Artificial Intelligence (AI) has a snowballing effect in the field of eye care. As a result, AI-enabled devices have started surfacing in eye clinics for diagnosing and screening ophthalmic diseases. In traditional settings, eye testing requires people to physically attend a clinic, and an eye health care professional to perform the test. However, this serves as a barrier to rural, elderly, and many indigenous or mobility-impaired patients seeking access to eye care. To overcome this barrier, researchers have recently developed an online vision test that replaces conventional Snellen charts. Most importantly, the AI-based online test produces more accurate results and can help patients track their vision health over time. Moreover, AI-based systems are not only limited to eye screening.
Ophthalmological diseases, with a high occurrence rate such as diabetic retinopathy, age-related macular degeneration, glaucoma, retinopathy of prematurity, age-related or congenital cataract, and few with retinal vein occlusion, have become the interest area for AI researchers worldwide. Over the next decade, Artificial Intelligence will likely transform how we observe patterns in data and how it can correlate with the treatment of eye health. This article explores some real-time applications of this nascent revolution.
AI in the Diabetic Retinopathy Screening System
Diabetic retinopathy damages the retina, the light-sensitive layer of tissue at the back of the eye. The tissue is not regenerative. In other words, once the retina is damaged it cant be cured. It is a major cause of blindness. Above all, it results in up to 24,000 cases each year in adults in the United States. However, diabetic retinopathy, if diagnosed before its symptoms appear, is manageable. Therefore, regular screening is crucial in decreasing the long-term effects of the disease. Yet, the number of patients is far more in number than are ophthalmologists in the world. For instance, ophthalmologists recommend more than half of diabetic patients do an eye check-up once a year. Still, assessing more than 30 million patients seems an invincible challenge for a limited number of eye specialists. To solve this problem, researchers took refuge with telemedicine known as IDx-DR.
The IDx-DR system photographs the retina tissue and then analyses the resulting images to detect early signs of diabetic retinopathy such as hemorrhaging. Additionally, after receiving only four hours of training, anyone with a secondary school education could operate this autonomous AI system. Moreover, within a span of a few minutes, it can tell whether a person has a severe case of diabetic retinopathy or not. Most importantly, the Food and Drug Administration (FDA) approved IDx -DR as the first device to provide a screening decision without the availability of a clinician. Eventually, the system will improve the speed and accuracy of large-scale screening programs.
In case, you missed reading Top AI-based Medical Imaging Tools In 2021
Artificial Intelligence in Self Eye Diagnosis
Researchers at DeepMind and Moorfields Eye Hospital developed an AI algorithm that learns to make referral decisions for 50 common eye conditions. In addition, it spots signs of eye disease in an OCT retinal scan and then decides the urgency with which a person should see a specialist. The sheer volume of cases ophthalmologists deal with every year is unimaginable. Hence, this system could ease the load for eye specialists.
Artificial Intelligence in the Prevention of Glaucoma
Glaucoma, a leading cause of irreversible blindness, affects ~64.3 million patients aged 40–80 years worldwide. This number is expected to increase to 112 million by 2040.
Glaucoma is an optic nerve disease, which progressively erodes vision. Although, lowering of intraocular pressure (IOP) in the eye has shown signs of delay in glaucoma progression. Meanwhile, Kazemian et al developed a clinical forecasting tool. It uses tonometry (an instrument for estimating intraocular pressure) and Visual Field data to predict disease paths at different target IOPs.
Nonetheless, researchers are working on the refinement of this tool. So that it integrates other ophthalmic and non-ophthalmic data to better establish target IOP. Finally, the use of machine learning algorithms will foretell the risk of requiring invasive surgery or losing functional vision from glaucoma.
In conclusion, future research is crucial in evaluating the clinical deployment and cost-effectiveness of different AI-based systems in healthcare practice. Existing applications are a decisive step to develop more AI systems based on the needs of patients. Although there are still challenges ahead, Artificial Intelligence will most likely impact the practice of ophthalmology in the coming decades.
