Myocardial perfusion SPECT
is a widely-used, non-invasive method for the diagnosis and
management of patients with coronary disease. However,
non-uniform photon attenuation, Compton scatter, limited and
depth-dependent spatial resolution, as well as image noise,
limit the ability of SPECT to obtain images that reliably
represent the true tracer distribution. The non-uniform
attenuation of the thorax is the most significant factor
limiting the diagnostic efficacy of myocardial SPECT.
- Receptor Quantification as a Function of Uptake Ratio(s)
- Partial Volume Effects
- Scatter and Attenuation - Correction Schemes
- Threshold for Changes in Uptake(s)
- Comparison of Different Acquisition Modes, e.g. 2D vs. 3D Pet(s)
- Design of Different Reconstruction Strategies
- Testing and Validation of Image Registration Techniques
- Design of Imaging Protocol for Patients
currently used attenuation, scatter and resolution correction
methods are suboptimal, since they do not provide improvement
in the 25% false-negative findings in a group of about 100
patients with luminal diameter stenoses of at least 50%(1).
Furthermore, the ability to detect multivessel disease was 70%
without and 47% with corrections. This finding implies that
myocardial SPECT can seriously underestimate the extent of
disease in high-risk patients. On the other hand, the
false-positive findings in the group with a low probability of
coronary disease were reduced from 14% without corrections to
3% with corrections.
Obviously, further improvements in both hardware and software
for myocardial SPECT are necessary before this important
diagnostic technique can realize its full potential. These
improvements must be carefully evaluated on realistic,
anthropomorphic phantoms to improve results in clinical
The thorax is molded of polyurethane, modified for
tissue-equivalence, with a mass density of 1.10 g/cc. The
narrow beam linear attenuation coefficient measured at 140 keV
(Tc-99m) is 0.160 cm -1.
The skeleton, embedded in the soft tissue, extends from the
suprasternal notch down to L2. The RSD materials closely meet
the standards of the International Commission on Radiation
Units and Measurement (ICRU) Report No. 44 (Tissue Substitutes
in Radiation Dosimetry and Measurement, 1989) for both the
cortical and spongiosa components of the human skeleton. The
mass densities are 1.88 g/cc for cortical bone and 1.16 g/cc
for spongiosa. The narrow beam linear attenuation coefficient
for the cortical component, measured at 140 keV, is 0.280
The volume of the thoracic cavity, when all organs (heart,
lungs, and liver) are inserted, is about 8,200 ml. It is
filled from the top, with either an inert or a radioactive
solution, to simulate the thoracic background.
A valve is installed at the base for conveniently draining the
phantom. The residue on the walls of the cavity can be easily
flushed with the fittings provided at the top of the phantom.
A second, smaller fitting is also provided as an air-bleed
An accurately anatomic heart model is based on vacuum-formed
shells. It was designed using high resolution,
contrast-enhanced, ultrafast CT data from a normal patient,
slightly modified to facilitate its use.
left and right chambers are connected at the atrium region to
make a single compartment which can be filled and flushed
independently using two ports labeled HC (heart chambers). The
right ventricle is slightly modified to allow air to escape
during filling. The myocardial wall (MW) has two ports,
flushing and independent filling. The volume of the heart
chambers is 284 ml, while the volume of the myocardial wall is
238 ml, without inserted defects.
The standard model includes three defects with volumes of 8.9,
13.5, and 41.7 ml, respectively. Each of the defects can be
Defects of different dimensions can be ordered at no added
cost. A disassembled heart is sent on request, so that
dimensions of a special set can be established. Note that
different defects cannot be retrofitted in the assembled
Perfusable lungs are molded in hollow, vacuum-formed shells,
filled with styofoam beads. The final mass density of 0.40
g/cc is attained by adding an inert or radioactive solution
through a filling port at the top of each lung shell. Extra
sets of lungs can also be furnished for work continuity. The
volumes of the left and right lung shells are 907 ml and 1,134
A liver with a volume of 980 ml is included to evaluate the
effect of its uptake on quantitative myocardial imaging. It is
a vacuum-formed shell, mounted on acrylic tubes to minimize
artifacts. The liver can be filled with an inert or
A set of fillable capsules is provided to serve as external
markers. Capsules can be filled with a radioactive solution
and attached to the external surface of the phantom. The
phantom can then be imaged, using SPECT or PET modalities to
compare image-registration techniques.
The thoracic phantom without the overlay simulates an average
male patient. The overlay, with or without breasts, simulates
a large female or a still larger male patient, respectively.
It is then possible to evaluate the effect of additional
attenuation and scatter on quantitative myocardial imaging.
The volume of each vacuum-formed breast is 972 ml. A tumor,
filled with a radioactive solution can be inserted to evaluate
the planar and tomographic imaging techniques used for
mammoscintigraphy-(2). A set of breast tumors (3, 6, 9, 12 and
15 mm diameters) is included. They are supported by thin,
threaded nylon rods which pass through ports and are sealed by
O-rings. They can be bent by hand to reach any part of a
The thorax is mounted on a base plate with an O-ring seal.
Four rubber feet provide space under the phantom for drain
fittings the base plate is removed readily to provide access
to the interior of the phantom.
A knob at the top of the neck secures a rod which passes
through to the heart/lung subassembly for disassembly.
**Applicable to both Heart/Thorax(h) and Striatal(s)
Phantoms unless otherwise indicated.
Nuclear Medicine orders are handled by Capintec Inc.
For International contact Baris Kalyoncu:
7 Vreeland Road
Florham Park, NJ 07932
For North American contact Kristian Koschal:
7 Vreeland Road
Florham Park, NJ 07932
back to top