Subtyping of mild cognitive impairment using a deep learning model based on brain atrophy patterns.
Ontology highlight
ABSTRACT:
SUBMITTER: Kwak K
PROVIDER: S-EPMC8714856 | biostudies-literature | 2021 Dec
REPOSITORIES: biostudies-literature
Publications
Subtyping of mild cognitive impairment using a deep learning model based on brain atrophy patterns.
Kwak Kichang K Giovanello Kelly S KS Bozoki Andrea A Styner Martin M Dayan Eran E
Cell reports. Medicine 20211221 12
Trajectories of cognitive decline vary considerably among individuals with mild cognitive impairment (MCI). To address this heterogeneity, subtyping approaches have been developed, with the objective of identifying more homogeneous subgroups. To date, subtyping of MCI has been based primarily on cognitive measures, often resulting in indistinct boundaries between subgroups and limited validity. Here, we introduce a subtyping method for MCI based solely upon brain atrophy. We train a deep learnin ...[more]
Similar Datasets
Project description:IntroductionMotoric cognitive risk (MCR) and amnestic mild cognitive impairment (aMCI) syndromes are each reliable predictors of incident Alzheimer's disease (AD), but MCR may be a stronger predictor of vascular dementia than AD. This study contrasted cortical and hippocampal atrophy patterns in MCR and aMCI.MethodsCross-sectional data from 733 older adults without dementia or disability (M age = 73.6; 45% women) in the multicountry MCR consortium were examined. MCR was defined as presence of slow gait and cognitive concerns. Amnestic MCI was defined as poor episodic memory performance and cognitive concerns. Cortical thickness and hippocampal volumes were quantified from structural MRIs. Multivariate and univariate general linear models were used to examine associations between cortical thickness and hippocampal volume in MCR and aMCI, adjusting for age, sex, education, total intracranial volume, white matter lesions, and study site.ResultsThe prevalence of MCR and aMCI was 7.64% and 12.96%, respectively. MCR was associated with widespread cortical atrophy, including prefrontal, insular, cingulate, motor, parietal, and temporal atrophy. aMCI was associated with hippocampal atrophy.ConclusionDistinct patterns of atrophy were associated with MCR and aMCI. A distributed pattern of cortical atrophy - that is more consistent with VaD or mixed dementia- was observed in MCR. A more restricted pattern of atrophy - that is more consistent with AD - was observed in aMCI. The biological underpinnings of MCR and aMCI likely differ and may require tailored interventions.IntroductionMotoric cognitive risk (MCR) and amnestic mild cognitive impairment (aMCI) syndromes are each reliable predictors of incident Alzheimer's disease (AD), but MCR may be a stronger predictor of vascular dementia than AD. This study contrasted cortical and hippocampal atrophy patterns in MCR and aMCI.MethodsCross-sectional data from 733 older adults without dementia or disability (M age = 73.6; 45% women) in the multicountry MCR consortium were examined. MCR was defined as presence of slow gait and cognitive concerns. Amnestic MCI was defined as poor episodic memory performance and cognitive concerns. Cortical thickness and hippocampal volumes were quantified from structural MRIs. Multivariate and univariate general linear models were used to examine associations between cortical thickness and hippocampal volume in MCR and aMCI, adjusting for age, sex, education, total intracranial volume, white matter lesions, and study site.ResultsThe prevalence of MCR and aMCI was 7.64% and 12.96%, respectively. MCR was associated with widespread cortical atrophy, including prefrontal, insular, cingulate, motor, parietal, and temporal atrophy. aMCI was associated with hippocampal atrophy.ConclusionDistinct patterns of atrophy were associated with MCR and aMCI. A distributed pattern of cortical atrophy - that is more consistent with VaD or mixed dementia- was observed in MCR. A more restricted pattern of atrophy - that is more consistent with AD - was observed in aMCI. The biological underpinnings of MCR and aMCI likely differ and may require tailored interventions.
| S-EPMC11794591 | biostudies-literature
Project description:BackgroundHistopathological studies and animal models suggest that hippocampal subfields may be differently affected by aging, Alzheimer's disease (AD), and other diseases. High-resolution images at 4 Tesla depict details of the internal structure of the hippocampus allowing for in vivo volumetry of different subfields. The aims of this study were as follows: (1) to determine patterns of volume loss in hippocampal subfields in normal aging, AD, and amnestic mild cognitive impairment (MCI). (2) To determine if measurements of hippocampal subfields provide advantages over total hippocampal volume for differentiation between groups.MethodsNinety-one subjects (53 controls (mean age: 69.3 ± 7.3), 20 MCI (mean age: 73.6 ± 7.1), and 18 AD (mean age: 69.1 ± 9.5) were studied with a high-resolution T2 weighted imaging sequence aimed at the hippocampus. Entorhinal cortex (ERC), subiculum, CA1, CA1-CA2 transition zone (CA1-2), CA3 & dentate gyrus (CA3&DG) were manually marked in the anterior third of the hippocampal body. Hippocampal volume was obtained from the Freesurfer and manually edited.ResultsCompared to controls, AD had smaller volumes of ERC, subiculum, CA1, CA1-2, and total hippocampal volumes. MCI had smaller CA1-2 volumes. Discriminant analysis and power analysis showed that CA1-2 was superior to total hippocampal volume for distinction between controls and MCI.ConclusionThe patterns of subfield atrophy in AD and MCI were consistent with patterns of neuronal cell loss/reduced synaptic density described by histopathology. These preliminary findings suggest that hippocampal subfield volumetry might be a better measure for diagnosis of early AD and for detection of other disease effects than measurement of total hippocampus.
| S-EPMC2943433 | biostudies-literature
Project description:The objective of this study was to evaluate the A/T/N biomarker scheme in relation with brain atrophy patterns in individuals with mild cognitive impairment (MCI). Of the 154 participants with MCI, 74 progressed to AD within 36-months, and 80 remained stable. In addition, 101 cognitively healthy participants and 102 participants with AD were included. The A/T/N classification was assessed with cerebrospinal fluid markers. Each individual was rated as either positive (abnormal) or negative (normal) on each biomarker. Brain atrophy was assessed with visual ratings from magnetic resonance imaging. None of the individuals with MCI progressed to AD if they had a negative "A" biomarker in conjunction with minimal atrophy. In contrary, several individuals with MCI progressed to AD if they had a positive "A" biomarker in conjunction with minimal atrophy. Numerous individuals with MCI showed inconsistency in the neurodegeneration domain ("N") regarding t-tau and atrophy. The assessment of the A/T/N classification scheme in addition with brain atrophy patterns in MCI, increases the knowledge of the clinical trajectories and the variability within the neurodegeneration domain. This emphasises that individuals with MCI display heterogeneous longitudinal patterns closely connected to their biomarker profiles, which could have important clinical implications.
| S-EPMC5976713 | biostudies-literature
Project description:BackgroundPredicting future Alzheimer's disease (AD)-related cognitive decline among individuals with subjective cognitive decline (SCD) or mild cognitive impairment (MCI) is an important task for healthcare. Structural brain imaging as measured by magnetic resonance imaging (MRI) could potentially contribute when making such predictions. It is unclear if the predictive performance of MRI can be improved using entire brain images in deep learning (DL) models compared to using pre-defined brain regions.MethodsA cohort of 332 individuals with SCD/MCI were included from the Swedish BioFINDER-1 study. The goal was to predict longitudinal SCD/MCI-to-AD dementia progression and change in Mini-Mental State Examination (MMSE) over four years. Four models were evaluated using different predictors: 1) clinical data only, including demographics, cognitive tests and APOE e4 status, 2) clinical data plus hippocampal volume, 3) clinical data plus all regional MRI gray matter volumes (N=68) extracted using FreeSurfer software, 4) a DL model trained using multi-task learning with MRI images, Jacobian determinant images and baseline cognition as input. Models were developed on 80% of subjects (N=267) and tested on the remaining 20% (N=65). Mann-Whitney U-test was used to determine statistically significant differences in performance, with p-values less than 0.05 considered significant.ResultsIn the test set, 21 patients (32.3%) progressed to AD dementia. The performance of the clinical data model for prediction of progression to AD dementia was area under the curve (AUC)=0.87 and four-year cognitive decline was R2=0.17. The performance was significantly improved for both outcomes when adding hippocampal volume (AUC=0.91, R2=0.26, p-values <0.05) or FreeSurfer brain regions (AUC=0.90, R2=0.27, p-values <0.05). Conversely, the DL model did not show any significant difference from the clinical data model (AUC=0.86, R2=0.13). A sensitivity analysis showed that the Jacobian determinant image was more informative than the MRI image, but that performance was maximized when both were included.ConclusionsThe DL model did not significantly improve the prediction of clinical disease progression in AD, compared to regression models with a single pre-defined brain region.
| S-EPMC10659533 | biostudies-literature
Project description:BackgroundPredicting future Alzheimer's disease (AD)-related cognitive decline among individuals with subjective cognitive decline (SCD) or mild cognitive impairment (MCI) is an important task for healthcare. Structural brain imaging as measured by magnetic resonance imaging (MRI) could potentially contribute when making such predictions. It is unclear if the predictive performance of MRI can be improved using entire brain images in deep learning (DL) models compared to using pre-defined brain regions.MethodsA cohort of 332 individuals with SCD/MCI were included from the Swedish BioFINDER-1 study. The goal was to predict longitudinal SCD/MCI-to-AD dementia progression and change in Mini-Mental State Examination (MMSE) over four years. Four models were evaluated using different predictors: (1) clinical data only, including demographics, cognitive tests and APOE ε4 status, (2) clinical data plus hippocampal volume, (3) clinical data plus all regional MRI gray matter volumes (N = 68) extracted using FreeSurfer software, (4) a DL model trained using multi-task learning with MRI images, Jacobian determinant images and baseline cognition as input. A double cross-validation scheme, with five test folds and for each of those ten validation folds, was used. External evaluation was performed on part of the ADNI dataset, including 108 patients. Mann-Whitney U-test was used to determine statistically significant differences in performance, with p-values less than 0.05 considered significant.ResultsIn the BioFINDER cohort, 109 patients (33%) progressed to AD dementia. The performance of the clinical data model for prediction of progression to AD dementia was area under the curve (AUC) = 0.85 and four-year cognitive decline was R2 = 0.14. The performance was improved for both outcomes when adding hippocampal volume (AUC = 0.86, R2 = 0.16). Adding FreeSurfer brain regions improved prediction of four-year cognitive decline but not progression to AD (AUC = 0.83, R2 = 0.17), while the DL model worsened the performance for both outcomes (AUC = 0.84, R2 = 0.08). A sensitivity analysis showed that the Jacobian determinant image was more informative than the MRI image, but that performance was maximized when both were included. In the external evaluation cohort from ADNI, 23 patients (21%) progressed to AD dementia. The results for predicted progression to AD dementia were similar to the results for the BioFINDER test data, while the performance for the cognitive decline was deteriorated.ConclusionsThe DL model did not significantly improve the prediction of clinical disease progression in AD, compared to regression models with a single pre-defined brain region.
| S-EPMC10949809 | biostudies-literature
Project description:AimsCognitive impairment is a common symptom in the trajectory of Parkinson's disease (PD). However, the pathological underpinning is not fully known. We aimed to explore the critical structural alterations in the process of cognitive decline and its relationships with the dopaminergic deficit and the level of related cerebrospinal fluid (CSF) proteins.MethodsNinety-four patients with PD and 32 controls were included in this study. Neuropsychological tests were performed at baseline and after 28 months to identify which patients had normal cognition and which ones developed PD-MCI after follow-up ("converters"). Gray matter atrophy was assessed in cross-sectional and longitudinal analyses, respectively. The associations between altered GMV with dopamine transporter (DAT) results and the level of CSF proteins were assessed.ResultsAmong the 94 patients with normal cognition at baseline, 24 (mean age, 63.1 years) developed PD-MCI after 28 months of follow-up, and 70 (mean age, 62.3 years) remained nonconverters. The converters showed significant right temporal atrophy at baseline and extensive atrophy in temporal lobe at follow-up. Progressive bilateral frontal lobe atrophy was found in the converters. Baseline right temporal atrophy was correlated with the striatal dopaminergic degeneration in the converters. No correlation was found between the right temporal atrophy and the alterations of CSF proteins.ConclusionEarly atrophy in temporal lobes and progressive atrophy in frontal lobes might be a biomarker for developing multidomain impairment of cognition and converting to PD-MCI. Furthermore, cognition-related temporal atrophy might be associated with dopaminergic deficit reflected by DAT scan but independent of CSF proteins in patients with PD who convert to PD-MCI.
| S-EPMC6930819 | biostudies-literature
Project description:ImportanceIt remains unclear which risk factors accelerate brain atrophy along with a progression from normal cognition to mild cognitive impairment (MCI).ObjectiveTo examine risk factors associated with the acceleration of brain atrophy and progression from normal cognition to MCI based on long-term longitudinal data for middle-aged and older adults.Design, setting, and participantsData for this cohort study were extracted from the Biomarkers for Older Controls at Risk for Dementia (BIOCARD) cohort, initiated at the National Institutes of Health from January 1, 1995, to December 31, 2005, and continued at Johns Hopkins University from January 1, 2015, to October 31, 2023. All participants were cognitively normal at baseline. The participants whose structural magnetic brain imaging (MRI) of the brain and cerebrospinal fluid (CSF) measures were available for over 10 years were included.ExposuresLongitudinal structural MRI of the brain and measurement of CSF biomarkers for Alzheimer disease pathology (ratio of amyloid β peptide 42 [Aβ42] to Aβ40, tau phosphorylated at threonine 181, and total tau).Main outcomes and measuresAnnual change rates of segmental brain volumes, Kaplan-Meier survival curves plotting time to event for progression to MCI symptom onset, and hazard ratios (HRs) determined by Cox proportional hazards regression models.ResultsA total of 185 participants (mean [SD] age, 55.4 [8.4] years; 116 women [63%]) were included and followed up for a maximum of 27 years (median, 20 [IQR, 18-22] years). The groups with high levels of atrophy in the white matter and enlargement in the ventricles had an earlier progression from normal cognition to MCI symptom onset (HR for white matter, 1.86 [95% CI, 1.24-2.49]; P = .001; HR for ventricles, 1.71 [95% CI, 1.19-2.24]; P = .009). Diabetes was associated with progression to MCI (HR, 1.41 [95% CI, 1.06-1.76]; P = .04), as was a low CSF Aβ42:Aβ40 ratio (HR, 1.48 [95% CI, 1.09-1.88]; P = .04), and their combination had a higher HR of 1.55 (95% CI, 1.13-1.98]; P = .03), indicating a synergic association of diabetes and amyloid pathology with MCI progression.Conclusions and relevanceIn this cohort study of middle-aged and older adults, higher rates of volume change in the white matter and ventricles, along with the presence of diabetes and a low CSF Aβ42:Aβ40 ratio, were identified as important risk factors for the progression to MCI. These results support the importance of identifying individuals who have accelerated brain atrophy to optimize preventive strategies for progression to MCI.
| S-EPMC11525609 | biostudies-literature
Project description:Recent genetic and proteomic studies demonstrate that clusterin/apolipoprotein-J is associated with risk, pathology, and progression of Alzheimer's disease (AD). Our main aim was to examine associations between plasma clusterin concentration and longitudinal changes in brain volume in normal aging and mild cognitive impairment (MCI). A secondary objective was to examine associations between peripheral concentration of clusterin and its concentration in the brain within regions that undergo neuropathological changes in AD. Non-demented individuals (N=139; mean baseline age 70.5 years) received annual volumetric MRI (912 MRI scans in total) over a mean six-year interval. Sixteen participants (92 MRI scans in total) were diagnosed during the course of the study with amnestic MCI. Clusterin concentration was assayed by ELISA in plasma samples collected within a year of the baseline MRI. Mixed effects regression models investigated whether plasma clusterin concentration was associated with rates of brain atrophy for control and MCI groups and whether these associations differed between groups. In a separate autopsy sample of individuals with AD (N=17) and healthy controls (N=4), we examined the association between antemortem clusterin concentration in plasma and postmortem levels in the superior temporal gyrus, hippocampus and cerebellum. The associations of plasma clusterin concentration with rates of change in brain volume were significantly different between MCI and control groups in several volumes including whole brain, ventricular CSF, temporal gray matter as well as parahippocampal, superior temporal and cingulate gyri. Within the MCI but not control group, higher baseline concentration of plasma clusterin was associated with slower rates of brain atrophy in these regions. In the combined autopsy sample of AD and control cases, representing a range of severity in AD pathology, we observed a significant association between clusterin concentration in the plasma and that in the superior temporal gyrus. Our findings suggest that clusterin, a plasma protein with roles in amyloid clearance, complement inhibition and apoptosis, is associated with rate of brain atrophy in MCI. Furthermore, peripheral concentration of clusterin also appears to reflect its concentration within brain regions vulnerable to AD pathology. These findings in combination suggest an influence of this multi-functional protein on early stages of progression in AD pathology.
| S-EPMC3425349 | biostudies-literature
Project description:BackgroundPeripheral (plasma) and central (cerebrospinal fluid, CSF) measures of tau are higher in Alzheimer's disease (AD) relative to prodromal stages and controls. While elevated CSF tau concentrations have been shown to be associated with lower grey matter density (GMD) in AD-specific regions, this correlation has yet to be examined for plasma in a large study.ObjectiveDetermine the neuroanatomical correlates of plasma tau using voxel-based analysis.MethodsCross-sectional data for 508 ADNI participants were collected for clinical, plasma total-tau (t-tau), CSF amyloid (Aβ42) and tau, and MRI variables. The relationship between plasma tau and GMD and between CSF t-tau and GMD were assessed on a voxel-by-voxel basis using regression models. Age, sex, APOEɛ4 status, diagnosis, and total intracranial volume were used as covariates where appropriate. Participants were defined as amyloid positive (Aβ+) if CSF Aβ42 was <192 pg/mL.ResultsPlasma tau was negatively correlated with GMD in the medial temporal lobe (MTL), precuneus, thalamus, and striatum. The associations with thalamus and striatum were independent of diagnosis. A negative correlation also existed between plasma tau and GMD in Aβ+ participants in the MTL, precuneus, and frontal lobe. When compared to CSF t-tau, plasma tau showed a notably different associated brain atrophy pattern, with only small overlapping regions in the fusiform gyrus.ConclusionPlasma tau may serve as a non-specific marker for neurodegeneration but is still relevant to AD considering low GMD was associated with plasma tau in Aβ+ participants and not Aβ-participants.
| S-EPMC5523909 | biostudies-literature
Project description:BackgroundAmyloid pathology in subjects with mild cognitive impairment (MCI) is an important risk factor for progression to dementia due to Alzheimer's disease. Predicting the onset of dementia is challenging even in the presence of amyloid, as time to progression varies considerably among patients and depends on the onset of neurodegeneration. Survival analysis can account for variability in time to event, but has not often been applied to MRI measurements beyond singular predefined brain regions such as the hippocampus. Here we used a voxel-wise survival analysis to identify in an unbiased fashion brain regions where decreased gray matter volume is associated with time to dementia, and assessed the effects of amyloid on these associations.MethodsWe included 276 subjects with MCI (mean age 67 ± 8, 41% female, mean Mini-Mental State Examination 26.6 ± 2.4), baseline 3D T1-weighted structural MRI, baseline cerebrospinal fluid (CSF) biomarkers, and prospective clinical follow-up. We fitted for each voxel a proportional Cox hazards regression model to study whether decreased gray matter volume predicted progression to dementia in the total sample, and stratified for baseline amyloid status.ResultsDementia at follow-up occurred in 122 (44%) subjects over an average follow-up period of 2.5 ± 1.5 years. Baseline amyloid positivity was associated with progression to dementia (hazard ratio 2.4, p < 0.001). Within amyloid-positive subjects, decreased gray matter volume in the hippocampal, temporal, parietal, and frontal regions was associated with more rapid progression to dementia (median (interquartile range) hazard ratio across significant voxels 1.35 (1.32-1.40)). Repeating the analysis in amyloid-negative subjects revealed similar patterns (median (interquartile range) hazard ratio 1.76 (1.66-1.91)).ConclusionsIn subjects with MCI, both abnormal amyloid CSF and decreased gray matter volume were associated with future progression to dementia. The spatial pattern of decreased gray matter volume associated with progression to dementia was consistent for amyloid-positive and amyloid-negative subjects.
| S-EPMC5596472 | biostudies-literature