Publications Repository - Helmholtz-Zentrum Dresden-Rossendorf

1 Publication

A Python-based interface for batch processing of image reconstruction jobs for Philips PET systems

Maus, J.; Schramm, G.; Oehme, L.; Hofheinz, F.; Petr, J.; van den Hoff, J.

Retrospective manual processing of large amounts of patient data is time consuming and error prone. This is especially true for repeated image reconstructions since the vendor software usually does not allow batch processing. To overcome this limitation we have developed a batch processing interface for our Philips PET system that is also suitable for application in a clinical context.
We have used Python3 to develop a framework that allows to run reconstructions with vendor-provided protocols and to modify parameters beforehand. The framework includes functions to list patient studies, run/abort a reconstruction, copy patient data from/to servers and to modify incorrectly entered parameters like injection time or dose. The framework targets current Philips PET systems. It has been tested with a Ingenuity-TF PET/MR system consisting of a EBW workstation and PRS+CIRS recon servers. In addition to basic image reconstruction tasks the framework also allows to add custom user defined data processing. In our case we added a MR-based attenuation segmentation algorithms as well as amplitude-based respiratory gating methods which we developed in-house. Furthermore, to facilitate the use by clinical staff the framework also comes with PyQt5-based graphical user interface.
The framework transparently interacts with the recon servers of the vendor. Images reconstructed with both interfaces were verified to be identical. MR-based attenuation segmentation and respiratory gating were also validated. The framework can be used in batch processing environments either via direct use of the implemented functions, the supplied command-line tools or via the graphical user interface.
The developed framework/tools are compatible with all common Philips PET systems and greatly facilitates reprocessing of large amounts of patient data.

  • Poster
    54. Jahrestagung der Deutschen Gesellschaft für Nuklearmedizin, 20.-23.04.2016, Dresden, Deutschland

Publ.-Id: 23552