The OpenNeuroPET Archive – A Molecular Neuroimaging Archive

Gitte Moos Knudsen, Rigshospitalet
Grant amount: DKK 10,144,473

Advanced imaging technologies, such as positron emission tomography (PET) and single photon emission tomography (SPECT), can inform critically about brain mechanisms in health and disease, and the volume of studies and data is increasing exponentially. As with other fields of science, there is a growing concern about the validity and reproducibility of outcomes from PET studies. Some of those inconsistencies are likely due to insufficient statistical power in the individual studies, to patient population heterogeneity, and to differences in software and data analysis methods. Since PET studies are extremely costly and involve exposure to radiation, better use of the resulting data is warranted from a both a financial and ethical point of view. To fully realize the potential of these new technologies, there is a need to share and aggregate data from many different studies to robustly link patterns observed in the images to diagnosis, mechanistic understanding, and clinical outcomes.

Our project will establish OpenNeuroPET as an open, international shared data repository and processing platform for PET imaging data. We are working in close collaboration with European and US based scientists to generate such a common platform that is found useful for all users. This will provide the international research community with standardized, expert-annotated data sets to facilitate meta-analysis and advanced computational modeling. Researchers will be able to analyze and share their imaging data along with relevant patient meta-data, in full compliance with ethical guidelines and legislation protecting patient rights and privacy (e.g., GDPR). The new infrastructure will adhere to the FAIR data principles and help to strengthen transparency, reproducibility, and collaboration in the field of medical imaging.

Project participants
Gitte Moos-Knudsen, Professor
Rigshospitalet, Neurobiology Research Unit

Melanie Ganz-Benjaminsen, Assistant Professor
University of Copenhagen, Department of Computer Science