Longitudinal analysis of retinal microvascular changes and its association with diabetic retinopathy

Abstract

Purpose : Diabetic retinopathy (DR) is the leading cause of blindness in the working-age population. Identifying early retinal microvascular changes relevant to DR has the potential to detect DR prior to physical manifestations in the eye. This study aims to determine a set of microvascular biomarkers that are responsible for the development and progression of DR.

Methods : A retrospective study cohort consisting of baseline and follow-up images of 30 healthy and 30 diabetic participants randomly chosen from the LANDMark study. Diagnosis of DR was based on Early Treatment of Diabetic Retinopathy Study scale. A set of 84 geometric microvascular parameters (GMP) were generated from macular centred images using a semi-automated, multi-step web-based Vascular AnalySis Program (VASP). VASP automatically identified vessel boundaries, categorized vessels into arteries or veins, and divided the fundus into multiple zones. An experienced grader confirmed and adjusted the outputs produced by VASP (when necessary), and GMPs were then generated. We analysed the four evolution of all 84 GMPs returned by VASP and compared each parameter between the two groups. Demographic comparisons were performed using a Chi x² test and parameter analysis was performed using a Generalised Linear Model to correct for confounders.

Results : The mean age of diabetic cohort was 57 ± 11 years, with their average BMI significantly greater than the healthy cohort (28 ± 6 kg/m², p=0.03). Supine systolic and diastolic blood pressure were significantly higher (p< 0.001 & 0.03) in diabetes. From the 84 generated GMPs, five were significantly different between the two groups. In relation to the diabetic cohort, these topographic differences can be summarised as: increased tortuosity for all vessels in Zone C (p=0.002), increased tortuosity thickness for arteries and veins in Zone B,C (p=0.02 & 0.03), increased asymmetric width ratio for vessel branches in Zone C (p=0.01), and the combined width of 2 major vein trees in Zone C increased (p=0.04).

Conclusions : Our study found a significant increase in five novel GMPs exclusive to diabetes. Following further research, these objective measurements could be used as biomarkers to identify patients at risk of developing referable DR and enable timely treatment, thus reducing the risk of blindness.

This is a 2021 ARVO Annual Meeting abstract.

 

Steps of parameter generation

 

GMPs compared across healthy and diabetic groups.