How AI Can Help to Manage Juvenile-Onset Myopia

Share

Tweet

Share

Share

Email

July 1, 2025

By Marcus Ang, MBBS, MMED, MCI, FRCS, PhD

The incidence of juvenile-onset myopia is rising. This is exacerbated by reduced outdoor time and more screen time indoors—not only in Asia, but worldwide.¹ As myopia is irreversible, myopia control interventions are only effective if they are instituted in a timely and appropriate manner.² Myopia management strategies include myopia prevention, early detection of myopia progression and personalized myopia control.³ Environmental and behavioral modifications, such as increasing outdoor time and reducing indoor near work or screen time, remain essential for myopia prevention. Myopia control interventions such as atropine eye drops, specialized contact lenses and myopia control spectacles, all shown to be safe and effective, should also be introduced to slow myopia progression.⁴ 

The Role of AI

Recently, artificial intelligence (AI) has emerged as a promising adjunctive tool to aid clinicians in the management of myopia.^5,6 These algorithms could be used to support various aspects of myopia management, including detection of myopia or its complications; prediction of future myopia progression; and treatment response—allowing for personalized myopia management.

For example, multi-modal AI algorithms with more than 80-90% predictive performance have been described that rely on a single baseline fundus image to identify which children developed high myopia five years later on in life, in a multi-ethnic population aged between 6 to 12 years old.⁷ Similar AI models have been replicated and shown to have potential for implementation into school-based screening programs to identify at-risk children.⁸

Emerging AI Technologies

There are also emerging novel AI technologies to support myopia management in children. These include: 

• Explainable AI (XAI): AI systems designed so humans can see how and why they make decisions—think of it as “opening the black box” of a model to build trust and spot errors.
• Automated machine learning: Software that handles the tedious parts of building machine learning models (like choosing algorithms, tuning settings and testing performance) so you don’t have to be an expert to get good results.
• Federated learning (FL): A way to train a single AI model across many devices (phones, hospitals, etc.) without moving the private data off of each device—only model updates (not raw data) are shared and aggregated.
• Synthetic AI technology, such as generative adversarial networks: AI techniques for creating realistic—but entirely artificial—data or content (images, text, sensor readings, etc.). It’s used to train, test or simulate systems when real data is scarce or sensitive.
• Natural language processing (NLP) utilizing large-language models (LLMs):⁹ The branch of AI that lets computers understand, interpret and generate human language—everything from chatbots and translation tools to sentiment analysis and voice assistants.

Research is Needed

While exciting, these AI technologies are still developing in the research realm and have not been clinically deployed. Clinicians still have to decide and advise patients on the appropriate myopia management strategy and which interventions to recommend.

There are also widespread challenges and barriers to implementation—infrastructural, regulatory and data security issues, as well as difficulty with health systems integration. Public awareness of myopia and its potential complications are inadequate in many societies around East Asia, which may also hinder acceptance of these novel digital technologies. Nonetheless, novel AI modalities, including LLM and FL, could play an important role in the future by overcoming these barriers and improving digital literacy or myopia awareness. There is a now global recognition that AI could connect myopia care providers, facilitate collaborative efforts across sectors, and improve patient access to evidence-based myopia management.¹⁰

Conclusion

In summary, AI has the potential to support myopia management by providing more consistent, rapid and scalable evaluation, which could lead to earlier detection and intervention. In the future, AI could help to identify children at-risk of developing high myopia.

Finally, AI could help optimize treatment strategies that enhance personalization of therapies, ensuring the most effective intervention is chosen for each individual. With the further development and implementation of AI into health care systems, these various adjunctive clinical tools could play a major role in shaping myopia management in the future. 

 

Associate Professor Dr. Marcus Ang, MBBS, MMED, MCI, FRCS, PhD is Senior Consultant Ophthalmologist & Head of the Cornea and External Eye Disease Service; & Head of Refractive Service, Singapore National Eye Center (SNEC). He also serves as Advisor at the SNEC Myopia Center.

 

References

 

1. Eppenberger LS, Davis A, Resnikoff S, et al. Key strategies to reduce the global burden of myopia: consensus from the international myopia summit. Br J Ophthalmol 2025;109(5):535-542.
2. Jonas JB, Ang M, Cho P, et al. IMI Prevention of Myopia and Its Progression. Invest Ophthalmol Vis Sci 2021;62(5):6.
3. Ang M, Flanagan JL, Wong CW, et al. Review: Myopia control strategies recommendations from the 2018 WHO/IAPB/BHVI Meeting on Myopia. Br J Ophthalmol 2020;104(11):1482-1487.
4. Eppenberger LS, Grzybowski A, Schmetterer L, Ang M. Myopia Control: Are We Ready for an Evidence Based Approach? Ophthalmol Ther 2024;13(6):1453-1477.
5. Foo LL, Ang M, Wong CW, et al. Is artificial intelligence a solution to the myopia pandemic? Br J Ophthalmol 2021;105(6):741-744.
6. Li Y, Foo LL, Wong CW, et al. Pathologic myopia: advances in imaging and the potential role of artificial intelligence. Br J Ophthalmol 2023;107(5):600-606.
7. Foo LL, Ng WY, Lim GYS, Tan TE, Ang M, Ting DSW. Artificial intelligence in myopia: current and future trends. Curr Opin Ophthalmol 2021;32(5):413-424.
8. Qi Z, Li T, Chen J, et al. A deep learning system for myopia onset prediction and intervention effectiveness evaluation in children. NPJ Digit Med 2024;7(1):206.
9. Ng Yin Ling C, Zhu X, Ang M. Artificial intelligence in myopia in children: current trends and future directions. Curr Opin Ophthalmol 2024;35(6):463-471.
10. Wong CW, Foo LL, Morjaria P, et al. Highlights from the 2019 International Myopia Summit on ‘controversies in myopia’. Br J Ophthalmol 2021;105(9):1196-1202.