Background and aim: Liver biopsy (LB) has lost popularity to stage liver fibrosis in the era of highly effective anti-hepatitis C virus (HCV) therapy, yet diagnosis of persistent cirrhosis may have important implications following HCV eradication. As performance of serological non-invasive tests (NITs) to predict residual fibrosis in non-viremic HCV patients is unknown, we investigated accuracy of NITs to predict residual fibrosis in cirrhotics after a sustained virological response (SVR) to interferon (IFN).
Methods: Thirty-eight patients with a pre-treatment histological diagnosis of cirrhosis and a 48-104 months post-SVR LB were tested with APRI, CDS, FIB-4, FibroQ, Forns Score, GUCI Index, King Score, Lok Index, PLF, ELF. In 23 (61%) patients, cirrhosis had histologically regressed.
Results: All NITs values declined after SVR without any significant difference between regressors and non-regressors (AUROC 0.52-0.75). Using viremic cut-offs, PPV ranged from 34% to 100%, with lower NPV (63% - 68%). NITs performance did not improve using derived cut-offs (PPV: 40% - 80%; NPV: 66% - 100%). PLF, which combines several NITs with transient elastography, had the best diagnostic performance (AUROC 0.75, Sn 61%, Sp 90%, PPV 80%, NPV 78%). After treatment, none of the NITs resulted significantly associated with any of the histological features (activity grade, fibrosis stage, area of fibrosis).
Conclusions: The diagnostic estimates obtained using both viremic and derived cut-off values of NITs were suboptimal, indicating that none of these tests helps predicting residual fibrosis and that LB remains the gold standard for this purpose.