Identifying liver cirrhosis in patients with chronic hepatitis B: an interpretable machine learning algorithm based...

Area of Science:

• Hepatology
• Medical Informatics
• Diagnostic Imaging

Background:

• Chronic hepatitis B (CHB) is a leading cause of liver cirrhosis (LC), a condition with poor prognosis.
• Early and accurate diagnosis of LC in CHB patients is critical for timely intervention.

Purpose of the Study:

• To enhance the diagnostic accuracy of liver cirrhosis in CHB patients.
• To integrate liver stiffness measurement (LSM) with traditional indicators using machine learning.

Main Methods:

• Development and validation of machine learning models, including random forest (RF), logistic regression, k-nearest neighbors, artificial neural network, support vector machine, and eXtreme Gradient Boosting.
• Feature selection using LASSO and RF-RFE.
• Performance evaluation via ROC curves, calibration curves, and decision curve analysis.
• SHAP analysis for model interpretability.

Main Results:

• The RF model demonstrated high accuracy (>0.80) and AUC (>0.80) in internal and external validation sets.
• LSM was identified as the most significant contributor to the model's predictive power.
• The integrated model showed strong discriminative power, even when using LSM or traditional indicators alone.

Conclusions:

• Machine learning models, especially RF, effectively identify liver cirrhosis in CHB patients.
• Integrating LSM with traditional indicators significantly enhances diagnostic performance for LC in CHB.