Ken Fujiwara

Prediction of Outcomes in Mild Cognitive Impairment by Using 18F-FDG-PET: A Multicenter Study

BACKGROUNDn18F-FDG-PET is defined as a biomarker of neuronal injury according to the revised National Institute on Aging–Alzheimer’s Association criteria.nnnOBJECTIVEnThe objective of this multicenter prospective cohort study was to examine the value of 18F-FDG-PET in predicting the development of Alzheimer’s disease (AD) in patients with mild cognitive impairment (MCI).nnnMETHODSnIn total, 114 patients with MCI at 9 participating institutions underwent clinical and neuropsychological examinations, MRI, and 18F-FDG-PET at baseline. The cases were visually classified into predefined dementia patterns by three experts. Anautomated analysis for 18F-FDG-PET was also performed to calculate the PET score. Subjects were followed periodically for 3 years, and progression to dementia was evaluated.nnnRESULTSnIn 47% of the patients with MCI, progression of symptoms justified the clinical diagnosis of “probable AD”. The PET visual interpretation predicted conversion to AD during 3-year follow-up with an overall diagnostic accuracy of 68%. Overall diagnostic accuracy of the PET score was better than that of PET visual interpretation at all follow-up intervals, and the optimized PET score threshold revealed the best performance at the 2-year follow-up interval with an overall diagnostic accuracy of 83%,a sensitivity of 70%, and a specificity of 90%. Multivariate logistic regression analysis identified the PET score as the most significant predictive factor distinguishing AD converters from non-converters.nnnCONCLUSIONnThe PET score is the most statistically significant predictive factor for conversion from MCI to AD, and the diagnostic performance of the PET score is more promising for rapid converters over 2 years.

ONE-YEAR OBSERVATION IN FDG-PET IMPROVES PREDICTION OF PROGRESSIVE MCI TO ALZHEIMER'S DISEASE WITHIN 5 YEARS: LONGITUDINAL RESULTS FROM SEAD-J

Background: Recently, we introduced metrics for Alzheimer’s disease (AD) risk assessment called AD pattern similarity (AD-PS) scores. These are class-conditional probabilities generated using regularized logistic regression (RLR) estimated in a high-dimensional voxel space. Here we report exploratory analyses designed to determine if the AD-PS scores sensitivity would improve when (a) we use measures of cortical thickness (CT) and/or regional volumetric (Vols) measures as predictors instead of voxels, or (b) RLR is replaced by Random Forest (RF), a highly nonlinear classifier that also generates conditional probabilities. Methods: Baseline data from 359 participants of the AD Neuroimaging Initiative study were analyzed: 188 were cognitively normal (CN) and 171 had AD. FreeSurfer (FS) measures of CT and volumes and gray matter images were used to create four different sets of predictors (CT, Vols, combined CT and Vols (CT + Vols) and voxels) for both classifiers. Thus, we compared eight different versions of these classifiers when discriminating CN from AD. We varied the sample size from 20 to 280. Mean classification performance was estimated using 30 random data partitions in training and testing datasets. GLMNET and randomForest libraries versions of RLR and RF were used in our implementation. Results: For larger samples, RF and RLR relative classification accuracy was similar for the sets of predictors based on FS measures. RF clearly performed worse for the high-dimensional voxel space. For larger sample sizes, RLR estimated in the voxel space and both RLR and RF estimated in the combined CT+Vols sets of predictors, produced similar classification accuracy and were the best three performers. Both methods trained on the CT+Vols predictors often performed better than their counterparts trained independently on CT or Vols. Conclusions: When distinguishing CN from AD, given a sufficiently large sample size, we did not find evidence that using sets of predictors of lower dimensionality instead of voxel space improved the classification performance of RLR. RF non-linearity also did not yield advantages. These analyses suggest that RF or FS measures will not improve performance of the AD-PS scores. However, they do provide useful information about the impact of dimension and sample size on these two machine learning methods.

Image interpretation method of [C-11]BF-227 amyloid PET in J-ADNI

Nobuyuki Okamura, Takashi Kato, Ken Fujiwara, Kengo Ito, Michio Senda, Ryozo Kuwano, Kenji Ishii, Kazunari Ishii, Takeshi Iwatsubo The Japanese Alzheimer’s Disease Neuroimaging Initiative, Tohoku University School of Medicine, Sendai, Japan; National Center for Geriatrics and Gerontology, J-ADNI PET Core, Obu, Aichi, Japan; NCGG, J-ADNI PET Core, Obu, Japan; National Center for Geriatrics and Gerontology, J-ADNI PET Core, Obu, Japan; J-ADNI PET Core, Kobe, Japan; Niigata University, J-ADNI Biomarker Core, Niigata, Japan; Tokyo Metropolitan Geriatric Hospital, J-ADNI PET Core, Tokyo, Japan; Kinki University, J-ADNI PET Core, Osakasayama, Japan; The University of Tokyo, J-ADNI, Tokyo, Japan; J-ADNI, Tokyo, Japan. Contact e-mail: tkato@ncgg.go.jp

Association of cognitive decline with cerebral metabolism and education in amnestic MCI: Implications for the cognitive reserve hypothesis

CEREBRAL METABOLISM AND EDUCATION IN AMNESTIC MCI: IMPLICATIONS FOR THE COGNITIVE RESERVE HYPOTHESIS Takashi Kato, Kengo Ito, Ken Fujiwara, Takashi Yamada, Akinori Nakamura, Yutaka Arahata, Yukihiko Washimi SEAD-J Study Group, National Center for Geriatrics and Gerontology, Obu, Japan; National Center for Geriatrics and Gerontology, Obu, Japan; NCGG, Obu, Japan; 4 Department of Clinical and Experimental Neuroimaging National Center for Geriatrics and Gerontology, Obu, Japan; 5 National Center for Geriatrics and Gerontology, Obu, Japan; SEAD-J, Obu, Japan.

Relationship between cerebral glucose metabolism and [C-11]BF-227 accumulation in the stages from cognitively normal to Alzheimer's disease

relationship of atrophy in central olfactory structures to their functional deficit in AD by quantitatively determine the relationship of olfactory fMRI activation with local atrophy in the POC and hippocampus. Methods: 12 AD and 20 age-matched normal controls (NC) participated in this study. All AD and NC participants completed the University of Pennsylvania Smell Identification Test (UPSIT). The anatomical and fMRI images were acquired using a 3T MRI scanner. The olfactory stimulation paradigm consisted of three concentrations (0.10%, 0.32% and 1.0%) of the odorant (lavender) administered sequentially with three repetitions for each concentration. Each odor stimulation lasted for 6s, followed by 42s of baseline with odorless air. The hippocampus and POC were manually segmented and were saved as ROIs for subsequent fMRI activation voxels in those two local regions. Results: The UPSIT scores were significantly different (21.17 6 7.94 for AD and 30.85 6 5.69 for NC). The volumes and activation voxels of POC and hippocampus are showed in Fig. 1. The AD group showed prominent atrophy in both hippocampus and POC. Compared to NC group, the average volumes of the POC and hippocampus in the AD group were reduced by 39% and 44%. There was a high correlation between the atrophy of POC and hippocampus (p <0.001). Olfactory activations in the corresponding structures show a much greater reduction in AD: 98% in POC and 95% in hippocampus. The activation reduction and local atrophy in these two regions were significantly correlated (P 1⁄4 0.008 for POC and P 1⁄4 0.033 for hippocampus). Fig. 2 demonstrated olfactory fMRI activation difference in POC, hippocampus and insula between the two groups. Conclusions: In this study, we determined the morphological and functional changes in the brain areas most susceptible to AD pathology. We revealed that the reduction of BOLD response due to the disease in POC and hippocampus was much greater than the structural changes in the corresponding areas. These results indicate that olfactory fMRI can be a more sensitive marker for detection and evaluation of AD.