| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Nuclear Medicine |
1 From the Division of Nuclear Medicine, Ghent University Hospital, De Pintelaan 185, B-9000 Ghent, Belgium. Received January 9, 2001; revision requested March 5; revision received April 20; accepted May 21. Supported by a special research grant from Ghent University and the Flemish government (BOZF 01104699). Address correspondence to K.J.V.L. (e-mail: koen.vanlaere@rug.ac.be).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
MATERIALS AND METHODS: Eighty-one healthy volunteers underwent both technetium 99m ethylene cysteine dimer SPECT and three-dimensional magnetization preparation rapid acquisition gradient-echo magnetic resonance (MR) imaging. Statistical parametric mapping was used to conduct VBM analysis of the morphologic data, which were compared voxel by voxel with the results of a similar analysis of the perfusion data and more specifically in brain areas showing significant perfusion changes.
RESULTS: VBM data, as compared with perfusion changes, indicated a more symmetric age-related gray matter volume decrease along the Sylvian fissure and in subcortical regions (P < .001). The combination of functional and structural changes indicated a relatively lower functional decrease with aging, as compared with the structural atrophy in the visual, parietal, sensorimotor, and right prefrontal cortices. Significant relative morphologic sex-based differences were found in the cerebellar and temporal cortices, but the comparison did not reveal significant differences between the functional and morphometric data.
CONCLUSION: Age-related perfusion changes are paralleled by similar more symmetric changes in gray matter concentration, which are more prominent than the perfusion changes in some regions. No sex-based differences between perfusion and gray matter concentration were found.
Index terms: Brain, atrophy, 10.83 • Brain, MR, 13.12141, 13.121412 • Brain, perfusion • Brain, radionuclide studies, 13.12162, 13.12163 • Brain, SPECT, 13.12162
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
99mTc-labeled ethylene cysteine dimer (ECD) and hexamethyl-propyleneamine oxime act as chemical microspheres and are retained in the brain in a fixed distribution that reflects the cerebral perfusion pattern, but these agents have different uptake mechanisms and patterns in healthy (3) and pathologic states (7,10). Because 99mTc ECD has higher in vitro stability and cerebral retention, lower radiation burden, rapid blood clearance of metabolites, and rapid elimination from extracerebral tissue, this radioligand has generated increased interest. However, the data on the properties of this agent in healthy populations are scarce in the literature (11,12).
For optimal clinical and research analysis sensitivity, the detailed characterization of covariate factors such as age and sex, which may affect interindividual physiologic uptakes, is very important. Although the results of many studies indicate that aging alters both regional blood flow and glucose metabolism with a specific frontotemporal pattern (8), to our knowledge, none of these functional topographic studies have included regional structural data in the analysis. This is particularly interesting because the results of numerous computed tomographic (CT) and magnetic resonance (MR) imaging studies have demonstrated an age-related decrease in brain size, an enlargement of cortical sulci, and an increase in cerebrospinal fluid space (CSF) after the age of 30 (13–17). Therefore, atrophy is considered a confounding variable in functional studies because it causes a partial volume effect that increases with lower spatial resolution. Since, to our knowledge, no clinically practical voxel-based atrophy correction schemes have been validated (18–20) yet, it was our aim to investigate brain perfusion at SPECT as a function of age and sex in healthy adult volunteers for 6 age decades and to correlate these perfusion findings with gray matter concentrations determined by using voxel-based morphometry (VBM).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
A total of 81 adult subjects (41 women, 40 men; age range, 20–81 years; average age, 44.2 years; all age decades stratified for sex) in whom good-quality MPRAGE images were available were included in this prospective cross-sectional study. The study was approved by the local ethics committee of Ghent University Hospital. Written informed consent was obtained from all the volunteers prior to the MR imaging and SPECT studies, after the nature of the procedures had been fully explained.
Background (Previous Study) Information
All selected volunteers underwent perfusion SPECT imaging with 925 MBq of 99mTc ECD (Dupont Pharmaceuticals, Brussels, Belgium). Acquisition and reconstruction, with incorporation of nonuniform attenuation and triple-energy window scatter correction, were performed as described previously (12). The previous study also included SPECT data analysis performed with special software used to test hypotheses about neurologic imaging data (SPM99; Wellcome Department of Cognitive Neurology, University College, London). In short, nonlinear warping with use of the corresponding MPRAGE MR images was performed. In that study (12), a highly significant decrease with age in the following tissue volumes was found (P < .001): anterior cingulate gyrus; left prefrontal, left lateral frontal, and left superior temporal and insular cortices; bilateral basal ganglia; and medial thalami. The contrast of relative increased perfusion with age indicated substantial clusters at the left and right cunei in the occipital cortex, in accordance with a relative sparing of these areas (ie, proportional scaling effect). As for sex-related differences, at SPECT, men had higher perfusion in the cerebellum bilaterally and in the left anterior temporal and orbitofrontal cortices, with extension to the anterior cingulate. On the other hand, women had significantly higher perfusion in the right parietal cortex than men (P < .001), with extension to the posterior temporal cortex.
MR Imaging
In the present study, MR images were obtained within 3 months after the first SPECT scans were obtained in all subjects (mean interval between date of MR imaging and date of SPECT, -1 week ± 3.5; range, -11 to 7 weeks) with a 1.5-T unit (Magnetom SP4000; Siemens, Erlangen, Germany). High-spatial-resolution anatomic imaging was performed with a three-dimensional MPRAGE sequence (9.7/4.0 [repetition time msec/echo time msec], 8° flip angle, 178 sections, 0.9-mm section thickness, 230 x 256 matrix, 250-mm field of view, one signal acquired). This sequence yielded T1-weighted sagittal images with a nearly isotropic resolution: The voxel size was 0.98 x 0.98 x 0.90 mm.
VBM Analysis
To investigate underlying age- and sex-related structural changes, a VBM study also was conducted (21,22). With the SPM software, automated segmentation based on a combination of intensity-driven and Bayesian classifications was carried out on normalized MR images, with subdivision of the gray matter, white matter, CSF, and extracerebral structures. Inhomogeneity corrections were incorporated because they have been shown to substantially improve the segmentation process (22). From this segmentation software, the relative CSF, gray matter, and white matter volumes normalized to each individual’s intracranial volume were extracted and evaluated as functions of age and sex by using linear regression analysis (SPSS v9.0; SPSS, Heverlee, Belgium).
For segmentation, nonlinear normalization was performed with the same parameters as those described earlier but with a final voxel size of 1.5 x 1.5 x 1.5 mm. Afterward, the binary segmented gray matter image was converted to a 3 x 3 x 3-mm voxel image. The gray matter image was then smoothed with an anisotropic Gaussian kernel approximating that of the SPECT spatial resolution, as measured with point sources. The in-plane (ie, transaxial) full width at half maximum was 7.9 mm, and the z-axis (ie, transverse) full width at half maximum was 8.9 mm (23). This differential smoothing to the same smoothness allows the assumption of an equal spatial autocovariance structure of both of the two data sets (24). Then, the same isotropic kernel as that used for the perfusion data was applied before analysis with the SPM software. Data were also adjusted by means of proportional scaling to a value of 50 and with a gray matter threshold of 0.10. To test the influence of the nonlinear normalization procedure, the same analysis was conducted with affine transformation (ie, no nonlinear parameters).
Age-related differences were studied on a voxel-wise basis in a correlation design. Linear voxel-wise regression with the subject’s age as a covariate was investigated. Contrasts were defined for age-related regional perfusion decrease and increase.
We also followed an approach used by Büchel et al (25), in which nonlinear regression was used with second-order polynomial expansion by defining the squared values of the subject’s age as a second covariate of interest. To assess the magnitude of the nonlinear effect of age, the squared values were used as covariates of interest in the study design, with the linear evolution of age as a covariate of no interest (26).
Sex-based differences were studied by comparing age-matched studies in a categorical population-comparison design with one scan per subject (voxel-wise t test). Contrasts were defined to examine the areas where perfusion was higher in men than in women and vice versa. The resultant set of statistical values for each contrast constituted a statistical parametric map of the t statistic, which was transformed into the normal distribution of the unit, SPM{Z} (27). This SPM{Z} map was interrogated at the height threshold corresponding to a P value (Phgt) of .001 and at the extent threshold corresponding to a P value (Pext) of less than .05, which were corrected for multiple comparisons unless stated otherwise.
To compare the differential structural adjusted response with the functional adjusted response, a single-subject, two-condition design in which perfusion was one condition, gray matter concentration was the second (24), and age and sex were covariates was explored. The covariate interactions were explicitly modeled and centered around the mean value. The variable-covariate interaction effects (eg, changes as functions of age) were then contrasted in the statistical analysis with the same thresholds as those defined earlier (ie, Phgt = .001; Pext = .05). This approach assumes implicit linearity between gray matter concentration and perfusion. For this analysis, a gray matter threshold of 0.40 was used.
The structural and functional adjusted responses were directly compared by using a 10-mm volume of interest centered at the most prominent voxel in the following three brain areas: the left insula, right insula, and left caudate head. These volume-of-interest data were analyzed by using linear regression with the software package described earlier (SPSS). All results for specific regions were defined probabilistically on the Montreal Neurological Institute template, with visual control of the normalized template images.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
|
|
|
Age-related Morphologic Changes Compared with Perfusion
The results of a direct comparison of gray matter concentration versus perfusion changes with age, which were studied in a categorical design with comparisons of age-condition interactions, showed a relatively smaller perfusion decrease with age in the parietal, sensorimotor, and occipital cortices, bilaterally distributed in equal extent and significance (Fig 3). A relatively smaller perfusion decrease with age was seen also in the right prefrontal cortex. The coordinates of the locations with significant differences and the corresponding P values are listed in Table 2.
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
On the other hand, the regional loss of gray matter volume with aging in our study was most prominent in the temporal and lateral frontal cortices and in the precentral, postcentral, and parietal cortices. A gray matter concentration decrease was found subcortically in the hippocampus, medial thalamus, and caudate heads. These voxel-based topographic results are in conjunction with those obtained in several region-of-interest–based volumetric MR studies (14,17,30–32). The relative gray and white matter and CSF volume changes with age in the adults in the current series correspond to data in a recent volumetric MR study (13) involving 116 volunteers examined from childhood to late adulthood. The results of these studies and of investigations by other authors have also shown an age-related dilatation of CSF ventricular spaces with aging. This dilatation consistently affects the median nuclei of the thalamus and superficial cortical sulci in the frontal, parietal, and parasagittal regions (13,31,33). The age-related reductions in cerebellar volume observed at lower levels of significance in this study, albeit mild compared with the cerebral shrinkage, have also been described in other studies (34,35).
However, there has been ample discussion in the literature as to whether neuronal loss subtending gross structural changes in healthy volunteers fully explains the regional changes measured with functional imaging. Contrary to widely held belief, neuronal concentration remains essentially constant during normal aging, but neuron size decreases—presumably as a result of reduced dendritic arborization and dendritic spine concentration (36)—and the number of glial cells decreases (8). These phenomena occur with substantial variability according to genetic factors, education, profession, lifestyle, intellectual and physical activity, and general physical condition (37,38). From our study data, we hypothesize that the majority of observed age-related regional perfusion changes can be attributed to underlying changes in gray matter volume (ie, increase in atrophy). In one recent positron emission tomography 15O-H2O perfusion study (39) in which region-of-interest–based partial volume corrections were applied on predefined regions in healthy volunteers, a similar conclusion was made.
Moreover, from the direct voxel-wise correlation contrast between 99mTc ECD uptake and gray matter concentration, it was determined that visual and sensorimotor regions show relatively increased sparing of perfusion with advancing age; this was also the case for the right periinsular and prefrontal cortices. The latter effect of functional lateralization toward the right hemisphere with advancing age has been observed in earlier studies (11,12, 40,41) but not in all investigations (29). A relative increase in basic neuronal perfusion delivery in the right, predominantly anterior, hemisphere consistently adheres to the emerging principle of greater left-to-right functional vulnerability with aging and neurodegeneration (42).
Sex-Specific Differences
The influence of sex on regional metabolism and perfusion is somewhat controversial and is probably marginal, but a limited number of studies have addressed this issue with reasonable statistical accuracy (29,41).
In this study, we found discrete but significantly higher perfusion in the left anterior temporal cortex and orbitofrontal cortex in men; to our knowledge, these findings have not been described before. Higher cerebellar perfusion in men has been observed before (41), but it is unclear whether it was related to relative differences in cerebellar volume in other independent volumetric studies (34). On the basis of our study data in the same subjects, it is likely that both gray matter volume and perfusion in the cerebellar hemispheres are greater in men to an equal extent, and thus, functional differences are based on structural properties.
Women had higher perfusion in the right inferior parietal cortex; this finding is in accordance with the results of a study involving the use of 99mTc hexamethyl-propyleneamine oxime in 120 adult subjects aged 50–90 years (43). MR volumetric study findings have shown no sex-related difference in parietal cortex volume in healthy volunteers after correction for global hemispheric differences (44), in accordance with our VBM analysis results. It is known that men relatively excel at visuoconstructional tasks (45); this finding has been connected to the parietal cortex. Further studies may be warranted to investigate a tentative relationship to the observed differential functional perfusion demand during the resting state (43).
The trend toward increased right-side temporal gray matter concentration in women indicated by our VBM analysis results is in line with the results of other recent MR volumetric studies (46,47); however, this finding is not in accordance with the results of a study with 465 healthy volunteers (48). In a histologic study (49), women had greater neuron concentration bilaterally in the planum temporale of the posterior temporal cortex. The physiologic implications of these findings, however, remain indeterminate and require further investigation.
Methodological Aspects and Study Limitations
There are a few issues concerning the use of VBM in this study that need addressing. First, since proportional scaling was used for global normalization of voxel values, a constant relationship between gray matter concentration and perfusion in each brain region was implicitly assumed. This also led to a proportional scaling of the error variance, which is a requirement for studying interaction effects. This assumption is reasonable since under normal circumstances, tracer uptake is proportional to local gray matter volume.
Second, the most important possible confounding factor in the direct comparison between perfusion and gray matter concentration is the risk of spillover effects between gray and white matter (24). Non–gray matter was given a value of zero at VBM, so no account was given to the contribution from spill in from white matter with regard to perfusion. for. It is known that white matter perfusion is approximately one-third to one-fourth of gray matter (8), and a more sophisticated approach would incorporate a linear combination of gray matter, white matter, and possibly CSF segmented image partitions. Whether underestimation of white matter perfusion spill in is an explanation for the apparently preserved 99mTc ECD uptake relative to gray matter volume in some regions may depend on the local tissue characteristics. Cortical spill-in effects should be relatively homogeneous, except in regions where more gray matter is present, such as the periinsular area. If present, however, such confounding effects are not expected to invalidate the observed asymmetric effect in the right anterior regions.
Third, the uptake pattern of 99mTc ECD is known to be nonlinear with respect to true blood flow, especially in high-perfusion areas such as the visual cortex (50), and as a consequence, activation or age-related changes may be less pronounced in these regions. Therefore, the observed finding that 99mTc ECD uptake decreases less than does atrophy in the occipital area may be at least partly confounded by the saturation effect in the relation between 99mTc ECD uptake and perfusion.
Finally, in this study, only indirectly partial volume effects were studied. A more rigid approach to correction for underlying changes would involve voxel-based partial volume correction. However, to our knowledge, such methods have not been established and well validated in static clinical SPECT or PET data (19,20) and therefore were not attempted in the current investigation. Apart from the underlying physiologic implications, the results of this study indicate the absolute necessity to interpret functional SPECT data in light of structural changes to obtain nondistorted analysis data with maximal sensitivity.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, K.J.V.L., R.A.D.; study concepts and design, K.J.V.L.; literature research, K.J.V.L.; clinical and experimental studies, K.J.V.L.; data acquisition and analysis/interpretation, K.J.V.L.; statistical analysis, K.J.V.L.; manuscript preparation, K.J.V.L.; manuscript definition of intellectual content, K.J.V.L., R.A.D.; manuscript editing, K.J.V.L.; manuscript revision/review and final version approval, K.J.V.L.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. Villain, B. Desgranges, F. Viader, V. de la Sayette, F. Mezenge, B. Landeau, J.-C. Baron, F. Eustache, and G. Chetelat Relationships between Hippocampal Atrophy, White Matter Disruption, and Gray Matter Hypometabolism in Alzheimer's Disease J. Neurosci., June 11, 2008; 28(24): 6174 - 6181. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Chetelat, B. Desgranges, B. Landeau, F. Mezenge, J. B. Poline, V. de la Sayette, F. Viader, F. Eustache, and J.-C. Baron Direct voxel-based comparison between grey matter hypometabolism and atrophy in Alzheimer's disease Brain, January 1, 2008; 131(1): 60 - 71. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Marti-Bonmati, J. J. Lull, G. Garcia-Marti, E. J. Aguilar, D. Moratal-Perez, C. Poyatos, M. Robles, and J. Sanjuan Chronic Auditory Hallucinations in Schizophrenic Patients: MR Analysis of the Coincidence between Functional and Morphologic Abnormalities Radiology, August 1, 2007; 244(2): 549 - 556. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. H. ten Dam, D. M. J. van den Heuvel, A. J. M. de Craen, E. L. E. M. Bollen, H. M. Murray, R. G. J. Westendorp, G. J. Blauw, and M. A. van Buchem Decline in Total Cerebral Blood Flow Is Linked with Increase in Periventricular but Not Deep White Matter Hyperintensities Radiology, April 1, 2007; 243(1): 198 - 203. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. D. Devous Sr., D. Altuna, N. Furl, W. Cooper, G. Gabbert, W. T. Ngai, S. Chiu, J. M. Scott III, T. S. Harris, J. K. Payne, et al. Maturation of Speech and Language Functional Neuroanatomy in Pediatric Normal Controls. J Speech Lang Hear Res, August 1, 2006; 49(4): 856 - 866. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A.-S. Pirson, T. Vander Borght, K. Van Laere, K. Takahashi, and S. Yamaguchi Age and Gender Effects on Normal Regional Cerebral Blood Flow AJNR Am. J. Neuroradiol., June 1, 2006; 27(6): 1161 - 1163. [Full Text] [PDF]
|
|
|
|
 |
|
 V. H. ten Dam, F. M.A Box, A. J.M. de Craen, D. M.J. van den Heuvel, E. L.E.M. Bollen, H. M. Murray, M. A. van Buchem, R. G.J. Westendorp, G. Jan Blauw, and on behalf of the PROSPER Study Group Lack of Effect of Pravastatin on Cerebral Blood Flow or Parenchymal Volume Loss in Elderly at Risk for Vascular Disease Stroke, August 1, 2005; 36(8): 1633 - 1636. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. F. Fotenos, A. Z. Snyder, L. E. Girton, J. C. Morris, and R. L. Buckner Normative estimates of cross-sectional and longitudinal brain volume decline in aging and AD Neurology, March 22, 2005; 64(6): 1032 - 1039. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Matsuda, T. Ohnishi, T. Asada, Z.-j. Li, H. Kanetaka, E. Imabayashi, F. Tanaka, and S. Nakano Correction for Partial-Volume Effects on Brain Perfusion SPECT in Healthy Men J. Nucl. Med., August 1, 2003; 44(8): 1243 - 1252. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G B Frisoni, C Testa, A Zorzan, F Sabattoli, A Beltramello, H Soininen, and M P Laakso Detection of grey matter loss in mild Alzheimer's disease with voxel based morphometry J. Neurol. Neurosurg. Psychiatry, December 1, 2002; 73(6): 657 - 664. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
), white matter (
), and CSF (*), which are expressed as percentage fractions of intracranial volume and plotted as functions of age. Data were obtained from automated segmentation at T1-weighted MPRAGE MR imaging performed in the healthy volunteers. For each volume group, linear fits with individual 95% regression prediction intervals are given.
) and gray matter concentration at VBM (
) in the (a) left and (b) right insulae and (c) the left caudate head as functions of age, with a Phgt threshold of less than .001. Both data sets were normalized to an average value of 50. The lines represent a linear regression through the data; perfusion is represented by the dashed line; and VBM, by the nondashed line. L = left, R = right.
