Radiological Society of North America/Quantitative Imaging Biomarker Alliance Shear Wave Speed Bias Quantification in Elastic and Viscoelastic Phantoms

Abstract

Objectives: To quantify the bias of shear wave speed (SWS) measurements between different commercial ultrasonic shear elasticity systems and a magnetic resonance elastography (MRE) system in elastic and viscoelastic phantoms.

Methods: Two elastic phantoms, representing healthy through fibrotic liver, were measured with 5 different ultrasound platforms, and 3 viscoelastic phantoms, representing healthy through fibrotic liver tissue, were measured with 12 different ultrasound platforms. Measurements were performed with different systems at different sites, at 3 focal depths, and with different appraisers. The SWS bias across the systems was quantified as a function of the system, site, focal depth, and appraiser. A single MRE research system was also used to characterize these phantoms using discrete frequencies from 60 to 500 Hz.

Results: The SWS from different systems had mean difference 95% confidence intervals of ±0.145 m/s (±9.6%) across both elastic phantoms and ± 0.340 m/s (±15.3%) across the viscoelastic phantoms. The focal depth and appraiser were less significant sources of SWS variability than the system and site. Magnetic resonance elastography best matched the ultrasonic SWS in the viscoelastic phantoms using a 140 Hz source but had a - 0.27 ± 0.027-m/s (-12.2% ± 1.2%) bias when using the clinically implemented 60-Hz vibration source.

Conclusions: Shear wave speed reconstruction across different manufacturer systems is more consistent in elastic than viscoelastic phantoms, with a mean difference bias of < ±10% in all cases. Magnetic resonance elastographic measurements in the elastic and viscoelastic phantoms best match the ultrasound systems with a 140-Hz excitation but have a significant negative bias operating at 60 Hz. This study establishes a foundation for meaningful comparison of SWS measurements made with different platforms.

Keywords: Quantitative Imaging Biomarker Alliance; acoustic radiation force; elasticity; phantom; shear wave; viscoelasticity.

© 2021 American Institute of Ultrasound in Medicine.

Figures

Figure 1.

Calibration measurements on all of the softer (E1786) and stiffer (E1787) elastic ultrasound phantoms and the phantom set designated for comparison to MRE measurements (E1788) using a research scanner sequence at 3 different focal depths (40 [blue], 60 [red], and 80 [green] mm). The dashed orange line in each plot represents the grand mean of all measurements made in the ultrasound phantoms for each plot: 0.907 ± 0.033 (3.7%) m/s and 2.025 ± 0.051 (2.5%) m/s for the soft and stiff phantoms, respectively. A given phantom set’s mean difference from these grand means was used as a corrective factor to normalize for this fabrication variability between different phantom pairs.

Figure 2.

Aggregate SWS data in the soft (blue) and stiff (green) elastic phantoms measured at different sites, where some sites had multiple systems available for measurement. Each system at each site was used by 3 appraisers who made 10 replicate measurements at each of the focal depths (30, 45, and 70 mm) in each phantom. In some cases (sites D, E, F, J, K, L, and M), coarser quantization (rounding to the nearest 0.1 m/s) of the reported SWS by some or all of the site systems is apparent.

Figure 3.

All of the elastic phantom data grouped by unique system. Some systems were used at only a single measurement site, whereas other systems were used at multiple measurement sites. Note that a single system (B) appears to report SWS with coarser quantization (0.1 m/s) compared to the other systems.

Figure 4.

Tukey mean difference plots for the aggregated phase I systems (top) and sites (bottom), using the normalization data (Figure 1) as the reference measurement for each phantom. The colors in each plot represent the same system/site, respectively. Note that system/site biases are not necessarily consistent across the soft and stiff phantoms (eg, a system with a negative bias in the soft phantom may have a positive bias in the stiff phantom).

Figure 5.

Group and phase SWS measurements in one pair of the phase I elastic phantoms made by using the Verasonics research scanner sequences and processing code at a focal depth of 45 mm, derived from shear wave displacement (Disp) and from shear wave velocity (Vel). The circles in each plot represent the mean of 10 independent acquisitions, and the error bars represent the 95% CI for each measurement. Magnetic resonance elastographic measurements were made at discrete frequencies of 140, 180, 200, 300, 400, and 500 Hz. The slopes of linear fits to these phase velocities, which are indicative of undesired dispersion (frequency-dependent phase velocity) in these elastic phantoms, are summarized in Table 4.

Figure 6.

Phase II phantoms measured with different systems with 3 different focal depth configurations (30, 45, and 70 mm). The orange line on each plot represents the grand mean value across all systems for each phantom.

Figure 7.

Tukey mean difference plots for the aggregated phase II systems using the data from the calibration Verasonics system as the reference measurements for each phantom. The colors in each plot represent the same system. Note that system biases are not necessarily consistent across the different phantoms (eg, a system with a negative bias in one phantom may have a positive bias another phantom).

Figure 8.

Group and phase velocities calculated in the 3 phase II viscoelastic phantoms that were distributed to all of the measurement sites. The error bars represent the 95% CI over 16 independent measurements. As expected, these viscoelastic phantoms have higher group SWS estimated when by using velocity (Vel) data instead of displacement (Disp) data (left plots). This same trend is seen in the positive slope of the corresponding phase velocity curves (right plots). In the phase velocity plots, note that the frequency range of the reconstructed phase velocities increases as a function of increasing stiffness, and the variance of the estimated phase velocity increases at higher frequencies because of the lower signal-to-noise ratio at these higher frequencies. The slopes of the linear-fit phase velocity lines are summarized in Table 4.

Figure 9.

Comparison of group SWS calculated with displacement and velocity data in the phase I and II phantoms compared to equivalent processing of in vivo human data at varying fibrosis stages. The dashed line represents a unity ratio between the velocity- and displacement-based group SWSs that would be indicative of an elastic material; data points above this line would indicate a dispersive material. In the phase II phantoms, the group SWS calculated by using velocity data was 32% ± 1.9% greater than by using displacement data, whereas in the human data, the velocity-based group SWS was 27% ± 5.6% greater than the displacement-based group SWS across all fibrosis stages.

Figure 10.

Violin distributions of aggregate ultrasound SWS data across all systems and sites at a focal depth of 45 mm for each phase II phantom compared to discrete MRE measurements made at frequencies ranging from 60 to 200 Hz. The black box within each violin plot represents the interquartile range of the data, with the white circle representing the median value. Vertical lines extend away from each violin distribution to represent 1.5× the standard deviation of the data. The surrounding shape represents the probability density of the data.

Figure 11.

Measurements showing the longitudinal stability of the phase II phantoms using the group SWS calculated by using displacement and velocity data as representative metrics. The error bars represent the standard deviation over 16 independent measurements.