Objective Diagnosis & Quantification of Retinal Disease

描述(申请人提供)：目前视网膜疾病的诊断和量化，如老年性黄斑变性，依赖于训练有素的专家对彩色视网膜图像和荧光素血管造影的主观解释。诊断过程的主观性使临床试验结果在临床实践中的适当应用变得复杂起来。随着视网膜疾病基因分型的快速发展和针对分子水平疾病的治疗的发展，对准确、可重复和精确的视网膜疾病表型的需求变得至关重要。高分辨率数字眼底成像系统的发展、先进的数学和高速计算现在使视网膜病变的完全量化成为可能。为了解决这一重要问题，已经组建了一支由视网膜图像解释专家、数学家、计算机科学家和图像处理专家组成的团队。视网膜图像分析和量化的过程分为三个主要部分：(1)图像选择和准备，(2)图像对齐或配准，(3)图像分割、识别和量化。在本提案的第一阶段(R21)，目标是开发一个原型系统，以回答与这三个组成部分中的每一个相关的重要可行性问题。要实现这一目标，将需要开发图像质量评估指标、阴影校正算法、全局翘曲函数和各种视网膜结构的强度分布数据库。在该提案的第二阶段(R33)，R21阶段的原型系统将根据现有的黄金标准(专家人类解释)和其他客观和主观临床参数进行精炼、优化和验证。在这个拟议的研究项目结束时，将开发出一种经过验证的技术，用于对视网膜疾病进行可重复性、准确性和客观性的诊断、分类和量化。这项技术将在未来的临床和临床前研究以及日常临床实践中具有重要价值。

DESCRIPTION (provided by applicant): The current diagnosis and quantification of retinal diseases, such as age-related macular degeneration, relies on the subjective interpretation of color retinal images and fluorescein angiograms by trained experts. The subjectivity of the diagnostic process has complicated the appropriate application of clinical trial results to clinical practice. With rapid advances in the genotyping of retinal disease and the development of therapies which target disease at the molecular level, the need for accurate, reproducible, and precise phenotyping of retinal disease has become critical. The development of high-resolution digital fundus imaging systems, advanced mathematics, and high-speed computing has now made complete quantification of retinal lesions feasible. To solve this important problem, a team of retinal image interpretation experts, mathematicians, computer scientists, and image processing experts has been assembled. The process of retinal image analysis and quantification is divided into three major components: (1) image selection and preparation, (2) image alignment or registration, and (3) image segmentation, identification, and quantification. In the first (R21) phase of this proposal, the aim is to develop a prototype system to answer important feasibility questions related to each of these three components. Accomplishing this aim will require the development of image quality assessment metrics, shade-correction algorithms, global warping functions, and a database of intensity profiles for various retinal structures. In the second (R33) phase of this proposal, the prototype system from the R21 phase will be refined, optimized, and validated against the existing gold standard (expert human interpretation) and other objective and subjective clinical parameters. At the end of this proposed research project, a validated technology will have been developed for the reproducible, accurate, and objective diagnosis, classification, and quantification of retinal disease. This technology will be of great value in future clinical and pre-clinical studies, and in everyday clinical practice.