Citations

Abstract Background The HEAD study aims to collect a large dataset of multiple tau‐PET tracers to provide robust anchor values for tau‐PET harmonization. Here, we tested the hypothesis that anchoring two tau tracer uptake values using head‐to‐head measurements has the potential to generate an accurate universal tau‐PET scale, named Uniτ(tau). Methods We assessed 200 individuals across the aging and AD spectrum (Training: HEAD data freeze 2.0, n=185; Testing: UPitt dataset (Gogola et al.), n=15) with [ 18 F]Flortaucipir and [ 18 F]MK‐6240 tau‐PET. SUVRs were processed to a common 8mm FWHM, with inferior cerebellar gray matter as the reference region (Pascoal et al.). Uniτ explored two anchoring/harmonization methods. First, we examined within‐tracer anchoring by creating anchor values based on the mean SUVR of Youngs (<25 years) and 95th percentile voxels from cognitively impaired individuals. Second, we explored within‐ plus between‐tracer anchoring, employing linear (e.g., piecewise) and non‐linear regressions. To address the inherent problem of discontinuity of piecewise transformations, we implemented two smoothing methods at the inflection point between equations, transforming them into a continuous function. Results Uniτ scale anchoring within‐tracer resulted in similar estimates for high values, but less accurate in the lower range (Figure 1,2). Anchoring within‐ plus between‐tracer improved estimate consistency, with the piecewise transformation generating the best results. The piecewise smoothing equation yielded estimates comparable to those obtained from the piecewise method without smoothing. This allowed for the use of a single formula. In addition, this leads to more robust results when the goal is to study longitudinal changes in the scale (data not shown). UPitt testing dataset showed similar results to the training set (Figure 3). Conclusion Our preliminary findings suggest that anchoring tau‐PET values both within and between tracers has the potential to harmonize tau‐PET tracers, while preserving their characteristics. Currently, piecewise smoothing is the preferred method for Uniτ, but we are continuously fine‐tuning scale parameters as we acquire more data. The final scale parameters will be determined based on extensive training and testing data from multiple tracers. This cautious methodology holds the promise of delivering reliable, robust, and reproducible results, ensuring safe usage of the scale in clinical trials, and potentially paving the way for future use in clinical practice.

AbstractBackgroundHarmonization of the two most commonly used Tau PET tracers, 18F‐Flortaucipir and 18F‐MK6240 has proven to be complex. Unlike the centiloid scale for amyloid tracers, Tau PET SUVRs of the two tracers are not linearly comparable and vary markedly in dynamic range and sensitivity.MethodTau PET SUVRs for Braak stage (1‐6) in 18F‐MK6240 and 18F‐Flortaucipir were obtained from the Longitudinal multicenter head‐to‐head harmonization of tau‐PET tracers (HEAD) project. For training, we used 13 amyloid confirmed AD, 24 amyloid confirmed MCI and 66 amyloid negative controls. To assess linearity and correlations with neuropsychological tests, we included the training data as well as 74 subjects of unknown or inconsistent amyloid status. Two methods of generating outcome measures were used. (1) As a stand‐in for Tau load, similar to the centiloid scale, a score of 100, 50 or 0 was given to participants of the training set by multiplying CDR score by 100. Hierarchical and ridge regression were performed for each tracer using PET‐Braak SUVRs as predictive variables. Regression coefficients were used to calculate an outcome measure for all participants. (2) Principal component analysis (PCA) was applied for each tracer to the vectors of PET‐Braak regional SUVR of the training subset. Projections along the principal axis were calculated for both modalitiesResultHierarchical, ridge regression and PCA all yielded composite measures with good linear dependence between the two tracers (r2 for hierarchical, ridge and PCA were 0.84, 0.80, 0.80, respectively). Regions remaining in the hierarchical regression analyses varied between tracers: predictive regions for 18F‐MK6240 were PET‐Braak 1, 4 and 5, while 18F‐Flortaucipir used only PET‐Braak 1 and 5. Composite measures correlated well with MMSE and total MOCA score with all composite measures r < ‐0.61.ConclusionWhile both tracers may have different sensitivities to regional Tau distributions and disease stages, weighting multiple regions differently for each tracer gives a composite score linearly comparable across tracers that correlates with neuropsychological measures. This study is limited by the low total tau load and impairment level of all participants and would benefit from the inclusion of patients with more disease progression.

AbstractBackgroundLarge‐scale studies comparing sporadic early‐onset AD (EOAD, age<65) and late‐onset AD (LOAD, age≥65) are lacking. We compared amyloid‐PET outcomes (positivity rate and amyloid burden) between patients clinically diagnosed with sporadic EOAD vs LOAD, leveraging data from the Longitudinal Early‐Onset AD Study (LEADS) and the Alzheimer’s Disease Neuroimaging Initiative 3 (ADNI3).Method731 patients meeting the 2011 NIA‐AA criteria for AD dementia or MCI were included (505 early‐onset from LEADS, 226 late‐onset from ADNI3, Table 1). All participants underwent amyloid‐PET with [18F]Florbetaben or [18F]Florbetapir. Amyloid positivity was centrally determined by a process involving a visual read by a trained expert and PET‐only quantification; in case of a discrepancy, a read from an independent physician acted as a tiebreaker. Logistic regressions in each cohort examined relations between amyloid positivity and age, sex, MMSE and APOE4 genotype. Amyloid burden was independently quantified in Centiloids using an MRI‐based pipeline. Mean Centiloids in LEADS and ADNI were compared with two‐way ANOVA, for visually positive and visually negative scans.ResultAmyloid positivity rate was higher in LEADS (76%) than ADNI (64%, p<0.001, Figure 1A). Lower MMSE and APOE4 genotype increased odds of amyloid positivity in both cohorts, although the APOE4 effect was stronger in ADNI than LEADS (OR=10.1 versus 2.4, p=0.007, Table 2). Amyloid positivity was more common in females across cohorts, but this effect was only statistically significant in LEADS (Table 2). Centiloids were bimodally distributed in both cohorts, although the separation between positive and negative scans was more prominent in LEADS (Figure 1B). Visually positive scans had significantly higher Centiloids in LEADS than in ADNI, whereas no cohort difference was observed for visually negative scans (Figure 1C). Sensitivity analyses showed that this effect was driven by patients with MCI (CDR≤0.5; Figure 1D‐E).ConclusionThe lower amyloid positivity rate in ADNI might be due to AD‐mimicking pathologies being more common at an older age. The higher amyloid burden in early‐onset, amyloid‐positive patients could reflect younger patients being diagnosed later in the disease course compared to typical, late‐onset patients. Alternatively, younger patients might tolerate higher neuropathology burden due to higher brain reserve or fewer co‐pathologies.

AbstractBackgroundIn vivo studies using the tau PET tracers have shown high performance for the diagnosis of Alzheimer’s disease dementia and patterns of tracer uptake that resemble those observed in post‐mortem studies. However, tau tracers present distinct patterns of binding that might influence their performance in detecting AD pathology. In a head‐to‐head study, we investigated the performance of [18F]MK6240 and [18F]Flortaucipir for the diagnosis of AD.MethodWe assessed 132 individuals from the HEAD study (58 CU Aβ‐, 15 CU Aβ+, 14 MCI Aβ‐, 32 MCI Aβ+ and 13 AD dementia) with Aβ‐PET, [18F]MK6240 and [18F]Flortaucipir. Voxel‐wise receiver operating characteristic curves (ROC) of the two tau tracers were used to contrast groups provided the area under the curve (AUC) for disease diagnosis or biomarkers positivity.ResultThe brain maps showed numerically higher and more spread AUC for [18F]Flortaucipir to discriminate CU Aβ‐ from CU Aβ+ (Figure 1A). The difference between tracers’ AUC was greater in Braaks IV and V (Figure 2A), both regions that are not expected to have tau accumulation in CU Aβ+ individuals, reflecting a potential off‐target binding for [18F]Flortaucipir. To differentiate CU Aβ‐ from MCI Aβ+ individuals, [18F]MK6240 presented a numerically higher AUC than [18F]Flortaucipir in Braak I and II and similar AUC in other Braak regions (Figure 1B, 2B). We observed a high performance of [18F]MK6240 and [18F]Flortaucipir in differentiating AD dementia from CU Aβ‐ individuals. However, [18F]MK6240 exhibits a higher AUC than [18F]Flortaucipir in all Braak regions, especially Braak V‐VI (Figure 1C, 2C). Finally, [18F]MK6240 presented higher AUC in all Braak regions to discriminate MCI Aβ‐ from MCI Aβ+ individuals (Figure 1D, 2D).ConclusionOur results indicate that [18F]MK6240 and [18F]Flortaucipir present high accuracy to discriminate AD from CU Aβ‐ individuals. However, [18F]MK6240 presents higher AUC to discriminate AD and MCI Aβ+ from CU Aβ‐ individuals than [18F]Flortaucipir. Together, our head‐to‐head study sheds light on the distinct patterns of binding for Tau‐PET tracers.

AbstractBackgroundIdentification of cognitively unimpaired (CU) individuals who may progress to mild cognitive impairment (MCI), is a pressing issue in the Alzheimer’s disease (AD) field, since therapeutic interventions may be more effective in the absence of cognitive impairment and neurodegeneration. CU individuals positive for amyloid and tau PET are very likely in the AD pathway. In out‐patient cognitive screening, we use rapid and simple tests such as The Montreal Cognitive Assessment (MoCA) ‐ a composite of executive, visuospatial, naming, attention, language, abstraction, delayed recall, and orientation performances. We hypothesize that sub‐categorizing the CU group based on the presence of amyloid (A) and tau (T) pathologies, as measured by PET, can reveal subtle cognitive deficits, even in a clinical test with low sensitivity such as MoCA.MethodWe included 88 CU [defined as CDR global score of 0] older adults with available head‐to‐head MK6240 tau‐PET and Flortaucipir tracers from the ongoing HEAD multi‐site observational study. They also underwent Amyloid PET imaging using PiB or NAV4694. MoCA testing was done around the imaging time. To increase similarity with clinical practice, Amyloid and Tau PET positivity was defined using visual reads. Based on tracer uptake status, we categorized the individuals into A‐T‐, A‐T+, A+T‐ and A‐T‐ groups. Unpaired t‐test and one‐way ANOVA were used to test for significant differences between groups.Result2‐tailed p‐values were significant between the A‐T‐ group and the A+T+ group for MoCA total score [p: 0.0025 and p: 0.0025]. When we stratified MoCA in the different cognitive domains, we found that the results were driven by MoCA (memory) Delayed Recall score [p: 0.0007 and p: 0.0162]. A‐T+ vs A+T+ was also significant [p: 0.0099] for MoCA Delayed Recall. Both tau tracers (MK6240 and Flortaucipir) showed similar performance for determining T+ and identifying cognitive decline.ConclusionMoCA, a simple and commonly administered in‐office neuropsychological test, can detect subtle cognitive dysfunction in CU older adults who are Aβ and tau positive (A+/T+). These CU A+/T+ individuals are likely on the path to developing AD dementia.

AbstractBackgroundThe HEAD study aims to collect a large dataset of multiple tau‐PET tracers to provide robust anchor values for tau‐PET harmonization. Here, we tested the hypothesis that anchoring two tau tracer uptake values using head‐to‐head measurements has the potential to generate an accurate universal tau‐PET scale, named Uniτ(tau).MethodsWe assessed 200 individuals across the aging and AD spectrum (Training:HEAD data freeze 2.0, n=185; Testing:UPitt dataset (Gogola et al.), n=15) with [18F]Flortaucipir and [18F]MK‐6240 tau‐PET. SUVRs were processed to a common 8mm FWHM, with inferior cerebellar gray matter as the reference region (Pascoal et al.). Uniτ explored two anchoring/harmonization methods. First, we examined within‐tracer anchoring by creating anchor values based on the mean SUVR of Youngs (<25 years) and 95th percentile voxels from cognitively impaired individuals. Second, we explored within‐ plus between‐tracer anchoring, employing linear (e.g., piecewise) and non‐linear regressions. To address the inherent problem of discontinuity of piecewise transformations, we implemented two smoothing methods at the inflection point between equations, transforming them into a continuous function.ResultsUniτ scale anchoring within‐tracer resulted in similar estimates for high values, but less accurate in the lower range (Figure 1,2). Anchoring within‐ plus between‐tracer improved estimate consistency, with the piecewise transformation generating the best results. The piecewise smoothing equation yielded estimates comparable to those obtained from the piecewise method without smoothing. This allowed for the use of a single formula. In addition, this leads to more robust results when the goal is to study longitudinal changes in the scale (data not shown). UPitt testing dataset showed similar results to the training set (Figure 3).ConclusionOur preliminary findings suggest that anchoring tau‐PET values both within and between tracers has the potential to harmonize tau‐PET tracers, while preserving their characteristics. Currently, piecewise smoothing is the preferred method for Uniτ, but we are continuously fine‐tuning scale parameters as we acquire more data. The final scale parameters will be determined based on extensive training and testing data from multiple tracers. This cautious methodology holds the promise of delivering reliable, robust, and reproducible results, ensuring safe usage of the scale in clinical trials, and potentially paving the way for future use in clinical practice.

AbstractBackgroundThe association between [18F]Flortaucipir (FTP) and [18F]MK6240, two commonly used tau‐PET tracers in Alzheimer’s disease (AD), varies due to distinct binding properties and off‐target signal regions. Our study aims to elucidate the biological factors influencing this association and evaluate the applicability of a common equation across different on‐target regions.Method113 individuals from the HEAD dataset (11 young, 58 cognitively unimpaired elderly, and 44 cognitively impaired) underwent [18F]MK6240, [18F]FTP and Aß‐PET scans. Images were processed using the inferior cerebellar grey (CG) as the reference region, with subjects categorized by Aß status. Stepwise linear fitting identified off‐target (OFF) regions impacting tracer associations within the whole cortex, Braak and temporal‐meta regions of interest (ROIs). OFF regions included were a composite mask (pallidum, putamen, caudate ‐ PPC) to address collinearity in subcortical nuclei, as well as the region with the greatest partial volume effects (PVE) on the ROI (meninges, choroid plexus, and the hotspot medial to entorhinal cortices (EC)). Akaike information criteria (AIC) analyses tested linear models with various off‐target regions.ResultPPC and PVE regions influenced tracer association in on‐target regions. PPC binding is associated with age. PPC and meninges retention, along with Aß status interaction, demonstrated the best model fit for the whole cortex (Figure 1). Weighting factors derived from the whole cortex model were applied to Braak regions and temporal meta‐ROI. Braak III‐VI and temporal‐meta ROI showed optimal models including OFF regions and Aß status interaction (Figure 2). Forcing the whole equation gave optimal results. For Braak I (EC), adding the OFF covariates only was the best model. For Braak II (hippocampus), the best model included OFF regions and the interaction with Aß status (Figure 2). Forcing the whole cortex equation onto Braak I and Braak II yielded suboptimal fits.ConclusionHarmonizing tau‐PET tracers involves considering significant covariates. Integrating meningeal and age‐related off‐target regions enhances [18F]MK6240 and [18F]FTP association in on‐target regions, further improved by Aß status inclusion. While the whole cortex model may not suit early ROI, it proves effective for Braak III and beyond. Creating an equation for tracer translation necessitates incorporating meningeal and subnuclei off‐target retention.

AbstractBackgroundNormative models (NM) of brain metrics based on large, diverse populations offer novel strategies to detect individual brain abnormalities. To create an age‐dependent statistical model of brain microstructure over the human lifespan, we built the largest multi‐site NM of white matter (WM) diffusion tensor imaging (DTI) metrics based on 54,591 subjects. We used state‐of‐the‐art tools to adjust for site‐dependent effects.MethodWe used Hierarchical Bayesian Regression (HBR) to determine the age trajectory and lifespan centile curves (ages 3‐100 years) by merging data from different sections of the lifespan. We analyzed 19 international public datasets with the ENIGMA‐DTI protocol. Mean fractional anisotropy (FA), mean, axial, and radial diffusivity (MD, AxD, RD) were extracted for 21 bilateral ROIs from the JHU‐WM atlas and the whole WM skeleton. Regressions were run with each metric per ROI as a function of age and sex, and the DTI protocol as the batch effect (37 protocols). Z‐scores were derived for 81 patients with Alzheimer’s disease (AD; mean age: 77y±8.4, 46M/35F) and 225 MCI participants (mean age: 75.1y±8.1, 127M/98F) and extreme deviations (Z>|2|) were quantified for each condition. ROI‐wise areas under the ROC curve were calculated to determine the classification accuracy of the Z‐scores.ResultThe average FA for the overall WM skeleton reached its peak value at 27 years, with minima for MD, RD and AxD at 47, 42 and 54 years, respectively. We found extreme FA deviations (Z>2) for MCI (AUC=0.67) and AD (AUC=0.66) in the genu of the corpus callosum, extreme RD deviations (Z<2) in the tapetum (AUC=0.73) and Extreme AxD deviations (Z>2) for MCI in the superior longitudinal fasciculus (AUC=0.65). See Figures 1 & 2.ConclusionNM of brain microstructure with HBR yielded lifespan trajectories for widely‐used DTI metrics. By pooling large scale multi‐site dMRI to create a statistical reference, single subject deviations are detectable across the lifespan for AD and MCI. These NM will be publicly available to the research community.

AbstractBackgroundThe HEAD study focuses on collecting an extensive dataset from various tau‐PET tracers, aiming to establish robust anchor values, which are essential for harmonizing tau‐PET measurements. Here, we aim to showcase the capability of converting 3D tau‐PET images into a common scale using the Universal Tau‐PET Scale, Uniτ (tau), and to use these 3D images to subsequently obtain ROIs as needed.MethodsWe assessed 185 individuals across the aging and AD spectrum from the HEAD study, with [18F]Flortaucipir and [18F]MK‐6240 tau‐PET tracers. Tau‐PET SUVR images were standardized to a common 8mm FWHM, using the inferior cerebellar gray matter as the reference region. We generated Uniτ Tau‐PET 3D images using a single formula based on Meta‐temporal parameters in two steps: first, within‐tracer anchoring based on young (<25 years) and cognitively impaired individuals; second, between‐tracer anchoring using a piecewise transformation from [18F]Flortaucipir to [18F]MK‐6240, with smoothing at the piecewise inflection point. Subsequently, we extracted mean Uniτ values from these 3D images for key ROIs, including meta‐temporal, mesial, temporo‐parietal, frontal, and Braak stages III‐VI. Finally, we correlated Uniτ values across ROIs between the two tracers to evaluate the accuracy of estimates from the voxel‐wise transformation.ResultsThe original SUVR images present large differences between the two tau‐PET tracers (Figure 1). However, upon applying the Uniτ piecewise transformation with smoothing to all brain voxels, we were able to reasonably harmonize these images to the Uniτ scale, substantially reducing visual variability. Notably, the mean Uniτ values for the ROIs extracted from these harmonized 3D tau‐PET brain images demonstrated a high level of association between the two tracers (Figure 2). Furthermore, estimates generated from ROIs or extracted from our 3D parametric Uniτ model yielded identical estimates (Figure 3).ConclusionThe strong associations between tracers after directly harmonizing 3D images to Uniτ scale using the piecewise transformation with smoothing, underscore the effectiveness of the proposed method. This approach provides a reliable and standardized way to compare tau‐PET data across different tracers. Our results indicate the feasibility of harmonizing 3D tau‐PET images without relying on pre‐established ROIs, overcoming the limitation of restricting the analysis to only few brain regions.

AbstractBackgroundDefault mode network (DMN) resting state connectivity has been correlated with heightened amyloid and tau – hallmarks of Alzheimer’s Disease (AD). Tau is postulated to impact a meta‐temporal area including DMN‐associated regions like amygdala, entorhinal cortex, fusiform gyrus, parahippocampus, inferior temporal, and middle temporal gyrus. We recruited individuals with varying cognitive status to undergo resting state connectivity and imaging with two tau tracers (Flortaucipir and MK6240). We investigated the association between DMN connectivity and both global and regional (meta‐temporal) tau.MethodsWe studied 149 adults who were young cognitively healthy (n = 10, 21‐25 years, 5F), older cognitively healthy (n = 85, 50‐83 years, 51F), had MCI (n = 43, 53‐84 years, 23F), or had AD (n = 11, 62‐84, 6F). All individuals underwent resting state imaging and PET imaging with both Flortaucipir and MK6240 tau tracers. Resting data was preprocessed and DMN connectivity was estimated using bilateral posterior cingulate gyrus as region of interest. Tau tracer uptake was preprocessed (Pascoal. et al), and voxel‐wise standardized uptake value ratio (SUVR) was computed using inferior cerebellar gray matter as reference. We computed universal and meta‐temporal uptake for both tracers. We conducted four voxel‐wise associations between DMN connectivity and uptake in each tracer. We corrected for multiple comparisons using SnPM13 to estimate non‐parametric p‐values (10,000 permutations), we used a cluster forming threshold of p = 0.001 and corrected for multiple comparisons by controlling the family‐wise error rate (FWE) at p = 0.05.ResultsGreater universal Flortaucipir uptake was associated with lower within‐network DMN connectivity (bilateral posterior cingulate, precuneus, angular/supramarginal gyrus, and inferior parietal) but greater between‐network DMN connectivity (superior parietal, postcentral gyrus; Figure 1A). Greater universal MK6240 uptake was associated only with greater DMN‐cerebellum connectivity (Figure 1B). Greater meta‐temporal Flortaucipir was associated with lower within‐network DMN connectivity similar to universal uptake with addition of medial prefrontal cortex and middle temporal gyrus (Figure 2A). Greater meta‐temporal MK6240 was only associated with greater DMN‐superior parietal connectivity (Figure 2B).ConclusionsThese preliminary findings show that uptake using different tau tracers may lead to varying conclusions on association with connectivity. This may have implications for understanding the role of tau in altering resting state connectivity findings.