Classification subject age and gender breakdown.

Provide a brief description of the statistical map you want to upload, such as "Brain parcellation results using the Desikan atlas on a single subject from the ADNI dataset"

Collection description

Subject species

homo sapiens

Modality

Structural MRI

Analysis level

single-subject

Cognitive paradigm (task)

Surface properties of object paradigms

Map type

Pa

MRI and amyloid PET fusion classification accuracies (%).

MRI and FDG PET fusion classification accuracies (%).

Objectives: Alzheimer's disease (AD) is the most prevalent neurodegenerative disorder and the most common form of dementia in the older people. Some types of mild cognitive impairment (MCI) are the clinical precursors of AD, while other MCI forms tend to remain stable over time and do not progress to AD. To discriminate MCI patients at risk of AD from stable MCI, we propose a novel deep-learning radiomics (DLR) model based on 18F-fluorodeoxyglucose positron emission tomography (18F-FDG PET) images and combine DLR features with clinical parameters (DLR+C) to improve diagnostic performance.Methods:18F-fluorodeoxyglucose positron emission tomography (PET) data from the Alzheimer's disease Neuroimaging Initiative database (ADNI) were collected, including 168 patients with MCI who converted to AD within 3 years and 187 patients with MCI without conversion within 3 years. These subjects were randomly partitioned into 90 % for the training/validation group and 10 % for the independent test group. The proposed DLR approach consists of three steps: base DL model pre-training, network features extraction, and integration of DLR+C, where a convolution network serves as a feature encoder, and a support vector machine (SVM) operated as the classifier. In comparative experiments, we compared our DLR+C method with four other methods: the standard uptake value ratio (SUVR) method, Radiomics-ROI method, Clinical method, and SUVR + Clinical method. To guarantee the robustness, 10-fold cross-validation was processed 100 times.Results: Under the DLR model, our proposed DLR+C was advantageous and yielded the best classification performance in the diagnosis of conversion with the accuracy, sensitivity, and specificity of 90.62 ± 1.16, 87.50 ± 0.00, and 93.39 ± 2.19%, respectively. In contrast, the respective accuracy of the other four methods reached 68.38 ± 1.27, 73.31 ± 6.93, 81.09 ± 1.97, and 85.35 ± 0.72 %. These results suggested the DLR approach could be used successfully in the prediction of conversion to AD, and that our proposed DLR-combined clinical information was effective.Conclusions: This study showed DLR+C could provide a novel and valuable method for the computer-assisted diagnosis of conversion to AD from MCI. This DLR+C method provided a quantitative biomarker which could predict conversion to AD in MCI patients.

Changes in cognitive status occurred within 36 months of follow-up.AD = Alzheimer's disease; CN = cognitively normal; cMCI = mild cognitive impairment in subjects who converted to AD; ncMCI = MCI subjects who remained stable; BMI- Body mass index; GDS – Geriatric Depression Scale; FAQ = Functional Assessment Questionnaire; MMSE = Mini-Mental State Exam.

Classification results in AD versus NC under three kinds of features(%).

Classification results in EMCI versus AD under three kinds of features(%).

Classification results in EMCI versus LMCI under three kinds of features(%).

Classification results in LMCI versus AD under three kinds of features(%).

BackgroundThe metabolic basis of Alzheimer disease (AD) is poorly understood, and the relationships between systemic abnormalities in metabolism and AD pathogenesis are unclear. Understanding how global perturbations in metabolism are related to severity of AD neuropathology and the eventual expression of AD symptoms in at-risk individuals is critical to developing effective disease-modifying treatments. In this study, we undertook parallel metabolomics analyses in both the brain and blood to identify systemic correlates of neuropathology and their associations with prodromal and preclinical measures of AD progression.Methods and findingsQuantitative and targeted metabolomics (Biocrates AbsoluteIDQ [identification and quantification] p180) assays were performed on brain tissue samples from the autopsy cohort of the Baltimore Longitudinal Study of Aging (BLSA) (N = 44, mean age = 81.33, % female = 36.36) from AD (N = 15), control (CN; N = 14), and “asymptomatic Alzheimer’s disease” (ASYMAD, i.e., individuals with significant AD pathology but no cognitive impairment during life; N = 15) participants. Using machine-learning methods, we identified a panel of 26 metabolites from two main classes—sphingolipids and glycerophospholipids—that discriminated AD and CN samples with accuracy, sensitivity, and specificity of 83.33%, 86.67%, and 80%, respectively. We then assayed these 26 metabolites in serum samples from two well-characterized longitudinal cohorts representing prodromal (Alzheimer’s Disease Neuroimaging Initiative [ADNI], N = 767, mean age = 75.19, % female = 42.63) and preclinical (BLSA) (N = 207, mean age = 78.68, % female = 42.63) AD, in which we tested their associations with magnetic resonance imaging (MRI) measures of AD-related brain atrophy, cerebrospinal fluid (CSF) biomarkers of AD pathology, risk of conversion to incident AD, and trajectories of cognitive performance. We developed an integrated blood and brain endophenotype score that summarized the relative importance of each metabolite to severity of AD pathology and disease progression (Endophenotype Association Score in Early Alzheimer’s Disease [EASE-AD]). Finally, we mapped the main metabolite classes emerging from our analyses to key biological pathways implicated in AD pathogenesis. We found that distinct sphingolipid species including sphingomyelin (SM) with acyl residue sums C16:0, C18:1, and C16:1 (SM C16:0, SM C18:1, SM C16:1) and hydroxysphingomyelin with acyl residue sum C14:1 (SM (OH) C14:1) were consistently associated with severity of AD pathology at autopsy and AD progression across prodromal and preclinical stages. Higher log-transformed blood concentrations of all four sphingolipids in cognitively normal individuals were significantly associated with increased risk of future conversion to incident AD: SM C16:0 (hazard ratio [HR] = 4.430, 95% confidence interval [CI] = 1.703–11.520, p = 0.002), SM C16:1 (HR = 3.455, 95% CI = 1.516–7.873, p = 0.003), SM (OH) C14:1 (HR = 3.539, 95% CI = 1.373–9.122, p = 0.009), and SM C18:1 (HR = 2.255, 95% CI = 1.047–4.855, p = 0.038). The sphingolipid species identified map to several biologically relevant pathways implicated in AD, including tau phosphorylation, amyloid-β (Aβ) metabolism, calcium homeostasis, acetylcholine biosynthesis, and apoptosis. Our study has limitations: the relatively small number of brain tissue samples may have limited our power to detect significant associations, control for heterogeneity between groups, and replicate our findings in independent, autopsy-derived brain samples.ConclusionsWe present a novel framework to identify biologically relevant brain and blood metabolites associated with disease pathology and progression during the prodromal and preclinical stages of AD. Our results show that perturbations in sphingolipid metabolism are consistently associated with endophenotypes across preclinical and prodromal AD, as well as with AD pathology at autopsy. Sphingolipids may be biologically relevant biomarkers for the early detection of AD, and correcting perturbations in sphingolipid metabolism may be a plausible and novel therapeutic strategy in AD.

Cerebrospinal fluid (CSF) 42 amino acid species of amyloid beta (Aβ42) and tau levels are strongly correlated with the presence of Alzheimer's disease (AD) neuropathology including amyloid plaques and neurodegeneration and have been successfully used as endophenotypes for genetic studies of AD. Additional CSF analytes may also serve as useful endophenotypes that capture other aspects of AD pathophysiology. Here we have conducted a genome-wide association study of CSF levels of 59 AD-related analytes. All analytes were measured using the Rules Based Medicine Human DiscoveryMAP Panel, which includes analytes relevant to several disease-related processes. Data from two independently collected and measured datasets, the Knight Alzheimer's Disease Research Center (ADRC) and Alzheimer's Disease Neuroimaging Initiative (ADNI), were analyzed separately, and combined results were obtained using meta-analysis. We identified genetic associations with CSF levels of 5 proteins (Angiotensin-converting enzyme (ACE), Chemokine (C-C motif) ligand 2 (CCL2), Chemokine (C-C motif) ligand 4 (CCL4), Interleukin 6 receptor (IL6R) and Matrix metalloproteinase-3 (MMP3)) with study-wide significant p-values (p

Classification results in NC versus EMCI under three kinds of features(%).

Classification results in NC versus LMCI under three kinds of features(%).

Additional file 1: Supplementary Table 1. Quality control (QC) information on pre-processing of DNA methylation samples and probes. Supplementary Table 2. Estimates of inflation and bias in the analysis CpG to CSF biomarker associations in Alzheimer's disease (AD) and cognitively normal (CN) samples. The bacon approach (PMID: 28129774) was implemented using the R package bacon. Conventional approach for inflation estimate is based on the method described in Devlin and Roeder (PMID: 11315092). Supplementary Table 3. Summary of number of significant CpGs and DMRs associated with CSF AD biomarkers. Supplementary Table 4. Significant CpGs associated with CSF total tau in cognitive normal (CN) and Alzheimer's disease (AD) subjects. Highlighted in red are significant DNAm to CSF biomarker associations with P-value < 10-5or FDR < 0.05, or disease by DNAm interaction with P-value < 0.05. Also highlighted in red are gene promoter regions mapped with significant CpGs. The significant DNAm were compared to analysis results of brain samples in Zhang et al. (2020) (PMID: 33257653) and Shireby et al. (2022) (PMID: 36153390). Supplementary Table 5. Significant CpGs associated with CSF Aβ42 in cognitive normal (CN) and Alzheimer's disease (AD) subjects. Highlighted in red are significant associations with P-value < 10-5, FDR < 0.05, or disease by DNAm interaction with P-value < 0.05. Highlights in yellow indicates overlap with significant DNAm in previous literature. Also highlighted in red are gene promoter regions mapped with significant CpGs. The significant DNAm were compared to analysis results of brain samples in Zhang et al. (2020) (PMID: 33257653) and Shireby et al. (2022) (PMID: 36153390). Supplementary Table 6. Significant CpGs associated with CSF phosphorylated tau181 in cognitive normal (CN) and Alzheimer's disease (AD) subjects. Values highlighted in red are significant associations with P-value < 10-5, FDR < 0.05, or disease by DNAm interaction with P-value < 0.05. Also highlighted in red are genes with significant CpGs in the promoter regions. Highlights in yellow indicates overlap with significant DNAm in previous literature. The significant DNAm were compared to analysis results of brain samples in Zhang et al. (2020) (PMID: 33257653) and Shireby et al. (2022) (PMID: 36153390). Supplementary Table 7. Significant DMRs at 5% Sidak adjusted P-value (z_sidak_p) associated with CSF Aβ42 in cognitively normal (CN) subjects and Alzheimer's disease (AD) subjects. For each DMR, annotations include location of the DMR based on hg19/GRCh37 genomic annotation (chr, start, end), nearby genes based on GREAT (GREAT_annotation), and Illumina gene annotations (UCSC_RefGene_Name), location with respect to CpG islands (Relation_to_Island), and overlap with enhancers described in Nasser et al. (2021) study (PMID: 33828297). Direction indicates positive or negative association between DNA methylation at a CpG located within the DMR and CSF biomarker. Highlights in yellow indicates overlap with significant DNAm in previous literature. Highlights in red indicate gene promoter regions mapped with significant DMRs. The significant DNAm were compared to analysis results of brain samples in Zhang et al. (2020) (PMID: 33257653) and Shireby et al. (2022) (PMID: 36153390). Supplementary Table 8. Significant DMRs at 5% Sidak adjusted P-value (z_sidak_p) associated with CSF pTau in cognitively normal (CN) subjects and Alzheimer's disease (AD) subjects. For each DMR, annotations include location of the DMR based on hg19/GRCh37 genomic annotation (chr, start, end), nearby genes based on GREAT (GREAT_annotation) and Illumina gene annotations (UCSC_RefGene_Name), location with respect to CpG islands (Relation_to_Island), and overlap with enhancers described in Nasser et al. (2021) study (PMID: 33828297). Direction indicates positive or negative association between DNA methylation at a CpG located within the DMR and CSF biomarker. Highlights in yellow indicates overlap with significant DNA methylation loci in previous literature. The significant DNAm were compared to analysis results of brain samples in Zhang et al. (2020) (PMID: 33257653) and Shireby et al. (2022) (PMID: 36153390). Supplementary Table 9. Significant DMRs at 5% Sidak adjusted P-value (z_sidak_p) associated with CSF total Tau in cogntively normal (CN) subjects and Alzheimer's disease (AD) subjects. For each DMR, annotations include location of the DMR based on hg19/GRCh37 genomic annotation (chr, start, end), nearby genes based on GREAT (GREAT_annotation) and Illumina gene annotations (UCSC_RefGene_Name), location with respect to CpG islands (Relation_to_Island), and overlap with enhancers described in Nasser et al. (2021) study (PMID: 33828297). Direction indicates positive or negative association between DNA methylation at a CpG located within the DMR and CSF biomarker. Highlights in yellow indicates overlap with significant DNA methylation loci in previous literature. The significant DNAm were compared to analysis results of brain samples in Zhang et al. (2020) (PMID: 33257653) and Shireby et al. (2022) (PMID: 36153390). Supplmentary Table 10. Results of pathway analysis using methylGSA (PMID: 30346483). In cognitively normal (CN) subjects, at 25% FDR, a total of 13 KEGG pathways and 76 Reactome pathways were enriched with CSF biomarker-associated CpGs, among which 3 KEGG pathways also reached 5% FDR. In AD samples, at 25% FDR, 10 KEGG pathways and 3 Reactome pathways were enriched with CSF biomarker-associated CpGs, among which 2 Reactome pathways also reached 5% FDR. Shown in red are FDR values less than 0.25, values of FDR < 0.05 are additionally highlighted in bold. Description of pathways that reached FDR < 0.25 in both CN and AD samples are also highlighted in red. Supplementary Table 11. Significant associations between CSF biomarker-associated CpGs and DMRs with target genes in blood samples of cognitively normal (CN) subjects and Alzheimer's disease (AD) subjects. We analyzed CpGs located in the promoter regions and distal regions separately. More specifically, for CpGs located in the promoter region (within ± 2 kb around the transcription start sites (TSS)), we tested the association between CpG methylation with expression levels of the target genes; for CpGs in the distal regions (> 2 kb from TSS), we tested associations between CpG methylation with expression levels of ten nearest genes upstream and downstream from the CpG.Supplementary 12. A total of 30127 CpG - mQTL pairs, associated with 16 unique CpGs, were significant in both cognitively normal (CN) and Alzheimer's disease (AD) sample analyses. The blood mQTLs were obtained from the GoDMC database. Definitions for columns under "Blood mQTLs" can be obtained from README file at http://mqtldb.godmc.org.uk/downloads.Supplementary Table 13. In cognitively normal (CN) samples, a total of 1518 mQTLs overlapped with the 24 GWAS nominated LD blocks in Kunkle et al. (2019) (PMID: 30820047). The mQTLs in blood were obtained from the GoDMC database. Annotations for CpGs include location of the CpG based on hg19/GRCh37 genomic annotation (Chr, Position), Illumina gene annotation (UCSC_RefGene_Name), the type of associated genomic feature (UCSC_RefGene_Group), and location with respect to CpG islands (Relation_to_Island).Supplementary Table 14. In Alzheimer's disease (AD) samples, a total of 41 mQTLs overlapped with the 24 GWAS nominated LD blocks in Kunkle et al. (2019) (PMID: 30820047). The mQTLs in blood were obtained from the GoDMC database. Annotations for CpGs include location of the CpG based on hg19/GRCh37 genomic annotation (Chr, Position), Illumina gene annotation (UCSC_RefGene_Name), the type of associated genomic feature (UCSC_RefGene_Group), and location with respect to CpG islands (Relation_to_Island).Supplementary Table 15. Sensitivity analysis for model that adjust cell type proportions estimated by the IDOL algorithm (PMID: 29843789), as implemented by estimateCellCounts2 function in R package FlowSorted.Blood.EPIC. All Aβ42-associated CpGs remained highly significant, with P-values ranging from 1.10 x 10-10 to 1.81 x 10-4. Supplementary Table 16. Sensitivity analysis for model that adjust cell type proportions estimated by the IDOL algorithm (PMID: 29843789), as implemented by estimateCellCounts2 function in R package FlowSorted.Blood.EPIC. All pTau181-associated CpGs remained highly significant, with P-values ranging from 1.39 x 10-8 to 2.92 x 10-3. Supplementary Table 17. CpGs with significant associations to both CSF AD biomarkers (in ADNI dataset) and Braak stage (in London dataset). Highlighted in red are associations that reached P < 10-5 in ADNI dataset analysis and P < 0.05 in London dataset. Supplementary Table 18. CSF biomarker-associated DMRs with significant associations to both CSF AD biomarkers (in ADNI dataset) and brain pathology (Braak stage in the independent London dataset).

Prediction performance of the LSTM classification model using various image input types. The highlighted values are the best per evaluation metric. Evaluation was calculated on the k = 5-fold cross validation sets from the training set with n = 561 participants of the ADNI Phase 1 dataset. The values are the average and standard deviation rates across all k = 5-fold cross validation sets. Visual representation were extracted using the ChestX-Ray8 database [16].

Descriptive statistics (mean±sd for continuous variables, percentage for binary) of the variables included in the models for European Prevention of Alzheimer’s Dementia Longitudinal Cohort Study (EPAD) and Alzheimer’s Diseasing Neuroimaging Initiative (ADNI).

The ORs of AD in CSF AD cases compared with population controls: multivariable logistic regression analysis.

Demographic characteristics of study samples.

CSF AD cases were compared with population controls and with CSF controls using logistic regression analysis.