Chapter 3 - Susceptibility-weighted imaging

Introduction As has been discussed in Ch. 2, imaging cerebral microbleeds (CMB) is possible in MRI thanks to gradient-recalled echo (GRE) imaging methods. These sequences are particularly sensitive to blood products such as hemosiderin (non-heme iron) and to changes in the level of deoxyhemoglobin (heme iron). In the early days of clinical MRI, GRE scanning was usually done in two dimensions (2D) with echo times (TE) limited to a range of 15 to 25 ms at 1.5 T. These limitations were imposed because longer TE values combined with the thick slice 2D scans led to very poor image quality. With the advent of three-dimensional (3D) GRE imaging, it became possible to scan with much smaller voxel sizes and longer TE periods (40, 60 and 80 ms at 1.5 T) while still maintaining good image quality [1], thanks to the decreased dephasing across the now smaller voxels. This early work also recognized the importance of filtered phase images and was the foundation from which the concept of susceptibility-weighted imaging (SWI) was to emerge [2]. Phase data are usually discarded in conventional GRE imaging. This is because phase is sensitive not only to local field inhomogeneities, which are of great interest, but also to all forms of unwanted magnetic field inhomogeneity, including eddy current effects, main magnet effects and air-tissue interface effects. Most of the unwanted phase effects are of low spatial frequency in nature. Therefore, with the appropriate high-pass filtering [2-4], the phase data potentially become quite useful.