Biomedical Engineering Reference
In-Depth Information
(for a review of transmission scanning in emission tomography, see Bailey
63
).
The transmission data are similar to that measured in a CT scan, except that
in emission tomography a gamma emitter is used for the source rather than
an x-ray tube. Ideally, the data are acquired simultaneously with the emission
data to reduce scanning duration and ensure accurate registration. The cor-
rections are applied before reconstruction in PET and either before, during,
or after reconstruction in SPECT.
5.4.1.2.2 Reconstruction Artifacts
The reconstruction algorithm may also produce artifacts, especially where
there are extreme differences in the radioactive concentration between adja-
cent structures. An example often seen is when a radiotracer is excreted via
the kidneys, and the bladder may contain high concentrations of urine com-
pared with the concentration of the radiotracer in surrounding tissues. Itera-
tive reconstruction methods deal with this situation better, in general, than
analytical filtered back projection.
5.5
Spatial Inaccuracies in 3D Ultrasound Imaging
When performing an ultrasound scan, the transducer is moved so that a 3D
image of the tissue structure is built up in the brain of the operator. However
in order to analyze this information or for anybody else to visualize it, meth-
ods for the acquisition and display of 3D ultrasound (3D-US) images have
been developed. An additional benefit of 3D-US is the display of image
planes which are not accessible in 2D ultrasound; i.e., those corresponding to
transducer positions and orientations which are physically unrealizable. This
might include, for instance, looking down the neck of a patient as if the head
had been chopped off!
A 3D ultrasound image is created by moving an ultrasound beam through a
volume of tissue and acquiring a series of 2D images (slices) at a fixed time inter-
val. For each slice, the position and orientation of the transducer is recorded.
There are two principal acquisition methods used, which will be referred to as
freehand
and
fixed
geometry. In freehand scanning the transducer movement is
controlled by the operator in the same way as for conventional 2D ultrasound.
The position and orientation of the transducer is recorded by a position sensor
attached to the transducer. In fixed geometry scanning the transducer move-
ment is partially or completely controlled by the scanner or other device con-
trolling the acquisition. In this case a position sensor is not used.
After acquisition has completed, the ultrasound data must be reconstructed
so that a series of parallel slices is produced. The algorithm used for reconstruc-
tion depends on the acquisition method as well as a number of other factors.
More information on acquisition and reconstruction methods is contained in
Barry et al.
64
