Nomograms for Risk of Hepatocellular Carcinoma in Patients With Chronic Hepatitis B Virus Infection

Counseling patients with chronic hepatitis B virus (HBV) on their individual risk of liver disease progression is challenging. This study aimed to develop nomograms for predicting hepatocellular carcinoma risk in patients with chronic hepatitis B. Two thirds of the Risk Evaluation of Viral Load Elevation and Associated Liver Disease/Cancer-Hepatitis B Virus (REVEAL-HBV) study cohort was allocated for model derivation (n = 2,435), and the remaining third was allocated for model validation (n = 1,218). Previously confirmed independent risk predictors included in three Cox proportional hazards regression models were sex, age, family history of hepatocellular carcinoma, alcohol consumption habit, serum ALT level, hepatitis B envelope antigen (HBeAg) serostatus, serum HBV DNA level, and HBV genotype. Regression coefficients were rounded into integer risk scores, and predicted risk over 5- and 10-year periods for each risk score was calculated and depicted in nomograms. The predictive accuracy was evaluated using the area under the receiver operating characteristic curve (AUROC) and the correlation between predicted and observed hepatocellular carcinoma risk. All selected risk predictors were statistically significant in all models. In each model, either HBeAg seropositivity or HBeAg seronegativity with high viral load (HBV DNA level >or= 100,000 copies/mL) and genotype C infection had the highest risk scores. All AUROCs for risk prediction nomogram were >or= 0.82 in both model derivation and validation sets. The correlation coefficients between the observed hepatocellular carcinoma risk and the nomogram-predicted risk were greater than 0.90 in all model derivation and validation sets. These easy-to-use nomograms based on noninvasive clinical characteristics can accurately predict the risk of hepatocellular carcinoma in patients with chronic hepatitis B. They may facilitate risk communication between patients and clinicians.