Artificial intelligence can help bridge care gaps and facilitate detection and follow-up for patients with vision-threatening macular degeneration.
Age-related macular degeneration (AMD), a common retinal disease, is the leading cause of blindness in people over the age of 50 in the developed world. The retina is the layer that lines the inner part of the eye which receives light and converts it to electrical signals, which are then sent to the brain for further processing. The macula, an area of the retina responsible for central vision, undergoes a variety of benign changes as we age; however AMD is the result of complex aging processes which can lead to retinal dysfunction and vision loss.
There are two forms of the disease, dry and wet AMD. In dry AMD, fatty deposits accumulate in the retina with or without macular atrophy, but no abnormal blood vessel formation occurs. Wet AMD, on the other hand, is characterized by abnormal blood vessel growth in addition to the dry AMD changes. This form of the disease is referred to as ‘wet’ because the new blood vessels may leak and cause retinal swelling (i.e., macular edema). According to the American Academy of Ophthalmology, it is estimated that 12.5 million individuals are affected by AMD in the United States. An estimated 71,000 new cases of wet AMD develop each year in North America.
The treatment of the wet form of AMD has been revolutionized with the introduction of ophthalmic injections of anti-vascular endothelial growth factor (anti VEGF) agents. These agents inhibit blood vessel growth and were originally developed to treat cancer by depriving the tumor of its blood supply. They later have been shown to be particularly effective in halting the growth of new blood vessels in the retina. Today they are in wide clinical use both in oncology as well as in the treatment of various eye diseases. They are injected into the eye in a short and nearly painless procedure, to inhibit the formation and development of new abnormal blood vessels in AMD.
In the “dry” form of the disease, select patients can significantly reduce the risk for progression simply by taking food supplements (specific formulas containing vitamins and a few other micronutrients such as zinc). The safety and efficacy of these supplements was demonstrated in large scale clinical trials. Treatments for severe cases of the dry form of the disease are currently being investigated, and some promising agents are expected to hit the market in the foreseeable future.
AI can be incorporated to facilitate the different phases in the effort to prevent vision loss caused by AMD:
AMD is currently diagnosed by an ophthalmologist in a dilated fundus examination; i.e the patient’s pupils are pharmacologically dilated to enable the physician a good view of the retina using a slit lamp and a magnifying lens. The American Academy of Ophthalmology recommends receiving a complete eye exam to screen for common eye disease once at the age of 40, and every one to two years after the age of 65. Despite this, practice shows that low percentages of healthy individuals actually adhere to these guidelines. Presumably, this is due to multiple factors including psychological barriers, mobility issues, insufficient medical coverage and the busy modern lifestyle.
The use of AI may contribute to improved rates of screening. Some AI-based diagnostic solutions can accurately detect AMD using a single image obtained using a portable retinal camera. This technology can make AMD screening possible while visiting the primary care physician, at the pharmacy, or even at home. Accessible screening will help in closing a significant care gap, providing opportunities for early treatment (i.e spotting patients with early AMD who can benefit from supplements). AI could also be used to accurately predict the likelihood of developing AMD from the retinal images themselves and also using additional patient data such as age, lifestyle, and genetic predisposition. This information could then be used to guide decisions related to preventative treatment strategies. Patients who are found to be suffering from AMD or who have already been diagnosed should have regular follow-up visits with an ophthalmologist. Different imaging modalities are used during these visits to monitor disease progression and response to treatment. The main imaging modality in use is Optical Coherence Tomography (OCT), which provides a detailed view of the retinal layers. Interpreting OCT images takes time and expertise.
AI algorithms can assist clinicians in analyzing images to save time and make the clinical encounter more efficient, as well as support clinical decision making. AI could also be used to monitor the progression of AMD over time and to determine the effectiveness of treatments. This information can be used to develop personalized treatment plans and drug choice to improve patient outcomes.These changes will leave the physician with more time for direct patient interaction, optimize patient care, and help reduce the treatment burden.
In summary, AMD is a major cause of vision loss in the elderly, and a growing global concern given the constant increase in lifespan. Once diagnosed, the disease burden can be alleviated with proper care, making early detection and routine follow-ups essential. The integration of AI into healthcare can make disease detection more accessible and efficient. By leveraging AI, physicians will also be able to provide patients with AMD better, more accurate care.