Improving ultrasound display by retrieving high-frequency sonographic information

The ability of an ultrasound system to distinguish between two points along the direction of pulse propagation at a certain depth in tissue is denoted as the axial resolution of the system, which is an important factor in image quality. The axial resolution of ultrasonic B-mode images has been studied by Wagner et al. (1983) and shown to be on the order of the pulse bandwidth. However, a broad pulse bandwidth, which is obtained at high transmission frequency, has limited depth of penetration because of tissue attenuation. The goal of this dissertation is to optimize the ultrasound axial resolution by applying all of the information available in the received echo signal. Recent results by Nguyen et al. (2013) may be used to infer spatial resolution of the RF data. Specifically, they describe the ability of the acquisition stage to capture spatial frequency components of the object into the RF data via a spectrum they call "acquisition information spectrum'' (AIS). The spatial frequency representation of AIS contains two lobes: a baseband lobe and a high-frequency lobe centered at twice the central pulse frequency. Study of AIS reveals that the RF data contains spatial frequency components of the object within a bandwidth twice the pulse bandwidth. In other words, the detection bandwidth of the RF data goes up to twice the detection bandwidth of the B-mode image for a 100 percent bandwidth pulse-echo profile. Extending the analysis to B-mode images was a challenge because of the non-linear demodulation process involved. The unique contribution of this dissertation research was to use ultrasound intensity (B-mode squared) statistics to demonstrate the transfer of spatial frequency components of the object in the traditional ultrasound display and use the resulting insight to improve display processing that results in improved visibility of an important breast cancer marker, microcalcifications. Our new results in analyzing intensity statistics provided a linear system framework to pursue our goal. Our analysis shows that intensity signals only contain low spatial-frequency components of the object, while high-frequency components that are originally captured in the RF data are lost during the demodulation. This loss of high spatial-frequency components of the object is what limits the axial resolution in ultrasound display. To circumvent this limitation, we proposed an alternative display stage processing that we call "complement intensity processing'' to transfer and display high spatial-frequency components of the object. As an important application where high spatial-frequency (high-resolution) components of the object have significant diagnostic value, we consider detection of microcalcifications. Microcalcifications are very small high-contrast deposits that form in soft tissues. Computer simulation results validate the effectiveness of complement intensity processing in capturing information associated with these small calcifications. Practical challenges of implementing this method are addressed in the designed lab experiments, which also prove the usefulness of complement intensity processing in improving the resolution of ultrasound display. The contributions of this dissertation are listed below: * exploring the acquisition information spectrum in 2-D spatial frequency domain * analyzing the transfer of diagnostic information through intensity mean and variance * demonstrating that high spatial-frequency information is not available in the conventional echo-intensity signal * introducing a new, simple, and cost-effective processing method in the display stage to display high-frequency information efficiently * simulating and analyzing the improvement achieved by this method for the application of microcalcification detection * verifying the improvements predicted by theory and simulations using lab-generated models for lesions containing microcalcification