Search

Background: Tau tangles are one of the pathological hallmarks of Alzheimer's disease (AD) and can be quantified using PET. It remains unclear which molecular processes are related to tau aggregation. Protein levels in CSF can be used to study such underlying processes. This study aims to identify biological processes related to tau‐PET using unbiased CSF proteomics in participants along the AD continuum. Method: We included 89 participants with CSF proteomic and [18F]flortaucipir (tau)‐PET data available from the Amsterdam Dementia Cohort and the EMIF‐preclin AD study (N = 68 with abnormal CSF Aβ42 [64.7% cognitively impaired, 67.6% tau‐PET visual read positive] and N = 21 controls [normal cognition and normal CSF AD biomarkers]) (table 1). [18F]flortaucipir BPND was quantified in the temporal meta‐ROI (Braak I, Braak III, Braak IV). CSF proteomics was measured using untargeted LC‐MS/MS based on TMT labelling. We included proteins available in the whole sample (n = 1421 proteins). Protein concentrations were normalized to the control group. We determined the relationship between tau‐PET BPND (outcome) with protein levels (determinant) using linear regressions with sex and age as covariates, stratified for group. GO database was used for biological pathway enrichment analyses within the AD group for proteins thresholded at p<0.05 for positive and negative associations separately. Result: Higher CSF tau levels were associated with higher tau‐PET BPND (β = 0.092, p = 0.002). In total, 458 proteins were associated with tau‐PET BPND in AD. A higher tau‐PET BPND was related to higher levels of 140 proteins and lower levels of 318 proteins. Proteins that showed higher levels with increasing tau‐PET BPND were mainly involved in immune system processes (figure 1), while the proteins with lower levels were predominantly related to synaptic processes and cell signaling (figure 2). No significant associations were found within controls. Conclusion: Using CSF proteomics, we identified 458 proteins associated with tau‐PET in AD. Higher tau burden was associated with increased levels of proteins involved in immune activation suggesting a role of the immune system in tau accumulation. Furthermore, higher tau burden was associated with lower levels of proteins related to synaptic process hinting at a role for tau in synaptic degeneration. [ABSTRACT FROM AUTHOR]

Background: The association between tau pathology and neurodegeneration (atrophy) in Alzheimer's disease (AD) is not fully understood. This study aimed to investigate whether tau pathology is associated with concurrent and longitudinal cortical thickness in individuals along the AD continuum. Method: We included 92 amyloid‐positive cognitively impaired individuals (age 65±7, 48% female, MMSE 23±4) with AD dementia (n = 73) or mild cognitive impairment due to AD (MCI, n = 19) and 61 cognitively normal individuals (CN; age 66±8, 48% female, MMSE 29±1) with subjective cognitive decline (SCD, n = 55, 19 amyloid‐positive) or healthy controls (HC, n = 6, amyloid‐status unknown). All participants underwent dynamic [18F]flortaucipir PET and structural 3T‐MRI at baseline. A subset underwent two‐year follow‐up structural 3T‐MRI (CN: n = 35, time‐interval 24±8 months; AD: n = 26; time‐interval 28±4 months). Parametric [18F]flortaucipir binding potential (BPND) images were generated using receptor parametric mapping (cerebellar gray matter reference region) to quantify tau pathology. Cortical thickness (over time) was measured using Freesurfer (version 6.0.1). Regional associations between [18F]flortaucipir BPND and cortical thickness were assessed using linear mixed models in three commonly used a priori defined regions‐of‐interest, including entorhinal, middle‐and inferior temporal, and neocortical areas. Models were adjusted for age, sex and amyloid status (in case of CN individuals). Result: For cognitively impaired individuals, higher [18F]flortaucipir BPND (in all three regions) was associated with lower concurrent cortical thickness in the entorhinal cortex [Table‐1&Fig‐1]. Higher middle‐ and inferior temporal and neocortical BPND was associated with cortical thinning over time within the same region (intra‐regionally). Higher middle‐ and inferior temporal BPND additionally showed associations with cortical thinning over time in the neocortex (inter‐regionally). For cognitively normal individuals, we found no associations between [18F]flortaucipir BPND and cortical thickness (over time) [Table‐1&Figure‐1]. Conclusion: Tau pathology is associated with both concurrent and longitudinal atrophy in a region‐specific manner in cognitively impaired individuals (MCI/AD), but not in cognitively normal individuals (HC/SCD). The regional‐specificity of these associations might suggest that there are similarities between tau‐related progression patterns and patterns of neurodegeneration. Taking into account a temporal delay between the development of region‐specific tau pathology and neurodegeneration, this potentially suggests that tau pathology is a (partial) driver of neurodegeneration in AD. [ABSTRACT FROM AUTHOR]

Background: Different tau‐PET patterns have been observed in Alzheimer's disease (AD) which may be associated with demographic factors, copathology and resilience. We and others previously found that not all Aβ‐positive AD patients are tau‐PET positive. It is currently not known whether tau pathology develops in these groups over time. Therefore, we aimed to examine longitudinal tau binding in tau‐PET negative individuals with symptomatic AD. Method: We included 97 Aβ‐positive participants with clinical diagnosis of mild cognitive impairment or dementia due to AD (MCI/AD) of which 12 (13%) were tau‐PET negative. To compare, we included 131 tau‐PET negative cognitively unimpaired individuals (CU) of which 93 (71%) were Aβ‐negative and 38 (29%) Aβ‐positive. Tau‐status was defined by [18F]flortaucipir PET consensus visual read of two nuclear physicians according to US Food and Drug Administration‐approved guidelines. [18F]flortaucipir SUVr at baseline (N = 225) and after ±2 years (N = 81) was quantified in an early tau region reflecting Braak I (entorhinal cortex), the medial temporal lobe (MTL), and a combined visual read region of interest (ROI) (parietal, occipital, and posterior lateral temporal lobe). We used age‐and sex‐adjusted linear mixed models to compare tau accumulation. Result: Demographics are shown in table 1. Compared to MCI/AD A+T+, MCI/AD A+T‐ were older (p<0.001) and more often male (p = 0.015). MCI/AD A+T‐ did not differ from CU A‐T‐ in baseline tau‐PET SUVr, but had a slightly lower SUVr in Braak I compared to CU A+T‐ (p = 0.040) (figure 1). After ±2 years, all MCI/AD A+T‐ participants who underwent follow up remained visual read tau negative. However, MCI/AD A+T‐ showed a steeper increase in Braak I and MTL tau‐PET SUVr compared to CU A‐T‐ (figure 2A,B) and in Braak I compared to CU A+T‐ (β = 0.31, p = 0.036). Interestingly, MCI/AD A+T‐ showed a significantly steeper tau‐PET SUVr increase compared to MCI/AD A+T+ in Braak I (β = 0.29, p = 0.029) (figure 2A). Conclusion: Although visual read tau‐PET status remains stable over time, our results indicate that tau‐negative MCI/AD patients do accumulate tau over time, most pronounced in early tau regions. This highlights the relevance of quantifying tau regionally over time in tau‐PET negative individuals with clinical AD diagnosis. [ABSTRACT FROM AUTHOR]

Background: Limited information is available on the genetic and environmental contributions to early AD pathology. Within a sample of cognitively unimpaired genetically identical older twins we aimed to assess 1) effects of amyloid‐ß pathology on [18F]flortaucipir (tau) binding and 2) spatial similarities in tau patterns within twin pairs. Method: To date, ten twin pairs (i.e. twenty twins) from the on‐going EMIF‐AD PreclinAD study underwent dynamic (0‐30min and 80‐100min) [18F]flortaucipir PET. Two twin pairs were concordant amyloid‐ß positive, four twin pairs concordant amyloid‐ß negative and four twin pairs amyloid‐ß discordant (i.e. one twin positive and co‐twin negative) based on [18F]flutemetamol visual read. [18F]flortaucipir binding potential (BPND) images were generated using receptor parametric mapping (reference region: cerebellar gray matter). First, we compared [18F]flortaucipir BPND in an early (entorhinal), intermediate (limbic) and late (neocortical) stage tau pathology region between amyloid‐ß positive and negative twins using generalized estimated equation correcting for twin dependency. Next, we examined spatial tau pattern similarities within twin pairs by correlating (Spearman) [18F]flortaucipir BPND across 46 ROIs from the Hammers template between each twin and their co‐twin for each true twin pair. Correlations were also performed for all non‐twin "random" pairs. Correlation coefficients within true twin pairs were compared to the correlation coefficients within random pairs, correcting for age and twin amyloid‐ß status. Results: Amyloid‐ß positive twins showed higher [18F]flortaucipir BPND in entorhinal (p=0.01) and limbic (p=0.04) regions(Table‐1). Fig‐1 shows representative [18F]flortaucipir images of an amyloid‐ß concordant positive, concordant negative and discordant pair, with regional similarities observed in amyloid‐ß concordant pairs as well as dissimilarities in amyloid‐ß discordant pairs (Fig‐1;Fig‐2). Between‐twin correlations across regional [18F]flortaucipir BPND, to assess spatial tau pattern similarities, were higher for true twin pairs (r=0.71±0.17) compared to random pairs (r=0.45±0.23) (p=0.002)(Table‐2). Conclusion: In line with previous studies, amyloid‐ß positive cognitively unimpaired twins showed higher tau deposition in early and intermediate tau stage regions. Furthermore, we observed higher spatial tau pattern similarities within true twin pairs compared to random pairs, suggesting a moderate genetic contribution to mechanisms underlying tau spread. With the upcoming larger sample we aim to further assess genetic and environmental contributions to early AD pathology. [ABSTRACT FROM AUTHOR]

Background: A distinct subset of patients with Alzheimer's disease (AD) experience early and predominant behavioral and personality changes, while the neurobiological origins of these symptoms are largely unknown. In this study, we examined the distribution of tau pathology using PET imaging or post‐mortem evaluation in patients with the behavioral variant of AD (bvAD). Method: For the PET study, seven amyloid‐β positive clinically diagnosed patients with bvAD from the Amsterdam Dementia Cohort (ADC, n=2), University of California San Francisco (UCSF, n=3) and the Swedish BioFINDER study (n=2) underwent [18F]flortaucipir or [18F]RO948 (tau) PET (Table 1). Regional standardized uptake value ratios (SUVR) in entorhinal, temporoparietal, frontal, and insular cortex as well as mean cortical SUVR and frontal‐to‐parietal SUVR ratios were contrasted against patients with typical (memory‐predominant) AD (tAD; n=55, n=60 and n=90, respectively). For the post‐mortem component of the study, the percentages of AT8 (tau)‐positive areas in hippocampus CA1, temporal, parietal, frontal and insular cortices were compared between 9 cases with bvAD and 7 with tAD. Result: Visual inspection of subject‐specific tau PET patterns revealed more pronounced frontal involvement in 6/7 bvAD patients compared to patients with tAD (Figure 1). Regional SUVR analyses revealed higher‐than‐average tau uptake in 3/7 bvAD cases in the temporoparietal cortex, 5/7 in the frontal cortex, and 3/7 in the insular cortex compared to patients with tAD (Figure 2). Furthermore, 4/7 cases showed higher‐than‐average mean cortical SUVR and 6/7 cases had higher‐than‐average frontal‐to‐parietal SUVR ratios (Figure 2 & Table 2). Postmortem evaluation revealed that bvAD cases had a lower tau load in the CA1 region of the hippocampus compared to patients with tAD (p<.05); no differences were found, however, in temporal, parietal, frontal or insular regions (all p>.05, Figure 3 & Table 3). Conclusion: This multicenter case series with tau PET and neuropathological examination demonstrates a heterogeneous distribution of tau pathology across patients with bvAD, ranging from pronounced anterior involvement to a more temporoparietal pattern. Further work is needed to explore alternative pathophysiological mechanisms underlying the bvAD phenotype. [ABSTRACT FROM AUTHOR]

Background: Dementia with Lewy Bodies (DLB) is characterized by the presence of neuronal inclusions containing alpha‐synuclein proteins, and often co‐occurs with Alzheimer's disease (AD) pathology. We aimed to determine the pattern of tau pathology and relative cerebral blood flow (rCBF) in DLB, compared to AD and controls, and their association with cognitive impairment using a single dynamic [18F]flortaucipir PET scan. Method: Eighteen patients with DLB (66 ± 8 years, MMSE 25 ± 4, 44% amyloid positive (abnormal CSF (Aβ42 <813 pg/mL)/visual read amyloid‐PET), 100% FP‐CIT SPECT abnormal, Table 1), 65 amyloid positive cognitively impaired patients (MCI‐AD (n = 13), AD (n = 52)) and 50 controls underwent a dynamic 130‐minute [18F]flortaucipir PET scan. Receptor parametric mapping (cerebellar gray matter reference region) was used to extract (regional, Hammers based) binding potential (BPND) and R1, which reflect tau pathology and rCBF, respectively. We performed voxel‐wise comparisons (Puncorrected < 0.001) of [18F]flortaucipir BPND and R1 between diagnostic groups using SPM, adjusted for age and sex. DLB patients underwent extensive neuropsychological assessment covering memory, executive functioning, language, attention and visuospatial domains. For DLB only, we performed linear regression analyses between [18F]flortaucipir BPND, R1 and cognition in the following regions‐of‐interest (ROIs); medial and lateral temporal/ parietal, occipital and frontal cortex, adjusted for age, sex and education. Result: Averaged [18F]flortaucipir BPND images across groups showed visually minimal tau uptake in the inferior temporal lobe in DLB (Figure 1). Voxel‐wise comparisons showed lower [18F]flortaucipir R1 in the occipital lobe in DLB compared to AD and controls (Figure 2). Regional [18F]flortaucipir BPND was lower in DLB compared to AD and comparable with tau binding in controls (Figure 3). Occipital and lateral parietal R1 was lower in DLB compared to AD and controls (all p<0.01, Figure 3). R1 but not BPND was related to cognition (language only); lower medial temporal (stß = 0.76, p = 0.001), medial parietal (stß = 0.64, p = 0.02) and frontal (stß = 0.64, p = 0.02) R1 was related to lower language score (Figure 4). Conclusion: In our sample, tau load in patients with DLB did not differ from controls, but there were DLB‐specific occipital and lateral parietal flow reductions compared to both controls and AD patients. Our results indicate that rCBF may be of more clinical relevance than tau pathology in DLB. [ABSTRACT FROM AUTHOR]

Background: Several studies validated R1 as a marker for relative cerebral blood flow (rCBF). However, the similarities or differences of R1 estimates obtained from different PET tracers has never been investigated. Aim: The aim of the current study was head‐to‐head comparison of R1 estimates, derived from dynamic [18F]florbetapir and [18F]flortaucipir PET scans. Method: Fifty subjects with subjective cognitive decline (Aβ‐;N = 31, Aβ+;N = 19) underwent both [18F]florbetapir and [18F]flortaucipir PET (interval: 50.6 ± 34.7 days). Receptor parametric mapping (RPM) was used to generate R1 maps with cerebellar grey matter as reference. For voxel‐wise analyses, Statistical Parametric Mapping software (SPM12) was used. Furthermore, different regions‐of‐interest were assessed (in subject‐space). Paired t‐tests, coefficients of determination (r2) and ICC were used to compare R1 estimates from [18F]florbetapir and [18F]flortaucipir PET. Results: Voxel‐wise analysis revealed only a few small significant clusters in part of the thalamus, insular gyri, midbrain and part of the cerebellum with higher R1 estimates for [18F]florbetapir and higher R1 estimates in the hippocampus, vermis, midbrain and calcarine sulcus for [18F]flortaucipir in Aβ‐ subjects (Figure 1a & 1b). In Aβ+ subjects, [18F]florbetapir showed higher R1 estimates only in part of the thalamus and [18F]flortaucipir showed higher R1 estimates in the hippocampus and striatum (Figure 1c & 1d). Despite the small amount of significant voxels, [18F]florbetapir R1 showed good correlation and excellent ICC with [18F]flortaucipir for whole brain (Aβ‐:r2 = 0.76, ICC:0.92; Aβ+:r2 = 0.60,ICC = 0.87). Conclusion: To conclude, [18F]flortaucipir and [18F]florbetapir showed similar R1 estimates in cortical regions, illustrating that R1 is a robust marker for flow independent of tracer of interest. [ABSTRACT FROM AUTHOR]

Background: The mechanisms contributing to synaptic loss in Alzheimer's disease (AD) are poorly understood and may be associated with tau pathology. In this combined positron emission tomography (PET) and magnetoencephalography (MEG) study, we therefore investigated associations between tau ([18F]flortaucipir PET), synaptic density (synaptic vesicle 2A [11C]UCB‐J PET) and synaptic function (MEG) in AD. Method: Seven amyloid‐positive AD subjects (age 64.3±8.2; 3/7 female; MMSE 24.1±1.8) were included from the Amsterdam Dementia Cohort. All subjects underwent dynamic 130‐min [18F]flortaucipir and 60‐min [11C]UCB‐J PET with arterial sampling. Six subjects underwent 10‐min eyes‐closed 306‐channel MEG. [18F]flortaucipir binding potential (BPND) was determined with receptor parametric mapping (reference region: cerebellar gray matter) with partial volume correction (Hypr‐IDM‐Hypr). [11C]UCB‐J distribution volume ratio (DVR, plasma input‐derived) was obtained using the centrum semi‐ovale as reference region. [18F]flortaucipir BPND, [11C]UCB‐J DVR and MEG spectral measures (total power (0.5‐48Hz), peak frequency, and relative delta (0.5‐4Hz), theta (4‐8Hz), alpha (8‐13Hz) power) were calculated in 12 bilateral a priori defined temporal, parietal, frontal and occipital region‐of‐interests (ROIs). We used generalized estimating equations (GEE), correcting for multiple ROIs per subject, to investigate associations between [18F]flortaucipir, [11C]UCB‐J and spectral MEG. GEE analyses with MEG were performed separately per brain lobe. Additionally, within‐subject regional correlations (Spearman) between [18F]flortaucipir and [11C]UCB‐J were performed. Result: Higher [18F]flortaucipir BPND was associated with lower [11C]UCB‐J DVR (β=‐0.35, p<0.001) (Figure 1A,B). Within subjects, negative associations were observed when neocortical tau was high, while positive associations were observed when neocortical tau was low (Figure 1C). Additionally, higher [18F]flortaucipir was strongly associated with oscillatory slowing, reflected by negative associations with peak frequency and alpha power, and positive associations with delta or theta power (Table 1). In contrast, lower [11C]UCB‐J binding was associated with lower peak frequency and alpha power, and higher delta power, most clearly in the occipital lobe (Table 1). Conclusion: Across subjects, higher regional tau pathology was associated with synaptic loss, while within subjects, this association depended on subjects' neocortical tau levels. Moreover, higher tau pathology and synaptic loss were associated with oscillatory slowing. This suggests that tau is associated with synaptic loss and reduced synaptic activity, yet that a tau‐threshold may need to be reached before regional synaptic loss occurs. [ABSTRACT FROM AUTHOR]

Background: Tau pathology can be assessed by [18F]flortaucipir positron emission tomography (PET). Dynamic PET protocols allow for accurate quantification of [18F]flortaucipir specific binding. However, dynamic acquisitions can be challenging or burdensome to patients given the long required scan duration of 130 min, especially in persons with cognitive impairment due to Alzheimer's disease (AD). Therefore, the current study assessed the effect of shorter scan protocols for [18F]flortaucipir PET on its quantitative accuracy. Method: For this study, two study samples were used: 10 Alzheimer disease (AD) patients and 10 healthy controls (HC) with arterial blood sampling, and 8 AD patients and 6 HCs with a test‐retest scan protocol but without arterial sampling. All subjects underwent a 130‐minute dynamic [18F]flortaucipir PET scan with a 20 min break after 60 minutes. First, the second part of the scan was shortened to 80‐120, 80‐110 and 80‐100 minutes p.i.. Second, the first part of the scan was shortened to 0‐50, 0‐40, and 0‐30 min p.i.. To fill the gap of missing time frames in the reference region (i.e. grey matter cerebellum) time activity curves, four different interpolation methods were tested: 2T4k_VB model using population‐based input‐function (2T4k_POP‐IP), cubic, linear and exponential interpolations. Regional BPND values estimated using SRTM and/or RPM and different scan protocols were compared to the respective values estimated using the original scan acquisition. In addition, test‐retest repeatability (TRT) of the RPM BPND values estimated using the optimal shortest scan duration was assessed. Result: RPM BPND values obtained using a shortened second part of the PET scan had negligible effect on the accuracy of parameter estimates (Table 1). The RPM BPND values using 0‐30/80‐100 min with 2T4k_POP‐IP interpolation had the best correlation with original DVR‐1, SRTM and RPM BPND (r2=0.95, slope=0.91; r2=0.97, slope=0.88; r2=0.99, slope=0.96). The TRT remained similar using the shortest scan duration (Figure 1). Conclusion: The current study demonstrated that accurate [18F]flortaucipir quantification can be obtained using a scan protocol of 0‐30/80‐100 min p.i.. With this reduction in total scan duration the scan will be less burdensome while remaining quantitatively accurate. [ABSTRACT FROM AUTHOR]

Background: Carotid occlusive disease (COD) is a risk factor for cognitive decline. One of the proposed underlying mechanisms is that cerebral hypoperfusion may result in accelerated accumulation of amyloid‐β in the brain. We hypothesized that patients with longstanding, unilateral COD have an asymmetrical burden of cerebral amyloid‐β, with more amyloid‐β in the hemisphere ipsilateral to the COD than in the contralateral hemisphere. Method: The AMYCODE study is a cross‐sectional observational study. Inclusion criterion is a unilateral occlusion of the internal carotid artery. Exclusion criteria are: a contralateral stenosis of > 70% and a history of vascular reconstructive surgery. All patients underwent a neuropsychological assessment, MRI‐scan and dynamic 18F Florbetaben positron emission tomography scan (PET). Global relative perfusion (R1) and binding potential (BPND) were determined from the PET‐images using PMOD and a simplified reference tissue model with the cerebellar grey matter as reference region. Distribution Volume ratio (DVR) was calculated as BPND + 1. We performed Wilcoxon signed‐rank tests to examine differences between hemispheres within subjects. Result: To date, we included eight participants (seven males, age 66.1±7.3 years, MMSE score 28.1±2.0; inclusion ongoing, target n = 20). Mean global DVR was 1.03±0.03 in the ipsilateral hemisphere (i.e. side of the occlusion) and 1.04±0.02 in the contralateral hemisphere (p = 0.064). Mean global R1 was 0.89±0.06 in the ipsilateral hemisphere and 0.91±0.05 in the contralateral hemisphere (p = 0.033). Conclusion: Preliminary analyses suggest that patients with unilateral COD had similar amyloid‐β distribution, despite lower cerebral perfusion in the hemisphere ipsilateral to the COD than in the contralateral hemisphere. [ABSTRACT FROM AUTHOR]

Background: Dynamic [18F]flortaucipir PET imaging allows for more accurate quantification of specific binding compared to static imaging in Alzheimer’s disease (AD). Besides, with dynamic imaging flow is taken into account, which is especially important in longitudinal studies, because disease progression or drug intervention may lead to flow changes. We aimed to compare semi‐quantitative (SUVr) and quantitative (BPND) parameters for [18F]flortaucipir PET in a longitudinal study. Method: We included subjects with subjective cognitive decline (SCD; n=38) and AD patients (n=24). All subjects underwent baseline (BL) and 2‐year follow‐up (FU) dynamic [18F]flortaucipir PET. BPND and R1 (relative cerebral blood flow) were calculated (RPM; reference region cerebellar gray). SUVr was derived 80‐100 min p.i.. (Trans)entorhinal, limbic and neocortical regions‐of‐interest (ROIs) were assessed. For each parameter %change was calculated (100*(FU‐BL)/BL). Regional SUVr was compared against respective DVR (BPND+1) using paired T‐tests. Additionally, simulations were performed to model effects of flow changes on BPND and SUVr in high, medium or low binding categories. For each category, R1 values were varied to mimic 5‐25% increase or decrease in flow. Thereafter, %bias for SUVr with respect to underlying binding and its association with respect to change in flow were evaluated. Result: In SCD, there was a significant difference between %change in DVR (2.56%) and SUVr (1.85%) in the (trans)entorhinal ROI (p<0.05) only. In AD, differences in %change were found in the limbic ROI (DVR 6.61% vs. SUVr 7.52%; p<0.05) only. R1 changes were small (+0.7% in SCD and ‐1.6% in AD). Simulations showed varying %bias in SUVr depending on underlying binding. With increasing flow a decreased bias in low binding conditions was observed, whereas in high binding conditions an increased bias was seen. DVR remained relatively stable under different flow conditions, irrespective of underlying binding. Conclusion: Compared to BPND, SUVr provided a good estimate of specific binding for [18F]flortaucipir over a two‐year follow‐up. However, simulations showed that [18F]flortaucipir SUVr is subject to variable bias in case of larger flow changes, which seemed to be depending on underlying tau load. Therefore, BPND should be preferred over SUVr, especially in the context of pharmacotherapeutic interventions that could induce significant changes in flow. [ABSTRACT FROM AUTHOR]

Background: Early‐onset Alzheimer's disease (AD) dementia is characterized by a heterogeneous cognitive profile consisting of language and visuo‐spatial impairment. Spatial patterns of tau pathology measured with [18F]flortaucipir may help to explain the heterogeneity within early‐onset AD. Method: Seventy‐one patients with Alzheimer's disease dementia patients (n=28 early‐onset age=58, MMSE=22; n=43 late‐onset age=71, MMSE=23), of whom 6 patients with posterior cortical atrophy, 2 behavioral of AD (bvAD). All patients had available 130 minutes [18F]flortaucipir PET scans and extensive neuropsychological assessment. We generated parametric images and non‐displaceable binding potentials (BPnd) were extracted using receptor parametric mapping in the following brain regions; frontal, occipital, parietal, medial and lateral temporal cortex. We performed principal component analyses to reduce 12 cognitive tests into latent variables which reflect (uncorrelated) cognitive functions. ANOVA and linear regressions were performed to investigate between [18F]flortaucipir group differences and associations with cognition functions. Result: Early‐onset AD dementia patients showed increased [18F]flortaucipir BPnd in occipital, parietal and frontal cortex (all p<0.001) compared to late‐onset AD. Across all AD patients, adjusted for education, sex and age, frontal [18F]flortaucipir BPnd was associated with attention (p=0.003), occipital [18F]flortaucipir BPnd with visual scanning and memory (both p=0.01), parietal [18F]flortaucipir BPnd with attention and visual scanning (p=0.002 and p=0.02 respectively). Lateral temporal [18F]flortaucipir BPnd was associated with attention (p=0.006) and medial temporal [18F]flortaucipir BPnd with memory (p=0.01). Interaction effects were found for occipital and parietal [18F]flortaucipir BPnd and visual scanning for early‐onset AD, and parietal [18F]flortaucipir BPnd and attention for late‐onset. Frontal [18F]flortaucipir BPnd was associated with visual scanning in late‐onset AD. In an exploratory analysis, we found higher [18F]flortaucipir BPnd in parieto‐occipital cortices in PCA compared early‐ and late‐onset AD (both p<0.001). Conclusion: Differential spatial patterns of tau pathology were found depending on age‐of‐onset. Tau pathology may help to explain the clinical phenotype of early and late‐onset AD. [ABSTRACT FROM AUTHOR]