Purpose
To quantify retinal capillary density and morphology in uveitis using SD-OCTA.
Design
Cross-sectional, observational study
Methods
Healthy and uveitic subjects were recruited from two tertiary care eye centers. Prototype SD-OCTA devices (Cirrus, Carl Zeiss Meditec, Inc., Dublin, CA) were used to generate 3×3 mm2 OCTA images centered on the fovea. Subjects were placed into 3 groups based on the type of optical microangiography (OMAG) algorithm used for image processing (intensity-and/or phase) and type of retinal segmentation (automatic or manual). A semi-automated method was used to calculate skeleton density (SD), vessel density (VD), fractal dimension (FD), and vessel diameter index (VDI). Retinal vasculature was assessed in the superficial retinal layer (SRL), deep retinal layer (DRL), and non-segmented retinal layer (NS-RL). A generalized estimating equations model was used to analyze associations between the OCTA measures and disease status within each retinal layer. A P value < 0.05 was accepted as significant. Reproducibility and repeatability were assessed using the Intraclass Correlation Coefficient (ICC).
Results
The SD, VD, and FD of the parafoveal capillaries were lower in uveitic eyes compared to healthy eyes in all retinal segments. In addition, SD and VD were significantly lower in the DRL of subjects with uveitic macular edema. There was no correlation in any capillary parameters and anatomic classification of uveitis.
Conclusions
Quantitative analysis of parafoveal capillary density and morphology in uveitis demonstrates significantly lower capillary density and complexity. SD-OCTA algorithms are robust enough to detect these changes and can provide a novel diagnostic index of disease for uveitis subjects.