A K-Space Filtering Technique for Vibration Sonoelastography

Abstract

The aim of this project was to investigate the feasibility of a k-space filtration technique to produce elasticity maps with stationary vibration fields. The method was implemented in a MATLAB routine and tested against numerical and experimental data gathered from an imaging phantom. The method is unable to correctly detect the cyst at frequencies less than 1000 Hz, both in experimental and numerical data. With numerical data at 2500 Hz the method successfully extracts both the speed and the shape of the cyst. Our results indicate that, while the k-filtering approach might be feasible for detection of breast cancer at high frequency, further work is needed. Primarily a more powerful source of vibration must be used. Secondly a 3D acquisition scheme must be used to deal with transversal vibration modes. Finally, gated operation is recommended to reduce discomfort for both patient and operator.