Repository logo

In vivo pathology markers in tauopathies: prognostic and diagnostic implications



Change log



The neurodegenerative tauopathies include Alzheimer’s disease, frontotemporal dementia (half due to tauopathy), and progressive supranuclear palsy (PSP). In addition to abnormal accumulation of tau protein, they are each characterised by neuroinflammation, with increasing evidence that the neuroinflammation plays a role in the onset and progression of these diseases, rather than simply being a response to neuronal injury. PET and MRI allow the in vivo visualisation and quantification of markers of neuropathology and neurodegeneration. However, the value of these in vivo measures at different stages of disease, and their value for predicting the clinical progression in patients, remains undetermined. In my thesis, I examine the prognostic value of multimodal in vivo PET imaging for predicting longitudinal clinical and cognitive decline in Alzheimer’s disease, frontotemporal dementia and PSP. I use the PET ligand [11C]PK11195 PET as a marker of microglial activation, and [18F]AV-1451 PET to quantify tau pathology (or in the case of frontotemporal dementia, tau and/or TDP-43 pathology), with structural 3T MRI as an index of atrophy. In patients with Alzheimer’s dementia (n=12) and its prodrome, mild cognitive impairment with amyloid positivity (n=14), I applied latent growth curve models (LGCMs), multiple linear regression and Bayesian regression analyses to test the prognostic value of PET and MRI, alone and in combination, to predict cognitive decline over three years. Tau burden and microglial activation in temporo-parietal cortical regions were identified as significant predictors of cognitive decline, over and above baseline atrophy. In frontotemporal dementia, I undertook three complementary analyses. First, in pre-symptomatic carriers of genetic mutations associated with frontotemporal dementia (i.e. in MAPT, GRN, or C9orf72 gene), I applied LGCMs to test for associations between cross-sectional and longitudinal changes in atrophy, apathy and cognition. I compared gene mutation carriers (n=304) to non-carrier relatives as controls (n=296). My results suggest that apathy occurs early in disease progression and worsens over time in those considered to be still asymptomatic carriers. Worsening of apathy over time was associated with early brain changes in frontal and cingulate cortex, and predicted subsequent deterioration of cognitive performance. Second, in a case series of post-symptomatic frontotemporal dementia mutation carriers (n=7), I assessed [11C]PK11195 and [18F]AV-1451 PET uptake, and discuss their association with clinical phenotype. The genetic cases each showed neuroinflammation in frontotemporal regions, and [11C]PK11195 distribution reflected clinical symptoms in each patient. Last, in a larger group of genetic and sporadic frontotemporal dementia cases (n=30), I applied a linear mixed effects model and regression analyses to test the prognostic value of [11C]PK11195, structural MRI and apathy assessment on cognitive decline up to 4 years of follow-up. Across the frontotemporal dementia spectrum, in vivo atrophy and neuroinflammation markers in frontal regions, and apathy severity at baseline, were associated with faster cognitive decline. Finally, I examined the prognostic value of MRI and PET markers in PSP (n=17), as a model of primary non- Alzheimer’s tauopathy. First, I studied the in vivo association between tau accumulation and neuroinflammation, finding that they co-localise in subcortical and cortical regions, previously described as affected by PSP-related neuropathology. Next, I applied linear mixed effects and regression models to test the prognostic value of baseline PET markers for microglial activation and tau pathology, and atrophy-related MRI measures on clinical progression over time. Inflammation in subcortical regions was strongly associated with cross-sectional clinical severity and faster clinical worsening, similar to [18F]AV-1451 binding in the same regions. However, atrophy measures were not related to clinical progression over time. Together these studies demonstrate that inflammation occurs in Alzheimer’s disease, frontotemporal dementia and PSP. Inflammation as measured by [11C]PK11195 PET in regions known to be affected pathologically in the disease can predict cognitive decline and clinical progression over time, independent of tau burden and atrophy. In frontotemporal dementia, apathy and structural brain changes also resulted to be predictive of clinical changes in pre- and post-symptomatic patients. This work provides evidence supporting future trials of immune-mediated strategies, with the aim of modifying the course of disease, which might be coupled with improved stratification or individualised treatment approaches based on cognitive and imaging assessment at baseline. I suggest these therapeutic approaches would be more effective if given early in the disease course.





O'Brien, john
Rowe, James


Tauopathies, Positron Emission Tomography (PET), Magnetic Resonance Imaging (MRI), Neuroinflammation, Microglial activation, Tau pathology, Neurodegeneration, Apathy, Longitudinal data analysis, Predictive modelling, Cognitive decline, Clinical progression, Alzheimer's disease, Frontotemporal dementia, Progressive supranuclear palsy


Doctor of Philosophy (PhD)

Awarding Institution

University of Cambridge
Cambridge University Hospitals NHS Foundation Trust (CUH) (146281)
Wellcome Trust (103838/Z/14/Z)
Cambridge Trust Vice-Chancellor’s Award and Sidney Sussex College Scholarship