PET is widely adopted in clinical oncology to investigate the biochemical characteristics of malignant lesions. This thesis focused mainly on the quantitative analysis of PET data, which could be of value to clinical physicians and researchers. Chapter 1 briefly introduced the background and outlines of the studies conducted in this thesis. In Chapter 2, we focused on the development and evaluation of a novel and robust method for automatic segmentation using an active contour model (MASAC), providing a useful tool to delineate metabolically active tumor volume (MATV) in PET images. Chapter 3 investigated the variability and repeatability of quantitative metrics as a function of segmentation method, user interaction, uptake interval and reconstruction protocol, to understand the potential relationships among these aspects. In Chapter 4, we developed a novel dynamic whole-body (WB) anthropomorphic PET simulation framework to assess the potential benefits of dynamic PET imaging. Chapter 5 investigated the impact of tissue classification in MRI-guided attenuation correction (AC) on WB Patlak PET/MRI. In summary, this thesis provided novel and useful tools for PET imaging. It also carried quantitative analysis on both conventional SUV and parametric Ki images. The proposed tools and results presented in this thesis may be used as a guide to quantify various types of oncological malignancies in PET imaging. Further studies are required to establish a benchmark with different imaging procedures on various PET scanners to evaluate their performance in different clinical scenarios.
|Qualification||Doctor of Philosophy|
|Place of Publication||[Groningen]|
|Publication status||Published - 2019|