Development of a high-resolution and high efficiency single photon detector for cardiovascular
diseases study in mice. SPECT assessment of left ventricular perfusion using different routes of
delivery of 99mTc-MIBI
E. Cisbani, F. Cusanno, S. Colilli, R. Fratoni, F. Garibaldi, F, Giuliani, M. Gricia, R. Fratoni, M.
Lucentini, M. L. Magliozzi, F. Santanvenere
Dipartimento TESA, Istituto Superiore di Sanita’ and INFN - gr. Sanita’ – Rome, Italy
G. Marano, M. Musumeci
Dipartimento del Farmaco, Istituto Superiore di Sanita’ - Roma
M. Baiocchi, L. Vitelli
Istituto Superiore di Sanita’ Dipartimento di Oncologia - Roma
G. De Vincentis
Universita’ La Sapienza, Roma
S. Majewski
University of West Virginia – Morgantown
B. Tsui, Y. Wang
Johns Hopkins University, Baltimore, MD, USA
Abstract–SPECT systems using pinhole
collimator apertures permit radiolabeled
molecular distributions to be in principle
imaged with high resolution in vivo in small
animals. However, studying cardiovascular
diseases by means of small animal models is
very challenging. Submillimeter spatial
resolution and high sensitivity, plus good
energy resolution are required in this case at
the same time. We designed what we consider
the “optimal” radionuclide detector system
for this task. It should allow studying both
detection of unstable atherosclerotic plaques
and monitoring the effect of therapies. Using
mice is particularly challenging in situations
that require several intravenous injections of
radiotracers, possibly for weeks or even
months, in chronically ill animals, during
longitudinal studies. Thus, alternative routes
of delivering the radiotracer in a tail vein are
under investigation. In this study we have
performed preliminary measurements of
detection of atherosclerotic plaques in
genetically modified mice with the highresolution prototype detector. We have also
evaluated the feasibility of assessing left
ventricular perfusion by intraperitoneal
delivery of 99mTc-MIBI in healthy mice.
1. Introduction
Cardiovascular disease (CVD) is the leading
cause of disability and mortality in the
developed countries. Atherosclerosis is a
systemic disease that develops slowly and
often asymptomatically, so that for many
patients its first manifestation is sudden
cardiac death, stroke, or myocardial
infarction. The clinical challenge is not just in
identifying the patient with atheroma but in
recognizing specific lesions likely to cause
clinical events, that means “vulnerable”
plaque. Also monitoring novel treatment
strategies, for example delivery of different
variants of stem cells. For these reasons, the
assessment of myocardial perfusion plays an
important role in the diagnostic work-up of
patients as well as in the assessment of
prognosis and guiding the therapy [1-3].
Studies with mice are very important due to
the similarities of the disease onset and
progression with human coronary artery
diseases. Mice models both genetically
modified and artificially induced are
available. From the imaging side, molecular
imaging by radionuclides is the most reliable
non invasive technique for myocardial
perfusion studies. SPECT is the technique of
choice here over PET. In fact SPECT
techniques have a special role in small
animal imaging research [6]; they have
limited sensitivity due to the collimation but
1
the PET has intrinsic limitations in terms of
spatial resolution [4]. A large spectrum of
SPECT radiotracers is accessible, and,
provided the detector has good energy
resolution, multi-labelling allows studying
different phenomena simultaneously. For
example, dual tracer small animal SPECT
would allow simultaneous imaging of 99mTclabeled stem cells to assess myocardial
perfusion and of 111In labelled stem cells to
delineate stem cells engraftment [3,5]. It has
been shown [14,17] that after careful
calibration, using standard nuclear medicine
software, perfusion ECG gated SPECT in
mice permits quantification of LV volumes
and motion. This would allow evaluating the
effects of therapy in the limit of the
sensitivity attained by the system. In fact the
magnitude of 99mTc-MIBI uptake predicts the
response of myocardium with abnormal
function to subsequent revascularization in
the chronic coronary artery disease, and the
recovery of myocardium after reperfusion
therapy for acute MI.
Studying cardiovascular diseases by means of
small animal models is very challenging
because submillimeter spatial resolution,
good energy resolution and high sensitivity
are required. The goal of the experiment
dictates the spatial resolution and the
sensitivity required for the imaging system.
Many devices have been proposed or
developed, each with different performance
characteristics [20]. Most of them are based
on the standard Anger camera concept with
pinholes and multipinholes [20]. Due to the
high efficiency and to the magnification
factor that can be employed, good
performances have been showed, especially
using multipinhole techniques. Nevertheless,
the Anger camera based systems have
limitations [21] for several reasons. The ideal
system should have an “open” and flexible
design, to be integrated in a multimodality
system with other detectors (MRI, CT,
optical). This is not possible with the standard
Anger camera based systems.
Our group started research studies on
detection of vulnerable atherosclerotic
plaques and of stem cell therapy on mice.
From the review of prior art, many clinical
trials have been performed recently on this
subject but the results are contradictory.
Therefore, indeed more basic studies with
small animals have to be performed [2]. Also,
we found that repeated injections of
radiotracers in mice modelled for infarction,
possibly for weeks or even months, when
needed, is not possible using the usual route
of delivery (tail vein). Alternative delivery
routes have to be developed.
This paper describes the research started by
our collaboration in outlining the best
detectors suited for these studies, the
preliminary measurements, with a high
resolution detector prototype, in detection of
atherosclerotic plaques on genetically
modified mice, and perfusion measurements
comparing the uptake of 99mTc-MIBI with the
two injection options: via tail vein and via
peritoneum.
2. Materials and Methods
We designed what we consider the “optimal”
radionuclide detector system for this task,
flexible enough to be integrated in a
multimodality system:
8 detectors to
optimize the trade-off between spatial
resolution and sensitivity. One of these
special modules is a detector with spatial
resolution in the range of 300-500 m,
sensitivity of 0.3 cps/kBq, and active area
100 x 100 mm2, using tungsten pinhole
collimator(s) and a high granularity pixellated
scintillator (CsI(Na) with 0.8 mm pitch - the
smallest so far achieved for SPECT detectors)
or a continuous LaBr3. Details on the basic
studies on detector prototypes can be found in
[7-12] (Stan, you should add some of your
references)
The performances of such a detector system
compared to what can be obtained with Anger
camera based systems are shown in Fig.1
2
Fig. 2. Flood raw image (57Co) and pixel
identification of the CsI(Na) 0.8 mm pitch
array coupled to a Hamamatsu H9500 flat
panel PMT (see text).
Fig. 1. Spatial resolution and efficiency for
high resolution based system and Anger
camera based system
2.1 Detector layout
The arguments for the selection of the
particular layout come from the fact that in
the SPECT with multipinholes, with 3D
reconstruction, a sufficient number of
“resolution elements” has to be used [13]
(Ben should check if this is the right
reference). This translates in the requirement
of approximately 120 pixels in a 100 mm
dimension of the detector, that means an
intrinsic spatial resolution of ri = 0.8 mm.
Scintillator arrays composed of very small
pixels have to be used and identifying these
small pixels is challenging. It requires the use
of multichannel readout to fully exploit the
detector characteristics. We have designed
and built multichannel electronics capable of
reading out up to ~ 4096 channels
individually at 20 KHz [9, 16].
Preliminary test measurements with small
samples (see Fig.2) show that very good pixel
identification is obtained coupling such a
scintillator array to a Hamamatsu H9500 flat
panel PMT (with 3 x 3 mm2 anode pixels).
Nevertheless this layout could be not
optimized in the dead areas between the
PSPMT’s [9,12]. Moreover, good energy
resolution across the detector field of view
would be necessary to be able to use the
multilabeling technique. For this reason
careful comparison has to be done between
CsI(Tl) 0.8 mm pitch and LaBr3(Ce) that,
provided thin enough plates (4 mm) (or small
pixels) can be obtained. This solution would
allow to minimize the dead area problem and
would provide uniformly very good energy
resolution.
2.1 Detector prototypes
Two prototypes with performances close to
the needs in terms of spatial resolution were
available, a NaI(Tl) 1.5 mm pitch 100 x 100
mm2, coupled to PSPMT Hamamatsu H8500
(6 x 6 mm2 anode pixel) and CsI(Tl) 1 mm
pitch coupled to PSPMT Hamamatsu H9500
(3 x 3 mm2 anode pixel). These prototypes
allowed performing the measurements
described in this paper. We used pinhole
collimator that according to the magnification
allows obtaining a FOV of ~ 30 x 30 mm2
(M=3) sufficient for imaging the fraction of
the mouse body relevant for studying the
stem cell trafficking, or a FOV of 25 x 25
mm2, sufficient for heart perfusion imaging.
The spatial resolution and sensitivity depend
strongly on the pinhole dimension. The
selected pinhole diameter was 0.5 mm.
This allowed us to study the basic problems
of the detection system and animal handling.
2.2 Micro SPECT system prototype
The prototype SPECT system is equipped
with a 2.5-cm-diameter acrylic cylindrical
3
bed-holder (3 mm thick) that keeps the mouse
horizontal (see Fig. 3).
Fig. 3 The SPECT system prototype.
Two detectors are mounted on a motorized
gantry that can rotate around the animal bed
(only one detector was used for the
measurement described here). The bed holder
stays in a fixed position. The system could be
manually adjusted to optimize the distance
between the pinhole and the axis of rotation,
giving the possibility to resize the camera
parameters depending on measurement
requirements. The detector characteristics and
performance parameters are listed in Table 1.
Tab.1
Pinhole Diameter (mm)
NaI (Tl) Scintillator:
- pitch (mm)
- Thickness (mm)
0.5
1.5
6
- Dimension (mm)
100 × 100
Photomultiplier Array
(2 × 2) H8500
Resolution (mm)
Efficiency (cps/MBq)
Magnification Factor
Field of View (mm)
< 0.8
35
3
33
2.4 Animal procedures, Anaesthesia,
and Tracer administration
Two three-month-old adult VFB/N male
mice, weighing 30 g, were intraperitoneally
anesthetized. For one of the mice, the
radiotracer (MIBI-99mTc) was injected into the
tail vein. Care was taken to minimize, as
much as possible, the volume of injected
tracers around 0.02-0.05 ml to avoid
significant changes in the whole blood
volume of the mice. The single pinhole
projection data were acquired in 60 angular
intervals over 360 degrees. Thoracic bone
scan was performed to evaluate system’s
image quality (a mouse was injected with 2
99m
mCi
of
Tc-MDP).
Tomographic
acquisitions started 2 hours after tracer
administration. Projection data were acquired
at 2 min/projection. Myocardial perfusion
scan was performed. Live mouse was injected
with 6.7 mCi of 99mTc-MIBI; acquisitions
started 1 hour after tracer administration to
ensure a better contrast of heart to soft
tissues. Projection data were acquired at 60
sec/projection. The same procedure was used
for the second mouse but it was injected with
6.7 mCi of 99mTc-MIBI intraperitoneally. To
assure high-resolution and artefacts free
SPECT image reconstruction, mechanical
calibration of the imager was needed. For this
reason, tomographic acquisition of a set of 2
point sources (~ 1 mm in size) positioned as
far as possible both along the axis of rotation
and away of it, was also performed.
2.5 Image reconstruction technique
2.3 Phantom studies
To test the tomographic spatial resolution
capabilities of our imager, a miniature acrylic
resolution phantom was manufactured, as
shown in Fig. 4. It consists of 6 sectors, each
containing equally sized sets of small
diameter holes (0.8, 0.9, 1.0, 1.1, 1.2, 1.3
mm). The overall phantom diameter was 25
mm. The total activity in all filled capillary
holes was ~ 4.5 mCi of 99mTc. The single
pinhole projection data were acquired in 60
angular intervals over 360 degrees at 2
min/projection.
Upper
head
projection
data
were
reconstructed using a 3D pinhole OS-EM
algorithm. Which takes into account
geometric misalignment parameters of the
system, including the centre-of-rotation error,
the tilt angles between the axis-of-rotation
and the detector plane in 3D space. Size of
the reconstruction matrix was 90°×°90°×°90
with a voxel size of 0.25 mm. A 3D
Butterworth filter was used for the postreconstruction.
4
FOV at a distance of 10 mm. The energy
resolution was 14%@122 keV.
2.6 Myocardial perfusion analysis
There is no true standard for quantification of
SPECT[15]. We used the Standardized uptake
value, SUV, also referred to the dose uptake
ratio, DUR, calculated as a ratio of tissue
radioactivity concentration (in units kBq/ml)
at time T, CPET(T) and injected dose (in
units MBq) at the time of injection divided by
body weight (in kg units).
SUV = CPET(T)/(Injected dose/animal's weight).
If radiotracer is uniformly distributed, and
taking into account the delay time, we
calculated it as Regional Uptake Value
(RUV) for the region of interest (heart).
2.7 Detection of atherosclerotic plaques
Another detector prototype using CsI(Tl)
scintilator array, 1.0 mm pitch and the same
setup and the same procedure has been used
for detecting atherosclerotic plaques.
3.2 Perfusion images
Three-month-old 30g VFB/N male mice,
were intraperitoneally anesthetized. Care was
taken to minimize the volume of injected
tracers into the tail vein (0.02-0.05 ml) to
avoid significant changes in the whole blood
volume of the mice. Fig. 5 shows a perfusion
image obtained with the detector prototype
described (NaI(Tl) 1.5 mm pitch). Midventricular short-axis slice (left) and
horizontal long-axis slice (right) obtained
from reconstructed projection of myocardial
perfusion images are shown. Left and right
ventricular cavities and corresponding walls
can be easily identified.
5 mm
5 mm
LV
LV
RV
3. Results
3.1 Phantom studies, Spatial resolution,
Sensitivity
For the resolution phantom as well as for
myocardial perfusion we used a FOV of
33 mm. The spatial resolution of the
system is then ~0.8 mm.
1.3 mm
0.8 mm
1.2 mm
0.9 mm
R
V
Fig. 5. Mutual perpendicular cross-section
through 3-dimensional myocardial perfusion
image volume of a live mouse. Short-axis
slice (left) shows myocardial perfusion in
right ventricular (RV) and left ventricular
(LV) walls. Horizontal long-axis slice is
shown at right.
The need for a different route of delivery
brought us to a new scan with comparison of
two different modalities, injection through the
tail vein and injection through the
peritoneum. The same procedure was
adopted. In Fig. 5 we show images of the
mouse injected trough the tail vein;
transversal, sagittal and coronal views are
shown.
1.1 mm
1.0 mm
Fig. 4 Miniature acrylic resolution phantom (left),
and reconstructed image (right), sum of 21 trans-
axial slices. 0.8 mm capillaries are clearly
separated in the image.
The sensitivity of the system was ~ 35
cps/MBq. It was measured by using a 370
kBq source of 57Co placed in the centre of the
Fig. 6. Transversal, sagittal and coronal heart
views. Tail vein injection.
5
The second mouse had the radiotracer
injected peritoneally. All other procedures
were the same. Fig.7 shows the obtained
perfusion images.
Fig. 8a. APOE(+/-) mouse 6 weeks old
Fig. 7. The same as Fig. 3 except that for the
mouse injected peritoneally.
3.3 Uptake
Tab.2 shows the results of the calculated
uptake for the two delivery modalities.
Tab.2
Fig.8b 25 weeks
Control
Fig.8c APOE(+/-)
25 weeks old
4. Conclusions and Outlook
SUV
Peritoneum
Tail vein
Transversal
0.47
1.09
Coronal
0.39
1.22
Sagittal
0.41
1.31
A reduction of uptake occurred but the
ventricular cavities are identified in both
cases.
3.4 Detecting atherosclerotic plaques
Mice of different ages were scanned.
Preliminary results are shown in Fig. 8.
Young mouse (Fig. 8a) did not show plaque
uptake, at the limit of the sensitivity of the
detector. Uptake of Annexin V by liver is
seen. Fog. 8b shows the result for the control
mouse, 25 weeks old. It doesn’t show any
suspicious spots. Fig. 8c shows the result for
the APOE (+/-) mouse 25 weeks old.
Suspicious spots seem to be seen in this
image. No definitive conclusions can be
extracted from this preliminary analysis.
Hystological findings confirmed eventually
the detection of atherosclerotic plaques.
A single head high-resolution detector
prototype has been built for studying
optimized SPECT system for researches on
the detection of atherosclerotic vulnerable
plaques and on the diffusion of stem cells,
their fate and the effect of the therapy.
The scope was to determine the detector
characteristics and to study animal handling
issues.
The spatial resolution of the prototype
showed to be sufficient for perfusion studies.
The energy resolution allows using dual
tracer techniques.
The sensitivity would have been significantly
higher using a larger dimension of the pinhole
(up to 1.5 mm [14]).
We demonstrated that injecting the
radiotracer through the peritoneum instead of
the tail vein allows obtaining a good
perfusion.
However in this case the price to be paid is a
reduction of the uptake by the heart muscle.
The sensitivity of the system has to be
increased to compensate this reduction. This
can be obtained by fine-tuning the parameters
(the
pinhole
dimension
and
the
magnification), using the multipinhole
technique and adding as many detector heads
as practically possible to the system.
It has to be underscored that to fully
accomplish the objectives of the study will
probably require the integration of other
modalities [3], essentially optical and MRI
with significant modifications of the layout,
6
and of the materials and components, starting
with substitution of PSPMT’s with Silicon
Photomultipliers (SiPMs) insensitive to the
magnetic fields. Research in this direction is
ongoing.
References
[13] B.M.W. Tsui and Y. Wang, J. Nucl.
Cardiol. 12 (2005), p. 262
[14] A. Costantinesco et al. Assessment of
ventricular Perfusion, Volumes and Motion in
Mice Using Pinhole Gated SPECT, JNM
2005; 46:1005-1011
[15] Seo et al. JNM, Vol. 45, N. 9, September
2004
[16] A.G. Argentieri et al., "A Novel Modular
and Flexible Readout Electronics for Photon
imaging Applications", in proceedings of the
IEEE NSS-MIC 2008 conference, 2008
[17] B.B.Chin et al. Left ventricular
functional assessment in Mice: feasibility of
high spatial and temporal resolution ECGgates blood poll SPECT. Radiology, Vol 245.
N. 2 November 2007 , pg. 440
[18] F. Garibaldi et al., High resolution
detectors for molecular imaging of
cardiovascular diseases (in preparation)
[19] F. Beekman et al. Optimizing
multipinhole SPECT geometries using an
analytical model, P.B.M 21 (207) 2561
[20] M.C.Wu et al. Pinhole single photon
Emission Computed Tomography for
Myocardial Perfusion Imaging of Mice,
JACC vol. 42, N.3, 2003
[21] MT. Masden et al, Recent advances in
SPECT imaging, J Nucl Med 2007; 48:661–
673
[22] Metzler SD, Jaszczak RJ, Patil NH,
Vemulapalli S, Akabani G, Chin BB.
Molecular imaging of small animals with a
triple-head SPECT system using pinhole
collimation. IEEE Trans Med Imaging.
2005;24:853–862.
[1] Orlic, D., et al., Bone marrow cells
regenerate infarcted myocardium. Nature,
2001. 410(6829): p. 701-5.
Zhang, S., et al., Purified human bone
marrow multipotent mesenchymal stem cells
regenerate
infarcite
myocardium
in
experimental rats. Cell Transplant, 2005.
14(10): p. 787-98.
Assmus, B., et al., Transcoronary
transplantation of progenitor cells after
myocardial infarction. N.Engl J Med, 2006.
355(12): p. 1222-32.
Wollert KC, Meyer GP, Lotz J, et al.
Intracoronary autologous bone-marrow cell
transfer after myocardial infarction: the
BOOST randomised controlled clinical trial.
Lancet 2004;364: 141–8
[2] Segers Vincent F.M. & Richard T. Lee,
Stem cell therapy for cardiac diseases,
Nature, Vol. 451, 21 February 2008,
doi:10.1038/nature06800
[3] Bengel, F.M., V. Schachinger, and S.
Dimmeler, Cell-based therapies and imaging
in cardiology, Eur J Nucl Med Mol Imaging,
2005. 32 Suppl 2: p. S404-16.
[4]T.Lewellen,Phys.Med.Biol.53(2008)R287
–R317
[5] Acton, P.D. and R. Zhou, Imaging
reporter genes for cell tracking with PET and
SPECT. Q J Nucl Med Mol Imaging, 2005.
49(4): p. 349-60.
[6] R. Meikle et al. Small animal SPECT and
its place in the matrix of molecular Imging
Technologies, Phys. Med. Biol. 50 (2005)
R45-R61)
[7] F. Garibaldi et al. Small animal imaging
by single photon emission using pinhole and
coded apertures collimation, IEEE TNS
52(3)2005,. 573-579
[8] F. Garibaldi et al. NIM A 569, 2006, 286 – 290
[9] E. Cisbani et al. NIM A 571, 2007, 169-172
[10] F. Cusanno et al. NIM A 569, 2006, 193-196, and
references quoted therein
[11] F. Garibaldi et al. Nucl. Instr. Meth A 471, 2001,
222-228
[12] M.L. Magliozzi et al. Proceedings of 9th ICATPP
Conference, Como 2005.
7