Artifacts in liver stiffness evaluation

Overview of artifacts in liver shear wave elastography

The use of shear wave elastography techniques proves valuable in providing qualitative and quantitative elasticity maps of the liver. There are, however, common disturbances that can affect the tissue measurements of stiffness. Image artifacts, signals that appear present on the image but are not present in the body, can result in overestimation, underestimation, and highly variable calculations of tissue stiffness maps. Several artifacts can influence stiffness measurements for most techniques, most notably ultrasound-based techniques. These include:

  • Poor acoustic windows that result in both limited push and tracking penetration and cause high variability in elasticity measurement and rib or lung shadowing, which produce further instabilities in elasticity values. Poor signal-to-noise ratio in the shear wave data or the resultant shear wave speed reconstructions is commonly reported on commercial systems through quality metrics and images.

  • Cardiac pulsatility artifacts that result from both the creation of natural shear waves from a high heart rate and the presence of pulsating vessels near the region of interest.

  • Liver boundary, or Glisson capsule, artifacts due to interactions with portal vessels and the distortion of shear wave motion around the liver capsule.

  • Tissue viscosity that causes both vibration-frequency dependent shear wave speeds, especially when comparing measurements between modalities, and overestimation bias due to reconstruction assumptions of pure elasticity.

  • Tissue nonlinearity such that stiffness changes nonlinearly with strain due to extracellular matrix heterogeneity and is underestimated when calculated with linear assumptions.

  • Motion artifacts due to patient motion and natural motion from breathing.

  • Shear wave reflection artifacts due to stiffness gradients within the liver.

  • Nonperpendicular probe placement that results in nonlinear signal decay and liver stiffness overestimation.

  • Compression with the probe/nonlinear tissue response.

Vibration controlled transient elastography

Vibration controlled transient elastography (VCTE) relies upon external, mechanical vibration to induce shear waves within tissue. The shear waves are tracked via Doppler-like ultrasound techniques from which Young’s modulus can be calculated, assuming homogenous, isotropic, and elastic behavior. There are, however, limitations with VCTE and difficulties in interpretation. For example, without controlled probe compression force, the excitation can distort and decrease in center-frequency, resulting in lower stiffness measurements. Excessive compression can also lead to nonlinear stiffening of the compressed tissues. Similarly, without proper damping of the mechanical vibration, duplicate shear waves may be induced, and regression algorithms may misrepresent actual tissue elasticity. These artifacts that were observed with the first version of the FibroScan are less evident in the subsequent versions by controlling the vibration.

Acoustic radiation force impulse techniques artifacts

Both general ultrasound limitations and limitations of acoustic radiation force impulse (ARFI) techniques result in tissue stiffness variability estimations. The most common sources of artifacts in ARFI are a poor acoustic window ( Fig. 6.1 ), limited penetration ( Fig. 6.2 ), organ shadows ( Fig. 6.3 ), vessel pulsation ( Fig. 6.4 ), loss of signal ( Fig. 6.5 ), reverberations within the liver ( Figs. 6.6 and 6.7 ), and motion artifacts ( Fig. 6.8 ). Proper placement of the region of interest within the liver can reduce many potential artifacts. Additionally, ARFI is more susceptible to shear wave dispersion effects than fixed excitation frequency methods such as VCTE and magnetic resonance elastography (MRE); in a viscoelastic medium, shear wave velocity is dispersive, meaning velocities vary with frequency, and ARFI shear wave speed estimations can be biased based on the specific imaging system used.

Fig. 6.1

Example of a poor acoustic window. Shadowing is present on both the right and left side of the B-mode image. Of the elasticity map (on the right) , 50% has no coloring due to the shadowing. Note that the liver capsule is not a bright white line, suggesting that the transducer is also not parallel to the liver capsule. The confidence map on the left confirms poor quality of the elasticity map. The measurement box was placed in an area with medium-high confidence on the confidence map. However, with the poor acoustic window, the measurement should not be used.

Fig. 6.2

Two-dimensional shear wave elastography measurement in an obese patient. Signal dropout due to limited penetration. There is a thick subcutaneous tissue (>3 cm) that decreases the strength of the push-pulse, leading to a low signal-to-noise in the estimate of tissue stiffness. For this reason, a large part of the elasticity map is void of color. Note that the region of interest (ROI) has been positioned 1.5 cm below the Glisson capsule, as per guidelines; however, the center of the ROI is already at 5.5 cm from the transducer: this is not an optimal location because the maximum strength of the ultrasound is at 4–4.5 cm from the transducer. Tips: in cases like this one, the ROI can be moved closer to the liver capsule, making sure that reverberation artifacts are not present in the proximal portion of the ROI. However, it should be acknowledged that there is a rate of failures in the liver stiffness measurement.

Fig. 6.3

Image taken in the supine position at rib level where lung artifact and rib artifact are present (A). Stiffness measurement in the same patient position but moving down two rib levels (B). Note that, when correctly taken, the liver stiffness is 3.6 kPa (B), whereas, taken incorrectly (A), the liver stiffness is 10.9 kPa.

Fig. 6.4

In this image the liver capsule is a bright white line and almost parallel to the top of the elasticity map. Note that there is no coloring around the portal vessel, which was included in the field of view. This is due to both vessel pulsations and reverberation of the shear waves off the vessel wall. The small color-coded (blue and green) area on the left side of the elasticity map is mainly due to artifacts generated by the vessel pulsation. The confidence map (on the left side) detects the majority of artifacts, showing them in red and yellow , except the one due to vessel pulsation on the bottom left side of the field of view (green) . The measurement in this area is incorrect and leads to an overestimation of liver stiffness.

Fig. 6.5

Although the acoustic radiation force impulse pulse generates the shear waves, it is the B-mode imaging that tracks the displacements caused by the shear waves and uses that information to estimate the shear wave speed. In this case the B-mode image is of poor quality with low signal intensity leading to poor quality liver stiffness measurement.

Mar 12, 2023 | Posted by in ULTRASONOGRAPHY | Comments Off on Artifacts in liver stiffness evaluation

Full access? Get Clinical Tree

Get Clinical Tree app for offline access