190 Chapter 7 DISCUSSION We assessed the associations between tau pathology and neuronal injury (reflected by measures of cortical thickness and rCBF) over time. We showed that higher tau pathology at baseline (especially in the Braak III/IV region) was associated with faster cortical thinning over time in Aβ+ individuals. Annual change in tau pathology did not show associations with cortical thinning over time, but larger increases in tau pathology did show associations with larger increases in R1 over time in inferolateral and superior parietal regions in Aβ+ individuals. Our results are in line with disease models proposing that tau load is a key driver of local cortical thinning and stress the need for future longitudinal studies into the role of rCBF in the pathophysiological process of AD. Furthermore, our results indicates that a single tau PET scan at baseline best predicts cortical thinning over time when compared to longitudinal tau PET imaging. This in turn highlights the potential of a single tau PET to improve the prognosis and selection of the right target population for clinical trials. Our findings of higher tau pathology at baseline predicting cortical thinning over time in Aβ+ individuals are in line with previous studies, demonstrating local and non-local associations between tau pathology and longitudinal cortical thinning in Aβ+ individuals [2, 5, 11, 12, 40]. Regions most commonly showing associations between tau pathology and cortical thinning over time include frontotemporal and occipitoparietal regions [2, 11, 12, 40], which is similar to our findings, showing strongest effects for Braak III/IV regional tau pathology and increased cortical thinning in the frontotemporal-parietal regions. We also assessed the association between (annual) change in tau pathology and (annual) change in cortical thickness. Whereas others reported (moderate) associations between larger increases in tau pathology and larger decreases in cortical thickness in individuals with mild cognitive impairment or atypical AD [12, 13], we found no associations surviving correction for multiple testing between change in tau pathology and longitudinal cortical thickness. A potential reason for these differences across studies might be the variance in parameters used. In our study we used fully quantitative tau PET BPND to precisely measure (change in) tau pathology, whereas other studies used the semi-quantitative standardized uptake value ratio (SUVr), which is more sensitive to blood flow-introduced bias, specifically in longitudinal settings [41]. However, recently it has been demonstrated that SUVr provides an accurate estimate of specific binding for [18F]flortaucipir over a two-year follow-up during which changes in flow are small [42], making this unlikely to be a sole explanation for differences in results between studies. A difference in statistical power might also have played a role, since we did find some associations between larger increases in tau pathology with decreases in cortical thinning in Aβ+ individuals when using a more liberal statistical thresh-
RkJQdWJsaXNoZXIy MjY0ODMw