Retinal imaging and cardiovascular health
The eye offers a unique window through which blood vessels in the retina (the layer of light sensitive cells at the back of the eye) can be observed without complex or risky intervention. Physicians routinely examine the eyes of their patients using a hand-held ophthalmoscope to check for abnormalities that might indicate the presence of high blood pressure or diabetes. Research by Professor Simon Thom and his team at ICCH has shown that computer analysis of photographs of the retina can provide additional information on cardiovascular health, possibly in advance of other clinical indications. This has practical advantages as retinal photographs can be obtained easily in just a few minutes using a special camera, widely available in clinics throughout the world.
In collaboration with the University of Wisconsin and using an image analysis technique developed by Nick Witt, the team showed that abnormalities in the geometry of retinal vessels, such as the relationship between their diameters where the vessels divide, can predict the subsequent development of heart disease, independently of other risk factors. The geometry of the retinal vessels is regulated by the endothelium, a thin layer of cells lining the vessels. Dysfunction of this layer is believed to give rise to the altered geometry, and also to increase the risk of atherosclerotic lesions that may lead to heart disease and stroke.
The retina has a higher oxygen requirement than any other organ in the body, and this makes it exquisitely sensitive to diseases which impair its blood supply. This blood supply is delivered by a network of blood vessels - a tree of small arteries - that lies flat on the retinal surface. We know that the shape of this tree becomes deranged at early stages in the development of hypertension and diabetes. It is likely that even minor derangements in successive branches of the tree will have a big impact on blood supply when multiplied across the whole network. Dan Liu is building mathematical models that capture the geometric features of the eye in normal and diseased state. This approach will help us to determine the impact of hypertension and diabetes on the efficiency of retinal blood flow and to identify which abnormalities are most influential. In turn this should offer new clues to early clinical detection.