AI will be a powerful adjunct in ophthalmologists’ offices
Click Here to Manage Email Alerts
Artificial intelligence and its partner, machine learning, are defined in Wikipedia as “any system that perceives its environment and takes actions that maximize its chance of achieving its goals.”
As I was preparing this commentary, I took a few minutes to remember the times I had used AI in the past week. I thought of several web searches on my computer, use of the GPS in my phone (which gave me three alternative routes to an unfamiliar destination, including an estimate of the time in minutes required with each route), a few shows I was directed to on Netflix, a couple of questions answered by Siri, multiple targeted advertisements received on my computer, phone and by mail, fingerprint recognition as I opened my computer, and iris recognition as I passed through CLEAR security at the airport on my way to a business meeting, along with the knowledge that advanced AI was being used by my pilot to safely operate the aircraft I was about to enter. Artificial intelligence is a significant part of each of our daily activities, and despite some concerns to the contrary, in every case, except unwanted spam advertising, it has enhanced quality of life.
What about the practice of medicine and especially ophthalmology? Our electronic medical records and computerized billing systems make good use of AI, especially for appointments, prescriptions, coding and billing. However, as I think about it, most of us are using AI much more frequently out of the office in our personal lives than while seeing patients. As I look to the future, this is certain to change.
The diagnosis and proper management of glaucoma is to me especially complex and would benefit greatly from the application of AI, especially when we start to use genetic profiles to help calculate individual risk and expected efficacy of alternative therapies. We already have the Hill-RBF calculator in cataract surgery. In the future, AI applied to preoperative OCT images might well help us program our phacoemulsification machines to ideal parameters for each individual cataract. AI is widely used in MRI and CAT scans by neuro-ophthalmology. The proper diagnosis and management of retinopathy of prematurity seems ideal for the use of widefield retinal photography combined with AI-assisted interpretation, often by an examiner remote from the patient. Retinal imaging in general is a natural for AI assistance in interpretation. This will become even more helpful as home diagnostic devices become more prevalent.
In the not too distant future, the ophthalmologist may have hundreds if not thousands of data inputs to review from patients using home diagnostic units in glaucoma, diabetic retinopathy and age-related macular degeneration. This task will be nearly impossible without the application of AI and machine learning to screen those data sets and select the ones that require careful physician review or an urgent office appointment. Retinal photography may well prove to be an ideal window into the brain and vascular system for diagnosis and treatment of Alzheimer’s disease, hypertension as well as arteriosclerotic heart and vascular disease.
Much like the advanced AI systems used by commercial jet pilots that allow them to transport their passengers more safely, AI will be an adjunct to the well-trained ophthalmologist when treating patients, not a replacement. Investment of human and financial capital into the application of AI to medicine and ophthalmology is significant, and the power of the innovation cycle will bring us some remarkable advances in this decade.
Collapse
Read more about
Click Here to Manage Email Alerts