Prediction of myopia onset and shift in premyopic school-aged children - a machine learning-based algorithm

Q: What is the main purpose of this study?

To examine longitudinal ocular parameter changes in premyopic children and develop machine learning models that predict myopia onset and myopia shift within 1 year.

Q: How were the participants followed and what measurements were taken?

A prospective cohort enrolled 320 premyopic children aged 6–12 years, with measurements at baseline and every 6 months for 12 months. Outcomes included cycloplegic spherical equivalent, axial length, corneal curvature, and subfoveal choroidal thickness, among others.

Q: Which factors were most important for predicting myopia onset?

In the optimal model, baseline spherical equivalent was the most significant predictor, followed by parental myopia, subfoveal choroidal thickness, and age.

Aug 4, 2026

Upload by Genevieve

Longitudinal changes in ocular parameters in premyopic children were investigated to support early identification of myopia risk. A prospective cohort enrolled 320 children aged 6–12 years and repeatedly measured uncorrected visual acuity, cycloplegic spherical equivalent, axial length, corneal curvature, and subfoveal choroidal thickness at baseline and every 6 months for 12 months. Machine learning models predicted 1-year myopia onset and myopia shift and used SHAP for interpretation, achieving high discrimination (AUC-ROC 0.963).

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Introduction

Premyopia concept and definition

Epidemiology and clinical significance

Rationale for predicting onset and shift

Methods

Study design and participants

Measurements and predictive variables

Machine learning approach and SHAP interpretation

Results

Longitudinal ocular parameter changes

Model performance for onset prediction

Conclusion

OPEN ACCESS EDITED BY Gábor Jandó, University of Pécs, Hungary REVIEWED BY Bhim Bahadur Rai, Australian National University, Australia Brian Vohnsen, University College Dublin, Ireland *CORRESPONDENCE Qi Zhao [zhaoqidmu@126.com](zhaoqidmu@126.com); [zhaoqi0219@126.com](zhaoqi0219@126.com)[ ](zhaoqi0219@126.com)RECEIVED 13 June 2025 ACCEPTED 24 October 2025 PUBLISHED 17 November 2025 CITATION Gao M, Hou Y, Lu Y, Shi Z and Zhao Q (2025) Prediction of myopia onset and shift in premyopic school-aged children: a machine learning-based algorithm. Front. Med. 12:1646277. doi: 10.3389/fmed.2025.1646277 COPYRIGHT © 2025 Gao, Hou, Lu, Shi and Zhao. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. TYPE Original Research PUBLISHED 17 November 2025 DOI 10.3389/fmed.2025.1646277 Prediction of myopia onset and shift in premyopic school-aged children: a machine learning-based algorithm Mingjun Gao, Yanhua Hou, Yutong Lu, Zhanhua Shi and Qi Zhao* Department of Ophthalmology, The Second Affiliated Hospital of Dalian Medical University, Dalian, China Purpose: This study aimed to investigate longitudinal changes in ocular parameters and develop a machine learning-based model for predicting myopia onset and shift within 1 year in school-aged premyopic children. Methods: This prospective cohort study enrolled 320 premyopic children aged 6–12 years from the Ophthalmology Clinic of The Second Affiliated Hospital of Dalian Medical University. Uncorrected visual acuity (logMAR), cycloplegic spherical equivalent (SE), axial length (AL), average corneal curvature (CC), and subfoveal choroidal thickness (SFCT) were measured at baseline and 6-month intervals for 12 months. Premyopia was defined as -0.50 D < SE ≤ + 0.75 D. A multivariable analysis evaluated predictive factors including age, gender, parental myopia, baseline SE, AL, CC, axial length/corneal radius (AL/CR), and SFCT. Machine learning algorithms were used to predict 1-year myopia onset and myopia shift, along with Shapley Additive exPlanations (SHAP) interpretation. Results: Among 284 participants (88 . 8% retention rate), 141 children (49.3%) developed myopia. The cohort exhibited an annual SE progression of −0.695 ± 0.222 D and AL elongation of 0.356 ± 0.122 mm. The AL/CR increased from 2.986 ± 0.061 to 3.029 ± 0.072 (p < 0.001), while SFCT demonstrated a significant reduction of 21.535 ± 9.731 μm (p < 0.001) . The optimal model achieved an AUC-ROC of 0.963 (95% CI: 0.930–0.997) for myopia onset prediction, with baseline SE emerging as the most significant predictor, followed by parental myopia, SFCT, and age. Meanwhile, our algorithm also achieved clinically acceptable 1-year predictions of SE. Conclusion: Premyopic children exhibited accelerated myopic progression. Our machine learning-based predictive models showed promising performance for myopia onset and myopia shift, providing clinically valuable risk stratification for targeted prevention strategies. KEYWORDS premyopia, myopic progression, subfoveal choroidal thickness, machine learning, prediction model Frontiers in Medicine 01 [frontiersin.org](frontiersin.org) 1 Introduction Myopia has become a major global public health issue, especially in East Asia. Among Chinese children and adolescents, its prevalence shows an annual increase, increasing from 55.5% in 2010 to 60.1% in 2019, with the peak age of onset dropping from 12 years old in 2010 to 7 years old in 2019 ( 1). An earlier onset of myopia is associated with a higher risk of developing more severe myopia in adulthood. The r

FAQs

What is the main purpose of this study?

To examine longitudinal ocular parameter changes in premyopic children and develop machine learning models that predict myopia onset and myopia shift within 1 year.

How were the participants followed and what measurements were taken?

A prospective cohort enrolled 320 premyopic children aged 6–12 years, with measurements at baseline and every 6 months for 12 months. Outcomes included cycloplegic spherical equivalent, axial length, corneal curvature, and subfoveal choroidal thickness, among others.

Which factors were most important for predicting myopia onset?

In the optimal model, baseline spherical equivalent was the most significant predictor, followed by parental myopia, subfoveal choroidal thickness, and age.

Aug 4, 2026

Upload by Genevieve