Longitudinal Tau PET Using 18 F-Flortaucipir: The Effect of Relative Cerebral Blood Flow on Quantitative and Semiquantitative Parameters

Semiquantitative PET measures such as SUV ratio (SUVr) have several advantages over quantitative measures, such as practical applicability and relative computational simplicity. However, SUVr may potentially be affected by changes in blood flow, whereas quantitative measures such as nondisplaceable binding potential (BP ) are not. For F-flortaucipir PET, the sensitivity of SUVr for changes in blood flow is currently unknown. Therefore, we compared semiquantitative (SUVr) and quantitative (BP ) parameters of longitudinal F-flortaucipir PET scans and assessed their vulnerability to changes in blood flow. Subjects with subjective cognitive decline ( = 38) and Alzheimer disease patients ( = 24) underwent baseline and 2-y follow-up dynamic F-flortaucipir PET scans. BP and relative tracer delivery were estimated using receptor parametric mapping, and SUVr at 80-100 min was calculated. Regional SUVrs were compared with corresponding distribution volume ratio (BP + 1) using paired tests. Additionally, simulations were performed to model effects of larger flow changes in different binding categories. Results in subjective cognitive decline and Alzheimer disease showed only minor differences between SUVr and BP changes over time. Relative tracer delivery changes were small in all groups. Simulations illustrated a variable bias for SUVr depending on the amount of binding. SUVr provided an accurate estimate of changes in specific binding for F-flortaucipir over a 2-y follow-up during which changes in flow were small. Notwithstanding, simulations showed that large(r) flow changes may affect F-flortaucipir SUVr. Given that it is currently unknown to what order of magnitude pharmacotherapeutic interventions may induce changes in cerebral blood flow, caution may be warranted when changes in flow are potentially large(r), as in clinical trials.