Gabriela Torres

Temporal artifact minimization in sonoelastography through optimal selection of imaging parameters

Sonoelastography is an ultrasonic technique that uses Kasais autocorrelation algorithms to generate qualitative images of tissue elasticity using external mechanical vibrations. In the absence of synchronization between the mechanical vibration device and the ultrasound system, the random initial phase and finite ensemble length of the data packets result in temporal artifacts in the sonoelastography frames and, consequently, in degraded image quality. In this work, the analytic derivation of an optimal selection of acquisition parameters (i.e., pulse repetition frequency, vibration frequency, and ensemble length) is developed in order to minimize these artifacts, thereby eliminating the need for complex device synchronization. The proposed rule was verified through experiments with heterogeneous phantoms, where the use of optimally selected parameters increased the average contrast-to-noise ratio (CNR) by more than 200% and reduced the CNR standard deviation by 400% when compared to the use of arbitrarily selected imaging parameters. Therefore, the results suggest that the rule for specific selection of acquisition parameters becomes an important tool for producing high quality sonoelastography images.

Effects of aberration in crawling wave sonoelastography

Quantitative sonoelastography, through the formation of crawling waves, allows the estimation of elastic parameters in tissues using pulsed wave Doppler techniques. However, this techniques performance may be compromised by aberration effects during in vivo applications. In this study, an experimental evaluation of the effects of aberration when estimating shear wave speed from homogeneous phantoms was performed. The evaluations were performed using a commercial ultrasound scanner and gelatin-agar aberration layers of 33.26 ns, 62 ns and 116.73 ns RMS strength, and 2.00 mm, 3.40 mm and 6.70 mm of correlation length, respectively. The estimated speed values were obtained as a function of the vibration frequency for both the non-aberrated and aberrated cases. The estimated mean shear wave speed values in the presence of aberration showed only a 3% variation from those obtained in the non-aberrated cases. These experimental results suggest that crawling wave sonoelastography is not significantly affected by aberration effects in conditions typically observed in applications such as abdominal and breast imaging.

ARFI variance of acceleration (VoA) for noninvasive characterization of human carotid plaques in vivo

Rather than degree of stenosis, assessing plaque structure and composition is relevant to discerning risk for plaque rupture with downstream ischemic event. The structure and composition of carotid plaque has been assessed noninvasively using Acoustic Radiation Force Impulse (ARFI) ultrasound imaging. In particular, ARFI-derived peak displacement (PD) estimations have been demonstrated for discriminating soft (lipid rich necrotic core (LRNC) or intraplaque hemorrhage (IPH)) from stiff (collagen (COL) or calcium (CAL)) plaque features; however, PD did not differentiate LRNC from IPH or COL from CAL. The purpose of this study is to evaluate a new ARFI-based measurement, the variance of acceleration (VoA), for differentiating among soft and stiff plaque components. Both PD and VoA results were obtained in vivo for a human carotid plaque acquired in a previous study and matched to a histological standard analyzed by a pathologist. With VoA, plaque feature contrast was increased by an average of 60% in comparison to PD.

In vivo delineation of human carotid plaque features with ARFI variance of acceleration (VoA)

Stroke is commonly caused by thromboembolic events originating from ruptured carotid plaque, with rupture potential related to plaque composition. We have previously shown that soft (intraplaque hemorrhage/necrotic core) and stiff (collagen/calcium) plaque components are differentiated with high sensitivity and specificity by peak displacement (PD) in Acoustic Radiation Force Impulse (ARFI) imaging. However, PD showed low performance for distinguishing between soft and between stiff features. We hypothesize that soft and stiff feature delineation will be improved using a new ARFI outcome metric — the variance of acceleration (VoA). We test our hypothesis using histological validation.

In vivo delineation of carotid plaque features with ARFI variance of acceleration (VoA): Clinical results

Stroke is commonly caused by thromboembolic events originating from a ruptured carotid plaque. Histological studies have shown that plaque rupture potential is related to plaque composition rather than degree of stenosis. The use of Acoustic Radiation Force Impulse (ARFI) ultrasound imaging has enabled assessment of carotid plaque composition and structure. In particular, a previous study has shown that ARFI-derived peak displacement (PD) can differentiate between soft (intraplaque hemorrhage/necrotic core) and stiff (collagen/calcium) plaque components with high sensitivity and specificity. However, PD showed low performance for distinguishing between soft and between stiff features. This study evaluates an alternative ARFI-derived parameter, variance of acceleration (VoA), for intraplaque feature delineation. This study analyzed 20 carotid plaques imaged in vivo in patients undergoing carotid endarterectomy (CEA). After imaging, CEA specimens were harvested for histological validation. VoA was statistically significantly different (Wilcoxon, p < 0.01) between all examined plaque features, including between intraplaque hemorrhage and necrotic core and between collagen and calcium. Results suggest that VoA analysis improves ARFI discrimination between soft and between stiff carotid plaque components that are related to vulnerability for rupture.

Automated system for point shearwave elastography (pSWE) in rodent livers

Point shear wave elastography (pSWE) provides noninvasive measures of mechanical stiffness of biological tissue. Over the past decade, research has demonstrated that pSWE is effective at diagnosing a multitude of pathologies such as liver fibrosis and cancer. Life science researchers stand to benefit tremendously from access to pSWE but a number of challenges have limited its use in small animals. First, high frequency pSWE imaging suffers from transducer heating challenges and bandwidth requirements. Second, pSWE measurements are highly susceptible to variations in the environment and operator (e.g. precompression of tissue, transducer placement, etc.) which bias measurements and introduce error. Therefore, the objective of this work is to develop a robotically controlled, non-contact pSWE system capable of measuring liver stiffness in rodents.

Tissue characterization using simultaneous estimation of backscatter coefficient and elastic shear modulus

Tissue characterization using quantitative ultrasound (QUS) parameters has received significant attention in recent years due to its potential to improve the detection and diagnosis of diseased states. However, the vast majority of studies in QUS tissue typing have used parameters derived from either longitudinal or shear waves in isolation, thereby discarding potentially useful complementary information these parameters may carry. In this study, the simultaneous estimation of backscatter coefficients (derived from longitudinal waves) and shear modulus (derived from shear waves) was implemented on data from a clinical scanner. Both parameters were estimated from five ex vivo porcine kidney samples and used to calculate the anisotropy ratio in the parameters when analyzing the middle and pole regions of the kidneys. For all samples, the estimated parameters were higher in the pole regions than in the middle region, with anisotropy ratios of 1.42±0.11 and 3.07±0.70 for the shear modulus and the backscatter coefficient, respectively. Therefore, these results demonstrate that QUS parameters derived from both longitudinal and shear waves can be estimated simultaneously and may be used in conjunction to track changes in tissue structure and composition.Tissue characterization using quantitative ultrasound (QUS) parameters has received significant attention in recent years due to its potential to improve the detection and diagnosis of diseased states. However, the vast majority of studies in QUS tissue typing have used parameters derived from either longitudinal or shear waves in isolation, thereby discarding potentially useful complementary information these parameters may carry. In this study, the simultaneous estimation of backscatter coefficients (derived from longitudinal waves) and shear modulus (derived from shear waves) was implemented on data from a clinical scanner. Both parameters were estimated from five ex vivo porcine kidney samples and used to calculate the anisotropy ratio in the parameters when analyzing the middle and pole regions of the kidneys. For all samples, the estimated parameters were higher in the pole regions than in the middle region, with anisotropy ratios of 1.42±0.11 and 3.07±0.70 for the shear modulus and the backscatter coefficient, respectively. Therefore, these results demonstrate that QUS parameters derived from both longitudinal and shear waves can be estimated simultaneously and may be used in conjunction to track changes in tissue structure and composition.

Comparative evaluation of aberration effects in crawling wave sonoelastography and comb-push ultrasound shear elastography

Various quantitative ultrasound techniques allow the estimation of elastic parameters in tissues using mechanical or acoustic radiation forces. However, it has previously been reported that the presence of phase aberration compromises elasticity measurement results when propagating in inhomogeneous media. In this study, a comparative evaluation of the effects of phase aberration when estimating shear wave speed from homogeneous and inclusion phantoms was performed using comb-push ultrasound shear elastography (CUSE) and crawling wave sonoelastography (CrW). In CUSE, acoustic radiation force induced shear wave fields are applied and in CrW, pulsed wave Doppler techniques are used to monitor the shear waves produced by mechanical vibrators. The evaluations were performed using a Verasonics ultrasound scanner and four propanediol-based aberration layers. The estimated mean shear wave speed values in the presence of aberration had maximum coefficient of variation of 37% in CrW and 96% in CUSE. The experimental results suggest that CrW outperformed CUSE in homogeneous phantoms, and both reconstructions from CUSE and CrW were adversely affected as the aberrator strength increased.

Simultaneous estimation of shear elastic modulus and backscatter coefficient: Phantom and in human liver in vivo study

The use of quantitative ultrasound (QUS) parameters to characterize tissue has shown potential to improve current clinical diagnosis. However, most studies in QUS have used parameters derived from either ultrasonic compressional or low frequency shear wave in isolation, thereby discarding additional information these parameters may carry. In this study, the feasibility of estimating both shear wave speed (SWS) and backscatter coefficient (BSC) when using a unique data set is demonstrated both with physical phantoms and in vivo human liver data. The phantom results showed contrast values of 8.16 and 1.71 for BSC and SWS. The in vivo liver results for BSC and SWS were in good agreement with previously reported values in the literature. Therefore, these results demonstrate that multi-parameter tissue characterization using BSC and SWS have a promising potential to track changes in both microscopic and macroscopic properties of tissue.

Effects of data acquisition parameters on the quality of sonoelastographic imaging.

Sonoelastography is an ultrasonic technique that provides qualitative and quantitative images of tissue elasticity. Even though the Kasai variance estimator is a key part of the sonoelastographic image formation, there are no studies that demonstrate that its performance using discrete time signals and finite sized ensemble lengths is optimal. In this work, the influence of the selection of acquisition parameters (pulse repetition frequency or PRF, vibration frequency, and ensemble length) on the quality of the elastograms is studied. Simulations are carried out to define the optimal PRF and ensemble length given a vibration frequency in order to avoid artifacts which can severely degrade image quality. This empirical criterion is supported by sonoelastography experiments performed using two commercial scanners, where the variability increased from 4% to 42% at the worst selection of acquisition parameters. Although a further mathematical proof of the empirical findings is required, these results suggest that careful selection of PRF, vibration frequency and ensemble lengths is required to ensure unbiased sonoelastograms.