Study examines use of deep machine learning for detection of diabetic retinopathy

November 29, 2016

In an evaluation of retinal photographs from adults with diabetes, an algorithm based on deep machine learning had high sensitivity and specificity for detecting referable diabetic retinopathy, according to a study published online by JAMA.

Among individuals with diabetes, the prevalence of diabetic retinopathy is approximately 29 percent in the United States. Most guidelines recommend annual screening for those with no retinopathy or mild diabetic retinopathy and repeat examination in 6 months for moderate diabetic retinopathy. Retinal photography with manual interpretation is a widely accepted screening tool for diabetic retinopathy.

Automated grading of diabetic retinopathy has potential benefits such as increasing efficiency and coverage of screening programs; reducing barriers to access; and improving patient outcomes by providing early detection and treatment. To maximize the clinical utility of automated grading, an to detect referable diabetic retinopathy is needed. Machine learning (a discipline within computer science that focuses on teaching machines to detect patterns in data) has been leveraged for a variety of classification tasks including automated classification of diabetic retinopathy. However, much of the work has focused on "feature-engineering," which involves computing explicit features specified by experts, resulting in algorithms designed to detect specific lesions or predicting the presence of any level of diabetic retinopathy. Deep learning is a machine learning technique that avoids such engineering and allows an algorithm to program itself by learning the most predictive features directly from the images given a large data set of labeled examples, removing the need to specify rules explicitly. Application of these methods to medical imaging requires further assessment and validation.

In this study, Lily Peng, M.D., Ph.D., of Google Inc., Mountain View, Calif., and colleagues applied to create an algorithm for automated detection of diabetic retinopathy and diabetic macular edema in retinal fundus (the interior lining of the eyeball, including the retina, optic disc, and the macula) photographs. A specific type of network optimized for image classification was trained using a data set of 128,175 retinal images, which were graded 3 to 7 times for diabetic retinopathy, diabetic macular edema, and image gradability by a panel of 54 U.S. licensed ophthalmologists and ophthalmology senior residents between May and December 2015. The resultant algorithm was validated using 2 separate data sets (EyePACS-1, Messidor-2), both graded by at least 7 U.S. board-certified ophthalmologists.

The EyePACS-1 data set consisted of 9,963 images from 4,997 patients (prevalence of referable diabetic retinopathy [RDR; defined as moderate and worse diabetic retinopathy, referable diabetic macular edema, or both], 8 percent of fully gradable images; the Messidor-2 data set had 1,748 images from 874 patients (prevalence of RDR, 15 percent of fully gradable images). Use of the algorithm achieved high sensitivities (97.5 percent [EyePACS-1] and 96 percent [Messidor-2]) and specificities (93 percent and 94 percent, respectively) for detecting referable diabetic retinopathy.

"These results demonstrate that deep neural networks can be trained, using large data sets and without having to specify lesion-based features, to identify diabetic retinopathy or in retinal fundus images with high sensitivity and high specificity. This automated system for the detection of offers several advantages, including consistency of interpretation (because a machine will make the same prediction on a specific image every time), and specificity, and near instantaneous reporting of results," the authors write.

"Further research is necessary to determine the feasibility of applying this algorithm in the clinical setting and to determine whether use of the algorithm could lead to improved care and outcomes compared with current ophthalmologic assessment."

Explore further: Barriers for diabetic retinopathy screening vary

More information: JAMA, DOI: 10.1001/jama.2016.17216

Related Stories

Barriers for diabetic retinopathy screening vary

October 20, 2016
(HealthDay)—Patients and health care providers have markedly divergent perceptions of barriers to diabetic retinopathy screening, according to a study published in the Oct. 6 issue of the U.S. Centers for Disease Control ...

Treated diabetic retinopathy rare in children with T1DM

October 19, 2016
(HealthDay)—Treated diabetic retinopathy (DR) is extremely rare among children with type 1 diabetes, according to research published online Oct. 7 in Diabetes Care.

Lucentis approved for diabetic retinopathy

February 9, 2015
(HealthDay)—A drug to treat the most common eye disease among diabetics has been approved by the U.S. Food and Drug Administration.

Study seeks to improve diabetic eye health

October 11, 2016
While diabetes is the most common cause of blindness in the working-age population in the U.S., only about 55% of lower income adults living with the disease undergo the retinal screening needed to detect and help prevent ...

Vision-threatening stages of diabetic retinopathy associated with higher risk of depression

July 7, 2016
In a study published online by JAMA Ophthalmology, Gwyneth Rees, Ph.D., of the University of Melbourne, Australia, and colleagues examined the association between severity of diabetic retinopathy and diabetic macular edema ...

Fewer ophthalmologists equals less diabetic eye care

April 15, 2014
(HealthDay)—In areas with less access to ophthalmologists, fewer individuals with diabetes, diabetic retinopathy, and age-related macular degeneration (ARMD) receive care, according to a study published in the April issue ...

Recommended for you

World's blind population to soar: study

August 3, 2017
The world's blind will increase threefold from about 36 million today to 115 million in 2050 as populations expand and individuals grow ever older, researchers said Thursday.

Simulations signal early success for fractal-based retinal implants

July 27, 2017
Computer simulations of electrical charges sent to retinal implants based on fractal geometry have University of Oregon researchers moving forward with their eyes focused on biological testing.

Scientists regenerate retinal cells in mice

July 26, 2017
Scientists have successfully regenerated cells in the retina of adult mice at the University of Washington School of Medicine in Seattle.

Genome editing with CRISPR-Cas9 prevents angiogenesis of the retina

July 24, 2017
A research team from the Schepens Eye Research Institute of Massachusetts Eye and Ear has successfully prevented mice from developing angiogenesis of the retina—the sensory tissue at the back of the eye—using gene-editing ...

Too little vitamin D may hinder recovery of injured corneas

July 24, 2017
Injury or disease in combination with too little vitamin D can be bad for the window to your eyes.

Combination of type 2 diabetes and sleep apnoea indicates eyesight loss within four years

July 4, 2017
Research led by the University of Birmingham has discovered that patients who suffer from both Type 2 diabetes and obstructive sleep apnoea are at greater risk of developing a condition that leads to blindness within an average ...

0 comments

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.