UC Berkeley

Magnetic particle imaging (MPI) is an emerging tracer imaging modality. In this work, we develop novel scanning strategies to improve the imaging performance of MPI and explore novel methods to use MPI for theranostic applications. A major focus is inventing and exploring different scanning strategies to overcome or leverage the magnetic relaxation dynamics associated with the tracers used in MPI in order to find optimal trade-offs in imaging performance metrics such as spatial resolution, SNR and scanning time. In the first part of this thesis, we perform an experimental study to underscore the significant discrepancy between experimental performance of large core size tracers and their theoretical performance, and demonstrate that this is a major obstacle to improving MPI performance. Subsequently, we describe the hardware design and construction of a frequency-flexible, arbitrary waveform tabletop scanner to enable the investigation of novel scanning strategies to help address this issue. Conventional MPI uses a single frequency excitation wave around 20 kHz. With this new device, we show the first study of non-sinusoidal excitation waveforms such as square, trapezoidal, triangle waveforms, and based on this, we design a novel approach to scanning and signal encoding in MPI which we name pulsed MPI (pMPI). Subsequently, we show that pMPI unlocks the potential of large core size nanoparticles by suppressing deleterious relaxation-based blurring effects. This enabled the first viable use of these nanoparticles which previously have been unusable in MPI. We also exploit the frequency-flexibility to optimize parameters for continuous (sinusoidal) wave MPI and elucidate an optimal waveform that achieves both good resolution and SNR where previous work has suboptimally traded-off one for another.

In the second part, we perform extensive preclinical studies as the first in vivo proof-of-concept of theranostic MPI. We show that MPI is unique as a theranostic modality due to many key advantages such as precise localization of therapy in vivo, an ability to receive measure and predict dosage based on image-guidance, and finally elucidate an approach for future development of real-time feedback for fine control and definite quantification of dosage. Our in vivo proof-of-concept results show robust localization of the heating dose deposited as well as therapeutic outcomes in a dual tumor xenograft rodent model. Importantly, we prove that this localization method can address one key challenge in Magnetic Hyperthermia, which is to avoid collateral damage to off-target organs such as the liver which tends to accumulate magnetic nanoparticles and is prone to unintended heat damage. Furthermore, we investigate aerosolized magnetic nanoparticles as a method to image the lungs with MPI. We demonstrate that MPI lung imaging with magnetic aerosol can be a viable alternative to clinically established radioaerosol procedures. This lays important groundwork for use of magnetic aerosols and MPI for safer lung imaging and lung theranostics in combination with the abovementioned work.