IDx’s automated diabetic retinopathy screening software scheduled to complete FDA clinical trial by end of summer 2017 – July 20, 2017

After 20 years of work spearheaded by retina specialist Dr. Michael Abràmoff (IDx President & Director), IDx recently announced that IDx-DR, an automated diabetic retinopathy screening software, is slated to complete a multi-center FDA clinical trial by the end of summer 2017. The trial is taking place at 10 sites across seven states, enrolling more than 850 people with diabetes. In October, the company published impressive results of its deep learning algorithm in Investigative Ophthalmology & Visual Science, demonstrating detection with 97% sensitivity and 87% specificity. While this FDA clinical trial is IDx’s first foray into the US market, IDx-DR version 2.0, a secure cloud version of the software, recently received CE certification as a Class IIa Medical Device and is currently for sale in Europe. This exciting area of automatic, computer-driven retinopathy detection has received a fair amount of press in recent months. One of IDx’s key partners, IBM Watson Health, has designed its own diabetic retinopathy detection software, recently demonstrating 86% accuracy in classifying eye scans by severity. As partners, IBM plans to distribute the IDx software in the 31 countries that comprise the European Economic Area – distribution may expand to Australia, Canada, and the US, pending regulatory approval. Google is also developing an automated deep learning algorithm aimed at detecting retinopathy and macular edema from images, with results published in JAMA in December 2016 demonstrating over 90% sensitivity and specificity. We’re excited to see expansion in this promising field, most recently noted by the notable NIDDK Director Dr. Judith Fradkin, who gave an inspiring talk at this year’s ADA detailing numerous success stories of early detection and treatment of retinopathy using machine learning. We love how these technologies work to break down geographic barriers, boosting screening rates and allowing clinicians to focus less on computer screens and more on in-person care. We see a lot of potential here to help doctors do an even better job with higher predictability.


-- by Maeve Serino, Brian Levine, Adam Brown, and Kelly Close