A recent study has leveraged machine learning to develop models for accurate prediction of the long-term visual acuity (VA) for highly myopic eyes based on clinical and imaging information.
High myopia is said to be a global concern due to its escalating prevalence and the potential risk of severe visual impairment caused by pathologic myopia. Australia is expected to have 4.1 million high myopes by 2050, unless myopia management is widely implemented, which is four times the number of high myopes in 2020 (1.1 million).
In the study, the researchers hypothesised that using artificial intelligence to estimate future VA could help clinicians to identify and monitor patients with a high risk of vision reduction in advance.
Published in JAMA Ophthalmolology, the retrospective cohort study recruited 1,616 eyes from a total of 967 patients with high myopia whose best-corrected (BC) VA at three and five years was known.
The researchers developed regression models to predict BCVA at three and five years, and a binary classification model to predict the risk of developing visual impairments at five years.
They found that support vector machines presented the best prediction of BCVA at three years and random forest at five years. Meanwhile, logistic regression presented the best prediction of visual impairment at five years.
“The findings suggest that machine learning models have the potential to be used for clinical assessments and future visual acuity monitoring in patients with high myopia,” the researchers said.
More reading
Western Australia study finds computer screens the real culprit in myopia
Don’t forget about myopia progression in adults
Outdoor light exposure – the first step in myopia management
