HIFU lesion characterization on liver: Acquisition and results

Journal article

A key problem in the practical use of high intensity focused ultrasound (HIFU) as a tool for cancer treatment is the non-invasive characterization of the regions of tissue that have been successfully necrosed. Previously, we proposed an approach to image guidance based on the use of echo data obtained from a diagnostic ultrasound transducer and a shape-based inverse scattering approach. Here an experimental apparatus is described to acquire data for in vitro phantom experiments and ex vivo experiments with HIFU-induced lesions in porcine liver. The in vitro studies employed tissue-mimicking phantoms with inclusions embedded to simulate HIFU lesions. Using just 5 elements of a 3.5-MHz ultrasound array, the inversion was able to recover the properties of lesions down to a size of 20 mm long by 4 mm in diameter. Experiments were then carried out with fresh excised porcine liver. A 1.1-MHz HIFU transducer was used to create lesions in the liver, and the tissue was then scanned with the ultrasound array. After imaging the tissue was sliced and the true lesion geometry assessed optically. The echo data were then processed to determine the lesion from the data. It was found for lesions with cavitation that the hyperechogenecity from the bubbles made the inversion unable to find the lesion. For the HIFU lesions formed with no cavitation, the lesions had a diameter of less than 3 mm and were too small to be detected with the 5-element array. These results indicate that the shape-based inversion process is promising for detecting purely thermal lesions but that more elements may be needed for detection of HIFU lesions. © 2009 American Institute of Physics.

Authors: Durning, B; Raymond, J; Church, C; Cleveland, R; Miller, E

• Journal: AIP Conference Proceedings
• Volume: 1113
• Pages: 96-100
• Publication date: 2009-01-01
• DOI: 10.1063/1.3131479
• EISSN: 1551-7616
• ISSN: 0094-243X
• Language: English
• Keywords: HIFU detection; Inverse problem; Model-based