Dynamic [18F]PI-2620 recording facilitates prediction of β-amyloid positivity and simultaneous staging of tau and neurodegeneration

Background

Patients with Alzheimer’s disease (AD) and clinically overlapping neurodegenerative diseases are classified molecularly using the A/T/N classification system. Apart from fluid biomarkers and structural MRI, A/T/N assessment incorporates β-amyloid-PET (A), tau-PET (T), and [18F]FDG-PET (N). We evaluated if dynamic features of tau-PET with [18F]PI-2620 allow prediction of β-amyloid positivity and simultaneous staging of tau and neurodegeneration in individual patients using a single imaging session.

Methods

We studied 129 patients with tauopathies, comprising 47 patients with β-amyloid-positive 3/4-repeat tauopathy and 82 patients with β-amyloid-negative primary 4-repeat tauopathies, alongside 17 healthy controls. Participants underwent 60-minute dynamic [18F]PI-2620 tau-PET. Kinetic modelling simultaneously provided tracer efflux rate (K2a) as a predictor for β-amyloid status, distribution volume ratio (DVR) for regional tau burden, and relative perfusion (R1) as a marker of neurodegeneration. These parameters were validated against β-amyloid-PET, [18F]FDG-PET, volumetric MRI, and cerebrospinal fluid biomarkers using receiver operating characteristic and correlation analyses and visual assessments. [18F]PI-2620K2a-based prediction of β-amyloid positivtiy was additionally tested in an independent, clinically heterogeneous validation cohort of 97 individuals.

Results

[18F]PI-2620K2a differentiated tau isoform compositions despite clinically overlapping presentations and tau-PET patterns, demonstrating reduced cortical tracer clearance in 3/4-repeat compared with 4-repeat tauopathies. [18F]PI-2620K2a remained sensitive in individuals with visually negative or low [18F]PI-2620DVR signals and outperformed both tau burden and perfusion for predicting β-amyloid status, achieving an area under the curve of 0.99, a positive predictive value of 91.5%, and a negative predictive value of 95.1%. Performance remained robust in the independent validation cohort, with an area under the curve of 0.98, a positive predictive value of 87.9%, and a negative predictive value of 95.3%. [18F]PI-2620DVR quantified regional tau patterns, with cortical predominance in 3/4-repeat tauopathy and subcortical involvement in 4-repeat tauopathies. [18F]PI-2620R1 exhibited strong quantitative and visual associations with established markers of neurodegeneration. Integration of K2a, DVR, and R1 in disease-specific composite regions enabled individualized three-dimensional A/T/N staging.

Conclusions

Dynamic [18F]PI-2620 PET imaging facilitates assessment of the β-amyloid status and regional staging of tau and neurodegeneration during a single acquisition. This one-stop-shop approach may reduce radiation exposure, streamline diagnostic workflows, and facilitate personalized disease profiling.