Quantitative Techniques for PET/CT: A Clinical Assessment of the Impact of PSF and TOF

Tomographic reconstruction has been a challenge for many imaging applications, and it is particularly problematic for count-limited modalities such as Positron Emission Tomography (PET). Recent advances in PET, including the incorporation of time-of-flight (TOF) information and modeling the variation of the point response across the imaging field (PSF), have resulted in significant improvements in image quality. While the effects of these techniques have been characterized with simulations and mathematical modeling, there has been relatively little work investigating the potential impact of such methods in the clinical setting. The objective of this work is to quantify these techniques in the context of realistic lesion detection and localization tasks for a medical environment. Mathematical observers are used to first identify optimal reconstruction parameters and then later to evaluate the performance of the reconstructions. The effect on the reconstruction algorithms is then evaluated for various patient sizes and imaging conditions. The findings for the mathematical observers are compared to, and validated by, the performance of three experienced nuclear medicine physicians completing the same task.