|Abstract:||Whole-body positron emission tomography (PET) has been widely used for cancer diagnosis and treatment, but its low spatial resolution limits the management of cancers in specific organs or regions. To promote the pre-clinical and clinical research and improve the diagnosis accuracy and treatment out- comes, organ-dedicated PET systems, such as brain-dedicated scanners, or breast-dedicated scanners, have been proposed and built.
Head and neck cancer (HNC) is collectively a group of cancers that usu- ally begin in mucosal surfaces inside the head and neck. Due to the complex anatomical structure and vital physiological role of the tumor-involved regions, the goal of HNC treatment is not only to improve survival outcomes but also to preserve organ function. A higher spatial resolution HNC imaging will allow radiation oncologists to accurately measure the boundaries of tumors, design the planned target volume dose, and thus offer more freedom to choose from treatment options such as surgery, radiation therapy, chemotherapy and targeted therapy.
To achieve better management of HNC, this thesis proposes an organ- dedicated PET scanner with a focus on HNC. Based on the features of previous published whole-body systems and organ-dedicated systems, the HNC system is designed to have a two-panel geometry and a follow-on scan protocol. System performance including photon sensitivity, noise equivalent count (NEC) rate, spatial resolution, and hot rod visualization are evaluated through Monte Carlo simulation. The results show that superior performance is achieved when compared with a whole-body system. Specifically, the sensitivity is 0.71%. Given a 2-mm depth of interaction (DOI) resolution, the system can achieve a 1-mm in-panel and 2-mm orthogonal-panel spatial resolution. The NEC rate is 10.4 kcps at 5.7 kBq/cm3, and 2-mm diameter hot rod is visualizable.
Based on the simulation result, a high-resolution dual-ended readout PET detector is designed. The crystal is lutetium-yttrium oxyorthosilicate (LYSO), and the crystal size is 1×1×20 mm3. The detector is silicon photomultiplier (SiPM), and the SiPM channel size is 3×3 mm2. A light guide is inserted between the LYSO and the SiPM for crystal identification, and the optimal light guide thickness is evaluated as 1.2 mm. The detector is cooled by a Peltier element, and the detector performance is characterized at 10 depths. The results show that all crystal can be resolved, and the energy, timing, and DOI resolution are measured as 15.66% at 511 keV , 602.98 ps, and 2.33 mm respectively. To optimized the detector geometry, two dual readout cables are further designed to put the readout electronics on the same side of the LYSO. The energy, timing, and DOI resolution of the optimized detector is measured as 16.13% at 511 keV , 658.03 ps, and 2.62 mm respectively. When comparing with previously published high-resolution dual-ended read- out PET detectors, our results show good energy and DOI resolution, and the best timing resolution.
The optimized detector is further scaled-up to a two-panel sub-scanner. Each panel has 2×2 detectors, and the panel size is 53.8×57.8 mm2. The panel distance is 107.4 mm in this preliminary experiment. Spatial resolution at the field of view (FOV) center, 5, 10, 15, 20 mm away from the center along the in-panel and orthogonal-panel axes are measured. Specifically, the sub-scanner can achieve a 1.94 mm in-panel and 4.44 mm orthogonal-panel spatial resolution at the FOV center. Compared with other organ-dedicated PET systems, good in-panel spatial resolution is achieved. In the future, the modular panel can be utilized to scale up to more complex systems in the future.