AbstractBackgroundImaging tau burden in vivo using Positron Emission Tomography (PET) has tremendous potential for advancing our biological understanding of Alzheimer’s disease. Towards this direction, there is a need for post‐acquisition imaging data ‘harmonization’ tools and analysis methods that reduce variance across studies, sites, and scanners, while retaining the power to study complex disease related effects. The Global Alzheimer’s Association Interactive Network (GAAIN), the Laboratory of Neuro Imaging (LONI), and the Critical Path for Alzheimer’s Disease (CPAD) Consortium have undertaken a collaborative effort to provide a harmonized approach to the analysis of tau PET images.MethodThe GAAIN/LONI Pipeline is a software suite that allows the design and execution of neuroimaging workflows through flexible integration of different image preprocessing software and parallel processing via a computer grid. Each workflow has a distinct fingerprint that can be executed locally on different platforms or remotely via GAAIN, thus increasing interoperability and allowing the use of “harmonized” workflows across different datasets. Tau PET workflows were designed and deployed on the GAAIN/LONI Pipeline based on existing literature. In turn, these tau PET workflows were executed on CPAD tau PET to derive tau PET burden. Clinical validation was performed by quantifying tau PET burden on different clinical groups. Analytical validation was performed by correlating the LONI Pipeline results with independently developed pipelines.ResultWe first show feasibility results as the tau PET workflows are deployed and executed on CPAD data. For clinical validation, we tested whether derived metrics can distinguish between clinical groups. For analytical validation we tested correlations of results with existing methods that have been independently developed. These results provide a unified analysis model that can be utilized from GAAIN/LONI and CPAD to conduct harmonized analyses on multiple datasets. The full capabilities of this model will be explored in subsequent work.ConclusionThe proposed pipeline, which is going to be part of the publicly available GAAIN/LONI platform, should provide researchers with an optimized roadmap that describes how tau PET imaging data should be processed and analyzed for reliability and reproducibility in the context of drug development and regulatory decision making.