Varun Ranganathan, MCOptom

Clinical Optometrist

An OCULAR Interface Exclusive

Synopsis:

With myopia reaching alarming levels globally, early detection and effective management are more critical than ever. This article explores how artificial intelligence (AI) is in the forefront of the fight against myopia by enabling earlier diagnosis, predictive risk assessment, and personalised intervention strategies. The article also highlights how AI helps monitor progression, improve treatment adherence, and support research into the environmental and genetic factors contributing to myopia.

Introduction:

In 2010, it was estimated that myopia is the most common form of uncorrected refractive error causing visual impairment affecting 108 million people globally. But this has changed drastically, and the new projection is alarmingly high and by 2050 more than 50% and 10% of the world will have myopia or high myopia respectively. Higher amounts of myopia have the potential to cause vision impairment by myopic macular degeneration or its comorbidities, cataract, retinal detachment, and glaucoma, ¹ the risk of which increase with any increase in myopia. The combination of vision impairment from uncorrected myopia and irreversible vision loss from myopia-related complications make accurate global estimates of the prevalence and temporal trends critical for planning care and services.

The main drawbacks in tackling myopia are the lack of reliable risk prediction model for high and pathologic myopia, as well as the individual differences in the progression of myopia can make it difficult to provide timely intervention. We will discuss the ability of AI in predictive analysis making early diagnosis and management.

How AI can help in myopia management:

In clinical management of myopia, it is often necessary to evaluate multi-faceted data including myopia-related risk factors (e.g., near work time, outdoor activity time, genetics, race, gender, etc.), best-corrected visual acuity, refraction, axial length, and some ocular metrics (e.g., intraocular pressure, ocular surface conditions). ² Given the size of the data and the complexity of myopia, it is extremely difficult to perform manual analysis. As myopia progresses further, the axial length (AL) increases, the optic disc begins to tilt and twist, and irreversible retinal chorioretinopathy may develop.  There are some of the fundus lesions associated with high myopia. Among these patients, about 3.1% eventually develop different types of myopic maculopathy with a characteristic set of pathological changes. ³

The detection of fundus lesions and myopia-related complications in high myopia is an important need, for which deep learning methods such as Convolutional Neural Networks (CNNs) have high accuracy. ⁴ Compared to the manual, deep learning methods take only a few hours to a few days in the training phase of the model and can produce instant results when interpreting images. In addition to detection, classification of these myopic lesions is equally important in pathological myopia. New methods have effectively classified lesions in patients with high myopia based on color fundus images. The results showed that the classification accuracy for each lesion specified in meta-analysis of pathologic myopia (META-PM), a widely accepted classification standard for pathologic myopia, was comparable to that of experts, reaching 97.03%–99.41%. ⁵

AI can provide new ideas for morphological changes in myopic eyes. On tasks that ophthalmologists cannot perform (e.g., predicting refraction based on fundus images), deep learning methods can be done on fundus photos, ultra-wide fundus photos, or OCT images. ⁶ In many mammalian models, choroidal thickness can rapidly change in both directions when images are focused anteriorly or posteriorly to the retina before axial changes. ⁷ AI methods can effectively predict choroidal thickness based on fundus images rather than OCT images, which facilitates the assessment of early pathological changes in highly myopic eyes and guides early diagnosis and treatment. ⁸

Conclusion:

The involvement of an eye care professional is still preferred over a complete AI-based screening and diagnosis and Explainable AI (XAI) can be a potential solution in gaining patients’ trust. Entire human replacement AI algorithms can be a preferred solution only in automated tasks, but this step can save time and resources, improving quality of treatment.

References:

1. Spaide, Ohno-Matsui, A. Yannuzzi., Pathologic Myopia https://doi.org/10.1007/978-1-4614-8338-0.
2. Chen, Y. X., Liao, C. M., Tan, Z., and He, M. G. (2021). Who needs myopia control? Int. J. Ophthalmol. 14 (9), 1297–1301. doi:10.18240/ijo.2021.09.01.
3. Chang, L., Pan, C. W., Ohno-Matsui, K., Lin, X., Cheung, G. C., Gazzard, G., et al. (2013). Myopia-related fundus changes in Singapore adults with high myopia. Am. J. Ophthalmol. 155 (6), 991–999. doi:10.1016/j.ajo.2013.01.016.
4. Shao, L., Zhang, Q. L., Long, T. F., Dong, L., Zhang, C., Da Zhou, W., et al. (2021). Quantitative assessment of fundus tessellated density and associated factors in fundus images using artificial intelligence. Transl. Vis. Sci. Technol. 10 (9), 23. doi:10.1167/tvst.10.9.23.
5. Lu, L., Ren, P., Tang, X., Yang, M., Yuan, M., Yu, W., et al. (2021). AI-model for identifying pathologic myopia based on deep learning algorithms of myopic maculopathy classification and “plus” lesion detection in fundus images. Front. Cell Dev. Biol. 9, 719262. doi:10.3389/fcell.2021.719262.
6. Artificial intelligence technology for myopia challenges: A review Frontiers in Cell and Developmental Biology Volume 11 – 2023 https://www.frontiersin.org/journals/cell-and-developmental-biology/articles/10.3389/fcell.2023.1124005 DOI=10.3389/fcell.2023.1124005
7. Read, S. A., Fuss, J. A., Vincent, S. J., Collins, M. J., and Alonso-Caneiro, D. (2019). Choroidal changes in human myopia: Insights from optical coherence tomography imaging. Clin. Exp. Optom. 102 (3), 270–285. doi:10.1111/cxo.12862.
8. Sun, D., Du, Y., Chen, Q., Ye, L., Chen, H., Li, M., et al. (2021). Imaging features by machine learning for quantification of optic disc changes and impact on choroidal thickness in young myopic patients. Front. Med. (Lausanne). 8, 657566. doi:10.3389/fmed.2021.657566.