A machine learning model for 90-day mortality prediction in hepatitis B virus-related acute-on-chronic liver failure: the pivotal role of CALLY index - Scorecard - MDSpire
Advertisement
A machine learning model for 90-day mortality prediction in hepatitis B virus-related acute-on-chronic liver failure: the pivotal role of CALLY index
Clinical Scorecard: A machine learning approach for predicting 90-day mortality in acute-on-chronic liver failure associated with hepatitis B virus: emphasizing the significance of the CALLY index
At a Glance
Category
Detail
Condition
Hepatitis B virus-related acute-on-chronic liver failure (HBV-ACLF)
Key Mechanisms
Integration of the C-reactive protein-albumin-lymphocyte (CALLY) index for mortality prediction
Target Population
Patients aged 18–80 with HBV-ACLF
Care Setting
ICU and critical care settings
Key Highlights
90-day mortality rate in HBV-ACLF is high and requires early risk stratification.
The CALLY index improves risk stratification compared to traditional models like MELD.
LightGBM model demonstrated strong discriminative performance with AUCs up to 0.940.
SHAP analysis identified INR as the strongest predictor of mortality.
A user-friendly online tool was developed for clinical implementation.
Guideline-Based Recommendations
Diagnosis
Diagnosis of HBV-ACLF according to the 2018 Chinese liver failure criteria.
Management
Early risk stratification and allocation of ICU resources based on predictive models.
Monitoring & Follow-up
Continuous assessment of key laboratory parameters including CALLY index components.
Risks
High short-term mortality associated with rapid liver deterioration in HBV-ACLF patients.
Patient & Prescribing Data
Patients hospitalized with HBV-ACLF from 2015 to 2025.
Integration of machine learning models for personalized interventions and organ support.
Clinical Best Practices
Utilize the CALLY index for improved risk assessment in HBV-ACLF.
Implement machine learning frameworks for predicting mortality in critical care settings.
Ensure timely decision-making based on early risk stratification.
