Details

Autor(en) / Beteiligte

• Wei, Shouyi

Titel

Development of PET Imaging Technologies for Organ-Specific Applications

Ort / Verlag

ProQuest Dissertations & Theses

Erscheinungsjahr

2020

Quelle

ProQuest Dissertations & Theses A&I

Beschreibungen/Notizen

• Positron emission tomography (PET) is an important medical imaging modality that provides in vivo detection and measurement of radioisotope-labeled molecules. While whole-body PET scanners are widely available, they are expensive and not optimized for the imaging of specific regions of the body. Organ-specific PET imaging generally takes advantage of higher-resolution detector technologies and has been applied to clinical applications such as breast cancer detection and brain imaging. Our group has been dedicated to the advancement of organ-specific PET imaging technologies in PET system design, data processing, image reconstruction, and novel clinical applications. In this work, we first present the design and performance measurement of a high-resolution organ-specific PET system—VersaPET. The scanner has achieved ~1.9mm spatial resolution at the center of field of view (FOV), and overall good sensitivity, count rate performance and detector gain stability required by clinical applications such as breast PET imaging. The image reconstruction for VersaPET becomes challenging because the scanner has block detector architecture which causes substantial transaxial and axial gaps in its geometry, and increased parallax error due to its compact geometry. In order to cope with these hurdles, we developed and evaluated a GPU-accelerated image reconstruction framework for VersaPET which incorporates the exact block geometry as well as point spread function (PSF) modeling with full data correction including normalization, attenuation correction and scatter correction. We then examined the ability of the VersaPET system to measure human blood input function simultaneously with 18F-FDG PET/MRI brain imaging, in order to provide a noninvasive alternative to existing invasive arterial blood sampling. We collected prospective human data, developed appropriate data processing methods, and compared performance between the methods at the levels of input function as well as final outcome measures determined by kinetic modeling. Finally, we advanced the application of breast cancer imaging by collecting pilot human data with VersaPET and developing new detector concepts to improve imaging performance and expand the field of view to axillary lymph nodes which are critical in the assessment of breast cancer. We evaluated novel scanner geometries with limited angle sampling and taking advantage of modern detector approaches, including time of flight (TOF) and depth of interaction (DOI) readouts, using Monte-Carlo simulation and detection-based tasks using channelized Hotelling observer (CHO).