Machine learning sees into the future to prevent sight loss in humans

(Image Credit: AdobeStock/knssr)

As machine learning has been refined, it has been used to predict the outcomes of many health conditions. A team of researchers from Japan has found a way to predict whether people with severe shortsightedness will have good or bad vision in the future.

The researchers from the Tokyo Medical and Dental University (TMDU) developed a machine-learning model that could effectively predict—and visualize—the risk of visual impairment over the long term. The results of their study were published in JAMA Ophthalmology.¹

Ultimately, contacts, glasses, or surgery can be used to correct the vision of patients diagnosed with myopia, but having high myopia is not just inconvenient; half of the time it leads to a condition called pathologic myopia, and complications from pathologic myopia are a leading causes of blindness.²

“We know that machine-learning algorithms work well on tasks such as identifying changes and complications in myopia,” Yining Wang, MD, lead author of the study, said in a TMDU news release. “In this study, we wanted to investigate something different, namely how good these algorithms are at long-term predictions.”

Researchers conducted a cohort study and looked at the visual acuity of 967 Japanese patients at TDMU’s Advanced Clinical Center for Myopia after 3 and 5 years had passed. With this information, they constructed a dataset from 34 variables that are often collected during ophthalmic examinations, such as age, current visual acuity, and the diameter of the cornea. The researchers then reviewed several p machine-learning models such as random forests and support vector machines. Of these models, the logistic regression-based model performed the best at predicting visual impairment at 5 years.¹

The researchers found that being able to predict outcomes proved to be only one part of the equation.

Kyoko Ohno-Matsui, MD, PhD, pointed out in the TMDU news release it is important to present the model’s output in a way that is convenient for making medical decisions and easily understood by patients.

To do this, the researchers used a nomogram to visualize the classification model. The researchers noted that each variable is assigned a line with a length that indicates how important it is for predicting visual acuity. These lengths can be converted into points that can be added up to obtain a final score explaining the risk of visual impairment in future.

Ultimately, patients who end up losing their vision often suffer physically and financially as a result of their loss of independence. Moreover, the drop in global productivity resulting from severe visual impairment was estimated to be $94.5 billion in 2019.² While the research team’s model still needs to be tested on a wider population, the results of the study have shown that machine-learning models have good potential to help address this increasingly important public health concern, which will benefit both individuals and society as a whole.

Reference:

1. Wang Y, Du R, Xie S, et al. Machine Learning Models for Predicting Long-Term Visual Acuity in Highly Myopic Eyes. JAMA Ophthalmol. Published online October 26, 2023. Accessed December 14, 2023. doi:10.1001/jamaophthalmol.2023.4786

2. Tokyo Medical and Dental University. Machine learning sees into the future to prevent sight loss in humans. Eurekalert. Published December 13, 2023. Accessed December 14, 2023. https://www.eurekalert.org/news-releases/1011015