Research on Establishing Corneal Edema after Phacoemulsification Prediction Model Based on Variable Selection with...

Area of Science:

• Ophthalmology
• Medical Statistics
• Predictive Modeling

Background:

• Corneal edema (CE) is a significant factor impacting phacoemulsification outcomes.
• Accurate prediction of CE post-phacoemulsification is crucial for surgical planning and patient management.

Purpose of the Study:

• To develop and validate a nomogram for predicting corneal edema after phacoemulsification.
• To evaluate the utility of copula entropy in enhancing nomogram-based prediction models.

Main Methods:

• Multivariate logistic regression was employed to construct an initial CE nomogram using data from 178 patients.
• Copula entropy was utilized for variable selection to refine the prediction model.
• Model performance was assessed using predictive accuracy, Area Under the Receiver Operating Characteristic Curve (AUC), and Decision Curve Analysis (DCA).

Main Results:

• The refined Copula nomogram, utilizing variables like cumulative dissipated energy (CDE) and best corrected visual acuity (BCVA), demonstrated comparable predictive accuracy (0.9098) to the initial CE nomogram (0.9039).
• No significant difference was observed in AUCs between the two models (0.9512 vs. 0.9637), indicating robust predictive performance.
• DCA suggested that the Copula nomogram possesses practical clinical applicability.

Conclusions:

• A high-performing nomogram was developed for predicting CE following phacoemulsification.
• Copula entropy offers a valuable method for improving nomogram models in clinical prediction.