Abstract
Purpose :
The prevalence of peripheral retinal findings is unclear. We assess the prevalence of peripheral retinal abnormalities in retinal patients in an academic setting to understand how they associate with retinal diagnoses.
Methods :
We performed a retrospective, observational clinical study. Ultra-widefield images (Optos P200DTx) were collated from the right eyes of consecutive retina clinic patients with demographic data attending Shiley Eye Institute, UCSD between 1st January and 30th November 2021. Images were included with unobstructed views of >60% of each quadrant and visibility of fourth order arterioles. The images were examined by a masked retinal specialist who recorded abnormal peripheral findings by quadrant (Fig, 1). Statistical analysis was performed using R software with fitted logistic regression models testing each peripheral change. Pearson’s Chi-squared test was used for samples lacking logarithmic power.
Results :
A total of 556 images were collated. After initial quality control, 312 patient images (mean age 61.9±19.5 years, 55.1% female) were analyzed. 152 (48.7%) patients had peripheral pathology, most commonly in the temporal quadrant. The most common peripheral findings were drusen (n=73, 23.4%) and laser marks (n=28, 9.0%). Diabetic retinopathy (DR) was associated with peripheral blot hemorrhages (P<.001), dot hemorrhages (P<.001), laser marks (P=.002), pigmentation (P=.049) and sclerosis of vessels (P<.001). Age-related macular degeneration (AMD) was associated with peripheral drusen (P=.001). Age was associated with general degeneration including chorioretinal atrophy (P<.001)(Fig. 2).
Conclusions :
Peripheral retinal findings were common in our retinal patients, drusen being most prevalent. The findings were validated by associations of DR with common DR peripheral findings. AMD was correlated with drusen, validated by recent literature. New correlations between DR and pigmentation and vessel sclerosis need to be further investigated. This data is useful to retinal specialists using ultra-widefield imaging and can provide a platform for training AI algorithms in detecting peripheral pathology in ultra-widefield imaging.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.