Ultrasonic tracking of shear waves using a particle filter

Particle filter efficiency under limited communication

Sequential Monte Carlo methods are typically not straightforward to implement on parallel architectures. This is because standard resampling schemes involve communication between all particles. The α-sequential Monte Carlo method was proposed recently as a potential solution to this which limits communication between particles. This limited communication is controlled through a sequence of stochastic matrices known as α-matrices. We study the influence of the communication structure on the convergence and stability properties of the resulting algorithms. In particular, we quantitatively show that the mixing properties of the α-matrices play an important role in the stability properties of the algorithm. Moreover, we prove that one can ensure good mixing properties by using randomized communication structures where each particle only communicates with a few neighbouring particles. The resulting algorithms converge at the usual Monte Carlo rate. This leads to efficient versions of distributed sequential Monte Carlo.

A kernel smoothing algorithm for ablation visualization in ultrasound elastography

Three-dimensional visualization of tumor ablation procedures have significant clinical value because the ability to accurately visualize ablated volumes can help clinicians gauge the extent of ablated tissue necrosis and plan future treatment steps. Better control over ablation volume can prevent recurrence of tumors treated using ablative procedures. This paper presents a kernel based smoothing algorithm called MatérnSmooth to reconstruct shear wave velocity maps from data acquired through ultrasound electrode vibration elastography. Shear wave velocity estimates are acquired on several intersecting imaging planes that share a common axis of intersection collinear with the ablation needle. An objective method of choosing smoothing parameters from underlying data is outlined through simulations. Experimental validation was performed on data acquired from a tissue mimicking phantom. Volume estimates were found to be within 20% of the true value.

Assessment of Mechanical Properties of Tissue in Breast Cancer-Related Lymphedema Using Ultrasound Elastography

Breast cancer-related lymphedema is a consequence of a malfunctioning lymphatic drainage system resulting from surgery or some other form of treatment. In the initial stages, minor and reversible increases in the fluid volume of the arm are evident. As the stages progress over time, the underlying pathophysiology dramatically changes with an irreversible increase in arm volume most likely due to a chronic local inflammation leading to adipose tissue hypertrophy and fibrosis. Clinicians have subjective ways to stage the degree and severity such as the pitting test which entails manually comparing the elasticity of the affected and unaffected arms. Several imaging modalities can be used but ultrasound appears to be the most preferred because it is affordable, safe, and portable. Unfortunately, ultrasonography is not typically used for staging lymphedema, because the appearance of the affected and unaffected arms is similar in B-mode ultrasound images. However, novel ultrasound techniques have emerged, such as elastography, which may be able to identify changes in mechanical properties of the tissue related to detection and staging of lymphedema. This paper presents a novel technique to compare the mechanical properties of the affected and unaffected arms using quasi-static ultrasound elastography to provide an objective alternative to the current subjective assessment. Elastography is based on time delay estimation (TDE) from ultrasound images to infer displacement and mechanical properties of the tissue. We further introduce a novel method for TDE by incorporating higher order derivatives of the ultrasound data into a cost function and propose a novel optimization approach to efficiently minimize the cost function. This method works reliably with our challenging patient data. We collected radio frequency ultrasound data from both arms of seven patients with stage 2 lymphedema, at six different locations in each arm. The ratio of strain in skin, subcutaneous fat, and skeletal muscle divided by strain in the standoff gel pad was calculated in the unaffected and affected arms. The p -values using a Wilcoxon sign-rank test for the skin, subcutaneous fat, and skeletal muscle were 1.24×10^-5 , 1.77×10^-8 , and 8.11×10^-7 respectively, showing differences between the unaffected and affected arms with a very high level of significance.

Efficient 3-D Reconstruction in Ultrasound Elastography via a Sparse Iteration Based on Markov Random Fields

Percutaneous needle-based liver ablation procedures are becoming increasingly common for the treatment of small isolated tumors in hepatocellular carcinoma patients who are not candidates for surgery. Rapid 3-D visualization of liver ablations has potential clinical value, because it can enable interventional radiologists to plan and execute needle-based ablation procedures with real time feedback. Ensuring the right volume of tissue is ablated is desirable to avoid recurrence of tumors from residual untreated cancerous cells. Shear wave velocity (SWV) measurements can be used as a surrogate for tissue stiffness to distinguish stiffer ablated regions from softer untreated tissue. This paper extends the previously reported sheaf reconstruction method to generate complete 3-D visualizations of SWVs without resorting to an approximate intermediate step of reconstructing transverse C planes. The noisy data are modeled using a Markov random field, and a computationally tractable reconstruction algorithm that can handle grids with millions of points is developed. Results from simulated ellipsoidal inclusion data show that this algorithm outperforms standard nearest neighbor interpolation by an order of magnitude in mean squared reconstruction error. Results from the phantom experiments show that it also provides a higher contrast-to-noise ratio by almost 2 dB and better signal-to-noise ratio in the stiff inclusion by over 2 dB compared with nearest neighbor interpolation and has lower computational complexity than linear and spline interpolation.