Abstract
Purpose :
Microvascular changes in diabetic retinopathy compromise the blood supply to the retina and eventually cause retinal ischemia, resulting in blindness. Early detection of retinal hypoxia can allow for early diagnosis and treatment, and the possibility of preventing irreversible damage and blindness. Our purpose is to determine if a hyperspectral retinal oximetry system can detect hypoxic changes in the retina of patients with diabetes. Our hypothesis is that hyperspectral-driven retinal oximetry will provide a more sensitive, earlier biomarker of diabetic complications in the eye, which would predate changes detected by current, less sensitive oximetry techniques as well as clinically visible microvascular changes.
Methods :
6 diabetic patients with various levels of ocular manifestations (average age 49.3), and 6 healthy controls (average age 32.1) have been imaged thus far. The ultimate goal is 10 patients in each group. Retinal oximetry measurements are taken of major vessels near the optic disk with a single fundus image. Oximetry measurements are made with a hyperspectral fundus camera with 16 bands of spectral resolution (470 – 630 nm), by least-squares fitting of oxy and deoxy hemoglobin spectra using the collected spectra.
Results :
In our comparisons thus far, diabetic patients have a statistically insignificant trend toward a lower artery-vein differential (0.230 in the diabetic group vs. 0.301 in the control group; p = 0.098). Oxygenation saturation maps show the distribution of oxygenated blood in the arteries and veins near the optic disc.
Conclusions :
Although the current data suggests a statistically insignificant trend, a larger sample size (such as our goal of 10 control and 10 diabetics) may reveal significance. If the trend proves significant, a hyperspectral-based oximetry camera should be investigated further for its ability to detect diabetic changes in the eye.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.