Lyduine E. Collij

INTRODUCTION: Early Alzheimer's disease (AD) involves subtle cortical changes that may precede atrophy. Magnetic resonance imaging (MRI) microstructural markers may detect earlier pathology than classical morphometry. METHODS: We analyzed cross‐sectional MRI and amyloid‐β (Aβ) positron emission tomography (PET) data from 1323 non‐demented AMYPAD participants. Cortical volume, thickness, gray–white matter contrast (GWC), and mean diffusivity (MD) were related to global Aβ burden and estimated time to Aβ‐positivity using regression, correlation, and change‐point analyses. RESULTS: Microstructural measures showed stronger age associations than macrostructural measures, whereas all measures were unaffected by apolipoprotein E ( APOE) ‐ε4 carriership. GWC and MD showed minimal overlap with volume and thickness. Higher Aβ burden was most strongly associated with reduced GWC and cortical thinning. Change‐point analyses showed GWC alterations preceded Aβ‐positivity by several years. DISCUSSION: Cortical microstructural MRI, particularly GWC, changes earlier than atrophy and may serve as an early in vivo marker of AD pathology.

Structural pathways of the brain facilitate functional communication, and their disruption in preclinical Alzheimer's disease may reflect network vulnerability and compensatory brain maintenance. However, it remains unclear how early amyloid-β affects structure-function alignment, whether effects are explained by functional network organisation, how they relate to cognition, and which biological processes contribute to their development. We included 460 older adults without dementia from AMYPAD-PNHS with functional MRI, diffusion MRI, and amyloid-β PET. Structure-function coupling was quantified using the structural-decoupling index (SDI) at global, sub-network, and regional scales. Linear models investigated the effect of amyloid-β burden on SDI. Mediation analyses evaluated whether functional graph topology explained amyloid-associated SDI effects and whether SDI mediated the relationship between amyloid-β burden and cognition. Regional gene expression data were integrated to assess transcriptomic determinants of amyloid-related structure-function coupling. Amyloid-positive individuals exhibit higher global SDI, driven by visual cortices. Mediation analyses demonstrate that amyloid-related SDI alterations are explained by reductions in local clustering, indicating less segregated processing. Despite higher SDI in amyloid-positive individuals, elevated SDI in visual regions mitigates the negative effect of amyloid-β burden on cognition. Amyloid-related SDI changes correlate with genes associated with amyloid-β metabolism, microglial activation, and synaptic remodelling. Early amyloid-β pathology is associated with decoupling of brain structure and function, primarily in visual cortices, mediated by network reconfiguration and shaped by regional molecular architecture. These findings suggest that lower structure-function coupling may represent a compensatory mechanism in preclinical Alzheimer's disease and highlight SDI as a biomarker for stratification and monitoring in prevention trials.

Limbic-predominant age-related TDP-43 encephalopathy (LATE) is characterized by an amnestic- and limbic-predominant phenotype, which can mimic Alzheimer disease (AD). In a memory clinic cohort, we tested whether clinical criteria for LATE can detect a clinical profile of LATE that is distinct from AD. In this retrospective examination of a longitudinal memory clinic cohort from the Alzheimer Center Amsterdam, we included individuals with mild cognitive impairment (MCI) and dementia (aged >50 years). We classified individuals based on baseline data on cognition, atrophy, amyloid-status, and tau-status into Probable- and Possible-LATE, co-occurring LATE and AD (LATE-AD), and AD (without LATE). Next, we compared these groups on demographics, clinical features, cognition, and atrophy. Of 3,606 individuals (mean age at baseline 66 [SD 6], 49.2% female) available for classification, we classified 56 (1.6%) as Probable-LATE, 115 (3.2%) as Possible-LATE, 127 (3.5%) as LATE-AD, and 1,675 (46.5%) as AD. Individuals with Probable-LATE progressed slower than AD on mini-mental state examination (MMSE) (sβ [SE] = 0.12 [0.05], p = 0.02), memory (sβ [SE] = 0.11 [0.5], p = 0.01), attention (sβ [SE] = 0.12 [0.16], p = 0.05), executive functioning (sβ [SE] = 0.09 [0.04], p = 0.03), and visuospatial functioning (sβ [SE] = 0.10 [0.05], p = 0.05). Individuals with LATE-AD progressed faster than AD on MMSE (sβ [SE] = -0.12 [0.05], p = 0.01), attention (sβ [SE] = -0.13 [0.06], p = 0.04), and executive functioning (sβ [SE] = -0.10 [0.05], p = 0.03). Mortality risk, compared to AD, was lower in individuals with Probable-LATE (hazard ratio [HR] 0.70 [0.49-0.99], p = 0.04) and higher in Possible LATE-AD (HR 1.25 [1.01-1.53], p = 0.04). Compared to AD, at baseline, individuals with Probable-LATE and Possible-LATE had higher inferior temporal-to-hippocampus ratios (indicating limbic-predominant atrophy; sβ [SE] = 0.59 [0.16], p < 0.01; sβ [SE] = 0.40 [0.13], p < 0.01), and Probable-LATE, Possible-LATE, and LATE-AD all showed smaller amygdalar volumes at baseline than AD (sβ [SE] = -0.55 [0.15], p < 0.01; sβ [SE] = -0.43 [0.12], p < 0.01; sβ [SE] = -0.62 [0.11], p < 0.01). Individuals with LATE-AD had thinner cortex at baseline in an "AD-signature" composite region compared to AD (sβ [SE] = -0.73 [0.11], p < 0.01). Using an operationalization of clinical criteria for LATE, 8.2% of participants with MCI or dementia from our tertiary memory clinic were classified as Possible-LATE, Probable-LATE, or LATE-AD. Probable-LATE was characterized by a milder disease course than AD, whereas LATE-AD was characterized by a more aggressive disease course. This underscores the value of the proposed clinical criteria in identifying individuals with suspected LATE, who have distinct clinical trajectories from AD. Our findings, therefore, support the use of these criteria to improve diagnostic and prognostic accuracy in the memory clinic.

To efficiently implement plasma and cerebrospinal fluid (CSF) biomarkers for staging and prognosis of Alzheimer disease (AD), we must understand their dynamics across disease progression. We analyzed participants from the Swedish BioFINDER-2 study with mass spectrometry measurements of plasma and CSF tau species, including eMTBR-tau243/MTBR-tau243 and phosphorylation occupancies (%p-tau). Disease duration was estimated using Aβ-PET and tau-PET with the SILA algorithm. Bootstrapped LOESS models showed that %p-tau217 changes earliest, increasing just before Aβ-PET positivity. Other p-tau species changed later, with smaller dynamic ranges and earlier ceiling effects. %p-tau205 and MTBR-tau243 changes aligned with tau-PET positivity onset, while MTBR-tau243—especially plasma eMTBR-tau243—tracked cortical tau burden in later stages. Non-phosphorylated mid-region tau may serve as a late-stage biomarker. Taken together, concurrent assessments of plasma or CSF %p-tau217, %p-tau205, and (e)MTBR-tau243 provides information about different biological events in the disease cascade, which can benefit clinical trials and patient management in clinical practice.

Background: To efficiently implement Alzheimer disease (AD) cerebrospinal fluid (CSF) and plasma tau biomarkers, it is key to understand how they change over the course of disease progression. We therefore examined the temporal trajectories of mass spectrometry‐based measurements of microtubule‐binding region of tau (MTBR‐tau243) and tau phosphorylation occupancies (%p‐tau) in both CSF and plasma against disease time in relation to Aβ‐PET and tau‐PET positivity. Method: We included 784 participants with plasma and 446 with CSF biomarker data from the Swedish BioFINDER‐2, ranging from cognitively unimpaired (CU), subjective cognitive decline (SCD), mild cognitive impairment (MCI) or dementia. Aβ‐PET and tau‐PET chronicity ( i.e. , disease duration) was derived using the sampled iterative local approximation (SILA) algorithm, with positivity defined as 20 Centiloid and 1.36 SUVR, respectively. All biomarkers and continuous tau‐PET burden in the temporal meta‐ROI and Braak regions were z‐scored based on the CU Aβ‐negative group (CSF: n = 97, plasma: n = 116). For CSF Aβ _42/40 , data were inverted such that higher z‐scores related to higher abnormality (for consistency across all biomarkers). For visualization purposes of the plasma biomarkers, all biomarkers and tau‐PET burden were also z‐scored based on the whole population, as plasma MTBR‐tau 243 has essentially zero variance in CU Aβ‐negative individuals. Result: The CSF and plasma cohorts were on average 71.4(±8.5) and 72.2 (±9.3) years of age, 49.1%/50.8% was female, and 263/453 (0.59%/51.8%) were cognitively impaired, respectively. In both CSF and plasma, only %p‐tau217 changed early, with significantly increased levels just prior to Aβ‐PET positivity (Figure 1A‐D, Figure 2A/B). Other p ‐tau species, such as %p‐tau181, changed later in CSF and plasma, had smaller dynamic ranges, and earlier ceiling effects (Figure 1E‐G, Figure 2C/D). Changes in CSF and plasma %p‐tau205 were closely associated with tau‐PET positivity onset (Figure 1H, Figure 2C/D). MTBR‐tau243 trajectories, especially in plasma, were closely associated with trajectories of cortical tau‐PET burden (Figure 1C, Figure 2C). Finally, CSF and plasma non‐phosphorylated mid‐region tau could be a potential late‐stage biomarker. Conclusion: CSF or plasma p ‐tau217, p ‐tau205, and MTBR‐tau243 provide information about different biological events in the disease cascade (Figure 3), which can benefit clinical trials and patient management in clinical practice.

Despite extensive research, open questions about the biological underpinnings of Alzheimer’s disease (AD) remain. Neuroimaging biomarkers based on positron emission tomography (PET) and magnetic resonance imaging (MRI) offer in vivo insights into these complex biological changes and interactions. However, most evidence to date comes from cross-sectional studies, limiting our understanding of disease progression. Longitudinal studies enable the investigation of biological changes within individuals, revealing how pathology evolves over time. With this review, we provide an overview of how longitudinal imaging biomarker studies have advanced the field and how they can contribute to future research. We highlight longitudinal biomarker studies that have provided critical insights into disease trajectories, staging, and individual variability. We further assess longitudinal multimodal studies which have elucidated interactions between AD-specific pathology, amyloid-β and tau, and broader biological changes like neurodegeneration, neuronal dysfunction, vascular disease, and inflammation. Further, we discuss associations of brain changes with symptomatology and clinical outcomes and conclude with challenges and future directions.

Limbic-predominant age-related TDP-43 encephalopathy (LATE) is characterized by an amnestic-and limbic-predominant phenotype, which can mimic the clinical presentation of Alzheimer’s disease (AD). Differentiating between LATE and AD is imperative in the era of disease-modifying treatments for AD, but there are no established biomarkers to detect TDP-43. In this observational study, we characterized individuals fulfilling the recently proposed clinical criteria for LATE in a tertiary memory clinic setting in terms of cognition, atrophy, and disease course. From the Amsterdam Dementia Cohort (ADC), we included 3606 individuals aged above 50 years old who had a diagnosis of MCI or dementia. We first classified individuals based on criteria for memory impairment, atrophy, amyloid-status and tau-status, into (Probable and Possible) LATE, co-occurring LATE and AD (Possible LATE-AD), and AD without LATE. Next, we compared these four groups on demographic, clinical, cognitive, and brain atrophy data. Longitudinal cognitive data were available for a subset. We classified 56/3606 (1.6% of the total group) as Probable LATE, 115 (3.2%) as Possible LATE, 127 (3.5%) as Possible LATE-AD, and 1675 (46.5%) as AD. The remaining 1633 (45.3%) were not classified within this framework (e.g., frontotemporal dementia). At baseline, individuals with Probable LATE had lower memory scores but higher scores on all other domains and MMSE than AD, while individuals with Possible LATE-AD had worse scores across all cognitive scores than AD (pin vivo .

With the rapid development of blood biomarkers (BBMs) related to amyloid‐β (Aβ) pathology in Alzheimer's disease (AD), the question arises whether these can replace the accepted reference standard, positron emission tomography (PET), for assessing Aβ burden. BBMs can differentiate Aβ status reliably in cognitively impaired patients, but two‐threshold strategies to further improve their performance demonstrate the need for pathological confirmation by Aβ‐PET in a non‐negligible portion of individuals (∼10%–40%), especially in early‐stage disease where BBM performance declines due to lower AD prevalence, reducing the test's positive predictive value (PPV) and increasing the risk for false‐positives. This may be increasingly relevant in the future, considering the development of (very) early interventions against AD. Further, BBMs do not accurately reflect the actual Aβ load or change after immunotherapy. Consequently, there are clear remaining needs for Aβ‐PET in several clinically important settings as (i) a confirmatory test and (ii) to determine treatment response. Highlights: When used as stand‐alone tests, blood biomarkers (BBMs) demonstrate good sensitivity and specificity (84%–90%) relative to amyloid‐β positron emission tomography (Aβ‐PET). Two‐threshold strategies improve BBM performance but require confirmatory testing by, for example, Aβ‐PET in a non‐negligible portion of patients that fall in an intermediate range or “gray‐zone” (∼10%–40%). In early/preclinical populations, BBM performance declines due to lower AD prevalence, reducing the test's positive predictive value (PPV) and increases the gray‐zone population. Currently available BBMs cannot reliably estimate Aβ‐PET burden or track Aβ‐plaque removal post‐immunotherapy.

The spatial distribution of tau pathology, the core driver of neurodegeneration in Alzheimer's disease (AD), varies markedly across individuals. While tau is thought to spread along brain networks, the role of inter-individual variability in shaping these patterns remains underexplored. Using resting-state fMRI and tau- PET from 805 participants across the AD continuum, we studied whether subject-specific functional connectivity (FC) profiles enhance the characterization of tau deposition patterns. A hybrid approach integrating individual and group- average FC outperformed both alone, particularly in symptomatic individuals and at finer spatial resolutions, the latter underscoring a critical but often overlooked role of spatial scale. Individualized FC also better captured individual tau topographies than canonical tau-PET maps derived from cohort-level data. These effects were specific to tau, and not seen for β-amyloid, and their predictive power increased with spatial granularity. Furthermore, baseline FC also predicted future tau accumulation at the individual level, supporting its prognostic value. Together, these findings provide strong evidence that individual functional brain architecture shapes tau propagation in humans, supporting the network spread hypothesis by showing that variability in connectivity translates into heterogeneity in tau distribution. This work advances biological understanding of tau propagation in AD, highlighting functional connectivity as a mechanistic substrate that supports prognostic assessment of tau trajectories.