British Journal of Radiology 75 (2002),S53-S59 © 2002 The British Institute of Radiology
Simultaneous PET and NMR
P K Marsden1,
D Strul1,
S F Keevil1,
S C R Williams2 and
D Cash2
1 Division of Radiological Sciences, Guy's, King's and St Thomas' School of Medicine, Guy's Hospital, London SE1 9RT and 2 Neuroimaging Research Group, Institute of Psychiatry, London SE5 8AF, UK
 |
Abstract
|
|---|
There is currently great interest in combining data from different imaging modalities, either by image registration methods that are performed after the data has been acquired or using new devices that can acquire data from two modalities simultaneously, or near simultaneously. In this paper a small prototype NMR-compatible PET scanner capable of acquiring PET images simultaneously with either NMR images or NMR spectra is described. In an associated paper [1], Pamela Garlick describes some investigations of cardiac metabolism that have been made using this system. One of the main challenges in constructing an NMR-compatible PET scanner is that photomultiplier tubes, which are an essential element of nearly all current PET systems, will not function in a high magnetic field. In collaboration with Simon Cherry and the Crump Institute of Biomedical Imaging at UCLA Medical School, a small (5.4 cm diameter) NMR-compatible PET scanner that will operate within the bore of an NMR magnet has been developed. Long optical fibres are used to transport light from the scintillation crystals that form the detector head to photomultiplier tubes situated in a low magnetic field region several metres from the magnet. This system has been used to perform simultaneous PET and NMR spectroscopy measurements with a 9.4T spectroscopy system, and has also been used to obtain simultaneous PET and MR images in several MRI scanners including a 4.7T small bore animal imaging system. Current efforts in the development of this technology are directed at experimental studies on small animals, both because this is less demanding technically and because it is in this area that applications are likely to appear first. However, there is no reason in principle why human PET-MR would not be feasible. Below, work with the prototype system and the next stage in its development are described, and some of the future possibilities and challenges are discussed.
|
Why combine PET and NMR?
|
|---|
Spatial correlation
The most straightforward advantage of performing PET and NMR imaging simultaneously is that in principle it is possible to obtain "perfect" spatial registration of functional PET and anatomical (or functional) MR images. In addition to accurate anatomical localization, highly accurate image registration has other advantages, including the possibility of using the MR image to correct for PET partial volume effects and to aid reconstruction of the PET image. Spatial registration of independently acquired PET and MR images is currently performed retrospectively using sophisticated software, and recent techniques can account, to some extent, for non-rigid tissue deformation that may occur between the two acquisitions. Incorporation of PET and MR scanners into a single gantry would keep subject motion and tissue deformation between PET and MR acquisitions to an absolute minimum, as is the approach adopted for the combined PET and CT systems described elsewhere in this special issue [2, 3].
Temporal correlation
More exciting than accurate spatial correlation is the potential for temporal correlation of functional PET and functional MR data. This principle is illustrated in Figure 1
, which shows temporally correlated 31P NMR spectra and radiotracer (18F-fluorodeoxyglucose (FDG)) uptake data acquired from an isolated perfused rat heart. Changes in both the NMR and PET data are recorded simultaneously in response to changes made to the perfusate supplying the heart. The radiotracer uptake data was acquired using a simple non-imaging gamma-ray detector, and the NMR spectra from a small coil surrounding the heart, in a 9.4T magnet [4]. The principle can be readily extended to imaging studies where regional functional MR data (e.g. perfusion or the distribution of a contrast agent) could be acquired simultaneously with regional PET data showing the evolution of the distribution of a labelled compound. True simultaneous imaging would also allow the cross-validation of MR and PET methodologies, and would remove confounding factors in complex experimental protocols.
View larger version (12K):
[in this window]
[in a new window]
|
Figure 1. Temporally correlated 31P NMR spectra and radiotracer (18F-fluorodeoxyglucose, FDG) uptake data acquired from an isolated perfused rat heart. Changes in both the NMR and PET data are recorded simultaneously as the heart changes from normoxic (A) to hypoxic (B) and then becomes normoxic again (C). The radiotracer uptake data were acquired using a simple non-imaging gamma-ray detector.
|
|
In studies such as those with the isolated perfused rat heart where there is consderable variation between individual experimental runs, simultaneous acquisition will allow the use of paired statistics in data analysis, thus significantly increasing the statistical power of the study. In experimental protocols involving animals this means that the total number of animals required to obtain a significant result is reduced. Looking forward to human studies, simultaneous acquisition would reduce the total imaging time for patients, and for children would remove the need for sedatives to be administered for two separate scanning sessions. Whilst simultaneous acquisition should provide improved image registration, there remain many outstanding problems associated with both periodic and non-periodic subject motion (e.g. respiration and peristalsis) during PET acqusition—simultaneous acquisition of dynamic MR images may provide a means of addressing these.
True simultaneous PET and MR acquisition requires close integration of the PET and MR scanners, and for a standard cylindrical magnet the most straightforward approach is to construct a PET scanner that will operate within the bore of the NMR magnet.
Reduction of positron range in a strong magnetic field
The idea of acquiring PET and NMR data simultaneously is relatively recent; however, the idea of performing PET in a strong magnetic field is not. In 1986 Iida et al [5] published results of a simulation showing that the reduction in positron range that occurs in a strong magnetic field could in principle lead to an improvement in the resolution obtainable with PET, and experimental measurements of this effect have been made by Hammer et al [6]. In order to have a significant effect a field of 5–10T is required and the resolution improvement appears only to be significant for positron emitting radionuclides with relatively high positron energies (e.g. 15O, 68Ga), which are not those in most frequent use. However, the spatial resolution of the latest generation of small animal PET scanners is around 1 mm full width at half-maximum (FWHM), and small animal MR imagers may have field strengths of 7T or higher. In this regime, the effect may yet prove relevant to improving the resolution beyond the limits imposed by positron range.
|
Prototype MR–PET system
|
|---|
The two main challenges in constructing an NMR-compatible PET scanner are firstly that the PET scanner must not adversely effect the operation of the MR scanner and secondly that the MR scanner must not adversely effect the PET. Magnetic or conducting materials placed within the bore of an MR scanner are likely to produce distortions and artefacts in MR images, and conversely the operation of most photomulitplier tubes, which are one of the main components of all conventional PET scanners, is adversely affected by even a relatively weak magnetic field. Ideally, the performance of both modalities would be comparable with that obtainable with state of the art dedicated systems.
The prototype MR-compatible PET system we have developed is shown schematically in Figure 2. The system is based on the technology used for the "microPET" small animal PET scanner developed by Simon Cherry at UCLA [7]. The microPET scanner is not itself MR-compatible; however, its design incorporates short optical fibres to transport light from the scintillation crystals that comprise the detector head to multichannel photomultiplier tubes situated 24 cm away. The MR-compatible prototype scanner [8] consists of a 54 mm diameter annulus of 72 small lutetium oxyorthosilicate (LSO) scintillation crystals. Each crystal is individually connected to one of three multichannel photomultiplier tubes by an optical fibre; however the length of the fibres has been extended to 4 m. This allows the photomultiplier tubes to be placed in a low field region away from the NMR magnet whilst the crystal annulus itself is placed within the magnet. The feasibility of this approach depends on how much the scintillation light is attenuated by the long fibres, and how far away from the magnet the photomultiplier tubes need to be. The combination of high light output LSO scintillation crystals, Photonis XP1722 multichannel photomultiplier tubes and 4 m of fibre has been shown to be effective for a range of magnets with field strengths ranging from 0.2T to 9.4T.
View larger version (14K):
[in this window]
[in a new window]
|
Figure 2. Prototype NMR-compatible PET scanner. The scintillation crystals are connected to the photomultiplier tubes by 4 m long optical fibres. This allows the photomultipliers to be situated in a low field region away from the NMR magnet.
|
|
|
Results from prototype system
|
|---|
Simultaneously acquired images of small NMR and PET visible test phantoms such as those shown in Figure 3 have demonstrated that the presence of the PET scanner causes negligible degradation to the NMR image quality, even for very fast MR pulse sequences [9]. Likewise, the operation of the PET scanner is not affected by either static magnetic or RF fields from the MR scanner. In experiments with a 9.4T spectroscopy system, a 30% degradation in the 31P line width was seen when the PET scanner was in position; however, this was considered to be not significant relative to the degradations that are introduced whenever biological samples are introduced into the magnet [10].
View larger version (45K):
[in this window]
[in a new window]
|
Figure 3. Images of small phantoms obtained simultaneously with PET and MR. PET images were acquired with 18F in 15 min (total counts  300k, activity in phantom 10 MBq). MR images were acquired on a 4.7T, 30 cm bore system in 15 min (spin echo, TE 30 ms, TR 2000 ms). For the hot-spot phantom—spot diameter 2 mm, spot separation 6 mm. The spatial resolution of the PET scanner is about 2 mm FWHM.
|
|
Dynamic images of FDG uptake in the isolated perfused rat heart have been acquired for several studies of cardiac metabolism. An example PET image is shown in Figure 4. For these studies each side of the rat heart (diameter 1 cm) is perfused independently, allowing one side to act as a control while the composition of the perfusate supplying the alternate side is varied. The myocardium is clearly visualized and independent tracer time–activity curves can be plotted for each side of the heart. Small, independent MR coils mounted on opposing sides of the heart allow spectra corresponding to each side to be acquired sequentially. A PET image with an adequate signal to noise ratio can be obtained in 5–10 min during which time NMR spectra from both sides can also be obtained. Examples of results obtained with this system are presented in the accompanying paper by Pamela Garlick [1].
View larger version (53K):
[in this window]
[in a new window]
|
Figure 4. Example FDG uptake image of an isolated perfused rat heart acquired with the prototype scanner. Most of the uptake is in the left ventricle, and a faint outline of the right ventricle can just be seen. These images took 10 min to acquire.
|
|
The prototype scanner has also been used to obtain simultaneous PET and MR images of the mouse brain. The image shown in Figure 5 was obtained on a 4.7T small bore MR scanner. The PET images were acquired 30 min after the administration of FDG and the expected distribution in the mouse brain is clearly seen. A 3D volume was acquired rapidly (<10 min) by the NMR scanner without moving the mouse. To obtain a volume PET image it was necessary to move the mouse in 2 mm steps through the single slice PET scanner, each slice taking 10 min to acquire. An MR-visible fiducial marker aligned with the PET imaging plane is attached to the PET scanner, thus allowing PET and MR image planes to be accurately registered.
View larger version (103K):
[in this window]
[in a new window]
|
Figure 5. Simultaneously acquired FDG–PET (right) and MR (left) images of the brain of a mouse. The MR image was acquired in a 4.7T, 30 cm bore MR system. The 3D PET volume image was acquired by stepping the mouse throught the PET scanner in 2 mm steps (10 min per step).
|
|
The prototype system appears to be compatible with all the MR systems we have looked at, although this result may not be generalizable to larger PET configurations where a much greater bulk of material, and other materials with potential compatibility problems, are likely to be required.
|
Proposed new developments
|
|---|
Whilst the prototype system has demonstrated that it is possible to obtain PET and MR images simultaneously, it has a poor overall performance relative to a dedicated small animal PET scanner such as microPET. In particular, it has a poor sensitivity and signal to noise ratio, and a very limited field of view (25 mm diameter). We are currently constructing a mkII version of the system based on the same technology, which uses a multilayer approach to address these limitations [11], and which will enable us to demonstrate some real applications of the technique. The new system is still a single slice design, but has an order of magnitude greater sensitivity which, combined with the use of NMR-compatible shielding materials [12], will result in a greatly improved signal to noise ratio. This will allow time–activity curves to be followed for small regions of interest. It is also designed to be portable so that it can be installed rapidly in different MR scanners. A schematic diagram of the new system is shown in Figure 6.
View larger version (29K):
[in this window]
[in a new window]
|
Figure 6. Schematic diagram of the new system currently under construction. This is a multilayer design which will have an order of magnitude greater sensitivity than the current prototype, allowing tracer uptake to be followed dynamically for small regions of interest.
|
|
|
Conclusion and possibilities
|
|---|
We have shown that simultaneous PET and MR imaging (or spectroscopy) is feasible for imaging of small animals. The particular detector configuration we have demonstrated is very simple and is well suited to the construction of a single slice imaging system. The next major challenge is to extend this to a multislice system that can image a whole animal in one go. It may be possible to adapt the current optical fibre/photomultiplier type readout to achieve this or it may be advantageous to move to a different detector technology such as avalanche photodiodes [13].
There is no reason in principle why the technology could not be scaled up to a human imaging system, although the problems associated with NMR compatibility (in particular the effect of the PET scanner materials on the MR) may become more serious. Certainly, a near simultaneous configuration such as that adopted for the recent combined PET–CT systems would be more straightforward to construct, but this would not be capable of the more exciting applications that are promised with true simultaneous PET–MR. The main issue for the construction of human system at this stage is probably the motivation to undertake such a project, and our philosophy at present is to construct a fairly conservative device that will allow us to develop the applications of this new technique. With new applications of both PET and MR emerging all the time it is likely that many potential applications for combined PET–MR systems will appear over the next few years.
|
Acknowledgments
|
|---|
Most of this work has been carried out in collaboration with our colleagues Simon Cherry (now UC Davis), Randy Slates, Yiping Shao and Robert Silverman at the Crump Institute for Biomedical Imaging, UCLA, and with Pamela Garlick and colleagues in the Division of Radiological Sciences, Guy's, King's & St. Thomas' School of Medicine, KCL. The work has been funded by the Special Trustees of Guy's and St Thomas' Hospital, the EPSRC, the Dunhill Trust and The Royal Society.
|
References
|
|---|
- Garlick PB. Simultaneous PET and NMR—initial results from isolated, perfused rat hearts. Br J Radiol 2002;75:S60–S66.[Free Full Text]
- Townsend DW, Beyer T. A combined PET/CT scanner: the path to true image fusion. Br J Radiol 2002;75:S24–S30.[Abstract/Free Full Text]
- Steinert HC, von Schulthess GK. Initial clinical experience using a new integrated in-line PET/CT system. Br J Radiol 2002;75:S36–S38.[Free Full Text]
- Buchanan M, Marsden PK, Mielke CH, Garlick PB. A system to obtain radiotracer uptake data simultaneously with NMR spectra in a high field magnet. IEEE Trans Nucl Sci 1996;43:2044–8.
- Iida H, Kanno I, Miura S, Murakami M, Takahashi K, Uemura K. A simulation study of a method to reduce positron-annihilation spread distributions using a strong magnetic-field in positron emission tomography. IEEE Trans Nucl Sci 1986;33:597–600.
- Hammer BE, Christensen NL, Heil BG. Use of a magnetic-field to increase the spatial-resolution positron emission tomography. Med Phys 1994;21:1917–20.[Medline]
- Cherry SR, Shao Y, Silverman RW, Siegel S, Meadors K, Mumcuoglu E, Young J, Jones WF, Moyers C, Andreaco M, Paulus M, Binkley D, Nutt R, Phelps ME. Micropet: a dedicated PET scanner for small animal imaging. J Nucl Med 1996;37:334.
- Shao Y, Cherry SR, Farahani K, Slates R, Silverman RW, Meadors K, Bowery A, Siegel S, Marsden PK, Garlick PB. Development of a PET detector system compatible with MRI/NMR systems. IEEE Trans Nucl Sci 1997;44:1167–71.
- Slates R, Farahani K, Shao Y, Marsden PK, Taylor J, Summers PE, Williams S, Beech J, Cherry SR. A study of artefacts in simultaneous PET and MR imaging using a prototype MR compatible PET scanner. Phys Med Biol 1999;44:2015–27.[Medline]
- Garlick PB, Marsden PK, Cave AC, Parkes HG, Slates R, Shao YP, Silverman RW, Cherry SR. PET and NMR dual acquisition (PANDA): applications to isolated, perfused rat hearts. NMR Biomed 1997;10:138–142.[Medline]
- Strul D, Slates RB, Dahlbom M, Cherry SR, Marsden PK. Use of an analytical model for optimizing the design of a small-animal PET scanner with DOI capability. Proceedings of 2000 IEEE Nuclear Science Symposium Conference; 2000 October 15–20; Lyon, France.
- Strul D, Cash D, Keevil SF, Halstead P, Williams SCR, Marsden PK. MR-compatible shields for 511keV gamma photons: a feasibility study. Proceedings of 2000 IEEE Nuclear Science Symposium Conference; 2000 October 15–20; Lyon, France.
- Ziegler SI, Fries O, Pichler B, Pimpl W, Roder F, Schwaiger M, Lorenz E. Prototype animal PET scanner with avalanche photodiode arrays and LSO crystals. J Nucl Med 1997;38:160.
This article has been cited by other articles:
|
|
|
|
 
C. Goetz, E. Breton, P. Choquet, V. Israel-Jost, and A. Constantinesco
SPECT Low-Field MRI System for Small-Animal Imaging
J. Nucl. Med.,
January 1, 2008;
49(1):
88 - 93.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
T. Higuchi, S. G. Nekolla, A. Jankaukas, A. W. Weber, M. C. Huisman, S. Reder, S. I. Ziegler, M. Schwaiger, and F. M. Bengel
Characterization of Normal and Infarcted Rat Myocardium Using a Combination of Small-Animal PET and Clinical MRI
J. Nucl. Med.,
February 1, 2007;
48(2):
288 - 294.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
P. K.E. Borjesson, Y. W.S. Jauw, R. Boellaard, R. de Bree, E. F.I. Comans, J. C. Roos, J. A. Castelijns, M. J.W.D. Vosjan, J. A. Kummer, C. R. Leemans, et al.
Performance of immuno-positron emission tomography with zirconium-89-labeled chimeric monoclonal antibody u36 in the detection of lymph node metastases in head and neck cancer patients.
Clin. Cancer Res.,
April 1, 2006;
12(7):
2133 - 2140.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
T Jones
Molecular imaging with PET - the future challenges
Br. J. Radiol.,
November 1, 2002;
75(90009):
S6 - 15.
[Full Text]
[PDF]
|
|
|
|
|
|
|
P B Garlick
Simultaneous PET and NMR--initial results from isolated, perfused rat hearts
Br. J. Radiol.,
November 1, 2002;
75(90009):
S60 - 66.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Why combine PET and...
