Abstract
Purpose :
Static objective retinal blood vessel analysis is widely used in research yet hasn’t transitioned into clinical practice. Reduced retinal artery and vein calibres have been associated with hypertension, diabetes and coronary heart disease-related deaths. Findings like this suggest retinal vessel calibre measurements have great potential as a clinical screening and monitoring tool. Despite strong evidence of repeatability in a research setting, there is little evidence to demonstrate that this approach is equally robust in a clinic (where conditions could be considered sub-optimal; e.g. undilated patients).
Methods :
20 subjects had 10 retinal images captured in quick succession (both undilated and dilated with Tropicamide 0.5%) on Day One with a Topcon 3D Maestro and were then re-imaged seven days later. Central Retinal Artery and Vein Equivalents (CRAE and CRVE) and Arterio-Venous Ratios (AVR) were calculated for each image using VesselMap v3.0 (Imedos Systems, Jena). Intra-image measurement variance was also calculated through multiple vessel analysis for a single image of each subject.
Results :
Mean age was 40 years (±14), with 11 females and 9 males. Overall average MSE was -1.20D (±2.22D). No subjects had existing cardiovascular disease. Average undilated baseline CRAE was 178.8µm (±15.8); CRVE was 211.0µm (±15.3) and AVR was 0.85 (±0.06). Average mydriatic baseline CRAE was 176.7µm (±15.8); CRVE was 210.0µm (±17.7) and AVR was 0.84 (±0.07). Both dilated and undilated measurements for CRAE and CRVE were compared in a Bland-Altman plot, with a bias of +2.07µm and +1.01µm respectively. Repeated measures ANOVA did not reveal a statistical significance between measurements taken on Day One or Day Seven, either dilated or undilated (p=>0.05).
Conclusions :
It has been demonstrated that there is no significant difference in measurements taken from images acquired with or without the use of mydriatics. Furthermore successive image capture showed a negligible effect induced by the cardiac pulse cycle. Finally, short-term time intervals (one week) did not impact significantly on retinal vessel measurements. In summary, it has been shown that vessel analyses conducted on images (both dilated and undilated) acquired in clinic with an OCT-fundus camera show similar levels of variance and repeatability to those previously reported.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.