Abstract
Purpose :
To understand how OCT-angiography (OCT-A) measured vessel density values correlate with the traditional clinical classification system of diabetic retinopathy using the Early Treatment Diabetic Retinopathy Study (ETDRS) grading scale.
Methods :
A retrospective, cross sectional clinical study comparing OCT-A vessel density values to traditional ETDRS grading scale scores in patients with diabetic retinopathy was performed. Subjects with diabetic retinopathy were imaged with a spectral domain OCT system (Avanti RTVue-XR; Optovue, Fremont, CA) and Optos wide-field fundus photography. Vessel densities of the 3x3mm superficial and deep OCT-A layers were measured using the built-in system software (Angiovue Version 2018.0.0.18). Vessel densities were measured at two regions of interest for each OCT-A layer: 1) whole image and 2) foveal region of 1mm in diameter. A standardized version of the ETDRS diabetic grading scale, the Digital Algorithmic Diabetic Retinopathy Severity Scoring System, was used to grade the wide-field fundus photos. Pearson correlation testing was used to compare the vessel densities to the scores provided by the digital ETDRS grading scale.
Results :
Seventy-two eyes of 72 diabetic patients were identified from 2015 to 2018. Mean age was 56.5 years and 79% patients were males. 38% eyes had an ETDRS score of 60, or had extensive old pan-retinal photocoagulation, and 22% with a score >60, or with neovascularization. Pearson correlations between vessel density and the digital ETDRS grading scale were -0.408 (p-value=0.01) for the whole superficial OCT-A layer, -0.195 (p-value=0.100) for the foveal region superficial OCT-A layer, -0.386 (p-value=0.01) for the whole deep OCT-A layer, and -0.241 (p-value=0.05) for the foveal region deep layer.
Conclusions :
Vessel densities of the whole superficial, whole deep and foveal region deep OCT-A layers had a significant inverse correlation with the scores provided by the digital algorithmic diabetic retinopathy severity scoring system. OCT-A may be used in the future to accurately stratify diabetic retinopathy and guide treatment.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.