Genentech's algorithm enhances accuracy of macular disease detection

Genentech Inc. said that by applying its deep learning algorithm to color fundus photographs (CFPs), it was able to detect diabetic macular edema (DME) with up to 97% accuracy. The results were published in Investigative Ophthalmology & Visual Science.

While optical coherence tomography (OCT) is the gold standard in DME detection, it is has a high cost and technical limitations. CFPs are less expensive and easier to obtain, but the human eye cannot easily identify macular thickening. Genentech sought to increase the accuracy of DME detection using CFPs by developing a deep learning algorithm to identify macular thickening in the photos...