V34(6) 1270-1281 2015, IEEE Transactions on Medical Imaging DOI:10.1109/TMI.2014.2383835
The interaction of ultrasonically-controlled microbubble oscillations (acoustic cavitation) with tissues and biological media has been shown to induce a wide range of bioeffects that may have significant impact to therapy and diagnosis of central nervous system diseases and disorders. However, the inherently non-linear microbubble oscillations combined with the micrometer and microsecond scales involved in these interactions and the limited methods to assess and visualize them transcranially hinder both their optimal use and translation to the clinics. To overcome these challenges, we present a noninvasive and clinically relevant framework that combines numerical simulations with multimodality imaging to assess and visualize the microbubble oscillations transcranially. In the present work, acoustic cavitation was studied with an integrated US and MR imaging guided clinical FUS system in non-human primates. This multimodality imaging system allowed us to concurrently induce and visualize acoustic cavitation transcranially. A high-resolution brain CT scan that allowed us to determine the head acoustic properties (density, speed of sound, and absorption) was also co-registered to the US and MR images. The derived acoustic properties and the location of the targets that were determined by the 3D-CT scans and the post sonication MRI respectively were then used as inputs to two- and three-dimensional Finite Difference Time Domain (2D, 3D-FDTD) simulations that matched the experimental conditions and geometry. At the experimentally-determined target locations, synthetic point sources with pressure amplitude traces derived by either a Gaussian function or the output of a microbubble dynamics model were numerically excited and propagated through the skull towards a virtual US imaging array. Then, using passive acoustic mapping that was refined to incorporate variable speed of sound, we assessed the losses and aberrations induced by the skull as a function of the acoustic emissions recorded by the virtual US imaging array. Next, the simulated passive acoustic maps (PAMs) were compared to experimental PAMs. Finally, using clinical CT and MR imaging as input to the numerical simulations, we evaluated the clinical utility of the proposed framework. The simulations indicated that the diverging pressure waves propagating through the skull lose 95% of their intensity as compared to propagation in water-only. Further, the incorporation of a variable speed of sound to the PAM back-projection algorithm indeed corrected the aberrations introduced by the skull and substantially improved the resolution. More than 94% agreement in the FWHM of the axial and transverse line profiles between the simulations incorporating microbubble emissions and experimentally-determined PAMs was observed. Finally, the results of the 2D simulations that used clinical datasets are promising for the prospective use of transcranial PAM in a human with an 82 mm aperture broadband linear array. Incorporation of variable speed of sound to the PAM back-projection algorithm appeared capable of correcting the aberrations introduced by the human skull. These results suggest that this integrated multimodal imaging approach can provide a physically accurate and thus clinically-relevant framework for developing a comprehensive treatment guidance for therapeutic applications of acoustic cavitation in the brain. Ultimately it may enable the quantification of the emissions and provide more control over this nonlinear process.
Copyright © 2015 IEEE
This article may be downloaded for personal use only. Any other use requires prior permission of the authors and publisher.
Clinic Ultrasound Laboratory (クレメント超音波研究室)
Cleveland Clinic (クリーブランド・クリニック),
Lerner Research Institute
Case Western Reserve University