Abstract
Purpose :
To characterize the local vessel orientation of retinal microvasculature using optical coherence tomography angiography (OCTA) and to investigate the diabetic retinopathy-associated changes in the vasculature organization.
Methods :
Macular OCTA scans were obtained from 3 subjects with healthy eyes and 3 subjects with diabetic retinopathy (DR) using a commercial OCTA module (SPECTRALIS, Heidelberg Engineering). OCTA images from the right eye of each subject were analyzed. The original angiogram was first enhanced using Hessian matrix-based vesselness filter (Frangi et al, Springer, 1998) prior to binary filter with globally determined threshold. Eigenvector of Hessian matrix was used to calculate the local orientation of each vessel in the region of interest (ROI), a circular disk centered at the macula (Fig.1). The preferred orientation of retinal microvasculature at each 45° sector of ROI, namely NS, SN, ST, TS, TI, IT, IN, and NI (N=nasal, S=superior, T=temporal, I=inferior), was quantified and compared with its sector axis.
Results :
The sectorial preferred orientation of retinal microvasculature varied within the same eye and appeared to be unaligned with its sector axis (Fig.2). In IT, the off angle is 63.1±9.1 and 73.9±29.1 for healthy and DR subjects, respectively; in ST, it is 57.9±2.9 and 63.08±29.0. More variations of vessel preferred orientation were observed in subjects with DR than healthy, particularly in the sectors of NS and NI (standard deviation=44.6 vs 1.8 and 35.3 vs 5.7). There appeared to be a consistent pattern of sectorial vasculature organization among healthy subjects.
Conclusions :
A novel approach for quantitative analysis of OCTA image was used and our preliminary results showed differential retinal microvasculature organizations between healthy and DR subjects in terms of the sectorial preferred orientation of vessels. Quantitative ophthalmic imaging could provide a useful tool for diagnosis of ocular disease and monitoring of treatment. Vessel orientation-related quantification may serve as a biomarker in ocular disease like DR.
This is a 2020 ARVO Annual Meeting abstract.