Image analysis ideal for AI in ophthalmology
Click Here to Manage Email Alerts
I wrote a perspective on the application of artificial intelligence and machine learning in ophthalmology a year ago. The interested reader can find that March 25, 2022, perspective in the Healio/OSN archives.
The specialty that utilizes AI the most is radiology. AI is especially well suited for evaluating images, making its broad use in radiology logical. Ophthalmology is also a very visual specialty, and today we can generate quality images of the periocular face, the anterior segment and the posterior segment. Posterior segment imaging led to the first FDA approval of an AI device in ophthalmology, the IDx-DR (Digital Diagnostics), which can evaluate a retinal photograph and determine if diabetic retinopathy is more than mild, triggering a referral to an eye care professional. This is a useful telehealth and screening tool, but it has not found broad application in ophthalmology offices.
I was surprised to learn that while we ophthalmologists have one FDA-approved indication for AI, the FDA has approved 521 AI devices in medicine, with 91 approved in 2022 alone.
In the application of AI, eye care is a late adopter. I believe there are many potential applications for AI in ophthalmology, especially on the diagnostic side. Because analysis of an image is ideal for AI, ocular surface disease, keratoconus diagnosis and evaluation for progression, cataract surgery biometry, glaucoma diagnosis, progression evaluation, perhaps even preferred treatment selection, and especially retinal diseases, including diabetic eye disease, exudative and dry age-related macular degeneration, geographic atrophy and even retinal dystrophies, are all good targets for AI. Advances in AI applications to ophthalmology will also be strengthened by broader application of genetic testing, interaction with other digital and virtual interfaces, and even the metaverse.
I remain convinced further application of AI will be a win for eye care professionals, their patients and the industry that supports us. Significant human and financial capital is being invested in this area, and I anticipate many more FDA-approved AI devices for our field in the next decade.
Collapse
Read more about
Click Here to Manage Email Alerts