Journal of Radiology in press. Simultaneous PET/MR Images, acquired with a Compact MRI Compatible PET Detector in a 7 Tesla Magnet

Transcription

1 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 BS 1, Ciprian Catana 2, Brian, K. Swann 3, Stefan Siegel PhD 3, Wulf-Ingo Jung PhD 4, Robert Nutt 3, Simon R. Cherry PhD 2, Claus D. Claussen MD 1 and Bernd J. Pichler PhD 1 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 1/2727

2 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 [ 18 F]-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. 2/2727

3 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 (5-9). 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 3/2727

4 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 4/2727

5 coincidence were used to acquire PET images of a [ 18 F]-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. 5/2727

6 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 (5-9). 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 6/2727

7 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. 7/2727

8 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 mm_ 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 mm_ crystal block (Siemens Preclinical Solutions, USA) comprising 12 x 12 individual 1.5 x 1.5 x 4.5 mm_ 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 mm_ (17). The APDs are operated on a negative bias voltage and reach break-down at approx 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 9- channel 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 8/2727

9 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 9/2727

10 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 68 Ge 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 PD2- MFS-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.) 10/2727

11 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 11/2727

12 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 mm_ inner dimensions and filled with 20 MBq of 18 F 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 12/2727

13 2.5 mm (FWHM) Gaussian filter. Imaging was performed by two authors working together (M.S.J and S.S) Simultaneous [ 18 F] 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 [ 18 F] 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.) 13/2727

14 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 14/2727

15 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 511 kev) increased from 14.6% ± 0.3% to 15.9% ± 0.7% when measured inside the magnet while applying MR imaging sequences. Simultaneous [ 18 F] 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). 15/2727

16 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 16/2727

17 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. 17/2727

18 References 1. Townsend DW, Carney JP, Yap JT, Hall NC. PET/CT today and tomorrow. J Nucl Med 2004; 45 Suppl 1:4S-14S. 2. Antoch G, Saoudi N, Kuehl H, et al. Accuracy of whole-body dual-modality fluorine fluoro-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: 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. 4. 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: Shao Y, Cherry SR, Farahani K, et al. Simultaneous PET and MR imaging. Phys Med Biol 1997; 42: 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; 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: 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: 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: Thompson CJ, Paus T, Clancy R. Magnetic shielding requirements for PET detectors during transcranial magnetic stimulation. IEEE Trans. Nucl Sci. 1998; 45: Lecomte R, Cadorette J, Rodrigue S, et al. Initial results from the Sherbrooke avalanche photodiode positron tomograph. IEEE Trans. Nucl Sci. 1996; 43: 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: 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: 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: 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: 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; 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: 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; 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: /2727

19 20. 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: Grazioso R, Aykac M, Casey M, Given G, Schmand M. APD Performance in light sharing PET applications. IEEE Tran. Nucl. Sci. 2005; 52: 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: 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: 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: /2727

20 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 ± SD 1 (FWHM Mean peak position ± SD 1 (normalized) Outside scanner centre crystals Outside scanner corner crystals Inside scanner centre crystals Inside scanner corner crystals 14.6 ± 0.32 % 1.00 ± ± 3.34 % 0.82 ± ± 0.70 % 0.94 ± ± 4.97 % 0.77 ± SD = standard deviation 2 FWHM = full width at half maximum 20/2727

21 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 mm_) 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 [ 18 F] 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 21/2727

22 22/2727

23 Figure 2 23/2727

24 Figure 3 24/2727

25 Figure 4 25/2727

26 Figure 5 26/2727

27 Figure 6 27/2727