Early diabetic retinopathy diagnosis based on local retinal blood vesselsanalysis in optical coherence tomography angiography (OCTA) images

Abstract

urpose:This paper introduces a new computer-aided diagnosis (CAD) system for detecting early-stage diabetic retinopathy (DR) using optical coherence tomography angiography (OCTA) images.Methods:The proposed DR-CAD system is based on the analysis of new local features that describeboth the appearance and retinal structure in OCTA images. It starts with a new segmentationapproach that has the ability to extract the blood vessels from superficial and deep retinal OCTAmaps. The high capability of our segmentation approach stems from using a joint Markov–Gibbs ran-dom field stochastic model integrating a 3D spatial statistical model with a first-order appearancemodel of the blood vessels. Following the segmentation step, three new local features are estimatedfrom the segmented vessels and the foveal avascular zone (FAZ): (a) vessels density, (b) blood vesselcalibre, and (c) width of the FAZ. To distinguish mild DR patients from normal cases, the estimatedthree features are used to train and test a support vector machine (SVM) classifier with the radialbasis function (RBF) kernel.Results:On a cohort of 105 subjects, the presented DR-CAD system demonstrated an overall accu-racy (ACC) of 94.3%, a sensitivity of 97.9%, a specificity of 87.0%, the area under the curve (AUC)of 92.4%, and a Dice similarity coefficient (DSC) of 95.8%. This in turn demonstrates the promiseof the proposed CAD system as a supplemental tool for early detection of DR.Conclusion:We developed a new DR-CAD system that is capable of diagnosing DR in its earlystage.