Development and Validation of a Risk Score for Liver Cirrhosis Prediction in Untreated and Treated Chronic Hepatitis B

Application of Interpretable Machine Learning Models to Predict the Risk Factors of HBV-Related Liver Cirrhosis in CHB Patients Based on Routine Clinical Data: A Retrospective Cohort Study

Chronic hepatitis B (CHB) infection represents a significant global public health issue, often leading to hepatitis B virus (HBV)-related liver cirrhosis (HBV-LC) with poor prognoses. Early identification of HBV-LC risk is essential for timely intervention. This study develops and compares nine machine learning (ML) models to predict HBV-LC risk in CHB patients using routine clinical and laboratory data. A retrospective analysis was conducted involving 777 CHB patients, with 50.45% (392/777) progressing to HBV-LC. Admission data consisted of 52 clinical and laboratory variables, with missing values addressed using multiple imputation. Feature selection utilized Least Absolute Shrinkage and Selection Operator (LASSO) regression and the Boruta algorithm, identifying 24 key variables. The evaluated ML models included XGBoost, logistic regression (LR), LightGBM, random forest (RF), AdaBoost, Gaussian naive Bayes (GNB), multilayer perceptron (MLP), support vector machine (SVM), and k-nearest neighbors (KNN). The data set was partitioned into an 80% training set (n = 621) and a 20% independent testing set (n = 156). Cross-validation (CV) facilitated hyperparameter tuning and internal validation of the optimal model. Performance metrics included the area under the receiver operating characteristic curve (AUC), Brier score, accuracy, sensitivity, specificity, and F1 score. The RF model demonstrated superior performance, with AUCs of 0.992 (training) and 0.907 (validation), while the reconstructed model achieved AUCs of 0.944 (training) and 0.945 (validation), maintaining an AUC of 0.863 in the testing set. Calibration curves confirmed a strong alignment between observed and predicted probabilities. Decision curve analysis indicated that the RF model provided the highest net benefit across threshold probabilities. The SHAP algorithm identified RPR, PLT, HBV DNA, ALT, and TBA as critical predictors. This interpretable ML model enhances early HBV-LC prediction and supports clinical decision-making in resource-limited settings.

Machine learning-based plasma metabolomics for improved cirrhosis risk stratification

BACKGROUND

Cirrhosis is a leading cause of mortality in patients with chronic liver disease (CLD). The rapid development of metabolomic technologies has enabled the capture of metabolic changes related to the progression of cirrhosis.

METHODS

This study used proton nuclear magnetic resonance (1 H-NMR) serum metabolomics data from the UK Biobank (UKB) and employed elastic net-regularized Cox proportional hazards models to explore the role of metabolomics in cirrhosis risk stratification in patients with CLD. Metabolomic data were integrated with aspartate aminotransferase to platelet ratio index (APRI) and fibrosis-4 score (FIB-4) to construct predictive models for cirrhosis risk. The model performance was assessed in both the derivation and validation cohorts.

RESULTS

A total of 2,738 eligible patients were included in the analysis. Several metabolites showed an independent association with cirrhosis events (68 out of 168 metabolites after adjustment for age and sex, and 21 out of 168 metabolites after full adjustment). The integration of metabolomics with FIB-4 improved the predictive performance compared to FIB-4 alone (Harrell's C: 0.717 vs. 0.696, ΔC = 0.021, 95% confidence interval [CI] 0.014-0.028, Net Reclassification Improvement [NRI]: 0.504 [0.488-0.520]). Similarly, the combination of metabolomics with APRI also improved predictive performance compared to APRI alone (Harrell's C: 0.747 vs. 0.718, ΔC = 0.029, 95% CI 0.022-0.035, NRI: 0.378 [0.366-0.389]). Key metabolites, including branched-chain amino acids (BCAAs), lipids, and markers of oxidative stress, were identified as significant predictors. Pathway enrichment analysis revealed that disruptions in lipid and amino acid metabolism play a central role in the progression of cirrhosis.

CONCLUSION

1 H-NMR serum metabolomics significantly improves the prediction of cirrhosis risk in patients with CLD. The APRI + Metabolomics model demonstrated strong discriminatory power, with key metabolites involved in fatty acid and amino acid metabolism, providing a promising tool for the early screening of cirrhosis risk.

Comparison of models to predict incident chronic liver disease: a systematic review and external validation in Chinese adults

BACKGROUND

Risk prediction models can identify individuals at high risk of chronic liver disease (CLD), but there is limited evidence on the performance of various models in diverse populations. We aimed to systematically review CLD prediction models, meta-analyze their performance, and externally validate them in 0.5 million Chinese adults in the China Kadoorie Biobank (CKB).

METHODS

Models were identified through a systematic review and categorized by the target population and outcomes (hepatocellular carcinoma [HCC] and CLD). The performance of models to predict 10-year risk of CLD was assessed by discrimination (C-index) and calibration (observed vs predicted probabilies).

RESULTS

The systematic review identified 57 articles and 114 models (28.4% undergone external validation), including 13 eligible for validation in CKB. Models with high discrimination (C-index ≥ 0.70) in CKB were as follows: (1) general population: Li-2018 and Wen 1-2012 for HCC, CLivD score (non-lab and lab) and dAAR for CLD; (2) hepatitis B virus (HBV) infected individuals: Cao-2021 for HCC and CAP-B for CLD. In CKB, all models tended to overestimate the risk (O:E ratio 0.55-0.94). In meta-analysis, we further identified models with high discrimination: (1) general population (C-index ≥ 0.70): Sinn-2020, Wen 2-2012, and Wen 3-2012 for HCC, and FIB-4 and Forns for CLD; (2) HBV infected individuals (C-index ≥ 0.80): RWS-HCC and REACH-B IIa for HCC and GAG-HCC for HCC and CLD.

CONCLUSIONS

Several models showed good discrimination and calibration in external validation, indicating their potential feasibility for risk stratification in population-based screening programs for CLD in Chinese adults.

Clinical Consequences of Hepatitis B Surface Antigen Loss in Chronic Hepatitis B Infection: A Systematic Literature Review and Meta-Analysis

Background and Aims

Functional cure, which requires sustained hepatitis B surface antigen (HBsAg) loss after treatment cessation, is currently the optimal treatment endpoint for chronic hepatitis B virus infection. We performed a systematic literature review (SLR) and meta-analyses to assess the association between HBsAg loss and long-term clinical outcomes.

Methods

We performed a SLR of scientific literature published in Medline and Embase reporting the incidence of cirrhosis, hepatic decompensation (HD), hepatocellular carcinoma (HCC), liver-related mortality (LRM), and all-cause mortality (ACM) in relation to HBsAg status. Bayesian hierarchical commensurate prior meta-analyses synthesized evidence on the association between HBsAg loss and each outcome.

Results

Thirty-eight studies, comprising 50,354 patients with 350,734 patient-years of follow-up, were included in the meta-analyses, reporting on cirrhosis (n = 12), HD (n = 12), HCC (n = 36), LRM (n = 12), and ACM (n = 16). Pooled incidence rate ratios (IRRs; vs HBsAg persistence) and respective credible intervals (Crls) were 0.28 (0.060-1.070) for cirrhosis, 0.13 (0.013-0.38) for HD, 0.27 (0.11-0.53) for HCC, 0.17 (0.028-0.61) for LRM, and 0.64 (0.24-1.17) for ACM. Single-predictor-adjusted IRRs remained consistent with those from the primary analyses for all outcomes except cirrhosis and LRM. Outcome incidence rates were modified by selected study, patient and infection characteristics, but trended in the same direction of reduced risk after loss.

Conclusion

Overall, HBsAg loss was associated with a reduced risk of most clinically relevant outcomes. While the magnitude of the effect differed across subgroups, the direction of the association remained similar. Our results validate the need to develop new strategies to achieve HBsAg loss.

Association Between Metabolic Dysfunction-Associated Fatty Liver Disease and the Risk of Cirrhosis in Patients with Chronic Hepatitis B-A Retrospective Cohort Study

BACKGROUND

Metabolic dysfunction-associated fatty liver disease (MAFLD) is a novel proposed concept that is being recognized worldwide. Both chronic hepatitis B (CHB) and MAFLD have been independently attributed to an increased risk of disease development to cirrhosis. However, it is still unclear whether MAFLD is associated with an increased risk of cirrhosis in CHB patients.

AIM

This study aimed to analyze the impact of MAFLD on the risk of cirrhosis in CHB patients.

METHODS

In this retrospective cohort study, consecutive CHB patients with or without MAFLD were enrolled from January 1st, 2007, to May 1st, 2020, in Guangdong Provincial Hospital of Chinese Medicine. Inverse probability treatment weighting (IPTW) was performed to balance the covariates across groups. The weighted Kaplan-Meier analysis and Cox regression analysis were used to compare both groups for the risk of cirrhosis.

RESULTS

A total of 1223 CHB patients were included in this study during the median follow-up of 5.25 years; of these patients, 355 were CHB-MAFLD patients. After IPTW, the weighted Kaplan-Meier analysis showed that the weighted cumulative incidence of cirrhosis was significantly higher in patients with MAFLD than that in patients without MAFLD (12.6% versus 7.1%, P=0.015). In the weighted multivariate Cox analysis, coexisting MAFLD was related to an increased risk of cirrhosis [adjusted weighted hazard ratio (HR) 1.790; P =0.020]. Age (>40 years, adjusted weighted HR, 1.950; P=0.015), diabetes mellitus (adjusted weighted HR, 1.883; P=0.041), non-antiviral treatment (adjusted weighted HR, 2.037; P=0.013), and baseline serum HBV DNA levels (>2.4 log¹⁰ IU/mL, adjusted weighted HR, 1.756; P=0.045) were significant risk factors for cirrhosis.

CONCLUSION

We found that MAFLD was associated with a higher risk of cirrhosis in CHB patients.