Real-Time Detection Of Multiple Lesions During High Intensity Focused Ultrasound (HIFU) Treatments

This work describes the development of a non-invasive real-time technique to detect changes in tissue caused by the production of multiple lesions during a HIFU treatment sequence. It is based on estimation of relative changes in tissue properties derived from backscattered RF data, such as speed of sound, density, absorption coefficient, backscattering power, etc., as a function of HIFU exposure. The HIFU-induced changes were studied using a modified Sonablate HIFU device. It makes use of a confocal 4 MHz pulse-echo ultrasound imaging transducer coupled with a HIFU delivery source on the same crystal. During the HIFU exposure period, every 0.5s the dosage delivery was interrupted for a short time (<80ms) and RF echo signals were acquired using the confocal imaging transducer. Using thermocouples, the temperatures at several locations in tissue were measured to correlate with the changes in tissue parameters. The RF data were digitized using a 50 MHz, 12-bit A/D converter and used for realtime as well as off-line processing and analysis. Several timeand frequency-domain echo RF processing algorithms were developed and tested through in vitro chicken breast and in vivo canine prostate experiments to estimate changes in tissue parameters in real time. Among these, the time-domain algorithms showed better correlation to gross pathology of HIFU-induced lesions in tissue. Work is in progress to implement the algorithm-of-choice in the Sonablate device for clinical applications in human prostate cancer treatments.