Department of Ophthalmology, Kyung Hee University Hospital, Kyung Hee University, 23, Kyungheedae-ro, Dongdaemun-gu, Seoul, Korea.
Department of Mechanical Engineering, Korea Advanced Institute of Science and Technology, 291 Daehak-ro, Yuseong-gu, Daejeon, 34141, Korea.
Graefes Arch Clin Exp Ophthalmol. 2021 Aug;259(8):2103-2111. doi: 10.1007/s00417-021-05099-y. Epub 2021 Feb 2.
To investigate the diagnostic utility of microvascular parameters for grading the severity of diabetic retinopathy (DR) with a range of views using wide-field swept-source optical coherence tomography angiography (SS-OCTA).
This retrospective study grouped 235 eyes with diabetes into the five grades: diabetes without retinopathy (no-DR), mild non-proliferative DR (NPDR), moderate NPDR, severe NPDR, and proliferative DR (PDR). Foveal avascular zone (FAZ) metrics, vessel density (VD), and the capillary nonperfusion area (NPA) were quantified with a customized, semiautomatic software algorithm. Regions of interest were selected from three rectangular fields of different sizes (i.e., 3 × 3 mm, 6 × 6 mm, and 10 × 10 mm), perpendicular to the fovea-optic disc axis.
NPA obtained from the 6 × 6mm and 10 × 10mm areas was the only discriminating parameter for the three NPDR stages. ROC curve analysis revealed that NPA from the 10 × 10mm field exhibited the best performance for grading DR into five stages. The NPA cutoff values were 3.7% (area under the curve (AUC): 0.91), 4.7% (AUC: 0.94), 9.3% (AUC: 0.94), and 21.4% (AUC: 0.90) for grading no-DR, mild from moderate NPDR, moderate from severe NPDR, and severe NPDR from PDR, respectively.
Increasing DR severity as assessed by conventional grading systems is accompanied with increasing retinal ischemia on SS-OCTA. NPA measured from the larger 10 × 10 mm scan area showed the highest sensitivity for determining five-grade DR severity. In the future, the addition of quantitative NPA may provide a more clinically feasible DR grading system.
利用宽视野扫频源光相干断层扫描血管造影(SS-OCTA)对多种视场下的微血管参数进行研究,以评估其在糖尿病视网膜病变(DR)严重程度分级中的诊断价值。
本回顾性研究将 235 只糖尿病患者的眼分为 5 个等级:无糖尿病视网膜病变(无 DR)、轻度非增殖性糖尿病视网膜病变(NPDR)、中度 NPDR、重度 NPDR 和增殖性糖尿病视网膜病变(PDR)。使用定制的半自动软件算法对黄斑中心凹无血管区(FAZ)指标、血管密度(VD)和毛细血管无灌注区(NPA)进行量化。在垂直于黄斑-视盘轴的三个不同大小的矩形区域(即 3×3mm、6×6mm 和 10×10mm)中选择感兴趣区域。
来自 6×6mm 和 10×10mm 区域的 NPA 是区分 3 个 NPDR 阶段的唯一判别参数。ROC 曲线分析显示,来自 10×10mm 区域的 NPA 对 DR 5 级分级的性能最佳。NPA 的截断值分别为 3.7%(曲线下面积(AUC):0.91)、4.7%(AUC:0.94)、9.3%(AUC:0.94)和 21.4%(AUC:0.90),用于分级无 DR、轻度至中度 NPDR、中度至重度 NPDR 和重度 NPDR 至 PDR。
根据传统分级系统评估的 DR 严重程度增加,SS-OCTA 上视网膜缺血程度增加。来自较大的 10×10mm 扫描区域的测量 NPA 对确定 5 级 DR 严重程度具有最高的灵敏度。未来,定量 NPA 的增加可能提供更具临床可行性的 DR 分级系统。