Abstract
Purpose :
Diabetic retinopathy (DR) is characterized by changes to the retinal vasculature, with deposits of lipids and proteins within the retina, causing it to thicken. Simultaneously cells atrophy, leading to retinal thinning. To provide diagnostic information beyond total retinal thickness, we quantified retinal distortion for individual retinal layers.
Methods :
SDOCT images from 96 diabetic subjects were selected from 1750 diabetic patients from Aeon Imaging’s study of screening for DR in Alameda County California. A trained grader used color fundus photographs to determine the level of diabetic retinopathy for these subjects, split into 4 groups: males w/o signs of DR (n=18) (age 30-69 yr), females w/o signs of DR (n=35) (age 29-67 yr), males with DR (n=21) (age 37-65 yr), and females with DR (n=22) (age 35-67 yr). There was no statistical difference in age for either males or female groups (p=0.3414 and p=0.1779).
SDOCT images were automatically segmented using custom software in MATLAB, marking 9 retinal layer boundaries. This segmentation was verified, and manual correction was provided when necessary. Total thickness and retinal layer thickness were computed at 8 locations: 7, 5, 3, and 1 deg nasal to and temporal to the fovea. Total thickness was also computed at the fovea. Student’s T-Tests were used for statistical analysis.
Results :
In general, statistical significance was reached for males in the temporal retina for total thickness, outer nuclear layer (ONL) thickness, and inner segment (IS) thickness, with males with DR being thicker than males w/o DR. This held true even after removing those male subjects whose central retinal thickness was statistically significantly thicker that their control counterparts; for example, the ONL at temporal locations 7, 5, 3, and 1 deg from the fovea were thicker (p=0.0169, 0.0052, 0.0008, 0.0045). For females, while there were statistical differences between the 2 groups, the trends were neither as large nor as consistent.
Out of the males with DR, only 3 out of the 21 had retinal thickness that fell outside of the confidence interval for no DR controls. Yet, only 4 of the 18 male subjects with DR did not have at least 2 of 32 temporal retinal layer thickness measurements that differed from controls; 9 had between 3 and 6, and 4 had between 7 and 9.
Conclusions :
Individual retinal layer thicknesses differed for patients with DR, particularly for males in the temporal retina.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.