AURA : building an ‘A user repository of artifacts for perfusion imaging’ for quality control

Purpose - Perfusion imaging MRI techniques, including Arterial Spin Labeling (ASL), Dynamic Susceptibility Contrast (DSC), and Dynamic Contrast-Enhanced (DCE) imaging hold significant potential for research and clinical applications.1 During the ExploreASL2 online user meetings, it became clear that the prevalence of imaging artifacts and their type and typical manifestation became a major hurdle for researchers without extensive ASL expertise. Especially so because there is only limited availability of teaching materials, which do not cover the full spectrum of possible artifacts, and a comprehensive, publicly available database of all artifacts that occur in perfusion images is completely lacking. Such an overview is essential to guide both researchers and clinicians in recognizing and addressing artifacts effectively and standardizing the practice of artifact reporting in clinical studies. The AURA project aims to create a centralized user repository of artifacts for perfusion imaging to support researchers and clinicians in the evaluation of perfusion image quality. The database will provide an overview of general MRI artifacts within perfusion images, as well as method-specific artifacts across different vendors and imaging approaches. Additionally, a short explanation of each artifact, along with possible sources and solutions, will be provided. Methods - This project is part of the ISMRM Open Science Initiative for Perfusion Imaging (OSIPI).3 An overview of the AURA-project is provided in Figure 1. An artifact submission form, created using REDCap,4-5 will be shared in several MRI perfusion imaging communities, including OSIPI, the ISMRM Perfusion Study Group, and ESMRMB. Using this form, researchers will be able tocan upload images of artifacts and provide data acquisition information, as shown in Figure 1. Additionally, literature will be reviewed to collect information on the type, prevalence. and the source of perfusion imaging artifacts. Perfusion imaging sequence and analysis software developers will be contacted for additional input. Submitted artifacts will be screened and categorized by a team of experts in perfusion imaging acquisition, analysis, and QC. Based on the screening, submitters will be contacted to provide a BIDS- compliant version of the dataset. An open repository will be created and be accessible via the ISMRM OSIPI website.6 Results & discussion - The AURA -project will result in an online repository of existing artifacts in ASL, DCE, and DSC. For each artifact, examples from different vendors and sequences will be included, as well as information such as the source and solution. This database may create knowledge on the prevalence of perfusion imaging artifacts, allow training automatic QC algorithms, serve as an educational resource for radiographers and radiologists, and improve perfusion imaging guidelines. Conclusion - This community-driven repository will provide confidence to researchers, MR physicists, and clinicians on which perfusion images they can or cannot use. It may eventually improve the value of perfusion MRI for clinical practice, academic research, and clinical trials. References - 1Jahng, et al. (2014); 2Mutsaerts, et.al (2020); 3Bell, et al. (2023); 4-5Harris, et al. (2009, 2019); 5https://osipi.ismrm.org/