The shape of things to come: three ways OCT is evolving

Artificial intelligence

Optical coherence tomography (OCT) devices produce a vast expanse of information. Moorfields Eye Hospital alone takes around 1000 OCT scans each day across its UK sites, while a growing number of High Street practices are purchasing OCT devices. It is this data-rich nature of OCT that makes it a frontrunner for the use of machine learning. A partnership between DeepMind and Moorfields Eye Hospital has developed a method of diagnosing common eye conditions from OCT scans with an accuracy on par to world-leading clinicians, while IBM has used artificial intelligence (AI) to detect and classify the severity of diabetic eye disease and diagnose glaucoma. Where will AI head next? It is anyone’s guess.  

Binocular OCT

OCT angiography

If you think of a traditional OCT scan as a polaroid snapshot, then OCT angiography (OCT-A) is the sharp image taken with a digital camera. OCT-A is capable of displaying blood vessels at a finer level. And, unlike older techniques, it can do this without the need to inject a dye into the patient’s arm beforehand. This makes it useful for investigating vascular changes as a result of disease, and of course, is kinder to the person on the receiving end than the traditional method. Early research suggests that it may be helpful in detecting not only eye disease, but also for pinpointing warning signs of other conditions such as Alzheimer’s disease.

Image credit: Shutterstock