Background: Although hepatocellular carcinoma (HCC) occurs mostly in patients with chronic hepatitis B virus (HBV) or hepatitis C virus (HCV) infection or heavy alcohol use or cirrhosis, some patients develop HCC without these risk factors. Our objective in this study was to develop and validate a new HCC risk score that could stratify HCC risk in patients who develop HCC without known risk factors.
Methods: A new HCC risk score was developed using a nationwide, population-based cohort among individuals without chronic HBV infection, chronic HCV infection, heavy alcohol use or cirrhosis (n = 467 206, derivation cohort). The performance of the HCC risk score was validated using an independent Samsung Medical Center Health Promotion Center cohort (n = 91 357, validation cohort).
Results: Multivariable Cox regression analysis identified six independent risk factors: age, sex, smoking, diabetes, total cholesterol level and serum alanine aminotransferase level. A 19-point scale for HCC risk score was developed, with 10-year risk of HCC ranging from 0.0% to 6.16% for the lowest and highest risk scores, respectively. The area under the receiver operating characteristics curve values (AUROCs) to predict HCC development were 0.83 [95% confidence interval (CI): 0.77, 0.88)] and 0.92 (95% CI: 0.89, 0.95) at 10 years in the derivation and validation cohorts, respectively. Predicted risk was well correlated with the Kaplan-Meier observed HCC risk.
Conclusions: A simple-to-use, novel HCC risk score was developed for predicting HCC development in individuals without alleged risk factors. It can be used to assess the risk of HCC in this population so that decisions about their clinical management, including risk reduction interventions, can be subsequently made.
【저자키워드】 diabetes, risk score, Hepatocellular carcinoma, alanine aminotransferase, National Cohort,