Fatemi, Mostafa

Acoustic radiation force has been used in many interesting medical and industrial applications in the past two decades. The fact that radiation stress can be confined to a small focal area and exerted remotely inside or on an object allows one to use this phenomenon in a variety of ways to explore properties of the object. Vibro-acoustography is a material characterization and imaging method that uses harmonic or impulsive acoustic radiation force to stimulate the targeted material and receives the resulting acoustic response. The output acoustic signal is then used to extract important information about the object. Biomedical applications of vibro-acoustography and its closely related methods include imaging, flow detection, and tissue characterization. Other applications include non-destructive evaluation/testing of materials, modal analysis of small structures, and microscopy. This talk presents an overview of vibro-acoustography and its closely related techniques over the past 20 years. Various biomedical and industrial applications will be presented. The talk will conclude with a discussion on potential future applications of this technology.

Minimally invasive procedures using ultrasound are becoming prominent, especially in applications like biopsy, catheter guidance, and surgeries. In our previous work a catheter tip is fitted with a miniature omnidirectional piezoelectric crystal; the ultrasonic field from the imaging array and the piezoelectric crystal interact resulting in a symmetric Doppler shift. Previously, a frequency domain algorithm was used to identify pixels having the symmetric Doppler shift. The identified pixels can then be visualized on B-mode and color map with a unique color thus representing the location of the catheter tip. However, frequency domain algorithms are computationally expensive and require changes to data workflow. Using mean frequency estimators with symmetric Doppler shift results in the catheter tip being displayed as zero mean noise making it hard to distinguish from color Doppler noise. To overcome this limitation, complex demodulation can be used. Modulating the time domain signal with sine and cosine by the piezoelectric excitation frequency and summing them up results in cancellation of one of the symmetric Doppler components. With only one frequency component remaining mean frequency estimators can be used to identify the location of the catheter. The presented algorithm has lower computational complexity and requires minimal change in software design.

While conventional filtering techniques can effectively exclude tissue-related Doppler shifts from those resulting from blood motions in large vessels (flow rates ~ cm/s), significant spectral overlap can hinder proper clutter filtering in imaging slow flows (flow rates ~ mm/s or tenths of mm/s). Recently, several groups have proposed clutter filtering via processing of echo ensembles in the rank space instead of frequency domain. Though significant Doppler content can result from fast moving tissues, global coherence of the induced variations enables a low rank representation. In this talk, we first provide a brief overview of the low rank clutter removal methods. We then present a new framework for the analysis of Doppler ensembles using multi-linear decomposition of a tensor comprising Doppler ensembles at different rates. The model results in a high dimensional tensor from which, after proper rank selection, images of flow at different speeds can be extracted along different fibers. Using simulation models, we demonstrate the superior performance of the proposed method over singular value clutter filtering in imaging small vessels embedded in moving tissue undergoing both translational and shearing motions. We also present in vivo examples where the proposed method provides significant blood to clutter + noise ratio enhancement compared to singular value clutter filtering.

We hypothesize that pattern of microvasculature in the thyroid nodules can be a biomarker of thyroid cancer. In this study, morphological features of microvasculature networks are assessed as potential biomarkers for differentiation of benign, suspicious and malignant thyroid nodules. This prospective study included n = 81 patients, with a total of 61 benign, 12 suspicious and 8 malignant nodules verfied by pathology and cytology findings. Thyroid nodules were first identified using conventional B-mode ultrasound. A sequence of raw ultrasound data were acquired from each nodule site in two perpendicular planes at ~600 frames/sec. Power Doppler image of the microvasculature was formed using singular value decomposition clutter removal filtering. Gross images were further processed using a combination of morphology and Hessian-based vessel enhancement filtering. Background-free images of the microvasculature networks were segmented (vessels segment length >385μm and diameter >90μm) for morphological analysis. Using a Bayesian framework, measures of vessels density, diameter and tortuosity were tested for significant differences in the three groups. Among all measures, the probability of vessel diameter >600μm in benign cases was significantly higher than that in malignant cases with the p-value <0.01 using Wilcoxon-Rank-Sum test, suggesting reliable cancer detection performance of the proposed method. [Acknowledgment: NIH Grant EB17213.]

As shown in a recent study, estimation of the magnitude of the acoustic radiation force (ARF) in isotropic tissue-like solids is the key to measuring a third-order nonlinear modulus of elasticity (C)—responsible for coupling the shear and volumetric constitutive responses of a material. Due to experimental limitations, a direct measurement of the ARF in tissue is however difficult. To overcome this problem, we deploy 3D elastodynamic finite element (EFE) simulations of the ARF experiment in tissue-mimicking phantoms to estimate (i) the “background” shear modulus from the phase of the induced shear waves; (ii) the ARF magnitude from the shear wave amplitude, and (iii) the modulus C from the knowledge of the ARF. In this way, by interpreting the ARF-generated shear waves through the prism of 3D-EFE simulations, we facilitate local estimation of the nonlinear C modulus with spatial resolution equaling the size of the focal region. We call this method the C-Elastography (CE). In this study, the CE technique is applied to gelatin phantoms containing an agar-based inclusion. The profiles of the C-modulus in phantoms tested demonstrate good correlation with the phantom geometry, showing marked (and sharp) C-contrast at push points acting on the inclusion. [Acknowledgment: NIH-Grant EB23113.]

Early detection of osteopenia and osteoporosis is critical in identifying at-risk groups and manage necessary therapy. In this paper, a new noninvasive and quantitative vibro-acoustic method is proposed for detection of osteoporosis and osteopenia. In this method, we excite the bone by an ultrasound radiation force (URF) pulse. The URF pulse induces a vibration in the bone, resulting an acoustic wave that is received by a hydrophone placed on the skin. This prospective study included n = 23 volunteers. The number of volunteers with osteopenia and osteoporosis were 7, 6, respectively, and the number of volunteers with normal bone was 10. URF excitations were applied at four equi-distance points on the bone. The resulting acoustic signals were used for wave velocity estimation based on cross-correlation technique. In general, the hydrophone signal may be modeled as a superposition of different modes of bone vibration. Based on this model, we used variational mode decomposition (VMD) as an efficient technique to decompose the received signal into an ensemble of band-limited intrinsic mode functions. Statistical analysis demonstrates that osteopenia and osteoporosis bones can be differentiated from normal bone with p< 0.001 using Wilcoxon rank-sum test on estimated velocity of the modes with maximum power.

Abstract Kelvin–Voigt fractional derivative (KVFD) model parameters have been used to describe viscoelastic properties of soft tissues. However, translating model parameters into a concise set of intrinsic mechanical properties related to tissue composition and structure remains challenging. This paper begins by exploring these relationships using a biphasic emulsion materials with known composition. Mechanical properties are measured by analyzing data from two indentation techniques—ramp-stress relaxation and load-unload hysteresis tests. Material composition is predictably correlated with viscoelastic model parameters. Model parameters estimated from the tests reveal that elastic modulus E 0 closely approximates the shear modulus for pure gelatin. Fractional-order parameter α and time constant τ vary monotonically with the volume fraction of the material’s fluid component. α characterizes medium fluidity and the rate of energy dissipation, and τ is a viscous time constant. Numerical simulations suggest that the viscous coefficient η is proportional to the energy lost during quasi-static force-displacement cycles, E A . The slope of E A versus η is determined by α and the applied indentation ramp time T r . Experimental measurements from phantom and ex vivo liver data show close agreement with theoretical predictions of the η − E A relation. The relative error is less than 20% for emulsions 22% for liver. We find that KVFD model parameters form a concise features space for biphasic medium characterization that described time-varying mechanical properties.

Bladder wall mechanical properties are important indicators of bladder compliance. This study compares the effectiveness of the Kelvin-Voigt, Maxwell, and fractional Kelvin-Voigt rheological models in capturing the mechanical properties of normal (compliant) and aberrant (non-compliant) ex-vivo pig bladder walls. Bladders were filled to different volumes, varying their thickness and internal pressure, and excited by the acoustic radiation force. Pulse-echo ultrasound imaged anti-symmetrical Lamb waves traveling along the bladder wall. Two-dimensional Fourier analysis generated velocity-frequency dispersion curves for the Lamb wave propagation. Fitting these experimental curves against theoretical curves from the models yielded values for bladder elasticity and viscosity. Pressure, measured simultaneously, was used as a point of comparison with elasticity. Our findings indicate an increasing trend in both elasticity for the rheological models and recorded pressure with increasing volume. This rate was observed to be significantly higher in aberrant bladders than intact bladders for the Maxwell and fractional Kelvin-Voigt models (p-value of 0.0082). Additionally, all three models showed a similar strong correlation (greater than 0.9) between pressure and elasticity. These findings provide a foundation for selecting an appropriate bladder wall rheology model which can potentially be used in evaluation of human bladder mechanical properties. [Work supported by NIH grant DK99231.]