Ci-fGBD: Cluster-Integrated Fast Generalized Bruhat Decomposition for Multimodal Data Clustering in Alzheimer's Disease.

Vascular factors increase the risks of developing Alzheimer's disease (AD) and they contribute to AD pathology. Since amyloid beta (Aβ) deposits can be observed in both diseases, there is an overlap which impedes a clear discrimination and difficult clinical diagnosis. In the present study, we compared cerebrospinal fluid (CSF) profiles of neurodegenerative and inflammatory biomarkers in a patient cohort of controls (n = 50), AD (n = 65) and vascular dementia (VaD) (n = 31) cases. Main results were validated in a second cohort composed of AD (n = 26), rapidly progressive AD (rpAD) (n = 15), VaD (n = 21), and cognitively unimpaired patients with vascular encephalopathy (VE) (n = 25) cases. In the study, cohort significant differences were detected in tau, p-tau, and Aβ1-42 (Aβ42) levels between AD and VaD patients, but not for the neuron-specific enolase (NSE), S100B protein, 14-3-3 and YKL-40. Differential tau, p-tau, and Aβ42 levels between AD and VaD were confirmed in the validation cohort, which additionally showed no differences between AD and rpAD, nor between VaD and VE. The evaluation of the biomarker performance in discrimination between AD and VaD patients revealed that the best diagnostic accuracy could be obtained when tau, p-tau, and Aβ42 were combined in form of Aβ42/p-tau (AUC 0.84–0.90, sensitivity 77–81%, specificity 80–93%) and (tau × p-tau)/Aβ42 ratio (AUC 0.83–0.87, sensitivity 73–81%, specificity 78–87%). Altogether, our studies provided neurodegenerative biomarker profiles in two cohorts of AD and VaD patients favoring the combination of CSF biomarker to differentiate between diseases.

BackgroundThe neurological events that follow traumatic brain injury (TBI) can include cognitive decline, personality changes, and long‐term structural and functional effects. At the microscale, the neuropathological changes following TBI overlap partially with those seen in Alzheimer’s disease (AD). Nevertheless, our understanding of how mild TBI (mTBI) can lead to AD remains inadequate. The predictive potential of acute mTBI degeneration patterns in prognosticating the risk of AD‐like structural changes is also largely unexplored.MethodIn 66 AD patients, 33 mTBI patients and 72 healthy controls (HCs), diffusion‐weighted imaging (DWI) was used to calculate the fractional anisotropy (FA) of white matter (WM) connections and T1 ‐weighted magnetic resonance imaging (MRI) was used to estimate gray matter (GM) thickness. Support vector machines (SVMs) then leveraged acute mTBI data to predict the severity of AD‐like neurodegeneration measured six months after mTBI.ResultIn both AD and mTBI, WM degeneration patterns and their extent are indistinguishable, in a statistical sense, within the fornix, corpus callosum, and corona radiata. In these structures, 78%, 35%, and 56% of their respective voxels have mean FAs which differ significantly from those of HCs for both mTBI and AD cohorts (p < 0.05). The mean FAs of these same sets of voxels are statistically equivalent across the mTBI and AD groups (p < 0.05, two one‐sided tests of statistical equivalence). mTBI and AD participants exhibit statistically significant similarities in their GM degeneration patterns, particularly within temporal poles (memory) and orbitofrontal cortices (executive function). 47% and 75% of voxels in these structures, respectively, exhibit statistical equivalence of mean FA across mTBI and AD (p < 0.05); these sets of voxels’ mean FAs are also statistically significantly different from those of HCs. SVM predictions suggest that acute post‐traumatic cognitive assessments contain information that accurately predicts AD‐like neurodegeneration patterns and severity observed chronically after mTBI (sensitivity: 0.87±0.09, specificity: 0.93±0.09; precision: 0.88±0.2).ConclusionOur findings suggest that mTBI‐induced atrophy can relegate the brain onto AD‐like neurodegeneration trajectories and that post‐mTBI cognitive assessments may provide early indications of AD risk after TBI.

BackgroundEarly identification of individuals with mild cognitive impairment (MCI) and Alzheimer’s disease (AD) is a clinical and research imperative. Use of diagnostic codes for MCI and AD identification has limitations. We used clinical notes to supplement diagnostic codes in the Veterans Affairs Healthcare System (VAHS) electronic health records (EHR) to identify and establish cohorts of Veterans recorded with MCI or AD.MethodsTargeted keyword searches for MCI (“Mild cognitive impairment;” “MCI”) and AD (“Alz*”) were used to extract clinical notes from the VAHS EHR from fiscal year (FY) 2010 through FY 2019. Iterative steps of inclusion and exclusion were applied until searches achieved a positive predictive value ≥ 80%. MCI and AD cohorts were identified via clinical notes and/or diagnostic codes (i.e., including Veterans recorded by “Notes Only,” “Notes + Code,” or “Codes Only”).ResultsA total of 2,134,661 clinical notes from 339,007 Veterans met the iterative search criteria for MCI due to any cause and 4,231,933 notes from 572,063 Veterans met the iterative search criteria for AD. Over the 10-year study period, the number of clinical notes recording AD was generally stable, whereas the number for MCI more than doubled. More Veterans were identified for the MCI or AD cohorts via clinical notes than by diagnostic codes, particularly in the AD cohort. Among Veterans identified by having “Notes + Code” for MCI, the number first recorded by a code was lower than the number first recorded by a note until FY 2015 and then gradually became comparable after FY 2015. Among Veterans identified by having “Notes + Code” for AD, the number first recorded by a note was more than double the number first recorded by a code AD in each of the FYs.ConclusionsClinical note-based identification captured more Veterans recorded with MCI and AD than diagnostic code-based identification.Supplementary InformationThe online version contains supplementary material available at 10.1007/s40120-023-00540-2.

Recently, REST (RE1-silencing transcription factor) gene has been shown to be lost in Alzheimer’s disease (AD), and a missense minor REST allele rs3796529-T has been shown to reduce the rate of hippocampal volume loss. However, whether the REST rs3796529 genotype is associated with the rate of functional deterioration in AD is unknown. A total of 584 blood samples from Taiwanese patients with AD were collected from January 2002 to December 2013. The diagnosis of AD was based on the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association criteria. The allele frequency of rs3796529-T was compared between the AD cohort and 993 individuals from the general population in Taiwan. Kaplan-Meier analysis, the log rank test and a multivariate Cox model were then used to evaluate the association between rs3796529-T and functional deterioration in the AD cohort. The allele frequency of rs3796529-T was significantly lower in the AD cohort compared to the general population cohort (36.82% vs. 40.73%, p=0.029). Kaplan-Meier analysis and the log rank test showed that the AD patients carrying the rs3796529 T/T genotype had a longer progression-free survival than those with the C/C genotype (p=0.012). In multivariate analysis, the rs3796529 T/T genotype (adjusted HR=0.593, 95% CI: 0.401-0.877, p=0.009) was an independent protective factor for functional deterioration. The rs3796529 T/T genotype was associated with slower functional deterioration in patients with AD. This finding may lead to a to better understanding of the molecular pathways involved, and prompt further development of novel biomarkers to monitor AD.

Alzheimer's disease (AD) and Lewy body disease (LBD) are characterized by early and gradual worsening perturbations in speeded cognitive responses. Using simple and choice reaction time tasks, we compared two indicators of cognitive speed within and across the AD and LBD spectra: mean rate (average reaction time across trials) and inconsistency (within person variability). The AD spectrum cohorts included subjective cognitive impairment (SCI, n = 28), mild cognitive impairment (MCI, n = 121), and AD (n = 45) participants. The LBD spectrum included Parkinson's disease (PD, n = 32), mild cognitive impairment in PD (PD-MCI, n = 21), and LBD (n = 18) participants. A cognitively unimpaired (CU, n = 39) cohort served as common benchmark. We conducted multivariate analyses of variance and discrimination analyses. Within the AD spectrum, the AD cohort was slower and more inconsistent than the CU, SCI, and MCI cohorts. The MCI cohort was slower than the CU cohort. Within the LBD spectrum, the LBD cohort was slower and more inconsistent than the CU, PD, and PD-MCI cohorts. The PD-MCI cohort was slower than the CU and PD cohorts. In cross-spectra (corresponding cohort) comparisons, the LBD cohort was slower and more inconsistent than the AD cohort. The PD-MCI cohort was slower than the MCI cohort. Discrimination analyses clarified the group difference patterns. For both speed tasks, mean rate and inconsistency demonstrated similar sensitivity to spectra-related comparisons. Both dementia cohorts were slower and more inconsistent than each of their respective non-dementia cohorts.

Given study-specific inclusion and exclusion criteria, Alzheimer's disease (AD) cohort studies effectively sample from different statistical distributions. This heterogeneity can propagate into cohort-specific signals and subsequently bias data-driven investigations of disease progression patterns. We built multi-state models for six independent AD cohort datasets to statistically compare disease progression patterns across them. Additionally, we propose a novel method for clustering cohorts with regard to their progression signals. We identified significant differences in progression patterns across cohorts. Models trained on cohort data learned cohort-specific effects that bias their estimations. We demonstrated how six cohorts relate to each other regarding their disease progression. Heterogeneity in cohort datasets impedes the reproducibility of data-driven results and validation of progression models generated on single cohorts. To ensure robust scientific insights, it is advisable to externally validate results in independent cohort datasets. The proposed clustering assesses the comparability of cohorts in an unbiased, data-driven manner.

INPP5D expression is associated with risk for Alzheimer's disease and induced by plaque-associated microglia

A DATA-DRIVEN MODEL OF BIOMARKER CHANGES IN SPORADIC ALZHEIMER'S DISEASE

INTRODUCTIONBlood protein biomarkers demonstrate potential for Alzheimer's disease (AD) diagnosis. Limited studies examine the molecular changes in AD blood cells.METHODSBulk RNA‐sequencing of blood cells was performed on AD patients of Chinese descent (n = 214 and 26 in the discovery and validation cohorts, respectively) with normal controls (n = 208 and 38 in the discovery and validation cohorts, respectively). Weighted gene co‐expression network analysis (WGCNA) and deconvolution analysis identified AD‐associated gene modules and blood cell types. Regression and unsupervised clustering analysis identified AD‐associated genes, gene modules, cell types, and established AD classification models.RESULTSWGCNA on differentially expressed genes revealed 15 gene modules, with 6 accurately classifying AD (areas under the receiver operating characteristics curve [auROCs] > 0.90). These modules stratified AD patients into subgroups with distinct disease states. Cell‐type deconvolution analysis identified specific blood cell types potentially associated with AD pathogenesis.DISCUSSIONThis study highlights the potential of blood transcriptome for AD diagnosis, patient stratification, and mechanistic studies.HighlightsWe comprehensively analyze the blood transcriptomes of a well‐characterized Alzheimer's disease cohort to identify genes, gene modules, pathways, and specific blood cells associated with the disease.Blood transcriptome analysis accurately classifies and stratifies patients with Alzheimer's disease, with some gene modules achieving classification accuracy comparable to that of the plasma ATN biomarkers.Immune‐associated pathways and immune cells, such as neutrophils, have potential roles in the pathogenesis and progression of Alzheimer's disease.

Among dementia-like diseases, Alzheimer disease (AD) and vascular dementia (VD) are two of the most frequent. AD and VD may share multiple neurological symptoms that may lead to controversial diagnoses when using conventional clinical and MRI criteria. Therefore, other approaches are needed to overcome this issue. Machine learning (ML) combined with magnetic resonance imaging (MRI) has been shown to improve the diagnostic accuracy of several neurodegenerative diseases, including dementia. To this end, in this study, we investigated, first, whether different kinds of ML algorithms, combined with advanced MRI features, could be supportive in classifying VD from AD and, second, whether the developed approach might help in predicting the prevalent disease in subjects with an unclear profile of AD or VD. Three ML categories of algorithms were tested: artificial neural network (ANN), support vector machine (SVM), and adaptive neuro-fuzzy inference system (ANFIS). Multiple regional metrics from resting-state fMRI (rs-fMRI) and diffusion tensor imaging (DTI) of 60 subjects (33 AD, 27 VD) were used as input features to train the algorithms and find the best feature pattern to classify VD from AD. We then used the identified VD–AD discriminant feature pattern as input for the most performant ML algorithm to predict the disease prevalence in 15 dementia patients with a “mixed VD–AD dementia” (MXD) clinical profile using their baseline MRI data. ML predictions were compared with the diagnosis evidence from a 3-year clinical follow-up. ANFIS emerged as the most efficient algorithm in discriminating AD from VD, reaching a classification accuracy greater than 84% using a small feature pattern. Moreover, ANFIS showed improved classification accuracy when trained with a multimodal input feature data set (e.g., DTI + rs-fMRI metrics) rather than a unimodal feature data set. When applying the best discriminant pattern to the MXD group, ANFIS achieved a correct prediction rate of 77.33%. Overall, results showed that our approach has a high discriminant power to classify AD and VD profiles. Moreover, the same approach also showed potential in predicting earlier the prevalent underlying disease in dementia patients whose clinical profile is uncertain between AD and VD, therefore suggesting its usefulness in supporting physicians' diagnostic evaluations.

To determine if the addition of delayed recall (DR) assessment adds sensitivity to the cognitive subscale of the Alzheimer Disease Assessment Scale (ADAS-cog) in clinical trials in mild cognitive impairment (MCI) and Alzheimer Disease (AD). Memory, particularly DR, is the most sensitive test for early detection of AD and MCI. However, it is not clear that assessment of DR adds benefit for measuring change over time after a diagnosis is made or in clinical trials. The ADAS-cog is the most commonly used tool to assess treatment efficacy in AD clinical trials. In an attempt to improve sensitivity to change, assessment of DR after the 3-trial, 10-word list was added to the standard 11-item ADAS-cog. We examined the added value of the DR in participants with MCI and AD followed for at least 1 year. Data from 111 subjects with AD and 259 subjects with MCI who were randomly assigned to the placebo arm of 2 clinical trials were included. Participants with AD had Mini-Mental State Examination scores of 13 to 27 and those with MCI had 24 to 30. We calculated the ADAS-cog11 score based on the original 11 items (range: best to worse, 0 to 70), the DR item score (range: 0 to 10 words not recalled), and the ADAS-cog12 (range: 0 to 80). We assessed the rate of missing items for DR over time, the change scores, the association between scores and baseline performance, and used longitudinal mixed effects regression models to examine the rate of change. At baseline AD subjects were near floor on DR (8.93 ± 1.6 SD) and showed little change over 1 year (0.12 ± 1.34); the MCI subjects baseline DR was 6.2 ± 2.2 with 1-year change of 0.20±1.7. We compared standardized change (change/SD) for ADAS-cog11, and 12 in MCI and found a 10% improvement with ADAS-cog12; there was no improvement in the AD group. In a subset of MCI and AD cases with matching Mini-Mental State Examination (23 to 27), the ADAS-cog12 provided an 18% improvement in standardized change in MCI subjects, with no benefit in the AD cohort, primarily owing to increased variance. The addition of DR to the ADAS-cog score increased the ability to detect change in subjects with MCI over 1 year compared with the ADAS-cog11 but increased the variance in subjects with AD, even in those with mild impairment These findings speak to the need to tailor outcome measures to the specific study population and diagnosis for maximal efficiency and economy when conducting clinical trials.

BackgroundThe complex pathogenesis of Alzheimer’s disease (AD) has resulted in limited current biomarkers for its classification and diagnosis, necessitating further investigation into reliable universal biomarkers or combinations.MethodsIn this work, we collect multiple CSF proteomics datasets and build a universal diagnose model by SVM-RFECV method combined with equal sample size and standard normalization design. The model was training in 297_CSF and then test the effect in other datasets.ResultsUtilizing machine learning, we identify a 12-protein panel from cerebrospinal fluid proteomic datasets. The universal diagnosis model demonstrated strong diagnostic capability and high accuracy across ten different AD cohorts across different countries and different detection technologies. These proteins involved in various biological processes related to AD and shows a tight correlation with established AD pathogenic biomarkers, including amyloid-β, tau/p-tau, and the Montreal Cognitive Assessment score. The high accuracy in the model may due to multiple protein combination based on comprehensive pathogenesis and different AD progress. Furthermore, it effectively differentiates AD from mild cognitive impairment (MCI) and other neurodegenerative disorders, especially the frontotemporal dementia (FTD), which share similar pathogenesis as AD.ConclusionThis study highlights a high accuracy, robustness and compatibility model of 12-protein panel whose detection is even based on label-free, TMT and DIA mass spectrometry or ELISA technologies, implicating its potential prospect in clinical application.

Antipsychotic Use and Risk of Stroke Among Community-Dwelling People With Alzheimer's Disease.

Hyperactivity of the renin-angiotensin system (RAS) is associated with the pathogenesis of Alzheimer's disease (AD) believed to be mediated by angiotensin-II (Ang-II) activation of the angiotensin type 1 receptor (AT1R). We previously showed that angiotensin-converting enzyme-1 (ACE-1) activity, the rate-limiting enzyme in the production of Ang-II, is increased in human postmortem brain tissue in AD. Angiotensin-III (Ang-III) activates the AT1R and angiotensin type-2 receptor (AT2R), but its potential role in the pathophysiology of AD remains unexplored. We measured Ang-II and Ang-III levels by ELISA, and the levels and activities of aminopeptidase-A (AP-A) and aminopeptidase-N (AP-N) (responsible for the production and metabolism of Ang-III, respectively) in human postmortem brain tissue in the mid-frontal cortex (Brodmann area 9) in a cohort of AD (n = 90) and age-matched non-demented controls (n = 59), for which we had previous measurements of ACE-1 activity, Aβ level, and tau pathology (also in the mid-frontal cortex). We found that both Ang-II and Ang-III levels were significantly higher in AD compared to age-matched controls and that Ang-III, rather than Ang-II, was strongly associated with Aβ load and tau load. Levels of AP-A were significantly reduced in AD but AP-A enzyme activity was unchanged whereas AP-N activity was reduced in AD but AP-N protein level was unchanged. Together, these data indicate that the APA/Ang-III/APN/Ang-IV/AT4R pathway is dysregulated and that elevated Ang-III could contribute to the pathogenesis of AD.

Given the high prevalence of psychotropic medication use in people with dementia and the potential for different prescribing practices in men and women, our study aimed to investigate sex differences in psychotropic medication use in older adults with Alzheimer's disease (AD) living in the US and Finland. We used data collected between 2005 and 2011 as part of the National Alzheimer's Coordinating Center (NACC) in the US, and Medication use and Alzheimer's disease (MEDALZ) cohorts in Finland. We evaluated psychotropic medication use (antidepressant, antipsychotic, anxiolytic, sedative, or hypnotic) in participants aged 65years or older. We employed multivariable logistic regression adjusted for demographics, co-morbidities, and other medications to estimate the magnitude of the association (adjusted odds ratio [aOR] with 95% confidence intervals [CIs]) according to sex. We included 1099 NACC participants (502 [45.68%] men, 597 [54.32%] women), and 67,049 participants from the MEDALZ cohort (22,961 [34.24%] men, 44,088 [65.75%] women). Women were more likely than men to use psychotropic medications: US, 46.2% vs. 33.1%, p<0.001; Finland, 45.3% vs. 36.1%, p<0.001; aOR was 2.06 (95% CI 1.58-2.70) in the US cohort and 1.38 (95% CI 1.33-1.43) in the Finnish cohort. Similarly, of the different psychotropic medications, women were more likely to use antidepressants (aOR-US: 2.16 [1.44-3.25], Finland: 1.52 [1.45-1.58]) and anxiolytics (aOR-US: 2.16 [1.83-3.96], Finland: 1.17 [1.13-1.23]) than men. Older women with AD are more likely to use psychotropic medications than older men, regardless of study population and country. Approaches to mitigate psychotropic medication use need to consider different prescribing habits observed in older women vs. men with AD.