Telemedicine and artificial intelligence (AI) provide solutions to the challenges faced by ophthalmologists and healthcare professionals around the world. Diseases such as diabetic retinopathy (DR), retinopathy of prematurity (ROP), age-related macular degeneration (AMD), glaucoma and other anterior segment disorders could be more easily predicted and detected with the help of these new technologies. New digital tools and the development of fifth-generation (5G) wireless networks, artificial intelligence (AI) approaches such as machine learning (ML) and deep learning (DL) and the Internet of Things (IoT) , or blockchain, have created new opportunities for the healthcare sector that offer great scenarios for improving diagnoses and making patient care more comfortable.
Moreover, the pandemic that we have unfortunately had to live with for more than a year now is prompting us to speed up the process of widespread telemedicine. Particularly in less industrialised countries where hospitals are often very far from villages.
Problems to be solved in ophthalmology with the help of artificial intelligence
-The WHO estimates that 420 million people live with various retinal diseases.
-5 billion people are at risk of becoming nearsighted by 2050 and will suffer from visual fatigue due to daily and frequent use of backlit devices.
-Age-related macular degeneration (AMD) will increase by more than 25% over the next 10 years, from 196 million to 243 million. Similarly, diabetic retinopathy (DR), a major cause of blindness, currently affects 146 million diabetics.
- 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.
-The current number of ophthalmologists globally and current drug treatments are insufficient to cover the demand.
Benefits of artificial intelligence solutions
- When diagnosed before symptoms appear, the disease can usually be managed and the worst outcome avoided.
- Regular screening is possible.
- Administration of appropriate pharmacological treatment to prevent disease progression.
- Improving the performance of ophthalmic products and enhancing tools for testing and diagnosing visual health, for the benefit of every patient, everywhere.
- From OCT scans, the system indicates how urgently a specialist should be contacted.
Potential applications of a Deep Learning (DL ) solution for DR screening.
AI in ophthalmology involves the use of labelled images to train algorithms to classify images of, for example, the ocular fundus and it would be more accurate to talk about Deep Learning (DL) which involves using whole images labelled with clinical diagnosis by experts, so that the algorithm “self-learns” predictive features for classification of diagnosis or severity, with better error rates than traditionally accepted.
Clinically acceptable performance of these DL algorithms in classifying ophthalmic imaging data such as colour photography of the ocular fundus (CFP) for various eye diseases such as Diabetic Retinopathy (DR) has been found. Other successful applications include classification of optical coherence tomography (OCT) scans.
Screening for DR, combined with timely referral and treatment, is a universally accepted strategy for the prevention of blindness. DR screening can be performed by a variety of healthcare professionals, including ophthalmologists, optometrists, general practitioners, screening technicians and clinical photographers.
In the following diagram we can see an illustration of the application of a DL solution for DR screening using imaging, comparing existing clinical practice with a fully automated AI model (replacement) and a semi-automated AI model (triage). AI indicates artificial intelligence; DL, deep learning; DR, diabetic retinopathy.
Cloud based solutions
Cloud-based imaging for diabetic retinopathy, glaucoma or AMD screening would make eye examinations even more convenient and help with data training and research in the field, and AI would play an innovative and disruptive role in the distribution of drugs in remote areas of the globe with a low presence of ophthalmologists.
Assisting patients in remote locations requires the implementation of digital solutions in some rural regions. Poor internet coverage requires the development of artificial intelligence solutions located within the screening hardware to provide real-time results effectively offline that can be transmitted to patients by enabling decentralised care by uploading data to a cloud-based server.
DL is likely to have an impact on the practice of medicine and ophthalmology in the coming decades.In the next 10–20 years, it will dramatically change the relationship between doctor and patient because the AI will assist the optometrist but the patient interaction will always be a personalised affair.
In the future, techniques and algorithms will be developed that can assist doctors with diagnoses and predict the evolution of diseases in advance, being able to administer appropriate drug treatments to save eye health. I just hope that patients will put a lot of trust in these tools, especially if they interface with doctors at a distance.
Artificial intelligence and deep learning in ophthalmology
Artificial intelligence (AI) based on deep learning (DL) has sparked tremendous global interest in recent years. DL has…