Forecasting Myopic Maculopathy Risk Over a Decade: Development and Validation of an Interpretable Machine Learning...

Area of Science:

• Ophthalmology
• Medical Imaging
• Machine Learning

Background:

• High myopia is a significant risk factor for myopic macular degeneration (MMD).
• Predicting MMD progression is crucial for early intervention and preventing vision loss.

Purpose of the Study:

• To develop and validate a prediction model for myopic macular degeneration (MMD) progression in patients with high myopia.
• To identify key clinical and imaging indicators associated with MMD progression.

Main Methods:

• A cohort of 660 high myopia patients was used for model development, with external validation on 212 participants.
• Thirty-four clinical variables were analyzed using sequential forward selection and five machine learning algorithms.
• The eXtreme Gradient Boosting algorithm was employed to forecast MMD progression risk over 10 years.

Main Results:

• Over 10.9 years, 20.2% of patients showed MMD progression.
• Key predictors included thinner subfoveal choroidal thickness, longer axial length, worse visual acuity, older age, female gender, and shallower anterior chamber depth.
• The model achieved an AUROC of 0.87 in the training cohort and 0.80 in the validation cohort.

Conclusions:

• Machine learning effectively predicts MMD progression using clinical and imaging data.
• This predictive tool can identify high-risk individuals for early intervention.
• Timely intervention can aid in vision protection for high myopia patients.