An ensemble-based approach for estimating personalized intraocular lens power
Salissou Moutari1, Jonathan E Moore2,3
1School of Mathematics and Physics, Queens University Belfast University Road, Belfast, BT7 1NN, Northern Ireland, UK. s.moutari@qub.ac.uk.
Scientific Reports
|November 26, 2021
Summary
A new ensemble regression model improves intraocular lens (IOL) power calculation by providing more accurate effective lens position (ELP) predictions. This approach enhances refractive outcomes in cataract surgery patients.
Area of Science:
- Ophthalmology
- Biomedical Engineering
- Computational Statistics
Background:
- Accurate intraocular lens (IOL) power calculation is crucial for successful cataract surgery and optimal patient refractive outcomes.
- Current IOL power calculation formulas rely on predicting the effective lens position (ELP), which is challenging and impacts accuracy.
- Limitations in ELP prediction are a primary source of refractive errors post-cataract surgery.
Purpose of the Study:
- To introduce a novel approach for personalized IOL power calculation using an ensemble of regression models for improved ELP prediction.
- To enhance the accuracy and robustness of IOL power calculations, thereby improving patient refractive outcomes.
- To evaluate the performance of the new formula against established modern IOL power calculation methods.
Main Methods:
- Development of a new IOL power calculation formula utilizing an ensemble of regression models to predict ELP.
- Rigorous performance assessment using cross-validation, comparing the devised formula against widely used methods (Haigis, Holladay I, Hoffer Q, SRK/T).
- Evaluation of prediction accuracy across various axial lengths and for different types of IOLs (monofocal and multifocal).
Main Results:
- The proposed ensemble approach demonstrated superior performance compared to standard formulas, showing lower mean absolute prediction errors.
- The new formula achieved higher prediction accuracy, with a greater percentage of eyes falling within ±0.5D and ±1D prediction ranges.
- The formula exhibited robustness, effectively managing variations in axial length, anterior chamber depth, and keratometry readings.
Conclusions:
- The novel ensemble regression approach offers a more accurate and robust method for personalized IOL power calculation.
- This improved ELP prediction technique has the potential to significantly enhance refractive outcomes for patients undergoing cataract surgery.
- The formula's robustness makes it adaptable to variations in biometric measurements, mitigating the impact of measurement inaccuracies.
