Optimizing passive cavitation mapping by refined minimum variance-based beamforming method: Performance evaluations in macaque models

Abstract

Microbubble-mediated focused ultrasound (FUS) therapies harness mechanical and/or thermal effects to deliver drugs or ablate tissues. Passive acoustic mapping (PAM) enables the spatio-temporal monitoring of cavitation activity, which is critical for the clinical translation of this technique. Traditional PAM is based on delay-and-sum (DAS) beamforming, a method whose quality tends to deteriorate due to issues including multi-bubble interference, distortion in the wavefront caused by the presence of the skull, unmodeled variability of array elements, etc. To provide for robustness, here we consider the use of minimum variance adaptive beamforming to PAM and demonstrate significant improvement in image quality compared to DAS. The minimum variance distortionless response (MVDR) method was evaluated and further improved by adding diagonal loading and by using subarray covariance estimates. Results demonstrate improvements in both the resolution and image contrast compared to DAS using either traditional or a refined MVDR beamformer. The axial full width at half maximum of the microbubble activity at the focus was reduced to 79.5% and 38.5% of that in DAS image for traditional and refined MVDR beamformers, respectively. Moreover, the refined MVDR method greatly enhanced the robustness while traditional MVDR beamforming induced self-nulling effects. We anticipate that the proposed method will improve our ability to monitor and control FUS-induced cavitation-based therapies.