文献速递眼底病与葡萄膜炎

中视网膜层高反射异常区在糖尿病眼诊断鉴别中的价值

Hyperreflective Abnormal Areas in the Middle Retinal Layers on En Face Optical Coherence Tomography for Diagnostic Discrimination in Diabetic Eyes.
PubMed 文献 · Am J Ophthalmol · 2026-07-13

OBJECTIVE: To evaluate whether hyperreflective lesions on en face optical coherence tomography of the middle retinal layers, which we label as hyperreflective abnormal areas (HAAs), improve the distinction of referable and vision-threatening diabetic retinopathy (DR), beyond esta…

要点

  • HAA与需转诊DR独立相关,每增加0.1个百分点OR为3.42(95%CI 2.37-7.97)。

核心结论

  • 中视网膜层HAA可在既定临床因素之外识别需转诊DR眼。
  • HAA与威胁视力DR的关联方向一致,但未改善模型区分度。
  • HAA可能作为无中心性糖尿病黄斑水肿的高危糖尿病眼的候选生物标志物。

研究设计与样本

  • 研究为对修订转诊标准准确性的分析。
  • 在平均en face OCT中视网膜层板层上手动量化HAA,板层自内丛状层上10 µm至外丛状层下10 µm。
  • 采用广义估计方程逻辑回归构建基础模型与加入HAA的全模型。
  • 以AUC、内部交叉验证和校准评估模型性能。

主要发现

  • HAA与需转诊DR独立相关,每0.1个百分点增加OR为3.42(95%CI 2.37-7.97)。
  • 内部交叉验证支持需转诊DR区分度改善的稳健性。
  • HAA与威胁视力DR正相关,OR为1.23。

关键数值与推荐

  • HAA量化为可分析视网膜面积的百分比。
  • 基础模型包含年龄、性别、糖尿病病程、血红蛋白A1c和高血压。
  • 全模型在基础模型上额外纳入HAA。

临床怎么用

  • 在无中心性糖尿病黄斑水肿的糖尿病眼,可考虑用中视网膜层HAA辅助识别需转诊DR。
  • HAA对威胁视力DR的增量价值尚不明确,不宜单独用于该判断。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

OBJECTIVE: To evaluate whether hyperreflective lesions on en face optical coherence tomography of the middle retinal layers, which we label as hyperreflective abnormal areas (HAAs), improve the distinction of referable and vision-threatening diabetic retinopathy (DR), beyond established clinical factors. DESIGN: Accuracy analysis of revised treatment referral criteria. PARTICIPANTS: Two hundred seventy-five eyes from 185 patients with diabetes without center-involving diabetic macular edema, encompassing a broad range of DR severity. METHODS: HAAs were manually quantified on averaged en face optical coherence tomography of a middle-retina slab, segmented from 10 µm above the inner plexiform layer to 10 µm below the outer plexiform layer, and quantified as percentage HAA of the analyzable retinal area. Generalized estimating equation logistic regression models were constructed to distinguish referable and vision-threatening DR. A base model included clinical parameters of age, sex, duration of diabetes, hemoglobin A1c, and hypertension, while a full model additionally incorporated HAAs. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), internal cross-validation, and calibration. MAIN OUTCOME MEASURES: Model discrimination and effect size of HAAs for distinguishing referable and vision-threatening DR. RESULTS: For referable DR, the full model showed significantly improved discrimination compared with the base model (AUC 0.969 vs 0.865; ΔAUC 0.104, P = .002). HAAs were independently associated with referable DR, with an odds ratio of 3.42 per 0.1-percentage point increase (95% CI, 2.37-7.97). Internal cross-validation supported the robustness of the improved discrimination for referable DR. For vision-threatening DR, the full model showed higher discrimination (AUC 0.865 vs 0.778), but did not reach significance (P = .126). HAAs were positively associated with vision-threatening DR, with an odds ratio of 1.23. CONCLUSIONS: Middle retinal layer HAAs identified eyes with referable DR beyond established clinical factors. The association of HAAs with vision-threatening DR was directionally consistent but did not improve model discrimination beyond clinical factors. Because HAAs may capture lesion-agnostic structural signatures of microvascular pathology in the middle retina, they may serve as a promising biomarker for diagnostic discrimination of high-risk diabetic eyes in the absence of center-involving diabetic macular edema.

来源:Am J Ophthalmol,查看原文。版权归原作者所有。

引用

  1. Kakihara S, Shah J, Zhuang K, et al. Hyperreflective Abnormal Areas in the Middle Retinal Layers on En Face Optical Coherence Tomography for Diagnostic Discrimination in Diabetic Eyes. Am J Ophthalmol. 2026. PMID: 42442471. doi:10.1016/j.ajo.2026.07.014

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。