Teleretinal programs have expanded in recent years, primarily to screen for diabetic retinopathy. Accurate and consistent assessment of the retinal vasculature, which often sustains damage as a result of cardiovascular disease, would advance teleretinal screening in terms of robustness and cost. Unfortunately, assessing changes to the retinal vasculature and quantifying abnormalities in retinal images has proven more difficult. Semi-automatic methods are reported to reduce grader variability, but they are time-consuming, requiring extensive reader interaction, thus limiting the overall effectiveness of a teleretinal screening program. VisionQuest Biomedical and its collaborator, the Retina Institute of South Texas (RIST), will demonstrate a software tool for comprehensive assessment of retinal vasculature (CARV) to aid readers in the quantitative characterization of vascular abnormalities and identification of those images with features indicative of potential sight-threatening or life-threatening conditions. CARV will provide the reader real-time artery to vein ratio (AVR) measurements, an accepted clinical value for determining risk of future stroke or hypertensive events, as well as measurements associated with branching patterns and tortuosity. It will also incorporate detection algorithms for features such as copper and silver wiring, AV nicking, and emboli. Our goal is to provide an aid to the retinal grader for detecting and quantifying common retinal vessel abnormalities with high agreement with the gold standard, a retinal specialist. In this Phase I we will also implement clinically accepted quantitative measurements for vasculature parameters. CARV, which can be integrated into a teleretinal system for diabetics, will allow graders to consistently identify signs of hypertensive retinopathy and CVD-related conditions and providers to exploit the full potential of teleretinal screening.
Vascular changes in the retina are often the result of hypertension and other forms of cardiovascular disease. Manual grading of these changes can result in high rates of inter- and intra-grader variability. In this project we will demonstrate a software tool for comprehensive assessment of retinal vasculature (CARV), a fully automatic system to aid readers in the quantitative characterization of vascular abnormalities and identification of those images with features indicative of potential sight-threatening or life-threatening conditions.