Focus prediction of high-intensity focused ultrasound (HIFU) in biological tissues based on magnetic resonance images

Abstract

High-intensity focused ultrasound (HIFU) is a non-invasive tumor treatment method that works by converting acoustic energy into heat and killing diseased cells at high temperatures in diseased areas of human tissue. The ability to accurately predict the HIFU focus has a direct impact on the safety and efficacy of the treatment procedure. At present the medical community mainly uses MR for guidance during HIFU procedures through temperature feedback and tissue displacement feedback. These systems perform poorly if one wishes to determine whether the HIFU focus is accurately focused on the target lesion prior to the start of treatment. A method is therefore proposed in this paper to accurately predict the HIFU focus in biological tissues. The method mines the mapping relationship between grey-scale information and acoustic parameters of magnetic resonance images of biological tissues from a novel perspective; after which the HIFU focus shift is predicted and experimentally validated. First, we performed MRI imaging on pork tissues (muscle, fat, and skin), and fitted a mathematical relationship between the grayscale of the magnetic resonance image and the acoustic parameters. Second, we reconstructed a 3D structural model of pork tissue based on magnetic resonance images. Third, we predicted the HIFU focus by combining k-Wave and 3D structural model. Finally, we performed actual experiments to measure the focal position of the HIFU sound field and compared it with simulation results. The comparison shows that the deviation between the focal point simulated by k-Wave and the actual measured focus is less than 0.30 mm on all acoustic axes in the X-, Y- and Z-direction, which falls within the acceptable range.