Experimental characterization of the ISIS technique for volume selected NMR spectroscopy.

As clinical applications of MRS grow in number and complexity, there is a need for standardized methods for characterizing the performance of volume selection techniques. The results are presented of a thorough evaluation of a particular implementation of ISIS performed using a procedure which forms the basis of the method adopted by the European Community Concerted Action on MRS and MRI. We have found that ISIS localization is optimal when the volume of interest is slightly smaller than the region we wish to study. Contamination with extraneous signal has little T1 dependence so long as TR greater than T1 and the detection pulse angle is 90 degrees. However, a poorly optimized detection pulse results in T1-weighted contamination unless TR greater than 3T1. In the clinical context, this corresponds to a different degree of contamination for each peak in the spectrum. Adiabatic detection pulses were used in an attempt to overcome this problem without resorting to unacceptable TR values, but these were found to function less well than properly optimized rectangular pulses, even if the power was increased above the level determined by the system for B1 insensitivity. These detailed results pertain only to our system, but illustrate the importance of performing similar measurements as part of clinical spectroscopy programmes at other centres.