Advertisement

HCC risk stratification after cure of hepatitis C in patients with compensated advanced chronic liver disease

  • Georg Semmler
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria

    Vienna Hepatic Hemodynamic Lab, Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • Elias Laurin Meyer
    Affiliations
    Institute for Medical Statistics, Center for Medical Statistics, Informatics and Intelligent Systems, Medical University Vienna, Vienna, Austria
    Search for articles by this author
  • Karin Kozbial
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • Philipp Schwabl
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria

    Vienna Hepatic Hemodynamic Lab, Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • Stefanie Hametner-Schreil
    Affiliations
    Internal Medicine IV, Ordensklinikum Linz Barmherzige Schwestern, Linz, Austria
    Search for articles by this author
  • Alberto Zanetto
    Affiliations
    Gastroenterology and Multivisceral Transplant Unit, Department of Surgery, Oncology, and Gastroenterology, Padua University Hospital, Padua, Italy
    Search for articles by this author
  • David Bauer
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria

    Vienna Hepatic Hemodynamic Lab, Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • David Chromy
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria

    Vienna Hepatic Hemodynamic Lab, Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • Benedikt Simbrunner
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria

    Vienna Hepatic Hemodynamic Lab, Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • Bernhard Scheiner
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria

    Vienna Hepatic Hemodynamic Lab, Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • Albert F. Stättermayer
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria

    Vienna Hepatic Hemodynamic Lab, Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • Matthias Pinter
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria

    Vienna Hepatic Hemodynamic Lab, Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • Rainer Schöfl
    Affiliations
    Internal Medicine IV, Ordensklinikum Linz Barmherzige Schwestern, Linz, Austria
    Search for articles by this author
  • Francesco Paolo Russo
    Affiliations
    Gastroenterology and Multivisceral Transplant Unit, Department of Surgery, Oncology, and Gastroenterology, Padua University Hospital, Padua, Italy
    Search for articles by this author
  • Helena Greenfield
    Affiliations
    Vienna Hepatic Hemodynamic Lab, Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • Michael Schwarz
    Affiliations
    Department of Gastroenterology and Hepatology, Klinikum Ottakring, Vienna, Austria
    Search for articles by this author
  • Caroline Schwarz
    Affiliations
    Department of Gastroenterology and Hepatology, Klinikum Ottakring, Vienna, Austria
    Search for articles by this author
  • Michael Gschwantler
    Affiliations
    Department of Gastroenterology and Hepatology, Klinikum Ottakring, Vienna, Austria
    Search for articles by this author
  • Sonia Alonso López
    Affiliations
    Liver Unit, Hospital General Universitario Gregorio Marañón, Madrid, Spain

    Instituto De Investigación Sanitaria Gregorio Marañón (IiSGM), Madrid, Spain
    Search for articles by this author
  • Maria Luisa Manzano
    Affiliations
    Liver Unit, Hospital Universitario 12 De Octubre, Madrid, Spain
    Search for articles by this author
  • Adriana Ahumada
    Affiliations
    Liver Unit, Hospital General Universitario Gregorio Marañón, Madrid, Spain
    Search for articles by this author
  • Rafael Bañares
    Affiliations
    Liver Unit, Hospital General Universitario Gregorio Marañón, Madrid, Spain

    Instituto De Investigación Sanitaria Gregorio Marañón (IiSGM), Madrid, Spain

    Universidad Complutense de Madrid, Madrid, Spain

    Centro de Investigación Biomédica En Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain
    Search for articles by this author
  • Mònica Pons
    Affiliations
    Liver Unit, Department of Internal Medicine, Hospital Universitari Vall d’Hebron, Vall d’Hebron Institut de Recerca (VHIR), Vall d’Hebron Barcelona Hospital Campus, Universitat Autònoma de Barcelona, Barcelona, Spain
    Search for articles by this author
  • Sergio Rodríguez-Tajes
    Affiliations
    Centro de Investigación Biomédica En Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain

    Liver Unit, Hospital Clínic, Universitat de Barcelona, Barcelona, Spain

    August Pi i Sunyer Biomedical Research Institute (IDIBAPS), Universitat de Barcelona, Barcelona, Spain
    Search for articles by this author
  • Joan Genescà
    Affiliations
    Centro de Investigación Biomédica En Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain

    Liver Unit, Department of Internal Medicine, Hospital Universitari Vall d’Hebron, Vall d’Hebron Institut de Recerca (VHIR), Vall d’Hebron Barcelona Hospital Campus, Universitat Autònoma de Barcelona, Barcelona, Spain
    Search for articles by this author
  • Sabela Lens
    Affiliations
    Centro de Investigación Biomédica En Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain

    Liver Unit, Hospital Clínic, Universitat de Barcelona, Barcelona, Spain

    August Pi i Sunyer Biomedical Research Institute (IDIBAPS), Universitat de Barcelona, Barcelona, Spain
    Search for articles by this author
  • Michael Trauner
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • Peter Ferenci
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • Thomas Reiberger
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria

    Vienna Hepatic Hemodynamic Lab, Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
  • Mattias Mandorfer
    Correspondence
    Corresponding author. Address: Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Währinger Gürtel 18-20, 1090, Vienna, Austria; Tel.: +43 1 40400 47440, fax: +43 1 40400 47350.
    Affiliations
    Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria

    Vienna Hepatic Hemodynamic Lab, Division of Gastroenterology and Hepatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
    Search for articles by this author
Open AccessPublished:December 03, 2021DOI:https://doi.org/10.1016/j.jhep.2021.11.025

      Highlights

      • We studied de novo HCC development in patients with cACLD after SVR in a derivation cohort (n = 475) and validation cohort (n = 1,500).
      • Algorithms based on post-treatment age/albumin/LSM, and optionally, AFP and alcohol consumption, accurately stratified de novo HCC risk.
      • Approximately two-thirds of patients were identified as having an HCC risk <1%/year.
      • In these patients, HCC-surveillance might not be cost-effective.

      Background & Aims

      Hepatocellular carcinoma (HCC) is a major cause of morbidity and mortality in patients with advanced chronic liver disease (ACLD) caused by chronic hepatitis C who have achieved sustained virologic response (SVR). We developed risk stratification algorithms for de novo HCC development after SVR and validated them in an independent cohort.

      Methods

      We evaluated the occurrence of de novo HCC in a derivation cohort of 527 patients with pre-treatment ACLD and SVR to interferon-free therapy, in whom alpha-fetoprotein (AFP) and non-invasive surrogates of portal hypertension including liver stiffness measurement (LSM) were assessed pre-/post-treatment. We validated our results in 1,500 patients with compensated ACLD (cACLD) from other European centers.

      Results

      During a median follow-up (FU) of 41 months, 22/475 patients with cACLD (4.6%, 1.45/100 patient-years) vs. 12/52 decompensated patients (23.1%, 7.00/100 patient-years, p <0.001) developed de novo HCC. Since decompensated patients were at substantial HCC risk, we focused on cACLD for all further analyses.
      In cACLD, post-treatment-values showed a higher discriminative ability for patients with/without de novo HCC development during FU than pre-treatment values or absolute/relative changes. Models based on post-treatment AFP, alcohol consumption (optional), age, LSM, and albumin, accurately predicted de novo HCC development (bootstrapped Harrel’s C with/without considering alcohol: 0.893/0.836). Importantly, these parameters also provided independent prognostic information in competing risk analysis and accurately stratified patients into low- (~2/3 of patients) and high-risk (~1/3 of patients) groups in the derivation (algorithm with alcohol consumption; 4-year HCC-risk: 0% vs. 16.5%) and validation (3.3% vs. 17.5%) cohorts. An alternative approach based on alcohol consumption (optional), age, LSM, and albumin (i.e., without AFP) also showed a robust performance.

      Conclusions

      Simple algorithms based on post-treatment age/albumin/LSM, and optionally, AFP and alcohol consumption, accurately stratified patients with cACLD based on their risk of de novo HCC after SVR. Approximately two-thirds were identified as having an HCC risk <1%/year in both the derivation and validation cohort, thereby clearly falling below the cost-effectiveness threshold for HCC surveillance.

      Lay summary

      Simple algorithms based on age, alcohol consumption, results of blood tests (albumin and α-fetoprotein), as well as liver stiffness measurement after the end of hepatitis C treatment identify a large proportion (approximately two-thirds) of patients with advanced but still asymptomatic liver disease who are at very low risk (<1%/year) of liver cancer development, and thus, might not need to undergo 6-monthly liver ultrasound.

      Graphical abstract

      Keywords

      Linked Article

      Introduction

      Direct acting antiviral (DAA)-based interferon (IFN)-free therapies for chronic hepatitis C (CHC) are highly effective, achieving sustained virologic response (SVR; i.e., HCV cure) in almost all patients with advanced chronic liver disease (ACLD).
      • Mandorfer M.
      • Kozbial K.
      • Freissmuth C.
      • Schwabl P.
      • Stattermayer A.F.
      • Reiberger T.
      • et al.
      Interferon-free regimens for chronic hepatitis C overcome the effects of portal hypertension on virological responses.
      SVR following IFN-free treatment has not only been associated with improvements in surrogates of portal hypertension such as liver stiffness measurement (LSM) or von Willebrand factor (VWF) levels, but also with amelioration of portal hypertension as assessed by hepatic venous pressure gradient (HVPG).
      • Lens S.
      • Alvarado-Tapias E.
      • Marino Z.
      • Londono M.C.
      • Elba L.L.
      • Martinez J.
      • et al.
      Effects of all-oral anti-viral therapy on HVPG and systemic hemodynamics in patients with hepatitis C virus-associated cirrhosis.
      • Lens S.
      • Baiges A.
      • Alvarado E.
      • Llop E.
      • Martinez J.
      • Fortea J.I.
      • et al.
      Clinical outcome and hemodynamic changes following HCV eradication with oral antiviral therapy in patients with clinically significant portal hypertension.
      • Mandorfer M.
      • Kozbial K.
      • Schwabl P.
      • Freissmuth C.
      • Schwarzer R.
      • Stern R.
      • et al.
      Sustained virologic response to interferon-free therapies ameliorates HCV-induced portal hypertension.
      • Mandorfer M.
      • Kozbial K.
      • Schwabl P.
      • Chromy D.
      • Semmler G.
      • Stättermayer A.F.
      • et al.
      Changes in hepatic venous pressure gradient predict hepatic decompensation in patients who achieved sustained virologic response to interferon-free therapy.
      • Mauro E.
      • Crespo G.
      • Montironi C.
      • Londoño M.C.
      • Hernández-Gea V.
      • Ruiz P.
      • et al.
      Portal pressure and liver stiffness measurements in the prediction of fibrosis regression after sustained virological response in recurrent hepatitis C.
      Decreases in the severity of portal hypertension translate into reductions in hepatic decompensation
      • Mandorfer M.
      • Kozbial K.
      • Schwabl P.
      • Chromy D.
      • Semmler G.
      • Stättermayer A.F.
      • et al.
      Changes in hepatic venous pressure gradient predict hepatic decompensation in patients who achieved sustained virologic response to interferon-free therapy.
      and concordantly liver-related mortality.
      • Backus L.I.
      • Belperio P.S.
      • Shahoumian T.A.
      • Mole L.A.
      Impact of sustained virologic response with direct-acting antiviral treatment on mortality in patients with advanced liver disease.
      • Carrat F.
      • Fontaine H.
      • Dorival C.
      • Simony M.
      • Diallo A.
      • Hezode C.
      • et al.
      Clinical outcomes in patients with chronic hepatitis C after direct-acting antiviral treatment: a prospective cohort study.
      • Nahon P.
      • Bourcier V.
      • Layese R.
      • Audureau E.
      • Cagnot C.
      • Marcellin P.
      • et al.
      Eradication of hepatitis C virus infection in patients with cirrhosis reduces risk of liver and non-liver complications.
      Nevertheless, a considerable proportion of patients remains at risk of developing complications of ACLD. While the incidence of hepatic decompensation seems to be comparatively low
      • Pons M.
      • Rodríguez-Tajes S.
      • Esteban J.I.
      • Mariño Z.
      • Vargas V.
      • Lens S.
      • et al.
      Non-invasive prediction of liver-related events in patients with HCV-associated compensated advanced chronic liver disease after oral antivirals.
      and non-invasive markers such as LSM and VWF/platelet count ratio (VITRO) facilitate risk stratification,
      • Semmler G.
      • Binter T.
      • Kozbial K.
      • Schwabl P.
      • Hametner-Schreil S.
      • Zanetto A.
      • et al.
      Noninvasive risk stratification after HCV eradication in patients with advanced chronic liver disease.
      de novo HCC development remains a major concern. Specifically, the incidence of HCC ranged from 1.5-1.8
      • Pons M.
      • Rodríguez-Tajes S.
      • Esteban J.I.
      • Mariño Z.
      • Vargas V.
      • Lens S.
      • et al.
      Non-invasive prediction of liver-related events in patients with HCV-associated compensated advanced chronic liver disease after oral antivirals.
      ,
      • Kanwal F.
      • Kramer J.
      • Asch S.M.
      • Chayanupatkul M.
      • Cao Y.
      • El-Serag H.B.
      Risk of hepatocellular cancer in HCV patients treated with direct-acting antiviral agents.
      to 3.6/100 patient-years
      • Ioannou G.N.
      • Beste L.A.
      • Green P.K.
      • Singal A.G.
      • Tapper E.B.
      • Waljee A.K.
      • et al.
      Increased risk for hepatocellular carcinoma persists up to 10 Years after HCV eradication in patients with baseline cirrhosis or high FIB-4 scores.
      ,
      • Finkelmeier F.
      • Dultz G.
      • Peiffer K.H.
      • Kronenberger B.
      • Krauss F.
      • Zeuzem S.
      • et al.
      Risk of de novo Hepatocellular Carcinoma after HCV Treatment with Direct-Acting Antivirals.
      in patients with ACLD/cirrhosis. Of note, clinically significant portal hypertension (CSPH, as defined by an HVPG ≥10 mmHg) is accompanied by a 6-fold increased risk of HCC in compensated ACLD (cACLD), suggesting that the aforementioned surrogates of portal hypertension may also indicate HCC risk.
      • Ripoll C.
      • Groszmann R.J.
      • Garcia-Tsao G.
      • Bosch J.
      • Grace N.
      • Burroughs A.
      • et al.
      Hepatic venous pressure gradient predicts development of hepatocellular carcinoma independently of severity of cirrhosis.
      While no data on VITRO is available, the occurrence of de novo HCC has previously been associated with LSM as well as traditional HCC risk factors such as age, serum albumin, and alpha-fetoprotein (AFP).
      • Carrat F.
      • Fontaine H.
      • Dorival C.
      • Simony M.
      • Diallo A.
      • Hezode C.
      • et al.
      Clinical outcomes in patients with chronic hepatitis C after direct-acting antiviral treatment: a prospective cohort study.
      ,
      • Nahon P.
      • Bourcier V.
      • Layese R.
      • Audureau E.
      • Cagnot C.
      • Marcellin P.
      • et al.
      Eradication of hepatitis C virus infection in patients with cirrhosis reduces risk of liver and non-liver complications.
      ,
      • Chromy D.
      • Mandorfer M.
      • Bucsics T.
      • Schwabl P.
      • Bauer D.
      • Scheiner B.
      • et al.
      Prevalence and predictors of hepatic steatosis in patients with HIV/HCV coinfection and the impact of HCV eradication.
      Several risk prediction models have been proposed based on these and other factors; however, all of these previously published scores have yet to undergo external validation. Thus, no recommendation regarding the identification of a low-risk subgroup of patients with ACLD in whom HCC surveillance is not cost-effective/warranted has been implemented in recent guidelines on the management and follow-up (FU) of CHC.
      • Farhang Zangneh H.
      • Wong W.W.L.
      • Sander B.
      • Bell C.M.
      • Mumtaz K.
      • Kowgier M.
      • et al.
      Cost effectiveness of hepatocellular carcinoma surveillance after a sustained virologic response to therapy in patients with hepatitis C virus infection and advanced fibrosis.
      We investigated the incidence of de novo HCC and its prediction in a comprehensively characterized cohort of patients with ACLD from 3 tertiary centers and aimed to validate the prognostic algorithms we developed in a large, independent validation cohort comprising patients with cACLD from other European centers. In addition, we applied previously published risk prediction models to both cohorts to evaluate their prognostic accuracy.

      Patients and methods

      Derivation cohort

      All patients achieving SVR after DAA-based IFN-free treatment at the Medical University of Vienna, Padua University Hospital, and Ordensklinikum Linz Barmherzige Schwestern with pre-treatment ACLD (defined as baseline [BL]-LSM ≥10 kPa, HVPG ≥6 mmHg, or advanced fibrosis/cirrhosis on liver histology [F3/4]) were screened for eligibility for this retrospective study based on prospectively collected data.
      • de Franchis R.
      • Baveno V.I.F.
      Expanding consensus in portal hypertension: report of the Baveno VI Consensus Workshop: stratifying risk and individualizing care for portal hypertension.
      After excluding all patients with Child-Pugh stage C who were not candidates for liver transplantation (i.e., patients in whom surveillance is not recommended
      EASL Clinical Practice Guidelines
      Management of hepatocellular carcinoma.
      ), a history or a current diagnosis of HCC, porto-sinusoidal vascular disease, previous orthotopic liver transplantation (OLT), or an HCC diagnosis/OLT during treatment from the dataset, 527 patients were included. Notably, subgroups of these patients have been previously investigated with regard to changes in HVPG and their prognostic value,
      • Mandorfer M.
      • Kozbial K.
      • Schwabl P.
      • Freissmuth C.
      • Schwarzer R.
      • Stern R.
      • et al.
      Sustained virologic response to interferon-free therapies ameliorates HCV-induced portal hypertension.
      ,
      • Mandorfer M.
      • Kozbial K.
      • Schwabl P.
      • Chromy D.
      • Semmler G.
      • Stättermayer A.F.
      • et al.
      Changes in hepatic venous pressure gradient predict hepatic decompensation in patients who achieved sustained virologic response to interferon-free therapy.
      the diagnostic/predictive ability of non-invasive markers for portal hypertension and hepatic decompensation,
      • Semmler G.
      • Binter T.
      • Kozbial K.
      • Schwabl P.
      • Hametner-Schreil S.
      • Zanetto A.
      • et al.
      Noninvasive risk stratification after HCV eradication in patients with advanced chronic liver disease.
      the predictive value of VITRO for hepatic decompensation,
      • Schwarzer R.
      • Reiberger T.
      • Mandorfer M.
      • Kivaranovic D.
      • Hametner S.
      • Hametner S.
      • et al.
      The von Willebrand Factor antigen to platelet ratio (VITRO) score predicts hepatic decompensation and mortality in cirrhosis.
      the influence of genetic variants on liver disease regression,
      • Semmler G.
      • Binter T.
      • Kozbial K.
      • Schwabl P.
      • Chromy D.
      • Bauer D.
      • et al.
      Influence of genetic variants on disease regression and outcomes in HCV-related advanced chronic liver disease after SVR.
      as well as changes in coagulation after HCV cure.
      • Russo F.P.
      • Zanetto A.
      • Campello E.
      • Bulato C.
      • Shalaby S.
      • Spiezia L.
      • et al.
      Reversal of hypercoagulability in patients with HCV-related cirrhosis after treatment with direct-acting antivirals.
      However, none of these studies focused on HCC.

      Clinical and laboratory parameters and liver stiffness measurement

      Clinical and laboratory parameters were evaluated by chart review. Alcohol consumption above the threshold for non-alcoholic fatty liver disease was defined as >30 g/day and >20 g/day for males and females, respectively.
      European Association for the Study of the LEuropean Association for the Study of DEuropean Association for the Study of O
      EASL-EASD-EASO Clinical Practice Guidelines for the management of non-alcoholic fatty liver disease.
      Plasma VWF antigen levels were measured by a latex agglutination assay (STA LIATEST VWF, Diagnostica Stago, Asnieres, France). VITRO score was calculated by dividing VWF (%) over platelet count (PLT) (G × L−1), as described previously.
      • Maieron A.
      • Salzl P.
      • Peck-Radosavljevic M.
      • Trauner M.
      • Hametner S.
      • Schofl R.
      • et al.
      Von Willebrand Factor as a new marker for non-invasive assessment of liver fibrosis and cirrhosis in patients with chronic hepatitis C.
      Paired measurements of non-invasive markers were performed prior to antiviral therapy, as well as after the end of treatment (EoT). Due to the retrospective design of this study (and also for logistical reasons), the time points were not standardized. Vibration-controlled transient elastography (FibroScan; Echosens, Paris, France) was used for LSM. All measurements were performed after a minimum fasting period of 4 hours and in the absence of relevant amounts of ascites.

      HCV therapy

      All patients were treated with IFN-free therapies. The choice of the regimen was at the physicians’ discretion and depended on their availability, reimbursement policies, and national as well as international clinical practice guidelines at the time of treatment initiation.
      • Sarrazin C.
      • Berg T.
      • Buggisch P.
      • Dollinger M.M.
      • Hinrichsen H.
      • Hofer H.
      • et al.
      S3 guideline hepatitis C addendum.
      • Sarrazin C.
      • Zimmermann T.
      • Berg T.
      • Neumann U.P.
      • Schirmacher P.
      • Schmidt H.
      • et al.
      [Prophylaxis, diagnosis and therapy of hepatitis-C-virus (HCV) infection: the German guidelines on the management of HCV infection - AWMF-Register-No.: 021/012].
      EASL recommendations on treatment of hepatitis C 2015.
      EASL recommendations on treatment of hepatitis C 2016.
      EASL recommendations on treatment of hepatitis C 2018.
      Treatment duration ranged from 8 to 24 weeks.

      HCC surveillance

      All patients underwent HCC surveillance either by ultrasound, computed tomography, or magnetic resonance imaging on a 6-monthly basis. HCC was diagnosed based on EASL clinical practice guidelines at the time.
      EASL Clinical Practice Guidelines
      Management of hepatocellular carcinoma.
      ,
      EASL-EORTC clinical practice guidelines: management of hepatocellular carcinoma.

      Validation cohort

      Data was collected from 1,500 patients with cACLD and without a history of HCC/OLT treated at other European centers (Hospital General Universitario Gregorio Marañón and Hospital Universitario 12 De Octubre [Madrid, Spain], Hospital Universitari Vall d’Hebron and Hospital Clínic [Barcelona, Spain], and Klinikum Ottakring [Vienna, Austria]). All patients achieved SVR after DAA-based IFN-free treatments. For the Spanish cohorts, details regarding in- and exclusion criteria as well as study design are provided in the individual publications,
      • Pons M.
      • Rodríguez-Tajes S.
      • Esteban J.I.
      • Mariño Z.
      • Vargas V.
      • Lens S.
      • et al.
      Non-invasive prediction of liver-related events in patients with HCV-associated compensated advanced chronic liver disease after oral antivirals.
      ,
      • Alonso S.
      • Manzano M.L.
      • Gea F.
      • Gutiérrez M.L.
      • Ahumada A.M.
      • Devesa M.J.
      • et al.
      A model based on non-invasive markers predicts very low hepatocellular carcinoma risk after viral response in HCV-advanced fibrosis.
      whereas for patients from the other Viennese hospital (Klinikum Ottakring) contributing to the derivation cohort, criteria/design were similar to the validation cohort (Table S1). Patients with missing data on FU-LSM and FU-albumin were not considered for our analyses.

      Statistical analyses

      Statistical analyses were performed using IBM SPSS Statistics 25 (SPSS Inc., USA) and R 4.0.5. (R Core Team, R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were reported as mean ± standard deviation or median (interquartile range), while categorical variables were reported as proportion of patients with/without a certain characteristic. Student’s t test was used for group comparisons of normally distributed variables and Mann-Whitney U test for non-normally distributed variables, respectively. Group comparisons of categorical variables were performed using either Pearson's Chi-squared or Fisher’s exact test. The areas under the curve (AUC) and respective 95% CIs of receiver-operating characteristic (ROC) analyses were calculated for continuous variables using the R package ‘cutpointr’, applying Youden’s J-statistic to obtain the respective optimized cut-offs for classifying patients regarding HCC development. To increase the reliability of these cut-offs, we performed bootstrap resampling 5,000 times. Univariable and multivariable Cox regression analyses were performed using the R ‘survival’ package to investigate the association of individual (continuous and binary) parameters with HCC development. For further model development, backward elimination excluding variables with p >0.100 was applied to identify variables that provide certain information for HCC prediction. For these analyses, the time to event was calculated from the EoT, and patients were censored at OLT, death, or end of FU. Harrel’s C-indices for the respective models were derived using the R package ‘dynpred’ with bootstrap resampling performed 5000 times to increase the generalizability of these models. Fine and Gray competing risks regression models were calculated with the R package ‘cmprsk’ to test whether variables included in the final model were still independently associated with HCC when considering OLT and death as competing risks.
      • Fine J.P.
      • Gray R.J.
      A proportional hazards model for the subdistribution of a competing risk.
      Finally, a score was derived from respective adjusted subdistribution hazard ratios (aSHRs). Moreover, published prediction models were tested in our cohort using Gray's test for subdistribution hazards. A p value ≤0.05 was considered statistically significant. As p values served only descriptive purposes, no multiplicity correction was applied.

      Ethics

      This study (derivation cohort) was approved by the institutional review boards (IRB) of the Medical University of Vienna (EK 1947/2019), Upper Austria (K-49-14), and Padua University Hospital (3103/A0/14). Written informed consent was obtained, if the requirement was not waived by the local IRB. For the validation cohort, approval of local IRB (City of Vienna for Klinik Ottakring; as described previously for Madrid
      • Alonso S.
      • Manzano M.L.
      • Gea F.
      • Gutiérrez M.L.
      • Ahumada A.M.
      • Devesa M.J.
      • et al.
      A model based on non-invasive markers predicts very low hepatocellular carcinoma risk after viral response in HCV-advanced fibrosis.
      and Barcelona
      • Pons M.
      • Rodríguez-Tajes S.
      • Esteban J.I.
      • Mariño Z.
      • Vargas V.
      • Lens S.
      • et al.
      Non-invasive prediction of liver-related events in patients with HCV-associated compensated advanced chronic liver disease after oral antivirals.
      ) was obtained.

      Results

      Characteristics of the derivation cohort

      The mean age of patients in the derivation cohort was 57.3 ± 11.1 years (Table S2) and 328 patients (62.2%) were male; 475 (90.1%) had compensated ACLD (cACLD) while 52 (9.9%) had not previously experienced any hepatic decompensation (dACLD). Varices were prevalent in 122 (23.1%) and 44 patients (8.3%) had Child-Pugh stage B/C, while 483 (91.7%) had Child-Pugh stage A with a mean model for end-stage liver disease (MELD) score of 8.6 ± 2.8 points. During a median FU of 41 (32) months, 34 (6.5%) developed HCC corresponding to an HCC incidence of 1.78/100 patient-years. Of note, 22 patients with cACLD (4.6%) developed HCC (1.45/100 patient-years) vs. 12 patients with dACLD (23.1%, p <0.001, 7.00/100 patient-years). Since patients with dACLD were at very high risk of de novo HCC development, we abstained from merging them with patients with cACLD. Moreover, the limited number of patients precluded dedicated analyses on risk factors for HCC in patients with dACLD. Accordingly, all other analyses focused on cACLD.

      cACLD subgroup of the derivation cohort

      Characteristics of patients with cACLD with and without HCC during FU (median 41 [IQR 33] months) are presented in Table S3. Also, time points of FU measurements are shown in Fig. S1A which clustered around 12 weeks after EoT. Differences in patient characteristics were observed for age, presence of varices and non-invasive markers of portal hypertension (i.e., LSM, PLT, VWF, and VITRO), hepatic function (i.e., MELD and serum albumin), as well as aspartate aminotransferase, AFP, and composite scores (i.e., aspartate aminotransferase-to-platelet ratio index [APRI] and Fibrosis-4 [FIB-4]) both at BL and FU. Of note, none of the patients who developed HCC during FU had uncharacterized nodules at BL.
      Following univariable ROC analyses, a similar moderate accuracy (AUC <0.800) to identify patients with HCC was evident for several continuous variables (Table 2). Specifically, FU-albumin, FU-LSM, BL-VWF, BL-/FU-VITRO, FU-APRI, BL-/FU-FIB-4, BL-/FU-AFP showed an AUC of 0.700-0.800 with FU-AFP having the numerically highest AUC (0.796; 95% CI 0.726-0.866). Of note, FU variables tended to be more informative than BL parameters. Again, absolute and relative changes were considerably less accurate (AUC <0.700) with relative Δ LSM showing the highest AUC (0.674; 95% CI 0.570-0.778). These analyses indicated that single parameters are incapable of accurately predicting HCC development in the post-SVR setting.
      Table 1Comparison of patient characteristics at BL and FU in the derivation and validation cohorts.
      Patient characteristicsDerivation cohort cACLD∗, n = 475Validation cohort AFP∗∗, n = 691Validation cohort non-AFP∗∗∗, n = 1,500p value

      (∗ vs. ∗∗)
      Comparison of continuous variables was performed using Student’s t test for normally distributed variables and Mann Whitney U test for non-normally distributed variables, respectively. Group comparisons of categorical variables were performed using Pearson's Chi-squared or Fisher’s exact test, as applicable. Values in bold indicate p ≤0.05.
      p value

      (∗∗ vs. ∗∗∗)
      Comparison of continuous variables was performed using Student’s t test for normally distributed variables and Mann Whitney U test for non-normally distributed variables, respectively. Group comparisons of categorical variables were performed using Pearson's Chi-squared or Fisher’s exact test, as applicable. Values in bold indicate p ≤0.05.
      Age, years57.6 ± 11.359.1 ± 12.361.4 ± 11.70.024<0.001
      Sex
       Male296 (62.3%)435 (63.0%)840 (56.0%)0.8250.015
       Female179 (37.7%)256 (37.0%)660 (44.0%)
      Ethnicityn = 475n = 689n = 1,321--
       Caucasian438 (92.2%)654 (94.9%)1,278 (96.7%)
       African30 (6.3%)13 (1.9%)15 (1.1%)
       Asian7 (1.5%)20 (2.9%)22 (1.7%)
       Latin-American0 (0%)2 (0.3%)6 (0.5%)
      BL-albumin, g⋅L-141.5 ± 4.241.3 ± 4.541.3 ± 4.40.4520.343
      BL-LSM, kPa16.0 (14.3)15.0 (13.1)16.3 (12.5)0.7450.402
      BL-PLT, G⋅L-1157 ± 65155 ± 68150 ± 660.7090.048
      BL-AFP, ng⋅ml-16.5 (10.7)6.7 (9.0)6.7 (9.0)0.4130.388
      BMI, kg⋅m-226.9 ± 5.0 (n = 470)27.2 ± 4.5 (n = 503)26.9 ± 4.4 (n = 1,096)0.3050.942
       ≥30 kg⋅m-2108 (23.0%)115 (22.9%)221 (20.2%)0.9660.210
      Diabetes
      Fasting blood glucose >125 mg⋅dl-1, HbA1c ≥6.5%, or antidiabetic medication.
      79 (16.6%)118 (17.1%)274 (18.3%)0.8420.418
      Alcoholn = 475n = 654n = 1,287
       Below the threshold
      >30 g/day and >20 g/day for males and females, respectively.23
      443 (93.3%)605 (92.5%)1,206 (93.7%)0.6270.736
       Above the threshold
      >30 g/day and >20 g/day for males and females, respectively.23
      32 (6.7%)49 (7.5%)81 (6.3%)
      FU-albumin, g⋅L-143.2 ± 3.743.0 ± 4.443.3 ± 3.90.3380.855
      FU-LSM, kPa11.8 (10.9)10.4 (9.0)10.4 (8.7)0.009<0.001
      FU-PLT, G⋅L-1170 ± 69168 ± 66159 ± 680.4790.003
      FU-AFP, ng⋅ml-13.6 (3.3)3.3 (2.8)3.3 (2.8)0.2010.201
      FU time
      According to the reverse Kaplan-Meier method.
      42.4 (39.6-45.1)44.4 (42.7-46.0)40.4 (39.7-41.2)--
      HCC22 (4.6%)36 (5.2%)65 (4.3%)0.6550.783
      Incidence/100 patient-years1.451.741.40--
      AFP, alpha-fetoprotein; BL, baseline; cACLD, compensated advanced chronic liver disease; FU, follow-up; HCC, hepatocellular carcinoma; LSM, liver stiffness measurement; PLT, platelet count.
      1 Comparison of continuous variables was performed using Student’s t test for normally distributed variables and Mann Whitney U test for non-normally distributed variables, respectively. Group comparisons of categorical variables were performed using Pearson's Chi-squared or Fisher’s exact test, as applicable. Values in bold indicate p ≤0.05.
      2 Fasting blood glucose >125 mg⋅dl-1, HbA1c ≥6.5%, or antidiabetic medication.
      3 >30 g/day and >20 g/day for males and females, respectively.
      European Association for the Study of the LEuropean Association for the Study of DEuropean Association for the Study of O
      EASL-EASD-EASO Clinical Practice Guidelines for the management of non-alcoholic fatty liver disease.
      4 According to the reverse Kaplan-Meier method.
      Table 2AUC values of pre-treatment and post-treatment parameters, as well as their absolute and relative changes, for predicting hepatocellular carcinoma development in the derivation cohort.
      ParameterAUC (95% CI)ParameterAUC (95% CI)
      Age0.664 (0.546-0.782)--
      BL-albumin, g⋅L-10.691 (0.586-0.796)Absolute Δ albumin, g⋅L-10.541 (0.423-0.658)
      FU-albumin, g⋅L-10.714 (0.594-0.835)Relative Δ albumin, %0.534 (0.414-0.654)
      BL-LSM, kPa0.631 (0.522-0.741)Absolute Δ LSM, kPa0.610 (0.476-0.744)
      FU-LSM, kPa0.713 (0.621-0.805)Relative Δ LSM, %0.674 (0.570-0.778)
      BL-PLT, G⋅L-10.687 (0.598-0.776)Absolute Δ PLT, G⋅L-10.502 (0.381-0.622)
      FU-PLT, G⋅L-10.674 (0.580-0.768)Relative Δ PLT, %0.525 (0.389-0.662)
      BL-VWF, %0.723 (0.635-0.811)Absolute Δ VWF, %0.551 (0.427-0.675)
      FU-VWF, %0.687 (0.577-0.798)Relative Δ VWF, %0.502 (0.380-0.625)
      BL-VITRO0.750 (0.673-0.827)Absolute Δ VITRO0.601 (0.478-0.724)
      FU-VITRO0.713 (0.613-0.813)Relative Δ VITRO, %0.503 (0.373-0.633)
      BL-AST0.567 (0.445-0.689)Absolute Δ AST0.516 (0.378-0.653)
      FU-AST0.631 (0.519-0.744)Relative Δ AST, %0.516 (0.387-0.645)
      BL-ALT0.501 (0.376-0.626)Absolute Δ ALT0.543 (0.409-0.676)
      FU-ALT0.598 (0.483-0.714)Relative Δ ALT, %0.586 (0.463-0.709)
      BL-APRI0.677 (0.560-0.794)Absolute Δ APRI0.600 (0.463-0.738)
      FU-APRI0.702 (0.594-0.809)Relative Δ APRI, %0.509 (0.376-0.642)
      BL-FIB-40.730 (0.648-0.812)Absolute Δ FIB-40.627 (0.517-0.737)
      FU-FIB-40.720 (0.627-0.813)Relative Δ FIB-4, %0.543 (0.421-0.665)
      BL-AFP0.720 (0.655-0.785)Absolute Δ AFP0.631 (0.526-0.737)
      FU-AFP0.796 (0.726-0.866)Relative Δ AFP, %0.536 (0.421-0.652)
      AFP, alpha-fetoprotein; ALT, alanine aminotransferase; APRI, AST-to-platelet ratio index; AST, aspartate aminotransferase; AUC, area under the curve; BL, baseline; FU, follow-up; FIB-4, fibrosis-4; LSM, liver stiffness measurement; PLT, platelet count; VITRO, von Willebrand factor antigen/platelet count ratio; VWF, von Willebrand factor.
      We aimed at identifying cut-offs that denote a high vs. low risk for HCC for the most promising parameters. Applying Youden’s J-statistics and bootstrap resampling, the following cut-offs were identified: Age ≥59.27 years, FU-albumin <42 g⋅L-1, FU-LSM ≥19.0 kPa, FU-PLT <190 G⋅L-1, FU-VWF ≥186%, FU-VITRO ≥1.02, FU-FIB-4 ≥1.93, and FU-AFP ≥4.6 ng⋅ml-1 (Table 3, Fig. S2). We abstained from further analyzing APRI, as it basically contains the same information as FIB-4, but FU-FIB-4 yielded a higher AUC.
      Table 3AUC values of pre-treatment and post-treatment parameters for predicting hepatocellular carcinoma development, and the respective Youden-optimized cut-offs (both without and after bootstrapping), in the derivation cohort.
      ParameterAUC (95%CI)Youden-optimized cut-offBootstrapped AUC
      Median AUCs and median Youden-optimized cut-offs from 5000 bootstrap samples are reported.
      Bootstrapped Youden-optimized cut-off
      Median AUCs and median Youden-optimized cut-offs from 5000 bootstrap samples are reported.
      Age, years0.664 (0.546-0.782)59.270.6759.24
      FU-albumin, g⋅L-10.714 (0.594-0.835)42.00.7142.0
      FU-LSM, kPa0.713 (0.621-0.805)19.00.7119.0
      FU-PLT, G⋅L-10.674 (0.580-0.768)198.50.67190
      FU-VWF, %0.687 (0.577-0.798)1860.69186
      FU-VITRO0.713 (0.613-0.813)0.950.711.02
      FU-FIB-40.720 (0.627-0.813)1.700.721.93
      FU-AFP, ng⋅ml-10.796 (0.726-0.866)4.60.804.6
      AFP, alpha-fetoprotein; AUC, area under the curve; BL, baseline; FU, follow-up; FIB-4, fibrosis-4; LSM, liver stiffness measurement; PLT, platelet count; VITRO, von Willebrand factor antigen/platelet count ratio; VWF, von Willebrand factor.
      1 Median AUCs and median Youden-optimized cut-offs from 5000 bootstrap samples are reported.

      Cox regression analyses and model estimation in the derivation cohort

      Next, we performed Cox regression analyses including dichotomized FU-values of non-invasive parameters and age as a central risk factor, since these values were superior to or equally accurate as BL values, and the utilization of data obtained at a single time point may facilitate the clinical application of the resulting risk prediction model (Table 4). We also included alcohol consumption above the threshold, while we did not include metabolic factors, since no statistically significant associations with HCC development were evident (Table S4).
      Table 4Cox regression analyses of risk factors for hepatocellular carcinoma development in the derivation cohort.
      Univariate analyses
      ParameterHazard ratio (95% CI)p value
      Calculated from Cox regression analyses.
      Age ≥59 years5.454 (1.835-16.210)0.002
      FU-albumin ≤42 g⋅L-15.642 (2.067-15.400)<0.001
      FU-PLT ≤190 G⋅L-19.619 (1.290-71.750)0.027
      FU-VWF ≥186%3.491 (1.354-9.005)0.010
      FU-VITRO ≥1.023.704 (1.244-11.030)0.019
      FU-AFP ≥4.6 ng⋅ml-117.130 (3.988-73.570)<0.001
      FU-LSM ≥19.0 kPa4.739 (1.964-11.440)<0.001
      FU-FIB-4 ≥1.934.535 (1.334-15.420)0.016
      Alcohol consumption above the threshold
      >30 g/day and >20 g/day for males and females, respectively.23
      3.106 (1.044-9.244)0.042
      Multivariate analyses
      ParameterBackward elimination – First stepBackward elimination – last step
      Stepwise exclusion of variables with p >0.100.
      Harrel’s C
      Calculated from last step of backward elimination using bootstrapped Harrel’s C-indices.
      Adjusted hazard ratio (95% CI)p value
      Calculated from Cox regression analyses.
      Adjusted hazard ratio (95% CI)p value
      Calculated from Cox regression analyses.
      Model 1
      FU-FIB-4 ≥1.933.341 (0.970-11.500)0.0593.341 (0.970-11.500)0.0590.720
      FU-albumin <42.04.639 (1.682-12.800)0.0034.639 (1.682-12.800)0.003
      Model 2
      FU-FIB-4 ≥1.931.629 (0.455-5.832)0.453--
      FU-albumin <42.03.236 (1.140-9.188)0.0274.012 (1.463-11.000)0.0070.841
      FU-VWF >1851.721 (0.645-4.952)0.278--
      FU-AFP ≥4.611.575 (2.616-51.220)0.00113.797 (3.191-59.650<0.001
      Model 3
      Age ≥594.164 (1.392-12.455)0.0114.107 (1.374-12.280)0.0110.859
      FU-albumin <42.03.321 (1.172-9.409)0.0243.789 (1.379-10.410)0.010
      FU-VITRO ≥1.021.715 (0.549-5.358)0.354--
      FU-AFP ≥4.610.555 (2.389-46.625)0.00211.713 (2.694-50.930)0.001
      Model 4
      Age ≥596.500 (2.168-19.490)<0.0016.500 (2.168-19.490)<0.0010.815
      FU-albumin <42.04.068 (1.464-11.300)0.0084.068 (1.464-11.300)0.008
      FU-LSM ≥19.04.644 (1.886-11.430)<0.0014.644 (1.886-11.430)<0.001
      Model 5
      Age ≥594.864 (1.609-14.707)0.0054.792 (1.893-38.804)0.0050.874
      FU-albumin <42.02.983 (1.052-8.457)0.0403.124 (1.115-8.754)0.030
      FU-VITRO ≥1.021.342 (0.420-4.288)0.620--
      FU-AFP ≥4.68.251 (1.811-37.594)0.0068.571 (1.893-38.804)0.005
      FU-LSM ≥19.02.517 (0.978-6.480)0.0562.666 (1.058-6.719)0.038
      Model 6
      Age ≥599.203 (2.846-29.770)<0.0019.203 (2.846-29.770)<0.0010.836
      FU-albumin <42.03.865 (1.394-10.720)0.0093.865 (1.394-10.720)0.009
      FU-LSM ≥19.05.285 (2.103-13.280)<0.0015.285 (2.103-13.280)<0.001
      Alcohol consumption above the threshold5.252 (1.631-16.910)0.0055.252 (1.631-16.910)0.005
      Model 7
      Age ≥595.496 (1.781-16.959)0.0035.496 (1.781-16.959)0.0030.893
      FU-albumin <42.03.229 (1.165-8.945)0.0253.229 (1.165-8.945)0.025
      FU-AFP ≥4.69.806 (2.161-44.502)0.0039.806 (2.161-44.502)0.003
      FU-LSM ≥19.03.150 (1.240-8.000)0.0163.150 (1.240-8.000)0.016
      Alcohol consumption above the threshold6.758 (2.152-21.227)0.0016.758 (2.152-21.227)0.001
      Values in bold indicate p ≤0.05. Parameters have been dichotomized according to the Youden’s index-optimized cut-offs (see Table 3). AFP, alpha-fetoprotein; FU, follow-up; FIB-4, fibrosis-4; LSM, liver stiffness measurement; PLT, platelet count; VITRO, von Willebrand factor antigen/platelet count ratio; VWF, von Willebrand factor.
      1 >30 g/day and >20 g/day for males and females, respectively.
      European Association for the Study of the LEuropean Association for the Study of DEuropean Association for the Study of O
      EASL-EASD-EASO Clinical Practice Guidelines for the management of non-alcoholic fatty liver disease.
      2 Calculated from Cox regression analyses.
      3 Stepwise exclusion of variables with p >0.100.
      4 Calculated from last step of backward elimination using bootstrapped Harrel’s C-indices.
      Following significant univariable associations with HCC development, 7 different multivariable models were built based on combinations of these variables. All of them accurately predicted HCC, however, FU-FIB-4, FU-VWF, and FU-VITRO were not independently associated with HCC development. Following backward elimination, the models comprising age ≥59 years, FU-AFP ≥4.6 ng⋅ml-1, FU-LSM ≥19 kPa, and FU-albumin <42.0 g⋅L-1 with and without alcohol consumption above the threshold showed the highest predictive ability (Harrel’s C: 0.893 and 0.874), while the same models without FU-AFP ≥4.6 ng⋅ml-1 also showed a high discriminative ability (Harrel’s C: 0.836 and 0.815).

      Competing risk analysis and modelling of a score

      To test whether the parameters included in the final version of models 5 and 7 (including alcohol) provided independent information for the prediction of HCC, while accounting for OLT and death as competing risks, we performed a competing risk regression analysis (Table 5). Importantly, all parameters were independently associated with HCC development during FU.
      Table 5Adjusted SHR of dichotomized risk factors for hepatocellular carcinoma development during FU in the derivation cohort using competing risk analysis classifying liver transplantation and death as competing risks.
      ParameterAdjusted SHR (95% CI)p value
      Calculated from Fine and Gray competing risks regression.
      Model 5
      Age ≥59 years4.76 (1.72-13.18)0.003
      FU-albumin ≤42.0 g⋅L-13.12 (1.08-8.98)0.035
      FU-AFP ≥4.6 ng⋅ml-18.85 (1.77-44.22)0.008
      FU-LSM ≥19.0 kPa2.64 (1.04-6.71)0.042
      Model 7
      Age ≥59 years5.55 (2.25-13.68)<0.001
      FU-albumin ≤42.0 g⋅L-13.21 (1.11-9.25)0.031
      FU-AFP ≥4.6 ng⋅ml-19.94 (2.45-40.40)0.001
      FU-LSM ≥19.0 kPa3.15 (1.14-8.67)0.026
      Alcohol consumption above the threshold6.70 (1.79-25.05)0.005
      Values in bold indicate p ≤0.05. AFP, alpha-fetoprotein; FU, follow-up; LSM, liver stiffness measurement; SHR, subdistribution hazard ratio.
      1 Calculated from Fine and Gray competing risks regression.
      A simple score was derived from adjusted (subdistribution) hazard ratios assigning 3 points for FU-AFP ≥4.6 ng⋅ml-1, 2 points for age ≥59 years, 2 points for alcohol consumption above the threshold, 1 point for FU-LSM ≥19 kPa, and 1 point for FU-albumin <42 g⋅L-1 (0 points were assigned if the respective criterion was not met). Following this approach, the derivation cohort was stratified according to the number of assigned points (Fig. S3). The subdistribution hazard ratio (SHR) was 2.47 (95% CI 1.91-3.19, p <0.001) per point. Patients were then stratified into low-risk (0-3 points, n = 308 [65.8%]) and high-risk (4-9 points, n = 160 [34.2%]) groups. Of note, 61.4% of the low-risk group had a BL- or FU-LSM value ≥12.5 kPa,
      • Castéra L.
      • Vergniol J.
      • Foucher J.
      • Le Bail B.
      • Chanteloup E.
      • Haaser M.
      • et al.
      Prospective comparison of transient elastography, Fibrotest, APRI, and liver biopsy for the assessment of fibrosis in chronic hepatitis C.
      and thus, showed evidence of cirrhosis. This dichotomization identified patients at very low and substantial risk of HCC at 4 years: 0% vs. 16.5% (Fig. 1A; HCC incidence rate per 100 patient-years: 0 vs. 4.3).
      Figure thumbnail gr1
      Fig. 1Cumulative incidence of de novo hepatocellular carcinoma development in cACLD patients in the derivation cohort.
      (A) Cumulative incidence curves (using competing risk analysis) of de novo hepatocellular carcinoma development of the post-treatment (FU) AFP/age/alcohol/LSM/albumin-derived strata (low-risk [0-3 points] vs. high-risk [4-9 points]) in patients with cACLD in the derivation cohort. 3 points are assigned for FU-AFP ≥4.6 ng⋅ml-1, 2 points for alcohol consumption above the threshold, 2 points for age ≥59 years, 1 point for FU-LSM ≥19 kPa, and 1 point for FU-albumin <42 g⋅L-1 (0 points if the respective criterion is not met). (B) Similar analysis based on the age/alcohol/LSM/albumin-derived score (i.e., without FU-AFP). Cumulative incidences are displayed according to low risk (0-3 points) and high risk (4-9 points) assignment. 3 points are assigned for age ≥59 years, 2 points for alcohol consumption above the threshold, 2 points for FU-LSM ≥19 kPa, and 2 points for FU-albumin <42 g⋅L-1 (0 points if the respective criterion is not met). AFP, alpha-fetoprotein; cACLD, compensated advanced chronic liver disease; FU, follow-up; LSM, liver stiffness measurement; SHR, subdistribution hazard ratio.
      Since AFP is not routinely assessed at many centers, we also tested whether a simple score derived from the adjusted (subdistribution) hazard ratios of model 6 (i.e., without AFP) was also able to stratify HCC risk. We assigned 3 points for age ≥59 years, 2 points for alcohol consumption above the threshold, 2 points for FU-albumin <42 g⋅L-1, and 2 points for FU-LSM ≥19 kPa and stratified patients into low-risk (0-3 points, n = 322 [68.8%]) and high-risk (4-9 points, n = 146 [31.2%]) groups. The HCC risk at 4 years was 1.3% vs. 14.8% (SHR 13.70; 95% CI 4.02-46.40; p <0.001; Fig. 1B; HCC incidence per 100 patient-years: 0.3 vs. 3.9).
      Finally, both approaches also yielded a high discriminative ability without including alcohol consumption above the threshold as a variable, thereby acknowledging uncertainties regarding the quantification of alcohol consumption (AFP-based algorithm at 4 years: 0.5% vs. 16.7% [SHR 21.90; 95% CI 5.10-94.00; p <0.001]; non-AFP-based algorithm: 1.8% vs. 15.0% [SHR 10.90; 95% CI 3.67-32.40; p <0.001]; Fig. S4AB).

      External validation of proposed risk scores

      In an attempt to externally validate these findings, we tested these 4 scores in an independent validation cohort comprising patients from Madrid, Barcelona, and another Viennese center (validation cohort). Overall, 691 patients were included in the validation cohort for the FU-AFP-based algorithm, while 1,500 patients were included in the validation cohort for the algorithm without FU-AFP. Despite small differences existing for age and FU-LSM, disease severity, FU time and HCC incidence were comparable (Table 1). In the validation cohort, FU measurements clustered around 48 weeks after EoT (Fig. S1B).
      As depicted in Fig. 2, both approaches including alcohol consumption as a risk factor efficiently stratified the risk of HCC during FU, with a probability of 3.3% vs. 17.5% developing HCC within 4 years according to the AFP-based algorithm (SHR 5.11, 95% CI 2.54-10.30, p <0.001; HCC incidence per 100 patient-years 0.9 vs. 4.4) and 3.7% vs. 11.6% developing HCC within 4 years according to the algorithm without AFP (SHR 3.46; 95% CI 2.05-5.84; p <0.001; HCC incidence per 100 patient-years 0.9 vs. 3.0). Comparable results were achieved without considering alcohol consumption as a risk factor (Table S4C,D).
      Figure thumbnail gr2
      Fig. 2Cumulative incidence of de novo hepatocellular carcinoma development in patients with cACLD in the validation cohort. (A) Cumulative incidence curves (using competing risk analysis) of de novo hepatocellular carcinoma development of the post-treatment (FU) AFP/age/alcohol/LSM/albumin-derived strata (low-risk [0-3 points] vs. high-risk [4-9 points]) in patients with cACLD in the validation cohort. 3 points are assigned for FU-AFP ≥4.6 ng⋅ml-1, 2 points for alcohol consumption above the threshold, 2 points for age ≥59 years, 1 point for FU-LSM ≥19 kPa, and 1 point for FU-albumin <42 g⋅L-1 (0 points if the respective criterion is not met). (B) Similar analysis based on the age/alcohol/LSM/albumin-derived score (i.e., without FU-AFP). Cumulative incidences are displayed according to low risk (0-3 points) and high risk (4-9 points) assignment. 3 points are assigned for age ≥59 years, 2 points for alcohol consumption above the threshold, 2 points for FU-LSM ≥19 kPa, and 2 points for FU-albumin <42 g⋅L-1 (0 points if the respective criterion is not met). AFP, alpha-fetoprotein; cACLD, compensated advanced chronic liver disease; FU, follow-up; LSM, liver stiffness measurement; SHR, subdistribution hazard ratio.

      Sensitivity analysis stratifying patients according to time between EoT and FU-LSM

      Both scores considering alcohol maintained an adequate discriminative ability when combining the derivation and validation cohort and stratifying patients into tertiles of the time between EoT and FU-LSM (Fig. S5).

      Discussion

      In the present study, we investigated predictive factors for HCC development in patients with cACLD who achieved SVR on IFN-free therapies. We focused on patients with ACLD, since HCC incidence has been reported to be significantly higher in patients with ACLD/advanced liver fibrosis or cirrhosis
      • Carrat F.
      • Fontaine H.
      • Dorival C.
      • Simony M.
      • Diallo A.
      • Hezode C.
      • et al.
      Clinical outcomes in patients with chronic hepatitis C after direct-acting antiviral treatment: a prospective cohort study.
      ,
      • Kanwal F.
      • Kramer J.
      • Asch S.M.
      • Chayanupatkul M.
      • Cao Y.
      • El-Serag H.B.
      Risk of hepatocellular cancer in HCV patients treated with direct-acting antiviral agents.
      ,
      • Kanwal F.
      • Kramer J.R.
      • Asch S.M.
      • Cao Y.
      • Li L.
      • El-Serag H.B.
      Long-term risk of hepatocellular carcinoma in HCV patients treated with direct acting antiviral agents.
      (and this is also reflected by current European surveillance recommendations
      EASL recommendations on treatment of Hepatitis C 2020.
      ). We provided an easily applicable score that facilitates risk stratification in clinical routine, as it identified patients in whom HCC surveillance may not be cost-effective (low-risk) or in whom surveillance is clearly warranted (high-risk; i.e., AFP ≥4.6 ng⋅ml-1 OR age ≥59 years WITH either FU-LSM ≥19 kPa AND/OR FU-albumin <42 g⋅L-1) due to a considerable probability of de novo HCC despite SVR. Importantly – and in contrast to most previous attempts – our proposed algorithms underwent extensive external validation, which is critical due to the profound implications of a delayed diagnosis of HCC that may result from an unwarranted termination of surveillance due to an underestimation of HCC risk. The analysis of the multicenter validation cohort confirmed that approximately two-thirds of patients (i.e., those who do not meet the aforementioned high-risk criteria) are classified as low-risk and that these patients exhibit an HCC risk <1%/year. Accordingly, the incidence of HCC in these patients clearly falls below the cost-effectiveness threshold (at 50,000 USD/quality-adjusted life year) for HCC surveillance, which has been estimated at 1.32/year.
      • Farhang Zangneh H.
      • Wong W.W.L.
      • Sander B.
      • Bell C.M.
      • Mumtaz K.
      • Kowgier M.
      • et al.
      Cost effectiveness of hepatocellular carcinoma surveillance after a sustained virologic response to therapy in patients with hepatitis C virus infection and advanced fibrosis.
      ,
      • Ioannou G.N.
      HCC surveillance after SVR in patients with F3/F4 fibrosis.
      Of note, this low-risk group also included a large proportion (61.4%) of patients with evidence of cirrhosis, indicating that current recommendations to identify at-risk patients who should undergo ultrasound surveillance have very limited accuracy.
      Our approach has important advantages that may promote its application in the clinic. First, a ‘one-time’ assessment after treatment (e.g., around 12 weeks after EoT or up 48 weeks after EoT) is easily applicable in clinical routine, since patients can be stratified according to their individual risk of HCC while confirming SVR. In addition, HCC risk stratification can be combined with the evaluation of the probability of hepatic decompensation by additionally assessing FU-PLT and FU-VWF to calculate the FU-VITRO score.
      • Semmler G.
      • Binter T.
      • Kozbial K.
      • Schwabl P.
      • Hametner-Schreil S.
      • Zanetto A.
      • et al.
      Noninvasive risk stratification after HCV eradication in patients with advanced chronic liver disease.
      In contrast, if approaches rely on BL values, they cannot be applied later in case of incomplete pre-treatment work-up or unavailable information (due to changes in treatment center), leaving these patients unclassified. Similarly, consideration of absolute/relative changes seems particularly problematic, as it doubles the number of required variables, and thus, the chance of missing information.
      Secondly, our approach combines indicators of liver fibrosis and portal hypertension (i.e., LSM
      • Mandorfer M.
      • Hernández-Gea V.
      • García-Pagán J.C.
      • Reiberger T.
      Noninvasive diagnostics for portal hypertension: a comprehensive review.
      ) and hepatic dysfunction (i.e. serum albumin) with age (a strong driver of carcinogenesis in general
      • Laconi E.
      • Marongiu F.
      • DeGregori J.
      Cancer as a disease of old age: changing mutational and microenvironmental landscapes.
      ) and AFP – a broadly available biomarker that is commonly applied for HCC surveillance in clinical practice,
      • Galle P.R.
      • Foerster F.
      • Kudo M.
      • Chan S.L.
      • Llovet J.M.
      • Qin S.
      • et al.
      Biology and significance of alpha-fetoprotein in hepatocellular carcinoma.
      which is obligatory according to Asian Pacific Association for The Study of Liver, optional according to the American Association for the Study of Liver Disease, and not recommended due to concerns about cost-effectiveness by the European Association for the Study of the Liver (EASL) clinical practice guidelines.
      EASL Clinical Practice Guidelines
      Management of hepatocellular carcinoma.
      However, the latter clinical practice guidelines also emphasize that the use of AFP should be re-evaluated in patients who achieved etiological cure, as it may perform better after the amelioration of hepatic inflammation.
      EASL Clinical Practice Guidelines
      Management of hepatocellular carcinoma.
      Interestingly, AFP showed the highest individual AUC for HCC development during FU and was considered as a binary variable in our risk prediction model at a cut-off of ≥4.6 ng⋅ml-1. Interestingly, this AFP cut-off is considerably lower than the cut-offs proposed for HCC surveillance that were either 20 or 200 ng⋅ml-1.
      • Omata M.
      • Cheng A.L.
      • Kokudo N.
      • Kudo M.
      • Lee J.M.
      • Jia J.
      • et al.
      Asia-Pacific clinical practice guidelines on the management of hepatocellular carcinoma: a 2017 update.
      However, AFP usually decreases with HCV cure (e.g., -2.5 ng⋅ml-1 or -41.3% in our study) and the proposed application of AFP (i.e., for risk stratification) differs from its common use as a biomarker for HCC surveillance (i.e., diagnosis of [very] early-stage HCC).
      Considerable evidence supports the use of LSM for predicting HCC risk in patients who achieved SVR, and which was recently summarized by a meta-analysis.
      • You M.W.
      • Kim K.W.
      • Shim J.J.
      • Pyo J.
      Impact of liver-stiffness measurement on hepatocellular carcinoma development in chronic hepatitis C patients treated with direct-acting antivirals: a systematic review and time-to-event meta-analysis.
      However, specific cut-offs for identifying patients at relevantly increased risk varied substantially according to the studied population, ranging from ≥20 kPa
      • Ogasawara N.
      • Saitoh S.
      • Akuta N.
      • Sezaki H.
      • Suzuki F.
      • Fujiyama S.
      • et al.
      Advantage of liver stiffness measurement before and after direct-acting antiviral therapy to predict hepatocellular carcinoma and exacerbation of esophageal varices in chronic hepatitis C.
      and >21.5 kPa
      • Conti F.
      • Buonfiglioli F.
      • Scuteri A.
      • Crespi C.
      • Bolondi L.
      • Caraceni P.
      • et al.
      Early occurrence and recurrence of hepatocellular carcinoma in HCV-related cirrhosis treated with direct-acting antivirals.
      to >30 kPa
      • Degasperi E.
      • D'Ambrosio R.
      • Iavarone M.
      • Sangiovanni A.
      • Aghemo A.
      • Soffredini R.
      • et al.
      Factors associated with increased risk of de novo or recurrent hepatocellular carcinoma in patients with cirrhosis treated with direct-acting antivirals for HCV infection.
      ,
      • Rinaldi L.
      • Guarino M.
      • Perrella A.
      • Pafundi P.C.
      • Valente G.
      • Fontanella L.
      • et al.
      Role of liver stiffness measurement in predicting HCC occurrence in direct-acting antivirals setting: a real-life experience.
      pre-treatment and ≥10 kPa post-treatment.
      • Pons M.
      • Rodríguez-Tajes S.
      • Esteban J.I.
      • Mariño Z.
      • Vargas V.
      • Lens S.
      • et al.
      Non-invasive prediction of liver-related events in patients with HCV-associated compensated advanced chronic liver disease after oral antivirals.
      ,
      • Ogasawara N.
      • Saitoh S.
      • Akuta N.
      • Sezaki H.
      • Suzuki F.
      • Fujiyama S.
      • et al.
      Advantage of liver stiffness measurement before and after direct-acting antiviral therapy to predict hepatocellular carcinoma and exacerbation of esophageal varices in chronic hepatitis C.
      Although several studies proposed that absolute and relative changes in LSM are related to HCC development,
      • Ravaioli F.
      • Conti F.
      • Brillanti S.
      • Andreone P.
      • Mazzella G.
      • Buonfiglioli F.
      • et al.
      Hepatocellular carcinoma risk assessment by the measurement of liver stiffness variations in HCV cirrhotics treated with direct acting antivirals.
      they showed (if at all) modest prognostic value in our series of patients. Moreover, when compared to the non-invasive diagnosis of CSPH and prediction of hepatic decompensation, the predictive ability of LSM for HCC seems inferior,
      • Semmler G.
      • Binter T.
      • Kozbial K.
      • Schwabl P.
      • Hametner-Schreil S.
      • Zanetto A.
      • et al.
      Noninvasive risk stratification after HCV eradication in patients with advanced chronic liver disease.
      which argues for the consideration of additional variables in order to increase prognostic accuracy.
      The combination of high LSM and AFP with the traditional risk factors such as old age and low serum albumin might optimize previous approaches: these were often established in studies that were based on less thoroughly characterized patient cohorts, and were therefore unable to establish synergistic effects between these variables.
      • Pons M.
      • Rodríguez-Tajes S.
      • Esteban J.I.
      • Mariño Z.
      • Vargas V.
      • Lens S.
      • et al.
      Non-invasive prediction of liver-related events in patients with HCV-associated compensated advanced chronic liver disease after oral antivirals.
      ,
      • Alonso S.
      • Manzano M.L.
      • Gea F.
      • Gutiérrez M.L.
      • Ahumada A.M.
      • Devesa M.J.
      • et al.
      A model based on non-invasive markers predicts very low hepatocellular carcinoma risk after viral response in HCV-advanced fibrosis.
      ,
      • Tani J.
      • Morishita A.
      • Sakamoto T.
      • Takuma K.
      • Nakahara M.
      • Fujita K.
      • et al.
      Simple scoring system for prediction of hepatocellular carcinoma occurrence after hepatitis C virus eradication by direct-acting antiviral treatment: all Kagawa Liver Disease Group Study.
      ,
      • Calvaruso V.
      • Cabibbo G.
      • Cacciola I.
      • Petta S.
      • Madonia S.
      • Bellia A.
      • et al.
      Incidence of hepatocellular carcinoma in patients with HCV-associated cirrhosis treated with direct-acting antiviral agents.
      Importantly, we have also included alcohol consumption above the threshold as a (modifiable) risk factor. Alcohol consumption has previously been associated with the development of HCC after SVR
      • Kanwal F.
      • Kramer J.
      • Asch S.M.
      • Chayanupatkul M.
      • Cao Y.
      • El-Serag H.B.
      Risk of hepatocellular cancer in HCV patients treated with direct-acting antiviral agents.
      ,
      • Minami T.
      • Tateishi R.
      • Fujiwara N.
      • Nakagomi R.
      • Nakatsuka T.
      • Sato M.
      • et al.
      Impact of obesity and heavy alcohol consumption on hepatocellular carcinoma development after HCV eradication with antivirals.
      • Ganne-Carrié N.
      • Layese R.
      • Bourcier V.
      • Cagnot C.
      • Marcellin P.
      • Guyader D.
      • et al.
      Nomogram for individualized prediction of hepatocellular carcinoma occurrence in hepatitis C virus cirrhosis (ANRS CO12 CirVir).
      • Vandenbulcke H.
      • Moreno C.
      • Colle I.
      • Knebel J.-F.
      • Francque S.
      • Sersté T.
      • et al.
      Alcohol intake increases the risk of HCC in hepatitis C virus-related compensated cirrhosis: a prospective study.
      and – according to the Baveno VII recommendations – prohibits the discharge of patients with cACLD who achieved SVR from portal hypertension surveillance. The consideration of alcohol consumption highlights the importance of this co-factor for progressive liver disease after SVR,
      • Semmler G.
      • Binter T.
      • Kozbial K.
      • Schwabl P.
      • Hametner-Schreil S.
      • Zanetto A.
      • et al.
      Noninvasive risk stratification after HCV eradication in patients with advanced chronic liver disease.
      even resulting in increased liver-related mortality,
      • Alavi M.
      • Law M.G.
      • Valerio H.
      • Grebely J.
      • Amin J.
      • Hajarizadeh B.
      • et al.
      Declining hepatitis C virus-related liver disease burden in the direct-acting antiviral therapy era in New South Wales, Australia.
      which may raise awareness both for physicians and patients. Fully acknowledging uncertainties regarding the quantification of alcohol consumption, we have confirmed that our risk scores perform appropriately and do not require any modifications, even if alcohol consumption is not considered. Although diabetes and metabolic comorbidities have been discussed as other potential risk factors for HCC in the post-SVR context, we did not observe such a significant association.
      • Minami T.
      • Tateishi R.
      • Fujiwara N.
      • Nakagomi R.
      • Nakatsuka T.
      • Sato M.
      • et al.
      Impact of obesity and heavy alcohol consumption on hepatocellular carcinoma development after HCV eradication with antivirals.
      Of note, our score does not include composite variables such as VITRO (which had a higher AUC than its individual components, VWF and PLT) or FIB-4, since none of these scores was predictive of HCC, when also considering other variables. Currently, the EASL (2020) clinical practice guidelines for the treatment of hepatitis C do not recommend a personalized surveillance strategy,
      EASL recommendations on treatment of Hepatitis C 2020.
      and thus, a 6-monthly ultrasound surveillance is recommended in all patients with pre-treatment advanced liver fibrosis (F3) or cirrhosis (F4). Importantly, this approach might not be cost-effective, especially not in patients who only have advanced liver fibrosis (F3) pre-treatment.
      • Farhang Zangneh H.
      • Wong W.W.L.
      • Sander B.
      • Bell C.M.
      • Mumtaz K.
      • Kowgier M.
      • et al.
      Cost effectiveness of hepatocellular carcinoma surveillance after a sustained virologic response to therapy in patients with hepatitis C virus infection and advanced fibrosis.
      A recent analysis estimated that the number of HCC surveillance candidates with SVR will increase more than 6-fold from 2012 to 2030.
      • Chen Q.
      • Ayer T.
      • Adee M.G.
      • Wang X.
      • Kanwal F.
      • Chhatwal J.
      Assessment of incidence of and surveillance burden for hepatocellular carcinoma among patients with hepatitis C in the era of direct-acting antiviral agents.
      Therefore, personalized surveillance strategies are urgently needed to optimize resource utilization and these surveillance strategies should be based on a comprehensive evaluation of de novo HCC risk – such as our proposed algorithms – rather than the pre-treatment liver fibrosis stage.
      • Ioannou G.N.
      HCC surveillance after SVR in patients with F3/F4 fibrosis.
      Since a late diagnosis of HCC has serious implications for the outcome of an individual patient, extensive external validation is mandatory, before risk stratification approaches are applied in the clinic to identify low-risk patients in whom HCC surveillance can be deferred. Therefore, we also evaluated previously proposed approaches based on the cACLD subgroup of our derivation cohort and the validation cohort. In this context, several specific aspects of individual scoring/grading systems were notable, and these are extensively discussed in the supplementary information.
      Of note, the international multicenter design increases the generalizability of our findings. Since we only included specialized centers, we were able to acquire a large, comprehensively characterized derivation cohort that provided information on the vast majority of potential predictors of HCC development in patients with ACLD who achieved SVR. However, presumably the most important strength of our study is the external validation of our algorithms in up to 1,500 patients from different centers across Europe. Although there were statistically significant differences in patient characteristics, the differences between cohorts were very small (<10%) for variables considered in our prognostic models (age and FU-LSM). Moreover, slight variations between the derivation and validation dataset may even increase the generalizability of our models.
      The main limitation of our study is its retrospective design, which introduced considerable variability regarding the time point for the assessment of post-treatment data. These measurements were clustered around 12 weeks after EoT in our derivation cohort and around 48 weeks after EoT in the validation cohort. These differences are mainly due to the design of the contributing studies
      • Pons M.
      • Rodríguez-Tajes S.
      • Esteban J.I.
      • Mariño Z.
      • Vargas V.
      • Lens S.
      • et al.
      Non-invasive prediction of liver-related events in patients with HCV-associated compensated advanced chronic liver disease after oral antivirals.
      ,
      • Alonso S.
      • Manzano M.L.
      • Gea F.
      • Gutiérrez M.L.
      • Ahumada A.M.
      • Devesa M.J.
      • et al.
      A model based on non-invasive markers predicts very low hepatocellular carcinoma risk after viral response in HCV-advanced fibrosis.
      as well as limited patient compliance and capacity restrictions. However, the heterogeneity in the assessment time point may actually improve the robustness and generalizability of our risk stratification approach, as it showed an excellent discriminative ability both in the derivation and validation cohort, despite differences in the time point of assessment. In addition, sensitivity analyses using time point-dependent stratification revealed that the discriminative ability of our risk stratification approach was maintained at all assessment time points. Accordingly, the lack of standardization may also be seen as a strength of our study, as some variation in the time point of assessment will be unavoidable in ‘real-world’ clinical practice, and thus, risk stratification systems should show a robust performance under rather unstandardized conditions to ascertain external validity.
      Similar to the variability in the time point of post-treatment measurements, HCC surveillance prior to therapy and EoT was not standardized. However, all patients had at least one unsuspicious imaging after EoT. Currently, lifelong HCC surveillance is recommended since a similar HCC incidence over time has been reported after SVR to IFN-free
      • Kanwal F.
      • Kramer J.R.
      • Asch S.M.
      • Cao Y.
      • Li L.
      • El-Serag H.B.
      Long-term risk of hepatocellular carcinoma in HCV patients treated with direct acting antiviral agents.
      and IFN-based
      • Ioannou G.N.
      • Beste L.A.
      • Green P.K.
      • Singal A.G.
      • Tapper E.B.
      • Waljee A.K.
      • et al.
      Increased risk for hepatocellular carcinoma persists up to 10 Years after HCV eradication in patients with baseline cirrhosis or high FIB-4 scores.
      therapies. However, our study cannot provide information on the long-term risk of HCC, and this is an unavoidable limitation of all studies available to date. Accordingly, long-term studies are warranted: these should also address the question of whether a re-evaluation of the laboratory/elastography parameters at a later time point may refine risk stratification regarding events occurring during long-term FU. Since the vast majority of included patients were Caucasian (>90%), it remains to be shown whether our findings can be extrapolated to other ethnicities. Data from Asia will be important to confirm the accuracy of the proposed algorithms.
      In conclusion, based on our international multicenter study, we developed and externally validated simple algorithms for HCC prediction in patients with cACLD who achieved SVR to IFN-free treatments, comprising a set of broadly available parameters, which were all evaluated at a single post-treatment time point. Approximately two-thirds of patients were identified as having an HCC risk <1%/year, thereby clearly falling below the cost-effectiveness threshold for HCC surveillance.

      Abbreviations

      ACLD, advanced chronic liver disease; AFP alpha-fetoprotein; AUC, area under the curve; BL, baseline; cACLD, compensated ACLD; CHC chronic hepatitis C; CSPH, clinically significant portal hypertension; dACLD, decompensated ACLD; EoT, end of treatment; FIB-4, fibrosis-4; FU, follow-up; HCC, hepatocellular carcinoma; HVPG, hepatic venous pressure gradient; IFN, interferon; LSM, liver stiffness measurement; MELD, model of end-stage liver disease; OLT liver transplantation; PLT, platelet count; ROC, receiver-operating characteristic; SVR, sustained virologic response; VITRO, von Willebrand factor antigen/platelet count ratio; VWF, von Willebrand factor

      Financial support

      This work was supported by a grant from the Medical Scientific Fund of the Mayor of the City of Vienna (No. 17035) as well as the Andrew K. Burroughs short-term training fellowship of the European Association for the Study of the Liver .

      Authors’ contributions

      Study concept and design (M.M.), acquisition of data (all authors), analysis and interpretation of data (G.S., E.M., T.R., M.M.), drafting of the manuscript (G.S., T.R., M.M.), critical revision of the manuscript for important intellectual content (all authors), language editing (H.L.).

      Data availability statement

      Data are available from the corresponding author ( [email protected] ) upon reasonable request.

      Conflict of interest

      G.S. has nothing to disclose. E.M. received grants from Novartis. K.K. received travel support from AbbVie, Bristol-Myers Squibb, and Gilead. P.Sc. received consulting fees from PharmaIN, and travel support from Falk and Phenex Pharmaceuticals; S.H.-S. served as a speaker and/or consultant and/or advisory board member for AbbVie, Bristol-Myers Squibb, Eisai, Gilead, and Intercept and received travel support from AbbVie and Gilead; A.Z. has nothing to disclose; D.B. received travel support from AbbVie and Gilead; D.C. served as a speaker and/or consultant and/or advisory board member for AbbVie, Gilead, and MSD, and received travel support from AbbVie, MSD, ViiV Healthcare and Gilead; B.Si. received travel support from AbbVie and Gilead. B.Sch. received travel support from AbbVie, Ipsen and Gilead. A.F.S. served as a speaker and/or consultant and/or advisory board member for Boehringer Ingelheim, Gilead, and MSD. M.Pi. served as a speaker and/or consultant and/or advisory board member for Bayer, Bristol-Myers Squibb, Ipsen, Eisai, Lilly, MSD, and Roche, and received travel support from Bayer and Bristol-Myers Squibb. R.S. has nothing to disclose. F.P.R. served as a speaker and/or consultant and/or advisory board member for AbbVie, Biotest, Gilead, and MSD, and received travel support from AbbVie, Biotest, and Gilead. H.G. has nothing to disclose. M.S. received speaking honoraria from BMS and travel support from Bristol-Myers Squibb, AbbVie, and MSD. C.S. has nothing to disclose. M.G. received grants from AbbVie, Gilead, and MSD; speaking honoraria/advisory board fees from AbbVie, Gilead, MSD, Janssen, Roche, Intercept, Norgine, AstraZeneca, Falk, and Shionogi. S.A.L. served as a speaker and/or consultant and/or advisory board member for AbbVie, Gilead, and MSD and received grants from AbbVie and Gilead. M.L.M. has nothing to disclose. A.A. received consulting and speaker fees from Abbvie and Gilead. R.B. served as a speaker and/or consultant and/or advisory board member for AbbVie, Gilead, and Janssen. M.Po. has nothing to disclose. S.R.-T. has nothing to disclose. J.G. has nothing to disclose. S.L. received grants from Gilead, speaker and advisory fees from Gilead, Abbvie and MSD. M.T. served as a speaker and/or consultant and/or advisory board member for Albireo, Boehringer Ingelheim, Bristol-Myers Squibb, Falk, Genfit, Gilead, Intercept, MSD, Novartis, Phenex, Regulus and Shire, and received travel support from AbbVie, Falk, Gilead, and Intercept, as well as grants/research support from Albireo, Cymabay, Falk, Gilead, Intercept, MSD, and Takeda. He is also co-inventor of patents on the medical use of 24-norursodeoxycholic acid. P.F. has served as a speaker and/or consultant and/or advisory board member for AbbVie, Bristol Myer-Squibb, Gilead, and MSD and has received research funding from Gilead. T.R. served as a speaker and/or consultant and/or advisory board member for AbbVie, Bayer, Boehringer Ingelheim, Gilead, Intercept, MSD, Siemens, and W. L. Gore & Associates and received grants/research support from AbbVie, Boehringer Ingelheim, Gilead, MSD, Philips, and W. L. Gore & Associates as well as travel support from Boehringer Ingelheim and Gilead. M.M. served as a speaker and/or consultant and/or advisory board member for AbbVie, Bristol-Myers Squibb, Gilead, Collective Acumen, and W. L. Gore & Associates and received travel support from AbbVie, Bristol-Myers Squibb, and Gilead.
      Please refer to the accompanying ICMJE disclosure forms for further details.

      Supplementary data

      The following are the supplementary data to this article:

      References

        • Mandorfer M.
        • Kozbial K.
        • Freissmuth C.
        • Schwabl P.
        • Stattermayer A.F.
        • Reiberger T.
        • et al.
        Interferon-free regimens for chronic hepatitis C overcome the effects of portal hypertension on virological responses.
        Aliment Pharmacol Ther. 2015; 42: 707-718
        • Lens S.
        • Alvarado-Tapias E.
        • Marino Z.
        • Londono M.C.
        • Elba L.L.
        • Martinez J.
        • et al.
        Effects of all-oral anti-viral therapy on HVPG and systemic hemodynamics in patients with hepatitis C virus-associated cirrhosis.
        Gastroenterology. 2017; 153: 1273-1283.e1271
        • Lens S.
        • Baiges A.
        • Alvarado E.
        • Llop E.
        • Martinez J.
        • Fortea J.I.
        • et al.
        Clinical outcome and hemodynamic changes following HCV eradication with oral antiviral therapy in patients with clinically significant portal hypertension.
        J Hepatol. 2020;
        • Mandorfer M.
        • Kozbial K.
        • Schwabl P.
        • Freissmuth C.
        • Schwarzer R.
        • Stern R.
        • et al.
        Sustained virologic response to interferon-free therapies ameliorates HCV-induced portal hypertension.
        J Hepatol. 2016; 65: 692-699
        • Mandorfer M.
        • Kozbial K.
        • Schwabl P.
        • Chromy D.
        • Semmler G.
        • Stättermayer A.F.
        • et al.
        Changes in hepatic venous pressure gradient predict hepatic decompensation in patients who achieved sustained virologic response to interferon-free therapy.
        Hepatology (Baltimore, Md). 2020; 71: 1023-1036
        • Mauro E.
        • Crespo G.
        • Montironi C.
        • Londoño M.C.
        • Hernández-Gea V.
        • Ruiz P.
        • et al.
        Portal pressure and liver stiffness measurements in the prediction of fibrosis regression after sustained virological response in recurrent hepatitis C.
        Hepatology (Baltimore, Md). 2018; 67: 1683-1694
        • Backus L.I.
        • Belperio P.S.
        • Shahoumian T.A.
        • Mole L.A.
        Impact of sustained virologic response with direct-acting antiviral treatment on mortality in patients with advanced liver disease.
        Hepatology (Baltimore, Md). 2019; 69: 487-497
        • Carrat F.
        • Fontaine H.
        • Dorival C.
        • Simony M.
        • Diallo A.
        • Hezode C.
        • et al.
        Clinical outcomes in patients with chronic hepatitis C after direct-acting antiviral treatment: a prospective cohort study.
        The Lancet. 2019; 393: 1453-1464
        • Nahon P.
        • Bourcier V.
        • Layese R.
        • Audureau E.
        • Cagnot C.
        • Marcellin P.
        • et al.
        Eradication of hepatitis C virus infection in patients with cirrhosis reduces risk of liver and non-liver complications.
        Gastroenterology. 2017; 152: 142-156.e142
        • Pons M.
        • Rodríguez-Tajes S.
        • Esteban J.I.
        • Mariño Z.
        • Vargas V.
        • Lens S.
        • et al.
        Non-invasive prediction of liver-related events in patients with HCV-associated compensated advanced chronic liver disease after oral antivirals.
        J Hepatol. 2020; 72: 472-480
        • Semmler G.
        • Binter T.
        • Kozbial K.
        • Schwabl P.
        • Hametner-Schreil S.
        • Zanetto A.
        • et al.
        Noninvasive risk stratification after HCV eradication in patients with advanced chronic liver disease.
        Hepatology (Baltimore, Md). 2021; 73: 1275-1289
        • Kanwal F.
        • Kramer J.
        • Asch S.M.
        • Chayanupatkul M.
        • Cao Y.
        • El-Serag H.B.
        Risk of hepatocellular cancer in HCV patients treated with direct-acting antiviral agents.
        Gastroenterology. 2017; 153: 996-1005.e1001
        • Ioannou G.N.
        • Beste L.A.
        • Green P.K.
        • Singal A.G.
        • Tapper E.B.
        • Waljee A.K.
        • et al.
        Increased risk for hepatocellular carcinoma persists up to 10 Years after HCV eradication in patients with baseline cirrhosis or high FIB-4 scores.
        Gastroenterology. 2019; 157: 1264-1278.e1264
        • Finkelmeier F.
        • Dultz G.
        • Peiffer K.H.
        • Kronenberger B.
        • Krauss F.
        • Zeuzem S.
        • et al.
        Risk of de novo Hepatocellular Carcinoma after HCV Treatment with Direct-Acting Antivirals.
        Liver Cancer. 2018; 7: 190-204
        • Ripoll C.
        • Groszmann R.J.
        • Garcia-Tsao G.
        • Bosch J.
        • Grace N.
        • Burroughs A.
        • et al.
        Hepatic venous pressure gradient predicts development of hepatocellular carcinoma independently of severity of cirrhosis.
        J Hepatol. 2009; 50: 923-928
        • Chromy D.
        • Mandorfer M.
        • Bucsics T.
        • Schwabl P.
        • Bauer D.
        • Scheiner B.
        • et al.
        Prevalence and predictors of hepatic steatosis in patients with HIV/HCV coinfection and the impact of HCV eradication.
        AIDS patient care and STDs. 2019; 33: 197-206
        • Farhang Zangneh H.
        • Wong W.W.L.
        • Sander B.
        • Bell C.M.
        • Mumtaz K.
        • Kowgier M.
        • et al.
        Cost effectiveness of hepatocellular carcinoma surveillance after a sustained virologic response to therapy in patients with hepatitis C virus infection and advanced fibrosis.
        Clin Gastroenterol Hepatol: Off Clin Pract J Am Gastroenterol Assoc. 2019; 17: 1840-1849.e1816
        • de Franchis R.
        • Baveno V.I.F.
        Expanding consensus in portal hypertension: report of the Baveno VI Consensus Workshop: stratifying risk and individualizing care for portal hypertension.
        J Hepatol. 2015; 63: 743-752
        • EASL Clinical Practice Guidelines
        Management of hepatocellular carcinoma.
        J Hepatol. 2018; 69: 182-236
        • Schwarzer R.
        • Reiberger T.
        • Mandorfer M.
        • Kivaranovic D.
        • Hametner S.
        • Hametner S.
        • et al.
        The von Willebrand Factor antigen to platelet ratio (VITRO) score predicts hepatic decompensation and mortality in cirrhosis.
        J Gastroenterol. 2019;
        • Semmler G.
        • Binter T.
        • Kozbial K.
        • Schwabl P.
        • Chromy D.
        • Bauer D.
        • et al.
        Influence of genetic variants on disease regression and outcomes in HCV-related advanced chronic liver disease after SVR.
        J Personalized Med. 2021; 11: 281
        • Russo F.P.
        • Zanetto A.
        • Campello E.
        • Bulato C.
        • Shalaby S.
        • Spiezia L.
        • et al.
        Reversal of hypercoagulability in patients with HCV-related cirrhosis after treatment with direct-acting antivirals.
        Liver Int Off J Int Assoc Stud Liver. 2018; 38: 2210-2218
        • European Association for the Study of the L
        • European Association for the Study of D
        • European Association for the Study of O
        EASL-EASD-EASO Clinical Practice Guidelines for the management of non-alcoholic fatty liver disease.
        J Hepatol. 2016; 64: 1388-1402
        • Maieron A.
        • Salzl P.
        • Peck-Radosavljevic M.
        • Trauner M.
        • Hametner S.
        • Schofl R.
        • et al.
        Von Willebrand Factor as a new marker for non-invasive assessment of liver fibrosis and cirrhosis in patients with chronic hepatitis C.
        Aliment Pharmacol Ther. 2014; 39: 331-338
        • Sarrazin C.
        • Berg T.
        • Buggisch P.
        • Dollinger M.M.
        • Hinrichsen H.
        • Hofer H.
        • et al.
        S3 guideline hepatitis C addendum.
        Zeitschrift fur Gastroenterologie. 2015; 53: 320-334
        • Sarrazin C.
        • Zimmermann T.
        • Berg T.
        • Neumann U.P.
        • Schirmacher P.
        • Schmidt H.
        • et al.
        [Prophylaxis, diagnosis and therapy of hepatitis-C-virus (HCV) infection: the German guidelines on the management of HCV infection - AWMF-Register-No.: 021/012].
        Zeitschrift fur Gastroenterologie. 2018; 56: 756-838
      1. EASL recommendations on treatment of hepatitis C 2015.
        J Hepatol. 2015; 63: 199-236
      2. EASL recommendations on treatment of hepatitis C 2016.
        J Hepatol. 2017; 66: 153-194
      3. EASL recommendations on treatment of hepatitis C 2018.
        J Hepatol. 2018; 69: 461-511
      4. EASL-EORTC clinical practice guidelines: management of hepatocellular carcinoma.
        J Hepatol. 2012; 56: 908-943
        • Alonso S.
        • Manzano M.L.
        • Gea F.
        • Gutiérrez M.L.
        • Ahumada A.M.
        • Devesa M.J.
        • et al.
        A model based on non-invasive markers predicts very low hepatocellular carcinoma risk after viral response in HCV-advanced fibrosis.
        Hepatology (Baltimore, Md). 2020;
        • Fine J.P.
        • Gray R.J.
        A proportional hazards model for the subdistribution of a competing risk.
        J Am Stat Assoc. 1999; 94: 496-509
        • Castéra L.
        • Vergniol J.
        • Foucher J.
        • Le Bail B.
        • Chanteloup E.
        • Haaser M.
        • et al.
        Prospective comparison of transient elastography, Fibrotest, APRI, and liver biopsy for the assessment of fibrosis in chronic hepatitis C.
        Gastroenterology. 2005; 128: 343-350
        • Kanwal F.
        • Kramer J.R.
        • Asch S.M.
        • Cao Y.
        • Li L.
        • El-Serag H.B.
        Long-term risk of hepatocellular carcinoma in HCV patients treated with direct acting antiviral agents.
        Hepatology (Baltimore, Md). 2020; 71: 44-55
      5. EASL recommendations on treatment of Hepatitis C 2020.
        J Hepatol. 2020;
        • Ioannou G.N.
        HCC surveillance after SVR in patients with F3/F4 fibrosis.
        J Hepatol. 2021; 74: 458-465
        • Mandorfer M.
        • Hernández-Gea V.
        • García-Pagán J.C.
        • Reiberger T.
        Noninvasive diagnostics for portal hypertension: a comprehensive review.
        Semin Liver Dis. 2020;
        • Laconi E.
        • Marongiu F.
        • DeGregori J.
        Cancer as a disease of old age: changing mutational and microenvironmental landscapes.
        Br J Cancer. 2020; 122: 943-952
        • Galle P.R.
        • Foerster F.
        • Kudo M.
        • Chan S.L.
        • Llovet J.M.
        • Qin S.
        • et al.
        Biology and significance of alpha-fetoprotein in hepatocellular carcinoma.
        Liver Int Off J Int Assoc Stud Liver. 2019; 39: 2214-2229
        • Omata M.
        • Cheng A.L.
        • Kokudo N.
        • Kudo M.
        • Lee J.M.
        • Jia J.
        • et al.
        Asia-Pacific clinical practice guidelines on the management of hepatocellular carcinoma: a 2017 update.
        Hepatol Int. 2017; 11: 317-370
        • You M.W.
        • Kim K.W.
        • Shim J.J.
        • Pyo J.
        Impact of liver-stiffness measurement on hepatocellular carcinoma development in chronic hepatitis C patients treated with direct-acting antivirals: a systematic review and time-to-event meta-analysis.
        J Gastroenterol Hepatol. 2020;
        • Ogasawara N.
        • Saitoh S.
        • Akuta N.
        • Sezaki H.
        • Suzuki F.
        • Fujiyama S.
        • et al.
        Advantage of liver stiffness measurement before and after direct-acting antiviral therapy to predict hepatocellular carcinoma and exacerbation of esophageal varices in chronic hepatitis C.
        Hepatol Res Off J Jpn Soc Hepatol. 2019;
        • Conti F.
        • Buonfiglioli F.
        • Scuteri A.
        • Crespi C.
        • Bolondi L.
        • Caraceni P.
        • et al.
        Early occurrence and recurrence of hepatocellular carcinoma in HCV-related cirrhosis treated with direct-acting antivirals.
        J Hepatol. 2016; 65: 727-733
        • Degasperi E.
        • D'Ambrosio R.
        • Iavarone M.
        • Sangiovanni A.
        • Aghemo A.
        • Soffredini R.
        • et al.
        Factors associated with increased risk of de novo or recurrent hepatocellular carcinoma in patients with cirrhosis treated with direct-acting antivirals for HCV infection.
        Clin Gastroenterol Hepatol: Off Clin Pract J Am Gastroenterol Assoc. 2019; 17: 1183-1191.e1187
        • Rinaldi L.
        • Guarino M.
        • Perrella A.
        • Pafundi P.C.
        • Valente G.
        • Fontanella L.
        • et al.
        Role of liver stiffness measurement in predicting HCC occurrence in direct-acting antivirals setting: a real-life experience.
        Dig Dis Sci. 2019; 64: 3013-3019
        • Ravaioli F.
        • Conti F.
        • Brillanti S.
        • Andreone P.
        • Mazzella G.
        • Buonfiglioli F.
        • et al.
        Hepatocellular carcinoma risk assessment by the measurement of liver stiffness variations in HCV cirrhotics treated with direct acting antivirals.
        Dig Liver Dis: Off J Ital Soc Gastroenterol Ital Assoc Stud Liver. 2018; 50: 573-579
        • Tani J.
        • Morishita A.
        • Sakamoto T.
        • Takuma K.
        • Nakahara M.
        • Fujita K.
        • et al.
        Simple scoring system for prediction of hepatocellular carcinoma occurrence after hepatitis C virus eradication by direct-acting antiviral treatment: all Kagawa Liver Disease Group Study.
        Oncol Lett. 2020; 19: 2205-2212
        • Calvaruso V.
        • Cabibbo G.
        • Cacciola I.
        • Petta S.
        • Madonia S.
        • Bellia A.
        • et al.
        Incidence of hepatocellular carcinoma in patients with HCV-associated cirrhosis treated with direct-acting antiviral agents.
        Gastroenterology. 2018; 155: 411-421.e414
        • Minami T.
        • Tateishi R.
        • Fujiwara N.
        • Nakagomi R.
        • Nakatsuka T.
        • Sato M.
        • et al.
        Impact of obesity and heavy alcohol consumption on hepatocellular carcinoma development after HCV eradication with antivirals.
        Liver Cancer. 2021; 10: 309-319
        • Ganne-Carrié N.
        • Layese R.
        • Bourcier V.
        • Cagnot C.
        • Marcellin P.
        • Guyader D.
        • et al.
        Nomogram for individualized prediction of hepatocellular carcinoma occurrence in hepatitis C virus cirrhosis (ANRS CO12 CirVir).
        Hepatology (Baltimore, Md). 2016; 64: 1136-1147
        • Vandenbulcke H.
        • Moreno C.
        • Colle I.
        • Knebel J.-F.
        • Francque S.
        • Sersté T.
        • et al.
        Alcohol intake increases the risk of HCC in hepatitis C virus-related compensated cirrhosis: a prospective study.
        J Hepatol. 2016; 65: 543-551
        • Alavi M.
        • Law M.G.
        • Valerio H.
        • Grebely J.
        • Amin J.
        • Hajarizadeh B.
        • et al.
        Declining hepatitis C virus-related liver disease burden in the direct-acting antiviral therapy era in New South Wales, Australia.
        J Hepatol. 2019; 71: 281-288
        • Chen Q.
        • Ayer T.
        • Adee M.G.
        • Wang X.
        • Kanwal F.
        • Chhatwal J.
        Assessment of incidence of and surveillance burden for hepatocellular carcinoma among patients with hepatitis C in the era of direct-acting antiviral agents.
        JAMA Netw open. 2020; 3 (e2021173-e2021173)