3D OCT Angiography for quantitative characterization of diabetic retinopathy

摘要 在糖尿病视网膜病变(DR)的临床诊断中，荧光素血管造影(FA)是目前唯一的方法 常规用于检测和治疗缺血，但它是一种侵入性方法，对 视觉症状出现前的早期微血管改变。以便及早发现和管理 DR中的缺血，光学相干断层血管成像(OCTA)是一种新的方法，已经获得了 FDA于2015年批准。与FA相比，OCTA允许对神经感觉性视网膜进行无创成像 以及分辨率约为10m的视网膜血管。最近的一些研究已经成功地使用OCTA来 提取与DR严重性密切相关的量化2D指标。卡沙尼博士的团队开创了 用于定量评估毛细血管形态和密度的几种2D OCTA指标。而当 非常有价值的是，这些2D OCTA度量是从3D OCTA数据的面部投影中得出的，并且 因此不可避免地模糊了原始3D血管网络中的几何和拓扑信息。 为了克服这一根本限制，史博士和卡沙尼博士将在这个R21项目中合作开发 用于自动分析OCTA数据的真正3D度量，并将其应用于大型DR的早期诊断 规模眼科研究。在南加州大学神经成像实验室(LONI)，施博士的团队进行了脑图研究 开发了各种先进的计算工具，用于基于普遍适用的3D形状分析 来自固有几何学的原理。在这个项目中，我们将翻译和改编这些工具以用于3D视网膜 使用OCTA数据进行血管系统建模和分析。本项目有三个具体目标：(1)应用 在脑成像中开发尖端计算算法以开发新的3DOCTA指标 基于体积和表面的视网膜毛细血管密度和形态的定量。(2)明确关系 使用我们之前发表的队列研究2D-和3D-OCTA指标与DR临床严重性的可靠性 健康的和糖尿病的受试者。(3)验证2D-和3D-OCTA指标与DR严重性的关系 在NEI资助的非裔美国人眼病研究(AFEDS)的一个特征良好的人群中 并确定与目前无法检测到的(亚临床)视网膜病变相关的OCTA指标。考虑到有钱人 史博士在LONI的团队开发的计算工具和已经收集的大规模OCTA数据 (n=396)从AFEDS研究和Kashani博士发表的研究来看，这个项目的风险很低，但 由此产生的3D OCTA度量和相关软件工具将对研究具有很高的价值，并有可能 临床社区。我们将使所有在这个项目中开发的软件工具和源代码免费 可通过NITRC(http://www.nitrc.org)和LONI网站(http://www.loni.usc.edu/Software).)获得

Abstract In the clinical diagnosis of diabetic retinopathy (DR), fluorescein angiography (FA) is currently the only method used routinely for the detection and treatment of ischemia, but it is an invasive method and not sensitive to early microvascular changes before the onset of visual symptoms. For the early detection and management of ischemia in DR, Optical Coherence Tomography Angiography (OCTA) is a novel method that has obtained FDA approval in 2015. Compared with FA, OCTA allows non-invasive imaging of both the neurosensory retina as well as the retinal vasculature at ~10m resolution. Some recent studies have successfully used OCTA to extract quantitative 2D metrics that are well correlated with DR severity. Dr. Kashani’s group has pioneered several of the 2D OCTA metrics for the quantitative assessment of capillary morphology and density. While highly valuable, these 2D OCTA metrics were derived from the en face projection of the 3D OCTA data, and therefore inevitably obscure the geometric and topological information in the original 3D vasculature networks. To overcome this fundamental limitation, Drs. Shi and Kashani will collaborate in this R21 project to develop truly 3D metrics for the automated analysis of OCTA data and apply them for the early diagnosis of DR in large scale eye studies. For brain mapping research, Dr. Shi’s group at Laboratory of Neuro Imaging (LONI) of USC has developed various advanced computational tools for 3D shape analysis based on generally applicable principles from intrinsic geometry. In this project, we will translate and adapt these tools for 3D retinal vasculature modeling and analysis using OCTA data. There are three specific aims in this project: (1) Apply cutting-edge computational algorithms developed in brain imaging to develop novel 3D OCTA metrics for volume- and surface-based quantitation of retinal capillary density and morphology. (2) Define the relationship and reliability of 2D- and 3D-OCTA metrics with clinical severity of DR using our previously published cohorts of healthy and diabetic subjects. (3) Validate the relationship of 2D- and 3D-OCTA metrics with DR severity among a well-characterized population from the NEI funded African American Eye Diseases Study (AFEDS) and identify OCTA metrics associated with currently undetectable (sub-clinical) retinopathy. Given the rich computational tools developed by Dr. Shi’s group at LONI and the already collected, large-scale OCTA data (n=396) from the AFEDS study and Dr. Kashani’s published studies, the risk of this project is low, but the resulting 3D OCTA metrics and associated software tools will be highly valuable to the research and potentially clinical community. We will make all the software tools and source codes developed in this project freely available through the NITRC (http://www.nitrc.org) and LONI website (http://www.loni.usc.edu/Software).