A Risk Prediction Model for Hepatocellular Carcinoma in the General Population Without Traditional Risk Factors for Liver Disease.

Abstract

Background and aim: Existing hepatocellular carcinoma (HCC) prediction models for the general population without traditional risk factors for chronic liver disease are limited. This study aimed to develop an HCC prediction model for individuals lacking these traditional risk factors.

Methods: The total of 138 452 adult participants without chronic viral hepatitis or significant alcohol intake who underwent regular health checkup at a tertiary hospital in South Korea were followed up for the development of HCC. Risk factors for HCC development were analyzed using Cox regression analysis, and prediction model was developed using the risk factors.

Results: Significant predictors of HCC development included older age, male sex, higher body mass index, presence of diabetes mellitus, and levels of aspartate aminotransferase, total cholesterol, and platelet count. A new HCC prediction model using these variables was developed. Harrell's concordance index and Heagerty's integrated area under the receiver operating characteristics (AUROC) curve of the model were 0.88 (95% confidence interval [CI] 0.85-0.91) and 0.89 (95% CI 0.86-0.91), respectively. The 5- and 10-year AUROC were 0.89 (95% CI 0.88-0.89) and 0.87 (95% CI 0.87-0.88), respectively. This model significantly outperformed the FIB-4 scoring model in predicting HCC and effectively stratified individuals into low-, intermediate-, and high-risk groups with significantly different cumulative incidences of HCC.

Conclusions: The new model, based on clinical parameters, provides a valuable tool for clinicians to stratify HCC risk in the general population without risk factors for chronic liver disease.