March 08, 2018
1 min read
Save

Machine learning may open new opportunities for telemedicine, screening

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

NEW YORK — Advances in optic disc analysis software may soon offer new possibilities for telemedicine and screening patients for early signs of disease, a speaker said here.

“There is some evidence starting to appear that this sort of system can certainly detect optic nerve pathology and fairly well, apparently at a level that is equivalent to ophthalmologists,” Sameer Trikha, BM, BSc, FRCOphth, MBA, said at the American Glaucoma Society annual meeting.

Pegasus, a retinal analysis decision support system developed by the artificial intelligence company Visulytix, has shown a comparable ability in judging optic disc photographs for glaucoma. The system has non-mydriatic or mydriatic cameras, offers a field of vision of 30° to 200°, and includes optic disc assessment, Trikha said.

A 2010 study published in Ophthalmology evaluated 243 ophthalmologists across 11 countries to determine diagnostic accuracy of judging optic disc photographs for glaucoma. Ophthalmologists were asked to grade 40 healthy eyes and 48 glaucomatous eyes with varying severity on stereoscopic slides. The participants had no time limit to assess the slides, and the overall diagnostic accuracy was 80.5%, Trikha said.

The Pegasus system was tested “out of box” vs. a single image, and the area under the curve was found to be 0.87. The accuracy of system was 83% compared with 80% for the ophthalmologists, Trikha said.

These systems and machine learning will support what ophthalmologists do by aiding learning and decision-making, he said.

“AI and algorithms themselves, they don’t provide and deliver safe and effective care. It’s people and systems that do that. It’s my strong belief that in the coming year and 2 years’ time, that ophthalmic specialists who are using these sorts of analog-based decisions or systems will ultimately start to replace those that don’t,” Trikha said. by Robert Linnehan

 

Reference:

Trikha S. Can a computer perform optic nerve assessment better than a glaucoma specialist? Presented at: American Glaucoma Society annual meeting; Feb. 28 to March 4, 2018; New York.

Disclosure: Trikha reports he is the chief medical officer for Visulytix.