Longitudinal tau PET using flortaucipir: Comparison of (semi)quantitative parameters

PurposeSemi-quantitative PET measures can be affected by changes in blood flow, whereas quantitative measures are not. The aim of the study was to compare semi-quantitative(SUVr) and quantitative(R1, BPND) parameters of longitudinal tau PET scans with [18F]flortaucipir, with respect to changes in blood flow. MethodsSubjects with subjective cognitive decline(SCD; n=38) and Alzheimers disease(AD) patients(n=24) underwent baseline(BL) and 2-year follow-up(FU) dynamic [18F]flortaucipir PET scans. BPND and R1 were estimated using RPM and SUVr(80-100min) was calculated (cerebellar gray as reference). For each region-of-interest ((trans)entorhinal, limbic and neocortical) and parameter, %change was calculated. Regional SUVrs were compared to corresponding DVR (= BPND+1) using paired T-tests. Additionally, simulations were performed to model effects of flow changes on BPND and SUVr in different binding categories. Thereafter, %bias for SUVr with respect to underlying binding and flow were evaluated. ResultsIn SCD, there was a difference between %change in the (trans)entorhinal ROI (DVR 2.56% vs SUVr 1.85%) only. In AD, a difference was found in the limbic ROI (DVR 6.61% vs SUVr 7.52%) only. R1 changes were small(+0.7% in SCD and -1.6% in AD). Simulations illustrated with increasing flow a decreased %bias for SUVr in low binding conditions, whereas a slightly increased bias was observed in high binding conditions. ConclusionSUVr provided an accurate estimate of specific binding for [18F]flortaucipir over a two-year follow-up. However, simulations showed that flow changes can affect [18F]flortaucipir SUVr, hence DVR/BPND should be preferred in more advanced disease stages and/or conditions that could induce significant flow changes like pharmacotherapeutic interventions.