The Relationship of Optical Coherence Tomography Angiography (OCTA) Metrics to Progression of Diabetic Retinopathy: A 2-Year Prospective Study

Abstract

Purpose : Prognostic value of optical coherence tomography angiography (OCTA) in assessment of diabetic retinopathy (DR) progression has yet to be examined. We performed a prospective, observational study to investigate the relationship between quantitative OCTA metrics and DR progression in patients with diabetes.

Methods : All diabetic subjects underwent OCTA with a swept-source OCT (DRI-OCT Triton, Topcon, Inc, Tokyo, Japan) at baseline. Automated measurements of foveal avascular zone (FAZ) area, FAZ circularity, vessel density (VD), vessel density index (VDI), fractal dimension (FD) of superficial capillary plexus (SCP) and deep capillary plexus (DCP) were obtained by using a customized MATLAB program. DR progression was defined as an increase in severity of two or more steps according to modified Airlie House classification. Cox proportional hazard model was used for analysis.

Results : One hundred sixty-one eyes of 97 subjects followed up for ≥ 2 years were included in this study. Over a median follow-up of 24.56 months, progression of DR occurred in 23 (14.3%) of 161 eyes. Of SCP, no significant associations were found between any of the OCTA metrics and DR progression. Of DCP, univariate analysis showed that FAZ area (HR, 1.57; 95% CI, 1.22-2.01; p<0.001), VD (HR, 1.86; 95% CI, 1.25-2.76; p=0.002) and FD (HR, 3.67; 95% CI, 1.21-11.13; p=0.02) of DCP were significantly associated with DR progression. These associations remained significant in multivariate analysis while adjusting for age, gender, duration of diabetes, glycated hemoglobin and baseline DR severity (FAZ area: p=0.003; VD: p=0.008; FD: p=0.035).

Conclusions : Larger FAZ area, lower VD and lower FD of DCP were associated with higher risk of DR progression independent of traditional risk factors. Our findings suggest that retinal microvascular abnormalities in DCP, rather than that of SCP, provide prognostic information on DR progression.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.

 

Customized program: automated quantification. (A) (F) original images; (B) (G) denoising; (C) (H) binarization; (D) (I) FAZ; (E) (J) vessel density

 

Kaplan-Meier curves. Exploratory variables were categorized into tertiles. (left) between eyes with different levels of FAZ area on DCP. (right) between eyes with different levels VD on DCP. These curves were provided for demonstration purpose only. FAZ area and vessel density were continuous data in final analysis.