New technology predicts heart disease with just an eye scan

The fundus; the rear interior wall of the eye, reflects the body’s overall health. [iStockphoto]

Being able to predict your risk of heart disease in less than a minute without any blood tests or blood pressure checks is now possible thanks to a ground-breaking development.

Using a new AI-powered algorithm, cardiovascular events can now be predicted through an eye scan, signifying a potential departure from traditional diagnostic methods like CT scans, MRIs and X-rays.

The journey to this discovery began when a joint team of researchers embarked on a mission to develop an automated tool for detecting diabetic retinopathy, a leading cause of blindness globally.

The team from Google and Verily worked to develop an algorithm that could recognise the signs of the disease and deliver a diagnosis within seconds when supplied with a patient’s retinal photos. The algorithm aimed to change the landscape of eye disease detection and management.

Through an eye scan, the fundus; the rear interior wall of the eye, reflects the body’s overall health. The retina, a tissue found in the fundus responsible for capturing images and sending them to the brain, offers a window that enables the effectiveness of this technology.

The algorithm is able to predict heart disease by analysing the veins and arteries in the retina. By studying the appearance of the fundus, doctors can also infer critical information like an individual’s blood pressure, age, and smoking habits. All of which are important predictors of cardiovascular health.

To train their cardiovascular prediction algorithm, the scientists used machine learning to analyse a medical dataset of nearly 300,000 patients. This data included eye scans as well as general medical data. Neural networks were then used to mine this information for patterns, learning to associate tell-tale signs in the eye scans with the metrics needed to predict cardiovascular risk such as age and blood pressure.

When tested, the algorithm could distinguish between the retinal images of two patients. One who would suffer a cardiovascular event in the following five years and one who would not 70 per cent of the time. This performance was only slightly lower than the commonly used SCORE method of predicting cardiovascular risk which requires blood tests and makes correct predictions 72 per cent of the time.

This breakthrough could enable ophthalmologists and other health workers to carry out cardiovascular screening using a camera without the need for a blood test or blood pressure check. It offers a highly effective test that doesn’t have to be done in a clinic. The study is considered the world’s largest study of its kind using AI to predict heart disease through an eye scan.

The innovation signals a shift towards an AI-powered paradigm for scientific discovery. Google CEO, Sundar Pichai during his keynote speech at Google I/O 2018 said, “AI is one of the most important things humanity is working on. It is more profound than electricity or fire.”