3D Image Analysis Approach to Determine Severity and Cause of Optic Nerve Edema

描述（申请人提供）：目前临床上对视神经肿胀的评估，仅靠专家的主观验光评价来诊断和鉴别视盘水肿的原因。我们研究工作的长期目标是开发自动化的3D图像分析方法，以确定一组最佳的3D参数，以量化视神经水肿的严重程度，并帮助区分潜在的原因。本应用程序的总体目标是制定策略，使用光谱域光学相干断层扫描（SD-OCT），快速准确地确定诊断为视神经乳头水肿的患者视神经肿胀的严重程度，并确定将视神经乳头水肿与其他引起视神经水肿的疾病区分开来的形态学特征。中心假设是，从3D图像分析技术中获得的关于体积和形状参数的信息将提高准确评估视盘水肿严重程度和原因的能力，而不是使用Fris¿n量表或当前的2D OCT参数进行现有的主观视神经肿胀验光评估。提出这项研究的基本原理是，拥有这样的3D参数将极大地改善视盘肿胀的评估方式。将实现下列具体目标：开发和评估从SD-OCT计算肿胀视神经头的新体积和形状参数的方法。这将通过在视盘肿胀患者的SD-OCT体积中改进和评估我们新的基于3D图的分割算法来完成。2. 确定SD-OCT参数与乳头水肿患者严重程度的临床测量最佳相关，并制定连续的严重程度量表。这将通过使用机器学习方法将SD-OCT参数与专家定义的Fris - n量表等级（一种基于基础的严重程度衡量标准）联系起来来实现。预计体积三维参数将比二维参数与临床测量更密切相关，并将提供连续的严重程度量表。3. 确定SD-OCT参数，区分视盘肿胀（或视盘明显肿胀，如假性视盘水肿）与视盘肿胀的其他原因，并开发相应的预测分类器。我们的工作假设是三维形状参数，特别是靠近布鲁赫膜开口的形状参数，将在自动分化过程中贡献最大。该方法是创新的，因为申请人开发的3D图像分析方法能够确定3D体积和形状参数，并且代表了使用定性图像信息和2D OCT图像信息评估视盘肿胀的现状的重大改进。提出的研究是重要的，因为它将有助于建立一个急需的替代和更客观的方法来评估视盘肿胀的严重程度和原因。

DESCRIPTION (provided by applicant): Currently, the clinical assessment of optic nerve swelling is limited by the subjective ophthalmoscopic evaluation by experts in order to diagnose and differentiate the cause of the optic disc edema. The long-term goal of our research effort is to develop automated 3D image-analysis approaches for the identification of an optimal set of 3D parameters to quantify the severity of optic nerve edema over time and to help differentiate the underlying cause. The overall objective in this application is to develop strategies, using spectral-domain optical coherence tomography (SD-OCT), to rapidly and accurately determine the severity of optic nerve swelling in patients diagnosed with papilledema and to ascertain morphological features that differentiate papilledema from other disorders causing optic nerve edema. The central hypothesis is that information about volumetric and shape parameters obtainable from 3D image analysis techniques will improve the ability to accurately assess the severity and cause of optic disc edema over the existing subjective ophthalmoscopic assessment of optic nerve swelling using the Fris¿n scale or current 2D OCT parameters. The rationale for the proposed research is that having such 3D parameters will dramatically improve the way optic disc swelling is assessed. The following specific aims will be pursued: 1. Develop and evaluate the methodology for computing novel volumetric and shape parameters of a swollen optic nerve head from SD-OCT. This will be completed by refining and evaluating our novel 3D graph-based segmentation algorithms in SD-OCT volumes of patients with optic disc swelling. 2. Identify SD-OCT parameters that optimally correlate with clinical measurements of severity in patients with papilledema and develop a continuous severity scale. This will be accomplished by using machine-learning approaches to relate SD-OCT parameters to expert-defined Fris¿n scale grades (a fundus-based measure of severity). It is anticipated that volumetric 3D parameters will more closely correlate with clinical measures than 2D parameters and will provide a continuous severity scale. 3. Identify SD-OCT parameters that differentiate papilledema from other causes of optic disc swelling (or apparent optic disc swelling, as in pseudopapilledema) and develop a corresponding predictive classifier. Our working hypothesis is that 3D shape parameters, especially those near Bruch's membrane opening, will contribute the most in the automatic differentiation process. The approach is innovative because the 3D image-analysis methodology developed by the applicants enables novel determination of 3D volumetric and shape parameters and represents a significant improvement over the status quo of using qualitative image information and 2D OCT image information for assessing optic disc swelling. The proposed research is significant because it will help to establish a much-needed alternative and more objective method by which to assess the severity and cause of optic disc swelling.