Ultraschall Med 2013; 34 - WS_SL5_02
DOI: 10.1055/s-0033-1354827

Intraobserver and interobserver variability in the estimation of liver fibrosis with ShearWave elastography in chronic liver disease: A proposition for standardization through a modified protocol

PS Zoumpoulis 1, I Vafiadi 2, L Kaklamanis 3, I Theotokas 4, P Eleni 5
  • 1Hellenic Society of Ultrasound in Medicine and Biology, Kifissia, Greece
  • 2Laiko University Hospital, Athens, Greece
  • 3Onassio Hospital, Athens, Greece
  • 4Diagnostic Echotomography, Athens, Greece
  • 5Hellenic Society of Ultrasound in Medicine and Biology, Athens, Greece

Purpose: To estimate Intra-Inter Observer Variability (Ia/IeOVar) of measurements of liver fibrosis using ShearWave Elastography (SWE) and to propose efficient ways for an appropriate estimation of fibrosis in cirrhotic patients.

Material and methods: We studied 112 cirrhotic patients using SWE. All patients had liver biopsy performed in period not exceeding ± 4 months from US examination. A Transient Elastography TE (Fibroscan) examination has been performed to 82 of those patients in a period not exceeding ± 1 month from US/SWE examination.

All patients had US examination and SWE of the liver performed by two independent examiners. Each examiner performed a third measurement of SWE using a modified protocol in order to equalize the differences between neighbouring liver segments. The findings were compared for fibrous liver staging (F-stage) which was pathologically classified using liver biopsy.

The diagnostic performance of SWE and a TE (Fibroscan) was estimated comparing with F-stage diagnosis of liver biopsy.

Results: We observed important Ia/IeOVar differences in SWE measurements, calculated in kPa, during the first part of the study, using the initial protocol.

To reduce Ia/IeOVar: -We choose a SWE- representative liver ROI giving the possibility to the examiner for an arbitrary traced SWE ROI; six elastography images and analysed with a prototype analysis software calculating mean of relative strain value, standard deviation of relative strain value, ratio of blue area in the analysed region, kurtosis-skewness of strain histogram and entropy. Better correlation of the SWE measurement in all stages of histologically proven liver fibrosis: F1/F2/F3/F4 with revised protocol.

Conclusion: TE and SWE of the liver are effective and reliable ways to estimate liver stiffness and fibrosis and have good correlation with histologically proven F-stage. We propose a modified, tissue specific and adapted to the liver anatomic features SWE protocol in order to limit Ia/IeOVar.