Ultrasound imaging is an indispensable tool in medical diagnostics and offers several advantages over other imaging techniques. During last decades, ultrasound technology has achieved significant progress leading to improved image quality. However, conventional ultrasound envelope-detected images remain qualitative in nature and lack quantitative information on the tissue state and pathology. Therefore, there has been an increased interest in developing techniques that would enhance the diagnostic capabilities of ultrasound imaging. This field of study is referred to as ultrasound tissue characterization and aims to improve ultrasonic diagnostic capabilities by quantitatively measuring physical properties that can be linked to the tissue state. Current state-of-the-art methods for ultrasound tissue characterization focus on one particular parameter, and attempt to estimate it directly from the backscattered RF signals. Other acoustic effects are being corrected for by means of additional measurements, or simply neglected. The significant number of underlying assumptions in this case makes application of the existing techniques difficult in clinical practice. Moreover, poor understanding of the interaction mechanisms of ultrasound with tissue remains the main challenge in ultrasound tissue characterization. The aim of this thesis was to develop and validate a fundamentally different approach for reconstruction of the local acoustic properties, wherein the forward scattering problem is iteratively solved through computer simulations in order to match the synthetically generated ultrasound data to the experimentally observed ones. The advantage of such approach is that numerical modelling of the forward ultrasound wave propagation enables to study various ultrasound-tissue interactions as well as their combined effect in a controlled manner. The developed approach was used for the reconstruction of an ultrasound attenuation coefficient and was extensively validated at increasing level of complexity of the considered problem. This work demonstrates the applicability of the proposed model-based approach in the field of ultrasound tissue characterization and its ability to provide accurate attenuation estimates in various settings. Furthermore, a simulation tool was developed, that enables modelling of the forward ultrasound wave propagation and the spectral characteristics of the backscattered signals.
|Qualification||Doctor of Science|
|Date of Award||3 Jul 2017|
|State||Published - 3 Jul 2017|