Abstract
Abstract: :
Purpose:To develop a computer model of type 1, 2, and combined growth pattern choroidal neovascularization (CNV) in order to better understand the pathobiology of CNV and predict treatment effects. Methods:A graphical stochastic Monte–Carlo simulation of diffusion limited aggregation (DLA) that forms a fractal pattern on a spherical surface was used to model the growth of vessels in CNV. Fractals are structures that possess the same basic properties on all length scales. The model was programmed in C++ on a multiprocessor computer with extensive use of the OPENGL graphical library. Simulations of vessel growth were based on murine models of CNV growth and angiographically demonstrated human CNV growth. Results:Type 1 (sub–RPE), type 2 (subretinal), and combined CNV growth simulations showed that no two patterns were identical. Simulations included type 1 pattern: multiple ingrowth sites; type 2 pattern: expansion and regression of net; combined pattern: expansion and regression of net overlying multiple ingrowth sites. Based on the results of previous studies with antiangiogenic treatments such as adenovirus–encoded PEDF and photodynamic therapy (PDT) applied to established CNV, prediction of vascular regression was demonstrated in the model. Conclusions:It is possible to use fractal modeling to simulate types 1, 2, and combined CNV growth patterns and predict outcomes after intervention. The simulation used in this study effectively models the dynamic nature of angiogenesis as mediated by numerous cytokines (such as VEGF, IL8, PEDF, and others). In general, the temporal pattern of CNV follows initiation, inflammatory active, and involution stages. Because no two growth patterns of CNV are exactly the same, the computer simulation of CNV is of a stochastic (random) nature.
Keywords: choroid: neovascularization • computational modeling