Regional profiles of the candidate tau PET ligand 18F-AV-1451 recapitulate key features of Braak histopathological stages

PET imaging of neurofibrillary tangles with [F-18]-AV-1451 opens up the possibility of staging tau pathology in vivo. Schwarz et al. describe a new algorithm for automated assignment of estimated Braak stages from imaging data. Estimates replicate expected patterns of tau pathology and are associated with cognition and amyloid status.See Thal and Vandenberghe (doi:10.1093/brain/aww057) for a scientific commentary on this article. PET imaging of neurofibrillary tangles with [F-18]-AV-1451 opens up the possibility of staging tau pathology in vivo. Schwarz et al. describe a new algorithm for automated assignment of estimated Braak stages from imaging data. Estimates replicate expected patterns of tau pathology and are associated with cognition and amyloid status.Post-mortem Braak staging of neurofibrillary tau tangle topographical distribution is one of the core neuropathological criteria for the diagnosis of Alzheimer's disease. The recent development of positron emission tomography tracers targeting neurofibrillary tangles has enabled the distribution of tau pathology to be imaged in living subjects. Methods for extraction of classic Braak staging from in vivo imaging of neurofibrillary tau tangles have not yet been explored. Standardized uptake value ratio images were calculated from 80-100 minute F-18-AV-1451 (also known as T807) positron emission tomography scans obtained from n = 14 young reference subjects (age 21-39 years, Mini-Mental State Examination 29-30) and n = 173 older test subjects (age 50-95 years) comprising amyloid negative cognitively normal (n = 42), clinically-diagnosed mild cognitive impairment (amyloid positive, n = 47, and amyloid negative, n = 40) and Alzheimer's disease (amyloid positive, n = 28, and amyloid negative, n = 16). We defined seven regions of interest in anterior temporal lobe and occipital lobe sections corresponding closely to those used as decision points in Braak staging. An algorithm based on the Braak histological staging procedure was applied to estimate Braak stages directly from the region of interest profiles in each subject. Quantitative region-based analysis of F-18-AV-1451 images yielded region of interest and voxel level profiles that mirrored key features of neuropathological tau progression including profiles consistent with Braak stages 0 through VI. A simple set of decision rules enabled plausible Braak stages corresponding to stereotypical progression patterns to be objectively estimated in 149 (86%) of test subjects. An additional 12 (7%) subjects presented with predefined variant profiles (relative sparing of the hippocampus and/or occipital lobe). The estimated Braak stage was significantly associated with amyloid status, diagnostic category and measures of global cognition. In vivo F-18-AV-1451 positron emission tomography images across the Alzheimer's disease spectrum could be classified into patterns similar to those prescribed by Braak neuropathological staging of tau pathology.