Abstract
Purpose :
Oxygen in the lamina cribrosa (LC) is essential for maintaining functioning retinal ganglion cell axons and vision. Previous studies of LC oxygen (Chuangsuwanich et al. 2016) used generic models and did not account for factor interactions. Our goal was to determine the influence, independently and in interaction, of various factors potentially affecting LC oxygen concentration based on a more anatomically realistic model of the LC vessel network.
Methods :
We reconstructed a model of the monkey LC vessel network based on histological sections (Fig 1). An inflow hematocrit and an arterial inlet pressure were set at the periphery to simulate blood flow from the circle of Zinn-Haller, and a venous outlet pressure was set at the central retinal vein to simulate drainage. Using a Monte Carlo approach, 500 models were generated with varying (baseline±20%) vessel lumen diameter, arterial pressure, venous pressure, inflow hematocrit, and oxygen consumption rate. Models were simulated to estimate LC neural tissue oxygen concentration. ANOVA was used to identify factors and interactions with the largest influence on the minimum (10th percentile) oxygen concentration in the LC, a measure of susceptibility to hypoxia.
Results :
Three factors and their interactions accounted for the majority of the variance (84.6%) in LC oxygen concentration (Fig 2). These factors were the vessel lumen diameter, oxygen consumption rate, and arterial pressure. LC oxygen concertation increased with vessel lumen diameter and arterial pressure, and decreased with oxygen consumption rate. Interactions between these factors had more influence than that of venous pressure and inflow hematocrit.
Conclusions :
Our models predict that the vessel lumen diameter, oxygen consumption rate, arterial pressure and their interactions have the largest influence on the oxygen concentration in the LC. To understand the susceptibility to ischemia in retinal ganglion cell axon damage and vision loss, the influential factors and their covariances should be better characterized.
This is a 2020 ARVO Annual Meeting abstract.