Monitoring and correcting spatio-temporal variations of the MR scanner’s static magnetic field

• DOI: 10.1007/s10334-006-0050-2
• Publication Date: 2006-10-16T23:16:08Z
• Authors: AbdEl Monem El-Sh...; Michael Schär; Paul A. Bottomley; Ergin Atalar
• Keywords: Magnetic Resonance Spectroscopy; Time Factors; thermometry; proton nuclear magnetic resonance; Phantoms; Imaging; 03 medical and health sciences; Computer-Assisted; Electromagnetic Fields; 0302 clinical medicine; Models; 539; Image Interpretation, Computer-Assisted; Humans; MR thermometry; device; Image Interpretation; nuclear magnetic resonance spectroscopy; Brain Mapping; Models, Statistical; accuracy; Magnet stability; Phantoms, Imaging; article; Temperature; Brain; phantom; Equipment Design; Statistical; Field mapping; Magnetic Resonance Imaging; signal noise ratio; Radiography; Magnetic field; priority journal; Spectrophotometry; Field homogeneity; Protons; mathematical model

The homogeneity and stability of the static magnetic field are of paramount importance to the accuracy of MR procedures that are sensitive to phase errors and magnetic field inhomogeneity. It is shown that intense gradient utilization in clinical horizontal-bore superconducting MR scanners of three different vendors results in main magnetic fields that vary on a long time scale both spatially and temporally by amounts of order 0.8-2.5 ppm. The observed spatial changes have linear and quadratic variations that are strongest along the z direction. It is shown that the effect of such variations is of sufficient magnitude to completely obfuscate thermal phase shifts measured by proton-resonance frequency-shift MR thermometry and certainly affect accuracy. In addition, field variations cause signal loss and line-broadening in MR spectroscopy, as exemplified by a fourfold line-broadening of metabolites over the course of a 45 min human brain study. The field variations are consistent with resistive heating of the magnet structures. It is concluded that correction strategies are required to compensate for these spatial and temporal field drifts for phase-sensitive MR protocols. It is demonstrated that serial field mapping and phased difference imaging correction protocols can substantially compensate for the drift effects observed in the MR thermometry and spectroscopy experiments.