The potential applications of artificial intelligence (AI) in eye health are numerous, but two areas are already real and growing that will have an immediate impact and appeal to eye care in the UK. One is the use of AI in patient support. The simplest version is an AI chatbot that can provide an “always on” solution to patient questions. The other is the application of AI based on deep learning to medical images, including retinal images.
Detecting aberrant features in retinal vasculature, lesions, optic disc images, and optical coherence tomography (OCT) images can facilitate different treatment models while standardizing and improving treatment. Ocular imaging is key to audits, telemedicine, and virtual support, all of which enable top-tier clinicians to plan safe and effective care for their patients without having to see each individual individually. will be
AI is already within us. Chatbots are used in several medical fields, including ophthalmology, and in glaucoma and diabetic eye care, several options already exist to help identify probable abnormalities, all of which are We are not using deep learning. So far, the most useful role appears to be screening and detection in pathologies involving characteristic changes in fundus appearance.
The top question is always, “What does it mean for patient care?”
Comorbidities are fairly common in the elderly cohort, with hypertension and cardiovascular disease, as well as other ocular pathology, to name just two, so there is always a screening component at the pre-assessment visit. Clinicians should confirm suitability for surgery and whether there are any comorbidities that require prior management. Any tool that helps identify retinal disease, especially through hazy media, is welcome.
The preoperative evaluation also grades lens opacification, but this is subject to clinician interpretation. For example, if this could be automated to flag posterior polar cataracts, it could be beneficial. Posterior pole cataract patients are at high risk of intraoperative complications, so a solid knowledge of the disease directly influences surgical planning.
We are always looking for intelligent solutions that improve patient care and outcomes, whether it be specialized wheelchairs for use with ophthalmic diagnostic equipment or digital AI solutions. Having access to technology that easily identifies abnormal retinal images and arouses suspicion in clinicians before they see a patient could increase the effectiveness of the examination. Digital imaging systems already facilitate shared care and virtual review through the exchange of accurate information. Leveraging AI to weed out false positives and false negatives from screening and monitoring services could further improve collaboration and reduce unnecessary or inaccurate and frustrating referrals to secondary care.
The top question is always, “What does it mean for patient care?” All other considerations are based on disease detection rates, compared to clinicians alone, earlier disease detection, and reduced need for clinic visits. increase. Refinement of referrals, screening, monitoring of stable eye conditions, and identification of progression are all resource-intensive areas for ophthalmology and optometry, and AI technology can be part of the solution. I can’t wait to see where it takes us.
