Comparing Predictability of Non-invasive Tools for Hepatocellular Carcinoma in Treated Chronic Hepatitis C Patients.

BACKGROUND: Non-invasive tools including liver stiffness measurement (LSM) or FIB-4, assessed before or after direct acting antivirals (DAA), have been suggested to predict hepatocellular carcinoma (HCC). AIMS: This study aims to compare predictability of HCC by these methods at different time points, to validate the HCC surveillance suggestion by guidelines, and to propose personalized strategy. METHODS: Chronic hepatitis C whose LSM and FIB-4 were available at pretherapy and after sustained virological response (SVR) were enrolled. Advanced chronic liver disease (ACLD) was defined as pretherapy LSM ≥ 10 kPa or FIB-4 index ≥ 3.25 or ultrasound signs of cirrhosis plus platelet count < 150,000/µL. The predictabilities were compared by area under ROC. The cumulative HCC incidences were calculated by Kaplan-Meier analysis. RESULTS: Among 466 ACLD patients, 40 patients developed HCC during a follow-up duration of 26.8 months. Comparable predictive performances for HCC between LSM and FIB-4 at pretherapy and SVR were noted. By guidelines suggestion using pretherapy LSM = 10 kPa (advanced fibrosis) and 13 kPa (cirrhosis) for risk stratification, the annual HCC incidences of those with LSM of < 10, 10-12.9 and ≥ 13 kPa were 1.1, 3.6, and 5.0%, respectively. Combination of baseline LSM < 12 kPa and SVR FIB-4 < 3.7 could further stratify relatively low risk of HCC in ACLD patients of annal incidence of 1.2%. CONCLUSIONS: ACLD patients who met advanced fibrosis but not cirrhosis by guidelines' cut-offs still posed high risk of HCC. Baseline LSM with SVR FIB-4 can be applied to stratify low, intermediate, and high risk of HCC for personalizing surveillance strategies after SVR.