Shenbagam, Assistant Professor, Department of Optometry, Sushant University
In today’s era ,digital innovations which includes AI ,5g Telecommunication networks and the Internet of thigs (IoT) offers lots of opportunities not only to eye care but all allied health care sectors .These breakthroughs in technologies paves way to the early detection and prevention of the disease per se.
This article is focused only on Primary Eye Care.AI Plays humongous role to tackle Diabetic Retinopathy, Retinopathy of Prematurity ,Age related macular degeneration, Glaucoma ,refractive error correction and cataract.
Illustration of Artificial Intelligence (AI),Machine Learning (ML) & Deep Learning (DL)
Diabetic Retinopathy Screening
Firstly ,lets us try to understand the Diabetes. As we all know Diabetes has several factors including the genetic and environmental factors characterized by permanent rise in blood sugar however the major complications of DM Type 2 ,is it damages the heart,kidneys,brain, eyes and nerves.
The distance between the technology discovery and clinical practical technology adoption through AI helps easier to detect and treat the eye conditions before any loss of vision starts to occur3.
Diabetes is one of the major causes for retinal blindnes.Telemedicine and Teleophthalmology have been a real eye opener particularly in rural areas. Recent study on Artificial Intelligence Detection of Diabetic Retinopathy conducted in the year 2022 compared the general Ophthalmologists, retina specialists and Eye Art Artificial Intelligence system in detecting mild Diabetic Retinopathy. It was found that AI system had higher sensitivity(96.5%) for detecting mild Diabetic Retinopathy than either general Ophthalmologists or retina specialists compared with the clinical reference standard .It can potentially serve as a low -cost -point of care diabetic retinopathy detection tool and help address the diabetic eye screening burden4.
The advantage of AI Screening systems is mainly ,point-of-care access and the potentially lower operating cost as a result of automatic interpretation of the images and referral to an eye care specialist only as needed
Applications of AI in Glaucoma Screening, Referral, Diagnosis, and Forecasting
Glaucoma is the second leading cause of blindness worldwide7.Increased Intraocular pressure (IOP), Visual field loss and Optic neuropathy are the major changes in glaucoma and the disease per se is irreversible . Glaucoma is also calledas silent thief of sight. Diagnosis is the first step in preserving the central and peripheral Vision .Along with comprehensive eye examination ,Retinal imaging test and Visual Field examination forms the basis for the assessment and diagnosis of glaucoma.
Fig 2 : Evolution of AI in Glaucoma
In the real world where picture acquisition and interpreting takes time ,AI models could minimize subjectivity by interpreting and quantifying retinal and optic nerve images. AI can be also be utilized to optimize workflows and processes in glaucoma clinics that may lead to more time for clinicians to interact with patients thus enhancing overall care. AI can be used to quantify optic cup, disc, and rim characteristics in fundus images, retinal layers in OCT images, and patterns of VF loss. Such applications hold promise for providing improved glaucoma assessment, as well as forecasting, screening, diagnosing, and prognosing glaucoma6.
Clinical Applications of AI in Optometry
Online platforms utilizing AI can automate care where possible ,identify at-risk patients and provide the range from digital triage to higher levels of care for those patients when necessary. These platforms can also be utilized to deliver timely instruction and education before and after exams & thereby enhance remote patient management and treatment compliance
AI can serve as an assistant to help organize the clinic visit by triaging patients and providing a working diagnosis ,determine the appropriate testing and room requirements and anticipate the procedures or referrals (if any) that may be needed for a patient. The Optometrist can then conduct a more focused ,efficient ,effective evaluation and management of a patient.
Depending on the reason for the patients visit ,the severity, any condition they might present with and the patient’s location ,the optometrist may not always need to immediately be in the same room at the same time as patient. An image can be obtained by an assistant and that image cane be interpreted by an AI system before being forwarded to an optometrist.Ideally,the patient can then be scheduled to be seen by an optometrist at an appropriate time depending on severity.
There are few shortcomings of this asynchronous care model as a stand-aloneexam. To name a few, just as case history questions need to be adapted depending on the responses ,examination of the anterior segment, fundus ,use of vital dyes, manipulation of lids, extent of eye coordination testing all depend on the findings of the initial evaluation. Due to these artifacts synchronous model would be more advantageous.
Other than patient intake and triage ,AI and Virtual reality reality can be used for training ( example retinoscopy, gonioscopy) and also for patient education ,explaining conditions and prognosis of the diseases such as Cataract, Age related Macular degeneration, Glaucoma, Dry eyes and their respective treatment options.AI applications can be widely used in CL fitting including Ortho-K.
Tele-optometry in practice :
Tele-optometry came to the forefront across the globe as option for delivering clinical care during the COVID-19 lockdown era. Since then, Tele optometry is increasing and being viewed and considered as a viable option for providing clinical care to population groups with access challenge especially in remote and rural communities where the medical examinations are much limited.
Conclusion:
AI is all around us and will stay in health care .Several studies have shown that performance of AI is equal to and even superior to that of Optometrists &Ophthalmologists in many diagnostic and predictive tasks. However , future studies should be focused on issues such as real-worldperformance,generalizabilty and interpretability of AI. Thus ,AI will be transformative in health sciences and medicine by allowing automated analysis of telemetrically gathered data eventually leading to better eye and health care.
