Evaluation of the HRRT and the HR+ for the task of relative region analysis using a realistic head and brain phantom

PET imaging with scanners such as the HRRT and the HR+ enables functional images of the binding of radioligands to neuroreceptors to be obtained, which is of great utility in neuroscience research. Usually a time series of images is acquired in order to apply a kinetic model to the time-changing concentrations of the radiotracer within different volumes of interest. Whether fully quantitative kinetic modeling methods are employed or simpler reference tissue methods, it is of paramount importance to ensure that the regional distributions of the radiotracer concentration are accurate. However, it is common practice to use simple uniform cylinders as a form of routine quality control to assess reconstruction accuracy, or perhaps simplified phantoms (like the Hoffman brain phantom), but these do not have attenuation and scatter properties which are comparable to the case of imaging the human head. Instead, this work uses an anthropomorphic head and brain phantom with realistically complex attenuation and scatter (yet with a simple large uniform emission region) to assess regional uniformity of reconstructed images for the HRRT (when using 3D OP-OSEM reconstruction) and for the HR+ (using DIFT reconstruction). For the HRRT it was found necessary to change the reconstruction parameters for the mu map and also to specially tune a parameter for the single scatter simulation in order to improve the uniformity of the HRRT images such that a relative region analysis produced errors below 5%. For the HR+, the importance of transmission scanning without radioactivity present in the FOV is emphasized, in order to carry out reliable relative region analyses (with errors <;5%). Relative region analyses are expected to indicate performance for the task of kinetic modeling when using a reference region, such as the cerebellum for imaging of raclopride kinetics.