Kinetic Analysis of Dynamic PET for Molecular, Functional and Physiological Characterization of Diseases
Abstract
PET with targeted probes may better elucidate the molecular and functional basis of diseases. The widely used standardized uptake value from static imaging, however, cannot quantify the probe uptake processes like perfusion, permeability, binding to and disassociation (k4) from target. The overarching thesis goal is to develop a model to enable kinetic analysis of dynamic imaging to separate these processes.
As perfusion delivery is not modelled in the current standard two tissue compartment (S2TC) model, I developed a flow modified two tissue compartment (F2TC) model that incorporates the blood flow effect. The model’s performances were investigated with simulation. It was applied to derive kinetic parameters of [18F]FAZA binding to highly hypoxic pancreatic cancer. As a validation, the distribution volume (DV) of [18F]FAZA determined with the F2TC and S2TC model were compared with graphical analysis (GA). Kinetic analysis requires arterial concentration of the native probe to model the observed tissue uptake over time, therefore, a method was developed to correct for the metabolite contamination of arterial plasma.
Based on fractional Euclidean distance of estimated and simulated parameters, F2TC model performed better than S2TC model, particularly with longer mean transit time due to the neglect of perfusion effect in the latter model. Also, dynamic acquisition longer than 45 minutes did not improve the accuracy of estimated F2TC model parameters. In the pancreatic cancer study: (a) GA showed that [18F]FAZA was reversibly bound to hypoxic cells; (b) DV estimated by the F2TC and S2TC model was not and was significantly different from GA respectively; (c) k4 and DV estimated by F2TC model could distinguish normal and cancerous tissue with 95% sensitivity. TLC-autoradiography identified metabolites in 2µL of arterial plasma with radioactivity as low as 17Bq. This high sensitivity and the ability to measure multiple (8-12) samples simultaneously could allow metabolite correction of arterial plasma to be performed in individual studies.
Finally, the reversible binding of [18F]FAZA in hypoxic pancreatic tumor cells could be due to efflux of reduced products by the multidrug resistance protein. Therefore, kinetic analysis of dynamic [18F]FAZA PET could monitor both hypoxia and drug resistance for individualized treatment.