Models for Dynamic Whole-body Scans

The standard models in PXMOD assume that all slices have the same timing because they were acquired with the scanner field-of-view at a fixed location. Modern PET scanners, however, are introducing dynamic whole-body acquisition functionality. Hereby, the scanner FOV is dynamically moved over the relevant parts of the body during an acquisition. Consequently, the signal is only measured partially for each slice in the volume with temporal gaps when the scanner is not positioned over the slice. This is reflected in the timing of signal, which is unique for each slice. For proper modeling, the slice-dependent acquisition times have to be retrieved correctly from the data, and taken care of in the parametric mapping.

Two dedicated models have been developed for the parametric mapping of dynamic whole-body FDG PET data:

• Patlak analysis: It is essentially the same as the Patlak model for brain mapping, except that slice-dependent timing is allowed for and no VOI delineation is enforced.
• Irreversible 2-tissue compartment model: This is an irreversible variant of the 2-tissue compartment model calculated using basis functions. It supports the fitting of the blood fraction as a fourth parameter, which is obtained at the cost of additional noise.

Note that depending on the reconstruction settings spatial smoothing of the data is required to obtain parametric maps. This smoothing can be configured in the loading dialog window.