The Contribution Of ～(18)F-FDG PET Imaging In The Differential Diagnosis Of Alzheimer's Disease And Non-AD Degenerative Dementia

Objective: ①To validate the methodologicial aspects of rCMRglc measurements, such as visual inspection, ROI and voxel-based methods. ②To make an empirical validation of statistical thresholds to protect against false positive rCMRglc difference, and to validate the normal database that can be representative of the population being studied. ③To evulate the effects of age and sex on rCMRglc under resting state using a voxel-based method. ④To evaluate the differences in glucose metabolism according to dementia severity using voxel-based method.⑤To compare the glucose metabolism between early and late onset of Alzheimer's disease using voxel-based methods. ⑥To investigate the pattern of brain metabolism in non-AD dementia such as FTD, PDD, DLB and Parkinsonism plus syndromes and to assess the role of FDG PET imaging in the differential diagnosis of AD and non-AD degenerative dementia using statistical parametric mapping.Part one: Evaluation of brain glucose metabolism in normal and Alzheimer's disease patients using FDG-PET imaging:comparison of Visual, ROI and Statistical Parametirc Mapping methodsMethods and materials: 46normal controls and 23 AD patients are included in our studies, 6 subjects were removed from the normal control according their SPM results. 40 normal controls were divided into four subgroups according to the age and sex: 26 young controls and 14 old controls;23 male controls and 17 femal controls. To assess the relative cerebral glucose metabolic rate, FDG PET imaging was performed in all subjects. The results were evualted by three different analyticalmethods, including SPM (statistical parametric mapping), ROI (Region of Interst)and visual inspection.Results:(D Comparison of two normal controls: a reliable statistical thresholds (PO.001)and voxel thresholds (K=100voxels) can be used to protect against false positiverCMRglc differences.(2) Comparison of each subject to the remainder normal controls: 6 healthy subjects were removed from the normal controls.(3) In comparison of old controls to the young controls, in rCMRglc, significant age-related decreases were founded in inferior frontal cotex, and superior temporal cotex bilaterally;and left medial frontal cortex, left precuneus, left angular gyrus and right anterior cingulate cotex ipsilaterally. However, significant positive correlations of age with rCMRglc were observed in bilateral medial frontal cortex, cingulate cortex and cerebellum structures.(D In comparison of male controls to female controls: the females had significantly lower rCMRglc in left inferior parietal cortex, temporal cortex, antierior cingulate cortex and insula. However, the females also had significantly higher rCMRglc in the left frontal cortex, the right inferior parietal lobule and thalamus.(5) Based on the visual inspection, PET image indicated the focal hypometabolic areas in bilateral parietal, posterior cingulate cortex, precuneus, temporal cortex and global hypometabolism, relative to the sensorimotor and visual cortex, basal ganglia, thalamus and cerebellum.(6) ROI analysis demonstrated the rCMRglc reductions in the brain regions such as bilateral frontal, temporal, parietal, precuneus and the posterior cingulate cortex in AD group, as compared with NC group (PO.01). However, the rCMRglc of occipital primary visual cortex, basal ganglia, thalamus and cerebellum in AD group were relatively spared (P>0.01).? Voxel-based analysis between two groups revealed the following results: As compared with the NC group, regional reductions in rCMRglc and the related Brodmann areas(BA)of AD group included bilateral superior frontal gyrus(BA 6,10), middle frontal gyrus(BA 6,8,9,46), inferior parietal lobule(BA 39,40), left fusiform gyrus(BA 37) and inferior temporal cortex(BA 37), right precuneus(BA 7,19), middle temporal gyrus(BA21) and superior temporal gyrus(BA42), (PO.001, uncorrected). Conclusions:CD Assessment of regional cerebral glucose metabolism using the voxel-by-voxel analytical techniques such as SPM can be able to reproduce the rCMRglc reductions observed with ROI method in AD as compared with NC.(2) Although the ROI technique is a useful method, it only analyses selected areas, thus many brain regions may be left unexplored. In contrast, voxel-based analysis using statistical parametric mapping is expected to overcome this limitation and the relationship between the metabolic and its anatomical basis can be investigated more accurately.(3) Significant age-related decreases in rCMRglc were founded in inferior frontal cortex and superior temporal cotex bilaterally;and left medial frontal cortex, left precuneus, left angular gyrus and right anterior cingulate cotex ipsilaterally. However, significant positive correlations of age with rCMRglc were observed in bilateral medial frontal cortex, cingulate cortex and cerebellum structures.(4) The females had significantly lower rCMRglc in left inferior parietal cortex, temporal cortex, anterior cingulate cortex and insula. However, the females also had significantly higher rCMRglc in thalamus, especially.(5) The AD-associated PET pattern typically presents as focal cortical hypometabolism in bilateral parietal, temporal and/or frontal lobes and the posterior cingulate cortex. Our study showed that decline in FDG uptake in the posterior cingulate cortex, temporo-parietal and prefrontal association corticesallow identification of AD with a high sensitivity.Part two: The contribution of 18F-FDG PET imaging in the differential diagnosis of Alzheimer's disease and non-AD degenerative dementiaMethods and materials: 23 Alzheimer's disease(AD) patients and 22 non-AD degenerative dementia patients including 9 Parkinson's disease with dementia(PDD), 4 Frontotemporal dementia (FTD), 4 Dementia of Lewy bodies patients(DLB) and 5 Parkinsonism plus syndromes cases, 40 normal controls (NO who matched in age, sex and years of education ,were included in our study. AD groups and NC groups were divided into two subgroups according to the age onset, respectively, and 65yrs of age was been used as a cutoff value. To evaluate the rCMRglc, FDG PET imaging was performed in all subjects. Subsequently, Statistical comparison of PET data with the normal control was performed using Statistical Parametric Mapping, a voxel-based analysis method. Rusults:?Differences in glucose metabolism according to dementia severity: Hypometabolic areas in each AD groups compared with NC were analyzed according to each CDR stage (CDR=1~CDR=3). The hypometabolic areas appeared to increase as the level of CDR stage increases further, especially the frontal cortical areas, as CDR=1 compared with CDR=2 group, however, this trend was small between CDR=2 group and CDR=3 group. Additionally, our studies also showed that subcortical structures such as basal ganglia and thalamus were more hypometabolic in the CDR=3 group than CDR=1 and CDR=2 groups. This decrease in glucose metabolism may indicate spreading of overt metabolic abnormalities from the parieto-temporal and posterior cingulate cortices to the frontal cortical areas.?Comparison of glucose hypometabolism between early onset and late onset AD:When early onset AD group and late onset AD group were compared with the young conrols and old controls, respectively, hypometabolic brain region in thesetwo groups showed the similar PET pattern, indicating a typical hypometabolism in the parieto-temporal and posterior cingulate cortices, and/or frontal cortical areas. Howere, the hypometabolism observed in early onset AD group appeared much more extensive than in the late onset AD group. The results also indicated that reduction of cerebral glucose metabolism was significantly greater in early onset AD group than in the late onset group.(3) Comparison of glucose hypotabolism between FTD and NC: FTD group demonstrated significant regional metabolic reductions as compared with normal controls. With a height threshold of PO.001 (unconnected), there was significant hypometabolism in frontal areas, including bilateral superior, middle, inferior and medial frontal gyri and the cingulate gyri. Hypometabolic areas were also found in bilateral insular, superior parietal lobule, left precunes, and right inferior parietal lobule. There was also hypometabolism in the subcotrical structures including right putamen and right thalamic regions such as right medial dorsal nucles, ventral anterior nucleus.?Regional cerebral glucose metabolism in the PDD group was compared with that of normal control group, the PDD group showed similar reduction pattern to AD at the lateral parieto-temporal association cortex, lateral frontal cortex and cingulate cortex. The regional cerebral glucose metabolism was also decreased in subcotical structures including left caudate and right putamen. There was also hypometabolism in the dorsomedial thalamus, lateral posterior nucleus, and pulvinar. ?By the voxel-by-voxel comparison between the DLB group and normal control group, regional reductions in rCMRglc included foci in the bilateral middle occipital cortex, cuneus, precuneus, cingulate and posterior cingulate cortex, left anterior cingulate cortex, right superior parietal lobule, superior temporal cotex, middle frontal cortex. Hypometabolic areas were also found in the subcotical structures including putamen, caudate, lateral posterior nucleus, and pulvinar.Conclusions:(D The differences in glucose metabolism patterns with the increase in dementia severity indicated that the extentsion of hypometabolic areas increased as CDR transited from 1 to 2, but slightly increased as the dementia progressed further to CDR 3. Therefore, by a voxel-by-voxel analysis, early stage of AD can be differentiated from the late stage of AD using FDG PET imaging.(2) The glucose hypometabolism of early onset AD patients was much more severe in magnitude and extent than that of late onset AD patients. This difference may result not only from age-related cognitive reserve capacity but also from the different progression rate of these two groups.(3) The regional cerebral glucose metabolism was found to be different among AD and non-AD degenerative dementia patients. The AD-associated PET pattern typically presents as focal cortical hypometabolism in bilateral parietal, temporal and/or frontal lobes and the PCC, however these changes can also be seen in PDD and DLB. Howerver, greater hypometabolism was observed in partieto-temporal association cortex in AD group, as compared with PDD group. And the occipital hypometabolism can be used to differencial the DLB group from AD and PDD groups.? Using the voxel-based analytical method to evaluate FDG PET images of degenerative dementia is likely to be valuable in the future studies of dementia-related disorders.