PET tracers for pathological deposits of tau protein aggregates in the brain are currently undergoing rapid development. They are targeting pathological deposits of tau aggregates, which are present in several neurodegenerative disorders, the socalled tauopathies. The most frequent of these is Alzheimer's disease, which is characterised by a combination of neurofibrillary tangles (primarily tau) and plaques (primarily beta-amyloid). However, pathological deposits of tau (without beta-amyloid plaques) also are present in rarer diseases, including progressive supranuclear palsy (PSP), cortico-basal degeneration (CBD), some types of frontotemporal dementia (FTD), and secondarily also in posttraumatic encephalopathy [2,12]. Most publications on PET tau imaging relate to Alzheimer's disease only, while a paper by Chiotis et al. in this issue also included a few cases of non-AD tauopathies.The paper should be considered within the background of the main clinical research needs and expectations for tau imaging. The most obvious ones are related to the potential of tau PET as an imaging biomarker in clinical trials of therapeutic compounds to delay the onset or progression of dementia in AD. Alternatively, increased levels of tau in CSF are also being used as a biomarker of pathological tau deposition in such trials. CSF sampling includes the option to also assess abnormal tau phosphorylation as well as other potential biomarkers, such as inflammatory cytokines [5]. Longitudinal studies will need to assess the merits of imaging the regional brain distribution of tau PET deposits compared to the analysis of their molecular diversity in CSF. Most importantly, a tau biomarker should measure disease progression at a prodromal stage where symptoms are absent or mild and do not show much change within typical trial durations. Tau PET could also help with accurate staging of prodromal AD, analogous to the pathological staging proposed and validated by Braak and Braak [10]. From a diagnostic perspective, tau PET would be expected to separate tauopathies from diseases associated with other pathological protein deposits such as Lewy bodies, TDP43, FUS, and prion protein [11], and to discriminate between different tauopathies by their regional distribution patterns [8].What does the paper by Chiotis et al. demonstrate? A main result is an excellent discrimination between healthy controls and AD patients. However, patients were relatively young with a maximum age of 74 years (whereas most AD patients are older) and the sample of controls was quite small and even younger, including only four elderly controls with a maximum age of 65. Thus, the study could not address the issue of discriminating between primary age-related tauopathy (PART), which also increases substantially with old age, and preclinical or prodromal AD. PART is characterised by mesial temporal tau deposits in the absence of beta-amyloid plaques and is usually regarded as a relatively benign age-related condition that is different from AD [6]. This issue has been add...