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

Abstract

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). 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. 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 ≥ 10kPa 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. Among 466 ACLD patients, 40 patients developed HCC during a follow-up duration of 26.8months. Comparable predictive performances for HCC between LSM and FIB-4 at pretherapy and SVR were noted. By guidelines suggestion using pretherapy LSM = 10kPa (advanced fibrosis) and 13kPa (cirrhosis) for risk stratification, the annual HCC incidences of those with LSM of < 10, 10-12.9 and ≥ 13kPa were 1.1, 3.6, and 5.0%, respectively. Combination of baseline LSM < 12kPa and SVR FIB-4 < 3.7 could further stratify relatively low risk of HCC in ACLD patients of annal incidence of 1.2%. 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.