Development of a Prediction Model and Corresponding Scoring Table for Postherpetic Neuralgia Using Six Machine
Zheng Lin1,2, Lu-Yan Yu1,2, Si-Yi Pan1,2
1First Clinical Medical College, Zhejiang Chinese Medical University, No. 548 Binwen Road, Binjiang District, Hangzhou, 310006, Zhejiang, China.
Pain and Therapy
|June 4, 2024
Summary
This study developed a predictive model and scoring table to identify patients at risk of developing postherpetic neuralgia (PHN) after shingles. The gradient boosting model achieved high accuracy, aiding early clinical intervention for PHN.
Area of Science:
- Neurology
- Medical Informatics
- Pain Management
Background:
- Postherpetic neuralgia (PHN) is a debilitating complication of herpes zoster (shingles) that significantly impairs patient quality of life.
- Early intervention for acute herpetic neuralgia pain is crucial for potentially reducing PHN incidence or severity.
- Identifying patients at high risk for PHN is essential for timely and targeted clinical management.
Purpose of the Study:
- To develop and validate a predictive model for identifying patients at risk of developing PHN.
- To create a scoring table based on logistic regression for predicting PHN risk.
- To facilitate informed clinical decision-making in managing acute herpetic neuralgia.
Main Methods:
- Retrospective analysis of 524 hospitalized herpes zoster patients.
- Feature selection using Least Absolute Shrinkage and Selection Operator (LASSO) regression.
- Development and comparison of six machine learning models (SVM, logistic regression, random forest, k-NN, gradient boosting, neural network).
- Validation of model performance and creation of a nomogram-based predictive scoring table.
Main Results:
- Eight key predictors for PHN were identified: age, Numerical Rating Scale (NRS) pain score, treatment initiation time, rash recovery time, history of malignant tumor, history of diabetes, varicella-zoster virus IgM, and serum nerve-specific enolase.
- The gradient boosting model demonstrated superior performance with an Area Under the Curve (AUC) of 0.931 and accuracy of 0.886 in the test set.
- The developed predictive scoring table achieved an AUC of 0.820 and accuracy of 0.790.
Conclusions:
- A robust methodology using machine learning was established for predicting PHN development in shingles patients.
- The gradient boosting model and the logistic regression-based scoring table offer effective tools for risk stratification.
- These tools can support clinicians in making timely and personalized treatment decisions for patients with acute herpetic neuralgia.
