LEAP: learning embeddings for atlas propagation

Association of Vascular Risk Factors and Cerebrovascular Pathology With Alzheimer Disease Pathologic Changes in Individuals Without Dementia

BACKGROUND AND OBJECTIVES

Vascular risk factors (VRFs) and cerebral small vessel disease (cSVD) are common in patients with Alzheimer disease (AD). It remains unclear whether this coexistence reflects shared risk factors or a mechanistic relationship and whether vascular and amyloid pathologies have independent or synergistic influence on subsequent AD pathophysiology in preclinical stages. We investigated links between VRFs, cSVD, and amyloid levels (Aβ1-42) and their combined effect on downstream AD biomarkers, that is, CSF hyperphosphorylated tau (P-tau181), atrophy, and cognition.

METHODS

This retrospective study included nondemented participants (Clinical Dementia Rating < 1) from the European Prevention of Alzheimer's Dementia (EPAD) cohort and assessed VRFs with the Framingham risk score (FRS) and cSVD features on MRI using visual scales and white matter hyperintensity volumes. After preliminary linear analysis, we used structural equation modeling (SEM) to create a "cSVD severity" latent variable and assess the direct and indirect effects of FRS and cSVD severity on Aβ1-42, P-tau181, gray matter volume (baseline and longitudinal), and cognitive performance (baseline and longitudinal).

RESULTS

A total cohort of 1,592 participants were evaluated (mean age = 65.5 ± 7.4 years; 56.16% F). We observed positive associations between FRS and all cSVD features (all p < 0.05) and a negative association between FRS and Aβ1-42 (β = -0.04 ± 0.01). All cSVD features were negatively associated with CSF Aβ1-42 (all p < 0.05). Using SEM, the cSVD severity fully mediated the association between FRS and CSF Aβ1-42 (indirect effect: β = -0.03 ± 0.01), also when omitting vascular amyloid-related markers. We observed a significant indirect effect of cSVD severity on P-tau181 (indirect effect: β = 0.12 ± 0.03), baseline and longitudinal gray matter volume (indirect effect: β = -0.10 ± 0.03; β = -0.12 ± 0.05), and baseline cognitive performance (indirect effect: β = -0.16 ± 0.03) through CSF Aβ1-42.

DISCUSSION

In a large nondemented population, our findings suggest that cSVD is a mediator of the relationship between VRFs and CSF Aβ1-42 and affects downstream neurodegeneration and cognitive impairment. We provide evidence of VRFs indirectly affecting the pathogenesis of AD, highlighting the importance of considering cSVD burden in memory clinics for AD risk evaluation and as an early window for intervention. These results stress the role of VRFs and cerebrovascular pathology as key biomarkers for accurate design of anti-amyloid clinical trials and offer new perspectives for patient stratification.

Memory performance mediates subjective sleep quality associations with cerebrospinal fluid Alzheimer's disease biomarker levels and hippocampal volume among individuals with mild cognitive symptoms

Sleep disturbances are prevalent in Alzheimer's disease (AD), affecting individuals during its early stages. We investigated associations between subjective sleep measures and cerebrospinal fluid (CSF) biomarkers of AD in adults with mild cognitive symptoms from the European Prevention of Alzheimer's Dementia Longitudinal Cohort Study, considering the influence of memory performance. A total of 442 participants aged >50 years with a Clinical Dementia Rating (CDR) score of 0.5 completed the Pittsburgh Sleep Quality Index questionnaire and underwent neuropsychological assessment, magnetic resonance imaging acquisition, and CSF sampling. We analysed the relationship of sleep quality with CSF AD biomarkers and cognitive performance in separated multivariate linear regression models, adjusting for covariates. Poorer cross-sectional sleep quality was associated with lower CSF levels of phosphorylated tau and total tau alongside better immediate and delayed memory performance. After adjustment for delayed memory scores, associations between CSF biomarkers and sleep quality became non-significant, and further analysis revealed that memory performance mediated this relationship. In post hoc analyses, poorer subjective sleep quality was associated with lesser hippocampal atrophy, with memory performance also mediating this association. In conclusion, worse subjective sleep quality is associated with less altered AD biomarkers in adults with mild cognitive symptoms (CDR score 0.5). These results could be explained by a systematic recall bias affecting subjective sleep assessment in individuals with incipient memory impairment. Caution should therefore be exercised when interpreting subjective sleep quality measures in memory-impaired populations, emphasising the importance of complementing subjective measures with objective assessments.

Alzheimer's disease genetic pathways impact cerebrospinal fluid biomarkers and imaging endophenotypes in non-demented individuals

INTRODUCTION

Unraveling how Alzheimer's disease (AD) genetic risk is related to neuropathological heterogeneity, and whether this occurs through specific biological pathways, is a key step toward precision medicine.

METHODS

We computed pathway-specific genetic risk scores (GRSs) in non-demented individuals and investigated how AD risk variants predict cerebrospinal fluid (CSF) and imaging biomarkers reflecting AD pathology, cardiovascular, white matter integrity, and brain connectivity.

RESULTS

CSF amyloidbeta and phosphorylated tau were related to most GRSs. Inflammatory pathways were associated with cerebrovascular disease, whereas quantitative measures of white matter lesion and microstructure integrity were predicted by clearance and migration pathways. Functional connectivity alterations were related to genetic variants involved in signal transduction and synaptic communication.

DISCUSSION

This study reveals distinct genetic risk profiles in association with specific pathophysiological aspects in predementia stages of AD, unraveling the biological substrates of the heterogeneity of AD-associated endophenotypes and promoting a step forward in disease understanding and development of personalized therapies.

HIGHLIGHTS

Polygenic risk for Alzheimer's disease encompasses six biological pathways that can be quantified with pathway-specific genetic risk scores, and differentially relate to cerebrospinal fluid and imaging biomarkers. Inflammatory pathways are mostly related to cerebrovascular burden. White matter health is associated with pathways of clearance and membrane integrity, whereas functional connectivity measures are related to signal transduction and synaptic communication pathways.

Continuous β-Amyloid CSF/PET Imbalance Model to Capture Alzheimer Disease Heterogeneity

BACKGROUND AND OBJECTIVES

Discordance between CSF and PET biomarkers of β-amyloid (Aβ) might reflect an imbalance between soluble and aggregated species, possibly reflecting disease heterogeneity. Previous studies generally used binary cutoffs to assess discrepancies in CSF/PET biomarkers, resulting in a loss of information on the extent of discordance. In this study, we (1) jointly modeled Aβ-CSF/PET data to derive a continuous measure of the imbalance between soluble and fibrillar pools of Aβ, (2) investigated factors contributing to this imbalance, and (3) examined associations with cognitive trajectories.

METHODS

Across 822 cognitively unimpaired (n = 261) and cognitively impaired (n = 561) Alzheimer's Disease Neuroimaging Initiative individuals (384 [46.7%] females, mean age 73.0 ± 7.4 years), we fitted baseline CSF-Aβ42 and global Aβ-PET to a hyperbolic regression model, deriving a participant-specific Aβ-aggregation score (standardized residuals); negative values represent more soluble relative to aggregated Aβ and positive values more aggregated relative to soluble Aβ. Using linear models, we investigated whether methodological factors, demographics, CSF biomarkers, and vascular burden contributed to Aβ-aggregation scores. With linear mixed models, we assessed whether Aβ-aggregation scores were predictive of cognitive functioning. Analyses were repeated in an early independent validation cohort of 383 Amyloid Imaging to Prevent Alzheimer's Disease Prognostic and Natural History Study individuals (224 [58.5%] females, mean age 65.2 ± 6.9 years).

RESULTS

The imbalance model could be fit (pseudo-R2 = 0.94) in both cohorts, across CSF kits and PET tracers. Although no associations were observed with the main methodological factors, lower Aβ-aggregation scores were associated with larger ventricular volume (β = 0.13, p < 0.001), male sex (β = -0.18, p = 0.019), and homozygous APOE-ε4 carriership (β = -0.56, p < 0.001), whereas higher scores were associated with increased uncorrected CSF p-tau (β = 0.17, p < 0.001) and t-tau (β = 0.16, p < 0.001), better baseline executive functioning (β = 0.12, p < 0.001), and slower global cognitive decline (β = 0.14, p = 0.006). In the validation cohort, we replicated the associations with APOE-ε4, CSF t-tau, and, although modestly, with cognition.

DISCUSSION

We propose a novel continuous model of Aβ CSF/PET biomarker imbalance, accurately describing heterogeneity in soluble vs aggregated Aβ pools in 2 independent cohorts across the full Aβ continuum. Aβ-aggregation scores were consistently associated with genetic and AD-associated CSF biomarkers, possibly reflecting disease heterogeneity beyond methodological influences.

Amyloid-PET imaging predicts functional decline in clinically normal individuals

BACKGROUND

There is good evidence that elevated amyloid-β (Aβ) positron emission tomography (PET) signal is associated with cognitive decline in clinically normal (CN) individuals. However, it is less well established whether there is an association between the Aβ burden and decline in daily living activities in this population. Moreover, Aβ-PET Centiloids (CL) thresholds that can optimally predict functional decline have not yet been established.

METHODS

Cross-sectional and longitudinal analyses over a mean three-year timeframe were performed on the European amyloid-PET imaging AMYPAD-PNHS dataset that phenotypes 1260 individuals, including 1032 CN individuals and 228 participants with questionable functional impairment. Amyloid-PET was assessed continuously on the Centiloid (CL) scale and using Aβ groups (CL < 12 = Aβ-, 12 ≤ CL ≤ 50 = Aβ-intermediate/Aβ± , CL > 50 = Aβ+). Functional abilities were longitudinally assessed using the Clinical Dementia Rating (Global-CDR, CDR-SOB) and the Amsterdam Instrumental Activities of Daily Living Questionnaire (A-IADL-Q). The Global-CDR was available for the 1260 participants at baseline, while baseline CDR-SOB and A-IADL-Q scores and longitudinal functional data were available for different subsamples that had similar characteristics to those of the entire sample.

RESULTS

Participants included 765 Aβ- (61%, Mdnage = 66.0, IQRage = 61.0-71.0; 59% women), 301 Aβ± (24%; Mdnage = 69.0, IQRage = 64.0-75.0; 53% women) and 194 Aβ+ individuals (15%, Mdnage = 73.0, IQRage = 68.0-78.0; 53% women). Cross-sectionally, CL values were associated with CDR outcomes. Longitudinally, baseline CL values predicted prospective changes in the CDR-SOB (bCL*Time = 0.001/CL/year, 95% CI [0.0005,0.0024], p = .003) and A-IADL-Q (bCL*Time = -0.010/CL/year, 95% CI [-0.016,-0.004], p = .002) scores in initially CN participants. Increased clinical progression (Global-CDR > 0) was mainly observed in Aβ+ CN individuals (HRAβ+ vs Aβ- = 2.55, 95% CI [1.16,5.60], p = .020). Optimal thresholds for predicting decline were found at 41 CL using the CDR-SOB (bAβ+ vs Aβ- = 0.137/year, 95% CI [0.069,0.206], p < .001) and 28 CL using the A-IADL-Q (bAβ+ vs Aβ- = -0.693/year, 95% CI [-1.179,-0.208], p = .005).

CONCLUSIONS

Amyloid-PET quantification supports the identification of CN individuals at risk of functional decline.

TRIAL REGISTRATION

The AMYPAD PNHS is registered at www.clinicaltrialsregister.eu with the EudraCT Number: 2018-002277-22.

Cross-sectional associations between 24-hour time-use composition, grey matter volume and cognitive function in healthy older adults

BACKGROUND

Increasing physical activity (PA) is an effective strategy to slow reductions in cortical volume and maintain cognitive function in older adulthood. However, PA does not exist in isolation, but coexists with sleep and sedentary behaviour to make up the 24-hour day. We investigated how the balance of all three behaviours (24-hour time-use composition) is associated with grey matter volume in healthy older adults, and whether grey matter volume influences the relationship between 24-hour time-use composition and cognitive function.

METHODS

This cross-sectional study included 378 older adults (65.6 ± 3.0 years old, 123 male) from the ACTIVate study across two Australian sites (Adelaide and Newcastle). Time-use composition was captured using 7-day accelerometry, and T1-weighted magnetic resonance imaging was used to measure grey matter volume both globally and across regions of interest (ROI: frontal lobe, temporal lobe, hippocampi, and lateral ventricles). Pairwise correlations were used to explore univariate associations between time-use variables, grey matter volumes and cognitive outcomes. Compositional data analysis linear regression models were used to quantify associations between ROI volumes and time-use composition, and explore potential associations between the interaction between ROI volumes and time-use composition with cognitive outcomes.

RESULTS

After adjusting for covariates (age, sex, education), there were no significant associations between time-use composition and any volumetric outcomes. There were significant interactions between time-use composition and frontal lobe volume for long-term memory (p = 0.018) and executive function (p = 0.018), and between time-use composition and total grey matter volume for executive function (p = 0.028). Spending more time in moderate-vigorous PA was associated with better long-term memory scores, but only for those with smaller frontal lobe volume (below the sample mean). Conversely, spending more time in sleep and less time in sedentary behaviour was associated with better executive function in those with smaller total grey matter volume.

CONCLUSIONS

Although 24-hour time use was not associated with total or regional grey matter independently, total grey matter and frontal lobe grey matter volume moderated the relationship between time-use composition and several cognitive outcomes. Future studies should investigate these relationships longitudinally to assess whether changes in time-use composition correspond to changes in grey matter volume and cognition.

Genetic, vascular and amyloid components of cerebral blood flow in a preclinical population

Aging-related cognitive decline can be accelerated by a combination of genetic factors, cardiovascular and cerebrovascular dysfunction, and amyloid-β burden. Whereas cerebral blood flow (CBF) has been studied as a potential early biomarker of cognitive decline, its normal variability in healthy elderly is less known. In this study, we investigated the contribution of genetic, vascular, and amyloid-β components of CBF in a cognitively unimpaired (CU) population of monozygotic older twins. We included 134 participants who underwent arterial spin labeling (ASL) MRI and [18F]flutemetamol amyloid-PET imaging at baseline and after a four-year follow-up. Generalized estimating equations were used to investigate the associations of amyloid burden and white matter hyperintensities with CBF. We showed that, in CU individuals, CBF: 1) has a genetic component, as within-pair similarities in CBF values were moderate and significant (ICC > 0.40); 2) is negatively associated with cerebrovascular damage; and 3) is positively associated with the interaction between cardiovascular risk scores and early amyloid-β burden, which may reflect a vascular compensatory response of CBF to early amyloid-β accumulation. These findings encourage future studies to account for multiple interactions with CBF in disease trajectory analyses.

Spatial cognition is associated with levels of phosphorylated-tau and β-amyloid in clinically normal older adults

Spatial cognition is associated with Alzheimer's disease (AD) biomarkers in the symptomatic stages of the disease. We investigated whether cerebrospinal fluid (CSF) biomarkers (phosphorylated-tau [p-tau] and β-amyloid) are associated with poorer spatial cognition in clinically normal older adults. Participants were 1875 clinically normal adults (age 67.8 [8.5] years) from the European Prevention of Alzheimer's Dementia Consortium. Mixed effect models assessed the cross-sectional association between p-tau181, β-amyloid1-42 (Aβ1-42) and p-tau181/Aβ1-42 ratio and spatial cognition measured using semi-automated Supermarket Task and the 4 Mountains Task. Levels of p-tau181, Aβ1-42, and p-tau181/Aβ1-42 ratio were significantly associated with spatial cognition scores on both tasks. The p-tau181/Aβ1-42 ratio showed the largest effect sizes (β = -0.04/0.05, p < 0.001). Lower entorhinal cortical volume was associated with poorer outcomes on both tasks (β = 0.06, p < 0.002) and accounted for 18%-22% of the direct association between p-tau181 and spatial cognition scores. In conclusion, degeneration of the entorhinal cortex mediates a significant proportion of the association between p-tau181 and spatial assessments in cognitively normal adults. Future studies should focus on increasing the sensitivity of digital spatial assessments.

The amyloid imaging for the prevention of Alzheimer's disease consortium: A European collaboration with global impact

Background

Amyloid-β (Aβ) accumulation is considered the earliest pathological change in Alzheimer's disease (AD). The Amyloid Imaging to Prevent Alzheimer's Disease (AMYPAD) consortium is a collaborative European framework across European Federation of Pharmaceutical Industries Associations (EFPIA), academic, and 'Small and Medium-sized enterprises' (SME) partners aiming to provide evidence on the clinical utility and cost-effectiveness of Positron Emission Tomography (PET) imaging in diagnostic work-up of AD and to support clinical trial design by developing optimal quantitative methodology in an early AD population.

The AMYPAD studies

In the Diagnostic and Patient Management Study (DPMS), 844 participants from eight centres across three clinical subgroups (245 subjective cognitive decline, 342 mild cognitive impairment, and 258 dementia) were included. The Prognostic and Natural History Study (PNHS) recruited pre-dementia subjects across 11 European parent cohorts (PCs). Approximately 1600 unique subjects with historical and prospective data were collected within this study. PET acquisition with [¹⁸F]flutemetamol or [¹⁸F]florbetaben radiotracers was performed and quantified using the Centiloid (CL) method.

Results

AMYPAD has significantly contributed to the AD field by furthering our understanding of amyloid deposition in the brain and the optimal methodology to measure this process. Main contributions so far include the validation of the dual-time window acquisition protocol to derive the fully quantitative non-displaceable binding potential (BP _ND ), assess the value of this metric in the context of clinical trials, improve PET-sensitivity to emerging Aβ burden and utilize its available regional information, establish the quantitative accuracy of the Centiloid method across tracers and support implementation of quantitative amyloid-PET measures in the clinical routine.

Future steps

The AMYPAD consortium has succeeded in recruiting and following a large number of prospective subjects and setting up a collaborative framework to integrate data across European PCs. Efforts are currently ongoing in collaboration with ARIDHIA and ADDI to harmonize, integrate, and curate all available clinical data from the PNHS PCs, which will become openly accessible to the wider scientific community.

Age differences in the association between sleep and Alzheimer's disease biomarkers in the EPAD cohort

Introduction

We aimed to determine the independent association between sleep quality and Alzheimer's disease (AD) biomarkers, and whether the associations differ with age.

Methods

We included 1240 individuals aged ≥50, without dementia from the European Prevention of Alzheimer's Disease v1500.0 dataset. Linear regression was used to examine Pittsburgh Sleep Quality Index (PSQI) scores against cerebrospinal fluid (CSF) phosphorylated tau/β-amyloid ratio (p-tau/Aβ42) for the entire sample and via age tertiles. Models controlled for demographic, clinical, genetic, vascular, and neuroimaging variables.

Results

For the youngest age tertile, shorter sleep duration and higher sleep efficiency were associated with greater p-tau/Aβ42 ratio. For the oldest tertile, longer sleep latency was associated with greater p-tau/Aβ42.

Discussion

Differential relationships between sleep and AD pathology depend on age. Short sleep duration and sleep efficiency are relevant in middle age whereas time taken to fall asleep is more closely linked to AD biomarkers in later life.

Highlights

This study shows age differences in the link between sleep and AD biomarkers.Shorter sleep was associated with greater p-tau/Aβ42 ratio in middle age.The association was independent of genetic, vascular, and neuroimaging markers of AD.

Pepinemab antibody blockade of SEMA4D in early Huntington's disease: a randomized, placebo-controlled, phase 2 trial

SIGNAL is a multicenter, randomized, double-blind, placebo-controlled phase 2 study (no. NCT02481674) established to evaluate pepinemab, a semaphorin 4D (SEMA4D)-blocking antibody, for treatment of Huntington's disease (HD). The trial enrolled a total of 265 HD gene expansion carriers with either early manifest (EM, n = 179) or late prodromal (LP, n = 86) HD, randomized (1:1) to receive 18 monthly infusions of pepinemab (n = 91 EM, 41 LP) or placebo (n = 88 EM, 45 LP). Pepinemab was generally well tolerated, with a relatively low frequency of serious treatment-emergent adverse events of 5% with pepinemab compared to 9% with placebo, including both EM and LP participants. Coprimary efficacy outcome measures consisted of assessments within the EM cohort of (1) a two-item HD cognitive assessment family comprising one-touch stockings of Cambridge (OTS) and paced tapping (PTAP) and (2) clinical global impression of change (CGIC). The differences between pepinemab and placebo in mean change (95% confidence interval) from baseline at month 17 for OTS were -1.98 (-4.00, 0.05) (one-sided P = 0.028), and for PTAP 1.43 (-0.37, 3.23) (one-sided P = 0.06). Similarly, because a significant treatment effect was not observed for CGIC, the coprimary endpoint, the study did not meet its prespecified primary outcomes. Nevertheless, a number of other positive outcomes and post hoc subgroup analyses-including additional cognitive measures and volumetric magnetic resonance imaging and fluorodeoxyglucose-positron-emission tomography imaging assessments-provide rationale and direction for the design of a phase 3 study and encourage the continued development of pepinemab in patients diagnosed with EM HD.

The Effect of Self-Paced Exercise Intensity and Cardiorespiratory Fitness on Frontal Grey Matter Volume in Cognitively Normal Older Adults: A Randomised Controlled Trial

OBJECTIVE

Exercise has been found to be important in maintaining neurocognitive health. However, the effect of exercise intensity level remains relatively underexplored. Thus, to test the hypothesis that self-paced high-intensity exercise and cardiorespiratory fitness (peak aerobic capacity; VO_2peak) increase grey matter (GM) volume, we examined the effect of a 6-month exercise intervention on frontal lobe GM regions that support the executive functions in older adults.

METHODS

Ninety-eight cognitively normal participants (age = 69.06 ± 5.2 years; n = 54 female) were randomised into either a self-paced high- or moderate-intensity cycle-based exercise intervention group, or a no-intervention control group. Participants underwent magnetic resonance imaging and fitness assessment pre-intervention, immediately post-intervention, and 12-months post-intervention.

RESULTS

The intervention was found to increase fitness in the exercise groups, as compared with the control group (F = 9.88, p = <0.001). Changes in pre-to-post-intervention fitness were associated with increased volume in the right frontal lobe (β = 0.29, p = 0.036, r = 0.27), right supplementary motor area (β = 0.30, p = 0.031, r = 0.29), and both right (β = 0.32, p = 0.034, r = 0.30) and left gyrus rectus (β = 0.30, p = 0.037, r = 0.29) for intervention, but not control participants. No differences in volume were observed across groups.

CONCLUSIONS

At an aggregate level, six months of self-paced high- or moderate-intensity exercise did not increase frontal GM volume. However, experimentally-induced changes in individual cardiorespiratory fitness was positively associated with frontal GM volume in our sample of older adults. These results provide evidence of individual variability in exercise-induced fitness on brain structure.

The relationship between chronic PTSD, cortical volumetry and white matter microstructure among Australian combat veterans

BACKGROUND

Posttraumatic stress disorder (PTSD) has been associated with volumetric and white matter microstructural changes among general and veteran populations. However, regions implicated have greatly varied and often conflict between studies, potentially due to confounding comorbidities within samples. This study compared grey matter volume and white matter microstructure among Australian combat veterans with and without a lifetime diagnosis of PTSD, in a homogenous sample assessed for known confounding comorbidities.

METHODS

Sixty-eight male trauma-exposed veterans (16 PTSD-diagnosed; mean age 69 years) completed a battery of psychometric assessments and underwent magnetic resonance and diffusion tensor imaging. Analyses included tract-based spatial statistics, voxel-wise analyses, diffusion connectome-based group-wise analysis, and volumetric analysis.

RESULTS

Significantly smaller grey matter volumes were observed in the left prefrontal cortex (P = 0.026), bilateral middle frontal gyrus (P = 0.021), and left anterior insula (P = 0.048) in the PTSD group compared to controls. Significant negative correlations were found between PTSD symptom severity and fractional anisotropy values in the left corticospinal tract (R² = 0.34, P = 0.024) and left inferior cerebellar peduncle (R² = 0.62, P = 0.016). No connectome-based differences in white matter properties were observed.

CONCLUSIONS

Findings from this study reinforce reports of white matter alterations, as indicated by reduced fractional anisotropy values, in relation to PTSD symptom severity, as well as patterns of reduced volume in the prefrontal cortex. These results contribute to the developing profile of neuroanatomical differences uniquely attributable to veterans who suffer from chronic PTSD.

Regional associations of white matter hyperintensities and early cortical amyloid pathology

White matter hyperintensities (WMHs) have a heterogeneous aetiology, associated with both vascular risk factors and amyloidosis due to Alzheimer's disease. While spatial distribution of both amyloid and WM lesions carry important information for the underlying pathogenic mechanisms, the regional relationship between these two pathologies and their joint contribution to early cognitive deterioration remains largely unexplored. We included 662 non-demented participants from three Amyloid Imaging to Prevent Alzheimer's disease (AMYPAD)-affiliated cohorts: EPAD-LCS (N = 176), ALFA+ (N = 310), and EMIF-AD PreclinAD Twin60++ (N = 176). Using PET imaging, cortical amyloid burden was assessed regionally within early accumulating regions (medial orbitofrontal, precuneus, and cuneus) and globally, using the Centiloid method. Regional WMH volume was computed using Bayesian Model Selection. Global associations between WMH, amyloid, and cardiovascular risk scores (Framingham and CAIDE) were assessed using linear models. Partial least square (PLS) regression was used to identify regional associations. Models were adjusted for age, sex, and APOE-e4 status. Individual PLS scores were then related to cognitive performance in 4 domains (attention, memory, executive functioning, and language). While no significant global association was found, the PLS model yielded two components of interest. In the first PLS component, a fronto-parietal WMH pattern was associated with medial orbitofrontal-precuneal amyloid, vascular risk, and age. Component 2 showed a posterior WMH pattern associated with precuneus-cuneus amyloid, less related to age or vascular risk. Component 1 was associated with lower performance in all cognitive domains, while component 2 only with worse memory. In a large pre-dementia population, we observed two distinct patterns of regional associations between WMH and amyloid burden, and demonstrated their joint influence on cognitive processes. These two components could reflect the existence of vascular-dependent and -independent manifestations of WMH-amyloid regional association that might be related to distinct primary pathophysiology.

Reduced cortical cholinergic innervation measured using [18F]-FEOBV PET imaging correlates with cognitive decline in mild cognitive impairment

•

Topographic FEOBV binding correlates with domain-specific cognitive performance.

•

Global and regional reductions in cholinergic innervation are observed in MCI.

•

Global FEOBV SUVR is associated with basal forebrain and hippocampal volumes.

•

Our results provide proof of concept for FEOBV PET to assess cholinergic terminal integrity.

Dysfunction of the cholinergic basal forebrain (BF) neurotransmitter system, including cholinergic axon denervation of the cortex, plays an important role in cognitive decline and dementia. A validated method to directly quantify cortical cholinergic terminal integrity enables exploration of the involvement of this system in diverse cognitive profiles associated with dementia, particularly at a prodromal stage. In this study, we used the radiotracer [¹⁸F]-fluoroethoxybenzovesamicol (FEOBV) as a direct measure of cholinergic terminal integrity and investigated its value for the assessment of cholinergic denervation in the cortex and associated cognitive deficits. Eighteen participants (8 with mild cognitive impairment (MCI) and 10 cognitively unimpaired controls) underwent neuropsychological assessment and brain imaging using FEOBV and [¹⁸F]-florbetaben for amyloid-β imaging. The MCI group showed a significant global reduction of FEOBV retention in the cortex and in the parietal and occipital cortices specifically compared to the control group. The global cortical FEOBV retention of all participants positively correlated with the BF, hippocampus and grey matter volumes, but no association was found between the global FEOBV retention and amyloid-β status. Topographic profiles from voxel-wise analysis of FEOBV images revealed significant positive correlations with the cognitive domains associated with the underlying cortical areas. Overlapping profiles of decreased FEOBV were identified in correlation with impairment in executive function, attention and language, which covered the anterior cingulate gyrus, olfactory cortex, calcarine cortex, middle temporal gyrus and caudate nucleus. However, the absence of cortical atrophy in these areas suggested that reduced cholinergic terminal integrity in the cortex is an important factor underlying the observed cognitive decline in early dementia. Our results provide support for the utility and validity of FEOBV PET for quantitative assessment of region-specific cholinergic terminal integrity that could potentially be used for early detection of cholinergic dysfunction in dementia following further validation in larger cohorts.

Impact of cerebral blood flow and amyloid load on SUVR bias

Background

Despite its widespread use, the semi-quantitative standardized uptake value ratio (SUVR) may be biased compared with the distribution volume ratio (DVR). This bias may be partially explained by changes in cerebral blood flow (CBF) and is likely to be also dependent on the extent of the underlying amyloid-β (Aβ) burden. This study aimed to compare SUVR with DVR and to evaluate the effects of underlying Aβ burden and CBF on bias in SUVR in mainly cognitively unimpaired participants. Participants were scanned according to a dual-time window protocol, with either [¹⁸F]flutemetamol (N = 90) or [¹⁸F]florbetaben (N = 31). The validated basisfunction-based implementation of the two-step simplified reference tissue model was used to derive DVR and R₁ parametric images, and SUVR was calculated from 90 to 110 min post-injection, all with the cerebellar grey matter as reference tissue. First, linear regression and Bland–Altman analyses were used to compare (regional) SUVR with DVR. Then, generalized linear models were applied to evaluate whether (bias in) SUVR relative to DVR could be explained by R₁ for the global cortical average (GCA), precuneus, posterior cingulate, and orbitofrontal region.

Results

Despite high correlations (GCA: R² ≥ 0.85), large overestimation and proportional bias of SUVR relative to DVR was observed. Negative associations were observed between both SUVR or SUVR_bias and R_1, albeit non-significant.

Conclusion

The present findings demonstrate that bias in SUVR relative to DVR is strongly related to underlying Aβ burden. Furthermore, in a cohort consisting mainly of cognitively unimpaired individuals, the effect of relative CBF on bias in SUVR appears limited. EudraCT Number: 2018-002277-22, registered on: 25-06-2018.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13550-022-00898-8.

Differences Between MR Brain Region Segmentation Methods: Impact on Single-Subject Analysis

For the segmentation of magnetic resonance brain images into anatomical regions, numerous fully automated methods have been proposed and compared to reference segmentations obtained manually. However, systematic differences might exist between the resulting segmentations, depending on the segmentation method and underlying brain atlas. This potentially results in sensitivity differences to disease and can further complicate the comparison of individual patients to normative data. In this study, we aim to answer two research questions: 1) to what extent are methods interchangeable, as long as the same method is being used for computing normative volume distributions and patient-specific volumes? and 2) can different methods be used for computing normative volume distributions and assessing patient-specific volumes? To answer these questions, we compared volumes of six brain regions calculated by five state-of-the-art segmentation methods: Erasmus MC (EMC), FreeSurfer (FS), geodesic information flows (GIF), multi-atlas label propagation with expectation–maximization (MALP-EM), and model-based brain segmentation (MBS). We applied the methods on 988 non-demented (ND) subjects and computed the correlation (PCC-v) and absolute agreement (ICC-v) on the volumes. For most regions, the PCC-v was good (>0.75), indicating that volume differences between methods in ND subjects are mainly due to systematic differences. The ICC-v was generally lower, especially for the smaller regions, indicating that it is essential that the same method is used to generate normative and patient data. To evaluate the impact on single-subject analysis, we also applied the methods to 42 patients with Alzheimer’s disease (AD). In the case where the normative distributions and the patient-specific volumes were calculated by the same method, the patient’s distance to the normative distribution was assessed with the z-score. We determined the diagnostic value of this z-score, which showed to be consistent across methods. The absolute agreement on the AD patients’ z-scores was high for regions of thalamus and putamen. This is encouraging as it indicates that the studied methods are interchangeable for these regions. For regions such as the hippocampus, amygdala, caudate nucleus and accumbens, and globus pallidus, not all method combinations showed a high ICC-z. Whether two methods are indeed interchangeable should be confirmed for the specific application and dataset of interest.

Disease Modelling of Cognitive Outcomes and Biomarkers in the European Prevention of Alzheimer's Dementia Longitudinal Cohort

A key challenge for the secondary prevention of Alzheimer’s dementia is the need to identify individuals early on in the disease process through sensitive cognitive tests and biomarkers. The European Prevention of Alzheimer’s Dementia (EPAD) consortium recruited participants into a longitudinal cohort study with the aim of building a readiness cohort for a proof-of-concept clinical trial and also to generate a rich longitudinal data-set for disease modelling. Data have been collected on a wide range of measurements including cognitive outcomes, neuroimaging, cerebrospinal fluid biomarkers, genetics and other clinical and environmental risk factors, and are available for 1,828 eligible participants at baseline, 1,567 at 6 months, 1,188 at one-year follow-up, 383 at 2 years, and 89 participants at three-year follow-up visit. We novelly apply state-of-the-art longitudinal modelling and risk stratification approaches to these data in order to characterise disease progression and biological heterogeneity within the cohort. Specifically, we use longitudinal class-specific mixed effects models to characterise the different clinical disease trajectories and a semi-supervised Bayesian clustering approach to explore whether participants can be stratified into homogeneous subgroups that have different patterns of cognitive functioning evolution, while also having subgroup-specific profiles in terms of baseline biomarkers and longitudinal rate of change in biomarkers.

Safety and efficacy of anti-tau monoclonal antibody gosuranemab in progressive supranuclear palsy: a phase 2, randomized, placebo-controlled trial

A randomized, double-blind, placebo-controlled, 52-week study (no. NCT03068468) evaluated gosuranemab, an anti-tau monoclonal antibody, in the treatment of progressive supranuclear palsy (PSP). In total, 486 participants dosed were assigned to either gosuranemab (n = 321) or placebo (n = 165). Efficacy was not demonstrated on adjusted mean change of PSP Rating Scale score at week 52 between gosuranemab and placebo (10.4 versus 10.6, P = 0.85, primary endpoint), or at secondary endpoints, resulting in discontinuation of the open-label, long-term extension. Unbound N-terminal tau in cerebrospinal fluid decreased by 98% with gosuranemab and increased by 11% with placebo (P < 0.0001). Incidences of adverse events and deaths were similar between groups. This well-powered study suggests that N-terminal tau neutralization does not translate into clinical efficacy.

Split-Attention U-Net: A Fully Convolutional Network for Robust Multi-Label Segmentation from Brain MRI

Multi-label brain segmentation from brain magnetic resonance imaging (MRI) provides valuable structural information for most neurological analyses. Due to the complexity of the brain segmentation algorithm, it could delay the delivery of neuroimaging findings. Therefore, we introduce Split-Attention U-Net (SAU-Net), a convolutional neural network with skip pathways and a split-attention module that segments brain MRI scans. The proposed architecture employs split-attention blocks, skip pathways with pyramid levels, and evolving normalization layers. For efficient training, we performed pre-training and fine-tuning with the original and manually modified FreeSurfer labels, respectively. This learning strategy enables involvement of heterogeneous neuroimaging data in the training without the need for many manual annotations. Using nine evaluation datasets, we demonstrated that SAU-Net achieved better segmentation accuracy with better reliability that surpasses those of state-of-the-art methods. We believe that SAU-Net has excellent potential due to its robustness to neuroanatomical variability that would enable almost instantaneous access to accurate neuroimaging biomarkers and its swift processing runtime compared to other methods investigated.

One Shot Segmentation: Unifying Rigid Detection and Non-Rigid Segmentation Using Elastic Regularization

This paper proposes a novel approach for the non-rigid segmentation of deformable objects in image sequences, which is based on one-shot segmentation that unifies rigid detection and non-rigid segmentation using elastic regularization. The domain of application is the segmentation of a visual object that temporally undergoes a rigid transformation (e.g., affine transformation) and a non-rigid transformation (i.e., contour deformation). The majority of segmentation approaches to solve this problem are generally based on two steps that run in sequence: a rigid detection, followed by a non-rigid segmentation. In this paper, we propose a new approach, where both the rigid and non-rigid segmentation are performed in a single shot using a sparse low-dimensional manifold that represents the visual object deformations. Given the multi-modality of these deformations, the manifold partitions the training data into several patches, where each patch provides a segmentation proposal during the inference process. These multiple segmentation proposals are merged using the classification results produced by deep belief networks (DBN) that compute the confidence on each segmentation proposal. Thus, an ensemble of DBN classifiers is used for estimating the final segmentation. Compared to current methods proposed in the field, our proposed approach is advantageous in four aspects: (i) it is a unified framework to produce rigid and non-rigid segmentations; (ii) it uses an ensemble classification process, which can help the segmentation robustness; (iii) it provides a significant reduction in terms of the number of dimensions of the rigid and non-rigid segmentations search spaces, compared to current approaches that divide these two problems; and (iv) this lower dimensionality of the search space can also reduce the need for large annotated training sets to be used for estimating the DBN models. Experiments on the problem of left ventricle endocardial segmentation from ultrasound images, and lip segmentation from frontal facial images using the extended Cohn-Kanade (CK+) database, demonstrate the potential of the methodology through qualitative and quantitative evaluations, and the ability to reduce the search and training complexities without a significant impact on the segmentation accuracy.

Inter-method reliability between automatic region of interest analytic application with multi-atlas segmentation and FreeSurfer

AIM

Alzheimer's disease (AD) is a progressive neurodegenerative disease characterized by the aggregation of amyloid-β and phosphorylated tau proteins. Magnetic resonance imaging (MRI) is a useful means of detecting hippocampal atrophy. However, instead of visual inspection, objective and time-saving tools for automated region of interest (ROI) analysis are needed. Advances in MRI segmentation techniques have enabled a multi-atlas approach with fewer errors than a conventional single-atlas approach. To support the clinical application of multi-atlas segmentation, an automated ROI analytic application consisting of multi-atlas segmentation with joint label fusion and corrective learning was developed: T-Proto. In the present study, we evaluated the inter-method reliability between T-Proto and a reference ROI analytic software, FreeSurfer.

METHODS

This was a database study. MRI data from 30 patients with AD were selected, and the inter-method reliability was assessed in terms of the intra-class correlation coefficient (ICC). A post-hoc comparison according to the severity of AD was also performed.

RESULTS

Almost all the regional volumes estimated with T-Proto were smaller than those estimated with FreeSurfer. The regional ICC values between the two methods showed moderate to excellent reliability. A post-hoc comparison revealed a similar t-value and effect size between both methods for the hippocampus.

CONCLUSION

In the present study, we showed that automated regional analysis using T-Proto was reliable in the hippocampus in terms of ICC, compared with FreeSurfer.

MRI Segmentation and Classification of Human Brain Using Deep Learning for Diagnosis of Alzheimer's Disease: A Survey

Many neurological diseases and delineating pathological regions have been analyzed, and the anatomical structure of the brain researched with the aid of magnetic resonance imaging (MRI). It is important to identify patients with Alzheimer's disease (AD) early so that preventative measures can be taken. A detailed analysis of the tissue structures from segmented MRI leads to a more accurate classification of specific brain disorders. Several segmentation methods to diagnose AD have been proposed with varying complexity. Segmentation of the brain structure and classification of AD using deep learning approaches has gained attention as it can provide effective results over a large set of data. Hence, deep learning methods are now preferred over state-of-the-art machine learning methods. We aim to provide an outline of current deep learning-based segmentation approaches for the quantitative analysis of brain MRI for the diagnosis of AD. Here, we report how convolutional neural network architectures are used to analyze the anatomical brain structure and diagnose AD, discuss how brain MRI segmentation improves AD classification, describe the state-of-the-art approaches, and summarize their results using publicly available datasets. Finally, we provide insight into current issues and discuss possible future research directions in building a computer-aided diagnostic system for AD.

Quantitative amyloid PET in Alzheimer's disease: the AMYPAD prognostic and natural history study

INTRODUCTION

The Amyloid Imaging to Prevent Alzheimer's Disease (AMYPAD) Prognostic and Natural History Study (PNHS) aims at understanding the role of amyloid imaging in the earliest stages of Alzheimer's disease (AD). AMYPAD PNHS adds (semi-)quantitative amyloid PET imaging to several European parent cohorts (PCs) to predict AD-related progression as well as address methodological challenges in amyloid PET.

METHODS

AMYPAD PNHS is an open-label, prospective, multi-center, cohort study recruiting from multiple PCs. Around 2000 participants will undergo baseline amyloid positron emission tomography (PET), half of whom will be invited for a follow-up PET 12 at least 12 months later.

RESULTS

Primary include several amyloid PET measurements (Centiloid, SUVr, BP_ND , R₁ ), and secondary are their changes from baseline, relationship to other amyloid markers (cerebrospinal fluid and visual assessment), and predictive value of AD-related decline.

EXPECTED IMPACT

Determining the role of amyloid PET for the understanding of this complex disease and potentially improving secondary prevention trials.

Translating research findings into clinical practice: a systematic and critical review of neuroimaging-based clinical tools for brain disorders

A pivotal aim of psychiatric and neurological research is to promote the translation of the findings into clinical practice to improve diagnostic and prognostic assessment of individual patients. Structural neuroimaging holds much promise, with neuroanatomical measures accounting for up to 40% of the variance in clinical outcome. Building on these findings, a number of imaging-based clinical tools have been developed to make diagnostic and prognostic inferences about individual patients from their structural Magnetic Resonance Imaging scans. This systematic review describes and compares the technical characteristics of the available tools, with the aim to assess their translational potential into real-world clinical settings. The results reveal that a total of eight tools. All of these were specifically developed for neurological disorders, and as such are not suitable for application to psychiatric disorders. Furthermore, most of the tools were trained and validated in a single dataset, which can result in poor generalizability, or using a small number of individuals, which can cause overoptimistic results. In addition, all of the tools rely on two strategies to detect brain abnormalities in single individuals, one based on univariate comparison, and the other based on multivariate machine-learning algorithms. We discuss current barriers to the adoption of these tools in clinical practice and propose a checklist of pivotal characteristics that should be included in an "ideal" neuroimaging-based clinical tool for brain disorders.

An Evaluation of Atlas Selection Methods for Atlas-Based Automatic Segmentation in Radiotherapy Treatment Planning

Atlas-based automatic segmentation is used in radiotherapy planning to accelerate the delineation of organs at risk (OARs). Atlas selection has been proposed as a way to improve the accuracy and execution time of segmentation, assuming that, the more similar the atlas is to the patient, the better the results will be. This paper presents an analysis of atlas selection methods in the context of radiotherapy treatment planning. For a range of commonly contoured OARs, a thorough comparison of a large class of typical atlas selection methods has been performed. For this evaluation, clinically contoured CT images of the head and neck ( N=316 ) and thorax ( N=280 ) were used. The state-of-the-art intensity and deformation similarity-based atlas selection methods were found to compare poorly to perfect atlas selection. Counter-intuitively, atlas selection methods based on a fixed set of representative atlases outperformed atlas selection methods based on the patient image. This study suggests that atlas-based segmentation with currently available selection methods compares poorly to the potential best performance, hampering the clinical utility of atlas-based segmentation. Effective atlas selection remains an open challenge in atlas-based segmentation for radiotherapy planning.

Dynamic multiatlas selection-based consensus segmentation of head and neck structures from CT images

PURPOSE

Manual delineation of head and neck (H&N) organ-at-risk (OAR) structures for radiation therapy planning is time consuming and highly variable. Therefore, we developed a dynamic multiatlas selection-based approach for fast and reproducible segmentation.

METHODS

Our approach dynamically selects and weights the appropriate number of atlases for weighted label fusion and generates segmentations and consensus maps indicating voxel-wise agreement between different atlases. Atlases were selected for a target as those exceeding an alignment weight called dynamic atlas attention index. Alignment weights were computed at the image level and called global weighted voting (GWV) or at the structure level and called structure weighted voting (SWV) by using a normalized metric computed as the sum of squared distances of computed tomography (CT)-radiodensity and modality-independent neighborhood descriptors (extracting edge information). Performance comparisons were performed using 77 H&N CT images from an internal Memorial Sloan-Kettering Cancer Center dataset (N = 45) and an external dataset (N = 32) using Dice similarity coefficient (DSC), Hausdorff distance (HD), 95th percentile of HD, median of maximum surface distance, and volume ratio error against expert delineation. Pairwise DSC accuracy comparisons of proposed (GWV, SWV) vs single best atlas (BA) or majority voting (MV) methods were performed using Wilcoxon rank-sum tests.

RESULTS

Both SWV and GWV methods produced significantly better segmentation accuracy than BA (P < 0.001) and MV (P < 0.001) for all OARs within both datasets. SWV generated the most accurate segmentations with DSC of: 0.88 for oral cavity, 0.85 for mandible, 0.84 for cord, 0.76 for brainstem and parotids, 0.71 for larynx, and 0.60 for submandibular glands. SWV's accuracy exceeded GWV's for submandibular glands (DSC = 0.60 vs 0.52, P = 0.019).

CONCLUSIONS

The contributed SWV and GWV methods generated more accurate automated segmentations than the other two multiatlas-based segmentation techniques. The consensus maps could be combined with segmentations to visualize voxel-wise consensus between atlases within OARs during manual review.

Imaging markers associated with the development of post-stroke depression and apathy: Results of the Cognition and Affect after Stroke - a Prospective Evaluation of Risks study

Introduction

It has been suggested that the development of post-stroke apathy (PSA) and depression (PSD) may be more strongly associated with generalised brain pathology, rather than the stroke lesion itself. The present study aimed to investigate associations between imaging markers of lesion-related and generalised brain pathology and the development of PSA and PSD during a one-year follow-up.

Patients and methods

In a prospective cohort study, 188 stroke patients received 3-Tesla MRI at baseline (three months post-stroke) for evaluation of lesion-related, vascular, and degenerative brain pathology. Presence of lacunes, microbleeds, white matter hyperintensities, and enlarged perivascular spaces was summed to provide a measure of total cerebral small vessel disease (cSVD) burden (range 0-4). The Mini International Neuropsychiatric Interview and Apathy Evaluation Scale were administered at baseline and repeated at 6- and 12-month follow-up to define presence of PSD and PSA, respectively.

Results

Population-averaged logistic regression models showed that global brain atrophy and severe cSVD burden (score 3-4) were significantly associated with the odds of having PSA (OR_GEE 5.33, 95% CI 1.99-14.25 and 3.04, 95% CI 1.20-7.69, respectively), independent of stroke lesion volume and co-morbid PSD. Medium cSVD burden (score 2) was significantly associated with the odds of having PSD (OR_GEE 2.92, 95% CI 1.09-7.78), independent of stroke lesion volume, co-morbid PSA, and pre-stroke depression. No associations were found with lesion-related markers.

Conclusions

The results suggest that generalised degenerative and vascular brain pathology, rather than lesion-related pathology, is an important predictor for the development of PSA, and less strongly for PSD.

3D whole brain segmentation using spatially localized atlas network tiles

Detailed whole brain segmentation is an essential quantitative technique in medical image analysis, which provides a non-invasive way of measuring brain regions from a clinical acquired structural magnetic resonance imaging (MRI). Recently, deep convolution neural network (CNN) has been applied to whole brain segmentation. However, restricted by current GPU memory, 2D based methods, downsampling based 3D CNN methods, and patch-based high-resolution 3D CNN methods have been the de facto standard solutions. 3D patch-based high resolution methods typically yield superior performance among CNN approaches on detailed whole brain segmentation (>100 labels), however, whose performance are still commonly inferior compared with state-of-the-art multi-atlas segmentation methods (MAS) due to the following challenges: (1) a single network is typically used to learn both spatial and contextual information for the patches, (2) limited manually traced whole brain volumes are available (typically less than 50) for training a network. In this work, we propose the spatially localized atlas network tiles (SLANT) method to distribute multiple independent 3D fully convolutional networks (FCN) for high-resolution whole brain segmentation. To address the first challenge, multiple spatially distributed networks were used in the SLANT method, in which each network learned contextual information for a fixed spatial location. To address the second challenge, auxiliary labels on 5111 initially unlabeled scans were created by multi-atlas segmentation for training. Since the method integrated multiple traditional medical image processing methods with deep learning, we developed a containerized pipeline to deploy the end-to-end solution. From the results, the proposed method achieved superior performance compared with multi-atlas segmentation methods, while reducing the computational time from >30 h to 15 min. The method has been made available in open source (https://github.com/MASILab/SLANTbrainSeg).

Adaptive Bayesian label fusion using kernel-based similarity metrics in hippocampus segmentation

The effectiveness of brain magnetic resonance imaging (MRI) as a useful evaluation tool strongly depends on the performed segmentation of associated tissues or anatomical structures. We introduce an enhanced brain segmentation approach of Bayesian label fusion that includes the construction of adaptive target-specific probabilistic priors using atlases ranked by kernel-based similarity metrics to deal with the anatomical variability of collected MRI data. In particular, the developed segmentation approach appraises patch-based voxel representation to enhance the voxel embedding in spaces with increased tissue discrimination, as well as the construction of a neighborhood-dependent model that addresses the label assignment of each region with a different patch complexity. To measure the similarity between the target and training atlases, we propose a tensor-based kernel metric that also includes the training labeling set. We evaluate the proposed approach, adaptive Bayesian label fusion using kernel-based similarity metrics, in the specific case of hippocampus segmentation of five benchmark MRI collections, including ADNI dataset, resulting in an increased performance (assessed through the Dice index) as compared to other recent works.

Can Atlas-Based Auto-Segmentation Ever Be Perfect? Insights From Extreme Value Theory

Atlas-based segmentation is used in radiotherapy planning to accelerate the delineation of organs at risk (OARs). Atlas selection has been proposed to improve the performance of segmentation, assuming that the more similar the atlas is to the patient, the better the result. It follows that the larger the database of atlases from which to select, the better the results should be. This paper seeks to estimate a clinically achievable expected performance under this assumption. Assuming a perfect atlas selection, an extreme value theory has been applied to estimate the accuracy of single-atlas and multi-atlas segmentation given a large database of atlases. For this purpose, clinical contours of most common OARs on computed tomography of the head and neck ( N=316 ) and thoracic ( N=280 ) cases were used. This paper found that while for most organs, perfect segmentation cannot be reasonably expected, auto-contouring performance of a level corresponding to clinical quality could be consistently expected given a database of 5000 atlases under the assumption of perfect atlas selection.

Clinical validation of automated hippocampal segmentation in temporal lobe epilepsy

OBJECTIVE

To provide a multi-atlas framework for automated hippocampus segmentation in temporal lobe epilepsy (TLE) and clinically validate the results with respect to surgical lateralization and post-surgical outcome.

METHODS

We retrospectively identified 47 TLE patients who underwent surgical resection and 12 healthy controls. T1-weighted 3 T MRI scans were acquired for all subjects, and patients were identified by a neuroradiologist with regards to lateralization and degree of hippocampal sclerosis (HS). Automated segmentation was implemented through the Joint Label Fusion/Corrective Learning (JLF/CL) method. Gold standard lateralization was determined from the surgically resected side in Engel I (seizure-free) patients at the two-year timepoint. ROC curves were used to identify appropriate thresholds for hippocampal asymmetry ratios, which were then used to analyze JLF/CL lateralization.

RESULTS

The optimal template atlas based on subject images with varying appearances, from normal-appearing to severe HS, was demonstrated to be composed entirely of normal-appearing subjects, with good agreement between automated and manual segmentations. In applying this atlas to 26 surgically resected seizure-free patients at a two-year timepoint, JLF/CL lateralized seizure onset 92% of the time. In comparison, neuroradiology reads lateralized 65% of patients, but correctly lateralized seizure onset in these patients 100% of the time. When compared to lateralized neuroradiology reads, JLF/CL was in agreement and correctly lateralized all 17 patients. When compared to nonlateralized radiology reads, JLF/CL correctly lateralized 78% of the nine patients.

SIGNIFICANCE

While a neuroradiologist's interpretation of MR imaging is a key, albeit imperfect, diagnostic tool for seizure localization in medically-refractory TLE patients, automated hippocampal segmentation may provide more efficient and accurate epileptic foci localization. These promising findings demonstrate the clinical utility of automated segmentation in the TLE MR imaging pipeline prior to surgical resection, and suggest that further investigation into JLF/CL-assisted MRI reading could improve clinical outcomes. Our JLF/CL software is publicly available at https://www.nitrc.org/projects/ashs/.

Structural brain imaging in Alzheimer's disease and mild cognitive impairment: biomarker analysis and shared morphometry database

Magnetic resonance (MR) imaging is a powerful technique for non-invasive in-vivo imaging of the human brain. We employed a recently validated method for robust cross-sectional and longitudinal segmentation of MR brain images from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort. Specifically, we segmented 5074 MR brain images into 138 anatomical regions and extracted time-point specific structural volumes and volume change during follow-up intervals of 12 or 24 months. We assessed the extracted biomarkers by determining their power to predict diagnostic classification and by comparing atrophy rates to published meta-studies. The approach enables comprehensive analysis of structural changes within the whole brain. The discriminative power of individual biomarkers (volumes/atrophy rates) is on par with results published by other groups. We publish all quality-checked brain masks, structural segmentations, and extracted biomarkers along with this article. We further share the methodology for brain extraction (pincram) and segmentation (MALPEM, MALPEM4D) as open source projects with the community. The identified biomarkers hold great potential for deeper analysis, and the validated methodology can readily be applied to other imaging cohorts.

Multi-Atlas Segmentation Using Partially Annotated Data: Methods and Annotation Strategies

Multi-atlas segmentation is a widely used tool in medical image analysis, providing robust and accurate results by learning from annotated atlas datasets. However, the availability of fully annotated atlas images for training is limited due to the time required for the labelling task. Segmentation methods requiring only a proportion of each atlas image to be labelled could therefore reduce the workload on expert raters tasked with annotating atlas images. To address this issue, we first re-examine the labelling problem common in many existing approaches and formulate its solution in terms of a Markov Random Field energy minimisation problem on a graph connecting atlases and the target image. This provides a unifying framework for multi-atlas segmentation. We then show how modifications in the graph configuration of the proposed framework enable the use of partially annotated atlas images and investigate different partial annotation strategies. The proposed method was evaluated on two Magnetic Resonance Imaging (MRI) datasets for hippocampal and cardiac segmentation. Experiments were performed aimed at (1) recreating existing segmentation techniques with the proposed framework and (2) demonstrating the potential of employing sparsely annotated atlas data for multi-atlas segmentation.

Warping an atlas derived from serial histology to 5 high-resolution MRIs

Previous work from our group demonstrated the use of multiple input atlases to a modified multi-atlas framework (MAGeT-Brain) to improve subject-based segmentation accuracy. Currently, segmentation of the striatum, globus pallidus and thalamus are generated from a single high-resolution and -contrast MRI atlas derived from annotated serial histological sections. Here, we warp this atlas to five high-resolution MRI templates to create five de novo atlases. The overall goal of this work is to use these newly warped atlases as input to MAGeT-Brain in an effort to consolidate and improve the workflow presented in previous manuscripts from our group, allowing for simultaneous multi-structure segmentation. The work presented details the methodology used for the creation of the atlases using a technique previously proposed, where atlas labels are modified to mimic the intensity and contrast profile of MRI to facilitate atlas-to-template nonlinear transformation estimation. Dice's Kappa metric was used to demonstrate high quality registration and segmentation accuracy of the atlases. The final atlases are available at https://github.com/CobraLab/atlases/tree/master/5-atlas-subcortical.

Clinical validity of medial temporal atrophy as a biomarker for Alzheimer's disease in the context of a structured 5-phase development framework

Research criteria for Alzheimer's disease recommend the use of biomarkers for diagnosis, but whether biomarkers improve the diagnosis in clinical routine has not been systematically assessed. The aim is to evaluate the evidence for use of medial temporal lobe atrophy (MTA) as a biomarker for Alzheimer's disease at the mild cognitive impairment stage in routine clinical practice, with an adapted version of the 5-phase oncology framework for biomarker development. A literature review on visual assessment of MTA and hippocampal volumetry was conducted with other biomarkers addressed in parallel reviews. Ample evidence is available for phase 1 (rationale for use) and phase 2 (discriminative ability between diseased and control subjects). Phase 3 (early detection ability) is partly achieved: most evidence is derived from research cohorts or clinical populations with short follow-up, but validation in clinical mild cognitive impairment cohorts is required. In phase 4, only the practical feasibility has been addressed for visual rating of MTA. The rest of phase 4 and phase 5 have not yet been addressed.

Brain Image Labeling Using Multi-atlas Guided 3D Fully Convolutional Networks

Automatic labeling of anatomical structures in brain images plays an important role in neuroimaging analysis. Among all methods, multi-atlas based segmentation methods are widely used, due to their robustness in propagating prior label information. However, non-linear registration is always needed, which is time-consuming. Alternatively, the patch-based methods have been proposed to relax the requirement of image registration, but the labeling is often determined independently by the target image information, without getting direct assistance from the atlases. To address these limitations, in this paper, we propose a multi-atlas guided 3D fully convolutional networks (FCN) for brain image labeling. Specifically, multi-atlas based guidance is incorporated during the network learning. Based on this, the discriminative of the FCN is boosted, which eventually contribute to accurate prediction. Experiments show that the use of multi-atlas guidance improves the brain labeling performance.

Your algorithm might think the hippocampus grows in Alzheimer's disease: Caveats of longitudinal automated hippocampal volumetry

Hippocampal atrophy rate-measured using automated techniques applied to structural MRI scans-is considered a sensitive marker of disease progression in Alzheimer's disease, frequently used as an outcome measure in clinical trials. Using publicly accessible data from the Alzheimer's Disease Neuroimaging Initiative (ADNI), we examined 1-year hippocampal atrophy rates generated by each of five automated or semiautomated hippocampal segmentation algorithms in patients with Alzheimer's disease, subjects with mild cognitive impairment, or elderly controls. We analyzed MRI data from 398 and 62 subjects available at baseline and at 1 year at MRI field strengths of 1.5 T and 3 T, respectively. We observed a high rate of hippocampal segmentation failures across all algorithms and diagnostic categories, with only 50.8% of subjects at 1.5 T and 58.1% of subjects at 3 T passing stringent segmentation quality control. We also found that all algorithms identified several subjects (between 2.94% and 48.68%) across all diagnostic categories showing increases in hippocampal volume over 1 year. For any given algorithm, hippocampal "growth" could not entirely be explained by excluding patients with flawed hippocampal segmentations, scan-rescan variability, or MRI field strength. Furthermore, different algorithms did not uniformly identify the same subjects as hippocampal "growers," and showed very poor concordance in estimates of magnitude of hippocampal volume change over time (intraclass correlation coefficient 0.319 at 1.5 T and 0.149 at 3 T). This precluded a meaningful analysis of whether hippocampal "growth" represents a true biological phenomenon. Taken together, our findings suggest that longitudinal hippocampal volume change should be interpreted with considerable caution as a biomarker. Hum Brain Mapp 38:2875-2896, 2017. © 2017 Wiley Periodicals, Inc.

The Relationship between Cerebral Small Vessel Disease, Hippocampal Volume and Cognitive Functioning in Patients with COPD: An MRI Study

The neural correlates of cognitive impairment in chronic obstructive pulmonary disease (COPD) are not yet understood. Structural brain abnormalities could possibly be associated with the presence of cognitive impairment through cigarette smoke, inflammation, vascular disease, or hypoxemia in these patients. This study aimed to investigate whether macrostructural brain magnetic resonance imaging (MRI) features of cerebral small vessel disease (SVD) and hippocampal volume (HCV) are related to cognitive performance in patients with COPD. A subgroup of cognitively high and low-performing COPD patients of the COgnitive-PD study, underwent a brain 3T MRI. SVD as a marker of vascular damage was assessed using qualitative visual rating scales. HCV as a marker of neurodegeneration was assessed using the learning embedding for atlas propagation (LEAP) method. Features of SVD and HCV were compared between cognitively high and low-performing individuals using Mann Whitney U tests and independent samples t-tests, respectively. No group differences were reported between 25 high-performing (mean age 60.3 (standard deviation [SD] 9.7) years; 40.0% men; forced expiratory volume in first second [FEV₁] 50.1% predicted) and 30 low-performing patients with COPD (mean age 60.6 (SD 6.8) years; 53.3% men; FEV₁ 55.6% predicted) regarding demographics, clinical characteristics, comorbidities and the presence of the SVD features and HCV. To conclude, the current study does not provide evidence for a relationship between cerebral SVD and HCV and cognitive functioning in patients with COPD. Additional studies will be needed to determine other possible mechanisms of cognitive impairment in patients with COPD, including microstructural brain changes and inflammatory-, hormonal-, metabolic- and (epi)genetic factors.

Concatenated Spatially-localized Random Forests for Hippocampus Labeling in Adult and Infant MR Brain Images

Automatic labeling of the hippocampus in brain MR images is highly demanded, as it has played an important role in imaging-based brain studies. However, accurate labeling of the hippocampus is still challenging, partially due to the ambiguous intensity boundary between the hippocampus and surrounding anatomies. In this paper, we propose a concatenated set of spatially-localized random forests for multi-atlas-based hippocampus labeling of adult/infant brain MR images. The contribution in our work is two-fold. First, each forest classifier is trained to label just a specific sub-region of the hippocampus, thus enhancing the labeling accuracy. Second, a novel forest selection strategy is proposed, such that each voxel in the test image can automatically select a set of optimal forests, and then dynamically fuses their respective outputs for determining the final label. Furthermore, we enhance the spatially-localized random forests with the aid of the auto-context strategy. In this way, our proposed learning framework can gradually refine the tentative labeling result for better performance. Experiments show that, regarding the large datasets of both adult and infant brain MR images, our method owns satisfactory scalability by segmenting the hippocampus accurately and efficiently.

The impact of automated hippocampal volumetry on diagnostic confidence in patients with suspected Alzheimer's disease: A European Alzheimer's Disease Consortium study

INTRODUCTION

Hippocampal volume is a core biomarker of Alzheimer's disease (AD). However, its contribution over the standard diagnostic workup is unclear.

METHODS

Three hundred fifty-six patients, under clinical evaluation for cognitive impairment, with suspected AD and Mini-Mental State Examination ≥20, were recruited across 17 European memory clinics. After the traditional diagnostic workup, diagnostic confidence of AD pathology (DCAD) was estimated by the physicians in charge. The latter were provided with the results of automated hippocampal volumetry in standardized format and DCAD was reassessed.

RESULTS

An increment of one interquartile range in hippocampal volume was associated with a mean change of DCAD of -8.0% (95% credible interval: [-11.5, -5.0]). Automated hippocampal volumetry showed a statistically significant impact on DCAD beyond the contributions of neuropsychology, ¹⁸F-fluorodeoxyglucose positron emission tomography/single-photon emission computed tomography, and cerebrospinal fluid markers (-8.5, CrI: [-11.5, -5.6]; -14.1, CrI: [-19.3, -8.8]; -10.6, CrI: [-14.6, -6.1], respectively).

DISCUSSION

There is a measurable effect of hippocampal volume on DCAD even when used on top of the traditional diagnostic workup.

Improving Spleen Volume Estimation Via Computer-assisted Segmentation on Clinically Acquired CT Scans

OBJECTIVES

Multi-atlas fusion is a promising approach for computer-assisted segmentation of anatomic structures. The purpose of this study was to evaluate the accuracy and time efficiency of multi-atlas segmentation for estimating spleen volumes on clinically acquired computed tomography (CT) scans.

MATERIALS AND METHODS

Under an institutional review board approval, we obtained 294 de-identified (Health Insurance Portability and Accountability Act-compliant) abdominal CT scans on 78 subjects from a recent clinical trial. We compared five pipelines for obtaining splenic volumes: Pipeline 1 - manual segmentation of all scans, Pipeline 2 - automated segmentation of all scans, Pipeline 3 - automated segmentation of all scans with manual segmentation for outliers on a rudimentary visual quality check, and Pipelines 4 and 5 - volumes derived from a unidimensional measurement of craniocaudal spleen length and three-dimensional splenic index measurements, respectively. Using Pipeline 1 results as ground truth, the accuracies of Pipelines 2-5 (Dice similarity coefficient, Pearson correlation, R-squared, and percent and absolute deviation of volume from ground truth) were compared for point estimates of splenic volume and for change in splenic volume over time. Time cost was also compared for Pipelines 1-5.

RESULTS

Pipeline 3 was dominant in terms of both accuracy and time cost. With a Pearson correlation coefficient of 0.99, average absolute volume deviation of 23.7 cm(3), and time cost of 1 minute per scan, Pipeline 3 yielded the best results. The second-best approach was Pipeline 5, with a Pearson correlation coefficient of 0.98, absolute deviation of 46.92 cm(3), and time cost of 1 minute 30 seconds per scan. Manual segmentation (Pipeline 1) required 11 minutes per scan.

CONCLUSION

A computer-automated segmentation approach with manual correction of outliers generated accurate splenic volumes with reasonable time efficiency.