Observations of transport processes in dense material as challenge for PET imaging procedures and image reconstruction methods


Observations of transport processes in dense material as challenge for PET imaging procedures and image reconstruction methods

Kulenkampff, J.; Lippmann-Pipke, J.; Zakhnini, A.; Pietrzyk, U.

Understanding transport processes in soils and rocks is crucial for geosciences. These processes are simulated with sophisticated models. However, complex structure, inhomogeneous composition and randomness still produce unsatisfactory predictions. Experimental data is needed for calibration and validation of the model codes and parameters.
PET, as method for spatiotemporal quantitative observation of tracer propagation with unrivaled sensitivity and selectivity, is first choice for providing these data sets during transport experiments. During a decade of research we established “GeoPET” as profitable laboratory method in geosciences, with convincing results.
However, compared to medical imaging, high material density and extended observation periods cause some challenges: We have to consider strong attenuation, a high scatter fraction and interferences from high energy gamma-emissions typical for long living radionuclides, like I-124, Co-58 and Na-22. Consequently, we have to handle images that are prone to artefacts and blurring. These deficiencies are presumably exacerbated by the inhomogeneous sensitivity of our ClearPET-scanner in combination with a disadvantageous normalization method of the applied osmaposl-reconstruction algorithm from the STIR-library (Thielemans et al. 2012).
In order to identify causes of these deficiencies and to improve the methods for correction and reconstruction, we conducted OpenGATE-Monte-Carlo-simulations (Jan et al. 2011, Zakhnini et al. submitted) of our experiments, with according results of experiment and simulation. From the simulation we are able to track the fate of each event, including scattering and detection process. These simulations resulted in some practical advices for the energy window and detector geometry, and the development of an efficient scatter correction method. More involved enhancements suggested from these simulations are additional detectors for Compton-scattering, other detector geometries, and development or application of more robust reconstruction methods.

Keywords: PET; Monte-Carlo-Simulation; flow experiments; imaging

  • Lecture (Conference)
    2012 Workshop “Tomography, data processing and image reconstruction for medicine and engineering”, 10.-12.09.2012, Dresden, Deutschland

Permalink: https://www.hzdr.de/publications/Publ-17654
Publ.-Id: 17654