A sensitive proton magnetic resonance spectroscopy ((1)H MRS) acquisition scheme that is capable of discriminating between normal controls and a group of patients with early Alzheimer's disease (AD) is presented. The performance of this newly developed method, TE averaged PRESS (PRESS-J), in detecting neurodegeneration associated with early AD is compared with that of short echo time (TE) PRESS. A stepwise discriminant function analysis is used to construct discriminant functions for both pulse sequences. These functions are each composed of a single predictor: the N-acetyl aspartate (NAA)/creatine (Cr) ratio for PRESS-J, and the NAA/myoInositol (mI) ratio for PRESS. We observed lower P-values, higher areas under the receiver operating characteristic curves, and higher sensitivity at a given specificity for the PRESS-J discriminating function in comparison with the PRESS discriminating function. The higher sensitivity of PRESS-J is due to decreased variability when the singlets are fit in the spectra. This increased sensitivity enables new MR applications and, among other benefits, allows for smaller group sizes in drug trials, which can significantly reduce the cost of such trials.
Copyright 2005 Wiley-Liss, Inc.