Evaluation of tau deposition using 18F-PI-2620 PET in MCI and early AD subjects-a MissionAD tau sub-study

Abstract

Background: The ability of ¹⁸F-PI-2620 PET to measure the spatial distribution of tau pathology in Alzheimer's disease (AD) has been demonstrated in previous studies. The objective of this work was to evaluate tau deposition using ¹⁸F-PI-2620 PET in beta-amyloid positive subjects with a diagnosis of mild cognitive impairment (MCI) or mild AD dementia and characterize it with respect to amyloid deposition, cerebrospinal fluid (CSF) assessment, hippocampal volume, and cognition.

Methods: Subjects with a diagnosis of MCI due to AD or mild AD dementia and a visually amyloid-positive ¹⁸F-florbetaben PET scan (n=74, 76 ± 7 years, 38 females) underwent a baseline ¹⁸F-PI-2620 PET, T1-weighted magnetic resonance imaging (MRI), CSF assessment (Aβ42/Aβ40 ratio, p-tau, t-tau) (n=22) and several cognitive tests. A 1-year follow-up ¹⁸F-PI-2620 PET scans and cognitive assessments were done in 15 subjects.

Results: Percentage of visually tau-positive scans increased with amyloid-beta deposition measured in ¹⁸F-florbetaben Centiloids (CL) (7.7% (<36 CL), 80% (>83 CL)). ¹⁸F-PI-2620 standardized uptake value ratio (SUVR) was correlated with increased ¹⁸F-florbetaben CL in several regions of interest. Elevated ¹⁸F-PI-2620 SUVR (fusiform gyrus) was associated to high CSF p-tau and t-tau (p=0.0006 and p=0.01, respectively). Low hippocampal volume was associated with increased tau load at baseline (p=0.006 (mesial temporal); p=0.01 (fusiform gyrus)). Significant increases in tau SUVR were observed after 12 months, particularly in the mesial temporal cortex, fusiform gyrus, and inferior temporal cortex (p=0.04, p=0.047, p=0.02, respectively). However, no statistically significant increase in amyloid-beta load was measured over the observation time. The MMSE (Recall score), ADAS-Cog14 (Word recognition score), and CBB (One-card learning score) showed the strongest association with tau deposition at baseline.

Conclusions: The findings support the hypothesis that ¹⁸F-PI-2620 PET imaging of neuropathologic tau deposits may reflect underlying neurodegeneration in AD with significant correlations with hippocampal volume, CSF biomarkers, and amyloid-beta load. Furthermore, quantifiable increases in ¹⁸F-PI-2620 SUVR over a 12-month period in regions with early tau deposition are consistent with the hypothesis that cortical tau is associated with cognitive impairment. This study supports the utility of ¹⁸F-PI-2620 PET to assess tau deposits in an early AD population. Quantifiable tau load and its corresponding increase in early AD cases could be a relevant target engagement marker in clinical trials of anti-amyloid and anti-tau agents.

Trial registration: Data used in this manuscript belong to a tau PET imaging sub-study of the elenbecestat MissionAD Phase 3 program registered in ClinicalTrials.gov ( NCT02956486 ; NCT03036280 ).

Keywords: Alzheimer’s disease; Amyloid-beta; Florbetaben; Mild cognitive impairment; PI-2620; Tau.

Fig. 1

¹⁸F-PI-2620 SUVR PET images (scalp stripped) registered to the T1-weighted MRI illustrating the tracer distribution in a tau-negative subject (top row), a tau-positive subject with uptake in the mesial temporal cortex (center row), and a tau-positive subject with extensive neocortical uptake (bottom row)

Fig. 2

Histogram of the number of subjects across ¹⁸F-florbetaben CL values (tau-positive subjects are marked in gray; cutoff for established amyloid pathology (36 CL) and high amyloid-load (82 CL) are marked with red and gray dashed lines, respectively) (left panel) and percent of visually tau-positive subjects by amyloid-beta load category (right panel)

Fig. 3

Scatter plots of the ¹⁸F-PI-2620 SUVR versus amyloid-beta ¹⁸F-florbetaben CLs. Red and blue lines represent ¹⁸F-PI-2620 SUVR and ¹⁸F-florbetaben CL positivity cutoffs, respectively, derived from previous publications [14, 25]. Solid dots correspond to those subjects that underwent 1-year follow-up ¹⁸F-PI-2620 PET scans

Fig. 4

Scatter plot of CSF Aβ42/Aβ40 ratio versus ¹⁸F-florbetaben CL (left), CSF p-Tau versus ¹⁸F-PI-2620 SUVR (fusiform gyrus) (center) and CSF t-Tau versus ¹⁸F-PI-2620 SUVR (fusiform gyrus) (right). The subject marked with a cross was considered an outlier and not included in the linear regression

Fig. 5

Percent amyloid-beta and tau accumulation (percent SUVR change) over 1-year follow-up. Statistically significant changes (p-value < 0.05) obtained from the linear mixed effect model are marked with an asterisk (*)

Fig. 6

Voxel-wise spearman correlation maps assessing the association between ¹⁸F-PI-2620 deposition and cognition at baseline displayed on top of the SPM’s T1-weighted template in the MNI space

Fig. 7

P-values assessing the association of ¹⁸F-PI-2620 deposition and cognition at baseline