首页|期刊导航|实用临床医药杂志|人工智能技术在眼科疾病辅助诊断中的应用研究进展

人工智能技术在眼科疾病辅助诊断中的应用研究进展OA

Research progress on the application of artificial intelligence technology in auxiliary diagnosis of ophthalmic diseases

中文摘要英文摘要

人工智能技术(AI)正在深刻变革眼科疾病的诊断模式.深度学习算法在眼底图像分析领域展现出的性能,推动糖尿病视网膜病变、年龄相关性黄斑变性、青光眼等主要致盲眼病的筛查从人工判读向自动化、智能化方向转型.多模态人工智能通过融合眼底彩照、光学相干断层扫描、视野等多源数据,实现了比单模态系统更全面精准的疾病表征,其受试者工作特征曲线的曲线下面积较单模态提升4%~5%.视觉基础模型和视觉大语言模型等新技术范式的兴起,进一步拓展了眼科AI的泛化能力和交互方式.本文系统梳理了近5年AI在眼科疾病辅助诊断中的主要应用进展,深入分析当前面临的核心挑战,并展望未来发展方向.

Artificial intelligence(AI)technology is profoundly transforming the diagnostic para-digm of ophthalmic diseases.The outstanding performance of deep learning algorithms in the field of fundus image analysis has promoted the screening of major blinding eye diseases including diabetic retinopathy,age-related macular degeneration and glaucoma to transform from manual interpretation to automated and intelligent modes.By integrating multi-source data such as color fundus photography,optical coherence tomography and visual field examination,multimodal artificial intelligence achieves more comprehensive and accurate disease characterization than single-modal systems,whose area un-der the receiver operating characteristic curve is improved by 4%to 5%compared with that of single-modal models.The emergence of new technical paradigms such as vision foundation models and vision large language models further expands the generalization ability and interaction modes of ophthalmic AI.This paper systematically sorts out the main application progress of AI in the auxiliary diagnosis of ophthalmic diseases in the past 5 years,deeply analyzes the core challenges currently faced,and clar-ifies the future development directions.

江霞;薛朝华;金新政

华中科技大学同济医学院附属协和医院眼科,湖北武汉,430022华中科技大学同济医学院附属协和医院眼科,湖北武汉,430022华中科技大学健康信息资源研究中心,湖北武汉,430030

医药卫生

人工智能眼科学计算机辅助诊断深度学习多模态成像糖尿病视网膜病变青光眼视觉大语言模型

artificial intelligenceophthalmologycomputer-aided diagnosisdeep learningmultimodal imagingdiabetic retinopathyglaucomavision large language model

《实用临床医药杂志》 2026 (15)

144-148,5

湖北省卫生健康委科研计划项目(2023XHYN073)

10.7619/jcmp.20262328

评论