Kinetic Modeling of the Tau PET Tracer 18F-AV-1451 in Human Healthy Volunteers and Alzheimer Disease Subjects.

18F-AV-1451 is currently the most widely used of several experimental tau PET tracers. The objective of this study was to evaluate 18F-AV-1451 binding with full kinetic analysis using a metabolite-corrected arterial input function and to compare parameters derived from kinetic analysis with SUV ratio (SUVR) calculated over different imaging time intervals. Methods:18F-AV-1451 PET brain imaging was completed in 16 subjects: 4 young healthy volunteers (YHV), 4 aged healthy volunteers (AHV), and 8 Alzheimer disease (AD) subjects. Subjects were imaged for 3.5 h, with arterial blood samples obtained throughout. PET data were analyzed using plasma and reference tissue-based methods to estimate the distribution volume, binding potential (BPND), and SUVR. BPND and SUVR were calculated using the cerebellar cortex as a reference region and were compared across the different methods and across the 3 groups (YHV, AHV, and AD). Results: AD demonstrated increased 18F-AV-1451 retention compared with YHV and AHV based on both invasive and noninvasive analyses in cortical regions in which paired helical filament tau accumulation is expected in AD. A correlation of R2 > 0.93 was found between BPND (130 min) and SUVR-1 at all time intervals. Cortical SUVR curves reached a relative plateau around 1.0-1.2 for YHV and AHV by approximately 50 min, but increased in AD by up to approximately 20% at 110-130 min and approximately 30% at 160-180 min relative to 80-100 min. Distribution volume (130 min) was lower by 30%-35% in the YHV than AHV. Conclusion: Our data suggest that although 18F-AV-1451 SUVR curves do not reach a plateau and are still increasing in AD, an SUVR calculated over an imaging window of 80-100 min (as currently used in clinical studies) provides estimates of paired helical filament tau burden in good correlation with BPND, whereas SUVR sensitivity to regional cerebral blood changes needs further investigation.