On-scanner Reconstruction for Susceptibility Source Separation

Motivation: Susceptibility source separation provides valuable quantification of iron and myelin but currently requires offline analysis. Enabling an on-scanner fast and automated reconstruction would enable wider adoption. Goal(s): To implement and validate an inline reconstruction for susceptibility source separation on scanners. Approach: Susceptibility separation was packaged into a containerized environment and integrated inline using a vendor prototype for inline image processing on existing scanner hardware. The implemented pipeline produced paramagnetic and diamagnetic maps based on multi-echo R2*- and R2-sensitive acquisitions or R2* alone. Results: Separated susceptibility maps were successfully generated on a 3 T scanner and saved as DICOM images. Impact: The implemented reconstruction tool enables producing para- and diamagnetic maps directly on the manufacturer's scanners without the need for offline processing.