Publikationsrepositorium - Helmholtz-Zentrum Dresden-Rossendorf

Purpose: To create an inventory of automated image processing pipelines of Arterial Spin Labeling (ASL) and summarize their features and accessibility for users to choose an optimal pipeline to fit their needs.

Methods: Pipeline developers were invited to self-assess their pipelines using a questionnaire developed by the Task Force 1.1 of the Open Science Initiative for Perfusion Imaging (OSIPI). Additionally, publicly available pipelines were evaluated by two independent testers using an objective unified scoring system. The testing focused on the capability, flexibility, and ease of use of the pipelines on various datasets.

Results: The developers of twenty-one pipelines filled in the questionnaire. Most pipelines support data from the three major vendors, i.e., GE (n=15), Philips (n=15), and Siemens (n=16), are free for research (n=18), work with the standard neuroimaging data format NIfTI (n=15), and can process standard 3D single PLD pseudo-continuous ASL images (n=21). Pipelines mainly differed in their support of advanced sequences and advanced features. Nine publicly available pipelines were included in the independent testing. Whereas certain pipelines were easy to use for users without programming skills, other pipelines offered more flexibility for configuring advanced processing options.

Conclusion: ASL data from the most commonly used ASL sequences saved in the standard neuroimaging data formats can be easily processed using publicly available pipelines. A specific choice of a pipeline should consider specific requirements on features and users’ skills, and the ASL inventory can serve as a valuable guide to facilitate this choice.