UC San Diego

Detecting, quantifying and visualizing biochemical mechanism in a living system without perturbing function is the goal of the instrument and algorithms designed in this thesis. Biochemical mechanisms of cells have long been known to be dependent on the signals they receive from their environment. Studying biological processes of cells in-vitro can vastly distort their function, since you are removing them from their natural chemical signaling environment. Mice have become the biological system of choice for various areas of biomedical research due to their genetic and physiological similarities with humans, the relatively low cost of their care, and their quick breeding cycle. Drug development and efficacy assessment along with disease detection, management, and mechanism research all have benefited from the use of small animal models of human disease. A high resolution, high sensitivity, three-dimensional (3D) positioning positron emission tomography (PET) detector system was designed through device characterization and Monte Carlo simulation. Position-sensitive avalanche photodiodes (PSAPDs) were characterized in various packaging configurations; coupled to various configurations of lutetium oxyorthosilicate (LSO) scintillation crystals. Forty novelly packaged final design devices were constructed and characterized, each providing characteristics superior to commercially available scintillation detectors used in small animal imaging systems: 1̃mm crystal identification, 14-15% of 511 keV energy resolution, and averaging 1.9 to 5.6 ns coincidence time resolution. A closed-cornered box-shaped detector configuration was found to provide optimal photon sensitivity (1̃0.5% in the central plane) using dual LSO- PSAPD scintillation detector modules and Monte Carlo simulation. Standard figures of merit were used to determine optimal system acquisition parameters. A realistic model for constituent devices was developed for understanding the signals reported by the detector system and to create event construction and utilization algorithms. To increase quantitative accuracy in the reconstructed images acquired by the system, a component- based normalization algorithm was developed