The effects of fatty deposits on the accuracy of the Fibroscan® liver transient elastography ultrasound system

Abstract

A new generation of ultrasound transient elastography (TE) systems have emerged which exploit the well-known correlation between the liver's pathological and mechanical properties through measurements of the Young's elastic modulus; however, little work has been carried out to examine the effect that fatty deposits may have on the TE measurement accuracy. An investigation was carried out on the effects on the measurement accuracy of a TE ultrasound system, the Fibroscan®, caused by overlaying fat layers of varying thickness on healthy liver-mimicking phantoms, simulating in vivo conditions for obese patients. Furthermore, a steatosis effect similar to that in non-alcoholic fatty liver disease (NAFLD) and alcoholic liver disease (ALD) was simulated to investigate its effect on the TE system. A range of novel elastography fat-mimicking materials were developed using 6–10 wt% poly(vinyl alcohol) cryogel capable of achieving a range of acoustic velocities (1482–1530 m s−1) and attenuation coefficients (0.4–1 dB MHz–1 cm–1) for simulating different liver states. Laboratory-based acoustic velocities and attenuation coefficients were measured while the Young's modulus was established through a gold standard compression testing method. A significant variation of the Young's elastic modulus was measured in healthy phantoms with overlaying fat layers of thicknesses exceeding 45 mm, impinging on the scanners region of interest, overestimating the compression tested values by up to 11 kPa in some cases. Furthermore, Fibroscan® measurements of the steatosis phantoms showed a consistent overestimation (∼54%), which strongly suggests that the speed of sound mismatch between that of liver tissue and that assumed by the scanner is responsible for the high clinical cut-offs established in the case of ALD and NAFLD.