Iterative MR Image Reconstruction Using Sensitivity and Inhomogeneity Field Maps

Introduction Standard Fourier reconstruction of MRI data relies on the Fourier transform relation between acquired k-space data and the image as described by the signal equation. This relationship does not hold, however, in the presence of inhomogeneities of the magnetic field. The Fourier reconstruction also does not handle data from nonCartesian k-space sampling patterns, such as spirals, and information from coil sensitivity maps in parallel imaging experiments. In nonparallel imaging schemes, many methods have been proposed to compensate for field inhomogeneities when a nonrectilinear k-space scanning protocol is used [1,2,3,4,5]. Most of these rely on the assumption of a smoothly varying field map. Parallel imaging methods such as SENSE [6] have enjoyed increasing popularity over the last few years due to decreased scan times, despite suffering from the need for efficient reconstruction algorithms when non-Cartesian kspace trajectories are used. In this abstract, we use the readily available information of an inhomogeneity field map and sensitivity maps for the coils to perform iterative least squares image reconstruction in a spiral SENSE experiment.