Speaker
Description
Creating and maintaining hyperpolarized states is difficult. Unsurprisingly, many of the same limitations are also present when considering biological and medical applications. T1, T2, diffusion or transport to the catalytic site of interest and relaxation processes while bound can all affect our ability to extract useful information from the hyperpolarized sample in vivo. Even when these effects do not preclude the imaging agents that are accessible to hyperpolarization techniques, they may confound straightforward interpretation in terms of metabolic changes of diagnostic value.
For example, in the case of the most widely-used hyperpolarized agent, pyruvate, observation of accelerated intracellular reduction to lactate is observed in many cancer types, and can be used to identify tumors or assess their aggressiveness. However, as is the case with 18F-FDG PET and other molecular imaging methods, quantitative and even qualitative interpretation is very hard to do-- local changes in perfusion, vessel leakiness, vascularity, transport across cell membranes, inflammatory infiltration, intra- or extracellular pH, and local or systemic lactate concentration can all affect the apparent reduction rate, confounding and potentially masking clinically relevant alterations.
In this talk I will discuss our efforts to disentangle these effects using additional measurements and a rational agent selection approach. The process consists of first assessing protein expression and metabolite concentration differences that indicate the hyperpolarization-accessible agent with the best ability to distinguish between health and disease states of interest. We then test the agent in isolated organ or patient-derived xenograft (PDX) models to confirm the expected sensitivity. Next, we simulate agent metabolism using a multi-compartment model and use a Bayesian analysis approach to both support fusion with other data sources, and to evaluate / minimize remaining diagnostic uncertainty. Finally, we will apply this optimized and hopefully specific imaging method to a human study to verify clinical utility. I will detail this approach in the context of our current goal of subtyping Human Hepatocellular Carcinoma to inform precision treatment selection at Penn.
