Original contributionTwo-Dimensional Sonoelastographic Shear Velocity Imaging
Introduction
Palpation is a routine physical examination whereby physicians qualitatively assess the elastic properties of soft tissue. During palpation, stiff masses that are discrete and differ from the surrounding tissue are considered suspicious for malignancy. Although these discrete masses may dislocate or feel fixed within the tissue, subtle findings are much more difficult to interpret. Specifically, in many tumor cases, despite the difference in stiffness, the small size of a pathological lesion or its deep location impedes detection and evaluation by palpation. In addition, lesions may or may not exhibit physical properties that allow detection using conventional medical imaging modalities such as computed tomography, magnetic resonance or ultrasound.
For more than two decades, the development of ultrasound-based imaging techniques for depicting the elastic properties of biological tissues has been the focal point of an international research endeavor (Gao et al 1996, Ophir et al 1999, Greenleaf et al 2003). The universal goal of these initiatives revolves around mapping some tissue mechanical property in an anatomically meaningful manner to describe valuable clinical information. Because changes in tissue stiffness are typically indicative of an abnormal pathological process (Anderson and Kissane 1977), imaging parameters related to tissue elasticity may finally provide a reliable gateway for differentiating normal from abnormal tissues.
Vibrational sonoelastography is a tissue elasticity imaging technique that estimates the amplitude response of tissues under harmonic mechanical excitation using ultrasonic Doppler techniques (Lerner et al. 1988). Because of a mathematical relationship between particle vibrational response and received Doppler spectral variance (Huang et al. 1990), low-frequency (50–500 Hz) and low-amplitude (20–100 μm) shear waves propagating in tissue can be visualized using sonoelastography to detect regions of abnormal stiffness (Parker et al. 1998).
Recently, it was shown that interfering shear waves produce slowly propagating interference patterns with an apparent velocity much less than (but proportional to) the underlying true shear velocity (Wu et al. 2004). Termed crawling waves, they are generated using a pair of mechanical sources vibrating at slightly offset frequencies or by continuously phase shifting one of the source excitation signals. The resultant shear wave interference patterns can be visualized in real-time using sonoelastographic imaging techniques. In general, crawling wave images describe shear wave propagation patterns and allow local estimation of shear velocity distributions (Wu et al. 2006; McLaughlin et al 2006; Hoyt et al. 2006). Assuming that the local shear velocity values are proportional to the square root of the shear modulus, spatial mapping of either parameter allows production of quantitative tissue elasticity images.
A real-time sonoelastographic technique for estimating local shear velocities from crawling wave images was introduced by Hoyt et al. (2006). Specifically, a relationship between crawling wave phase derivatives and local shear wave velocity was derived, with phase derivatives estimated using an autocorrelation-based technique. However, this method computes local shear velocities using a 1-D kernel and estimator and is conducted independent of shear wave displacement data in neighboring depth regions. Despite promising initial results using this shear velocity imaging technique, a more accurate and robust estimator may be realized by utilizing a 2-D shear velocity estimation algorithm. In this paper, a 2-D sonoelastographic shear velocity imaging technique is introduced. Performance of this novel tissue elasticity imaging modality is compared in simulation and experiments to results obtained using its 1-D precursor, followed by a conclusion.
Section snippets
Elastic properties of soft tissue
We begin by considering the wave motion equation for a linear and isotropic medium in terms of the displacements as: where E, v, ρ, and t are the Young’s modulus, Poisson’s ratio, mass density, displacement vector and time variable, respectively. As described by Landau and Lifshitz (1986), eqn (1) can be decomposed into two decoupled motion equations: one governing longitudinal wave motion and the other governing shear wave motion. In a homogeneous
Methods
To evaluate shear velocity estimation techniques, sonoelastographic simulation programs were implemented using MATLAB 7.0 (Mathworks, Inc., Natick, MA, USA). The first model assumes plane wave conditions and that the instantaneous shear wave interference patterns (i.e., shear wave displacement vector) propagating in a locally homogeneous and isotropic medium can be described by eqn (5). The shear wave signal-to-noise ratio (SNR) was implemented by superimposing 25 dB of Gaussian noise onto the
Methods
Crawling wave datasets were collected from two homogeneous elasticity phantoms with differing shear moduli. The true shear velocities in these phantoms were determined using shear wave time-of-flight measurements. Shear velocity sonoelastograms were generated using the local 1-D and 2-D shear velocity estimation techniques described in the Theory section. From sequences of shear velocity sonoelastograms equating to one full cycle of shear wave motion (deduced from the underlying crawling wave
Conclusions
In this paper, a novel 2-D sonoelastographic technique for estimating local shear velocities from propagating shear wave interference patterns (termed crawling waves) is introduced. A relationship between the local crawling wave spatial phase derivatives and local shear wave velocity is derived with phase derivatives estimated using a 2-D autocorrelation technique. Comparisons were made between the 2-D sonoelastographic shear velocity estimation technique and its computationally simpler 1-D
Acknowledgements
We are grateful for helpful suggestions from Drs. Deborah J. Rubens, John Strang, Zhe Wu and Man Zhang, and for the loan of equipment and expertise from General Electric (GE) Medical Systems. This work was supported by NIH Grant 5 R01 AG16317-05.
References (19)
- et al.
Imaging of the elastic properties of tissue—A review
Ultrasound Med Biol
(1996) - et al.
- et al.
Semiautomatic measurement of thermally ablated lesions in sonoelastographic images
J Ultrasound Med
(2007) Radiofrequency ablation of malignant liver tumors
Oncologist
(2001)- et al.
Tissue characterization using magnetic resonance elastography: Preliminary results
Phys Med Biol
(2000) - et al.
Theory of elasticity
(1986) - et al.
Sonoelasticity: Medical elasticity images derived from ultrasound signals in mechanically vibrated targets
Acoust Imaging
(1988) - et al.
Selected methods for imaging elastic properties of biological tissues
Annu Rev Biomed Eng
(2003) - et al.
Phantom materials for elastography
IEEE Trans Ultrason Ferroelectr Freq Control
(1997)
Cited by (80)
A digital viscoelastic liver phantom for investigation of elastographic measurements
2020, Computers in Biology and Medicine3-D H-Scan Ultrasound Imaging and Use of a Convolutional Neural Network for Scatterer Size Estimation
2020, Ultrasound in Medicine and BiologyCitation Excerpt :The overarching challenge is to find hidden patterns in the US data that reveal more information on tissue function and pathology (Kelly et al. 2018; Opacic et al. 2018; Steifer and Lewandowski 2019). Several promising US-based tissue characterization methods have been introduced, namely, backscatter classification (Kurokawa et al. 2016), integrated backscatter (Takami et al. 2019), spectral feature extraction (Fang et al. 2018) and tissue elasticity imaging (Hoyt et al. 2008a, 2008b). A potential limitation for some of these tissue characterization methods is that they use a relatively large kernel (window) of US data during quantification, which can affect spatial resolution and make in vivo measurement of local changes problematic.
2-D Ultrasound Shear Wave Elastography With Multi-Sphere-Source External Mechanical Vibration: Preliminary Phantom Results
2020, Ultrasound in Medicine and BiologyCitation Excerpt :Lightweight and ergonomic EMV devices are required to enable clinicians to work efficiently and comfortably. To vibrate heavy masses (the shear wave generation sources) at high frequency, many of the aforementioned designs in Hoyt et al. (2008), Sandrin et al. (2002) and Mellema et al. (2016) use linear actuators with sizes significantly larger than the ultrasound transducer, which makes handheld operations challenging. Second, continuous vibration of a heavy vibration source that is in direct contact with the human body at high frequency (over 100 Hz) generates strong reaction forces.
Classification of sonoelastography images of prostate cancer using transformation- based feature extraction techniques
2018, Soft Computing Based Medical Image AnalysisSpatial Angular Compounding Technique for H-Scan Ultrasound Imaging
2018, Ultrasound in Medicine and BiologyPost-Procedure Evaluation of Microwave Ablations of Hepatocellular Carcinomas Using Electrode Displacement Elastography
2016, Ultrasound in Medicine and Biology