A voxel-based specific regional analysis system for Alzheimer’s disease (VSRAD) was used for quantitative analysis of parahippocampal atropy with 1.5-tesla (T) MRI in a voxel-wise manner. The analysis of images acquired under a different imaging condition provides an error factor that has a calculated value. Clinical application of 3T-MRI is necessary for establishing a normal data base (N-DB) specific for 3T-MRI data, which permits appropriate application of VSRAD. We established an N-DB specific for 3T-MRI for use in VSRAD. The “Z-score of the parahippocampal gyrus” was 0.79 ± 0.32, and the N-DB of each age group did not have a big deflection when we analyzed a group of physically unimpaired persons in an N-DB specific for 3T-MRI. Therefore, we were able to confirm the validity of the customized N-DB. The “Z-score of the parahippocampal gyrus” was 1.62 ± 0.47 for the N-DB of VSRAD. The numerical value was high for the group of physically unimpaired persons.

Customization of normal data base specific for 3-tesla MRI is mandatory in VSRAD analysis.

Background: The hippocampus is a center of cognitive function and therefore hippocampal atrophy is the major factor in cognitive decline. Analysis of hippocampal size may make it possible to predict progression in cognitive impairment. To address this point, the present study investigated the relationship between hippocampal atrophy and dementia using magnetic resonance (MR) images and the Hasegawa Dementia Scale-Revised (HDS-R).



Methods:  The present study was performed on 274 subjects (14–97 years old; average, 66 years; 106 male and 168 female) who had no focal neurological deficit. Hippocampal area and whole brain area were measured in three series of coronal MR images taken from a 5-mm slice rostrally along dorsal edge of the pons, and hippocampal size normalized by calculating summated hippocampal areas as percentages of summated whole brain areas. Dementia was screened for using HDS-R.



Results:  Hippocampal size decreased and HDS-R reduced with age. Hippocampal atrophy was highly correlated with cognitive deterioration; a critical normalized hippocampal size for HDS-R of less than 20, which corresponds with mild cognitive impairment and dementia, was found in 65% of over 60-years old subjects and 98% of subjects with HDS-R of less than 20 were over 60 years old.



Conclusion:  There is a high probability that over 60-year-old people with a normalized hippocampal size of less than 1.0 would develop dementia in the future, even though their HDS-R might presently be over 20. Measurement of hippocampal size with coronal MR imaging may therefore contribute to prospective diagnoses of age-related dementia.

https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1479-8301.2007.00192.x

Hippocampal size may contribute to prospective diagnosis of age‐related dementia

AIM We studied the influences of silent brain infarction (SBI) and hypertension on brain atrophy and its longitudinal progression in healthy adults. METHODS MRI scans were performed on 109 neurologically normal adults (mean age, 58.6+/-5.8 years), with follow-up at an average of 4.9 years later. Patient histories of hypertension, smoking habits, and alcohol consumption were examined. We evaluated brain atrophy using the brain atrophy index (BAI; the ratio of the brain area to the intracranial area) and the ventricular atrophy index (VAI; the ratio of the ventricular area to the brain area) on MRI T1-weighted images at the levels of the basal ganglia and lateral ventricle in horizontal sections. RESULTS There were no differences in age, sex, dyslipidemia, BMI, smoking habit, and alcohol consumption between the normal group and the SBI or hypertension group. The BAI was significantly lower at entry for the SBI (+) group than for the SBI (-) group at both the basal ganglia and lateral ventricle levels (basal ganglia level, p=0.02; and lateral ventricle level, p=0.05). Moreover, the VAI was significantly higher at entry for the SBI (+) group than for the SBI (-) group at the lateral ventricle level (p=0.03). Furthermore, the BAI was significantly lower at entry for the hypertensive group than for the non-hypertensive group at the basal ganglia level (p=0.007). There were no significant differences in the annual variations of the BAI and VAI between the normal group and the SBI (+) or hypertensive group. CONCLUSION The present results suggest that the SBI and hypertension are accelerating factors for brain atrophy and ventricular dilatation.

The influences of silent cerebral infarction and hypertension on brain atrophy in normal adults

Donepezil (Aricept): The World’s First Drug for Alzheimer’s Disease

One of the most well-known and complex applications of artificial intelligence (AI) is Alzheimer’s detection, which lies in the field of medical imaging. The complexity in this task lies in the three-dimensional structure of the MRI scan images. In this paper, we propose to use 3D Convolutional Neural Networks (3D-CNN) for Alzheimer’s detection. 3D-CNNs have been a popular choice for this task. The novelty in our paper lies in the fact that we use a deeper 3D-CNN consisting of 10 layers. Also, with effectively training our model consisting of Batch Normalization layers that provide a regularizing effect, we don’t have to use any transfer learning. We also use the simple data augmentation technique of flipping. Our model is trained for binary classification that distinguishes between Alzheimer’s and normal, as well as multiclass classification consisting of Alzheimer’s, Mild Cognitive Impairment, and normal classes. We tested our model on the ADNI dataset and achieved 94.17% and 89.14% accuracy for binary classification and multiclass classification, respectively.

Deep 3D Convolution Neural Network for Alzheimer’s Detection

In recent years, voxel-based morphometry (VBM) has become a popular tool for the early diagnosis of Alzheimer disease. The Voxel-Based Specific Regional Analysis System for Alzheimer's Disease (VSRAD), a VBM system that uses MRI, has been reported to be clinically useful. The able-bodied person database (DB) of VSRAD, which employs sagittal plane imaging, is not suitable for analysis by axial plane imaging. However, axial plane imaging is useful for avoiding motion artifacts from the eyeball. Therefore, we created an able-bodied person DB by axial plane imaging and examined its utility. We also analyzed groups of able-bodied persons and persons with dementia by axial plane imaging and reviewed the validity. After using the DB of axial plane imaging, the Z-score of the intrahippocampal region improved by 8 in 13 instances. In all brains, the Z-score improved by 13 in all instances.

Kyohhito Iida

Papers

早期アルツハイマー型痴呆診断支援システム (vsrad) における水平断画像の有用性