Biomedical Engineering Reference
In-Depth Information
used to identify the crystal element where the g interaction occurred. Each detector
is connected in coincidence with opposing detectors. The line connecting the two
detectors giving a signal is called a line of response (LOR). The events occurring
along each possible LOR are histogrammed into sinograms, or projections, which
are then converted into radioactivity images using analytical or statistical image
reconstruction algorithms. A number of small scintillation crystals are mounted
on PMT. Increased spatial resolution requires smaller scintillation crystals. The
imaging properties, such as uniformity and spatial resolution, have improved with
time as a result of a more efficient camera designs. Commercial tomographs using
BGO crystals reached a spatial resolution of approximately 4 mm. An LSO crystal
has a five fold higher light output and an eight times faster light decay time. This
allows the spatial resolution to be close to its physical limit of 1-2 mm, which is
determined by the positron range in tissue and the small noncollinearity of the an-
nihilation photons. The introduction of empirical digital corrections of energy and
position of events has also improved the image properties.
In order for PET images to provide quantitative information, which is nec-
essary to obtain biologically meaningful results, proportionality between image
count density and radiotracer concentration must be preserved. This requires the
tomograph to be calibrated for effects such as dead time and detector nonuniform-
ity. In addition, accurate corrections for effects that alter the perceived source dis-
tribution must be performed. These effects include attenuation of the
rays (which
leads to source intensity underestimation), Compton scattering of the detected
γ
γ
ray, and detection of random coincidences (which contribute to source position
misidentification by assigning events to incorrect LORs). When all the corrections
are appropriately applied to the data, PET images show the distribution of the ra-
diotracer in units of radioactivity concentration. In addition to providing static im-
ages of radioactivity distribution, PET can also provide information on the change
of radioactivity distribution as a function of time by performing sequential scans
(dynamic scanning). Compartmental modeling (described in Chapter 10) is used to
convert a process observed in terms of radioactivity distribution into biologically
meaningful variables.
8.5.2.3 Single-Photon Emission Computed Tomography (SPECT)
Conventional planar imaging techniques of PET suffer from artifacts and errors
due to superposition of underlying and overlying objects, which interfere with the
region of interest. The emission CT approach called SPECT has been developed
based on detecting individual photons emitted at random by the radionuclide to be
imaged. A major difference between the SPECT and PET is that SPECT uses a ra-
dioactive tracer such as technetium-99 (product of the longer lived molybdenum-99
(
t
1/2
-ray photon
instead of a positron-emitting substance. Photons of energy 140 KeV are released,
which are easily detected by gamma cameras. The same PET camera technology
is used in SPECT where conventional gamma cameras are mounted on a frame
that permits 360° rotation of the detector assembly around the patient. Further,
=
2.8 days) and has a half-life of about 6 hours) that emits a single
γ
