STRAIN IMAGING

Background

Ultrasound elastography is a technique to estimate the biological tissue elastic properties by measuring local tissue deformations (strain) when subjected to a (known) force. Tissue that undergoes large deformation/high strain can be considered soft (Figure) compared to hard tissue that shows low strain. Ultrasound elastography aims at providing an objective variant of palpation. Palpation is what physicians perform when they use their tactile senses to detect irregular structures and changes in stiffness in tissue by pressing on them. Those irregularities and changes are often strong indicators for diseases. Since palpation is limited by the accessibility of the organs, it is not quantitative but is strongly dependent on the expertise of the performing clinician. With ultrasound elastography also, noninvasive deformation measurements of more deeply located structures can be performed and visualized in so-called elastograms (or strain maps).

Strain.jpg

Ultrasound strain imaging (elastography) was first described by Ophir and his co-workers. For this technique, ultrasound data are acquired in a pre-deformation (relaxed) state and compared with ultrasound data acquired in a post-deformation state (activated). To create the (small) deformation, a force is required which can be applied externally or internally. Internal body forces can originate from respiratory motion, the beating of the heart, or variations in blood pressure as for example in carotid strain imaging. If the rate of deformation relative to the ultrasound acquisition frame rate is low, the method is also often referred to as quasi-static strain imaging. Negative strain indicates compressive deformation (i.e. squashing) while positive strain indicates tensile strain (i.e. stretching). Therefore, normal strain (ε), also known as engineering strain, is defined as the ratio of the change in length of a tissue, ΔL, divided by the original length of the tissue, L.


Research/developments


Displacement compounding

Ultrasound displacement and strain imaging can be very accurate in the direction of the ultrasound beam due to the availability of phase information. Hence the displacement and strain imaging accuracy perpendicular to beam direction is less due to the absence of phase. A method that circumvents the lack of phase information in lateral direction is called angular compounding or displacement compounding. In this technique, ultrasound beam steering is used to obtain multiple axial displacement estimates, which are projected in the lateral direction, as depicted in Figure 1.5. In this way, lateral displacement estimates benefit from the availability of phase information and become more accurate. In theory, larger beam steering angles provide more accurate lateral displacements because the steered beam becomes more aligned with the lateral direction. However, in practice, this is often not feasible because the ultrasound quality reduces due to a decrease in element sensitivity and pronounced grating lobe interference. Furthermore, geometrically the field of view size and depth are determined by the overlapping area of the steered beams, which depends on the width of the aperture width of the transducer. The figure below. displays the displacement compounding scheme for two steered displacements (d_α^ax,d_(-α)^ax ) estimated using both pre- and post-deformation states and projected into the lateral/horizontal direction (d_0^(Comp_lat )) according to equation 1.3.


Compounding.jpg


High performance beamforming grid

Using ultrafast image acquisition and plane wave imaging, grants the freedom to convert unfocused element data into focused RF data at any given location within the region insonified. Displacement compounding combines axial displacement estimates of at least two acquisition angles; hence, ideally, the underlying sampling grids for RF data of the multiple angles should coincide at points where a compound displacement estimate is to be estimated. Rotating the zero-degree beamforming grid to an arbitrary angle does not provide these coinciding points. Within our lab we designed a novel way to combine the zero-degree grid with the angular grids yielding the displacement grid defined by spatially overlapping specific sample locations of all grids. It turns out that it is only possible to create such a grid combination with orthogonality as a prerequisite, for certain transmit angles α.

Grid.jpg


Patient specific phantom of the carotid bifurcation

In order to perform feasibility tests of multiple ultrasound-based techniques like strain imaging and bloodflow imaging, our lab designed a tissue mimicking phantom of the carotid bifurcation. This phantom is based on a patient specific geometry and made of PVA-cryogel matching the acoustic and mechanical properties of real tissue. The vessel wall was surrounded by a second body PVA gel supporting the vessel-wall strength. The phantom was connected to a pulsatile circulation of blood mimicking fluid matching real physiological blood flow and pressure characteristics.

Phantom.jpg


Multi-slice in vivo gated ultrasound acquisitions

One of the main limitations of 2D strain imaging is the accuracy degradation of the strain map due to out of plane motion. Therefore, our lab developed an acquisition setup to combine multiple 2D images and process them into a dynamic 3D ultrasound volume of the carotid artery bifurcation. In this way we were able to calculate the carotid artery wall motion in 3D. This information was them finally processed into a 3D strain map allowing the assessment of the carotid wall mechanics. This information might hold the potential to classify the risk for plaque rupture better than the relative amount of vessel occlusion by the plaque which is mainly used today.

Setup.jpg

Clinical Applications


VASC
CARD
MUSCLE
BREAST

Our people

Scientific staff


Technical staff