factors because contrast-enhanced pixels overestimate attenuation. Some
investigators advocate not using contrast agents and others suggest the use
of water-based contrast agents to mitigate this effect.
Partial-Volume Effect
Partial-volume effects are inherent flaws of all imaging devices, because no
imaging device has perfect spatial resolution. When a “hot” spot relative to
a “cold” background is smaller than twice the spatial resolution of the
imaging device, the activity around the object is smeared over a larger area
than it occupies in the reconstructed image. Although the total counts are
preserved, the object appears to be larger and to have a lower activity
concentration than it actually has. Similarly, a small cold spot relative to a
hot background would appear smaller as if with higher activity concentra-
tion. Such underestimation and overestimation of activities around smaller
objects result from what is called the partial-volume effect.
The partial-volume effect is a serious problem for smaller structures in
images, and correction needs to be applied for the overestimation or under-
estimation of the activities in them. A correction factor, called the recovery
coefficient, is the ratio of the reconstructed count density to the true count
density of the region of interest that is smaller than twice the spatial reso-
lution of the system. The recovery coefficient can be determined by mea-
suring the count densities of different objects containing the same activity
but with sizes larger as well as smaller than the spatial resolution of the
system. Recovery coefficients are usually measured using phantoms which
may not truly be representative of the human body. The measured recov-
ery coefficients are then applied to the image data of the patient to correct
for partial volume effect. The recovery coefficient would be one for the
larger objects.
Center of Rotation
The center of rotation (COR) parameter is a measure of the alignment of
the opposite views (e.g., posterior versus anterior or right lateral versus left
lateral) obtained by the SPECT system. The COR must be accurately
aligned with the center of the acquisition matrix in the computer. If the
COR is misaligned, then a point source would be seen as a “donut” on the
image (Fig. 12.17). Thus, an incorrect COR in a SPECT system would result
in image degradation. For example, an error of 3 mm in the alignment of
COR is likely to cause a loss of resolution of ~30% in a typical SPECT
system (Todd-Pokropek, 1983).
The misalignment of COR may arise from improper shifting in camera
tuning, mechanics of the rotating gantry, and misaligned attachment of the
collimator to the detector. The COR off by more than one pixel may cause
degradation in the reconstructed image. It is essential that the COR align-
ment is assessed routinely in SPECT systems to avoid potential degrada-
Single Photon Emission Computed Tomography 175