目的 评估深浅层血管密度变化、视网膜高反射点等形态学指标与糖网黄斑水肿（DME）治疗反应的相关性。 方法 回顾性选取2020年4月至2023年11月就诊于眼科的DME患者（45例 59眼）。对比患者治疗前后深浅层血管密度变化、视网膜高反射点等指标变化。使用Pearson检验分析各指标与DME抗血管内皮生长因子（VEGF）治疗反应不佳的相关性。采用受试者工作特征曲线（ROC）分析基线视网膜深层毛细血管丛（DCP）、黄斑中心视网膜厚度（CMT）、视网膜高反射灶（HRF）数量及血流密度（FD300-VD）在诊断DME对抗VEGF治疗反应不佳的价值。 结果 与治疗前相比，治疗后患者的视网膜浅层毛细血管丛（SCP）、DCP、FD300-VD、深浅层血流比（DSFR）上升，CMT厚度明显降低（P < 0.05）。Pearson相关性检验显示，抗VEGF治疗反应不佳患者结局与基线SCP-VD、DCP-VD、DSFR、FD300-VD呈负相关（r = -0.458、-0.433、-0.604、-0.452，P < 0.05），与CMT呈正相关（r = 0.427，P < 0.05）。多因素logistic回归分析结果显示，SCP、DCP、FD300-VD、DSFR上升，CMT降低是影响DME抗VEGF治疗应答的相关因素（OR= 0.285、0.272、0.291、0.268、2.821，P < 0.05）。ROC曲线分析结果显示，基线DCP、CMT、HRF数量、FD300-VD的曲线下面积（AUC）分别为0.918、0.934、0.947、0.927，在诊断DME抗VEGF治疗反应不佳中具有良好的价值。 结论 深浅层血管密度变化、视网膜高反射点等形态学指标在评估DME治疗反应中具有良好的应用价值，可辅助临床治疗工作的开展。

Objective To evaluate the correlation between the changes of vascular density in deep and shallow layers， retinal hyperreflexia and the therapeutic response of diabetic macular edema （DME）. Methods Retrospectively selected DME patients （45 cases and 59 eyes） who visited the Department of Ophthalmology from April 2020 to November 2023 as the experimental group. Compare the change of deep and superficial blood vessel density and retinal hyperreflective points before and after treatment of the patients. The correlation between each index and poor response to anti-vascular endothelial growth factor （VEGF） treatment with DME was analyzed using the Pearson test. The value of baseline deep retinal capillary plexus （DCP）， central macular retinal thickness （CMT）， number of high-reflective foci （HRF） in the retina， and blood flow density （FD300-VD） in diagnosing poor response to anti-VEGF therapy in DME was analyzed using the subject operating characteristic curve （ROC）. Results Compared with pre-treatment， patients' superficial retinal capillary plexus （SCP）， DCP， FD300-VD， and deep-superficial flow ratio （DSFR） increased and CMT thickness significantly decreased after treatment （P < 0.05）. Pearson's correlation test showed that the outcome of patients with poor response to anti-VEGF therapy was negatively correlated with baseline SCP-VD， DCP-VD， DSFR， and FD300-VD （r = -0.458， -0.433， -0.604， and -0.452， P < 0.05） and positively correlated with CMT （r = 0.427， P < 0.05）. The results of multifactorial logistic regression analysis showed that a rise in SCP， DCP， FD300-VD， and DSFR， and a decrease in CMT were correlates affecting the response to anti-VEGF therapy in DME （OR = 0.285， 0.272， 0.291， 0.268， 2.821， P < 0.05）.The results of ROC curve analysis showed that the AUCs of baseline DCP， CMT， HRF quantity， and FD300-VD were 0.918， 0.934， 0.947， and 0.927， respectively， which were of good value in diagnosing poor response to anti-VEGF therapy in DME. Conclusion Morphologic indicators such as changes in the density of deep and superficial blood vessels and retinal hyperreflective spots have good application value in assessing the response to DME treatment and can assist in clinical treatment.