PET quantification of binding of a novel tau radioligand, 18F-AM-PBB3, in consideration of time-dependent changes in its plasma free fraction.

Objectives A novel tau PET ligand, 18F-AM-PBB3, has been developed by structural modifications of 11C-PBB3. Time-activity curves (TACs) in the brain obtained by human PET indicated a good radioligand uptake after injection consistent with reversible binding kinetics. Plasma parent TAC showed an initial rapid rise followed by a decline, which then plateaued with no further decreases. We hypothesized that the plasma parent free fraction may not be constant but rather decrease with time, and therefore analyzed kinetics of 18F-AM-PBB3 with plasma input functions corrected for sequentially measured parent free fractions. \nMethods One Alzheimer’s disease (AD) and one healthy control (HC) subject underwent a dynamic PET scan, arterial blood sampling, and free fraction measurements after injection of 18F-AM-PBB3. We estimated two plasma input functions with and without free fraction corrections. One- and two-tissue compartment analyses (1TCM, 2TCM) were performed for both input functions to estimate binding potential (BPND) using cerebellar gray matter as reference. A reference tissue model (MRTMO) was also applied to estimate BPND without using blood data. We compared curve fitting and BPND values between these methods to evaluate the validity of the input correction for the plasma free fraction in compartment modeling.\nResults Plasma parent free fractions decreased over time (AD: 2.3%, HC:3.0% at 3 min and AD: 0.37, HC: 0.41% at 180 min). 2TCM produced a better fit to tissue TACs than did 1TCM with corrected input functions. Neither 1TCM nor 2TCM with uncorrected inputs described tissue TACs well. BPND values estimated by 2TCM with corrected inputs and MRTMO were in an excellent agreement. \nConclusions Plasma inputs corrected for free fractions, which decreased over time, allowed quantification of 18F-AM-PBB3 binding using 2TCM, and a good correlation was found between estimates by this method and more practical reference tissue model requiring no plasma data.Brain and Brain PET 201