| The analytic nature of the radio frequency (RF) data from diagnostic pulse-echo ultrasound allows for fine axial displacement and strain estimation through correlation-based speckle tracking techniques. Speckle tracking allows for estimating axial tissue displacements in the sub-micron range with high spatial resolution. This was exploited in the development of ultrasound elastography and thermography. Elastography offers the promise of a higher contrast imaging compared to conventional ultrasound (e.g. in imaging stiff tumors) while thermography offers the promise of guiding and monitoring thermal therapies (e.g. in cancer treatment). At the relatively low frame rates of ultrasound, however, the quality of the speckle tracking may deteriorate (due to loss of correlation between subsequent frame data), which results in noisy displacement field estimates. These, in turn, may render the strain field estimates useless for the intended applications. This thesis research aims at developing a high frame rate pulse-echo ultrasound systems for displacement and strain estimation with application to both elastography and thermography.For elastography, a high frequency ultrasound system has been developed for viscoelastic property measurement of thin tissue samples (e.g. skin and subcutaneous tissue, tissue-engineered cardiovascular valves). A dual-element concave transducer was used for localized application of acoustic radiation force (ARF) pulses and monitoring of the induced axial shift. Synchronous silencing of ARF beam and coded excitation technique were used to allow the capture of the full dynamics of tissue deformation in response to the ARF pulse. A modulation unit was also designed to control the temporal behavior of the ARF beam to allow for the investigation of a variety of modulation schemes. A correlation-based speckle tracking algorithm was used to produce spatio-temporal axial displacement maps. The displacement maps were used as input to an on-line tracking of the viscoelastic properties using an extended Kalman Filter (EKF). The EKF simultaneously estimated the state and parameter values utilizing a second order constitutive model of viscoelastic response with unknown parameters. It also allowed for the incorporation of both input and measurement noise models. The former is significant since local variance in absorption and/or beam distortions cannot be predicted precisely in situ.For thermography, a diagnostic scanner was integrated with high intensity focused ultrasound (HIFU) sources to provide real-time 2D high frame rate data system suitable of performing real-time 2D temperature estimation. The front-end of the system was a commercially available scanner equipped with a research interface, which allowed the control of imaging sequence and access to the RF data in real-time. A high framerate 2D RF acquisition mode, M2D, was used to capture the transients of tissue motion/ deformations in response to pulsed HIFU. The M2D RF data was streamlined to the back-end of the system, where a 2D temperature imaging algorithm based on speckle tracking was implemented on a graphics processing unit (GPU). The real-time images of temperature change were computed on the same spatial and temporal grid of the M2D RF data, i.e. no decimation. Verification of the algorithm was performed by monitoring localized HIFU-induced heating of a tissue-mimicking elastography phantom. These results clearly demonstrated the repeatability and sensitivity of the algorithm. Furthermore, in vitro results were presented to demonstrate the possible use of this algorithm for imaging changes in tissue parameters due to HIFU-induced lesions. These results clearly demonstrated the value of the real-time data streaming and processing in monitoring and guidance of minimally-invasive thermo-therapy. |
