Modeling Macular Degeneration Progression Using Quantitative Phenotypes

Vanderbilt University, Nashville TN

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): Age-related macular degeneration (AMD) is one of the most common causes of visual impairment in the United States (US). Although a multitude of studies have shown that both genetic and environmental factors contribute to the pathogenesis of AMD, little is known on how genetic variation impacts the disease's rate of progression. One of the difficulties that current studies face in examining AMD progression is the splitting of disease severity into just a few categories that view progression as the change from one stage to the next. To solve this problem, we propose to examine rate of progression as a continuous measure through the quantification of drusen load and geographic atrophy (GA), two hallmarks of AMD, over a predetermined time period. We hypothesize that by creating a quantitative measure of AMD progression, we will increase our power to develop and test models that predict progression. In aim 1, fundus photographs taken from AMD cases will be used to develop a quantitative measure of drusen load and GA to assign progression rates and serve as our training data set. In aim 2, our training data will have 100 SNPs genotyped at known AMD risk loci and used to generate a genetic model of progression. Regression models will be built that address both the impact of genetic variation on drusen load and GA alone, as it is not biased toward self-reporting, and models that examine progression using known environmental and genetic risk factors. The models generated from these data will be validated in an independent data set. In aim 3, we will test the feasibility of the BioVU system in linking al necessary clinical information required to assign a progression rate. BioVU is Vanderbilt University's bio-repository system that was developed with the goal of linking DNA samples, extracted from blood, to de-identified electronic medical records to allow assessment of how genetic variation impacts clinical phenotypes. BioVU currently contains 765 Caucasian samples over the age of 65 with de-identified electronic medical records linked to their DNA, and that meet the criteria for AMD case status based on AMD specific ICD-9 codes. A subset of these samples will be used in a proof-of-concept study on the efficacy of the BioVU system in extracting a sufficient amount of clinical data to assign an ARED score and a progression rate. Samples of interest will have their electronic medical records pulled from BioVU and linked to fundus photographs when available. Samples will then be genotyped and used to build an electronic medical record derived data set to further test our models. The models generated here will further our understanding of the role genetic variants play in the major AMD phenotypes, drusen load and GA, and provide one of the many tools necessary to guide the personalized care for individuals affected with this disease.