Giovanna Ferraioli1, Laura Maiocchi1, Raffaella Lissandrin1, Carmine Tinelli2, Annalisa De Silvestri2, Carlo Filice1, and the Liver Fibrosis Study Group*
1) Ultrasound Unit, Infectious Diseases Department, Fondazione IRCCS Policlinico S. Matteo, Medical School University of Pavia;
2) Clinical Epidemiology and Biometric Unit, Fondazione IRCCS Policlinico S. Matteo, Pavia, Italy
ABSTRACT Background & Aims: Noninvasive assessment of liver stiffness has been increasingly used to evaluate fibrosis instead of liver biopsy, especially in patients with chronic viral hepatitis. The aim of this study was to assess the performance in staging liver fibrosis of the updated ElastPQ® technique (EPIQ7 ultrasound system, Philips Healthcare, Bothell, WA, USA) in the “real life” setting by using the FibroScan as the reference standard and to understand whether the use of the quality criteria improves the performance of the technique.
Methods: This was a cross-sectional study: 278 patients affected by chronic hepatitis C referred for liver stiffness measurement with the FibroScan® 502 Touch device (Echosens, Paris, France) underwent measurements also with the ElastPQ® technique. For the assessment of significant fibrosis (F≥2), advanced fibrosis (F≥3) and cirrhosis (F=4), respectively, we used the cutoffs of 7.0, 9.5 and 12.0 kPa. The diagnostic performance of ElastPQ® was assessed using the area under the ROC (AUROC) curve analysis and was evaluated overall and for cases with (a) 10 measurements and IQR/M≤30%, (b) 5 measurements and IQR/M ≤30%, (c) 10 measurements and IQR/M>30%, (d) 5 measurements and IQR/M>30%.
Results: The optimal cutoffs of ElastPQ® for significant fibrosis, advanced fibrosis and cirrhosis were 6.43, 9.54 and 11.34 kPa, respectively. For measurements with an IQR/M≤30%, there was no statistically significant decrease in sensitivity between 10 and 5 measurements (p=0.26, p=0.09, p=0.71, for F≥2, F≥3, and F=4, respectively).
Conclusion: The ElastPQ® technique is reliable and accurate for staging liver fibrosis. The number of measurements does not affect the performance.