UCLA

Image-guided thermal therapies, including radiofrequency ablation (RFA), microwave ablation (MWA), and high-intensity focused ultrasound (HIFU), offer a curative option for patients with early-stage liver cancer who are not candidates for surgical resection or transplantation. These therapies deposit thermal energy through an interstitial applicator or an external transducer under image guidance to induce coagulative necrosis within the tumor. Magnetic resonance imaging (MRI) is uniquely suited to guide these procedures, providing superior soft-tissue contrast for treatment planning and targeting, continuous device guidance without ionizing radiation, non-invasive temperature mapping by proton resonance frequency (PRF) shift thermometry, and immediate post-procedural assessment of the ablation zone. An MRI-based workflow of planning, monitoring, and evaluation has been established and has demonstrated early feasibility. However, wider adoption is limited by critical challenges at each stage of this workflow. For planning, computational models have been explored to simulate and predict the ablation zone, but they lack sufficient experimental validation. For monitoring, PRF thermometry of abdominal organs during energy delivery is compromised by respiratory motion and electromagnetic interference (EMI) from energy-delivery systems, both of which corrupt image quality and temperature measurements. For evaluation and targeting in therapies demanding high precision, such as MRI-guided HIFU, image contrast and sensitivity remain suboptimal for small features such as small tumors and tumor margins, and a contrast-enhancement technique operating in the clinically relevant therapeutic temperature range is lacking. As a result, the success of MRI-guided thermal therapy remains largely empirical and operator-dependent, restricting reproducibility and confining these procedures to a small number of specialized centers. This dissertation addresses these technical challenges by developing and evaluating computational modeling, MR thermometry, and contrast-enhancement techniques for guiding thermal therapies. First, a joint computational–experimental framework was developed to characterize MWA energy-delivery strategies, validating the computational model against experimental measurements and supporting protocol optimization. Second, a volumetric PRF MR thermometry framework was developed that combines stack-of-radial MRI, image-based motion tracking, and image-navigated multi-baseline correction, extending temperature monitoring to near-full-liver coverage at high spatiotemporal resolution during breathing. Third, a software-driven active EMI suppression technique was developed that removes microwave-generator interference in the k-space domain, restoring image quality and reliable in vivo thermometry without hardware modification or interrupted heating. Fourth, a contrast-enhancement technique was developed using thermoresponsive polymer-modified nanoparticles that reversibly modulate MRI contrast under HIFU stimulation, improving lesion detection at body temperature. Together, these advances could help move planning, monitoring, and evaluation from empirical practice toward a more quantitative, precise, and repeatable workflow, potentially supporting safer and more predictable MRI-guided thermal therapies and their broader clinical translation.