Smart Phone Adaptor Accurately Detects Diabetic Retinopathy

Researchers in India were able to successfully use retinal diagnostics via smartphone adaptors to detect a diabetes-related eye disease, according to a study published in Ophthalmology.

Diabetic retinopathy (DR) is the leading cause of visual impairment among working-age adults in the world. Around 80% of individuals with DR reside in low- and middle-income countries while it is estimated that 8 out of 10 people with diabetes worldwide live in developing and emerging countries.

Early screening and treatment can prevent irreversible blindness caused by the condition. However, a lack of resources largely prevents DR screening implementation in these regions, which often have a poorly resourced health system.

As a potential low-cost solution to the lack of screening, researchers tested various smartphone-based fundus imaging (SBFI) devices by conducting examinations among individuals in South India.

The study included examinations of 381 eyes from 193 patients with diabetes, recruited from outreach eye clinics in South India between April and May of 2017. Researchers carried out 4 technically different approaches of SBFI in terms of image quality and diagnostic accuracy.

Three approaches were based on direct ophthalmoscopy (using the Peek Retina, D-EYE, and a do-it-yourself solution) and one (the Paxos Scope adapter) was based on indirect ophthalmoscopy. The methods were compared based on image quality criteria, field-of-view, examination time, and diagnostic accuracy to detect DR. The indirect SBFI device was also equipped with a Pan Retinal 2.2 lens.

The average age of study participants was 56.64 (SD 10.85) years and 59% were female. Baseline mean visual acuity was measured at 0.43 (0.5) (20/50). Examinations consisted of documenting changes in the retina by filming the back of the eye with a camera. In practice, anyone with a smart phone could carry out the examination then send images to an ophthalmologist for diagnosis.

A total of 35 hours of video material and 1800 single images were acquired. Although all 4 approaches yielded sharp images for most of the examinations, SBFI based on indirect ophthalmoscopy yielded the best image quality (P <.01), the largest field-of-view, and the longest examination time (111 vs. 68—86 seconds, P <.0001).

Researchers also found “sensitivity/specificity to detect DR were highest for the indirect SBFI approach (0.79/0.99 for any and 1.0/1.0 for severe DR, 0.79/1.0 for diabetic maculopathy [95% CI for all]).”

The indirect method, consisting of an adapter with an additional lens attached to the smartphone, “allowed almost 80% of eyes with any retinal changes to be detected, even in the early stages. Advanced damage could even be diagnosed 100% of the time,” Maximilian Wintergerst, MD, a lead author of the study, said in a statement.

Working with the Sankara Eye Foundation in India, researchers are currently developing an app to create an encrypted electronic patient file for each patient on the smartphones used for the examination. The app will contain both the images and findings of the doctor who reviewed them. Artificial intelligence to pre-evaluate the images is also in development.

“Coronavirus disease 2019 (COVID-19) has further necessitated the need for us to explore methods of reducing patients visiting the hospital,” said coauthor Mahesh Shanmugam, PhD. “This modality is promising in increasing efficiency of screening for retinal changes in diabetics.”

Reference

Wintergerst MWM, Mishra DK, Hartmann L, et al. Diabetic retinopathy screening using smartphone-based fundus imaging in India [published online May 24, 2020]. Ophthalmology. doi:10.1016/j.ophtha.2020.05.025