人工智能在识别及分析糖尿病视网膜病变OCTA图像中视网膜无灌注区的研究进展OA

Diabetes Mellitus(DM)is a significant challenge in global public health,with its epidemiological charac-teristics showing a marked upward trend.Diabetic retinopathy(DR),a major blinding complication of diabetes,has non-perfusion area(NP)as a key indicator reflecting its severity.Optical coherence tomography angiography(OCTA),a non-invasive and high-resolution vascular imaging technique,has a significant advantage in displaying NP.In recent years,ar-tificial intelligence(AI)technology,especially deep learning,has become a powerful tool for analyzing OCTA images,en-abling automatic recognition and quantitative analysis of NP.This article reviews the technical methods,performance,clinical applications,and challenges of using AI to analyze NP in OCTA images of DR patients,with the aim of providing references for related research and clinical applications.