- Email: [email protected]
Glucose metabolism of the midline nuclei raphe in the brainstem observed by PET–MRI fusion imaging
NeuroImage 59 (2012) 1094–1097
Contents lists available at SciVerse ScienceDirect
NeuroImage journal homepage: www.elsevier.com/locate/ynimg
Glucose metabolism of the midline nuclei raphe in the brainstem observed by PET–MRI fusion imaging Young-Don Son a, 1, Zang-Hee Cho a,⁎, 1, Hang-Keun Kim a, Eun-Jung Choi a, Sang-Yoon Lee a, Je-Geun Chi b, Chan-Woong Park a, Young-Bo Kim a a b
Neuroscience Research Institute, Gachon University of Medicine and Science, Incheon, South Korea Department of Pathology, Seoul National University, Seoul, South Korea
a r t i c l e
i n f o
Article history: Received 17 March 2011 Revised 5 September 2011 Accepted 15 September 2011 Available online 22 September 2011 Keywords: PET–MRI Brainstem Raphe nuclei Fusion imaging
a b s t r a c t The brainstem contains various important monoaminergic neuronal centers, including the raphe nuclei which contain serotonergic neurons. The raphe nuclei, however, are not easily identifiable and located by conventional neuroimaging. Methods: Fluorodeoxyglucose positron emission tomography (PET) and magnetic resonance imaging (MRI) were performed in seven healthy subjects using a new PET–MRI, which consists of a high-resolution research tomograph (HRRT) PET and 7.0 T-MRI. Glucose metabolism of raphe nuclei was semiquantitatively measured and identified along the midline brainstem region in vivo. Results: Midline nuclei clustered in four groups appeared to be the raphe nuclei and could be clearly visualized; specifically, we identified the groups as the dorsal raphe, raphe reticularis centralis superior, raphe pontis, and raphe magnus group. Conclusion: FDG imaging of the midline raphe nuclei in vivo could potentially be an important tool for investigating brain diseases as well as conducting functional brain studies in the context of sleep disorders, depression, and neurodegenerative disease. © 2011 Elsevier Inc. All rights reserved.
Introduction Since the development of positron emission tomography (PET) in the mid-1970s (Cho et al., 1976; Phelps et al., 1978), it has been possible to observe molecular changes and metabolic function of the human brain in vivo. More recently, PET has become increasingly popular for the diagnosis of cancer and tumors, especially with markedly improved resolution and increased availability of the novel radiotracers. PET has also become an important tool for the study of functional and disease states of the human brain, such as the subcortical areas that hitherto was unable to observe. With the development of brain-dedicated highresolution PET, such as high-resolution research tomograph (HRRT) (Schmand et al., 1999; Wienhard et al., 2002), ultra-high-resolution brain imaging of cortical as well as subcortical areas and even subcortical monoaminergic nuclei became possible mainly due to the reduction of the partial-volume effect, one of the major stumbling blocks in the conventional PET for high-resolution brain imaging, especially in the
⁎ Corresponding author at: Neuroscience Research Institute, Gachon University of Medicine and Science, 1198 Kuwol-Dong, Namdong-Gu, Incheon, 405–760, South Korea. Fax: + 82 32 460 8230. E-mail address: [email protected] (Z.-H. Cho). 1 These authors contributed equally to this work. 1053-8119/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2011.09.036
brainstem. On another front, ultra-high-field magnetic resonance imaging (MRI), such as 7.0 T-MRI, also began to permit visualizing the human brain in vivo and to offer extensive detail of brainstem areas, such as the red nucleus and substantia nigra (Cho, 2009). These advances led us to develop a PET–MRI fusion system that combined two high-end imaging systems, HRRT-PET and 7.0 T-MRI, for the study of functional and metabolic activities of the brain in vivo, including the brainstem (Cho et al., 2008). The combination of the functional and neuro-molecular imaging of HRRT-PET and high-resolution structural imaging capability of 7.0 TMRI has opened a new avenue to study brain areas such as the brainstem once thought of as an invisible black box. Recently, the new PET– MRI system was successfully applied to study glucose metabolism of hippocampal substructures and thalamic nuclei of humans in vivo (Cho et al., 2010, 2011). Success in these studies encouraged us to conduct further research, such as in the brainstem, an intriguing and complex brain area that hosts most of the neuronal centers that secrete the major monoaminergic neurotransmitters, including serotonin. The simplest and perhaps most convincing study would involve the raphe nuclei, which is the center of the serotonergic neurons since they unmistakably reside along the dorsal midline of the brainstem starting from the uppermost part of the tagmentum and extending to the beginning of the medulla and are known to mediate numerous functions, such as sleep, appetite, and mood. However, MR images of the human brain in vivo usually have limited contrast and resolution for brainstem nuclei
Y.-D. Son et al. / NeuroImage 59 (2012) 1094–1097
such as the raphe nuclei, and these images have consistently failed to visualize those nuclei. Given the development of serotonergic ligands 11 such as the serotonin transporter ligand [ C]3-amino-4-(2-dimethyla11 minomethylphenylsulfanyl)-benzonitrile ( C-DASB) and the serotonin 11 5HT1A ligand [carbonyl- C]N-(2-(1-(4-(2-methoxyphenyl)-piperazinyl) 11 ethyl)-N-pyridinyl)cyclohexanecarboxamide ( C-WAY100635), PET could measure serotonergic nuclei in the brainstem more directly and convincingly. However, the results of previous studies could not visualize the metabolic activities of the individual raphe nuclei mainly due to the limit of the resolution in the existing PET scanners (Ichise et al., 2003; Kim et al., 2006). The aim of the present study was to visualize and quantitate glucose metabolism in the various individual raphe nuclei by fusion imaging, using the newly developed brain-dedicated high-resolution PET–MRI (Cho et al., 2008, 2010, 2011). Materials and methods This newly developed PET–MRI fusion system consists of HRRTPET and 7.0 T-MRI with a shuttle system through which two modality images are physically combined by the calibrated coordinates with high precision (Cho et al., 2008). Detailed specifications of the system are described in the Supplementary materials. Ten healthy volunteers, eight men and two women (22.9±1.73 years old), were recruited for the study. Among them, three participants (two men and one woman) were excluded due to their excessive motion during PET scanning. Each of the participants signed an informed consent form according to the protocol approved by the internal review board of 18 Gachon University of Medicine and Science. A bolus injection of [ F]fluor18 odeoxyglucose ( F-FDG) (185 MBq) was given to each subject immediately before 7.0 T-MRI. MRI was conducted for 30 min, and after 10 min, PET imaging was implemented and completed in 30 min. T2*-weighted two-dimensional gradient-echo images were obtained by 7.0 T-MRI with the following parameters: repetition time (TR) 750 ms; echo time (TE) 21 ms; flip angle 30°; voxel size 0.18×0.18×1.5 mm; number of slices, 17. During the PET scan, an additional transmission scan using cesi137 um 137 ( Cs) was also performed for attenuation correction. The PET image was reconstructed using the 3D-OP-OSEM algorithm (Hong et al., 2007). The reconstructed image had a matrix size of 256×256×207 with 1.22×1.22×1.22 mm3 iso-voxel resolution. After the scanning, both images were automatically fused by the system, which was precalibrated with submillimeter precision. For the PET glucose uptake evaluation, the standardized uptake value ratio (SUVR) was obtained by normalizing the uptake value of PET images over the uptake value of the cerebellum of each subject. Eight regions of interest (ROIs) were defined and drawn from the PET–MRI fusion images of each subject. Among the eight ROIs, four ROIs were the raphe nucleus groups, according to Naidich et al. (2009), and the rest consisted of four non-raphe ROIs, namely the thalamus, red nucleus, mammillary body, and inferior colliculus. The relative levels of glucose metabolism were compared in both the selected raphe and non-raphe nuclei. Finally, the SUVR of each ROI was obtained from the PET–MRI fusion image for the analysis. Results For the first time, using HRRT-PET and 7.0 T-MRI fusion imaging, we were able to visualize glucose uptake of well-separated clusters of the activities that appear to be the raphe nuclei (Figs. 1a–c). Several localized metabolic activity groups, which are likely to represent the mesencephalic raphe nuclei, were clearly observed in the brainstem by HRRT-PET and further identified by 7.0 T-MRI anatomical image. A unique feature of the serotonergic study is that the raphe nuclei are all located along the midline posterior-dorsomedial brainstem areas (Fig. 1d) and are distinctly identifiable. The first group (R1), including the dorsal nucleus raphe, was located in the upper midbrain area close to the inferior colliculus (landmark structures such as red nucleus
1095
and mammillary body further support that R1 includes the dorsal raphe). Dorsal raphe nuclei, however, were difficult to distinguish from other nuclei in the vicinity because the dorsal raphe is intermingled with the red nucleus, which has relatively higher activity than the dorsal raphe. The second group (R2) was located ventral to the inferior colluculus in the upper portion of the pons. The R2 group, also known as the nucleus reticularis centralis superior, consistently shows the highest level of glucose metabolism among the raphe nuclei groups. The third group (R3), located at the central and posterior aspect of the pons, appeared to correspond to the nucleus raphe pontis. As the smallest among the four raphe nuclei groups, R3 was not easily discernable in two out of seven subjects. The fourth group (R4) was found at the most posterior and inferior aspects of the pons and also at the junction point to the superior aspect of medulla. This group appeared to include the nucleus raphe magnus, nucleus raphe obscurus, and nucleus raphe pallidus. The level of glucose metabolism in the raphe nuclei was found to be much lower than in most of the cerebral cortex, except in the cerebellar cortex, where activities are somewhat higher than those in the brainstem but generally do not reach those of the cerebral cortices (Fig. 1e). Within the vicinity of the raphe nuclei, there are several well-known landmark structures that verified our affirmation of the clusters as the raphe nuclei: the inferior colliculus, red nucleus, and mammillary body. These landmark structures showed relatively higher levels of metabolism in the areas of the brainstem and always with certainty thereby provided us an additional information on the relative locations of the raphe nuclei, especially the high-resolution images obtained from 7.0 T-MRI. These observations were consistently reproducible for all the subjects as shown in Fig. 2a. Although the exact location of each raphe nucleus differed slightly between individuals, the general distribution pattern was very consistent. However, the individual differences resulted in the blurring of the image when the intersubject data were averaged. Nevertheless, the typical pattern of glucose metabolism in the brainstem area was well characterized even after averaging of the images. The spatially normalized PET images of those activity groups were also well matched among the subjects (see Fig. 2b). Discussions Using the new PET–MRI fusion imaging system, which consists of high-end PET (HRRT) and ultra-high-field MRI (7.0 T), glucose metabolism of several clusters of the individual brainstem raphe nuclei were clearly imaged and identified. Although MRI alone does not visualize the raphe nuclei because of low contrast, 7.0 T-MRI of the brainstem areas provided excellent images of several landmark anatomical structures from which it was possible to deduce the brainstem raphe nuclei groups. The results clearly identified the four groups of raphe nuclei in the brainstem area as described by Naidich et al. (2009). One of the unique features of the present study is that because the raphe nuclei are all located right in the midline of the brainstem, they are easy to locate, especially where the corresponding anatomy is clear due to a number of unmistakable landmark structures. Although some previous rat or monkey studies (Fançois et al., 2010; Lyons et al., 1996; Skelin et al., 2008) using high-resolution 14 autoradiography with C-deoxyglucose reported glucose metabolism of the raphe nuclei, to our knowledge, this is the first report of clear visualization of the four groups of human raphe nuclei in vivo using FDG PET. Most of animal studies investigated the dorsal raphe and medial raphe only, except Skelin et al. (2008). Unlike our findings, they found that rats showed highest metabolism in the raphe pontine among raphe nuclei. It could be due to the difference between human and animal or due to the anesthesia applied to rats for experiment. In addition, Heiss et al. (2004) imaged subcortical nuclei of the human brain using FDG-PET with HRRT-PET, but the individual raphe nuclei were not specifically found in that study. Their major interest was to
1096
Y.-D. Son et al. / NeuroImage 59 (2012) 1094–1097
Fig. 1. A representative fusion image set of the brainstem area from a single subject obtained by positron emission tomography–magnetic resonance imaging (PET–MRI). (a) A medial sagittal T2*-weighted image of the brainstem obtained by 7.0 T-MRI. This high-resolution image depicts various landmark structures in the midbrain and brainstem areas such 18 18 as the mammillary body (MB), thalamus, and red nucleus (RN). (b) [ F]Fluorodeoxyglucose ( F-FDG)-PET image obtained by high-resolution research tomograph (HRRT-PET) of a slice corresponding to (a). Many well-known localized activities are clearly visible, such as the MB, RN, and other structures, together with activity cluster groups that are believed to be the raphe nuclei. (c) Fusion image of (a) and (b). From this fusion image, together with histological information provided by Naidich et al. (2009) and ultra-high resolution image of 7.0 T-MRI, it was possible to identify four raphe nucleus groups (arrowheads): dorsal, superior central, pontine, and medullary. (d) Raphe nuclei previously identified in 18 the brain atlas of Naidich et al. (2009) and our new PET–MRI fusion image of the brainstem area obtained by F-FDG. Note the markedly clear and distinct correlation between the locations and distributions of the raphe nuclei shown in the reference of Naidich et al. and those of the raphe nuclei seen in the PET–MRI fusion image. (e) Relative glucose metabolic activities of each raphe nucleus as well as the other midbrain nuclei measured in seven healthy subjects. R2 shows the highest glucose metabolic activity among the raphe nuclei. The other non–raphe nucleus groups, compared with the raphe nucleus groups, the highest level of activities were found in TH followed by IC, RN, and MB. TH: thalamus, MB: mammillary body, IC: inferior colliculus, RN: red nucleus, R1–R4: raphe nuclei, B1: nucleus raphes pallidus, B2: nucleus raphes obscurus, B3: nucleus raphes magnus, B5: nucleus raphes pontis, B6, B8: nucleus centralis superior, B7: dorsal nucleus raphe.
Fig. 2. Individual PET–MRI (HRRT-PET and 7.0 T-MRI) fusion images of seven subjects and an averaged image. (a) Seven intersubject images at the middle sagittal section of the brainstem. All subjects showed similar distributions of the raphe nuclei as well as the surrounding areas. (b) Averaged PET image of the seven subjects spatially normalized to the Montreal Neurological Institute (MNI) template using the SPM8. Although spatial averaging resulted in a blurred image, the overall raphe distribution confirmed that the raphe nuclei distribution is consistent among the subjects.
Y.-D. Son et al. / NeuroImage 59 (2012) 1094–1097
find some of the small nuclei such as the substantia nigra, red nucleus, inferior colliculus, and etc., all of which are substantially larger structures than the raphe nuclei. We further extended their study to focus on the raphe nuclei by taking advantage of high-resolution capability of HRRT PET and 7.0 T-MRI fusion system. It is worth noting that HRRT-PET is a brain-dedicated research PET, which is designed to provide images of the highest sensitivity and resolution, down to 2.5 mm full width at half maximum iso-voxel resolution due to the smallest head-only gantry diameter, which effectively leads to the highest solid angle and, therefore, the highest sensitivity. This high resolution in turn reduces much error due to partial volume averaging effect. In addition, since HRRT-PET covers the whole brainstem area, it is ideally suitable for brainstem imaging. For the SUVR calculation, the cerebellum was chosen as the reference tissue. In order to minimize the variation due to different buildup and wash-out time of the tissue, FDG injection and scanning times were made uniform for all subjects. Glucose uptake of neurons represents neuronal activity, either directly or indirectly. Although the direct correlation between neurotransmitter release and glucose consumption has not yet been clearly established, changes in glucose consumption reflect the neuronal activities, including neurotransmitter and receptor dynamics. In many cases, instead of direct or indirect measurement of the neuroreceptors or neurotransporters, measurement of metabolic activities of the well-separated individual nuclei could be sufficient for the understanding of the role of nuclei, such as serotonergic neurons in the brainstem area. With regard to future work, a multi-tracer study combining FDG imaging and serotonergic receptor or transporter imaging of the same subject would further confirm that the nuclei we observed are indeed the serotonergic nuclei. For example, 11 recent imaging studies with C-DASB showed increased activities in the raphe nuclei (Ichise et al., 2003; Kim et al., 2006), although resolution was too poor to specifically identify any of the individual nuclei. Correla11 18 tion studies between C-DASB and F-FDG will further enable us to assuredly verify that the nuclei with high glucose uptake in the brainstem area are indeed the raphe nuclei. Additionally, quantitative 11 imaging study of FDG and serotonin receptors, such as C-WAY100635, would also provide further clues about the relationship between the neuronal activities of neurotransmitter and receptor and the glucose metabolism.
Conclusions FDG-PET images of the brainstem were obtained by using a new HRRT-PET and 7.0 T-MRI fusion imaging protocol. This highsensitivity and high-resolution molecular imaging provided, for the first time, clearly identifiable in vivo glucose metabolic activities of the raphe nuclei in the human brainstem. This observation was possible due to the improved resolution and sensitivity of HRRT-PET, which in turn reduced the partial-volume effect incorporated with detailed anatomical images of 7.0 T-MRI, effectively guiding us to accurately locate the raphe nuclei in the brain, in correlation with several landmark structures within the brainstem.
1097
Conflict of interest statement The authors declare no conflict of interest. Acknowledgments This work was supported by the National Research Foundation (NRF), the Ministry of Education, Science and Technology (2008–2004159). Appendix A. Supplementary data Supplementary data to this article can be found online at doi:10. 1016/j.neuroimage.2011.09.036. References Cho, Z.H., 2009. 7.0 Tesla MRI Brain Atlas: In Vivo Atlas with Cryomacrotome Correlation. Springer, New York, NY. Cho, Z.H., Chan, J.K., Eriksson, L., 1976. Circular ring transverse axial positron camera for 3-dimensional reconstruction of radionuclides distribution. IEEE Trans. Nucl. Sci. 23, 613–622. Cho, Z.H., Son, Y.D., Kim, H.K., Kim, K.N., Oh, S.H., Han, J.Y., Hong, I.K., Kim, Y.B., 2008. A fusion PET–MRI system with a high-resolution research tomograph-PET and ultrahigh field 7.0T-MRI for the molecular-genetic imaging of the brain. Proteomics 8, 1302–1323. Cho, Z.H., Son, Y.D., Kim, H.K., Kim, S.T., Lee, S.Y., Chi, J.G., Park, C.W., Kim, Y.B., 2010. Substructural hippocampal glucose metabolism observed on PET/MRI. J. Nucl. Med. 51, 1545–1548. Cho, Z.H., Son, Y.D., Kim, H.K., Kim, N.B., Choi, E.J., Lee, S.Y., Chi, J.G., Park, C.W., Kim, Y.B., Ogawa, S., 2011. Observation of glucose metabolism in the thalamic nuclei by fusion PET/MRI. J. Nucl. Med. 52, 401–404. Fançois, J., Koning, E., Ferrandon, A., Sandner, G., Nehlig, A., 2010. Metabolic activity in the brain of juvenile and adult rats with a neonatal ventral hippocampal lesion. Hippocampus 20, 841–851. Heiss, W.D., Habedank, B., Klein, J.C., Herholz, K., Wienhard, K., Lenox, M., Nutt, R., 2004. Metabolic rates in small brain nuclei determined by high-resolution PET. J. Nucl. Med. 45, 1811–1815. Hong, I.K., Chung, S.T., Kim, H.K., Kim, Y.B., Son, Y.D., Cho, Z.H., 2007. Ultra fast symmetry and SIMD-based projection-backprojection (SSP) algorithm for 3-D PET image reconstruction. IEEE Trans. Med. Imaging 26, 789–803. Ichise, M., Liow, J.S., Lu, J.Q., Takano, A., Model, K., Toyama, H., Suhara, T., Suzuki, K., Innis, R.B., Carson, R.E., 2003. Linearized reference tissue parametric imaging methods: application to [11C]DASB positron emission tomography studies of the serotonin transporter in human brain. J. Cereb. Blood Flow Metab. 23, 1096–1112. Kim, J.S., Ichise, M., Sangare, J., Innis, R.B., 2006. PET imaging of serotonin transporters with [11C]DASB: test-retest reproducibility using a multilinear reference tissue parametric imaging method. J. Nucl. Med. 47, 208–214. Lyons, D., Friedman, D.P., Nader, M.A., Porrino, L.J., 1996. Cocaine alters cerebral metabolism within the ventral striatum and limbic cortex of monkeys. J. Neurosci. 16, 1230–1238. Naidich, T.P., Duvernoy, H.M., Delman, B.N., Sorensen, A.G., Kollias, S.S., Haacke, E.M., 2009. Duvernoy's atlas of the human brain stem and cerebellum: high-field MRI, surface anatomy, internal structure, vascularization and 3 D sectional anatomy. Springer, Vienna. Phelps, M.E., Hoffman, E.J., Huang, S.C., Kuhl, D.E., 1978. ECAT: a new computerized tomographic imaging system for positron-emitting radiopharmaceuticals. J. Nucl. Med. 19, 635–647. Schmand, M., Casey, M., Wienhard, K., Eriksson, L., Jones, W., Lenox, M., Young, J., Baker, K., Miller, S., Reed, J., 1999. HRRT a new high resolution LSO-PET research tomograph. J. Nucl. Med. 40, 76. Skelin, I., Sato, H., Diksic, M., 2008. Olfactory bulbectomy reduces cerebral glucose utilization: 2-[14C]deoxyglucose autoradiographic study. Brain Res. Bull. 76, 485–492. Wienhard, K., Schmand, M., Casey, M.E., Baker, K., Bao, J., Eriksson, L., Jones, W.F., Knoess, C., Lenox, M., Lercher, M., 2002. The ECAT HRRT: performance and first clinical application of the new high resolution research tomograph. IEEE Trans. Nucl. Sci. 49, 104–110.
Contents lists available at SciVerse ScienceDirect
NeuroImage journal homepage: www.elsevier.com/locate/ynimg
Glucose metabolism of the midline nuclei raphe in the brainstem observed by PET–MRI fusion imaging Young-Don Son a, 1, Zang-Hee Cho a,⁎, 1, Hang-Keun Kim a, Eun-Jung Choi a, Sang-Yoon Lee a, Je-Geun Chi b, Chan-Woong Park a, Young-Bo Kim a a b
Neuroscience Research Institute, Gachon University of Medicine and Science, Incheon, South Korea Department of Pathology, Seoul National University, Seoul, South Korea
a r t i c l e
i n f o
Article history: Received 17 March 2011 Revised 5 September 2011 Accepted 15 September 2011 Available online 22 September 2011 Keywords: PET–MRI Brainstem Raphe nuclei Fusion imaging
a b s t r a c t The brainstem contains various important monoaminergic neuronal centers, including the raphe nuclei which contain serotonergic neurons. The raphe nuclei, however, are not easily identifiable and located by conventional neuroimaging. Methods: Fluorodeoxyglucose positron emission tomography (PET) and magnetic resonance imaging (MRI) were performed in seven healthy subjects using a new PET–MRI, which consists of a high-resolution research tomograph (HRRT) PET and 7.0 T-MRI. Glucose metabolism of raphe nuclei was semiquantitatively measured and identified along the midline brainstem region in vivo. Results: Midline nuclei clustered in four groups appeared to be the raphe nuclei and could be clearly visualized; specifically, we identified the groups as the dorsal raphe, raphe reticularis centralis superior, raphe pontis, and raphe magnus group. Conclusion: FDG imaging of the midline raphe nuclei in vivo could potentially be an important tool for investigating brain diseases as well as conducting functional brain studies in the context of sleep disorders, depression, and neurodegenerative disease. © 2011 Elsevier Inc. All rights reserved.
Introduction Since the development of positron emission tomography (PET) in the mid-1970s (Cho et al., 1976; Phelps et al., 1978), it has been possible to observe molecular changes and metabolic function of the human brain in vivo. More recently, PET has become increasingly popular for the diagnosis of cancer and tumors, especially with markedly improved resolution and increased availability of the novel radiotracers. PET has also become an important tool for the study of functional and disease states of the human brain, such as the subcortical areas that hitherto was unable to observe. With the development of brain-dedicated highresolution PET, such as high-resolution research tomograph (HRRT) (Schmand et al., 1999; Wienhard et al., 2002), ultra-high-resolution brain imaging of cortical as well as subcortical areas and even subcortical monoaminergic nuclei became possible mainly due to the reduction of the partial-volume effect, one of the major stumbling blocks in the conventional PET for high-resolution brain imaging, especially in the
⁎ Corresponding author at: Neuroscience Research Institute, Gachon University of Medicine and Science, 1198 Kuwol-Dong, Namdong-Gu, Incheon, 405–760, South Korea. Fax: + 82 32 460 8230. E-mail address: [email protected] (Z.-H. Cho). 1 These authors contributed equally to this work. 1053-8119/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2011.09.036
brainstem. On another front, ultra-high-field magnetic resonance imaging (MRI), such as 7.0 T-MRI, also began to permit visualizing the human brain in vivo and to offer extensive detail of brainstem areas, such as the red nucleus and substantia nigra (Cho, 2009). These advances led us to develop a PET–MRI fusion system that combined two high-end imaging systems, HRRT-PET and 7.0 T-MRI, for the study of functional and metabolic activities of the brain in vivo, including the brainstem (Cho et al., 2008). The combination of the functional and neuro-molecular imaging of HRRT-PET and high-resolution structural imaging capability of 7.0 TMRI has opened a new avenue to study brain areas such as the brainstem once thought of as an invisible black box. Recently, the new PET– MRI system was successfully applied to study glucose metabolism of hippocampal substructures and thalamic nuclei of humans in vivo (Cho et al., 2010, 2011). Success in these studies encouraged us to conduct further research, such as in the brainstem, an intriguing and complex brain area that hosts most of the neuronal centers that secrete the major monoaminergic neurotransmitters, including serotonin. The simplest and perhaps most convincing study would involve the raphe nuclei, which is the center of the serotonergic neurons since they unmistakably reside along the dorsal midline of the brainstem starting from the uppermost part of the tagmentum and extending to the beginning of the medulla and are known to mediate numerous functions, such as sleep, appetite, and mood. However, MR images of the human brain in vivo usually have limited contrast and resolution for brainstem nuclei
Y.-D. Son et al. / NeuroImage 59 (2012) 1094–1097
such as the raphe nuclei, and these images have consistently failed to visualize those nuclei. Given the development of serotonergic ligands 11 such as the serotonin transporter ligand [ C]3-amino-4-(2-dimethyla11 minomethylphenylsulfanyl)-benzonitrile ( C-DASB) and the serotonin 11 5HT1A ligand [carbonyl- C]N-(2-(1-(4-(2-methoxyphenyl)-piperazinyl) 11 ethyl)-N-pyridinyl)cyclohexanecarboxamide ( C-WAY100635), PET could measure serotonergic nuclei in the brainstem more directly and convincingly. However, the results of previous studies could not visualize the metabolic activities of the individual raphe nuclei mainly due to the limit of the resolution in the existing PET scanners (Ichise et al., 2003; Kim et al., 2006). The aim of the present study was to visualize and quantitate glucose metabolism in the various individual raphe nuclei by fusion imaging, using the newly developed brain-dedicated high-resolution PET–MRI (Cho et al., 2008, 2010, 2011). Materials and methods This newly developed PET–MRI fusion system consists of HRRTPET and 7.0 T-MRI with a shuttle system through which two modality images are physically combined by the calibrated coordinates with high precision (Cho et al., 2008). Detailed specifications of the system are described in the Supplementary materials. Ten healthy volunteers, eight men and two women (22.9±1.73 years old), were recruited for the study. Among them, three participants (two men and one woman) were excluded due to their excessive motion during PET scanning. Each of the participants signed an informed consent form according to the protocol approved by the internal review board of 18 Gachon University of Medicine and Science. A bolus injection of [ F]fluor18 odeoxyglucose ( F-FDG) (185 MBq) was given to each subject immediately before 7.0 T-MRI. MRI was conducted for 30 min, and after 10 min, PET imaging was implemented and completed in 30 min. T2*-weighted two-dimensional gradient-echo images were obtained by 7.0 T-MRI with the following parameters: repetition time (TR) 750 ms; echo time (TE) 21 ms; flip angle 30°; voxel size 0.18×0.18×1.5 mm; number of slices, 17. During the PET scan, an additional transmission scan using cesi137 um 137 ( Cs) was also performed for attenuation correction. The PET image was reconstructed using the 3D-OP-OSEM algorithm (Hong et al., 2007). The reconstructed image had a matrix size of 256×256×207 with 1.22×1.22×1.22 mm3 iso-voxel resolution. After the scanning, both images were automatically fused by the system, which was precalibrated with submillimeter precision. For the PET glucose uptake evaluation, the standardized uptake value ratio (SUVR) was obtained by normalizing the uptake value of PET images over the uptake value of the cerebellum of each subject. Eight regions of interest (ROIs) were defined and drawn from the PET–MRI fusion images of each subject. Among the eight ROIs, four ROIs were the raphe nucleus groups, according to Naidich et al. (2009), and the rest consisted of four non-raphe ROIs, namely the thalamus, red nucleus, mammillary body, and inferior colliculus. The relative levels of glucose metabolism were compared in both the selected raphe and non-raphe nuclei. Finally, the SUVR of each ROI was obtained from the PET–MRI fusion image for the analysis. Results For the first time, using HRRT-PET and 7.0 T-MRI fusion imaging, we were able to visualize glucose uptake of well-separated clusters of the activities that appear to be the raphe nuclei (Figs. 1a–c). Several localized metabolic activity groups, which are likely to represent the mesencephalic raphe nuclei, were clearly observed in the brainstem by HRRT-PET and further identified by 7.0 T-MRI anatomical image. A unique feature of the serotonergic study is that the raphe nuclei are all located along the midline posterior-dorsomedial brainstem areas (Fig. 1d) and are distinctly identifiable. The first group (R1), including the dorsal nucleus raphe, was located in the upper midbrain area close to the inferior colliculus (landmark structures such as red nucleus
1095
and mammillary body further support that R1 includes the dorsal raphe). Dorsal raphe nuclei, however, were difficult to distinguish from other nuclei in the vicinity because the dorsal raphe is intermingled with the red nucleus, which has relatively higher activity than the dorsal raphe. The second group (R2) was located ventral to the inferior colluculus in the upper portion of the pons. The R2 group, also known as the nucleus reticularis centralis superior, consistently shows the highest level of glucose metabolism among the raphe nuclei groups. The third group (R3), located at the central and posterior aspect of the pons, appeared to correspond to the nucleus raphe pontis. As the smallest among the four raphe nuclei groups, R3 was not easily discernable in two out of seven subjects. The fourth group (R4) was found at the most posterior and inferior aspects of the pons and also at the junction point to the superior aspect of medulla. This group appeared to include the nucleus raphe magnus, nucleus raphe obscurus, and nucleus raphe pallidus. The level of glucose metabolism in the raphe nuclei was found to be much lower than in most of the cerebral cortex, except in the cerebellar cortex, where activities are somewhat higher than those in the brainstem but generally do not reach those of the cerebral cortices (Fig. 1e). Within the vicinity of the raphe nuclei, there are several well-known landmark structures that verified our affirmation of the clusters as the raphe nuclei: the inferior colliculus, red nucleus, and mammillary body. These landmark structures showed relatively higher levels of metabolism in the areas of the brainstem and always with certainty thereby provided us an additional information on the relative locations of the raphe nuclei, especially the high-resolution images obtained from 7.0 T-MRI. These observations were consistently reproducible for all the subjects as shown in Fig. 2a. Although the exact location of each raphe nucleus differed slightly between individuals, the general distribution pattern was very consistent. However, the individual differences resulted in the blurring of the image when the intersubject data were averaged. Nevertheless, the typical pattern of glucose metabolism in the brainstem area was well characterized even after averaging of the images. The spatially normalized PET images of those activity groups were also well matched among the subjects (see Fig. 2b). Discussions Using the new PET–MRI fusion imaging system, which consists of high-end PET (HRRT) and ultra-high-field MRI (7.0 T), glucose metabolism of several clusters of the individual brainstem raphe nuclei were clearly imaged and identified. Although MRI alone does not visualize the raphe nuclei because of low contrast, 7.0 T-MRI of the brainstem areas provided excellent images of several landmark anatomical structures from which it was possible to deduce the brainstem raphe nuclei groups. The results clearly identified the four groups of raphe nuclei in the brainstem area as described by Naidich et al. (2009). One of the unique features of the present study is that because the raphe nuclei are all located right in the midline of the brainstem, they are easy to locate, especially where the corresponding anatomy is clear due to a number of unmistakable landmark structures. Although some previous rat or monkey studies (Fançois et al., 2010; Lyons et al., 1996; Skelin et al., 2008) using high-resolution 14 autoradiography with C-deoxyglucose reported glucose metabolism of the raphe nuclei, to our knowledge, this is the first report of clear visualization of the four groups of human raphe nuclei in vivo using FDG PET. Most of animal studies investigated the dorsal raphe and medial raphe only, except Skelin et al. (2008). Unlike our findings, they found that rats showed highest metabolism in the raphe pontine among raphe nuclei. It could be due to the difference between human and animal or due to the anesthesia applied to rats for experiment. In addition, Heiss et al. (2004) imaged subcortical nuclei of the human brain using FDG-PET with HRRT-PET, but the individual raphe nuclei were not specifically found in that study. Their major interest was to
1096
Y.-D. Son et al. / NeuroImage 59 (2012) 1094–1097
Fig. 1. A representative fusion image set of the brainstem area from a single subject obtained by positron emission tomography–magnetic resonance imaging (PET–MRI). (a) A medial sagittal T2*-weighted image of the brainstem obtained by 7.0 T-MRI. This high-resolution image depicts various landmark structures in the midbrain and brainstem areas such 18 18 as the mammillary body (MB), thalamus, and red nucleus (RN). (b) [ F]Fluorodeoxyglucose ( F-FDG)-PET image obtained by high-resolution research tomograph (HRRT-PET) of a slice corresponding to (a). Many well-known localized activities are clearly visible, such as the MB, RN, and other structures, together with activity cluster groups that are believed to be the raphe nuclei. (c) Fusion image of (a) and (b). From this fusion image, together with histological information provided by Naidich et al. (2009) and ultra-high resolution image of 7.0 T-MRI, it was possible to identify four raphe nucleus groups (arrowheads): dorsal, superior central, pontine, and medullary. (d) Raphe nuclei previously identified in 18 the brain atlas of Naidich et al. (2009) and our new PET–MRI fusion image of the brainstem area obtained by F-FDG. Note the markedly clear and distinct correlation between the locations and distributions of the raphe nuclei shown in the reference of Naidich et al. and those of the raphe nuclei seen in the PET–MRI fusion image. (e) Relative glucose metabolic activities of each raphe nucleus as well as the other midbrain nuclei measured in seven healthy subjects. R2 shows the highest glucose metabolic activity among the raphe nuclei. The other non–raphe nucleus groups, compared with the raphe nucleus groups, the highest level of activities were found in TH followed by IC, RN, and MB. TH: thalamus, MB: mammillary body, IC: inferior colliculus, RN: red nucleus, R1–R4: raphe nuclei, B1: nucleus raphes pallidus, B2: nucleus raphes obscurus, B3: nucleus raphes magnus, B5: nucleus raphes pontis, B6, B8: nucleus centralis superior, B7: dorsal nucleus raphe.
Fig. 2. Individual PET–MRI (HRRT-PET and 7.0 T-MRI) fusion images of seven subjects and an averaged image. (a) Seven intersubject images at the middle sagittal section of the brainstem. All subjects showed similar distributions of the raphe nuclei as well as the surrounding areas. (b) Averaged PET image of the seven subjects spatially normalized to the Montreal Neurological Institute (MNI) template using the SPM8. Although spatial averaging resulted in a blurred image, the overall raphe distribution confirmed that the raphe nuclei distribution is consistent among the subjects.
Y.-D. Son et al. / NeuroImage 59 (2012) 1094–1097
find some of the small nuclei such as the substantia nigra, red nucleus, inferior colliculus, and etc., all of which are substantially larger structures than the raphe nuclei. We further extended their study to focus on the raphe nuclei by taking advantage of high-resolution capability of HRRT PET and 7.0 T-MRI fusion system. It is worth noting that HRRT-PET is a brain-dedicated research PET, which is designed to provide images of the highest sensitivity and resolution, down to 2.5 mm full width at half maximum iso-voxel resolution due to the smallest head-only gantry diameter, which effectively leads to the highest solid angle and, therefore, the highest sensitivity. This high resolution in turn reduces much error due to partial volume averaging effect. In addition, since HRRT-PET covers the whole brainstem area, it is ideally suitable for brainstem imaging. For the SUVR calculation, the cerebellum was chosen as the reference tissue. In order to minimize the variation due to different buildup and wash-out time of the tissue, FDG injection and scanning times were made uniform for all subjects. Glucose uptake of neurons represents neuronal activity, either directly or indirectly. Although the direct correlation between neurotransmitter release and glucose consumption has not yet been clearly established, changes in glucose consumption reflect the neuronal activities, including neurotransmitter and receptor dynamics. In many cases, instead of direct or indirect measurement of the neuroreceptors or neurotransporters, measurement of metabolic activities of the well-separated individual nuclei could be sufficient for the understanding of the role of nuclei, such as serotonergic neurons in the brainstem area. With regard to future work, a multi-tracer study combining FDG imaging and serotonergic receptor or transporter imaging of the same subject would further confirm that the nuclei we observed are indeed the serotonergic nuclei. For example, 11 recent imaging studies with C-DASB showed increased activities in the raphe nuclei (Ichise et al., 2003; Kim et al., 2006), although resolution was too poor to specifically identify any of the individual nuclei. Correla11 18 tion studies between C-DASB and F-FDG will further enable us to assuredly verify that the nuclei with high glucose uptake in the brainstem area are indeed the raphe nuclei. Additionally, quantitative 11 imaging study of FDG and serotonin receptors, such as C-WAY100635, would also provide further clues about the relationship between the neuronal activities of neurotransmitter and receptor and the glucose metabolism.
Conclusions FDG-PET images of the brainstem were obtained by using a new HRRT-PET and 7.0 T-MRI fusion imaging protocol. This highsensitivity and high-resolution molecular imaging provided, for the first time, clearly identifiable in vivo glucose metabolic activities of the raphe nuclei in the human brainstem. This observation was possible due to the improved resolution and sensitivity of HRRT-PET, which in turn reduced the partial-volume effect incorporated with detailed anatomical images of 7.0 T-MRI, effectively guiding us to accurately locate the raphe nuclei in the brain, in correlation with several landmark structures within the brainstem.
1097
Conflict of interest statement The authors declare no conflict of interest. Acknowledgments This work was supported by the National Research Foundation (NRF), the Ministry of Education, Science and Technology (2008–2004159). Appendix A. Supplementary data Supplementary data to this article can be found online at doi:10. 1016/j.neuroimage.2011.09.036. References Cho, Z.H., 2009. 7.0 Tesla MRI Brain Atlas: In Vivo Atlas with Cryomacrotome Correlation. Springer, New York, NY. Cho, Z.H., Chan, J.K., Eriksson, L., 1976. Circular ring transverse axial positron camera for 3-dimensional reconstruction of radionuclides distribution. IEEE Trans. Nucl. Sci. 23, 613–622. Cho, Z.H., Son, Y.D., Kim, H.K., Kim, K.N., Oh, S.H., Han, J.Y., Hong, I.K., Kim, Y.B., 2008. A fusion PET–MRI system with a high-resolution research tomograph-PET and ultrahigh field 7.0T-MRI for the molecular-genetic imaging of the brain. Proteomics 8, 1302–1323. Cho, Z.H., Son, Y.D., Kim, H.K., Kim, S.T., Lee, S.Y., Chi, J.G., Park, C.W., Kim, Y.B., 2010. Substructural hippocampal glucose metabolism observed on PET/MRI. J. Nucl. Med. 51, 1545–1548. Cho, Z.H., Son, Y.D., Kim, H.K., Kim, N.B., Choi, E.J., Lee, S.Y., Chi, J.G., Park, C.W., Kim, Y.B., Ogawa, S., 2011. Observation of glucose metabolism in the thalamic nuclei by fusion PET/MRI. J. Nucl. Med. 52, 401–404. Fançois, J., Koning, E., Ferrandon, A., Sandner, G., Nehlig, A., 2010. Metabolic activity in the brain of juvenile and adult rats with a neonatal ventral hippocampal lesion. Hippocampus 20, 841–851. Heiss, W.D., Habedank, B., Klein, J.C., Herholz, K., Wienhard, K., Lenox, M., Nutt, R., 2004. Metabolic rates in small brain nuclei determined by high-resolution PET. J. Nucl. Med. 45, 1811–1815. Hong, I.K., Chung, S.T., Kim, H.K., Kim, Y.B., Son, Y.D., Cho, Z.H., 2007. Ultra fast symmetry and SIMD-based projection-backprojection (SSP) algorithm for 3-D PET image reconstruction. IEEE Trans. Med. Imaging 26, 789–803. Ichise, M., Liow, J.S., Lu, J.Q., Takano, A., Model, K., Toyama, H., Suhara, T., Suzuki, K., Innis, R.B., Carson, R.E., 2003. Linearized reference tissue parametric imaging methods: application to [11C]DASB positron emission tomography studies of the serotonin transporter in human brain. J. Cereb. Blood Flow Metab. 23, 1096–1112. Kim, J.S., Ichise, M., Sangare, J., Innis, R.B., 2006. PET imaging of serotonin transporters with [11C]DASB: test-retest reproducibility using a multilinear reference tissue parametric imaging method. J. Nucl. Med. 47, 208–214. Lyons, D., Friedman, D.P., Nader, M.A., Porrino, L.J., 1996. Cocaine alters cerebral metabolism within the ventral striatum and limbic cortex of monkeys. J. Neurosci. 16, 1230–1238. Naidich, T.P., Duvernoy, H.M., Delman, B.N., Sorensen, A.G., Kollias, S.S., Haacke, E.M., 2009. Duvernoy's atlas of the human brain stem and cerebellum: high-field MRI, surface anatomy, internal structure, vascularization and 3 D sectional anatomy. Springer, Vienna. Phelps, M.E., Hoffman, E.J., Huang, S.C., Kuhl, D.E., 1978. ECAT: a new computerized tomographic imaging system for positron-emitting radiopharmaceuticals. J. Nucl. Med. 19, 635–647. Schmand, M., Casey, M., Wienhard, K., Eriksson, L., Jones, W., Lenox, M., Young, J., Baker, K., Miller, S., Reed, J., 1999. HRRT a new high resolution LSO-PET research tomograph. J. Nucl. Med. 40, 76. Skelin, I., Sato, H., Diksic, M., 2008. Olfactory bulbectomy reduces cerebral glucose utilization: 2-[14C]deoxyglucose autoradiographic study. Brain Res. Bull. 76, 485–492. Wienhard, K., Schmand, M., Casey, M.E., Baker, K., Bao, J., Eriksson, L., Jones, W.F., Knoess, C., Lenox, M., Lercher, M., 2002. The ECAT HRRT: performance and first clinical application of the new high resolution research tomograph. IEEE Trans. Nucl. Sci. 49, 104–110.