Clinical Report: A machine learning approach for predicting 90-day mortality in HBV-ACLF

Overview

This study developed a machine learning framework integrating the CALLY index to predict 90-day mortality in patients with hepatitis B virus-related acute-on-chronic liver failure (HBV-ACLF). The LightGBM model demonstrated superior predictive performance compared to traditional models.

Background

Hepatitis B virus-related acute-on-chronic liver failure (HBV-ACLF) is associated with high short-term mortality and rapid clinical deterioration. Effective risk stratification is essential for timely intervention and resource allocation in critical care settings. Traditional prognostic scores often fail to capture the complex dynamics of this condition, necessitating innovative approaches such as machine learning.

Data Highlights

Key Findings

• The LightGBM model outperformed other machine learning algorithms in predicting 90-day mortality in HBV-ACLF patients.
• SHAP analysis identified INR as the strongest predictor of mortality, followed by the CALLY index, log-transformed TBIL, age, and creatinine.
• The CALLY-based model improved risk stratification compared to the MELD score, with a continuous net reclassification index of 0.6223.
• A user-friendly online tool was developed to facilitate clinical implementation of the mortality prediction model.
• The study included a cohort of 471 patients, with a training set of 329 and an internal validation set of 142.

Clinical Implications

The integration of the CALLY index into a machine learning framework provides a robust tool for predicting mortality in HBV-ACLF patients. This approach may enhance early risk stratification and inform clinical decision-making in intensive care settings.

Conclusion

The machine learning framework developed in this study offers a high-precision tool for predicting 90-day mortality in HBV-ACLF patients.

Related Resources & Content

1. Frontiers in Oncology, 2026 -- Machine learning prediction of 30-day all-cause mortality risk factors in HCC rupture
2. Frontiers in Medicine, 2026 -- Prediction models for mortality in patients with acute on chronic liver failure: systematic review and critical appraisal
3. Open Forum Infectious Diseases -- Creation and Assessment of a Machine Learning Algorithm for Identifying Individuals at High Risk for Hepatitis C Infection
4. Guidelines for the diagnosis and treatment of acute-on-chronic liver failure (2025) - PMC
5. Effects of different therapeutic methods on the 90-day prognosis of patients with HBV-ACLF: A systematic review and network meta-analysis - PMC
6. the asco post — Machine-Learning Model for HCC Risk Prediction May Outperform Current Methods
7. Guidelines for the diagnosis and treatment of acute-on-chronic liver failure (2025) - PMC
8. Effects of different therapeutic methods on the 90-day prognosis of patients with HBV-ACLF: A systematic review and network meta-analysis - PMC
9. Machine Learning Prediction of 90‐Day Mortality in HBV‐Related ACLF Using Olink‐Derived Inflammatory Protein Signatures - Zhang - 2025 - Liver International - Wiley Online Library