Abstract
We evaluated the value of secreted glycoprotein thrombospondin-2 (TSP-2) to predict hepatocellular carcinoma (HCC) occurrence in chronic hepatitis C (CHC) patients after Hepatitis C virus (HCV) elimination by direct-acting antiviral agents (DAAs). A total of 786 CHC patients without an HCC history who achieved a sustained virological response (SVR) with DAAs were randomly assigned 2:1, with 524 patients as the derivation cohort and 262 patients as the validation cohort. Serum TSP-2 levels at the end of treatment were measured by enzyme-linked immunosorbent assay (ELISA). In the derivation cohort, the cumulative HCC rate was significantly higher in the high TSP-2 group than in the low TSP-2 group. Multivariate Cox proportional hazards analysis revealed that TSP-2, α-fetoprotein (AFP), and the fibrosis-4 (FIB-4) index were independent HCC risk factors. The area under the receiver operating characteristic curve (AUROC) of the score calculated from these three factors (AFT score) for predicting HCC was 0.83, which was significantly higher than that of each factor alone (TSP-2: 0.70, AFP: 0.72, FIB-4: 0.69). The AFT score was used to stratify patients according to the risk of HCC occurrence in the validation cohort. Lastly, in patients with a FIB-4 index < 3.25, the serum TSP-2 levels could be used to identify those patients with a high risk of HCC occurrence. Serum TSP-2 levels are a predictive biomarker of HCC occurrence in CHC patients after HCV elimination by DAA treatment. The AFT score using TSP-2, AFP, and the FIB-4 index may identify those who require HCC surveillance.