●        The high performance AI assistive tool may help to identify young children who are at risk of developing high myopia later on during their teenage years, which may allow for appropriate myopia control intervention.
●        Ultimately, this may reduce the incidence of high myopia among Singaporean children.
 
Singapore, 8 February 2023 – The Singapore National Eye Centre’s (SNEC) Myopia Centre and Singapore Eye Research Institute (SERI) have developed a new artificial intelligence (AI) tool that can predict a child’s risk for developing high myopia (above -5.00 dioptres) during teenage years. Researchers behind the AI tool hope that the innovation will assist clinicians in identifying high risk children who may need early intervention. This would aid in the long-term goal of reducing the incidence of high myopia among Singaporean children, thereby preventing vision threatening complications related to high myopia during adulthood.
 
New AI tool could address unmet needs in myopia management
As the prevalence of myopia in Singapore and East Asia continues to rise, efforts to focus on children or individuals at risk of developing high myopia may be even more crucial, as such patients may benefit from timely intervention. The new AI tool has been developed with this in mind.
 
This first-of-its-kind tool employs deep learning algorithms that process a child’s (six to 12 years old) retinal image with clinical data (age, race, gender, refractive error of eyes and distance between the cornea and retina), and instantly predicts the likelihood of them developing high myopia later on in their teenage years (11 to 17 years old). At present, the tool is able to predict with at least 90% accuracy. Results of this study, conducted among 998 children aged six to 12 years old was published in Nature’s open-access journal, npj Digital Medicine in January 2023.
 

Myopia and high myopia in Singapore
At SNEC’s Myopia Centre, the number of patient visits has been steadily increasing over the last few years. Last year, there were over 6,000 patient visits, while 2021 saw 5,758 myopic patients, an increase from 3,206 in 2020 and 1,728 in 2019. It is estimated that 30% of these myopic patients have high myopia.
 
“More than 60% of children in Singapore become myopic by Primary 6, and it is estimated that more than 80% of young adults are myopic. Many often think that myopia is just part of growing up and children inevitably develop myopia,” said Professor Saw Seang Mei, Co-head of SERI’s Myopia Research Group.
 
“But while myopia is widespread in Singapore, high myopia is associated with potentially sight-threatening eye diseases later in life. These include glaucoma, cataract, and myopic maculopathy,” said Associate Professor Marcus Ang, Advisor, SNEC Myopia Centre and Principal Investigator of the AI tool.
 
In children, myopia tends to progress rapidly and increases the most in children aged five to 15, before stabilising at around 18 to 21 years old or older. Which means that the earlier a child develops myopia, the longer the runway for myopia progression, and the likelihood of developing high myopia increases. With the rates of myopia among young children increasing, especially during the Covid-19 pandemic, there are concerns that visual impairment from high myopia will become a significant public health problem in future.
 
Long-term goals
“The next phase of the study is to test the usability of the AI algorithm in a real world setting, and utilise techniques to explain what the AI system is thinking and seeing. This will help improve confidence in using this AI technology among primary eye care providers, patients and parents,” said Associate Professor Daniel Ting, Chief Data and Digital Officer at SNEC, and Head of the AI and Digital Innovations Research Group at SERI.
 
“This is an important step towards incorporating our technology as an assistive AI tool. We believe that there is great potential to reduce the incidence of high myopia in Singapore, which could thereby minimise the burden of visual impairment from myopia complications later in life,” said Assoc Prof Ang. If the validation phases are successful and the AI tool is feasible for implementation, there is also potential for it to be developed into a platform for community and primary eye care professionals. This would enable more children to be evaluated for risk of high myopia, and allow for appropriate referrals for early myopia management and control.