Quantification, spatial profiling, and mathematical modeling of pathology distributions in the human retina

Many retinal pathologies, including age-related macular degeneration (AMD), display characteristic spatial patterns. AMD predominately a ects the macula, the central conedominated region of the human retina responsible for high-acuity daytime vision. An understanding of why the macula is speci cally susceptible to age-related changes would likely prove invaluable to understanding the pathology of AMD and the development of preventative therapies. Unfortunately, such an understanding has thus far proven elusive. A large number of physiological and anatomical parameters vary signi cantly between the central and peripheral human retina, and many of these parameters are altered during retinal degenerative disorders. This produces a large number of spatial associations between various parameters, obscuring causal relationships and making identi cation of timing and initiating factors very challenging. To address this challenge we developed RetSpace, an analytic software package designed to generate and analyze quantitative maps of anatomic and pathologic parameters within individual human retinas. The RetSpace system was speci cally designed to analyze image sets generated with computational molecular phenotyping (CMP), a technique pioneered by our laboratory to characterize the immense cellular diversity of the neural and sensory retina. By combining the sensitivity and diversity of CMP with the analytic power of RetSpace, we have produced a novel mechanism to study the spatial distributions and regional variability of various measures of retinal anatomy and pathology, as well as the extent to which di erent pathologies show regional correlations potentially indicative of shared pathological origins. To demonstrate the utility of our approach we have analyzed the severity and spatial distributions of an extensive set of anatomic and pathologic parameters in a series of aging human donor retinas. In doing so we have identi ed novel metabolic changes in the RPE and photoreceptors that are spatially and quantitatively correlated with known pathological characteristics of AMD and may serve as sensitive markers of early stress in AMD and other retinal diseases. Mathematical models of the heterocellular diversity of these metabolic changes provide further insight into the mechanisms behind these changes and hint at the origins and spatial specificity of the disease.