UCSF

One major challenge in cancer biology is to monitor cancer metabolism in vivo with the goal of improved diagnosis. NMR detection of 13C isotope has the potential to provide the necessary information, but traditional methods have been hampered by the low concentration of the crucial metabolites. Newly emerging hyperpolarization techniques give dramatically increased MR signals, thereby enabling real-time investigation of injected 13C substrates and their downstream metabolic products. The transient nature of the hyperpolarized signal, however, introduces a number of challenges, necessitating the development of accelerated data acquisition schemes.

This dissertation first discusses the application of hyperpolarized 13C metabolic imaging for monitoring the tumor progression and regression in murine cancer models. Then, a number of new methods are presented that improve the research and future clinical value of this molecular imaging technique. In particular, novel data encoding and image reconstruction algorithms were developed based on state of the art techniques such as parallel imaging, compressed sensing, and low-rank matrix completion. The presented framework was applied to in vivo [1-13C]pyruvate metabolic imaging to accelerate data acquisition speed and improve the spatial and temporal resolution for the acquired metabolic data.