Trained Imagers Catch Most Diabetic Retinopathy at Point of Care

Veronica Hackethal, MD

June 15, 2015

Evaluation for diabetic retinopathy at the point of primary care, using ultrawide field images taken by trained, certified nonphysician imagers, misses less than 0.1% of patients who need referral to a specialist, according to a new study published online June 1 in Diabetes Care.

The new approach could save time and resources, with the researchers estimating it could reduce the workload of centralized reading centers by 60%.

"Appropriately trained and certified imagers following a defined imaging and grading protocol can accurately evaluate ultrawide field retinal images for the presence of either diabetic retinopathy or referable diabetic retinopathy at the time of imaging," commented first author Paolo Antonio Silva, MD, of Joslin Diabetes Center at Harvard Medical School, Boston, Massachusetts, and the University of the Philippines, Manila.

"Real-time evaluation at the time of retinal imaging may result in a substantial reduction of centralized reading-center burden and speed delivery of information and education to the patient," he added. "The accurate identification of referable diabetic retinopathy allows prompt eye-care referral, reducing the burden of false positives [to reading centers]."

Ultrawide retinal field imaging uses a scanning laser and an ellipsoid mirror to make high-resolution retinal images. The technique is more efficient and captures over twice the retinal surface as dilated seven-field photography, considered the gold standard of retinal photography, the researchers explain.

Dr Silva noted, however, that the cost of ultrawide field imaging devices at the present time "is prohibitive," especially in places with small populations and a limited number of diabetes patients. But prices will come down in time, he added.

Asked to comment, T Mark Johnson, MD, FRCSC, an attending surgeon at the Retina Group of Washington, DC, who was not involved in the study, agreed that the current costs of the cameras used in ultrawide field imaging, are "significant" and could therefore limit their use in screening programs.

And he claims that past studies have suggested that even three-field photography (just taking photos of the posterior pole of the retina) may be as good as ultrawide field imaging for screening for significant diabetic retinopathy. Therefore, more research is needed to justify the added costs of ultrawide field imaging, he pointed out.

Key to Success Is Proper Training of Imagers

Dr Silva and colleagues explain that patients with diabetes require lifelong ophthalmic care that generally includes an annual retinal evaluation and that, given the rapidly growing population affected by diabetes, it is estimated that in 20 years, around 2.7 million eyes worldwide will need to be evaluated each day just to fulfill these needs.

"This enormous task is unlikely to be accomplished by the current approaches of diabetes eye-care programs," they observe.

At the adult diabetes clinic at the Joslin Diabetes Center, they have developed a telemedicine retinal photography program using trained and certified nonphysician retinal imagers. This transitioned to use of ultrawide field imaging in 2012.

The current study was prospectively conducted from October 2013 to September 2014.

The imagers received 4 hours of training lectures, 12 hours of guided image review, and a 1-month trial period during which a specialist reviewed questionable images and findings. The retinal imagers also had prior experience using other retinal imaging devices.

They used the Optos P200MA and P200C devices to perform ultrawide field imaging and determined at the point of care whether patients had diabetic retinopathy that required referral to a specialist. The imagers assessed only the presence or absence of diabetic retinopathy (gradable, present, or referable), not its severity.

Dr Silva and colleagues then compared these real-time evaluations with masked evaluations at a centralized reading center.

The analysis included images from 1989 patients (3978 eyes). Reading-center evaluations showed that 3769 eyes (94.7%) were gradable for diabetic retinopathy: 1376 (36.5%) had retinopathy, and 580 (15.3%) had referable retinopathy.

Compared with reading-center evaluations, point-of-service evaluations had a high sensitivity for identifying more than minimal diabetic retinopathy, 0.95, and referable retinopathy, 0.99.

And point-of-service evaluations had a specificity for detecting more than minimal diabetic retinopathy and referable diabetic retinopathy of 0.84 and 0.76, respectively.

Trained nonphysician imager evaluation missed only three patients with referable diabetic retinopathy, with a false negative rate of 0.01.

"Retinal imagers were under the direct supervision of a retina specialist and received a standardized method of certification and training," Dr Silva pointed out.

"This aspect of the program is critical should similar studies or programs be implemented in other populations," he stressed.

Dr Johnson agrees that "the key to success is proper training. Both the readers in this study had extensive experience with fundus photography prior to undergoing training at the reading center. In addition, they were carefully supervised and provided with ongoing training and feedback."

Replacement Needed for Traditional Screening

Dr. Johnson agrees with the Joslin team that something has to change: currently, less than 60% of diabetic patients receive adequate screening for retinopathy, he told Medscape Medical News.

And the time lag between when images are taken at primary-care centers, interpreted at reading centers, and reported back to the primary-care provider means that patients with significant diabetic retinopathy may be lost to follow-up, he noted.

More immediate results after imaging could also improve patient education, he added.

"With the exploding population of diabetic patients, [sustaining] traditional screening methods for diabetic retinopathy is impossible," Dr Johnson emphasized, "The use of remote telescreening for diabetic retinopathy is of great importance in addressing this public-health need."

One of the two Optos P200 instruments used in this study was temporarily lent to Joslin Diabetes Center by Optos (Dunfermline, Fife, Scotland). The authors and Dr Johnson report no conflicts of interest.

Diabetes Care. Published online June 1, 2015. Abstract


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.