Motion compensation for non-periodic dynamic tracer distributions in multi-patch magnetic particle imaging

Abstract

Abstract Objective. While the spatial and temporal resolution of magnetic particle imaging is very high, the size of the field of view is limited due to physiological constraints. Multi-patch scans allow for covering larger areas by sequentially scanning smaller subvolumes, so-called patches. The visualization of tracer dynamics with a high temporal resolution are of particular interest in many applications, e.g. cardiovascular interventions or blood flow measurements. The reconstruction of non-periodic dynamic tracer distributions is currently realized by the reconstruction of a time-series of frames under the assumption of nearly static behavior during the scan of each frame. While this approach is feasible for limited velocities, it results in data gaps in multi-patch scans leading thus to artifacts for strong dynamics. In this article, we are aiming for the reconstruction of dynamic tracer concentrations with high velocities and the compensation of motion and multi-patch artifacts. Approach. We present a reconstruction method for dynamic tracer distributions using a dynamic forward model and representing the concentration within each voxel by a spline curve. The method is evaluated with simulated single- and multi-patch data. Main results. The dynamic model enables for the reconstruction of fast tracer dynamics from few frames and the spline approach approximates the missing data which reduces multi-patch artifacts. Significance. The presented method allows to compensate motion and multi-patch artifacts and to reconstruct fast dynamic tracer distributions with arbitrary motion patterns.