The utility of tau PET imaging in non-Alzheimer's disease (AD) tauopathies like behavioural frontotemporal dementia (bvFTD), which is mainly underlain by TDP-43 or tau pathology, remains debated. We aim to test the hypothesis that [18F]-AV-1451 tau PET using later than usual acquisition times, which have previously been shown in AD to allow to get closer to tracer equilibrium between the reference region and high-binding structures, and could be better suited to the lower affinity of this tracer for the straight tau filaments present in non-AD tauopathies, would allow to detect cortical tau pathology in a fraction of bvFTD patients and in patients with non-fluent primary progressive aphasia (nfPPA, most often underlain by tau pathology). Sixteen AD patients, 11 controls, 7 bvFTD patients (including a carrier of a GRN mutation leading to TDP-43 pathology) and 2 nfPPA patients were included. We compared SUVr obtained at the usual early time window for [18F]-AV-1451 PET acquisition (ET: 80-100 min) to a later acquisition window (LT: 190-210 min) between groups. Compared with ET, [18F]-AV-1451 LT uptake in AD patients was significantly higher in the temporo-parietal cortex, and lower in subcortical regions. The LT window allowed to detect significantly increased tau binding in the frontal or temporal cortex in 3 bvFTD patients and in the 2 nfPPA patients that was not detectable with ET. The GRN mutation carrier showed no significant increase of tracer binding. [18F]-AV-1451 late acquisition window could allow cortical binding to abnormal tau deposits to be revealed in a subset of bvFTD patients, which may distinguish them from the TDP-43 subtype.
Keywords: Frontotemporal dementia; Neurofibrillary tangles; TDP-43; Tau PET imaging; Tau pathology.