P2‐144: Model‐Based Comparison of Autosomal‐Dominant and Late‐Onset Alzheimer's Disease Progression in the Dian and ADNI Studies
Neil P. OxtobyAlexandra L. YoungMarco LorenziDavid M. CashPhilip S.J. WestonSébastien OurselinNick C. FoxJon M. SchottDaniel C. Alexander
4
Citation
0
Reference
10
Related Paper
Citation Trend
Abstract:
Understanding and identifying the earliest pathological changes of Alzheimer's disease (AD) is key to realizing disease-modifying treatments, which are likely to be most efficacious when given early. However, identifying individuals in the earliest, presymptomatic stage of typical, sporadic, late onset Alzheimer's disease (LOAD) is challenging. There is therefore of considerable interest to investigate the rarer autosomal-dominantly inherited forms of AD (ADAD). ADAD provides the opportunity to identify asymptomatic "at risk" individuals prior to the onset of cognitive decline, and has a predictable (if imperfect) age of symptom onset within families. This allows recruitment of presymptomatic individuals for observational/clinical trials. An important question is whether presymptomatic changes in ADAD mirror those in LOAD. We therefore performed a model-based analysis of biomarker changes in ADAD and LOAD. We used data-driven models to analyze biomarker data (magnetic resonance imaging, positron emission tomography, cerebrospinal fluid, cognitive test scores) from two multi-center observational studies: the Dominantly Inherited Alzheimer Network (DIAN) and the Alzheimer's Disease Neuroimaging Initiative (ADNI). From cross-sectional data we estimated pathological cascades using an event-based model1,2 (EBM). The EBM estimates a maximum-likelihood sequence of abnormality and its uncertainty (full posterior). To explore biomarker trajectories, we analyzed short-term longitudinal data using covariate-adjusted, nonparametric differential equation models3,4 (DEMs). Each trajectory enables estimation of a "transition time" between normal cognition and symptom onset, which combine to give a pathological cascade estimated from DEMs. We compared both estimated disease progression sequences across ADAD and LOAD. We fit DEMs to mutation carriers in DIAN and ApoE4-positive clinical progressors in ADNI. Both models estimate similar pathological cascades for ADAD and LOAD (Figure 1): accumulation of molecular pathology (amyloid, and tau where measured) followed by cognitive abnormalities, then brain hypometabolism/atrophy. Biomarker trajectories (e.g., Figure 2) accelerated from normal to abnormal levels for both ADAD and LOAD, with some evidence for plateauing/deceleration (consistent with sigmoidal behavior5) beyond symptom onset. AD progression sequences from data-driven models. (a)-(b) (upper): ADAD (DIAN-DF6). (c)-(d) (lower): LOAD (ADNI-1). Left: Event-based model (EBM). Biomarkers (imaging/molecular/cognitive) along the vertical axis are ordered by the maximum likelihood disease progression sequence (from top to bottom). The horizontal axis shows estimated variance in the posterior sequence, with positional likelihood per event given by grayscale intensity. Right: Differential equation models (DEMs). Biomarker trajectories estimated from differential data are anchored in time at the average point of symptom onset (right-hand edge). The disease progression sequence along the vertical axis (top to bottom) is determined by the average abnormality transition time from normal cognition to symptom onset (posterior median shown as red dots; median absolute deviation shown as red lines). The empirical probability of biomarker abnormality as a function of years prior to symptom onset is shown in grayscale intensity, increasing from left to right. Trajectories of adjusted bilateral hippocampus volume (in mm3) against years relative to symptom onset, estimated from data-driven differential equation models. (a) ADAD (left); (b) LOAD (right). Black dashed line is the average. Gray lines are sample trajectories from the posterior. Overlaid in blue (right-hand vertical axis on each plot) is the probability density for the abnormality transition time between normal level (shown as green lines with green box-plots to the left of each axis) and abnormal level (symptom onset; red lines and box-plots). a = asymptomatic; s = symptomatic; MC = mutation carrier (ADAD); N = normal (ADNI stable CN diagnosis); A = abnormal (symptom onset among ADNI clinical progressors). We found data-driven support for broadly similar biomarker progression of ADAD and LOAD (only minor differences), thus supporting research findings in ADAD being applicable to LOAD. Our data-driven model-based approaches increase understanding of pathological cascades with potential utility for patient staging and prognosis.Keywords:
Cognitive Decline
The Alzheimer's Disease Neuroimaging Initiative (ADNI) is a multisite, longitudinal study that assesses clinical, imaging, genetic, and biospecimen biomarkers through the process of normal aging to mild cognitive impairment and dementia. We present the creation of the Argentina‐ADNI—the first South American ADNI—and its effort to acquire data comparable with those gathered in other worldwide ADNI centers.
Cite
Citations (24)
Cite
Citations (2)
Alzheimer's disease (AD) and Mild Cognitive Impairment (MCI) are characterized by widespread pathological changes in the brain. At the same time, Alzheimer's disease is heritable with complex genetic underpinnings that may influence the timing of the related pathological changes in the brain and can affect the progression from MCI to AD. In this paper, we present a multivariate imaging genetics approach for prediction of conversion to Alzheimer's disease in patients with mild cognitive impairment. We employ multivariate pattern recognition approaches to obtain neuroimaging and polygenic discriminators between the healthy individuals and AD patients. We then design, in a linear manner, a composite imaging-genetic score for prediction of conversion to Alzheimer's disease in patients with mild cognitive impairment. We apply our approach within the Alzheimer's Disease Neuroimaging Initiative and show that the integration of polygenic and neuroimaging information improves prediction of conversion to AD.
Imaging genetics
Cite
Citations (9)
Cite
Citations (1,304)
Abstract Background Using the Alzheimer’s Disease Neuroimaging Initiative (ADNI) data archive we examined dementia‐related late‐onset cognitive impairment using neuroimaging and genetics biomarkers. Method We used FreeSurfer to parcellate the structural brain magnetic resonance imaging (MRI) data to derive homologous imaging signature vectors composed of 1,118 computed imaging biomarkers. The demographics of the 1,026 ADNI 1, ADNI GO and ADNI 2 arms of the ADNI study included participants ages 405 to 85, 266 normal CN’s (CDR = 0, Male:138, Female:128), 572 mild cognitively impaired MCI’s (CDR = 0.5, Male:227, Female:245), and 188 AD patients (CDR = 0.5/1, Male:102, Female:86). We extracted the AD‐related genetic markers and imaging markers for the network analysis. Using Plink, we modelled the genetics data and the using the Pipeline environment we identified single nucleotide polymorphisms (SNPs) associated with the clinical diagnosis. Network analyses were applied to examine multidimensional imaging‐genetics associations. Result The expected significant correlations between the SNPs and the neuroimaging phenotypes were confirmed using neuroimaging genetics networking analyses. These results may explain some of the differences among the AD, MCI and NC groups. We identified many associations between neuroimaging markers and genomic markers. For instance, cortical thickness, e.g., left and right hemispheres mean thickness, was more sensitive than regional volume morphometrics in capturing structural brain changes. We also identified salient neuroimaging (NI) markers that played important roles in neuroimaging networks discrimination between CN and MCI groups. However, regional volumes appeared to be more sensitive than thickness measures in discriminating between the MCI and AD groups. Conclusion Structural brain changes are important indicators of dementia progression. Network analysis pairing morphometric biomarkers with genetic indicators allows investigation of clinical and phenotypic associations that facilitate deep systematic understanding of genetic and environmental influences on aging and cognitive decline. Anatomical brain changes reflect the AD complex pathogenesis, their genetic associations, and their longitudinal propagation provide valuable clues to the progression of dementia and differences with normal aging. Further studies are necessary to untangle various deep brain‐networks and interpret their structural association with disease.
Imaging genetics
Genome-wide Association Study
Cite
Citations (0)
Imaging biomarker
Baseline (sea)
Cite
Citations (6)
Background: The North American Alzheimer's Disease Neuroimaging Initiative (NA-ADNI) was the first program to develop standardized procedures for Alzheimer's disease (AD) imaging biomarker collection. Objective: We describe the validation of acquisit
Imaging biomarker
Cite
Citations (25)
This article uses subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) to investigate and understand dementia-related late-onset cognitive impairment using neuroimaging and genetics biomarkers. The 1,245 subjects implemented using ADNI 1, ADNI GO and ADNI 2 were divided into three groups: those with Alzheimer's disease (AD), those with mild cognitive impairment (MCI), and those in the normal control (NC) group. Two hundred twenty eight of the subjects qualified for AD diagnosis at the baseline; 684 had MCI; and 333 were included in the NC group. The structural ADNI data were parcellated using FreeSurfer metrics, and all the SNPs for the 1,245 subject were extracted using Plink and the Pipeline environment. Network analyses were applied for all the subjects. Our previous study for the AD, MCI and NC subjects (using ADNI 1 - 808 subjects), indicated the significant associations between the SNPs and the neuroimaging phenotypes for the AD, MCI and NC subjects. The previous findings included that the set of 140 genes chosen (from 416 SNPs with p < 0.00005) represented commonly appearing genes in known AD gene networks. So we expect that we can get more meaningful result from the 1,245 subjects implemented using ADNI 1, ADNI GO and ADNI 2, because more big data have built up. We expect the significant correlations between the SNPs and the neuroimaging phenotypes in the 1,245 subjects in terms of neuroimaging genetics networking analyses. These analyses may explain some of the differences among the AD, MCI and NC groups.
Imaging genetics
Cite
Citations (0)
Here, we review progress by the Penn Biomarker Core in the Alzheimer's Disease Neuroimaging Initiative (ADNI) toward developing a pathological cerebrospinal fluid (CSF) and plasma biomarker signature for mild Alzheimer's disease (AD) as well as a biomarker profile that predicts conversion of mild cognitive impairment (MCI) and/or normal control subjects to AD. The Penn Biomarker Core also collaborated with other ADNI Cores to integrate data across ADNI to temporally order changes in clinical measures, imaging data, and chemical biomarkers that serve as mileposts and predictors of the conversion of normal control to MCI as well as MCI to AD, and the progression of AD. Initial CSF studies by the ADNI Biomarker Core revealed a pathological CSF biomarker signature of AD defined by the combination of Aβ1‐42 and total tau (T‐tau) that effectively delineates mild AD in the large multisite prospective clinical investigation conducted in ADNI. This signature appears to predict conversion from MCI to AD. Data fusion efforts across ADNI Cores generated a model for the temporal ordering of AD biomarkers which suggests that Aβ amyloid biomarkers become abnormal first, followed by changes in neurodegenerative biomarkers (CSF tau, F‐18 fluorodeoxyglucose‐positron emission tomography, magnetic resonance imaging) with the onset of clinical symptoms. The timing of these changes varies in individual patients due to genetic and environmental factors that increase or decrease an individual's resilience in response to progressive accumulations of AD pathologies. Further studies in ADNI will refine this model and render the biomarkers studied in ADNI more applicable to routine diagnosis and to clinical trials of disease modifying therapies.
Imaging biomarker
Biomarker Discovery
Cite
Citations (301)
To demonstrate measurement precision of cognitive domains in the Alzheimer's Disease Neuroimaging Initiative (ADNI) data set.
Data set
Cite
Citations (3)