Biomedical Engineering Reference
In-Depth Information
solution of MLEM [ 23 ]. The basic idea of this method is to use data blocks that
are successively employed to calculate the estimates of the image. One iteration is
divided into several steps that requires less data but sufficient to update the estimated
image. Hence, at the end of an iteration, which corresponds to the use of all available
data, the image was already updated as many times as the number of groups utilized,
allowing a significant acceleration of the process.
3
Motion Correction
One of the vital problems during the acquisition of biomedical images is to ensure
that the object is still relatively to the camera, to avoid image degradation. This
is particularly important in children or patients who have movement disorders.
Although patients are instructed to remain still during the scan, it is not always
possible and sometimes it is necessary to use restraints systems. This problem gets
worse on techniques that require long acquisition times, typically nuclear imaging
techniques such as PET and SPECT. Movement promotes significant reduction in
resolution, severe augmentation of partial volume effects, reducing the accuracy
of quantitative methods, alters the pattern definition of metabolic abnormalities
and impacts negatively on patient management. Consequently, motion correction
techniques become essential.
3.1
Effect of Motion in Emission Imaging
The effect usually attributed to motion is blurring of the moving objects, which
implies spread of the activity and loss of edge definition. These effects result both in
the appearance of artifacts and in the reduction of the detectability of small lesions.
Figure 12 illustrates the effect of the motion of a point source. The emission is
made from different points over time, which due to the temporal resolution of the
gamma detector, are impossible to discriminate. As a result the source is seen as a
larger source with less activity.
Another important aspect to consider is the relationship between the amplitude
of motion and the spatial resolution of the detector, since, for situations where the
amplitude is much smaller than the resolution distance, the movement will have a
negligible effect
Murase et al. [ 24 ] were the first to study systematically the impact of motion
in SPECT. They evaluated by simulation and with real tests, using adequate
phantoms, the impact of respiratory motion in SPECT imaging and concluded that
for an amplitude of 14 mm the contrast in lesions with 2 cm diameter decreased
approximately 20% when compared to the situation without motion.
Currently, it is well established that respiratory movements can cause artifacts
in myocardial perfusion SPECT exams with direct implications for the diagnosis
[ 25 - 33 ]. Concerning this result, Matsumoto et al. concluded that the artifacts created
Search WWH ::




Custom Search