Predicting Cycloplegic Spherical Equivalent Refraction Among Children and Adolescents Using Non-cycloplegic Data and
Keke Liu1,2, Ran Qin2, Huijuan Luo2
1School of Public Health, Capital Medical University, Beijing, China.
Insights
Machine learning accurately predicts cycloplegic refractive error in children using non-cycloplegic measurements and ocular biometrics. This offers a practical alternative for large-scale vision surveys.
Area of Science:
- Ophthalmology
- Biomedical Engineering
- Data Science
Background:
- Cycloplegic refraction is the standard for pediatric refractive error assessment.
- Logistical challenges limit its use in large-scale population surveys.
Purpose of the Study:
- To develop and validate machine learning models for estimating cycloplegic refractive error in children using non-cycloplegic data and ocular biometrics.
- To provide a feasible alternative for refractive error surveillance in pediatric populations.
Main Methods:
- Analysis of data from a nationwide ocular health survey (2020-2024) in China.
- Participants aged 5-18 years underwent non-cycloplegic and cycloplegic autorefraction, axial length (AL), and corneal radius (CR) measurements.
- Random forest and XGBoost models were trained to predict cycloplegic spherical equivalent (SE) using non-cycloplegic SE, uncorrected visual acuity (UCVA), and biometric parameters (AL, CR, AL/CR ratio).
Main Results:
- Both random forest (R²=0.88, RMSE=0.55 D) and XGBoost (R²=0.89, RMSE=0.54 D) models demonstrated high predictive accuracy.
- Key predictors included non-cycloplegic SE, AL/CR ratio, AL, and UCVA.
- Predicted SE showed strong agreement with cycloplegic SE, with minimal bias.
Conclusions:
- Machine learning models integrating non-cycloplegic SE and ocular biometrics can accurately estimate cycloplegic SE in children and adolescents.
- This approach offers a practical and efficient method for large-scale refractive error surveillance, overcoming limitations of traditional cycloplegic refraction.
Introduction:
Cycloplegic refraction is the gold standard for assessing refractive error in children. However, logistical constraints hinder its implementation in large-scale surveys.
Methods:
Data obtained from a nationwide ocular health survey conducted in ten provincial-level administrative divisions in China were analyzed (2020-2024). Participants aged 5-18 years underwent standardized non-cycloplegic and cycloplegic autorefraction, axial length (AL), corneal radius (CR), and AL/CR measurements. Random forest and XGBoost models were trained to predict the cycloplegic spherical equivalent (SE) using non-cycloplegic SE, uncorrected visual acuity (UCVA), and biometric parameters. Performance was evaluated using R2, root mean square error (RMSE), and Bland-Altman analysis.
Results:
Both models exhibited strong predictive performance. In the test set, random forest achieved R2=0.88 and RMSE=0.55 diopter (D), whereas XGBoost achieved R2=0.89 and RMSE=0.54 D. Non-cycloplegic SE, AL/CR ratio, AL, and UCVA were consistently the top predictors. The predicted SE exhibited strong agreement with the cycloplegic SE, with minimal residual bias.
Conclusion:
Machine learning models incorporating noncycloplegic SE and ocular biometrics accurately estimate cycloplegic SE in children and adolescents, providing a practical alternative for large-scale refractive-error surveillance when cycloplegia is impractical.
More Related Videos
13:12Partial Optic Nerve Transection in Rats: A Model Established with a New Operative Approach to Assess Secondary Degeneration of Retinal Ganglion Cells
Published on: October 15, 2017
08:42Monocular Visual Deprivation and Ocular Dominance Plasticity Measurement in the Mouse Primary Visual Cortex
Published on: February 8, 2020