Journal of Radiology
in press
Simultaneous PET/MR Images, acquired with a
Compact MRI Compatible PET Detector in a 7
Tesla Magnet
Martin S. Judenhofer BS1, Ciprian Catana2, Brian, K. Swann3, Stefan Siegel
PhD3, Wulf-Ingo Jung PhD4, Robert Nutt3, Simon R. Cherry PhD2, Claus D.
Claussen MD1 and Bernd J. Pichler PhD1
1
Laboratory for Preclinical Imaging and Imaging Technology, Clinic of
Radiology, University of Tübingen, Germany
2
Department of Biomedical Engineering, University of California, Davis, CA,
USA
3
Siemens Preclinical Solutions, Knoxville, TN, USA
4
Bruker BioSpin MRI, Ettlingen, Germany
1/27
Abstract:
The purpose of our study was to prospectively use compact APDs instead of
photomultiplier tubes to integrate a PET detector system within a 7 Tesla MR scanner.
. All animal experiments were performed according to the guidelines of the University of
Tübingen and the German law for the protection of animals. A very compact LSO-APD
PET detector was built and optimized so as to operate within a 7 Tesla MR scanner. The
detector performance was investigated outside and inside the magnet and MR image
quality was evaluated with and without the PET detector. Two PET detectors set up in
coincidence were used to acquire PET images of a [ 18F]-FDG injected mouse head
specimen in step and shoot mode. The performance of the PET detector when operated
inside the magnet during MR acquisition showed very little degradation in energy
resolution (increase from 14.6% to 15.9%). The MR imaging was not influenced by the
PET detector. The fused PET and MR images showed anatomy match and no
degradation of image quality. Thus, simultaneous PET and MR imaging in a 7 Tesla
system is feasible.
2/27
INTRODUCTION
Current detector research focuses on multimodality in vivo imaging to combine functional
and morphological information for clinical diagnosis and preclinical research with
laboratory animals. Combining positron emission tomography (PET) with X-ray computed
tomography (CT) (1) has already shown great value, especially in tumor diagnosis and
tumor staging (1-4). However, in contrast to CT, magnetic resonance (MR) imaging does
not give additional radiation and can, furthermore, produce images with high soft tissue
contrast even without the use of contrast agents. Dose from ionizing radiation and the
application of mass-levels of contrast agents can alter the biological process of interest
and should, therefore, be minimized. If standalone PET and MRI systems are used and
image data are fused manually, movement from one imaging device to another, or very
long scan times often make co-registration, especially of small regions such as lymph
nodes, impossible. Thus, current developments are fostering the combination of PET and
MRI for simultaneous data acquisition.
Attempts to combine PET and MRI have a history going back approximately 10 years (59). A limiting feature of early combined PET/MRI designs are the bulky photomultiplier
based PET detectors which are very sensitive to magnetic fields (10). Originally, the focus
was on an optical fiber based system, that channeled the scintillation light produced in the
PET detectors to photomultiplier tubes positioned outside the magnet and in the fringe
magnetic field (5, 7-9). The drawback of this concept is that the PET signal quality suffers
from light loss caused by the light transmission via optical fibers over several meters,
resulting in a reduced timing resolution, energy resolution, and crystal decoding accuracy.
However, current approaches combining PET and MRI are based on avalanche
photodiode (APD) technology (11-14) or complex MRI modifications such as field cycled
MRI, where the PET detector acquires data when the magnetic field is turned off (15) or
split magnets (16). APDs are light detectors which have proven to be successfully used in
3/27
PET technology (11, 17). Cherry et al used APDs combined with short fibers for realizing
a small animal PET/MR system, ensuring a MR field of view (FOV) without any metal or
electronic parts, by placing the APDs offset in the axial direction (18). Since only short
fibers are needed, the PET detector performance is only slightly degraded.
Our approach is to develop and optimize a compact full PET detector ring which can be
operated in the MR scanner making optical fiber coupling redundant. Both, the PET
scanner and the MR scanner should operate at their full performance potential without
influencing each other. In addition, the system should be designed for simultaneous
PET/MR imaging.
Initial measurements with PET detector technology based on lutetium oxyorthosilicate
(LSO) crystal blocks coupled to 3 x 3 APD arrays showed the feasibility of using an APD
based PET detector inside a high field MR system (18, 19). However, preliminary results
(19) showed substantial interference between the two imaging systems and degraded
system performance. Thus, the purpose of our study was to prospectively use compact
APDs instead of photomultiplier tubes to integrate a PET detector system within a 7 Tesla
MR scanner.
Abstract:
The purpose of our study was to prospectively use compact APDs instead of
photomultiplier tubes to integrate a PET detector system within a 7 Tesla MR scanner.
. All animal experiments were performed according to the guidelines of the University of
Tübingen and the German law for the protection of animals. A very compact LSO-APD
PET detector was built and optimized so as to operate within a 7 Tesla MR scanner. The
detector performance was investigated outside and inside the magnet and MR image
quality was evaluated with and without the PET detector. Two PET detectors set up in
4/27
coincidence were used to acquire PET images of a [ 18F]-FDG injected mouse head
specimen in step and shoot mode. The performance of the PET detector when operated
inside the magnet during MR acquisition showed very little degradation in energy
resolution (increase from 14.6% to 15.9%). The MR imaging was not influenced by the
PET detector. The fused PET and MR images showed anatomy match and no
degradation of image quality. Thus, simultaneous PET and MR imaging in a 7 Tesla
system is feasible.
5/27
INTRODUCTION
Current detector research focuses on multimodality in vivo imaging to combine functional
and morphological information for clinical diagnosis and preclinical research with
laboratory animals. Combining positron emission tomography (PET) with X-ray computed
tomography (CT) (1) has already shown great value, especially in tumor diagnosis and
tumor staging (1-4). However, in contrast to CT, magnetic resonance (MR) imaging does
not give additional radiation and can, furthermore, produce images with high soft tissue
contrast even without the use of contrast agents. Dose from ionizing radiation and the
application of mass-levels of contrast agents can alter the biological process of interest
and should, therefore, be minimized. If standalone PET and MRI systems are used and
image data are fused manually, movement from one imaging device to another, or very
long scan times often make co-registration, especially of small regions such as lymph
nodes, impossible. Thus, current developments are fostering the combination of PET and
MRI for simultaneous data acquisition.
Attempts to combine PET and MRI have a history going back approximately 10 years (59). A limiting feature of early combined PET/MRI designs are the bulky photomultiplier
based PET detectors which are very sensitive to magnetic fields (10). Originally, the focus
was on an optical fiber based system, that channeled the scintillation light produced in the
PET detectors to photomultiplier tubes positioned outside the magnet and in the fringe
magnetic field (5, 7-9). The drawback of this concept is that the PET signal quality suffers
from light loss caused by the light transmission via optical fibers over several meters,
resulting in a reduced timing resolution, energy resolution, and crystal decoding accuracy.
However, current approaches combining PET and MRI are based on avalanche
photodiode (APD) technology (11-14) or complex MRI modifications such as field cycled
MRI, where the PET detector acquires data when the magnetic field is turned off (15) or
split magnets (16). APDs are light detectors which have proven to be successfully used in
6/27
PET technology (11, 17). Cherry et al used APDs combined with short fibers for realizing
a small animal PET/MR system, ensuring a MR field of view (FOV) without any metal or
electronic parts, by placing the APDs offset in the axial direction (18). Since only short
fibers are needed, the PET detector performance is only slightly degraded.
Our approach is to develop and optimize a compact full PET detector ring which can be
operated in the MR scanner making optical fiber coupling redundant. Both, the PET
scanner and the MR scanner should operate at their full performance potential without
influencing each other. In addition, the system should be designed for simultaneous
PET/MR imaging.
Initial measurements with PET detector technology based on lutetium oxyorthosilicate
(LSO) crystal blocks coupled to 3 x 3 APD arrays showed the feasibility of using an APD
based PET detector inside a high field MR system (18, 19). However, preliminary results
(19) showed substantial interference between the two imaging systems and degraded
system performance. Thus, the purpose of our study was to prospectively use compact
APDs instead of photomultiplier tubes to integrate a PET detector system within a 7 Tesla
MR scanner.
7/27
MATERIALS AND METHODS
Design of the MR Compatible PET Scanner
The PET system is designed to be used within a 7 Tesla BioSpec 70/30 USR or a 7 Tesla
ClinScan (both Bruker BioSpin, Germany and Bruker BioSpin MR, Germany) MR system.
For these studies we used the BioSpec 70/30 USR scanner which operates at 300 MHz
resonance frequency. For all MR data acquisition and analysis, the ParaVision software
platform (Bruker BioSpin, Germany) was used. The full ring PET scanner is currently
under construction and will consist of 10 block detectors arranged in a ring of 60 mm
inner diameter. The crystal blocks have a size of 19 x 19 x 4.5 mm³ and form an axial
field of view (FOV) of 19 mm and a transaxial FOV of approx. 45 mm. This field of view
will be sufficient for imaging the brain, heart or abdominal region of a mouse. The outer
diameter of the PET ring is only 120 mm and will fit into the mini gradient set (B-GA 12,
Bruker BioSpin, Germany) of the MR scanner. The micro imaging radio frequency (RF)coil (Bruker BioSpin, Germany) with an outer diameter of 60 mm and an inner diameter of
36 mm will go inside the PET scanner (Fig 1).
PET Detector Design
Each PET detector consists of a 19 x 19 mm² crystal block (Siemens Preclinical
Solutions, USA) comprising 12 x 12 individual 1.5 x 1.5 x 4.5 mm³ crystals separated with
a highly reflective foil (17). The crystal block is coupled via a 3 mm thick light guide to a
monolithic 3 x 3 APD array (Hamamatsu, Japan) (Fig 2) where the individual APDs have
an active surface of 5 x 5 mm² (17). The APDs are operated on a negative bias voltage
and reach break-down at approx. -405 V. In comparison to the previous work (19) carried
out with large hybrid amplifiers, the 9 APD signals are now fed into a highly integrated 9channel charge sensitive preamplifier (Siemens Preclinical Solutions, USA) (17). The
output signals of the preamplifier are buffered and led outside the magnet by 6 m fully
8/27
shielded non magnetic coaxial cables (Leoni, Germany). All electronic parts are selected
to be non magnetic and mounted on a custom made 6-layer printed circuit board (PCB)
which has a flexible connection between the APD and the preamplifier in order to allow a
height adjustment of the LSO-APD detector in the radial position of the gantry (Fig 2).
The PCB was optimized for operation in a high magnetic field to keep eddy currents in
the power and ground planes low. While the LSO-APD detector, preamplifier, and buffers
(Fig 2) are residing inside the magnet, the 9 analog signals are processed, outside the 5
Gauss line, with a custom made 9 to 4 analog multiplexer, providing event position and
energy information from the 12 x 12 crystal block. For the test setup used in this work, the
analog signals were post processed with standard nuclear instrument modules (NIM) and
analog-to-digital conversion achieved by using a PD2-MFS-8 2M/14 data acquisition
(DAQ) board (United Electronic Industries Inc., USA) mounted in a standard PC (20). For
the final detector ring, dedicated PET electronics (Siemens Preclinical Solutions) will be
used to digitize the signals and perform the coincidence processing of all 10 detectors.
Based on problems seen in preliminary studies (19), we used double-sided printed circuit
board (PCB) material coated with 10 µm copper for electromagnetic shielding to protect
the PET front-end electronics from distortions induced by MR sequences. The copper
layer of this material was only 10 µm thin to avoid artifacts in MR images by eddy
currents induced in the material. The copper material was jet-cut into pieces and soldered
together at the edges to form a solid box (Fig 2).
Performance Test of the LSO-APD PET Detector outside the Magnetic Field
As described, several important optimizations of the front-end were performed based on
the experience reported in previous work (19) to minimize interference between the PET
and MR system. To ensure proper performance of the new front-end board layout, all 9
channels of the preamplifier have been tested by simulating a 100 pF APD capacitance
9/27
on the amplifier input. The results were compared to the performance parameters of the
amplifier published earlier (17). The measurements and the evaluation of data were
performed by three authors working together (M.S.J., B.K.S., and S.S.)
A complete assembled and shielded PET detector was tested outside the magnetic field.
The crystal block and light guide were glued with UV-curable optical glue (OP 20,
DYMAX, USA), whereas the light guide was coupled to the APD array with optical grease
(Bicron BC 630, Saint-Gobain Ceramics & Plastics, USA) allowing the use of the same
light guide and LSO block for all the boards tested. Position profiles showing the
individual crystals of the LSO -APD block detector were acquired by exposing the crystal
block to a 100 kBq
68Ge
point source. The APD array was biased at -380 V. All
measurements described here were carried out at room temperature without additional
cooling of the PET detector. For acquisition, the 4 signals from the multiplexer were
shaped with a semi-Gaussian filter (300 ns shaping time) and subsequently fed to a PD2MFS-8 2M/14 (20) data acquisition board for digitization and signal processing. To
generate an analog to digital conversion timing signal for the data acquisition board at the
peak of the analog pulse, the 4 signals from the multiplexer were split before shaping and
one signal path was processed with a custom made fast summing circuit and then fed to
a constant fraction discriminator (CFD) (CFD 103, PSI, Switzerland) and gate and delay
generator (DT 102, PSI, Switzerland) in order to generate an accurate timing signal.
Position profiles and mean peak to valley ratio for 3 centre crystals and 4 edge crystals (2
on each side) were calculated. Single crystal energy spectra of the PET detector and
energy resolution of centre and corner crystals were calculated using custom software
(20). The Quantitative values are reported as mean ± one standard deviation. Data were
acquired and evaluated by one author (M.S.J.)
10/27
Performance of the LSO-APD PET Detector inside the 7 Tesla MR Scanner
Comparable tests with an individual PET detector as described above were performed
inside the 7 Tesla scanner and while applying pulse sequences so as to evaluate whether
the PET detector (electronics and shielding) and the MR (gradients and the RF signals)
interfere with each other and whether there is a loss of PET or MR imaging quality. The
measurements outside the magnet and the evaluation of data were performed by one
author (M.S.J.)
For the MR measurements, a RF-coil with 72 mm inner diameter was used together with
the 200 mm gradient set (B-GA 20, Bruker BioSpin, Germany). A poly methyl
methacrylate (PMMA) cylinder (outer diameter of 28 mm and 100 mm length) filled with
silicone oil (M10, Roth, Germany) was used as a homogeneous phantom and placed
inside the RF-coil. A standard FLASH imaging sequence (TR = 400 ms, TE = 6 ms, Flip
angle = 30°, 256 x 256 pixels) was used to acquire axial images of the phantom. The PET
detector was positioned with its centre FOV aligned with the magnet isocentre and
radially at the outer edge of the RF-coil in the same way the PET detectors will be
arranged in the final setup (Fig 1). MR images were acquired with and without the PET
detector inside the magnetic field. As a measure of image quality, the ratio of the signal of
the phantom (S) to the signal of the background (BG) outside the phantom and the ratio
of the signal (S) to noise (N) were determined. S was measured as mean of five
concentric 20 mm regions placed on five adjacent axial image slices. To calculate the
signal to background (S/BG) and signal to noise (S/N), 4 regions with a diameter of
10 mm each were placed at the corners outside the phantom in the same axial image
slices where the signal regions had been placed. The mean of these background regions
was used as BG and their standard deviation was used as N. The PET detector was
11/27
switched on and data was acquired during the MR acquisition. The performance of the
PET detector inside the MR scanner and while applying MR sequences was assessed by
acquiring position profiles and energy spectra, and subsequently comparing them to the
measurements carried out outside the MR scanner.
Coincidence Detector Setup for Simultaneous PET/MR
Since the previous tests demonstrated good performance of the optimized LSO-APD PET
detector, even when operated simultaneously with the MR, two PET detector modules
were mounted with a distance of 69 mm between the crystal block surfaces on a gantry
made of PMMA (Fig 3) and set-up in coincidence to allow the acquisition of PET data.
The gantry had an inner diameter of 60 mm to hold the micro imaging RF-probe and an
outer diameter of 200 mm to fit inside the BG-A 20 gradient set of the MR scanner. The
signal processing was similar to the single detector set-up. Coincidences (12 ns time
window) were generated by feeding the fast summed analog output signal of each
detector block into a CFD (CFD 103, PSI, Switzerland) and both CFD outputs into a
coincidence unit (LRS 466, LeCroy, USA) triggering the 8 channel DAQ board. Setup
was prepared by one author (M.S.J).
PET imaging was performed by rotating the object using a step and shoot acquisition with
12 projections over 180°. After acquisition, all data were sorted into a sinogram and each
projection was normalized and corrected for decay by scaling the respective sinogram
values. Normalization data were gained by acquiring coincidence data from a flood
source phantom with 22 x 22 x 1 mm³ inner dimensions and filled with 20 MBq of
18F
which was placed in between the detectors. The normalized PET emission sinogram data
were reconstructed without further corrections using standard filtered back projection (cutoff frequency 0.5) into a 128 x 128 matrix. The images were post smoothed with a
12/27
2.5 mm (FWHM) Gaussian filter. Imaging was performed by two authors working together
(M.S.J and S.S)
Simultaneous [18F] FDG PET and MR Imaging of a Mouse Head
All animal experiments were performed according the guidelines of the University of
Tübingen for the use of living and dead animals in scientific studies and the German law
for the protection of animals. One female C57BL/6 mouse was intravenously injected with
200 MBq [18F] FDG and sacrificed 45 minutes after tracer uptake. Since the simultaneous
PET/MR measurements were carried out outside the University at Bruker BioSpin in
Ettlingen, Germany, federal regulations limited the maximum amount of activity handled
at their site. Thus, only the mouse head was transported to the company’s site having a
remaining activity of about 8 MBq at the start of the scan (4 h post injection). The head of
the mouse was placed in the 19 mm axial FOV of the PET scanner. A total of 12
projections with 6 min duration each were acquired. During each projection, coronal MR
images were acquired using a FLASH imaging sequence (TR = 394 ms, TE = 5.9 ms, Flip
Angle 40°, 6 Averages, 1 mm slice thickness, 256 x 256 pixels, 6 min acquisition time).
The PET data were reconstructed as described. Coronal views of the PET data were
fused with the MR images using the MiraView Software (Siemens Preclinical Solutions,
USA). The match of the PET data to the MR data was visually evaluated. MR image
evaluation was performed by three authors in consensus (M.S.J., W.I.J., and B.J.P.). PET
image acquisition, fusion, and evaluation were performed by two authors in consensus
(M.S.J., B.J.P.)
13/27
RESULTS
Performance of the LSO-APD PET Detector outside the Magnetic Field (Fig 4,
Table 1)
The measured preamplifier performance in combination with the front end board, modified
for use in an MR scanner was maintained and comparable to the results previously
reported (17). All crystals in the position profile acquired from the LSO-APD PET detector
outside the MR scanner could be separated and the peak to valley ratio measured from a
profile through a centre crystal row was 8.8 ± 2.9 for the centre crystals and 2.7 ± 1.3 for
the crystals at the edges. The energy spectra of one centre and one corner crystal clearly
show the 511 keV peak well separated from the Compton edge. The mean energy
resolution of 4 centre crystals was 14.6 ± 0.3 % and 19.9 ± 3.3 % for the crystals in the
corners of the block. The lower 511 keV photopeak position of the corner crystals indicate
a 21.5 % light loss compared to the centre crystals.
Performance of the LSO-APD PET Detector inside the 7 Tesla MR Scanner
The MR images, acquired with and without the operating PET detector mounted to the
RF-coil, show no considerable degradation in image quality, especially not towards the
bottom side of the phantom where the PET detector was located (Fig 5). The notch on
the top is due to an air bubble inside the phantom. The S/N of the phantom imaged
without the detector was 175, and the S/BG ratio was 92. When the detector was placed
in the scanner, the S/N was 177 and the S/BG 93. The quantitative analysis of the MR
images confirms that the MR image quality is maintained when the PET detector is used
inside the scanner. The position profile acquired from the PET detector when located
inside the MR scanner and while acquiring a FLASH imaging sequence (Fig 4) showed
only minor changes compared to the position profile acquired outside the scanner (Fig 4).
The peak to valley ratios dropped to 5.7 ± 1.9 for the centre crystals and 2.4 ± 1.3 for the
14/27
edge crystals. The mean 511 keV photopeak positions for corner and centre crystals
were decreased by 6% compared to the measurements outside the magnet (Table 1).
The energy resolution of the centre crystals (FWHM @ 511 keV) increased from
14.6% ± 0.3% to 15.9% ± 0.7% when measured inside the magnet while applying MR
imaging sequences.
Simultaneous [18F] FDG PET and MR Imaging of a Mouse Head
No degradation of image quality, in either PET or MR, could be observed in the
simultaneously acquired images of the mouse head (Fig 6). The fused PET-MR images
show a match of anatomy and FDG uptake, especially in the cortex and harderian glands
of the mouse (Fig 6).
15/27
DISCUSSION
Our test results of the preamplifier in combination with the dedicated multi-layer board,
modified for MR applications, showed superior performance than the hybrid amplifier
used in the prototype setup (19). The test of the fully assembled LSO-APD block detector
shows that all crystals can be resolved in the acquired position profiles. The mean energy
resolution of the block detector was 14.6 % for the centre crystals and 19.9 % for the
corner crystals. These results are comparable to other APD based block detectors (21)
and even to state-of-the-art PMT block readout schemes (21-23).
When the modified PET detector module was used inside the 7 Tesla scanner and while
acquiring MR images, the energy resolution increased from 14.6 % to 15.9 % for the
centre crystals and from 19.9 % to 21.9 % for the corner crystals. This is only an absolute
change of 2 % in energy resolution and is most probably a result of a slight increase in
temperature inside the MR system. Earlier tests have shown that the gain and noise of a
LSO-APD detector vary with the change of temperature by approximately 3%/K. The MR
phantom images, in sharp contrast to the results presented in (19), show no visible
interference when imaging is performed while an operating PET detector is located
around the RF-coil. The S/N and S/BG remained the same when MR imaging was carried
out with or without any PET detector material inside the MR scanner. This confirms that
the PCB material covered with a very thin copper layer on both sides proved to be a good
choice as PET detector shielding as well as a material that drastically reduces eddy
currents compared to other shielding materials tested (19).
The two coincident PET detectors showed good performance when used inside the
7 Tesla system. The first simultaneous PET/MR acquisition of a FDG mouse head
revealed no degradation of image quality for both MR and PET. To our knowledge, this is
the first example of simultaneous PET/MR imaging of a biological specimen with a
combined PET/MR scanner in which the entire PET detector resides in the active imaging
16/27
region of the MR. The fused images showed the expected FDG uptake in the brain and
harderian glands of the mouse which match the anatomical landmarks. Since the PET
FOV is physically aligned with the MR FOV, fusing of PET and MR images was not a
problem. Compared to other approaches realizing PET/MR systems (5, 7-9, 24), the
LSO-APD detector used in our study provides multiple PET slices covering an axial field
of view of 19 mm and is easily extended due to the very compact and modular nature of
the PET detector block design. In addition, the fully integrated detector design makes
light fibers redundant and provides better energy resolution (8), although further studies
with a full PET detector ring in the magnet are required to show whether MR image
quality is still maintained and spectroscopy is feasible.
Our study had limitations: First, we used only two PET detectors in coincidence which
prevents us from acquiring dynamic data. However the aim of our study was to evaluate
whether combined PET and MR acquisition is feasible with our detector design. Second,
potential artifacts from a full ring PET system need further evaluations from the MR
perspective. Third, the capability to perform more demanding MR sequences like echo
planar imaging (EPI) needs to be assessed in the presence of the full system.
In Conclusion, our study confirms that simultaneous PET and high field MR imaging with
LSO-APD based PET detectors is feasible without sacrificing image quality of either
system. Our next step is to focus on the evaluation of the full ring PET insert. While this
work concentrates on small animal PET/MR imaging, the results can be transferred to a
clinical system, where the MR usually works at much lower magnetic fields. The
combination of PET and MR imaging can open new opportunities in preclinical research
and clinical diagnosis. When using nuclear magnetic resonance (NMR) spectroscopy
together with MR imaging and PET, tri-modal imaging might be possible, therefore adding
even more information in biomedical research studies.
17/27
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
18/27
Townsend DW, Carney JP, Yap JT, Hall NC. PET/CT today and tomorrow. J Nucl Med
2004; 45 Suppl 1:4S-14S.
Antoch G, Saoudi N, Kuehl H, et al. Accuracy of whole-body dual-modality fluorine-18-2fluoro-2-deoxy-D-glucose positron emission tomography and computed tomography
(FDG-PET/CT) for tumor staging in solid tumors: comparison with CT and PET. J Clin
Oncol 2004; 22:4357-4368.
Beyer T, Antoch G, Muller S, et al. Acquisition protocol considerations for combined
PET/CT imaging. J Nucl Med 2004; 45 Suppl 1:25S-35S.
Reinhardt MJ, Joe AY, Jaeger U, et al. Diagnostic performance of whole body dual
modality 18F-FDG PET/CT imaging for N- and M-staging of malignant melanoma:
experience with 250 consecutive patients. J Clin Oncol 2006; 24:1178-1187.
Shao Y, Cherry SR, Farahani K, et al. Simultaneous PET and MR imaging. Phys Med Biol
1997; 42:1965-1970.
Pichler B, Lorenz E, Mirzoyan R, et al. Performance test of a LSO-APD PET module in a
9.4 Tesla magnet. In:Conference Record of the 1998 IEEE Nuclear Science Symposium
and Medical Imaging Conference. Piscataway, NJ: IEEE, 1998; 1237–1239.
Shao Y, Cherry SR, Farahani K, et al. Development of a PET detector system compatible
with MRI/NMR systems. IEEE Trans. Nucl Sci. 1997; 44:1167.
Slates R, Cherry S, Boutefnouchet A, Yiping S, Dahlborn M, Farahani K. Design of a
small animal MR compatible PET scanner. IEEE Trans. Nucl Sci. 1999; 46:565.
Yamamoto S, Takamatsu S, Murayama H, Minato K. A block detector for a multislice,
depth-of-interaction MR-compatible PET. IEEE Trans. Nucl Sci. 2005; 52:33.
Thompson CJ, Paus T, Clancy R. Magnetic shielding requirements for PET detectors
during transcranial magnetic stimulation. IEEE Trans. Nucl Sci. 1998; 45:1303.
Lecomte R, Cadorette J, Rodrigue S, et al. Initial results from the Sherbrooke avalanche
photodiode positron tomograph. IEEE Trans. Nucl Sci. 1996; 43:1952.
Yang Y, Dokhale PA, Silverman RW, et al. Depth of interaction resolution measurements
for a high resolution PET detector using position sensitive avalanche photodiodes. Phys
Med Biol 2006; 51:2131-2142.
Ziegler SI, Pichler BJ, Boening G, et al. A prototype high-resolution animal positron
tomograph with avalanche photodiode arrays and LSO crystals. Eur J Nucl Med 2001;
28:136-143.
Pichler B, Boning C, Lorenz E, et al. Studies with a prototype high resolution PET scanner
based on LSO-APD modules. IEEE Trans. Nucl Sci. 1998; 45:1298.
Gilbert KM, Handler WB, Scholl TJ, Odegaard JW, Chronik BA. Design of field-cycled
magnetic resonance systems for small animal imaging. Phys Med Biol 2006; 51:28252841.
Lucas A, Fryer T, Clark J, Ansorge R, Siegel S, Carpenter A. Development of a combined
microPET - MRI system. In:Molecular Imaging. Cologne, Germany: Decker, 2005; 266.
Pichler BJ, Swann BK, Rochelle J, Nutt RE, Cherry SR, Siegel SB. Lutetium
oxyorthosilicate block detector readout by avalanche photodiode arrays for high resolution
animal PET. Phys Med Biol 2004; 49:4305-4319.
Catana C, Wu Y, Judenhofer M, Pichler B, Cherrry S. A PSAPD-Based System for
Simultaneous Multi-Slice PET and MRI. In:Molecular Imaging and Biology, 2006; 82.
Pichler BJ, Judenhofer MS, Catana C, et al. Performance test of an LSO-APD detector in a
7-T MRI scanner for simultaneous PET/MRI. J Nucl Med 2006; 47:639-647.
20.
21.
22.
23.
24.
19/27
Judenhofer MS, Pichler BJ, Cherry SR. Evaluation of high performance data acquisition
boards for simultaneous sampling of fast signals from PET detectors. Phys Med Biol 2005;
50:29-44.
Grazioso R, Aykac M, Casey M, Given G, Schmand M. APD Performance in light sharing
PET applications. IEEE Tran. Nucl. Sci. 2005; 52:1413-1416.
Aykac M, Bauer F, Williams CW, Loope M, Schmand M. Timing Performance of Hi-Rez
Detector for Time-of-Flight (TOF) PET. IEEE Trans. Nucl Sci. 2006; 53:1084.
Soonseok K, Wai-Hoi W, Shuping X, et al. High resolution GSO block detectors using
PMT-quadrant-sharing design for small animal PET. IEEE Trans. Nucl Sci. 2006; 53:40.
Mackewn JE, Strul D, Hallett WA, et al. Design and development of an MR-compatible
PET scanner for imaging small animals. IEEE Trans. Nucl Sci. 2005; 52:1376.
Tables
TABLE 1
Energy Resolution and Peak Positions Measured with the LSO-APD Detector Inside and Outside the MRI
Scanner
Crystal location
Mean energy resolution
± SD1 (FWHM2@511keV)
Mean peak position ± SD1
(normalized)
Outside scanner
centre crystals
14.6 ± 0.32 %
1.00 ± 0.02
Outside scanner
corner crystals
19.9 ± 3.34 %
0.82 ± 0.11
Inside scanner
centre crystals
15.9 ± 0.70 %
0.94 ± 0.02
Inside scanner
corner crystals
21.9 ± 4.97 %
0.77 ± 0.10
1SD
= standard deviation
= full width at half maximum
2FWHM
20/27
Captions
Figure 1. Schematic view from the front (left) and side (right) of the magnet with the
gradient set, PET system, and RF-probe.
Figure 2. Top: Crystal block with light guide (a) and APD (b) mounted on the custom
made flex-rigid PCB board containing the preamplifier (d), buffers (e), and connectors (f).
The flexible APD receptacle (c) enables adaptation for accurate radial positioning of the
PET detector.
Centre: PCB board (28 x 192 mm²) and crystal mounted into detector housing.
Bottom: Fully enclosed PET detector.
Figure 3. Top: Prototype gantry, fitting inside the gradient set holding 2 PET detectors
and the RF-coil, used to acquire first simultaneous PET/MR images inside a 7 Tesla
magnet.
Bottom: Front view of PET gantry with RF-coil placed inside the scanner.
Figure 4. Top: Position profiles of the LSO-APD PET detector (12 x 12 crystals) acquired
inside and outside the scanner show that all 144 crystals can be identified.
Centre: Plot of energy histograms of centre and corner crystals acquired inside and
outside the scanner.
Bottom: Profiles through the position profile acquired inside and outside the scanner
comparing the peak to valley ratios.
Figure 5. MR images (fast low angle shot, no contrast media, TR = 400 ms, TE = 6 ms,
Flip angle = 30°, 256 x 256 pixels, axial slices), of a silicone oil phantom acquired with
and without the PET detector mounted to the RF-coil. The box marks the position of the
PET detector.
Figure 6. Simultaneously acquired PET (filtered back projection, 2.5 mm Gaussian post
smoothing filter) and MR (fast low angle shot, no contrast media, TR = 394 ms, TE = 5.9
ms, Flip Angle 40°, 6 Averages, 1 mm slice thickness, 256 x 256 pixels) images (coronal
views) of a mouse head injected with [18F] FDG. The fused images (centre row) show a
very good alignment of the two imaging modalities. The increased uptake in the PET
images maps to the location of the harderian glands behind the eyes in the MR images.
Figure 1
21/27
22/27
Figure 2
23/27
Figure 3
24/27
Figure 4
25/27
Figure 5
26/27
Figure 6
27/27