A model based on non-invasive markers predicts very low hepatocellular carcinoma risk after viral response in HCV-advanced fibrosis.

Background and aims Patients with HCV and advanced fibrosis remain at risk of hepatocellular carcinoma (HCC) after sustained viral response (SVR) and need life-long surveillance. Since HCC risk is not homogenous and may decrease with fibrosis regression, we aimed to identify patients with low HCC risk based on the prediction of non-invasive markers and its changes after SVR. Methods Multicenter cohort study including patients with HCV and compensated advanced fibrosis that achieved SVR after direct antivirals. Clinical and transient elastography (TE) data were registered at baseline, 1 and 3 yr. after the end of treatment (EOT). All patients underwent liver ultrasound scan every 6 months. Patients with clinical evaluation 1-yr after EOT were eligible. Univariate and multivariate Cox regression analysis were performed, and predictive models were constructed. HCC occurrence rates were evaluated by Kaplan-Meier. Results Nine hundred and ninety-three patients were eligible (56% M; 44% F; median age 62 yr.), 35 developed HCC (3.9 %), median follow-up 45 months (range 13-53). Baseline LSM (HR 1.040; 95% CI 1.017-1.064), serum albumin (HR 0.400; 95% CI 0.174-0.923), 1-yr delta LSM (HR 0.993; 95% CI 0.987-0.998) and 1-yr FIB-4 score (HR 1.095; 95% CI 1.046-1.146) were independent factors associated with HCC. The TE-based HCC risk model predicted 0% of HCC occurrence at 3 years in patients with score 0 (baseline LSM ≤17.3 KPa, albumin>4.2g/dl and 1-yr delta LSM >25.5%) vs 5.2% in patients with score 1-3. (Harrell´s C 0.779; log rank 0.002). An alternative model with FIB-4 similarly predict HCC risk. Conclusions A combination of baseline and dynamic changes in non-invasive markers may help to identify patients with a very low risk of HCC development after SVR.