Intra- and Interoperator Reproducibility of Acoustic Radiation Force Impulse (ARFI) Elastography–Preliminary Results

Abstract

Our study assessed acoustic radiation force impulse (ARFI) reproducibility and the factors influencing it. The intra- and interoperator reproducibility were studied in 33 and 58 patients, respectively. Intraclass correlation coefficient (ICC) was used to assess ARFI reproducibility. The overall intraoperator agreement was better than the interoperator one: ICC 0.90 vs. ICC 0.81. The correlation of repeated ARFI measurements was higher, but not significantly so, in cases in which intraoperator reproducibility was assessed compared with the ones in which interoperator reproducibility was studied: r = 0.848 vs. r = 0.694 (p = 0.08). For both intra- and interoperator reproducibility, the ICCs were smaller in women vs. men (0.88 vs. 0.91 and 0.67 vs. 0.86, respectively), in patients with high body mass index (BMI) ≥25 kg/m² vs. <25 kg/m² (0.88 vs. 0.91 and 0.79 vs. 0.82, respectively), in patients with ascites vs. no ascites (0.80 vs. 0.93 and 0.78 vs. 0.84, respectively) and in noncirrhotic vs. cirrhotic patients (0.77 vs. 0.82 and 0.70 vs. 0.83, respectively).

Introduction

Evaluation of liver fibrosis is important for the staging and prognosis of chronic hepatopathies. Liver biopsy (LB) is still considered the “gold standard” for liver fibrosis assessment (Perrillo, 1997, Saadeh et al., 2001), despite being an invasive method, not totally risk free (Rockey et al. 2009) and despite the fact that a significant sampling error can occur due to the small volume of the tissue sample (only approximately 1/50,000 of the liver volume) (Afdhal 2006). Moreover, some studies showed that the needle type and the fibrosis stage can also influence LB results (Sherman et al. 2007). In addition, inter- and/or intraobserver diagnostic discrepancies are estimated to affect up to 10%–20% of LBs (Bedossa and Poynard 1994).

For these reasons, several noninvasive methods for liver fibrosis assessment were developed in the last 10–15 years, to replace LB, at least partially. To be usable in clinical practice, they should be reproducible and accurate.

Serologic tests were the first ones developed, especially to evaluate liver lesions’ severity in patients with chronic hepatitis C (Rosenberg et al., 2004, Poynard et al., 2002). Transient elastography (TE) (FibroScan®) followed an ultrasound-based elastographic method for liver fibrosis assessment. It was validated first in patients with chronic hepatitis C, and later, in chronic hepatitis B, nonalcoholic steatohepatitis, in post transplant patients or in primary biliary cirrhosis (Castéra et al., 2005, Sporea et al., 2011, Rigamonti et al., 2008, Cardoso et al., 2012, Lupsor et al., 2010, Friedrich-Rust et al., 2010). A meta-analysis published in 2011 (Tsochatzis et al. 2011), showed good sensitivity (Se) and specificity (Sp) for TE in the diagnosis of significant fibrosis (F ≥ 2, according to Metavir scoring system) and of liver cirrhosis. Regarding reproducibility, Fraquelli et al. (2007) observed a very good interoperator agreement for TE. In a cohort of 200 patients, the intraclass correlation coefficient (ICC) was 0.98. In the same study, a reduced ICC was significantly associated with increased body mass index (BMI) (≥25 kg/m²), steatosis (>24% of fatty liver cells) and low staging grades (F0-1).

In the last years, two real-time elastographic methods (integrated into an ultrasound system) were developed: Real-time elastography (HI-RTE) (Tatsumi et al., 2008, Friedrich-Rust et al., 2007, Havre et al., 2008) and acoustic radiation force impulse (ARFI) elastography (Lupsor et al., 2009, Friedrich-Rust et al., 2009, Sporea et al., 2011, Piscaglia et al., 2011, Rizzo et al., 2011). A new real-time ultrasound-based elastographic method, supersonic shear wave elastography, was studied for the noninvasive evaluation of liver fibrosis in 2010–2011 (Bavu et al. 2011). An advantage of real-time elastographic methods, compared with TE, is the fact than they can be used in patients with ascites.

ARFI and supersonic shear wave elastography use shear waves, which are generated and propagate into the tissue, providing a quantitative evaluation of liver stiffness (LS), while in the case of HI-RTE tissue elasticity distribution is calculated by starting from the strain and stress induced into the examined tissue. HI-RTE does not provide a quantitative evaluation of LS, the results being represented as color-coded images superimposed over conventional B-mode images.

Similar to TE, ARFI elastography has a very good value in predicting severe fibrosis and liver cirrhosis (Friedrich-Rust et al., 2009, Fierbinteanu-Braticevici et al., 2009, Lupsor et al., 2009, Sporea et al., 2011, Rizzo et al., 2011). Most published studies show that the best ARFI cut-off values for the diagnosis of liver cirrhosis range between 1.8 m/s (Sporea et al., 2011, Sporea et al., 2011, Ebinuma et al., 2011) and 2 m/s (Lupsor et al., 2009, Rizzo et al., 2011).

The aim of this study was to assess the intra- and interoperator reproducibility of ARFI measurements and the factors that influence it.