The Eyes Have It

A new artificial intelligence (AI) system is designed to analyze images of the retina for early signs of degenerative disease — from dementia to heart disease to stroke. While diagnostic medicine has traditionally focused on distinct organs, diseases or physiological systems, this approach looks toward a more cohesive model of overall health. The ultimate goal is to reduce the need for expensive equipment and multiple tests that make preventative screening cost prohibitive.

In a project known as AlzEye, data from approximately 300,000 patients (age 40 and older) was gathered from hospitals throughout the U.K. over a period from 2008 through 2018. The data includes retinal scans as well as other health-related information, but excludes names and other readily identifiable personal information. In November of last year, the data sets were finally linked and the correlation process is now underway.

The researchers believe that the sheer number of data points will be the study’s greatest asset. Known as the Moorfields data set, the new database contains 12,000 major cardiac event cases and data on 11,900 stroke patients, compared with another widely used database that contains only 631 and 1,500 cases, respectively. Although there is no guarantee that correlations between retinal scans and Alzheimer’s disease will be revealed, the system will be able to check for other patterns as well. And if retinal scans prove to be a reliable window to other general health metrics, they could prove to be an inexpensive alternative to more expensive testing for early detection of chronic diseases.