Sensitivity–Specificity of Tau and Amyloid β Positron Emission Tomography in Frontotemporal Lobar Degeneration

Abstract

Objective: To examine associations between tau and amyloid β (Aβ) molecular positron emission tomography (PET) and both Alzheimer-related pathology and 4-repeat tau pathology in autopsy-confirmed frontotemporal lobar degeneration (FTLD). Methods: Twenty-four patients had [ 18 F]-flortaucipir-PET and died with FTLD (progressive supranuclear palsy [PSP], n = 10; corticobasal degeneration [CBD], n = 10; FTLD-TDP, n = 3; and Pick disease, n = 1). All but 1 had Pittsburgh compound B (PiB)-PET. Braak staging, Aβ plaque and neurofibrillary tangle counts, and semiquantitative tau lesion scores were performed. Flortaucipir standard uptake value ratios (SUVRs) were calculated in a temporal meta region of interest (meta-ROI), entorhinal cortex and cortical/subcortical regions selected to match the tau lesion analysis. Global PiB SUVR was calculated. Autoradiography was performed in 1 PSP patient, with digital pathology used to quantify tau burden. Results: Nine cases (37.5%) had Aβ plaques. Global PiB SUVR correlated with Aβ plaque count, with 100% specificity and 50% sensitivity for diffuse plaques. Twenty-one (87.5%) had Braak stages I to IV. Flortaucipir correlated with neurofibrillary tangle counts in entorhinal cortex, but entorhinal and meta-ROI SUVRs were not elevated in Braak IV or primary age-related tauopathy. Flortaucipir uptake patterns differed across FTLD pathologies and could separate PSP and CBD. Flortaucipir correlated with tau lesion score in red nucleus and midbrain tegmentum across patients, but not in cortical or basal ganglia regions. Autoradiography demonstrated minimal uptake of flortaucipir, although flortaucipir correlated with quantitative tau burden across regions. Interpretation: Molecular PET shows expected correlations with Alzheimer-related pathology but lacks sensitivity to detect mild Alzheimer pathology in FTLD. Regional flortaucipir uptake was able to separate CBD and PSP.

In Vivo 18F-Flortaucipir PET Does Not Accurately Support the Staging of Progressive Supranuclear Palsy

“…Off-target binding is well-characterised for 18 F-flortaucipir, but this problem alone would still leave open the utility to quantify tau pathology in areas without significant mono-amine oxidase levels or neuromelanin, such as cerebellum and medial frontal cerebral cortex (22). However, recent PET-to-autopsy correlational studies suggested that 18 F-flortaucipir PET does not reliably correspond to post-mortem tau pathology in non-Alzheimer's tauopathies (13,23). This suggests that 18 F-flortaucipir lacks sensitivity in non-Alzheimer tau pathology.…”

Autopsy Validation of Progressive Supranuclear Palsy‐Predominant Speech/Language Disorder Criteria

“…Neuropathological evaluations were performed in accordance with current diagnostic protocol, as previously described. 19,20 Pathological diagnoses were rendered according to established criteria: PSP was diagnosed based on the presence of characteristic neuronal (pretangles and tangles) and glial lesions (tufted astrocytes and oligodendroglial coiled bodies) in vulnerable cortical and subcortical regions, 21 and CBD by the presence of cortical and subcortical neuronal and glial lesions (astrocytic plaques) and thread-like processes in gray and white matter 22 ; FTLD-TDP type A had TDP-43-immunoreactive neuronal cytoplasmic inclusions, dystrophic neurites, and intranuclear inclusions, while type B had predominantly neuronal cytoplasmic inclusions. 23 PiD showed argyrophilic and tau-immunoreactive Pick bodies, which were negative on Gallyas silver stain.…”

Current directions in tau research: Highlights from Tau 2020

“…For example, current tau PET tracers cannot accurately distinguish between Braak stage 0 (absence of tauopathy) and Braak stage 1 (early tauopathy in the entorhinal cortex in the medial temporal lobe). 152,153,155,156 Until very recently, researchers also have been unable to pinpoint the precise area in the brain where tau accumulation in AD first begins using tau PET imaging. However, Sanchez et al recently reported findings that suggest that cortical tau pathology may begin in a region of the brain they call the "rhinal cortex"-a region defined anatomically using brain surface anatomy that is sampled during PET imaging.…”

“…The development of tau-specific PET ligands has made it possible to investigate the measurement of tau deposition not only in AD but also in a wide range of non-AD tauopathies, including PSP, 155,[159][160][161][162] CBD, 155,[163][164][165][166] and Pick's disease. 155 However, FTP and other currently available tau tracers, when used to image non-AD pathology, have generally demonstrated low-affinity binding, significant off-target binding, and other characteristics that have resulted in limited sensitivity and specificity. 165,[167][168][169] These in vivo results converge with autoradiography studies that have shown absent-to-low binding of FTP to non-AD tauopathies.…”

“…Although tau PET may be useful for accurately assessing advanced stages of tau pathology, it has not yet been proven sensitive enough to accurately detect early pathology. For example, current tau PET tracers cannot accurately distinguish between Braak stage 0 (absence of tauopathy) and Braak stage 1 (early tauopathy in the entorhinal cortex in the medial temporal lobe) 152,153,155,156 . Until very recently, researchers also have been unable to pinpoint the precise area in the brain where tau accumulation in AD first begins using tau PET imaging.…”