- Email: [email protected]
Cerebral glucose metabolism in corticobasal degeneration comparison with progressive supranuclear palsy using statistical mapping analysis
Neuroscience Letters 383 (2005) 22–27
Cerebral glucose metabolism in corticobasal degeneration comparison with progressive supranuclear palsy using statistical mapping analysis Rahyeong Juh a , Chi-Un Pae b,∗ , Tae-Suk Kim b , Chang-Uk Lee b , Boyoung Choe a , Taesuk Suh a,∗ a
Department of Biomedical Engineering, Kangnam St. Mary’s Hospital, The Catholic University of Korea College of Medicine, 505 Banpo-Dong, Seocho-Gu, Seoul 137-701, South Korea b Department of Neuropsychiatry, Kangnam St. Mary’s Hospital, The Catholic University of Korea College of Medicine, 505 Banpo-Dong, Seocho-Gu, Seoul 137-701, South Korea Received 19 November 2004; received in revised form 18 March 2005; accepted 21 March 2005
Abstract This study measured the cerebral glucose metabolism in patients suffering from corticobasal degeneration (CBD) and progressive supranuclear palsy (PSP). The aim was to determine if there is a different metabolic pattern using 18 F-labeled 2-deoxyglucose (18 F-FDG) positron emission tomography (PET). The regional cerebral glucose metabolism was examined in 8 patients diagnosed clinically with CBD (mean age 69.6 ± 7.8 years; male/female: 5/3), 8 patients with probable PSP (mean age 67.8 ± 4.5 years; male/female: 4/4) and 22 healthy controls. The regional cerebral glucose metabolism between the three groups was compared using statistical parametric mapping (SPM) with a voxel-byvoxel approach (p < 0.001, 200-voxel level). Compared with the normal controls, asymmetry in the regional glucose metabolism was observed in the parietal, frontal and cingulate in the CBD patients. In the PSP patients, the glucose metabolism was lower in the orbitofrontal, middle frontal, cingulate, thalamus and mid-brain than their age matched normal controls. A comparison of the two patient groups demonstrated relative hypometabolism in the thalamus, the mid-brain in the PSP patients and the parietal lobe in CBD patients. These results suggest that when making a differential diagnosis of CBD and PSP, voxel-based analysis of the 18 F-FDG PET images using a SPM might be a useful tool in clinical examinations. © 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Corticobasal degeneration; Progressive supranuclear palsy; Cerebral glucose metabolism; Voxel-based analysis
Corticobasal degeneration (CBD) is a slowly progressive movement disorder with an adult onset. It is characterized by cortical and basal ganglionic degeneration. Both CBD and progressive supranuclear palsy (PSP) are closely related parkinsonian disorders and have many common clinical features such as a poor levodopa response, eye movement abnormalities, cognitive impairments, apraxia, pyramidal signs, and dystonia. These similarities often make it difficult to differentiation them clinically, particularly in their early course [4,9]. The clinical characteristics of PSP and CBD are distinct. However, atypical cases with overlapping clinical and ∗
Corresponding author. Tel.: +82 2 590 2731; fax: +82 536 8744. E-mail addresses: [email protected] (C.-U. Pae), [email protected] (T. Suh). 0304-3940/$ – see front matter © 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2005.03.057
pathologic features have been reported. Moreover, these two diseases are difficult to discriminate based on clinical grounds only [2,12,16]. Neuroimaging methods, such as magnetic resonance imaging (MRI), single photon emission computed tomography (SPECT), and positron emission tomography (PET) might improve the diagnostic accuracy and help distinguish CBD from PSP. However, often, there are little or no characteristic findings in the early stages [6,11]. 18 F-labeled 2-deoxyglucose (18 F-FDG) PET has previously been used to estimate the regional glucose metabolism in CBD and PSP cases [3]. In this study, the 18 F-FDG PET measurements of the cerebral glucose metabolism in CBD and PSP patients were compared with those of 22 age-matched normal controls. In an attempt to characterize the distribution of the glucose metabolism in the CBD patients, the data obtained
R. Juh et al. / Neuroscience Letters 383 (2005) 22–27
23
van [8]. The unified Parkinson’s Disease Rating Scale (UPDRS) part III and the Hoehn and Yahr stage were evaluated while the patients were on their current dopaminergic medication. Twenty-two age matched healthy controls (mean age 67.8 ± 14.4 years; male/female: 9/13) were also examined. The healthy controls showed no clinical or neurological evidence of any cognitive deficits and were not taking any chronic medication. All the control subjects and patients provided informed consent, which was approved by the local ethics committee. All the patients and normal control subjects underwent 18 F-FDG PET. Table 1 summarizes the patients’ characteristics. None of the patients had their diagnosis proven by pathology. The 18 F-FDG PET scans was performed using a Biograph and ECAT HR plus (Siemens Medical System, Inc., Hoffman Estate, IL). A sequence of three 10-min frames was acquired 40 min after injecting the 214.6–444 MBq 18 F-FDG, which was combined into a single frame afterwards. The images were reconstructed by ordered subset expectation maximization (OSEM) using 16 subsets and 6 iteration reconstruction algorithms (Shepp filter, cut off frequency at 0.5 cycles per projection element). The dimensions of the reconstructed PET images were a 128 × 128 image matrix with a pixel size of 1.72 mm × 1.72 mm and an inter-slice distance of 2.43 mm (no gap space). Spatial preprocessing and statistical analysis were performed using the SPM implemented in Matlab 6.1 (The MathWorks, Inc., Natick, MA). All the reconstructed 18 F-FDG PET images were spatially normalized into Montreal Neurological Institute (MNI, McGill University, Montreal, Que., Canada) standard templates by an affine transformation (12 parameters for rigid transformations) and a non-linear transformation. The image sets were smoothed with an isotropic
from these patients and healthy subjects were compared using a voxel-by-voxel approach with statistical parametric mapping (SPM, Wellcome Department of Cognitive Neurology, Institute of Neurology, London, UK) [1,5]. The aim was to determine the difference in the glucose metabolism between the PSP and CBD patients using 18 F-FDG PET imaging. Eight CBD patients (mean age 69.6 ± 7.8 years; male/female: 5/3) and 8 PSP patients (mean age 67.8 ± 4.5 years; male/female: 4/4) were examined. The patients were admitted for a neurological work up. Eight consecutive patients were clinically diagnosed with CBD by neurologists who were experienced in movement disorders. The diagnosis included the clinical symptoms and the electrophysiological and magnetic stimulation studies. The criteria of the CBD Multicenter Case Control Study (CBD MCCS) were used to select those patients to be admitted for an examination of their cognitive impairment. The following criteria were used to diagnose CBD: slowly progressive asymmetric akinetic-rigid syndrome with one or more of the following signs of cortical impairment: ideomotor apraxia, myoclonus, cortical sensory loss, alien limb phenomenon, no beneficial effect of levodopa and no evidence of other diseases that could explain the symptoms. The patient’s left side was more affected than the right side. It should be noted that all those in the cortico-basal degeneration cohort had left-sided symptoms, which means that the asymmetry in the neuronal physiology can be interpreted unambiguously in relation to the unilateral pathology. The eight CBD patients also fulfilled the modified clinical criteria proposed by Lang et al. [7]. All the PSP patients presented with early-onset postural instability, gaze palsy, axial rigidity, bradykinesia and no notable response to dopaminergic drugs. They all fulfilled the clinical diagnostic criteria for probable PSP proposed by Lit-
Table 1 Clinical data from patients with corticobasal ganglia (CBD) and progressive supranuclear palsy (PSP) Patient Sex Age Most Limb affected rigidity side CBD PD1 PD2 PD3 PD4 PD5 PD6 PD7 PD8
M F M F M M M F
62 81 74 73 63 72 74 58
PSP PD1 PD2 PD3 PD4 PD5 PD6 PD7 PD8
F F M M F F M M
64 74 62 68 67 64 74 69
L L L L L L L L
Axial rigidity
Limb Limb Axial Postural SGP Apraxia Cortical Alien akinesia dystonia dystonia instability sensory limb loss
+ + + + + + + + + + + + + + + +
+ +
+ + + + + + + +
+ + + + + + + +
+ +
SGP: supranuclear gaze palsy; +: present; N/A: not available.
+ + +
+ + + +
+
+ + + + + + + + + + +
+
+ + + + + + +
+ + + + + + +
MRI/CT scan
Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy N/A Mild diffuse cortical atrophy Mid brain atrophy Mild diffuse cortical atrophy Mid brain atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mid brain atrophy
24
R. Juh et al. / Neuroscience Letters 383 (2005) 22–27
Gaussian filter (12 mm full-width at half-maximum) and the individual global counts were normalized by proportional scaling to a mean value of 50 ml/dl/min. This arbitrary value was derived from the mean cerebral blood flow but was used conventionally in this analysis. The images of the two patient groups were compared with those of the healthy group at every pixel, using a two-sample t test based on two contrasts to detect any regional decrease in metabolism (CBD < Normal, PSP < Normal, CBD < PSP and CBD > PSP). At a variable voxel height threshold (p < 0.001, 200-voxel level), clusters consisting of a minimum of 200 contiguous voxels were considered to be significantly different. The Talairach brain coordinates were estimated from a non-linear transformation from MNI space to Talairach space (Talairach Daemon Client, Version 1.1, Research Imaging Center, University of Texas Health Science Center at San Antonio) [5]. The differences between the CBD, PSP and the agematched healthy control groups were examined using an extent threshold level of 200 voxels with a p value <0.001 to
illustrate the group differences in the statistical voxel-based analysis, as well as for illustrating the result of the registration between the CBD, PSP patients and the normal controls (Table 2, Figs. 1 and 2). In the CBD patients group analysis, the glucose metabolism was significantly lower in the parietal (Z = 6.33), middle frontal (Z = 5.37, Brodmann 8), cingulate gyrus (Z = 5.15), precentral gyrus (Z = 4.34, Brodmann 9) and thalamus (Z = 4.46) than in the age-matched controls. In contrast, the PSP patients group analysis showed greater hypometabolism in the thalamus (Z = 8.22), middle frontal (Z = 6.52), medial frontal (Z = 6.23), anterior cingulate (Z = 5.82) and mid-brain (Z = 5.77). However, compared with the normal group, there was a significant hypermetabolic pattern observed in the both occipital areas. The voxel-based comparison with the SPM between the CBD and PSP group using a 200-voxel level extent threshold with p < 0.001, as shown in Fig. 3. It showed a decreasing pattern in the glass window. The PSP patient group showed greater hypometabolism in the thalamus (Z = 6.51),
Table 2 Comparison of the CBD, PSP and normal subject groups showing the locations where the glucose metabolism comparison results are lower in one group than another Voxel level
Voxel p(unc)
Z
Talairach coordinates (mm) x
y
z
Region
CBD < Normal
6.33 5.37 5.15 4.46 4.34 3.68
0.000 0.000 0.000 0.000 0.000 0.000
46 36 10 10 40 46
−20 24 −48 −28 24 28
48 38 28 6 36 −6
Right cerebrum, parietal lobe, postcentral gyrus Right cerebrum, frontal lobe, middle frontal gyrus, Brodmann area 8 Right cerebrum, limbic lobe, cingulate gyrus Right cerebrum, sub-lobar, thalamus, pulvinar Right cerebrum, frontal lobe, precentral gyrus, Brodmann area 9 Right cerebrum, frontal lobe, inferior frontal gyrus
PSP < Normal
8.22 6.63 6.52 6.23 5.82 5.77 5.60 5.05 3.99 3.30
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
−10 −14 −32 4 −2 2 −4 2 −22 −52
−21 10 18 56 38 −22 −22 26 2 −58
8 5 46 2 22 −4 −2 26 0 38
Left cerebrum, thalamus, gray matter, medial dorsal nucleus Left cerebrum, extra-nuclear Left cerebrum, frontal lobe, middle frontal gyrus Right cerebrum, medial frontal gyrus, Brodmann area 10 Left cerebrum, limbic lobe, anterior cingulate Right brainstem, midbrain Left brainstem, midbrain Right cerebrum, limbic lobe, cingulate gyrus Left cerebrum, lentiform nucleus, putamen Left cerebrum, parietal lobe, inferior parietal lobule, Brodmann area 40
PSP < CBD
6.51 5.46 4.31 3.70 2.84 2.38 1.97
0.000 0.000 0.002 0.003 0.004 0.004 0.011
−10 −52 −12 −44 −30 −16 −38
−18 −2 −22 −8 14 14 −12
2 −32 −2 48 −6 −2 0
Left cerebrum, thalamus, mammillary body Left cerebrum, frontal lobe, precentral gyrus, Brodmann area 6 Left brainstem, midbrain Left cerebrum, frontal lobe, precentral gyrus Left cerebrum, sub-lobar, insula Left cerebrum, sub-lobar, lentiform nucleus, putamen Left cerebrum, extra-nuclear
CBD < PSP
4.97 4.92 4.45 3.96 3.61 3.53 3.14 2.58 2.36
0.001 0.001 0.002 0.002 0.007 0.011 0.011 0.013 0.013
34 40 38 −24 46 30 −32 2 −26
−34 −34 −26 0 −26 −46 56 −54 −74
−8 −12 50 −26 58 58 40 0 −26
Right cerebrum, limbic lobe, sub-gyral Right cerebrum, limbic lobe, parahippocampal gyrus Right cerebrum, parietal lobe, postcentral gyrus, Brodmann area 3 Left cerebrum, limbic lobe, uncus Right cerebrum, parietal lobe, postcentral gyrus, Brodmann area 1 Right cerebrum, parietal lobe, superior parietal lobule, Brodmann area 7 Left cerebrum, parietal lobe, sub-gyral Right cerebellum Left cerebellum, posterior lobe, uvula
This table shows the location and peaks, as well as the Z value of the significant decreases in the metabolism in the CBD and PSP group compared with the normal control groups. There is significant hypometabolic region in the cerebral metabolism in the CBD patients compared with the PSP groups. The table lists the Z scores, the x, y, z, coordinates, the spatial extent of the cluster in 200 voxels, p < 0.001 uncorrected, and the Talairach coordinates.
R. Juh et al. / Neuroscience Letters 383 (2005) 22–27
25
Fig. 1. Specific voxels with significantly lower metabolism in the CBD group than in the age-matched control group at a threshold of a two-sample t test p < 0.001, 200-voxel level. The group comparison of the SPM result on the glass window between the CBD group and the age-matched control group. The glucose metabolism was significantly lower in the parietal lobe, middle frontal gyrus and cingulate gyrus than in the age matched control group. This image shows that the SPM, which was overlaid on the template MR supported MNI image of the region, can be used as a discriminating variable factor for distinguishing the CBD group from the control group.
precentral gyrus (Z = 5.46), mid-brain (Z = 4.31) and putamen (Z = 2.38) regions than the CBD patients. However, there was a decreasing pattern in the limbic lobe (Z = 4.97), parietal lobe (Z = 4.45, Brodmann 7), cerebellum (Z = 2.58) and uvula (Z = 2.36) in the CBD patients group (Table 2). The results of this voxel-based analysis method showed that the mid-brain, thalamus, limbic lobe and parietal lobe could be used as discriminating variable factors for distinguishing between CBD and PSP.
The aim of this study was to determine the major metabolic involvement of patients with probable CBD in comparison with both healthy subjects and patients with PSP using voxel-based analysis. Compared with the normal subjects, the CBD patients showed significantly lower cerebral glucose metabolism in the lateral hemisphere of the clinically most affected side in the superior parietal, medial frontal, and cingulate. In addition, a reduction in metabolism was also noted in the sensory motor system. In contrast, in the PSP patients,
Fig. 2. Specific voxels with significantly lower metabolism in the PSP group than in the age matched control group at a threshold of a two-sample t test p < 0.001, 200-voxel level. The statistically significant decrease in metabolism in the areas of the mid-brain, thalamus and basal ganglia is the discrimination factor.
26
R. Juh et al. / Neuroscience Letters 383 (2005) 22–27
Fig. 3. SPMs showing the spatial distribution of the significant region within the patients group. The group comparison of the SPM result between the CBD and PSP groups was displayed on the glass window. The figures are displayed in sagittal, coronal and transverse projections into the stereotactic space of MNI template. (a) Specific voxels with a significantly lower metabolism in the CBD group than in the PSP group at a threshold of a two-sample t test p < 0.001, 200-voxel level. The distribution of the decrease in the relative glucose uptake at rest is shown when the metabolic pattern of the CBD group was compared with that of the PSP group using mapping analysis. The hypometabolism region in the parietal lobule in the CBD group compared with that in the PSP patients group. (b) Specific voxels with significantly lower metabolism in the PSP group than in the CBD group. Hypometabolism region in the thalamus, the mid brain in the PSP group compared with that in the CBD patients group.
the glucose metabolism was lower in the orbitofrontal, middle frontal, thalamus, anterior cingulate and mid-brain than in the age-matched healthy controls. A comparison of the two patient groups demonstrated relative hypometabolism in the thalamus, the mid-brain in the PSP patients and the parietal lobe in the CBD patients. PSP represents at least 5% of the Parkinsonian syndrome cases. However, this percentage is probably an underestimate due to the difficulties associated with diagnosing this syndrome [11–13,17]. A diminished frontal cortex and anterior
cingulate as well as a lower striatal metabolism are consistently observed in PSP patients. However, this metabolic pattern is not specific to PSP patients, and this syndrome cannot be easily differentiated from CBD. A brain MRI of the PSP patients indicates that atrophy of the midbrain is more frequently associated with a third ventricle dilatation and T2periaqueductal hyperintensities [13,14]. This has been recognized in CBD patients with increased frequency. Both CT and MRI indicate cerebral atrophy contralateral to the clinically affected side. A PET study revealed an asymmetric reduction
R. Juh et al. / Neuroscience Letters 383 (2005) 22–27
in the cortical FDG metabolism and striatal tracer uptake [15]. Anatomic standardization and normalization for individual variations in the glucose metabolism could be suitable for global normalization because the regional cerebral glucose metabolism in the CBD patients was significantly lower in all the brain regions compared with the normal subjects. The relative glucose metabolism, which was normalized to the global cerebral cortical activity, was also compared thereby removing the inter-subject differences in metabolism. Recent voxel-based studies where this differentiation could be validly measured, showed this index to be particularly useful for differentiating CBD and PSP patients from the controls, as well as for distinguishing between CBD from PSP patients. The significant reduction in metabolism in the one side sensory motor system and the ipsilateral thalamus regions of the parietal cortex observed in the CBD patients was different from the PSP patients. It is believed that hypometabolism of this region can be responsible for the motor apraxia, which is one of the major symptoms of CBD patients. Moreover, asymmetrical hypometabolism in the thalamus was observed in the ipsilateral to the parietal cortex. It is difficult to explain why the hypometabolism observed in the thalamus might have been caused by the primary neural damage and degeneration [10]. The results of the neurological examinations showed that the CBD patients were more impaired. The deficit characteristics of CBD, such as cortical impairment apraxia, cortical sensory loss and alien limb phenomena, were attributed to the parietal dysfunction. On the other hand, the neurological features of the PSP patients included an impaired ocular motility, supranuclear gaze palsy, axial rigidity and limb akinesia [7,11]. The limitations of this study are as follows: there were an insufficient number of patients and a limited number of control subjects due to the limited availability of PET. In addition, there were no pathological correlations or post-mortem evaluations available. This means that a correlation between the voxel-based SPM analysis and ROI-based approaches could not be made. Because of the different pathophysiology in the cortical and subcortical brain structures, an assessment of the regional cerebral glucose metabolism using the SPM might be useful when making a differential diagnosis. These results clearly show that the most striking features of CBD are the metabolic deficits in the primary sensory motor system and parietal lobe. The differences in the regional glucose metabolism between CBD and PSP can be attributed to the different clinical features of these disease entities. This study evaluated the feasibility of using a voxel-based analysis for making a differential diagnosis of CBD and PSP via statistical mapping analysis. The comparison of the two patient groups demonstrated relative hypometabolism in the thalamus, the mid-brain in PSP patients and the parietal lobe in the CBD patients. This suggests that the parietal lobe and sensory motor system could be used as discriminating factors for distinguishing between CBD and PSP. Therefore, measuring the glucose metabolism by 18 F-FDG PET followed by voxel-by-voxel based SPM analysis and image registration
27
might be useful for distinguishing between CBD and PSP, and be helpful in establishing a differential diagnosis.
Acknowledgments This work was supported by Korea Institute of Science and Technology Evaluation and Planning (KISTEP) and Ministry of Science and Technology (MOST) of Korean government.
References [1] P.D. Acton, P.D. Mozley, H.F. Kung, Logistic discriminant parametric mapping: a novel method for the pixel-based differential diagnosis of Parkinson’s disease, Eur. J. Nucl. Med. 26 (1999) 1413–1423. [2] J. Ashburner, K.J. Friston, Voxel-based morphometry—the methods, Neuroimage 11 (2000) 805–821. [3] I. Coulier, J. Vries, K. Leenders, Is FDG-PET a useful tool in clinical practice for diagnosing corticobasal ganglionic degeneration? Mov. Disord. 18 (2003) 1175–1178. [4] D.W. Dickson, Neuropathologic differentiation of progressive supra nuclear palsy and cortical degeneration, J. Neurol. 246 (1999) II6–II15. [5] K.J. Friston, Commentary and opinion: II. Statistical parametric mapping: ontology and current issues, J. Cereb. Blood Flow Metab. 15 (1995) 361–370. [6] N. Hirono, K. Ishii, M. Sasaki, H. Kitagaki, M. Hashimoto, T. Imamura, S. Tanimukai, T. Hanihara, H. Kazui, E. Mori, Features of regional cerebral glucose metabolism abnormality in corticobasal degeneration, Dement. Geriatr. Cogn. Disord. 11 (2000) 139–146. [7] A.E. Lang, D.E. Riley, C. Bergeron, Corticobasal ganglionic degeneration, in: D.B. Calne (Ed.), Neurodegenerative Disease, Saunders, Philadelphia, 1994, pp. 877–894. [8] I. Litvan, Update on epidemiological aspects of progressive supranuclear palsy, Mov. Disord. 18 (2003) 43–50. [9] Y. Nagahama, H. Fukuyama, N. Turjanski, A. Kennedy, H. Yamauchi, Y. Ouchi, J. Kimura, D.J. Brooks, Cerebral glucose metabolism in corticobasal degeneration: comparison with progressive supra nuclear palsy and normal controls, Mov. Disord. 12 (1997) 691–696. [10] H. Nagasawa, T. Imamura, H. Nomura, M. Itoh, T. Ido, A case of corticobasal degeneration studied with positron emission tomography, Behav. Neurol. 6 (1993) 59–64. [11] B. Okuda, H. Tachibana, K. Keita, M. Takeda, M. Sugita, Comparison of brain perfusion in corticobasal degeneration and Alzheimer’s disease, Dement. Geriatr. Cogn. Disord. 12 (2001) 226–231. [12] W. Poewe, G. Wenning, The differential diagnosis of Parkinson’s disease, Eur. J. Neurol. 9 (2002) 23–30. [13] A. Schrag, Y. Ben-Shlomo, N.P. Quinn, Prevalence of progressive supranuclear palsy and multiple system atrophy: a cross-sectional study, Lancet 354 (1999) 1771–1775. [14] A. Schrag, C.D. Good, K. Miszkiel, H.R. Morris, C.J. Mathias, A.J. Lees, N.P. Quinn, Differentiation of atypical parkinsonian syndromes with routine MRI, Neurology 54 (2000) 697–702. [15] P. Soliveri, D. Monza, D. Paridi, Cognitive and magnetic imaging resonance imaging aspects of corticobasal degeneration and progressive supranuclear palsy, Neurology 53 (1999) 502–507. [16] M. Takeda, H. Tachibana, B. Okuda, K. Kawabata, M. Sugita, Electrophysical comparison between corticobasal degeneration and progressive supranuclear palsy, Clin. Neurol. Neurosurg. 100 (1998) 94–98. [17] S. Thobois, S. Guillouet, E. Broussolle, Contributions of PET and SPECT to the understanding of the pathophysiology of Parkinson’s disease, Neurophysiol. Clin. 31 (2001) 321–340.
Cerebral glucose metabolism in corticobasal degeneration comparison with progressive supranuclear palsy using statistical mapping analysis Rahyeong Juh a , Chi-Un Pae b,∗ , Tae-Suk Kim b , Chang-Uk Lee b , Boyoung Choe a , Taesuk Suh a,∗ a
Department of Biomedical Engineering, Kangnam St. Mary’s Hospital, The Catholic University of Korea College of Medicine, 505 Banpo-Dong, Seocho-Gu, Seoul 137-701, South Korea b Department of Neuropsychiatry, Kangnam St. Mary’s Hospital, The Catholic University of Korea College of Medicine, 505 Banpo-Dong, Seocho-Gu, Seoul 137-701, South Korea Received 19 November 2004; received in revised form 18 March 2005; accepted 21 March 2005
Abstract This study measured the cerebral glucose metabolism in patients suffering from corticobasal degeneration (CBD) and progressive supranuclear palsy (PSP). The aim was to determine if there is a different metabolic pattern using 18 F-labeled 2-deoxyglucose (18 F-FDG) positron emission tomography (PET). The regional cerebral glucose metabolism was examined in 8 patients diagnosed clinically with CBD (mean age 69.6 ± 7.8 years; male/female: 5/3), 8 patients with probable PSP (mean age 67.8 ± 4.5 years; male/female: 4/4) and 22 healthy controls. The regional cerebral glucose metabolism between the three groups was compared using statistical parametric mapping (SPM) with a voxel-byvoxel approach (p < 0.001, 200-voxel level). Compared with the normal controls, asymmetry in the regional glucose metabolism was observed in the parietal, frontal and cingulate in the CBD patients. In the PSP patients, the glucose metabolism was lower in the orbitofrontal, middle frontal, cingulate, thalamus and mid-brain than their age matched normal controls. A comparison of the two patient groups demonstrated relative hypometabolism in the thalamus, the mid-brain in the PSP patients and the parietal lobe in CBD patients. These results suggest that when making a differential diagnosis of CBD and PSP, voxel-based analysis of the 18 F-FDG PET images using a SPM might be a useful tool in clinical examinations. © 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Corticobasal degeneration; Progressive supranuclear palsy; Cerebral glucose metabolism; Voxel-based analysis
Corticobasal degeneration (CBD) is a slowly progressive movement disorder with an adult onset. It is characterized by cortical and basal ganglionic degeneration. Both CBD and progressive supranuclear palsy (PSP) are closely related parkinsonian disorders and have many common clinical features such as a poor levodopa response, eye movement abnormalities, cognitive impairments, apraxia, pyramidal signs, and dystonia. These similarities often make it difficult to differentiation them clinically, particularly in their early course [4,9]. The clinical characteristics of PSP and CBD are distinct. However, atypical cases with overlapping clinical and ∗
Corresponding author. Tel.: +82 2 590 2731; fax: +82 536 8744. E-mail addresses: [email protected] (C.-U. Pae), [email protected] (T. Suh). 0304-3940/$ – see front matter © 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2005.03.057
pathologic features have been reported. Moreover, these two diseases are difficult to discriminate based on clinical grounds only [2,12,16]. Neuroimaging methods, such as magnetic resonance imaging (MRI), single photon emission computed tomography (SPECT), and positron emission tomography (PET) might improve the diagnostic accuracy and help distinguish CBD from PSP. However, often, there are little or no characteristic findings in the early stages [6,11]. 18 F-labeled 2-deoxyglucose (18 F-FDG) PET has previously been used to estimate the regional glucose metabolism in CBD and PSP cases [3]. In this study, the 18 F-FDG PET measurements of the cerebral glucose metabolism in CBD and PSP patients were compared with those of 22 age-matched normal controls. In an attempt to characterize the distribution of the glucose metabolism in the CBD patients, the data obtained
R. Juh et al. / Neuroscience Letters 383 (2005) 22–27
23
van [8]. The unified Parkinson’s Disease Rating Scale (UPDRS) part III and the Hoehn and Yahr stage were evaluated while the patients were on their current dopaminergic medication. Twenty-two age matched healthy controls (mean age 67.8 ± 14.4 years; male/female: 9/13) were also examined. The healthy controls showed no clinical or neurological evidence of any cognitive deficits and were not taking any chronic medication. All the control subjects and patients provided informed consent, which was approved by the local ethics committee. All the patients and normal control subjects underwent 18 F-FDG PET. Table 1 summarizes the patients’ characteristics. None of the patients had their diagnosis proven by pathology. The 18 F-FDG PET scans was performed using a Biograph and ECAT HR plus (Siemens Medical System, Inc., Hoffman Estate, IL). A sequence of three 10-min frames was acquired 40 min after injecting the 214.6–444 MBq 18 F-FDG, which was combined into a single frame afterwards. The images were reconstructed by ordered subset expectation maximization (OSEM) using 16 subsets and 6 iteration reconstruction algorithms (Shepp filter, cut off frequency at 0.5 cycles per projection element). The dimensions of the reconstructed PET images were a 128 × 128 image matrix with a pixel size of 1.72 mm × 1.72 mm and an inter-slice distance of 2.43 mm (no gap space). Spatial preprocessing and statistical analysis were performed using the SPM implemented in Matlab 6.1 (The MathWorks, Inc., Natick, MA). All the reconstructed 18 F-FDG PET images were spatially normalized into Montreal Neurological Institute (MNI, McGill University, Montreal, Que., Canada) standard templates by an affine transformation (12 parameters for rigid transformations) and a non-linear transformation. The image sets were smoothed with an isotropic
from these patients and healthy subjects were compared using a voxel-by-voxel approach with statistical parametric mapping (SPM, Wellcome Department of Cognitive Neurology, Institute of Neurology, London, UK) [1,5]. The aim was to determine the difference in the glucose metabolism between the PSP and CBD patients using 18 F-FDG PET imaging. Eight CBD patients (mean age 69.6 ± 7.8 years; male/female: 5/3) and 8 PSP patients (mean age 67.8 ± 4.5 years; male/female: 4/4) were examined. The patients were admitted for a neurological work up. Eight consecutive patients were clinically diagnosed with CBD by neurologists who were experienced in movement disorders. The diagnosis included the clinical symptoms and the electrophysiological and magnetic stimulation studies. The criteria of the CBD Multicenter Case Control Study (CBD MCCS) were used to select those patients to be admitted for an examination of their cognitive impairment. The following criteria were used to diagnose CBD: slowly progressive asymmetric akinetic-rigid syndrome with one or more of the following signs of cortical impairment: ideomotor apraxia, myoclonus, cortical sensory loss, alien limb phenomenon, no beneficial effect of levodopa and no evidence of other diseases that could explain the symptoms. The patient’s left side was more affected than the right side. It should be noted that all those in the cortico-basal degeneration cohort had left-sided symptoms, which means that the asymmetry in the neuronal physiology can be interpreted unambiguously in relation to the unilateral pathology. The eight CBD patients also fulfilled the modified clinical criteria proposed by Lang et al. [7]. All the PSP patients presented with early-onset postural instability, gaze palsy, axial rigidity, bradykinesia and no notable response to dopaminergic drugs. They all fulfilled the clinical diagnostic criteria for probable PSP proposed by Lit-
Table 1 Clinical data from patients with corticobasal ganglia (CBD) and progressive supranuclear palsy (PSP) Patient Sex Age Most Limb affected rigidity side CBD PD1 PD2 PD3 PD4 PD5 PD6 PD7 PD8
M F M F M M M F
62 81 74 73 63 72 74 58
PSP PD1 PD2 PD3 PD4 PD5 PD6 PD7 PD8
F F M M F F M M
64 74 62 68 67 64 74 69
L L L L L L L L
Axial rigidity
Limb Limb Axial Postural SGP Apraxia Cortical Alien akinesia dystonia dystonia instability sensory limb loss
+ + + + + + + + + + + + + + + +
+ +
+ + + + + + + +
+ + + + + + + +
+ +
SGP: supranuclear gaze palsy; +: present; N/A: not available.
+ + +
+ + + +
+
+ + + + + + + + + + +
+
+ + + + + + +
+ + + + + + +
MRI/CT scan
Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy N/A Mild diffuse cortical atrophy Mid brain atrophy Mild diffuse cortical atrophy Mid brain atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mild diffuse cortical atrophy Mid brain atrophy
24
R. Juh et al. / Neuroscience Letters 383 (2005) 22–27
Gaussian filter (12 mm full-width at half-maximum) and the individual global counts were normalized by proportional scaling to a mean value of 50 ml/dl/min. This arbitrary value was derived from the mean cerebral blood flow but was used conventionally in this analysis. The images of the two patient groups were compared with those of the healthy group at every pixel, using a two-sample t test based on two contrasts to detect any regional decrease in metabolism (CBD < Normal, PSP < Normal, CBD < PSP and CBD > PSP). At a variable voxel height threshold (p < 0.001, 200-voxel level), clusters consisting of a minimum of 200 contiguous voxels were considered to be significantly different. The Talairach brain coordinates were estimated from a non-linear transformation from MNI space to Talairach space (Talairach Daemon Client, Version 1.1, Research Imaging Center, University of Texas Health Science Center at San Antonio) [5]. The differences between the CBD, PSP and the agematched healthy control groups were examined using an extent threshold level of 200 voxels with a p value <0.001 to
illustrate the group differences in the statistical voxel-based analysis, as well as for illustrating the result of the registration between the CBD, PSP patients and the normal controls (Table 2, Figs. 1 and 2). In the CBD patients group analysis, the glucose metabolism was significantly lower in the parietal (Z = 6.33), middle frontal (Z = 5.37, Brodmann 8), cingulate gyrus (Z = 5.15), precentral gyrus (Z = 4.34, Brodmann 9) and thalamus (Z = 4.46) than in the age-matched controls. In contrast, the PSP patients group analysis showed greater hypometabolism in the thalamus (Z = 8.22), middle frontal (Z = 6.52), medial frontal (Z = 6.23), anterior cingulate (Z = 5.82) and mid-brain (Z = 5.77). However, compared with the normal group, there was a significant hypermetabolic pattern observed in the both occipital areas. The voxel-based comparison with the SPM between the CBD and PSP group using a 200-voxel level extent threshold with p < 0.001, as shown in Fig. 3. It showed a decreasing pattern in the glass window. The PSP patient group showed greater hypometabolism in the thalamus (Z = 6.51),
Table 2 Comparison of the CBD, PSP and normal subject groups showing the locations where the glucose metabolism comparison results are lower in one group than another Voxel level
Voxel p(unc)
Z
Talairach coordinates (mm) x
y
z
Region
CBD < Normal
6.33 5.37 5.15 4.46 4.34 3.68
0.000 0.000 0.000 0.000 0.000 0.000
46 36 10 10 40 46
−20 24 −48 −28 24 28
48 38 28 6 36 −6
Right cerebrum, parietal lobe, postcentral gyrus Right cerebrum, frontal lobe, middle frontal gyrus, Brodmann area 8 Right cerebrum, limbic lobe, cingulate gyrus Right cerebrum, sub-lobar, thalamus, pulvinar Right cerebrum, frontal lobe, precentral gyrus, Brodmann area 9 Right cerebrum, frontal lobe, inferior frontal gyrus
PSP < Normal
8.22 6.63 6.52 6.23 5.82 5.77 5.60 5.05 3.99 3.30
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
−10 −14 −32 4 −2 2 −4 2 −22 −52
−21 10 18 56 38 −22 −22 26 2 −58
8 5 46 2 22 −4 −2 26 0 38
Left cerebrum, thalamus, gray matter, medial dorsal nucleus Left cerebrum, extra-nuclear Left cerebrum, frontal lobe, middle frontal gyrus Right cerebrum, medial frontal gyrus, Brodmann area 10 Left cerebrum, limbic lobe, anterior cingulate Right brainstem, midbrain Left brainstem, midbrain Right cerebrum, limbic lobe, cingulate gyrus Left cerebrum, lentiform nucleus, putamen Left cerebrum, parietal lobe, inferior parietal lobule, Brodmann area 40
PSP < CBD
6.51 5.46 4.31 3.70 2.84 2.38 1.97
0.000 0.000 0.002 0.003 0.004 0.004 0.011
−10 −52 −12 −44 −30 −16 −38
−18 −2 −22 −8 14 14 −12
2 −32 −2 48 −6 −2 0
Left cerebrum, thalamus, mammillary body Left cerebrum, frontal lobe, precentral gyrus, Brodmann area 6 Left brainstem, midbrain Left cerebrum, frontal lobe, precentral gyrus Left cerebrum, sub-lobar, insula Left cerebrum, sub-lobar, lentiform nucleus, putamen Left cerebrum, extra-nuclear
CBD < PSP
4.97 4.92 4.45 3.96 3.61 3.53 3.14 2.58 2.36
0.001 0.001 0.002 0.002 0.007 0.011 0.011 0.013 0.013
34 40 38 −24 46 30 −32 2 −26
−34 −34 −26 0 −26 −46 56 −54 −74
−8 −12 50 −26 58 58 40 0 −26
Right cerebrum, limbic lobe, sub-gyral Right cerebrum, limbic lobe, parahippocampal gyrus Right cerebrum, parietal lobe, postcentral gyrus, Brodmann area 3 Left cerebrum, limbic lobe, uncus Right cerebrum, parietal lobe, postcentral gyrus, Brodmann area 1 Right cerebrum, parietal lobe, superior parietal lobule, Brodmann area 7 Left cerebrum, parietal lobe, sub-gyral Right cerebellum Left cerebellum, posterior lobe, uvula
This table shows the location and peaks, as well as the Z value of the significant decreases in the metabolism in the CBD and PSP group compared with the normal control groups. There is significant hypometabolic region in the cerebral metabolism in the CBD patients compared with the PSP groups. The table lists the Z scores, the x, y, z, coordinates, the spatial extent of the cluster in 200 voxels, p < 0.001 uncorrected, and the Talairach coordinates.
R. Juh et al. / Neuroscience Letters 383 (2005) 22–27
25
Fig. 1. Specific voxels with significantly lower metabolism in the CBD group than in the age-matched control group at a threshold of a two-sample t test p < 0.001, 200-voxel level. The group comparison of the SPM result on the glass window between the CBD group and the age-matched control group. The glucose metabolism was significantly lower in the parietal lobe, middle frontal gyrus and cingulate gyrus than in the age matched control group. This image shows that the SPM, which was overlaid on the template MR supported MNI image of the region, can be used as a discriminating variable factor for distinguishing the CBD group from the control group.
precentral gyrus (Z = 5.46), mid-brain (Z = 4.31) and putamen (Z = 2.38) regions than the CBD patients. However, there was a decreasing pattern in the limbic lobe (Z = 4.97), parietal lobe (Z = 4.45, Brodmann 7), cerebellum (Z = 2.58) and uvula (Z = 2.36) in the CBD patients group (Table 2). The results of this voxel-based analysis method showed that the mid-brain, thalamus, limbic lobe and parietal lobe could be used as discriminating variable factors for distinguishing between CBD and PSP.
The aim of this study was to determine the major metabolic involvement of patients with probable CBD in comparison with both healthy subjects and patients with PSP using voxel-based analysis. Compared with the normal subjects, the CBD patients showed significantly lower cerebral glucose metabolism in the lateral hemisphere of the clinically most affected side in the superior parietal, medial frontal, and cingulate. In addition, a reduction in metabolism was also noted in the sensory motor system. In contrast, in the PSP patients,
Fig. 2. Specific voxels with significantly lower metabolism in the PSP group than in the age matched control group at a threshold of a two-sample t test p < 0.001, 200-voxel level. The statistically significant decrease in metabolism in the areas of the mid-brain, thalamus and basal ganglia is the discrimination factor.
26
R. Juh et al. / Neuroscience Letters 383 (2005) 22–27
Fig. 3. SPMs showing the spatial distribution of the significant region within the patients group. The group comparison of the SPM result between the CBD and PSP groups was displayed on the glass window. The figures are displayed in sagittal, coronal and transverse projections into the stereotactic space of MNI template. (a) Specific voxels with a significantly lower metabolism in the CBD group than in the PSP group at a threshold of a two-sample t test p < 0.001, 200-voxel level. The distribution of the decrease in the relative glucose uptake at rest is shown when the metabolic pattern of the CBD group was compared with that of the PSP group using mapping analysis. The hypometabolism region in the parietal lobule in the CBD group compared with that in the PSP patients group. (b) Specific voxels with significantly lower metabolism in the PSP group than in the CBD group. Hypometabolism region in the thalamus, the mid brain in the PSP group compared with that in the CBD patients group.
the glucose metabolism was lower in the orbitofrontal, middle frontal, thalamus, anterior cingulate and mid-brain than in the age-matched healthy controls. A comparison of the two patient groups demonstrated relative hypometabolism in the thalamus, the mid-brain in the PSP patients and the parietal lobe in the CBD patients. PSP represents at least 5% of the Parkinsonian syndrome cases. However, this percentage is probably an underestimate due to the difficulties associated with diagnosing this syndrome [11–13,17]. A diminished frontal cortex and anterior
cingulate as well as a lower striatal metabolism are consistently observed in PSP patients. However, this metabolic pattern is not specific to PSP patients, and this syndrome cannot be easily differentiated from CBD. A brain MRI of the PSP patients indicates that atrophy of the midbrain is more frequently associated with a third ventricle dilatation and T2periaqueductal hyperintensities [13,14]. This has been recognized in CBD patients with increased frequency. Both CT and MRI indicate cerebral atrophy contralateral to the clinically affected side. A PET study revealed an asymmetric reduction
R. Juh et al. / Neuroscience Letters 383 (2005) 22–27
in the cortical FDG metabolism and striatal tracer uptake [15]. Anatomic standardization and normalization for individual variations in the glucose metabolism could be suitable for global normalization because the regional cerebral glucose metabolism in the CBD patients was significantly lower in all the brain regions compared with the normal subjects. The relative glucose metabolism, which was normalized to the global cerebral cortical activity, was also compared thereby removing the inter-subject differences in metabolism. Recent voxel-based studies where this differentiation could be validly measured, showed this index to be particularly useful for differentiating CBD and PSP patients from the controls, as well as for distinguishing between CBD from PSP patients. The significant reduction in metabolism in the one side sensory motor system and the ipsilateral thalamus regions of the parietal cortex observed in the CBD patients was different from the PSP patients. It is believed that hypometabolism of this region can be responsible for the motor apraxia, which is one of the major symptoms of CBD patients. Moreover, asymmetrical hypometabolism in the thalamus was observed in the ipsilateral to the parietal cortex. It is difficult to explain why the hypometabolism observed in the thalamus might have been caused by the primary neural damage and degeneration [10]. The results of the neurological examinations showed that the CBD patients were more impaired. The deficit characteristics of CBD, such as cortical impairment apraxia, cortical sensory loss and alien limb phenomena, were attributed to the parietal dysfunction. On the other hand, the neurological features of the PSP patients included an impaired ocular motility, supranuclear gaze palsy, axial rigidity and limb akinesia [7,11]. The limitations of this study are as follows: there were an insufficient number of patients and a limited number of control subjects due to the limited availability of PET. In addition, there were no pathological correlations or post-mortem evaluations available. This means that a correlation between the voxel-based SPM analysis and ROI-based approaches could not be made. Because of the different pathophysiology in the cortical and subcortical brain structures, an assessment of the regional cerebral glucose metabolism using the SPM might be useful when making a differential diagnosis. These results clearly show that the most striking features of CBD are the metabolic deficits in the primary sensory motor system and parietal lobe. The differences in the regional glucose metabolism between CBD and PSP can be attributed to the different clinical features of these disease entities. This study evaluated the feasibility of using a voxel-based analysis for making a differential diagnosis of CBD and PSP via statistical mapping analysis. The comparison of the two patient groups demonstrated relative hypometabolism in the thalamus, the mid-brain in PSP patients and the parietal lobe in the CBD patients. This suggests that the parietal lobe and sensory motor system could be used as discriminating factors for distinguishing between CBD and PSP. Therefore, measuring the glucose metabolism by 18 F-FDG PET followed by voxel-by-voxel based SPM analysis and image registration
27
might be useful for distinguishing between CBD and PSP, and be helpful in establishing a differential diagnosis.
Acknowledgments This work was supported by Korea Institute of Science and Technology Evaluation and Planning (KISTEP) and Ministry of Science and Technology (MOST) of Korean government.
References [1] P.D. Acton, P.D. Mozley, H.F. Kung, Logistic discriminant parametric mapping: a novel method for the pixel-based differential diagnosis of Parkinson’s disease, Eur. J. Nucl. Med. 26 (1999) 1413–1423. [2] J. Ashburner, K.J. Friston, Voxel-based morphometry—the methods, Neuroimage 11 (2000) 805–821. [3] I. Coulier, J. Vries, K. Leenders, Is FDG-PET a useful tool in clinical practice for diagnosing corticobasal ganglionic degeneration? Mov. Disord. 18 (2003) 1175–1178. [4] D.W. Dickson, Neuropathologic differentiation of progressive supra nuclear palsy and cortical degeneration, J. Neurol. 246 (1999) II6–II15. [5] K.J. Friston, Commentary and opinion: II. Statistical parametric mapping: ontology and current issues, J. Cereb. Blood Flow Metab. 15 (1995) 361–370. [6] N. Hirono, K. Ishii, M. Sasaki, H. Kitagaki, M. Hashimoto, T. Imamura, S. Tanimukai, T. Hanihara, H. Kazui, E. Mori, Features of regional cerebral glucose metabolism abnormality in corticobasal degeneration, Dement. Geriatr. Cogn. Disord. 11 (2000) 139–146. [7] A.E. Lang, D.E. Riley, C. Bergeron, Corticobasal ganglionic degeneration, in: D.B. Calne (Ed.), Neurodegenerative Disease, Saunders, Philadelphia, 1994, pp. 877–894. [8] I. Litvan, Update on epidemiological aspects of progressive supranuclear palsy, Mov. Disord. 18 (2003) 43–50. [9] Y. Nagahama, H. Fukuyama, N. Turjanski, A. Kennedy, H. Yamauchi, Y. Ouchi, J. Kimura, D.J. Brooks, Cerebral glucose metabolism in corticobasal degeneration: comparison with progressive supra nuclear palsy and normal controls, Mov. Disord. 12 (1997) 691–696. [10] H. Nagasawa, T. Imamura, H. Nomura, M. Itoh, T. Ido, A case of corticobasal degeneration studied with positron emission tomography, Behav. Neurol. 6 (1993) 59–64. [11] B. Okuda, H. Tachibana, K. Keita, M. Takeda, M. Sugita, Comparison of brain perfusion in corticobasal degeneration and Alzheimer’s disease, Dement. Geriatr. Cogn. Disord. 12 (2001) 226–231. [12] W. Poewe, G. Wenning, The differential diagnosis of Parkinson’s disease, Eur. J. Neurol. 9 (2002) 23–30. [13] A. Schrag, Y. Ben-Shlomo, N.P. Quinn, Prevalence of progressive supranuclear palsy and multiple system atrophy: a cross-sectional study, Lancet 354 (1999) 1771–1775. [14] A. Schrag, C.D. Good, K. Miszkiel, H.R. Morris, C.J. Mathias, A.J. Lees, N.P. Quinn, Differentiation of atypical parkinsonian syndromes with routine MRI, Neurology 54 (2000) 697–702. [15] P. Soliveri, D. Monza, D. Paridi, Cognitive and magnetic imaging resonance imaging aspects of corticobasal degeneration and progressive supranuclear palsy, Neurology 53 (1999) 502–507. [16] M. Takeda, H. Tachibana, B. Okuda, K. Kawabata, M. Sugita, Electrophysical comparison between corticobasal degeneration and progressive supranuclear palsy, Clin. Neurol. Neurosurg. 100 (1998) 94–98. [17] S. Thobois, S. Guillouet, E. Broussolle, Contributions of PET and SPECT to the understanding of the pathophysiology of Parkinson’s disease, Neurophysiol. Clin. 31 (2001) 321–340.