The Nuclear Physics group at the University of Edinburgh are looking at new ways to analyse Positron Emission Tomography (PET) data. Working alongside medical physicists at the Edinburgh Imaging Facility QMRI their approach involves making use of signals that are recorded in PET machines, but are not at presently utilised in image reconstructions.
In PET imaging the noise created by false coincidences (the detection of two photons not originating from the same event) limits the rate of activity of the radioactive source that can be given to patients. This means that it typically takes about 30 minutes to acquire the volume of data required to create an image with sufficient spatial resolution to carry out diagnoses. Reducing the time required to perform a scan would mean that more patients could be scanned on the same machine each day thus bringing down the cost per patient.
The new method being investigated utilises physical processes that photons undergo in a way that is not currently made use of in PET imaging but is applied in nuclear physics research. Early results indicate that the idea will pass the proof of concept stage. The bulk of work is being done using Monte Carlo simulations in the GAMOS (Geant4 Architecture for Medical Orientated Simulations, framework. A small prototype is also being developed in the lab. In the future the new technique will be applied to data acquired at Edinburgh Imaging Facility QMRI.