Abstract
Background and motivation
The health of human tissue can be indicated by the stiffness of the tissue. It is known that the risk of a nodule being malignant is increased with the stiffness of the nodule. Elastography is an imaging mode capable of displaying the stiffness of the tissue. Static elastography with ultrasound consists of creating a pre- and post-compression ultrasound image where the tissue being imaged has been compressed between the images. The displacement of tissue is calculated along the axial dimension based on the assumption that speckle pattern follows tissue movement. Tissue strain, indicating the stiffness of tissue, can then be found from the displacement of the tissue.
Speckle statistics and the speckle pattern are different for images created with conventional and adaptive (Capon) beamforming. The speckle pattern created with adaptive beamforming has a smaller and more distinct pattern because of the improved resolution by adaptive beamforming. Hypothetically a more distinct pattern should result in better correlation and thus better displacement estimation.
Recently it has been shown that lateral oversampling is needed to achieve lateral shift-invariance between image frames when using adaptive beamforming. Shift-invariance between frames is especially important for elastography since the displacement estimate is based on correlation between two nearly identical frames.
Approach
To simulate static elastography two speckle images are created with Field II simulations based on the same scatter phantom, where the scatterers have been displaced axially to create pre- and post-compression ultrasound images. The images are created with the conventional beamformer and the adaptive beamformer with different parameters. In the middle of the phantom a circular object has constant displacement to mimic a hard malignant nodule in the tissue.
Results and conclusions
We show that lateral oversampling is necessary for single frame scenarios when doing adaptive beamforming and to achieve shift- invariant imaging of speckle. The speckle pattern from adaptive beamforming is more distinct, but our research shows that adaptive beamforming with certain parameters gave similar performance for axial correlation for displacement estimation as conventional beamforming and thus similar accuracy when doing static elastography.