Authors: Sonja Karst, Morgan Heisler, Natalia Page, Timothy Yu, Julian Lo, Mahyar Etminan, Simon Warner, Marinko Sarunic, David Maberley, Eduardo Navajas

Author Disclosure Block: S. Karst: Speakers honorarium, travel expense; Allergan, Bayer. M. Heisler: None. N. Page: None. T. Yu: None. J. Lo: None. M. Etminan:None. S. Warner: None. M. Sarunic: Research Fund/Shares; Seymore Vision. D. Maberley: None. E. Navajas: None.

Abstract Body:

Purpose: To find new imaging biomarkers of disease onset in patients with diabetes but without diabetic retinopathy.
Study Design: Prospective cross-sectional analysis of diabetic patients without diabetic retinopathy and healthy controls.
Methods: Seven standardized 3x3 mm areas were recorded with Swept Source Optical Coherence Tomography Angiography (OCTA, Plex Elite 9000 ©, Zeiss Meditec): one was centered on the fovea, three were recorded temporally of the fovea and three were recorded nasally to the optic disc. Raw data was exported to MATLAB for segmentation of the capillary network. The intermediate and deep capillary plexus (CP) were summarized as deep vascular complex (DVC). Capillary density of the superficial CP and DVC and the mean capillary diameter were generated for each area imaged, after subtracting the area occupied by blood vessels greater in diameter than capillaries. The fractal dimension was also generated for each image. In the 3x3 mm area centered on the fovea, size and circularity of the FAZ was analyzed. Statistical analyses were performed with SPSS (version 25). The study was reviewed and approved by the ethics review board of the University of British Columbia.
Results: 74 eyes of 28 patients and 19 controls (mean age 60 ± 15yrs, 26 female) were included in this study. In the healthy controls the mean vessel density of the superficial CP was significantly less in the temporal areas (0.469) compared to the nasal areas (0.484) or the fovea (0.497; p<0.000). There was no difference in the DVC between the areas imaged (p=0.207). Patients with diabetes but without DR had significantly less capillary density in the DVC in the superior temporal area and the nasal area compared to healthy controls. The vessel density in the superficial and deep CP around the fovea and the FAZ parameters were similar in both groups.
Conclusions: OCTA allowed a new insight into the pathogenesis of microvascular changes secondary to diabetes and might allow for the identification of patients with diabetes before clinical signs become apparent in the fundus biomicroscopy or fundus photography. Early changes seem to affect the deep CP first and areas outside the fovea might be more sensitive to early changes.