A new method to quantify tau pathologies with (11)C-PBB3 PET using reference tissue voxels extracted from brain cortical gray matter.

BACKGROUND Quantitative in vivo imaging of tau pathologies potentially improves diagnostic accuracy of neurodegenerative tauopathies and would facilitate evaluation of disease-modifying drugs targeting tau lesions in these diseases. Tau pathology can be quantified by reference tissue models without arterial blood sampling when reference tissue devoid of target binding sites is available. The cerebellar cortex has been used as a reference region in analyses of tau positron emission tomography (PET) data in Alzheimer's disease (AD). However, in a significant subset of tauopathies such as progressive supranuclear palsy and corticobasal degeneration, tau accumulation may occur in diverse brain regions including the cerebellar cortex. This hampers selection of a distinctive reference region lacking binding sites for a tau PET ligand. The purpose of this study was to develop a new method to quantify specific binding of a PET radioligand, (11)C-PBB3, to tau deposits using reference voxels extracted from cortical gray matter, which have a low likelihood of containing tau accumulations. METHODS We reanalyzed (11)C-PBB3 PET data of seven mild AD patients (ADs) and seven elderly healthy control subjects (HCs) acquired in a previous study. As a standard method, parametric images of binding potential ([Formula: see text]) were initially generated using reference tissue manually defined on the cerebellar cortex. We then constructed a frequency histogram of [Formula: see text] values in these parametric images and selected cortical gray matter voxels contained in a certain range of the histogram with a low likelihood of having (11)C-PBB3 binding sites. Finally, these reference voxels were used for generating new [Formula: see text] parametric images. RESULTS Reference tissue voxels defined by the histogram analysis spread throughout the cortical gray matter of AD and HC brains. The [Formula: see text] values determined by our new method correlated very well with those estimated by the standard method (r (2) = 0.94), although the binding estimates by the current method were slightly higher by ~0.14 than those by the standard method. CONCLUSIONS We developed and validated a new method enabling quantification of tau lesions that can accumulate in the cerebellum and other extensive brain areas. This method may be applicable to all tauopathy subtypes and various tau PET ligands besides (11)C-PBB3. TRIAL REGISTRATION The number is UMIN000009052.