Biomedical Engineering Reference
In-Depth Information
dows of actual state-of-the-art PET tomographs are around 4 ns. Therefore it
can happen that two photons that do not originate from the same annihila-
tion process are measured as a coincidence event (dark gray event, Figure 4.9).
These events are called random events or randoms. The rate of random coin-
cidence events along one line of response (LOR) is dependent on the rate of
single events for each detector of the LOR (ri i and r j ) and on the length of
the timing window ():
R ij = 2r i r j :
(4.6)
The overall random rate proportionally depends on the square of the activ-
ity concentration for a given activity distribution in the field of view. Usually,
randoms are distributed homogeneously over the field of view.
FIGURE 4.9: If photons of two independent annihilation processes are mea-
sured in two detectors (dark gray arrows) within the coincidence window of the
PET machine they will be considered as a coincidence event. An event along
a line of response which is not correlated to the two annihilation processes
will be used for image reconstruction (dashed line). Such wrong coincidence
events are called random events or randoms. The light gray arrows represent
a true coincidence event.
4.1.8 Collimator eects|Distance dependent spatial
resolution (SPECT only)
In SPECT collimators are used in front of the scintillation crystals for lo-
calization of incoming photons. In clinical routine parallel-hole collimators of
different materials, thickness and hole sizes depending on the application are
in use. For special applications such as high resolution small animal imaging
or cardiac imaging, other collimator types such as converging, diverging, and
pinhole collimators are in use. The spatial resolution of a SPECT system de-
 
Search WWH ::




Custom Search