Different materials placed in a homogeneous magnetic field (e.g. in an MR scanner), affect the magnetic field in different ways according to their magnetic susceptibility. For instance, water has weak negative susceptibility, namely it develops a small magnetization that acts to counteract the external field, while bone and air have near zero susceptibility with little effect on the magnetic field. When a structure composed of materials of different susceptibility, such as the head, is placed in the bore, the magnetic field becomes distorted and inhomogeneous . At higher magnetic fields, microscopic susceptibility from paramagnetic substances and blood products, and macroscopic susceptibility at the level of air-tissue and tissue-bone interfaces, such as near the air-filled sinuses and skull base, are sensibly increased. Consequently, magnetic field inhomogenei-ty and susceptibility artefacts will make it more difficult to obtain good-quality spectra, especially from lesions near the skull base or close to the calvaria. These problems can be alleviated mainly by the following expedients: using higher spatial resolution, optimizing RF pulse and coil designs [27,28], improving automatic local shimming methods [29, 30], and undistorting the images using magnetic field maps . Spectral distortion and loss in SNR may also be caused by eddy currents, which are more apparent at higher field strength due to increased speed and power of gradient coils. These eddy currents can create an additional magnetic field of duration much longer than the original gradient pulse that generated them , even though these effects can be reduced by dedicated hardware and software designs .
Field inhomogeneity, measured in Hz, has been found to be similar at 1.5 T and 3.0 T in phantom studies, but comparison of the linewidth and T2 values in vivo shows that the inhomogeneity contribution to the linewidth is greater at 3.0 T than at 1.5 T. For all three main metabolites (NAA, Cr and Cho), the average of the difference between the experimental linewidth and the estimated natural linewidth was 0.95 Hz at 1.5 T and 2.66 Hz at 3 T . Improved high-order shimming techniques may help to minimize this term at higher field strengths .
Another important effect found at 3.0 T is related to Bj inhomogeneities, usually referred to as dielectric resonance. As soon as the sample size approaches the dimension of the RF wavelength, the RF field becomes inhomogeneous. This can be observed as bright spots in the central area of head, body or phantom images acquired at 3.0 T and above. In MRS experiments, this effect makes it difficult or impossible to define the appropriate transmitter gain going along with the required flip angle, which will locally vary significantly, to apply the reciprocity theorem for spectrum quantification.
Was this article helpful?