Tuesday, 11 January 2011, 10.30 s.t., SFS
Fabian Gigengack works at the European Institute for Molecular Imaging (EIMI), University of Münster.
Title Motion Correction in Gated Cardiac PET
Abstract Different imaging techniques have been developed and improved during the last years and decades to get better insight into the human body. Compared to (primarily) morphological imaging techniques like CT or MRI, Positron Emission Tomography (PET) makes it possible to look at the metabolism, i.e. functionality, of the human body. This property of PET makes it possible to visualize for example the metabolism of glucose which helps to examine coronary artery diseases.
However, a problem of PET is image degradation due to patient and organ motion caused by relatively long acquisition times. This motion problem can be tackled with image processing techniques. For example list-mode-driven gating along with motion estimation techniques can be applied to compensate for respiratory and cardiac motion induced blurring.
While respiratory motion is more or less locally rigid, cardiac motion is highly non-linear, causing partial volume effect induced intensity irregularities. To overcome these intensity irregularities, the mass-preserving (respectively radioactivity- preserving) nature of PET images has to be taken into account. The preservation of radioactivity means that the total amount of radioactivity is approximately constant over all gates.
A variational registration framework will be presented where the radioactivity-preserving property of PET is considered in terms of a new transformation model. The method will be illustrated and validated on patient data.
Contact: Lars Ruthotto