Yanning Zhu

Ultrasonic elasticity imaging

In an ultrasound imaging system, a displacement vector is estimated for a pattern of samples throughout an imaged region of interest (ROI) by comparing two successive B-mode frames. The displacement vector is preferably estimated using block matching. Once displacement vectors are estimated for samples throughout the ROI, corresponding strain values are estimated, which indicate the degree of elasticity of the respective tissue portions. An image is then displayed showing the strain distribution within the ROI as it is stressed, for example, by the user pressing the ultrasound transducer against the patients body. The invention allows for both real-time and post-processed generation of elasticity displays, even based on the same body of acquired frame data. The real-time display is preferably generated using lower quality block matching whereas the post-processed elasticity calculations are carried out using high-quality techniques. Different techniques are included for adjusting the size and location of the search region of the block matching based on different measures of reliability of current displacement estimates.

A finite-element approach for Young's modulus reconstruction

Modulus imaging has great potential in soft-tissue characterization since it reveals intrinsic mechanical properties. A novel Youngs modulus reconstruction algorithm that is based on finite-element analysis is reported here. This new method overcomes some limitations in other Youngs modulus reconstruction methods. Specifically, it relaxes the force boundary condition requirements so that only the force distribution at the compression surface is necessary, thus making the new method more practical. The validity of the new method is demonstrated and the performance of the algorithm with noise in the input data is tested using numerical simulations. Details of how to apply this method under clinical conditions is also discussed.

Maximum-likelihood approach to strain imaging using ultrasound

A maximum-likelihood (ML) strategy for strain estimation is presented as a framework for designing and evaluating bioelasticity imaging systems. Concepts from continuum mechanics, signal analysis, and acoustic scattering are combined to develop a mathematical model of the ultrasonic waveforms used to form strain images. The model includes three-dimensional (3-D) object motion described by affine transformations, Rayleigh scattering from random media, and 3-D system response functions. The likelihood function for these waveforms is derived to express the Fisher information matrix and variance bounds for displacement and strain estimation. The ML estimator is a generalized cross correlator for pre- and post-compression echo waveforms that is realized by waveform warping and filtering prior to cross correlation and peak detection. Experiments involving soft tissuelike media show the ML estimator approaches the Cramer-Rao error bound for small scaling deformations: at 5 MHz and 1.2% compression, the predicted lower bound for displacement errors is 4.4 microns and the measured standard deviation is 5.7 microns.

Strain imaging with a deformable mesh

Ultrasonic strain imaging has drawn much attention recently because of its ability to noninvasively provide information on spatial variation of the elastic properties of soft tissues. Traditionally, local strain is estimated by scaling and cross correlating pre-and postcompression ultrasound echo fields. However, when the motion field generated by compression is more complex, scaling and cross correlation can no longer provide precise displacement estimates because of signal decorrelation. We introduce a new algorithm based on the deformable mesh method. This algorithm can accommodate more general forms of motion, namely, the motion that can be described by bilinear transformations. We applied the new algorithm to three sets of data in order to evaluate its performance. In the first set of data, primitive motions such as shearing and rotation are simulated. The second set of data is collected by compressing a tissue-mimicking phantom with three hard inclusions. The third experiment involves an ex vivo pig kidney embedded in a block of gelatin. The results from all three experiments show improvements with the new algorithm over other methods.

Lesion size ratio for differentiating breast masses

We are developing a clinical ultrasonic imaging system for real-time estimation and display of tissue elastic properties. We have demonstrated that real-time feedback of elasticity image is essential for obtaining high-quality data (consecutive images with high spatial coherence). The key element to successful scanning is real-time visual feedback which guides the patient positioning and compression direction. One of our findings, consistent with previous report is that benign breast masses are typically about the same size in B-mode images. In this work we continue testing that hypothesis with an increasingly large data set with greater diversity of breast mass types. Results from a single-observer ROC study demonstrate that the lesion size ratio is a useful criterion for classifying benign versus malignant breast masses.

Bioelasticity imaging:II. Spatial resolution

The large elasticity contrast possible with strain imaging promises new diagnostic information to augment x-ray, MRI, and ultrasound for the detection of tumors in soft tissue. In the past, we described the design of an elastographic system using the Fourier crosstalk concept introduced by Barrett and Gifford. The diagonal of the crosstalk matrix is related to the pre-sampled modulation transfer function (MTF) of the strain image. Another approach to measuring the spatial resolution of an elasticity image employs a linear frequency- modulated (chirp) strain pattern imposed upon a simulated ultrasonic echo field to study the strain modulation over a range of spatial frequencies in the image. In experiments, high contrast inclusions positioned at varying separations were imaged to apply the Rayleigh criterion for resolution measurement. We measured MTF curves that fell to 0.2 at a spatial frequency of 0.5 mm-1 to 1 mm-1 under realistic conditions. The spatial resolution for ultrasonic strain imaging strongly depends on the transducer properties and deformation patterns applied to the object. Experiments with tissue-like phantoms mimicking the properties of early breast cancer show that 2 mm spheres three times stiffer than the background can be readily resolved. Thus, the potential for using elasticity imaging to detect early breast cancers is excellent.

Deformable mesh algorithm for strain imaging with complex tissue deformation

This article introduces a framework for strain estimation using interpolated landmark displacement methods. Two implementations of this paradigm, namely, the block matching algorithm and the landmark displacement perturbation algorithm, are described. Results for experiments with phantoms and biological tissues are presented.

Noise reduction strategies in freehand elasticity imaging

We are developing a clinical ultrasonic imaging system for real-time estimation and display of tissue elastic properties. We have demonstrated that real-time feedback of elasticity images is essential for obtaining high-quality data (consecutive images with high spatial coherence). The key element to successful scanning is real-time visual feedback which guides the patient positioning and compression direction. Our data have clearly demonstrated nonlinearity in the strain properties of different tissue types. We have also demonstrated that a comparison of the area of a breast lesion observed in strain images versus B-mode images is a sensitive criterion for differentiating malignant from benign tumors. Frame-to-frame variability in strain images somewhat degrades the ability to observe these phenomena. Three strategies for reducing frame-to-frame strain image noise are described. The combination of these post-processing strategies provides a significant improvement in the quality of long sequences of strain images.

In vivo real-time freehand elasticity imaging

We are developing a system for real-time estimation and display of tissue elastic properties using a clinical ultrasonic imaging system. Results in phantoms are in excellent agreement with that predicted with finite element analysis. Results in volunteer patients have shown that high quality elasticity images are easily obtained in vivo in breast and thyroid pathologies. The key element to successful scanning is real-time visual feedback which guides the patient positioning and compression direction. Results show that the frame-to-frame changes in strain image contrast appear to be unique to specific lesion types. In addition, the size of a lesion displayed in a strain image, relative to that in a standard B-mode image, is about the same for benign lesions but the size is considerably larger for malignant lesions. The observations will likely significantly improve the discrimination of radiologically indeterminant lesions.

Spatial resolution in elasticity imaging with ultrasound

We are exploring the resolution limits of ultrasonic strain imaging techniques by building tissue-mimicking gelatin resolution phantoms with various targets imbedded. Through acquiring scan data with applied compression, we compute the strain and study the visibility and separability of targets with different sizes and separations. Our preliminary results have shown that targets with 4.2 mm in diameter separated by 1.6 mm are easily seen.