Optical Device for Remote Diagnoses Stores Retina Images in the Cloud

Phelcom Technologies has created a portable, low-cost device that is connected to a smartphone and that takes precise images of the retina to detect back-of-the-eye (fundus) disease. Called the Eyer, the optical device, which is connected to the smartphone’s camera, lights and images the retina. A mobile application sends the images over the internet to Eyer Cloud, which stores and manages patient files. If the user does not have Wi-Fi or access to a 3G or 4G network, the images are stored in the smartphone and sent to the Eyer Cloud when an internet connection becomes available.

A mobile application operating the optical device sends images of the retina over the internet to Eyer Cloud, which stores and manages patient files. Courtesy of Phelcom Technologies.

Users of the Eyer set up an account to which images can be saved automatically. “We had to make sure data privacy would be guaranteed and develop a means of transmitting images at high speed to the cloud, so that they can be viewed online regardless of the device’s location,” Phelcom CEO José Augusto Stuchi said. Most ophthalmoscopes in current use have to be connected to a computer to save data to a hard disk and are not web-enabled, the Phelcom team said.

The Eyer is designed to be used for remote diagnosis. “We invested significantly in optics and in design,” Stuchi said. “One challenge was producing a portable version of a device that is typically very large. Another was enabling nonmydriatic operation so that high-quality images of the retina can be captured without the need for pupil dilation.” A trained technician can produce the images, and an ophthalmologist who specializes in retinopathy can then analyze them and write an expert report at another location. Phelcom is currently partnering with ophthalmologists to develop the reporting component of the system.

The medical reports are fed into a database that in the future could be used to train a computer to find patterns associated with ocular fundus diseases, especially diabetic retinopathy. The firm currently has images of more than 10,000 retinas and it expects to examine 50,000 patients in 2020. The accuracy of Phelcom’s artificial intelligence (AI) system for detecting diabetic retinopathy without human intervention is currently close to 80%, Stuchi said. As the Phelcom database expands, accuracy could reach 95%. The accuracy of IDx-DR, the first AI algorithm for detecting diabetic retinopathy to be approved by the U.S. Food and Drug Administration (FDA), is currently rated at 89.5%.