The lab is developing a realistic Monte-Carlo (MC) simulation of a dedicated mini-positron emission tomography (PET) cylindrical scanner designed for the direct non-invasive tomographic measurement of the dynamic activity concentration of a PET radiotracer in the arteries, known as arterial input function, crossing through the human ankle or wrist. The MC model will be based on previously validated MC geometries with similar characteristics and will later be used to optimize the details of the ArterialPET scanner geometry before it is built.
Subsequently, the point spread function (PSF) response of the ArterialPET MC scanner model will be evaluated at different key points of the scanner’s transaxial and axial FOV (AFOV) from which the 3D model of the PSF resolution response will be estimated. Finally, a nested PSF-ordered subsets expectation maximization (OSEM) iterative reconstruction algorithm will be developed by nesting the spatially-variant PSF model within each iteration of the OSEM algorithm to accelerate the convergence and improve contrast-to-noise ratio. The findings from this simulation study and nested PSF reconstruction algorithm will be exploited to guide us in the construction of a similar PSF spatially variant model also for an actual ArterialPET scanner in human ankle and wrist.
The lab's main objective is to develop an accurate PSF resolution model that will improve the image quality and signal-to-noise ratio of the PET images, letting the team obtain a highly quantitative, image-derived, non-invasive estimate of the arterial input function in the human ankle or wrist. The lab’s project, as defined with the aforementioned objectives, is part of a larger multi-institutional project to design and develop a dedicated high-sensitivity, high-resolution human ankle or human wrist scanner capable of providing accurate image-derived input function signals to facilitate clinical translation of dynamic and multi-parametric PET imaging in routine standard-of-care PET exams.