Longitudinal multicenter head-to-head harmonization of tau PET tracers

Project Summary/Abstract: The development of tau positron emission tomography (PET) tracers has yielded the opportunity to better understand the tau accumulation associated with the development of Alzheimer’s disease (AD), improve the diagnostic accuracy of AD, and test the effects of therapeutic interventions in tau deposition. Our group and

others have shown that [18F]Flortaucipir has distribution patterns similar to those reported in postmortem studies and shows significant rates of tau accumulation over time despite brain off-target signal in white matter, basal ganglia, and choroid plexus. [18F]MK-6240, a second-generation tau tracer, has ~6-fold greater affinity for tau

tangles than [18F]Flortaucipir and negligible brain off-target signal. However, [18F]MK-6240 has off-target uptake in the meninges adjacent to the entire cortex, which can compromise the signal within brain regions. We have shown that [18F]MK-6240 has similar patterns of uptake as postmortem and [18F]Flortaucipir studies. However,

[18F]MK-6240 has a higher dynamic range and, although measured in different cohorts, the annual change in tau using [18F]MK-6240 appears to be greater than changes measured using [18F]Flortaucipir. To date, more than 95% of sites in the US performing tau PET studies are using one of these tracers. Although both tracers offer

robust tau estimates, these large differences in binding characteristics between them can lead to misleading interpretations of their outcomes and preclude the merging of their datasets using simple conversion methods. Thus, a well-powered longitudinal study assessing head-to-head [18F]Flortaucipir and [18F]MK-6240 scans has

the potential to: (1) standardize their analysis, allowing datasets to be combined, (2) compare their rates of longitudinal deposition to elucidate their advantages and limitations for research, trials, and practice, and (3) produce a benchmark dataset to be used by the scientific community to develop methods for PET quantification

and harmonization. Here, we propose a multi-site longitudinal study in which 620 subjects (40 young controls, 280 cognitively unimpaired elderly, 200 mild cognitive impairment, 100 AD dementia) will receive [18F]Flortaucipir and [18F]MK6240 scans at baseline and 18 months later. At each time point, subjects will also have an amyloid-

β (Aβ) PET scan, magnetic resonance imaging, cognitive tests, and a blood draw for plasma tau analyses. Methods will be harmonized with ADNI and the ADRC program. In Aim 1, we will (1) standardize processing methods, (2) convert to a common scale, (3) compare associations with Aβ, atrophy, and cognition, and (4)

compare Braak staging between tau tracers using cross-sectional data. In Aim 2, we will (1) ascertain the optimal processing method for longitudinal analysis and (2) compare longitudinal accumulation between tau tracers and its associations with changes in Aβ, atrophy, and cognition. As recent results from our group and others have

shown that the new p-tau plasma assays identify brain tau pathology, in exploratory Aim 3, we will compare cross-sectional and longitudinal [18F]Flortaucipir and [18F]MK6240 estimates with plasma p-tau outcomes.