Skip to content

Global Promise

Insights on International Collaborative Health

Global Promise Home Care Providers AI and Eye Health

AI and Eye Health

By Dr. T. Y. Alvin Liu

Did you know the retina is actually part of the brain? It’s a direct extension of the central nervous system. So when people say eyes are the window to one’s soul, it’s perhaps not that far of a stretch after all.

I’ve been interested in the retina since I was in medical school. Being an avid photographer, I’ve always thought of the retina as the film in a camera, and as you know, if the film doesn’t work, the camera won’t work, either. Similarly, besides being critical for generating sharp vision, the retina is also where many serious systemic and neurological conditions first manifest.

As we research ways to leverage innovation to treat various ophthalmic conditions, ophthalmologists—particularly those who specialize in retinal care—are better positioned than many of our colleagues to take advantage of recent advances in artificial intelligence (AI). Compared with other medical fields, we have relatively large, well-annotated image datasets that are critical for AI research and applications.

My research focuses on how ophthalmology can benefit from deep learning, a specific kind of AI that is exceptionally adept at pattern recognition. We can use deep learning to extract useful information from large databases of medical images to automate disease diagnosis and predict disease progression.

I’m encouraged by recent examples of medical AI applications making an impact in day-to-day life. For example, experts have developed an algorithm that can automatically determine—without any human input—which patients should be referred to an ophthalmologist for diabetic retinopathy. Such application is already FDA approved and has the potential to dramatically improve access to care, efficiency and cost-effectiveness.

As we look ahead, we will strive to develop next-generation algorithms that can accurately predict disease progression and patients’ response to treatments. Also, most of the current algorithms can only handle 2D images from one imaging modality. Eventually, we will need to develop algorithms capable of analyzing 3D images, or images generated from different modalities simultaneously.

I am a futurist and optimist, and I believe AI will completely transform eye health and other areas of medicine in the near future. Already, an AI algorithm can predict with high accuracy a person’s gender, age and blood pressure just from a single retinal color photograph. This is clearly “super-human” intelligence, though in a very narrow sense.

Considering we are only at the infancy of applying artificial intelligence in medicine, I can’t wait to see how this field will empower clinicians and researchers at Hopkins and around the world in 10, 20 or 30 years.

Y. Alvin Liu, M.D., is an assistant professor of ophthalmology at the Wilmer Eye Institute, Johns Hopkins Medicine. He is subspecialty trained in the medical and surgical treatment of vitreoretinal diseases, including diseases such as retinal detachment, macular holes, diabetic retinopathy, vein occlusions and age-related macular degeneration. Dr. Liu's research interests center on the application of artificial intelligence in predicting, screening, diagnosing and treating ophthalmic diseases. He supports Johns Hopkins Medicine’s global health care mission by speaking on the health system’s behalf at leading industry events around the world, including the recent Arab Health Congress, the largest health care conference in the Middle East.


Guest Author

3 thoughts on “AI and Eye Health”

  1. Informative article , do u that 3D images may help patients with coloboma birth defects and missing part of the retina ?

    1. Global Promise Editor

      Dear Mr. Scott, for now, the AI ophthalmology work focuses on neurodegenerative diseases like AMD. You read more about the endeavor at:

      And here is some information on research specifically related to OCA:

      We hope this is helpful and thank you for reading.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.