Imaging Aβ and tau in early stage Alzheimer’s disease with [18F]AV45 and [18F]AV1451

Azadeh Firouzian, Alex Whittington, Graham E. Searle, Ivan Koychev, Giovanna Zamboni, Simon Lovestone, Roger N. Gunn, on behalf of the Deep and Frequent Phenotyping study team

AD is a progressive neurodegenerative disorder that is associated with the accumulation of two different insoluble protein aggregates, Aβ plaques and hyperphosphorylated tau. This study aimed to investigate the optimal acquisition and quantification of [18F]AV45 and [18F]AV1451 to image Aβ and tau, respectively, in subjects with AD. Fifteen subjects with early stage AD underwent a T1-weighted structural MRI and two dynamic PET scans to image Aβ (60 min, [18F]AV45) and tau (120 min, [18F]AV1451). Both dynamic BPNDand static SUVR outcome measures were calculated and compared for 12 out of 15 subjects who completed 60 min of the Aβ PET scan and at least 110 min of the tau PET scan. The SRTM and reference Logan graphical analysis were applied to the dynamic data to estimate regional BPNDvalues and SUVR ratios from the static data. Optimal acquisition windows were explored for both the dynamic and static acquisitions. In addition, the spatial correlation between regional Aβ and tau signals was explored. Both the SRTM and graphical analysis methods showed a good fit to the dynamic data for both Aβ and tau dynamic PET scans. Mean regional BPNDestimates became stable 30 min p.i. for [18F]AV45 and 80 min p.i. for [18F]AV1451. Time stability analysis of static SUVR data showed that the outcome measure starts to become stable for scan windows of 30-50 min p.i. for [18F]AV45 and 80-100 min p.i. for [18F]AV1451. The results from these time windows correlated well with the results from the full dynamic analysis for both tracers (R2 = 0.74 for [18F]AV45 and R2 = 0.88 for [18F]AV1451). There was a high correlation between amyloid uptake estimate using both dynamic analysis methods in thalamus and tau uptake in thalamus, hippocampus and amygdala. Short static PET scans at appropriate time windows provided SUVR values which were in reasonable agreement with BPNDvalues calculated from dynamic scans using SRTM and reference Logan. These simplified methods may be appropriate for classification and intervention studies, although caution should be employed when considering interventional studies where blood flow and extraction could change.