Visulytix’s AI Tool Equivalent to Human Experts in Glaucoma Detection

Source: Visulytix

Visulytix announced that Pegasus-disc, its artificial intelligence (AI) decision support tool for the analysis of retinal images, has been shown to perform similarly to the consensus opinion of human experts when detecting glaucoma from optic disc images. This follows a study conducted on the images of 186 patients at a major Harvard University teaching hospital, according to a company news release.

A study performed in collaboration with Dr. Brian Song and Professor Louis Pasquale at Massachusetts Eye and Ear Infirmary, Boston, compared the diagnostic accuracy of optic disc image evaluation by a group of glaucoma specialists to Visulytix’s Pegasus-disc, an automated, deep learning-based decision support tool.

The study concluded that Pegasus’ artificial intelligence tool was equivalent to the consensus of two human experts in the detection of glaucoma, and in some cases may have higher sensitivity. It is now being extended to include up to 400 subjects. The study was presented at the Association of Research in Vision and Ophthalmology (ARVO) international meeting being held in Honolulu, Hawaii on Monday, 30th April 2018.

“Visulytix’s technology has enormous potential for high volume screening of patients in telemedicine programs” Professor Louis Pasquale of Harvard Medical School faculty, said in the news release. “With the prevalence of glaucoma, along with healthcare costs, rising, this technology will permit high quality care and make it accessible to patients around the world.”

The abstract was entitled Automated Evaluation of Optic Disc Images for Manifest Glaucoma Detection Using a Deep-Learning, Neural Network-Based Algorithm.”

“It was a great opportunity to present the findings of our study with Harvard at such a prestigious conference. The results of the study were extremely encouraging as we work towards taking Pegasus to commercialization worldwide,” said Sameer Trikha, chief medical offer of Visulytix.

Related Content