|M.Sc Student||Harpaz Shlomo|
|Subject||Differentiation between Benign and Malignant Lesions Using|
|Department||Department of Biomedical Engineering||Supervisor||Professor Emeritus Dan Adam|
|Full Thesis text|
_ Vibro-acoustography (VA) is a dynamic elasticity imaging method in which a small tissue region is excited by a sinusoidal low-frequency acoustic radiation force, causing this region to vibrate at the same frequency. These vibrations produce an acoustic emission field which is detected by a hydrophone. The differentiation between malignant and benign lesions is one of the most challenging tasks faced by clinicians. While this issue has been extensively dealt with in quasi-static elasticity imaging, it has not been addressed before by the VA literature.
In this study a new model was developed, of the acoustic emission from a spherical lesion obtained during VA-imaging. This model takes into account the dependency of the lesion's acoustic impedance on the nature of the lesion-tissue interface. Benign lesions are assumed to slip freely relative to the host tissue, due to their encapsulated nature. Malignant lesions are modeled as firmly bonded to their surroundings due to their highly infiltrative nature. The influence of the interfacial condition on the frequency-dependent value of the acoustic emission is analyzed, and compared to the dependency of this emission on several other parameters. Results reveal that the location of the sphere's resonance frequency is dependent on the nature of the interfacial condition, and that differences in resonance between malignant and benign lesions may be detected using the frequency-sweep method. Furthermore, it is shown that the nature of the interfacial condition must be taken into account in any attempt to use the frequency response of vibrating spheres in order to determine the density of the sphere or the mechanical properties of the embedding medium.
A 2D VA-image is formed by moving the focal point of the transducer along the desired plane, recording the complex amplitude of the acoustic emission received at each position, and then displaying the magnitude of the signal corresponding to each excitation point. This 2D imaging process is modeled in this study as a convolution between the imaging transducer's point spread function and the acoustic-emission-dependent object function. Simulations of the images obtained from a lesion-containing soft tissue region reveal that while the acoustic-emission magnitude images are similar for both malignant and benign lesions, the images of the imaginary and the real parts of this emission show halo patterns which may depend on the nature of the lesion-tissue interface. These results suggest that a novel approach to VA imaging may allow improved differentiation between benign and malignant lesions.