Sign In

Communications of the ACM

ACM TechNews

Artificial Intelligence in Medicine: Now There's an App For That

View as: Print Mobile App Share:
An image of a retina.

Diabetic retinopathy is a condition in which damage to the blood vessels supplying the retina impairs vision.

Credit: ALAMY

Nearly 80 percent of diabetics will develop retinal damage within a decade due to diabetic retinopathy, a condition in which damage to the blood vessels supplying the retina impairs vision. The condition can cause blindness if not detected and treated early and the California HealthCare Foundation (CHCF) believes artificial intelligence (AI) technology could help make detection easier by automating the screening process.

Lacking any AI expertise of its own, the organization turned to Kaggle, a website that helps organize competitions for statisticians and data scientists. CHCF provided $100,000 in prize money and uploaded thousands of retinal images, featuring both diseased and healthy retinas, to enable the website's members to develop an algorithm that could detect the disease. T

rained doctors agree with each other about whether or not an image of a retina shows signs of diabetic retinopathy about 84 percent of the time. Within five months, the contest's winner, University of Warwick statistician Benjamin Graham, was able to develop an algorithm that agreed with doctors 85 percent of the time.

Although CHCF wants to introduce the technology in clinics across California, an uncertain regulatory landscape regarding medical AI is making it hold the technology in reserve for now.

From The Economist
View Full Article


Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account