Monitoring imaging of lesions induced by high intensity focused ultrasound based on a matching pursuit method

We investigated the feasibility of a monitoring imaging method to visualize and evaluate the high intensity focused ultrasound (HIFU) induced lesion in vitro during and after their formation, which was based on a modified matching pursuit algorithm (MP). The method is based on the concept of sparse signal approximation in an overcomplete dictionary composed of unit-norm vectors that mimic the ultrasonic pulse-echo to be estimated. Each ultrasonic echo is principally represented by a unit-norm vector and therefore has a physical interpretation. we use the MP method to analyze noisy ultrasonic echo obtained in different stages during and after HIFU treatment and decompose the noisy echo into unit-norm vectors. The approximation of the echo can be obtained by the reconstruction result with several large coefficient unit-norm vectors. In the current research we compared lesion sizes obtained from our method, from measurement of physical lesions after exposure, ultrasonic integrated backscatter (IBS) images and from conventional B-mode images, no matter with or without hyperecho. It was found in freshly excised bovine livers that more precise definition of the treated area was obtained from the MP images than from the IBS images and the conventional B-mode images. When comparing to physical lesion sizes, there was no significant difference. We demonstrated the feasibility of our proposed MP-based method for the localization of HIFU-induced lesions immediately after HIFU treatment. Meanwhile our method can detect lesion locations even without the appearance of hyperecho.