Today's Edition
SARA ALZAABI (ABU DHABI)
Offering advantages in speed, precision, and scalability, artificial intelligence tools are becoming a necessity in detecting retinal diseases, according to research from Khalifa University of Science and Technology.
AI technologies can now identify early signs of retinal conditions, enabling timely interventions that help prevent visual impairments and even vision loss, explained Prof. Naoufel Werghi, Associate Professor in the Department of Electrical Engineering and Computer Sciences at Khalifa University (KU).
Prof. Werghi was among the university’s experts who conducted a comprehensive review of existing AI models for retinal disease screening. Other collaborators included with Dr. Bilal Hassan, now at New York University Abu Dhabi, and Dr. Taimur Hassan from Abu Dhabi University.
AI-powered tools plug certain gaps in diagnostics as they provide consistent, data-driven assessments, reducing the risk of human error or fatigue, Prof. Werghi told Aletihad.
“As humans, we are limited in our cognitive capacity. Clinicians can only process a limited amount of data at any time, whereas AI platforms can manage, integrate, and analyse hundreds or thousands of variables simultaneously,” he said.
Machine learning models achieve this by studying vast datasets made up of high-resolution images captured through techniques like fundus photography, OCT, and fluorescein angiography.
“These models — often based on deep learning — analyse pixel-level features to spot early signs of disease,” Prof. Werghi said.
Undiagnosed Retinal Diseases
AI innovations can become a game-changer in the management of retinal diseases, which are considered the leading cause of blindness globally. Out of the millions of cases of visual impairment worldwide, about half are preventable, according to the World Health Organization.
Among the most common diseases that AI tools can detect are diabetic retinopathy, macular degeneration, and glaucoma, Prof. Werghi said.
“If left untreated, these diseases can lead to severe visual impairment and even blindness,” he added.
Considering how irreversible vision loss can be, early diagnosis is crucial for these conditions, he said.
“Early detection allows timely interventions to halt or slow disease progression. In many cases, treatments are far more effective when retinal conditions are caught in their early stages.”
How the Tech Works
Explaining how the AI screening tools work, Prof. Werghi said: “Machine learning models are used in screening for retinal diseases by analysing medical imaging data, such as fundus photographs or optical coherence tomography (OCT) scans.”
“These models are trained by being exposed to a large number of labelled images-those reviewed by clinicians who identify the disease or mark the location of pathology. The machine adjusts its parameters over time until it can perform the required task with high accuracy.”
Once trained, the models are deployed to support clinicians by automatically identifying abnormalities, classifying disease severity, and highlighting regions of concern.
“Advanced models can quantify retinal pathology and assess treatment response-making diagnosis more precise and data-driven rather than subjective,” Prof. Werghi said.
Transforming Preventive Care
Machine learning models have demonstrated high accuracy —often comparable to or exceeding traditional methods, Prof. Werghi said.
“They also offer advantages in speed, quantification, scalability, and the ability to integrate data from multiple sources.”
Prof. Werghi believes that with AI’s rapid advancement, its role in health screenings is set to expand significantly.
“Over the next five to 10 years, we will see widespread adoption of AI systems for multi-disease detection, more portable diagnostic tools even used in pharmacies or homes, and tighter integration with telemedicine and wearable devices like smart glasses or contact lenses,” he said.
