Association of [18F]Flortaucipir-PET and plasma p-tau217 with tau neuropathology in AD and other neurodegenerative disorders

Abstract

Background and objectives: Evaluating tau biomarkers in patients with available autopsy data is critical for validating their sensitivity and specificity to detect Alzheimer's disease neuropathologic changes (ADNC). We examined the association between tau-PET (using [¹⁸F]Flortaucipir), plasma ptau-217, and measures of AD neuropathology in a group of clinically-impaired participants from a tertiary dementia center, including patients with AD and FTLD.

Methods: We analyzed Flortaucipir-PET (80-100 minutes post-injection acquisition, normalized to inferior cerebellar cortex) from 73 patients with a clinical diagnosis of various neurodegenerative diseases who underwent autopsy at the Neurodegenerative Disease Brain Bank (median [IQR] age: 67 [59, 73] years, 60% male, median [IQR] PET-to-autopsy duration: 3.9 years, [2.1, 5.1]). Standardized uptake value ratios (SUVRs) were extracted from the entorhinal cortex (an early tau region) and the temporal meta-ROI (a widely used AD signature region). A semi-quantitative rating of AD NFT burden in cortical areas was available for 56 participants. Plasma p-tau217 was quantified with Simoa (Janssen^®) in 56 participants (median [IQR] PET-to-plasma duration: 1.7 months [0.5, 4.1]).

Results: Our cohort included patients with a primary pathological diagnosis of AD (n=39), FTLD (tauopathies n=28, non-tauopathies n=4), and Lewy body dementia (n=2). Flortaucipir SUVRs were elevated in patients with a neuropathological diagnosis of AD compared with non-AD patients. Consistently elevated PET signal was detected in both the entorhinal cortex and temporal meta-ROI of patients with neurofibrillary tangle (NFT) at Braak stage VI and with high ADNC levels. No elevation of Flortaucipir SUVRs was observed at intermediate levels of ADNC. The burden of AD NFTs was correlated with local Flortaucipir SUVRs across cortical brain regions. Plasma p-tau217 concentrations were increased in patients at Braak stage V and VI and strongly correlated with Flortaucipir SUVRs across ROIs (r's≥0.75), driven by Braak V-VI cases. Flortaucipir SUVRs and plasma p-tau217 concentrations identified high Braak stages (V-VI) and high/intermediate ADNC levels with similarly high performance (area under the curve≥ 0.90).

Discussion: Flortaucipir-PET and plasma p-tau217 both displayed strong specificity for primary AD neuropathological diagnosis but lacked sensitivity to detect early AD-related tau co-pathology in patients with a non-AD diagnosis.

Keywords: Alzheimer’s disease; frontotemporal lobar degeneration; neurofibrillary tangle; neuropathology; plasma p-tau; positron emission tomography; tau.

Figure 1,. Association of Flortaucipir SUVRs with neuropathological diagnoses

A, Entorhinal FTP SUVR by neuropathologic diagnosis. B, Temporal meta-ROI SUVR by neuropathologic diagnosis. The horizontal line represents the cut-off for SUVR positivity derived from the non-autopsy control group (mean + 2SD). Individual points are color-coded by their NFT Braak stage. Abbreviations: AGD, argyrophilic grain disease; CBD, corticobasal degeneration; CTE, chronic traumatic encephalopathy; FTLD-FUS, frontotemporal lobar degeneration due to a FUS mutation; FTLD-MAPT, frontotemporal lobar degeneration due to MAPT mutation; FTLD TDP-43, frontotemporal lobar degeneration due to TDP-43; FTP, Flortaucipir; LBD, Lewy body dementia; PSP, progressive supranuclear palsy; SUVR, standardized uptake value ratio.

Figure 2,. Individual Flortaucipir SUVR images and w-maps for Braak stages III to VI cases

Individual images are presented as axial slices of SUVR images on the left column and individual w-maps on the right column, with each line being one case. *Cases described in Soleimani-Meigooni et al., Brain 2020. All Braak IV (n=3) and V (n=3) cases of the cohort are included. Visual read (VR) per FDA guidelines and quantification (SUVR) positivity (+) and negativity (−) are indicated for each case. W-maps are voxel-wise maps comparing single-subject images to control individuals, adjusting for age and sex (La Joie et al., Journal of Neurosciences 2012). Abbreviations: CBD, corticobasal degeneration; FTLD-MAPT, frontotemporal lobar degeneration due to MAPT mutation; FTP, Flortaucipir; LBD, Lewy body dementia; NFT, neurofibrillary tangles; PSP, progressive supranuclear palsy; SUVR, standardized uptake value ratio.

Figure 3,. Association of Flortaucipir SUVRs with AD neuropathological changes

Entorhinal cortex SUVR (top row) and temporal meta-ROI SUVR (bottom row) by A, NFT Braak stages; B, Thal phases; C, CERAD score and D, levels of ADNC. The horizontal line represents the cut-off for SUVR positivity derived from the control group (mean + 2SD). The n=2 CTE cases were NFT Braak non-conformant, and ADNC levels could not be determined for one of these cases. The Thal phase could not be determined for n=1 CBD case, and subsequently, ADNC levels could not be established. Abbreviations: ADNC, AD neuropathologic changes; FTP, Flortaucipir; NFT, neurofibrillary tangles; SUVR, standardized uptake value ratio.

Figure 4,. Association of Flortaucipir SUVRs and AD NFT burden

A, Scatter plot displaying the association of the global tau SUVR with AD NFT burden. Correlations were analyzed with Spearman’s rank correlations (r). Individual points are color-coded by NFT Braak stages. The global tau ROI corresponds to a large ROI encompassing all Braak stage regions. The average AD NFT burden corresponds to the average burden across 13 cortical brain regions. B, Bar chart displaying Spearman’s rank correlation coefficients between FTP SUVRs and AD NFT burden within the 13 cortical brain regions. Anterior midcingulate: P=0.016, CA2: P=0.0004. Other regions: P<0.0001. Abbreviations: FTP, Flortaucipir; NFT, neurofibrillary tangles; SUVR, standardized uptake value ratio.

Figure 5,. Plasma p-tau217 concentrations by diagnosis and by AD neuropathological scores

A, Plasma p-tau217 concentrations by neuropathological diagnosis. Plasma p-tau217 concentrations by B, NFT Braak stages; C, Thal phases; D, CERAD score and E, ADNC levels. Abbreviations: AD, Alzheimer’s disease; AGD, argyrophilic grain disease; CBD, corticobasal degeneration; CERAD, Consortium to Establish a Registry for Alzheimer’s Disease; CTE, chronic traumatic encephalopathy; FTLD-FUS, frontotemporal lobar degeneration due to a FUS mutation; FTLD-MAPT, frontotemporal lobar degeneration due to MAPT mutation; FTLD-TDP-43, frontotemporal lobar degeneration due to TDP-43; FTP, Flortaucipir; LBD, Lewy body dementia; NFT, neurofibrillary tangles; PSP, progressive supranuclear palsy.

Figure 6,. Association of plasma p-tau217 concentrations with Flortaucipir SUVRs and performance to identify AD pathology

A, Correlation between plasma p-tau217 concentrations and FTP SUVR ROIs. Correlations were studied using Spearman’s rank correlation (r) in the whole cohort and in the AD subgroup. B, ROC analysis of FTP SUVR and plasma p-tau217 performance to identify advanced Braak stages (V/VI) and intermediate/high levels of ADNC. Age, sex, and plasma/PET-to-autopsy delay were added as covariates. Comparison of AUC using the DeLong test: p>0.05. Abbreviations: ADNC, Alzheimer’s disease neuropathologic change; AUC, area under the curve; FTP, Flortaucipir; Se., sensitivity; Spe., specificity; SUVR, standardized uptake value ratio.