Corrections for Physical Factors - Correction Techniques in Emission Tomography - page 95

in scatter estimation. A refined approach taking this into account has been
proposed by pre-correcting the emission data for scatter using the convolution-
subtraction method [90].

5.5.4 Iterative reconstruction methods
Fast iterative image reconstruction algorithms, especially the ordered sub-
sets expectation maximization (OSEM) algorithm, have become important
competitors of analytical methods like filtered backprojection (FBP) during
the last years. One of their properties is the different convergence rate of im-
age features that have different spatial frequencies; low-frequency image com-
ponents will converge substantially faster than high-frequency components
as demonstrated by Fourier analysis of these algorithms [76]. As the scatter
distribution in emission tomography imaging is assumed to have only low-
frequency components, an early iteration in the reconstruction process may
be a good estimate of the spatial distribution of scattered events. This ap-
proach has been successfully introduced in both SPECT [56] and PET [90]
using one OSEM iteration.

5.6 Concluding remarks
Both PET and SPECT potentially offer the ability to absolutely quantify
tracer distributions in vivo. For obvious reasons, this can be achieved only by
a fundamental understanding and modeling of all relevant physical processes
from radiation formation to radiation detection. Especially the progress in
the field of new hybrid tomography systems, such as PET/CT and PET/MR
scanners, necessitates the development of new correction methods (specifically
methods correcting attenuation and scatter). Additionally, as the intrinsic
spatial resolution of emission tomography systems is continuing to increase,
proper correction for motion-related effects will be of growing importance in
both clinical practice and pre-clinical research. Furthermore, ongoing increase
of computational power allows more realistic models to be incorporated into
the reconstruction process. Therefore, corrections for physical factors will re-
main an important part of future research in emission tomography.

References
[1] R. Accorsi, L.E. Adam, M.E. Werner, and J.S. Karp. Optimization of
a fully 3D single scatter simulation algorithm for 3D PET.

Physics in

Medicine and Biology, 49:2577{2598, 2004.