2D-shear wave elastography is equivalent or superior to transient elastography for liver fibrosis assessment: Results from an individual patient data based meta-analysis
Publication in policy or professional journal

Times Cited
Web of Science6WOS source URL (as at 08/05/2021) Click here for the latest count
Altmetrics Information

Other information
AbstractBackground and Aims: 2D shear wave elastography (2D-SWE) based on supersonic shear imaging (SSI) has proven to be efficient for the evaluation of liver fibrosis in several small to moderate size trials. We aimed at running a larger scale analysis of individual data.
Methods: Clinical data from 1340 patients with 2D-SWE measurement and liver biopsy were collected retrospectively from 13 sites and analysed. Additionally, data on transient elastography (TE) were available in a subsample of 972 patients. The database was cleared by the French National Commission on Information Technology and Liberties and the study was registered on clinicaltrials.gov. The data were analysed using appropriate random effect models for ROC analysis as well as paired comparison of AUROC.
Results: Main etiologies were chronic hepatitis C (HCV, n = 470), hepatitis B (HBV, n = 420), non-alcoholic fatty liver disease (NAFLD, n = 172) or other liver diseases (n = 278). There was high heterogeneity between sites for liver disease etiologies and fibrosis stages. 40.8% of the patients had minimal or no fibrosis, 19.3% had significant fibrosis, 14.0% had severe fibrosis and 26.0% had cirrhosis.
Overall performance of 2D-SWE assessed by AUROC in patients with HCV, HBV and NAFLD was 86.3%, 91.6%, 85.9% for diagnosing significant fibrosis and 96.1%, 97.1% and 95.5% for diagnosing cirrhosis, respectively. Optimal cut-offs were 7.1 kPa for diagnosing significant fibrosis in all patients (75.7% correctly classified), 13.5 kPa for diagnosing cirrhosis in HCV and NAFLD patients, and 11.5 kPa for diagnosing cirrhosis in HBV patients (87% correctly classified).
Differences in AUROC were borderline significant for diagnosing significant fibrosis (95% CI for AUROC-2D-SWE minus AUROC-TE: [0.0004, 0.055], p = 0.047) and AUROC was significantly higher for 2D-SWE when diagnosing cirrhosis (95% CI for AUROC-2D-SWE minus AUROC-TE: [0.006, 0.036], p = 0.0058). 2D-SWE was superior for diagnosing cirrhosis specifically in HCV patients (AUROC difference: 0.018; p = 0.046) and HBV patients (AUROC difference: 0.062; p = 0.015). It was superior for diagnosing significant fibrosis in HBV patients (AUROC difference: 0.102; p = 0.0005) and severe fibrosis in NAFLD patients (AUROC difference: 0.152; p = 0.002).
Conclusions: 2D-SWE based on SSI showed good to excellent performance for non-invasive assessment of liver fibrosis, and was equivalent or superior to transient elastography.
All Author(s) ListE. Herrmann, V. de Lédinghen, C. Cassinotto, W.C.-W. Chu, V.Y.-F. Leung, G. Ferraioli, C. Filice, L. Castera, V. Vilgrain, M. Ronot, J. Dumortier, A. Guibal, S. Pol, J. Trebicka, C. Jansen, C. Strassburg, R. Zheng, J. Zheng, S. Francque, T. Vanwolleghem, L. Vonghia, E.K. Manesis, P. Zoumpoulis, I. Sporea, M. Thiele, A. Krag, M. Friedrich-Rust
Journal nameJournal of Hepatology
Volume Number62
Issue NumberSuppl. 2
PagesS199 - S200
LanguagesEnglish-United Kingdom

Last updated on 2021-09-05 at 01:25