Best-Corrected Visual Acuity Quantitative Prediction for Cataract Patients: AI-Assisted Clinical Diagnostics...

Area of Science:

• Ophthalmology
• Biomedical Engineering
• Artificial Intelligence in Medicine

Background:

• Cataract surgery outcomes are influenced by various patient-specific factors.
• Accurate prediction of best-corrected visual acuity (BCVA) is crucial for managing patient expectations and surgical planning.
• Existing methods may not fully capture the complex interplay of risk factors affecting visual outcomes.

Purpose of the Study:

• To quantitatively predict BCVA in cataract patients using a previously developed inverse problem algorithm (IPA).
• To identify significant biological risk factors influencing BCVA prediction.
• To assess the feasibility of an AI-assisted approach for pre-operative BCVA estimation.

Main Methods:

• A semi-empirical formula normalized seven risk factors (age, BMI, MAP, IOP, HbA1c, LDL-C, gender) to a -1.0 to +1.0 range.
• An IPA, implemented in STATISTICA 7.0, utilized a 29-term nonlinear equation incorporating risk factor interactions.
• The model was trained on 632 cataract patients and validated on 160 patients, achieving R² = 0.909.

Main Results:

• The IPA achieved a high prediction variance (0.929) on the training set and strong correlation (R² = 0.909) on the verification set.
• Age, body mass index (BMI), and intraocular pressure (IOP) were identified as significant predictors of BCVA.
• Mean arterial pressure (MAP), HbA1c, LDL-C, and gender showed insignificant contributions to the BCVA prediction model.

Conclusions:

• The developed IPA provides a robust method for predicting individual BCVA in cataract patients.
• This AI-assisted approach can aid in clinical diagnosis by offering pre-operative visual acuity estimations.
• The study highlights the potential of computational methods in enhancing ophthalmological diagnostic capabilities.