157 Longitudinal tau PET and cognition 4]. Annual change in tau pathology was also predictive of multi-domain cognitive change over time in CI individuals and in CU. These results are in line with a study in which they found that change in inferior temporal tau associated with cognitive changes on the preclinical Alzheimer cognitive composite (PACC; including memory, executive functioning and global cognition) [6, 41]. In addition, our findings match results from a previous study where they also found effects of tau pathology on MMSE in CI individuals [4]. Sensitivity analyses revealed that increased in early stage (Braak I) tau PET binding over time was associated with improvement on memory in CU individuals. It could be argued that tau pathology in the early stage entorhinal region is not yet sufficient to have a detrimental effect on cognition over time. The improvement in cognitive test scores in such cases likely reflects practice-effects independent of tau pathology, induced by repeated cognitive assessments. However, a recent paper including CU individuals from multiple cohorts showed that amyloid positive individuals who harbor tau pathology in early stage medial temporal regions are already destined to decline in cognition when assessed longitudinally, suggesting there is an effect of early stage tau pathology, specifically in those individuals also harboring amyloid pathology (which is the case for a subset of individuals in our CU group) [42]. Another explanation for our finding might be a compensatory mechanism acting upon early stage tau pathology, resulting in maintained or in our case even improved memory function despite presence of (early stage) pathology. A strength of this study is that we used fully quantitative dynamic 18F-flortaucpir PET scans able to obtain measures for both tau pathology and rCBF from a single scan. We did so in a cohort spanning the whole AD spectrum – from cognitively healthy individuals till AD dementia patients. Furthermore, we had longitudinal PET data, enabling us to assess associations between change in tau pathology and rCBF with cognitive change over time. Also, we had longitudinal cognitive data available, covering multiple cognitive domains relevant for AD. This study also has some limitations. First, our sample with longitudinal PET data available was relatively small and may have suffered some bias towards less severe AD cases. This is however often the case and insurmountable for longitudinal studies into neurodegenerative diseases. Second, R1 is a relative understudied parameter in current literature and may therefore be more poorly understood than measures like ASL. Third, we had no quantitative data for amyloid pathology available for all participants, not allowing us to investigate the role of amyloid pathology in the context of this study. In conclusion, baseline tau load as well as change in tau load are predictive for multi-domain longitudinal cognitive trajectories in both CI and CU individuals, especially when corrected for rCBF. This underlines that tau PET is a valuable tool in predicting domain-specific cognition, also in early stages of disease, which is the 6
RkJQdWJsaXNoZXIy MjY0ODMw