GPT-4还有助于推动医疗创新、总结最新的临床试验和提供有关伦理准则的相关信息。GPT-4在美国眼科学会基础和临床科学课程模拟考试中表现优异,对眼科问题回答的准确性显著优于眼科住院医师和执业医师 [ 12 ]。因此,眼科医师应持续关注不断发展的AI系统,随着这些系统的不断改进和完善,它们可能会越来越多地用于辅助医疗保健和医疗决策。另外,一项对StatPearls question bank中的眼科相关题目进行GPT-3.5、GPT-4和人类的对比得分评估显示,GPT-4的性能有显著提升 [ 13 ]。这一研究结果突显了先进的对话AI系统作为医学教育和实践的重要工具的潜力。
然而,如果没有对AGI进行适当的监督和管理,将LLM集成到医疗保健系统中可能会引入一系列风险。《眼科人工智能临床应用伦理专家共识(2023)》 [ 14 ]提出了AI在医学研究中应用的伦理规范,用于保障患者数据安全和个人权益。截至目前,尽管ChatGPT可以访问大量医学知识,但如果不使用网络工具,其无法提供最新的研究和指南,这对于快速变化的医学领域来说是一个明显的局限性;且ChatGPT提供的信息仅限于一般性知识和解释,不能对具体患者的情况提供个性化的诊疗建议;同时在处理复杂或高度专业化的医学问题时,如需要权衡多个诊断或详细讨论新型疗法,ChatGPT可能缺乏足够的专业深度和权威性。因此,在临床上,眼病患者在使用ChatGPT等LLM寻求相应诊治意见时,可能存在过时或不适合该患者病情的治疗建议。因此,临床医师和计算机背景的相关人员应重视并及时解决上述问题,以确保AGI在医疗领域和眼科专科中提供合适的帮助。截至2024年11月20日,我们通过在PubMed数据库检索("ophthalmology"OR"eye diseases"OR"ophthalmic disorders"OR"ocular health"OR"ocular diseases"OR"ophthalmic conditions") AND ("ChatGPT"OR"language model")关键词,ChatGPT在眼科方面的研究有243篇,眼科方向综述性研究27篇。Nath等 [ 15 ]提出GPT-3有能力改善眼科临床资源的合理使用,改善临床流程,并将AI解决方案的开发惠及临床医师和眼病患者。Salimi等 [ 16 ]提出ChatGPT在眼科研究中使用LLM的伦理问题,呼吁眼科期刊更新并制定论文写作过程中LLM伦理使用指南。Jin等 [ 17 ]提出了下一代眼科AI的LLM,临床医师和护理人员通过利用AI,获取丰富的信息,提高诊断的准确性,为患者提供个性化护理方案。本文从眼科图像分析、眼科疾病诊治处方、医疗用品监督和眼病手术治疗几个方面总结了未来GPT在眼科中的应用( 表1 )。
应用领域 描述
眼科图像分析 ChatGPT可用于自动分析眼科图像,识别眼底照片中的病变,检测视网膜疾病,并提供初步的筛查和辅助诊断
眼科疾病诊治处方 在给定患者病历和症状的情况下,ChatGPT可以为眼科医师生成诊断和治疗处方,提供定制的医疗建议
医疗用品监督 ChatGPT可用于监督医疗用品的生产、质量控制和供应链管理,确保眼科领域的药品和器械符合标准
眼病手术治疗 ChatGPT能够为眼科手术提供术前规划、手术技术说明和术后护理建议,帮助医师提高手术的精准性和安全性
ChatGPT在眼科领域的潜在应用
试读结束,您可以通过登录机构账户或个人账户后获取全文阅读权限。
参考文献
[1]
Vaswani A Shazeer N Parmar N et al. Attention is all you need[J/OL]. arXiv 201730[2024-05-15]. https://arxiv.org/pdf/1706.03762.
返回引文位置Google Scholar
百度学术
万方数据
[2]
Waisberg E Ong J Masalkhi M et al. GPT-4:a new era of artificial intelligence in medicine[J]. Ir J Med Sci 2023192(6)∶3197-3200. DOI: 10.1007/s11845-023-03377-8 .
返回引文位置Google Scholar
百度学术
万方数据
[3]
Ray P ChatGPT:a comprehensive review on background,applications,key challenges,bias,ethics,limitations and future scope[J]. Int Things Cyber-Physical Systems 20233121-154. DOI: 10.1016/j.iotcps.2023.04.003 .
返回引文位置Google Scholar
百度学术
万方数据
[4]
Zhao L Zhang L Wu Z et al. When brain-inspired AI meets AGI[J/OL]. Meta-Radiology 20231(1)∶100005[2024-05-15]. https://www.sciencedirect.com/science/article/pii/S295016282300005X?via%3Dihub. DOI: 10.1016/j.metrad.2023.100005 .
返回引文位置Google Scholar
百度学术
万方数据
[5]
Li J Dada A Puladi B et al. ChatGPT in healthcare:a taxonomy and systematic review[J/OL]. Comput Methods Programs Biomed 2024245108013[2024-11-15]. https://pubmed.ncbi.nlm.nih.gov/38262126/. DOI: 10.1016/j.cmpb.2024.108013 .
返回引文位置Google Scholar
百度学术
万方数据
[6]
Thrall JH Li X Li Q et al. Artificial intelligence and machine learning in radiology:opportunities,challenges,pitfalls,and criteria for success[J]. J Am Coll Radiol 201815(3)∶Pt B-504. 508DOI: 10.1016/j.jacr.2017.12.026 .
返回引文位置Google Scholar
百度学术
万方数据
[7]
Waisberg E Ong J Masalkhi M et al. Text-to-image artificial intelligence to aid clinicians in perceiving unique neuro-ophthalmic visual phenomena[J]. Ir J Med Sci 2023192(6)∶3139-3142. DOI: 10.1007/s11845-023-03315-8 .
返回引文位置Google Scholar
百度学术
万方数据
[8]
Waisberg E Ong J Paladugu P et al. Challenges of artificial intelligence in space medicine[J/OL]. Space Sci Technol 2022[2024-05-15]. https://www.semanticscholar.org/paper/Challenges-of-Artificial-Intelligence-in-Space-Waisberg-Ong/91850e845f0b5b2bf93f018875949b2dad197e55. DOI: 10.34133/2022/9852872 .
返回引文位置Google Scholar
百度学术
万方数据
[9]
Gunawan J Exploring the future of nursing:insights from the ChatGPT model[J]. Belitung Nurs J 20239(1)∶1-5. DOI: 10.33546/bnj.2551 .
返回引文位置Google Scholar
百度学术
万方数据
[10]
Dağci M Çam F Dost A Reliability and quality of the nursing care planning texts generated by ChatGPT[J/OL]. Nurs Educ 202349(3)∶E109-E114[2024-05-15]. https://journals.lww.com/nurseeducatoronline/abstract/2024/05000/reliability_and_quality_of_the_nursing_care.11.aspx. DOI: 10.1097/NNE.0000000000001566
返回引文位置Google Scholar
百度学术
万方数据
[11]
Waisberg E Ong J Masalkhi M et al. GPT-4 and ophthalmology operative notes[J]. Ann Biomed Eng 202351(11)∶2353-2355. DOI: 10.1007/s10439-023-03263-5 .
返回引文位置Google Scholar
百度学术
万方数据
[12]
Lin JC Younessi DN Kurapati SS et al. Comparison of GPT-3.5,GPT-4,and human user performance on a practice ophthalmology written examination[J]. Eye (Lond) 202337(17)∶3694-3695. DOI: 10.1038/s41433-023-02564-2 .
返回引文位置Google Scholar
百度学术
万方数据
[13]
Moshirfar M Altaf AW Stoakes IM et al. Artificial intelligence in ophthalmology:a comparative analysis of GPT-3.5,GPT-4,and human expertise in answering statpearls questions[J/OL]. Cureus 202315(6)∶e40822[2024-05-15]. https://pubmed.ncbi.nlm.nih.gov/37485215/. DOI: 10.7759/cureus.40822 .
返回引文位置Google Scholar
百度学术
万方数据
[14]
《眼科人工智能临床应用伦理专家共识(2023)》专家组中国医药教育协会数字影像与智能医疗分会中国医药教育协会智能医学专业委员会眼科人工智能临床应用伦理专家共识(2023)[J]. 中华实验眼科杂志 202341(1)∶1-7. DOI: 10.3760/cma.j.cn115989-20220905-00414 .
返回引文位置Google Scholar
百度学术
万方数据
Expert Workgroup of Expert consensus for ethics of clinical application of artificial intelligence in ophthalmology (2023)Digital Imaging and Intelligent Medicine Branch of China Medical Education AssociationIntelligent Medicine Special Committee of China Medical Education Association. Expert consensus for ethics of clinical application of artificial intelligence in ophthalmology (2023)[J]. Chin J Exp Ophthalmol 202341(1)∶1-7. DOI: 10.3760/cma.j.cn115989-20220905-00414 .
Goto CitationGoogle Scholar
Baidu Scholar
Wanfang Data
[15]
Nath S Marie A Ellershaw S et al. New meaning for NLP:the trials and tribulations of natural language processing with GPT-3 in ophthalmology[J]. Br J Ophthalmol 2022106(7)∶889-892. DOI: 10.1136/bjophthalmol-2022-321141 .
返回引文位置Google Scholar
百度学术
万方数据
[16]
Salimi A Saheb H Large language models in ophthalmology scientific writing:ethical considerations blurred lines or not at all?[J]. Am J Ophthalmol 2023254177-181. DOI: 10.1016/j.ajo.2023.06.004 .
返回引文位置Google Scholar
百度学术
万方数据
[17]
Jin K Yuan L Wu H et al. Exploring large language model for next generation of artificial intelligence i n ophthalmology [J/OL]. Front Med (Lausanne) 2023101291404[2024-05-16]. https://pubmed.ncbi.nlm.nih.gov/38076260/. DOI: 10.3389/fmed.2023.1291404 .
返回引文位置Google Scholar
百度学术
万方数据
[18]
杨卫华邵毅许言午眼科人工智能临床研究评价指南(2023)[J]. 国际眼科杂志 202323(7)∶1064-1071. DOI: 10.3980/j.issn.1672-5123.2023.7.03 .
返回引文位置Google Scholar
百度学术
万方数据
Yang WH Shao Y Xu YW et al. Guidelines on clinical research evaluation of artificial intelligence in ophthalmology(2023)[J]. Int Eye Sci 202323(7)∶1064-1071. DOI: 10.3980/j.issn.1672-5123.2023.7.03 .
Goto CitationGoogle Scholar
Baidu Scholar
Wanfang Data
[19]
王婷汪瑞昕林浩添基于人工智能的白内障新诊疗模式[J]. 中华实验眼科杂志 202139(9)∶832-836. DOI: 10.3760/cma.j.cn115989-20200608-00405 .
返回引文位置Google Scholar
百度学术
万方数据
Wang T Wang RX Lin HT . New diagnosis and treatment pattern for cataract based on artificial intelligence[J]. Chin J Exp Ophthalmol 202139(9)∶832-836. DOI: 10.3760/cma.j.cn115989-20200608-00405 .
Goto CitationGoogle Scholar
Baidu Scholar
Wanfang Data
[20]
" 一带一路 " 眼科人工智能研发国际联盟 " 2019医学人工智能数据库管理规范方案修订研讨会 " 专家组 国家重点研发计划 " 常见致盲、致畸、致死疾病的人工智能筛查诊断系统研发和临床试验 " 项目组 人工智能眼病图像数据库入库数据质量评价指南[J]. 中华实验眼科杂志 201937(12)∶937-941. DOI: 10.3760/cma.j.issn.2095-0160.2019.12.001 .
返回引文位置Google Scholar
百度学术
万方数据
" One Belt and One Road " International Ophthalmic Artificial Intelligence Research and Development Alliance " Medical Artificial Intelligence Database Management Specification Program Revision Seminar in 2019 " Expert Panel National Key Research and Development Program of China " Development and Clinical Trials of Artificial Intelligence Diagnostic System for Blindness-causing Diseases,Teratogenic Abnormalities,Life Threatening Disorders " Project Team,the Science and Technology Planning Projects of Guangdong Province " Construction of Medical Data Labeling Technique System,General Dataset,and Intelligent Application Platform " Project Team et al. Guidelines for image quality assessment of artificial intelligence ophthalmic disease imagingdatabase[J]. Chin J Exp Ophthalmol 2019,(12)∶937-941. DOI: 10.3760/cma.j.issn.2095-0160.2019.12.001 .
Goto CitationGoogle Scholar
Baidu Scholar
Wanfang Data
[21]
Adams LC Truhn D Busch F et al. Leveraging GPT-4 for post hoc transformation of free-text radiology reports into structured reporting:a multilingual feasibility study[J/OL]. Radiology 2023307(4)∶e230725[2024-05-16]. https://pubmed.ncbi.nlm.nih.gov/37014240/. DOI: 10.1148/radiol.230725 .
返回引文位置Google Scholar
百度学术
万方数据
[22]
Cheng K Guo Q He Y et al. Artificial intelligence in sports medicine:could GPT-4 make human doctors obsolete?[J]. Ann Biomed Eng 202351(8)∶1658-1662. DOI: 10.1007/s10439-023-03213-1 .
返回引文位置Google Scholar
百度学术
万方数据
[23]
Potapenko I Boberg-Ans LC Stormly Hansen M et al. Artificial intelligence-based chatbot patient information on common retinal diseases using ChatGPT[J]. Acta Ophthalmol 2023101(7)∶829-831. DOI: 10.1111/aos.15661 .
返回引文位置Google Scholar
百度学术
万方数据
[24]
Masalkhi M Ong J Waisberg E et al. ChatGPT to document ocular infectious diseases[J]. Eye (Lond) 202438(5)∶826-828. DOI: 10.1038/s41433-023-02823-2 .
返回引文位置Google Scholar
百度学术
万方数据
[25]
Pushpanathan K Lim ZW Er Yew SM et al. Popular large language model chatbots ' accuracy,comprehensiveness,and self-awareness in answering ocular symptom queries [J/OL]. iScience 202326(11)∶ 108 163 [2024-05-16]. https://pubmed.ncbi.nlm.nih.gov/37915603/. DOI: 10.1016/j.isci.2023.108163 .
返回引文位置Google Scholar
百度学术
万方数据
[26]
Cheng K Sun Z He Y et al. The potential impact of ChatGPT/GPT-4 on surgery:will it topple the profession of surgeons?[J]. Int J Surg 2023109(5)∶1545-1547. DOI: 10.1097/JS9.0000000000000388 .
返回引文位置Google Scholar
百度学术
万方数据
[27]
陈有信冯时赵清眼科人工智能研究的相关问题[J]. 中华眼底病杂志 202238(2)∶89-92. DOI: 10.3760/cma.j.cn511434-20220210-00077 .
返回引文位置Google Scholar
百度学术
万方数据
Chen YX Feng S Zhao Q Issues related to artificial intelligence research in ophthalmology[J]. Chin J Ocul Fundus Diseases 202238(2)∶89-92. DOI: 10.3760/cma.j.cn511434-20220210-00077 .
Goto CitationGoogle Scholar
Baidu Scholar
Wanfang Data
[28]
佚名《中国医疗人工智能发展报告(2020)》蓝皮书正式出版[J]. 中国数字医学 202116(6)∶1.
返回引文位置Google Scholar
百度学术
万方数据
[29]
Abramski K Citraro S Lombardi GSM . Cognitive network science reveals bias in GPT-3,GPT-3.5 turbo,and GPT-4 mirroring math anxiety in high-school students[J/OL]. Big Data Cogn Comput 20237(3)∶124[2024-05-17]. https://www.mdpi.com/2504-2289/7/3/124#metrics. DOI: 10.3390/bdcc7030124 .
返回引文位置Google Scholar
百度学术
万方数据
[30]
GBD 2019 Blindness and Vision Impairment CollaboratorsVision Loss Expert Group of the Global Burden of Disease Study. Trends in prevalence of blindness and distance and near vision impairment over 30 years:an analysis for the Global Burden of Disease Study[J/OL]. Lancet Glob Health 20219(2)∶e130-e143[2024-05-17]. https://pubmed.ncbi.nlm.nih.gov/33275950/. DOI: 10.1016/S2214-109X(20)30425-3 .
返回引文位置Google Scholar
百度学术
万方数据
备注信息
A
谭钢,Email: mocdef.labiamtoh99gnagnaT
B
邵毅,Email: mocdef.3ab6199eebeerf
C
所有作者均声明不存在利益冲突
D
国家自然科学基金 (82160195、82460203)
江西省双千计划科技创新高端领军人才项目 (jxsq2023201036)
江西省重大(重点)研发专项计划 (20223BBH80014)
评论 (0条)
注册
登录
时间排序
暂无评论,发表第一条评论抢沙发
MedAI助手(体验版)
文档即答
智问智答
机器翻译
回答内容由人工智能生成,我社无法保证其准确性和完整性,该生成内容不代表我们的态度或观点,仅供参考。
生成快照
文献快照

你好,我可以帮助您更好的了解本文,请向我提问您关注的问题。

0/2000

《中华医学会杂志社用户协议》 | 《隐私政策》

《SparkDesk 用户协议》 | 《SparkDesk 隐私政策》

网信算备340104764864601230055号 | 网信算备340104726288401230013号

技术支持:

历史对话
本文全部
还没有聊天记录
设置
模式
纯净模式沉浸模式
字号