Quantitative Elastography – Improving Lateral Displacement and Strain Measurement
L Garcia, J Fromageau, C Uff, NL Bush, JC Bamber, in collaboration with RJ Housden, G Treece, Engineering Department, University of Cambridge.
Source of funding: EPSRC, CRUK
Much of the work on elastography described on the Quantitative elasticity imaging – slip elastography page would be greatly improved if lateral displacement and strain were accurate and precise enough to be employed in image construction, particularly in the context of poroelastography and slip elastography. We have therefore explored two techniques that may improve lateral displacement and strain measurement: lateral over-sampling of array element spacing and ultrasound beam steering for measurement of the full displacement vector.
Beam steering has been found to provide the greatest improvement, and is the easiest to implement. Although beam steering has been investigated by other groups, they had implemented it using 1-D ultrasound tracking only, which provides the ultrasound-axial component of displacement at each of a number of steering angles, to reconstruct the displacement vector at each spatial location.
With the aim of recovering the correlation that is lost because mechanically-axial palpation with a linear ultrasound probe induces large ultrasound-lateral tissue displacements, we investigated 2-D speckle tracking. We found, however, that, rather than recovering lost correlation, 2-D speckle tracking suffers additional decorrelation. This turned out to be due to loss of orthogonality in the lateral and axial components of the point spread function, brought about by beam steering implemented within a linear scan format. A reshaping algorithm was therefore developed to process the radiofrequency scan data so as to reinstate this orthogonality. Following this, we have found, by simulation and experiment, that 2-D tracking provides an improvement over 1-D tracking, for both mechanical axial and lateral displacement measurement (Figure 6). Fig. 6. (a) Data in ultrasound acquisition coordinates demonstrates loss of axial-lateral orthogonoality (separability) of the point spread function (inset picture) when a beam steering angle of 10 degrees is employed. Reshaping of the radiofrequency ultrasound lines using the known steering angle recovers most of the orthogonality (inset picture). (c) Finite element simulation (FEM) and simulated ultrasound measured lateral displacements for various combinations of 1D and 2D speckle tracking, and lateral component of the displacement vector formed from two steered beams and two different beam steering angles, demonstrating the improvement obtained with 2D tracking once the reshaping algorithm has been employed.