Artificial Intelligence With Deep Learning Technology Looks Into Diabetic Retinopathy Screening

Gulshan and colleagues report the use of deep learning technology for diabetic retinopathy screening. Using large data sets of images to first "train" an algorithm, then using 2 separate data sets to "test" this algorithm, the authors showed that this novel diabetic retinopathy s...

Full description

Saved in:
Bibliographic Details
Published in:JAMA : the journal of the American Medical Association 2016-12, Vol.316 (22), p.2366-2367
Main Authors: Wong, Tien Yin, Bressler, Neil M
Format: Article
Language:English
Subjects:
Artificial Intelligence
Diabetic retinopathy
Diabetic Retinopathy - diagnosis
Edema
Humans
Learning
Mass Screening
Medical screening
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Gulshan and colleagues report the use of deep learning technology for diabetic retinopathy screening. Using large data sets of images to first "train" an algorithm, then using 2 separate data sets to "test" this algorithm, the authors showed that this novel diabetic retinopathy screening software based on deep learning techniques had an 87% to 90% sensitivity, 98% specificity, and an area under the receiver operating characteristic curve of 0.99 for detecting referable diabetic retinopathy, which was defined as moderate or worse diabetic retinopathy or referable diabetic macular edema. Here, Wong and Bressler present several challenges to immediate adoption of the software for clinical translation and utility in diabetic retinopathy screening programs.
ISSN:0098-7484
1538-3598
DOI:10.1001/jama.2016.17563