基于OCTA微血流参数定量分析的糖尿病视网膜病变早期风险分层模型构建及临床验证
CSTR:
作者:
作者单位:

蚌埠医科大学第一附属医院

作者简介:

通讯作者:

中图分类号:

基金项目:

2025年主持安徽省慢性疾病免疫学基础与临床重点实验室开放课题(MBPY202503);2025年主持数字医学与智慧健康安徽省重 点实验室开放课题(AHCM2025Z002);2024年承担蚌埠医科大学重点支持和培育学科领域专项(2024bypy015);2026年蚌埠市 社会科学一般项目(BB26B034);2025年安徽省卫生健康科技项目(AHWJ2025BAp20011)。


Construction and clinical validation of early risk stratification model for diabetic retinopathy based on quantitative analysis of OCTA microflow parameters
Author:
Affiliation:

The First Affiliated Hospital of Bengbu Medical University

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    目的 探讨光学相干断层扫描血管成像(optical coherence tomography angiography,OCTA)微血流参数在糖尿病视网膜病变(diabetic retinopathy,DR)早期无症状阶段的定量变化规律,构建多参数联合的DR早期风险分层模型并验证其临床效能。方法 回顾性纳入2023年1月至2025年12月于蚌埠医科大学第一附属医院就诊的2型糖尿病患者80例(80眼),根据早期DR危险因素分为低风险组(n=32)、中风险组(n=28)和高风险组(n=20)。所有受试者均行 OCTA扫描并采集OCTA参数,包括浅层毛细血管丛(superficial capillary plexus,SCP)与深层毛细血管丛(deep capillary plexus,DCP)的血管密度(vessel density,VD)、灌注密度(perfusion density,PD)、中心凹无血管区面积(foveal avascular zone area,FAZ-A)、周长(FAZ-P)及FD-300区毛细血管密度。同时采集空腹血糖(fasting blood glucose,FBG)、糖化血红蛋白(glycosylated hemoglobin,HbA1c)、糖尿病病程等临床指标。结果 随风险层级升高,SCP-VD、DCP-VD、FD-300(SCP)、6 mm×6 mm外层VD呈阶梯式下降,FAZ-A、FAZ-P呈阶梯式扩大(均P<0.05);而3 mm×3 mm PD在组间差异未达显著性(P=0.062)。经最小绝对收缩与选择算子(least absolute shrinkage and selection operator,LASSO)回归从16项候选参数中筛选出SCP-VD、DCP-VD、6 mm×6 mm外层VD、FAZ-P及糖尿病病程共5项独立变量纳入最终模型。训练集模型曲线下面积(area under the curve,AUC)为0.842(95%CI:0.761~0.923),验证集AUC为0.817(95%CI:0.728~0.906);校准曲线Bootstrap校正后Brier评分为0.138;决策曲线分析(decision curve analysis,DCA)显示,在阈值概率0.20~0.70范围内,模型净获益优于“全部干预”或“全部不干预”策略。结论 OCTA微血流参数可敏感反映DR早期微循环损害,基于OCTA参数联合糖尿病病程构建的Logistic风险分层模型对DR早期进展风险具有较好的预警价值,可为临床制订个体化随访间隔及开展早期干预提供依据。

    Abstract:

    Objective To investigate the quantitative changes of microflow parameters of optical coherence tomography angiography (OCTA) in the early asymptomatic stage of diabetic retinopathy (DR), and to construct a multi-parameter combination of DR early risk stratification model and verify its clinical efficacy. Methods A total of 80 patients (80 eyes) with type 2 diabetes who presented at the First Affiliated Hospital of Bengbu Medical University from January 2023 to December 2025, were retrospectively included. They were divided into low-risk group (n=32), medium-risk group (n=28) and high-risk group (n=20) according to the risk factors of early DR. All subjects underwent OCTA scanning, and OCTA parameters were collected, including the vessel density (VD) and perfusion density (PD) of the superficial capillary plexus (SCP) and deep capillary plexus (DCP), the area of the foveal avascular zone (FAZ-A), the perimeter of the foveal avascular zone (FAZ-P), and the capillary density in the FD-300 area. At the same time, clinical indicators such as fasting blood glucose (FBG), glycosylated hemoglobin (HbA1c) and diabetes course of the subjects were collected. Results As the risk level increased, SCP-VD, DCP-VD, FD-300 (SCP), and 6 mm×6 mm outer-layer VD showed a stepwise decrease, while FAZ-A and FAZ-P showed a stepwise increase (all P<0.05); however, there was no significant difference in 3 mm×3 mm PD between groups (P=0.062). Through the least absolute shrinkage and selection operator (LASSO) regression, five independent variables, namely SCP-VD, DCP-VD, 6 mm×6 mm outer-layer VD, FAZ-P, and the diabetes course, were selected from 14 candidate parameters and incorporated into the final model. The training set model AUC was 0.842 (95%CI: 0.761?0.923), the validation set AUC was 0.817 (95%CI: 0.728?0.906), the calibration curve Bootstrap corrected Brier score was 0.138, and the decision curve analysis (DCA) showed that the net benefit of the model was better than the“ all intervention” or“ all no intervention” strategy in the threshold probability range of 0.20?0.70. Conclusion The microflow parameters of OCTA can sensitively reflect the early microcirculation damage of DR. The Logistic risk stratification model based on microflow parameters of OCTA combined with the course of diabetes has good early warning value for the risk of early progression of DR, and can provide quantitative basis for the formulation of clinical individualized follow-up interval and early intervention.

    参考文献
    相似文献
    引证文献
引用本文

蒋胜群.基于OCTA微血流参数定量分析的糖尿病视网膜病变早期风险分层模型构建及临床验证[J].生物医学工程学进展,2026,(4):27-31

复制
分享
相关视频

文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2026-07-14
  • 最后修改日期:2026-07-29
  • 录用日期:2026-08-01
  • 在线发布日期: 2026-10-20
  • 出版日期:
文章二维码