BACKGROUND Few studies have compared factors related to cognitive function among people with simi... more BACKGROUND Few studies have compared factors related to cognitive function among people with similar genetic backgrounds but different lifestyles. OBJECTIVE We aimed to identify factors related to lower cognitive scores among older Japanese men in two genetically similar cohorts exposed to different lifestyle factors. METHODS This cross-sectional study of community-dwelling Japanese men aged 71-81 years included 2,628 men enrolled in the Kuakini Honolulu-Asia Aging Study based in Hawaii and 349 men in the Shiga Epidemiological Study of Subclinical Atherosclerosis based in Japan. We compared participant performance through Cognitive Abilities Screening Instrument (CASI) assessment in Hawaii (1991-1993) and Japan (2009-2014). Factors related to low cognitive scores (history of cardiovascular disease, cardiometabolic factors, and lifestyle factors) were identified with questionnaires and measurements. Multivariable logistic regression analysis was used to calculate the adjusted odds ra...
2016 9th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI), 2016
This paper presents super resolution (SR) of 3D MR images which is effective for brain segmentati... more This paper presents super resolution (SR) of 3D MR images which is effective for brain segmentation as practical application. Brain segmentation is one of important tasks to analyze brain morphometry. An accurate brain segmentation helps improve accuracy of post-processing. The segmentation is affected by the resolution of 3D MR images. In particular, the resolution in slice-select direction is much lower than the in-plane direction in some type of 3D MR images (e.g. T2-weighted images and proton density images). Therefore, we apply a learning-based SR method using sparse representation and self-similarity for generating high-resolution (HR) in slice-select direction to be same resolution in in-plane direction. With the visualized evaluation, it can be seen that the segmentation accuracy in SR results are improved by reducing partial volume effect. With the quantitative evaluation, one can confirm that the DICE value of SR results are higher than conventional interpolation method.
Background: The number of people with dementia is increasing in the world and it is of great publ... more Background: The number of people with dementia is increasing in the world and it is of great public health concern. Previous observational studies, mainly from western countries, have reported that modifiable factors for development of dementia were cardiometabolic risk factors, related to genetically determined metabolic abilities and lifestyle. There are no studies comparing which modifiable factors relate to low cognitive scores among populations with the same genetic background but different lifestyle. Objective: The purpose of the present study is to describe factors related to low cognitive scores in Japanese men in two cohorts that are genetically similar but with different lifestyle exposures. Methods: Population-based cross-sectional data from the Kuakini Honolulu-Asia Aging Study (HAAS) in Hawaii and Shiga Epidemiological Study of Subclinical Atherosclerosis (SESSA) in Japan were analyzed. Participants were Japanese men aged 71-81 years (2,628 men in Hawaii and 349 men in ...
Fluorine-19 magnetic resonance imaging (19F MRI) could be a promising approach for imaging amyloi... more Fluorine-19 magnetic resonance imaging (19F MRI) could be a promising approach for imaging amyloid deposition in the brain. However, the required features of a 19F MRI probe for amyloid detection remain unclear. In the present study, we investigated a series of compounds as potent 19F probes that could prevent the reduction in MR signal when bound to amyloid plaques in the brain. Each compound consists of styrylbenzoxazole as a core structure linked by a different length of polyethylene glycol (PEG) chain to one of three types of fluorine-labeled group: a trifluoroethoxy group, a hexafluoroisopropoxy group, or a 3',5'-bis(trifluoromethyl)benzylamino group. Among these compounds, 6-(3',6',9',15',18',21'-heptaoxa-23',23',23'-trifluoro tricosanyloxy)-2-(4'-dimethylaminostyryl)benzoxazole [compound 3b (m = 6)], which has a trifluoroethoxy group with seven ethylene glycol groups in the PEG chain, showed significant 19F MR signals in the bra...
Annals of clinical and translational neurology, 2014
We sought to identify the prevalence of MRI features of disproportionately enlarged subarachnoid ... more We sought to identify the prevalence of MRI features of disproportionately enlarged subarachnoid space hydrocephalus in possible idiopathic normal pressure hydrocephalus (DESH-iNPH) and to describe the clinico-radiological features and outcomes of a community-based investigation (The Vienna Trans-Danube Aging study). Of the 697 inhabitants (all 75 years old), 503 completed extensive neurological examinations at baseline and were followed up every 30 months thereafter with MRIs, mini-mental state examination (MMSE), and the Unified Parkinson Disease Rating Scale-Motor Section (UPDRSM). The DESH-iNPH participant data were compared with the data from participants with Evans index ratios >0.3 (ex vacuo hydrocephalus), cerebral small-vessel diseases, and normal MRIs. The widening of perivascular space was also evaluated by MRI in these groups. Eight participants with DESH-iNPH (1.6%) and 76 with ex vacuo hydrocephalus (16.1%) at baseline were identified. The mean MMSE in DESH-iNPH, ex...
Object. Controversy exists about correlations between histological tumor grade and magnetic reson... more Object. Controversy exists about correlations between histological tumor grade and magnetic resonance (MR) spectroscopy data. The authors studied single-voxel proton MR spectroscopy as a noninvasive way to evaluate grade of malignancy in intracranial meningiomas.Methods. The authors compared the results of MR spectroscopy with those derived by the MIB-1 staining index (SI) in 29 meningiomas. Proton MR spectroscopy was performed using stimulated echo acquisition and volume-localized solvent-attenuated proton nuclear MR sequences before surgery or other therapy.Twenty-four tumors were histologically benign (13 meningothelial, three fibrous, four transitional, three angiomatous, and one chordoid); four were atypical (Grade II), and one was papillary (Grade III). The mean MIB-1 SI in the benign group was significantly lower than those in the other groups (p = 0.0041). The mean choline-containing compound (Cho)/creatine and phosphocreatine (Cr) ratios in the benign and nonbenign groups w...
Annals of clinical and translational neurology, 2014
To identify the prevalence of MRI features of Binswanger's disease (BD), specifically MRI wit... more To identify the prevalence of MRI features of Binswanger's disease (BD), specifically MRI with diffuse white matter lesions and scattered multiple lacunes (BD-MRI), and to describe neurological features and pathological outcomes of a community-based cohort study. Of 697 participants (all 75 years old), 503 completed neurological examinations at baseline and were followed-up every 30 months thereafter with MRIs, the mini-mental state examination (MMSE) and the Unified Parkinson Disease Rating Scale-Motor Section (UPDRSM). Data from participants with BD-MRI were compared with those from participants with predominant white matter lesions (WML-MRI), scattered multiple lacunes (ML-MRI), or normal MRIs. Fourteen BD-MRI patients (2.8%) were detected at baseline. The mean MMSE scores in the BD-MRI, WML-MRI, ML-MRI, and normal MRIs groups were 26.4, 28.2, 28.4, and 28.5, respectively, and the mean UPDRSM scores were 9.1, 1.3, 3.1, and 1.7, respectively. At the 30-month follow-up, mortali...
Multimodal magnetic resonance images (e.g., T1-weighted image (TIWI) and T2-weighted image (T2WI)... more Multimodal magnetic resonance images (e.g., T1-weighted image (TIWI) and T2-weighted image (T2WI)) are used for accurate medical imaging analysis. Different modal images have different resolution depending on pulse sequence parameters under limited data acquisition time. Therefore, interpolation methods are used to match the low-resolution (LR) image with the high-resolution (HR) image. However, the interpolation causes blurring that affects analysis accuracy. Although some recent works such as non-local-means (NLM) filter have manifested impressive super-resolution (SR) performance with available HR modal images, the filter has high computational cost. Therefore, we propose a fast SR framework with iterative-guided back projection, which incorporates iterative back projection with a guided filter (GF) method for resolution enhancement of LR images (e.g., T2WI) by referring HR images in another modality image (e.g., T1WI). The proposed method not only achieves both high accuracy tha...
Proton magnetic resonance spectroscopy (MRS) can be used to investigate metabolite changes and sh... more Proton magnetic resonance spectroscopy (MRS) can be used to investigate metabolite changes and shows potential for clinical study and diagnosis of degenerative disorders such as mild cognitive impairment (MCI), Alzheimer’s disease (AD), dementia with Lewy bodies (DLB), and frontotemporal dementia (FTD). The common features of those degenerative disorders include a decrease in the concentration of NAA or in the NAA/Cr ratio and increase in the concentration of mIns or in the mIns/Cr ratio. Moreover, these findings tend to be regionally specific to the diseases, for example, to the hippocampus and the posterior cingulate and precuneal cortices in cases of AD, to the occipital lobe in cases of DLB, and to the frontal lobe in cases of FTD. Therefore MRS has potential clinical utility to differentiate those disorders. However, this approach may not be thoroughly established due to the inherent difficulty of quantitative measurement of metabolites, and therefore its utility is still limit...
Artificial intelligence (AI) shows promises in terms of diagnostic support on neuroimaging. We de... more Artificial intelligence (AI) shows promises in terms of diagnostic support on neuroimaging. We developed a software that predicts Alzheimer's disease (AD) using support vector machines (SVM) through three-dimensional brain MR images. Here, we will explain the general idea of voxel-based morphometry and SVM. We used the Alzheimer's Disease Neuroimaging Initiative (ADNI) database for SVM training and tested it on the Australian Imaging Biomarkers and Lifestyle Study of Ageing (AIBL) and the Japanese ADNI database. AI shows higher accuracy for predicting AD than a method of conventional statistical analysis, indicating potential clinical use for diagnostic support.
Background and Purpose— Few community-based studies have reported the prevalence of intracranial ... more Background and Purpose— Few community-based studies have reported the prevalence of intracranial artery stenosis (ICAS) assessed with magnetic resonance angiography. The aim was to determine the prevalence of ICAS using magnetic resonance angiography in a general population of Japanese men and to investigate the associations between ICAS and conventional cardiovascular risk factors. Methods— The Shiga Epidemiological Study of Subclinical Atherosclerosis randomly recruited and examined participants from Kusatsu City, Shiga, Japan, in 2006 to 2008 (baseline); 740 men returned for follow-up and underwent 1.5 T brain magnetic resonance angiography in 2012 to 2015. Participants were categorized as having no-ICAS, mild-ICAS (1 to <50%), or severe-ICAS (≥50%) in any of the arteries examined. After excluding the men with a history of stroke, 709 men were analyzed using multivariable logistic regression to assess independent associations of conventional cardiovascular risk factors with re...
Claudio Cuello, Pedro Rosa-Neto, McGill Center for Studies in Aging, Montreal, Quebec, Canada; Mc... more Claudio Cuello, Pedro Rosa-Neto, McGill Center for Studies in Aging, Montreal, Quebec, Canada; McGill Centre for Studies in Aging, Montreal, Quebec, Canada; Biospective Inc., Montreal, Quebec, Canada; McGill Center for Studies in Aging, Montreal, Quebec, Canada; Montreal Neurological Institute, Montreal, Quebec, Canada; McGill Center for Studies for Aging,Montreal, Quebec, Canada; McGill Center for Studies in Aging, Montreal, Quebec, Canada; McGill University, Montreal, Quebec, Canada; McGill Center for Studies in Aging, Porto Alegre, Brazil. Contact e-mail: erzimmer@gmail.com
Abstract Introduction We developed machine learning (ML) designed to analyze structural brain mag... more Abstract Introduction We developed machine learning (ML) designed to analyze structural brain magnetic resonance imaging (MRI), and trained it on the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. In this study, we verified its utility in the Japanese population. Methods A total of 535 participants were enrolled from the Japanese ADNI database, including 148 AD, 152 normal, and 235 mild cognitive impairment (MCI). Probability of AD was expressed as AD likelihood scores (ADLS). Results The accuracy of AD diagnosis was 88.0% to 91.2%. The accuracy of predicting the disease progression in non‐dementia participants over a 3‐year observation was 76.0% to 79.3%. More than 90% of the participants with low ADLS did not progress to AD within 3 years. In the amyloid positron emission tomography (PET)–positive MCI, the hazard ratio of progression was 2.39 with low ADLS, and 5.77 with high ADLS. When high ADLS was defined as N+ and Pittsburgh compound B (PiB) PET positivity was defined as A+, the time to disease progression for 50% of MCI participants was 23.7 months in A+N+, whereas it was 52.3 months in A+N‐. Conclusion These results support the feasibility of our ML for the diagnosis of AD and prediction of the disease progression.
Background: With the growing momentum for the adoption of machine learning (ML) in medical field,... more Background: With the growing momentum for the adoption of machine learning (ML) in medical field, it is likely that reliance on ML for imaging will become routine over the next few years. We have developed a software named BAAD, which uses ML algorithms for the diagnosis of Alzheimer's disease (AD) and prediction of mild cognitive impairment (MCI) progression.Methods: We constructed an algorithm by combining a support vector machine (SVM) to classify and a voxel-based morphometry (VBM) to reduce concerned variables. We grouped progressive MCI and AD as an AD spectrum and trained SVM according to this classification. We randomly selected half from the total 1,314 subjects of AD neuroimaging Initiative (ADNI) from North America for SVM training, and the remaining half were used for validation to fine-tune the model hyperparameters. We created two types of SVMs, one based solely on the brain structure (SVMst), and the other based on both the brain structure and Mini-Mental State Ex...
BACKGROUND Few studies have compared factors related to cognitive function among people with simi... more BACKGROUND Few studies have compared factors related to cognitive function among people with similar genetic backgrounds but different lifestyles. OBJECTIVE We aimed to identify factors related to lower cognitive scores among older Japanese men in two genetically similar cohorts exposed to different lifestyle factors. METHODS This cross-sectional study of community-dwelling Japanese men aged 71-81 years included 2,628 men enrolled in the Kuakini Honolulu-Asia Aging Study based in Hawaii and 349 men in the Shiga Epidemiological Study of Subclinical Atherosclerosis based in Japan. We compared participant performance through Cognitive Abilities Screening Instrument (CASI) assessment in Hawaii (1991-1993) and Japan (2009-2014). Factors related to low cognitive scores (history of cardiovascular disease, cardiometabolic factors, and lifestyle factors) were identified with questionnaires and measurements. Multivariable logistic regression analysis was used to calculate the adjusted odds ra...
2016 9th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI), 2016
This paper presents super resolution (SR) of 3D MR images which is effective for brain segmentati... more This paper presents super resolution (SR) of 3D MR images which is effective for brain segmentation as practical application. Brain segmentation is one of important tasks to analyze brain morphometry. An accurate brain segmentation helps improve accuracy of post-processing. The segmentation is affected by the resolution of 3D MR images. In particular, the resolution in slice-select direction is much lower than the in-plane direction in some type of 3D MR images (e.g. T2-weighted images and proton density images). Therefore, we apply a learning-based SR method using sparse representation and self-similarity for generating high-resolution (HR) in slice-select direction to be same resolution in in-plane direction. With the visualized evaluation, it can be seen that the segmentation accuracy in SR results are improved by reducing partial volume effect. With the quantitative evaluation, one can confirm that the DICE value of SR results are higher than conventional interpolation method.
Background: The number of people with dementia is increasing in the world and it is of great publ... more Background: The number of people with dementia is increasing in the world and it is of great public health concern. Previous observational studies, mainly from western countries, have reported that modifiable factors for development of dementia were cardiometabolic risk factors, related to genetically determined metabolic abilities and lifestyle. There are no studies comparing which modifiable factors relate to low cognitive scores among populations with the same genetic background but different lifestyle. Objective: The purpose of the present study is to describe factors related to low cognitive scores in Japanese men in two cohorts that are genetically similar but with different lifestyle exposures. Methods: Population-based cross-sectional data from the Kuakini Honolulu-Asia Aging Study (HAAS) in Hawaii and Shiga Epidemiological Study of Subclinical Atherosclerosis (SESSA) in Japan were analyzed. Participants were Japanese men aged 71-81 years (2,628 men in Hawaii and 349 men in ...
Fluorine-19 magnetic resonance imaging (19F MRI) could be a promising approach for imaging amyloi... more Fluorine-19 magnetic resonance imaging (19F MRI) could be a promising approach for imaging amyloid deposition in the brain. However, the required features of a 19F MRI probe for amyloid detection remain unclear. In the present study, we investigated a series of compounds as potent 19F probes that could prevent the reduction in MR signal when bound to amyloid plaques in the brain. Each compound consists of styrylbenzoxazole as a core structure linked by a different length of polyethylene glycol (PEG) chain to one of three types of fluorine-labeled group: a trifluoroethoxy group, a hexafluoroisopropoxy group, or a 3',5'-bis(trifluoromethyl)benzylamino group. Among these compounds, 6-(3',6',9',15',18',21'-heptaoxa-23',23',23'-trifluoro tricosanyloxy)-2-(4'-dimethylaminostyryl)benzoxazole [compound 3b (m = 6)], which has a trifluoroethoxy group with seven ethylene glycol groups in the PEG chain, showed significant 19F MR signals in the bra...
Annals of clinical and translational neurology, 2014
We sought to identify the prevalence of MRI features of disproportionately enlarged subarachnoid ... more We sought to identify the prevalence of MRI features of disproportionately enlarged subarachnoid space hydrocephalus in possible idiopathic normal pressure hydrocephalus (DESH-iNPH) and to describe the clinico-radiological features and outcomes of a community-based investigation (The Vienna Trans-Danube Aging study). Of the 697 inhabitants (all 75 years old), 503 completed extensive neurological examinations at baseline and were followed up every 30 months thereafter with MRIs, mini-mental state examination (MMSE), and the Unified Parkinson Disease Rating Scale-Motor Section (UPDRSM). The DESH-iNPH participant data were compared with the data from participants with Evans index ratios >0.3 (ex vacuo hydrocephalus), cerebral small-vessel diseases, and normal MRIs. The widening of perivascular space was also evaluated by MRI in these groups. Eight participants with DESH-iNPH (1.6%) and 76 with ex vacuo hydrocephalus (16.1%) at baseline were identified. The mean MMSE in DESH-iNPH, ex...
Object. Controversy exists about correlations between histological tumor grade and magnetic reson... more Object. Controversy exists about correlations between histological tumor grade and magnetic resonance (MR) spectroscopy data. The authors studied single-voxel proton MR spectroscopy as a noninvasive way to evaluate grade of malignancy in intracranial meningiomas.Methods. The authors compared the results of MR spectroscopy with those derived by the MIB-1 staining index (SI) in 29 meningiomas. Proton MR spectroscopy was performed using stimulated echo acquisition and volume-localized solvent-attenuated proton nuclear MR sequences before surgery or other therapy.Twenty-four tumors were histologically benign (13 meningothelial, three fibrous, four transitional, three angiomatous, and one chordoid); four were atypical (Grade II), and one was papillary (Grade III). The mean MIB-1 SI in the benign group was significantly lower than those in the other groups (p = 0.0041). The mean choline-containing compound (Cho)/creatine and phosphocreatine (Cr) ratios in the benign and nonbenign groups w...
Annals of clinical and translational neurology, 2014
To identify the prevalence of MRI features of Binswanger's disease (BD), specifically MRI wit... more To identify the prevalence of MRI features of Binswanger's disease (BD), specifically MRI with diffuse white matter lesions and scattered multiple lacunes (BD-MRI), and to describe neurological features and pathological outcomes of a community-based cohort study. Of 697 participants (all 75 years old), 503 completed neurological examinations at baseline and were followed-up every 30 months thereafter with MRIs, the mini-mental state examination (MMSE) and the Unified Parkinson Disease Rating Scale-Motor Section (UPDRSM). Data from participants with BD-MRI were compared with those from participants with predominant white matter lesions (WML-MRI), scattered multiple lacunes (ML-MRI), or normal MRIs. Fourteen BD-MRI patients (2.8%) were detected at baseline. The mean MMSE scores in the BD-MRI, WML-MRI, ML-MRI, and normal MRIs groups were 26.4, 28.2, 28.4, and 28.5, respectively, and the mean UPDRSM scores were 9.1, 1.3, 3.1, and 1.7, respectively. At the 30-month follow-up, mortali...
Multimodal magnetic resonance images (e.g., T1-weighted image (TIWI) and T2-weighted image (T2WI)... more Multimodal magnetic resonance images (e.g., T1-weighted image (TIWI) and T2-weighted image (T2WI)) are used for accurate medical imaging analysis. Different modal images have different resolution depending on pulse sequence parameters under limited data acquisition time. Therefore, interpolation methods are used to match the low-resolution (LR) image with the high-resolution (HR) image. However, the interpolation causes blurring that affects analysis accuracy. Although some recent works such as non-local-means (NLM) filter have manifested impressive super-resolution (SR) performance with available HR modal images, the filter has high computational cost. Therefore, we propose a fast SR framework with iterative-guided back projection, which incorporates iterative back projection with a guided filter (GF) method for resolution enhancement of LR images (e.g., T2WI) by referring HR images in another modality image (e.g., T1WI). The proposed method not only achieves both high accuracy tha...
Proton magnetic resonance spectroscopy (MRS) can be used to investigate metabolite changes and sh... more Proton magnetic resonance spectroscopy (MRS) can be used to investigate metabolite changes and shows potential for clinical study and diagnosis of degenerative disorders such as mild cognitive impairment (MCI), Alzheimer’s disease (AD), dementia with Lewy bodies (DLB), and frontotemporal dementia (FTD). The common features of those degenerative disorders include a decrease in the concentration of NAA or in the NAA/Cr ratio and increase in the concentration of mIns or in the mIns/Cr ratio. Moreover, these findings tend to be regionally specific to the diseases, for example, to the hippocampus and the posterior cingulate and precuneal cortices in cases of AD, to the occipital lobe in cases of DLB, and to the frontal lobe in cases of FTD. Therefore MRS has potential clinical utility to differentiate those disorders. However, this approach may not be thoroughly established due to the inherent difficulty of quantitative measurement of metabolites, and therefore its utility is still limit...
Artificial intelligence (AI) shows promises in terms of diagnostic support on neuroimaging. We de... more Artificial intelligence (AI) shows promises in terms of diagnostic support on neuroimaging. We developed a software that predicts Alzheimer's disease (AD) using support vector machines (SVM) through three-dimensional brain MR images. Here, we will explain the general idea of voxel-based morphometry and SVM. We used the Alzheimer's Disease Neuroimaging Initiative (ADNI) database for SVM training and tested it on the Australian Imaging Biomarkers and Lifestyle Study of Ageing (AIBL) and the Japanese ADNI database. AI shows higher accuracy for predicting AD than a method of conventional statistical analysis, indicating potential clinical use for diagnostic support.
Background and Purpose— Few community-based studies have reported the prevalence of intracranial ... more Background and Purpose— Few community-based studies have reported the prevalence of intracranial artery stenosis (ICAS) assessed with magnetic resonance angiography. The aim was to determine the prevalence of ICAS using magnetic resonance angiography in a general population of Japanese men and to investigate the associations between ICAS and conventional cardiovascular risk factors. Methods— The Shiga Epidemiological Study of Subclinical Atherosclerosis randomly recruited and examined participants from Kusatsu City, Shiga, Japan, in 2006 to 2008 (baseline); 740 men returned for follow-up and underwent 1.5 T brain magnetic resonance angiography in 2012 to 2015. Participants were categorized as having no-ICAS, mild-ICAS (1 to <50%), or severe-ICAS (≥50%) in any of the arteries examined. After excluding the men with a history of stroke, 709 men were analyzed using multivariable logistic regression to assess independent associations of conventional cardiovascular risk factors with re...
Claudio Cuello, Pedro Rosa-Neto, McGill Center for Studies in Aging, Montreal, Quebec, Canada; Mc... more Claudio Cuello, Pedro Rosa-Neto, McGill Center for Studies in Aging, Montreal, Quebec, Canada; McGill Centre for Studies in Aging, Montreal, Quebec, Canada; Biospective Inc., Montreal, Quebec, Canada; McGill Center for Studies in Aging, Montreal, Quebec, Canada; Montreal Neurological Institute, Montreal, Quebec, Canada; McGill Center for Studies for Aging,Montreal, Quebec, Canada; McGill Center for Studies in Aging, Montreal, Quebec, Canada; McGill University, Montreal, Quebec, Canada; McGill Center for Studies in Aging, Porto Alegre, Brazil. Contact e-mail: erzimmer@gmail.com
Abstract Introduction We developed machine learning (ML) designed to analyze structural brain mag... more Abstract Introduction We developed machine learning (ML) designed to analyze structural brain magnetic resonance imaging (MRI), and trained it on the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. In this study, we verified its utility in the Japanese population. Methods A total of 535 participants were enrolled from the Japanese ADNI database, including 148 AD, 152 normal, and 235 mild cognitive impairment (MCI). Probability of AD was expressed as AD likelihood scores (ADLS). Results The accuracy of AD diagnosis was 88.0% to 91.2%. The accuracy of predicting the disease progression in non‐dementia participants over a 3‐year observation was 76.0% to 79.3%. More than 90% of the participants with low ADLS did not progress to AD within 3 years. In the amyloid positron emission tomography (PET)–positive MCI, the hazard ratio of progression was 2.39 with low ADLS, and 5.77 with high ADLS. When high ADLS was defined as N+ and Pittsburgh compound B (PiB) PET positivity was defined as A+, the time to disease progression for 50% of MCI participants was 23.7 months in A+N+, whereas it was 52.3 months in A+N‐. Conclusion These results support the feasibility of our ML for the diagnosis of AD and prediction of the disease progression.
Background: With the growing momentum for the adoption of machine learning (ML) in medical field,... more Background: With the growing momentum for the adoption of machine learning (ML) in medical field, it is likely that reliance on ML for imaging will become routine over the next few years. We have developed a software named BAAD, which uses ML algorithms for the diagnosis of Alzheimer's disease (AD) and prediction of mild cognitive impairment (MCI) progression.Methods: We constructed an algorithm by combining a support vector machine (SVM) to classify and a voxel-based morphometry (VBM) to reduce concerned variables. We grouped progressive MCI and AD as an AD spectrum and trained SVM according to this classification. We randomly selected half from the total 1,314 subjects of AD neuroimaging Initiative (ADNI) from North America for SVM training, and the remaining half were used for validation to fine-tune the model hyperparameters. We created two types of SVMs, one based solely on the brain structure (SVMst), and the other based on both the brain structure and Mini-Mental State Ex...
Uploads
Papers by Akihiko Shiino