NORMALIZATION OF METABOLITE IMAGES IN H-1-NMR SPECTROSCOPIC IMAGING

Citation
Jm. Wild et I. Marshall, NORMALIZATION OF METABOLITE IMAGES IN H-1-NMR SPECTROSCOPIC IMAGING, Magnetic resonance imaging, 15(9), 1997, pp. 1057-1066
Citations number
19
Categorie Soggetti
Radiology,Nuclear Medicine & Medical Imaging
Journal title
ISSN journal
0730725X
Volume
15
Issue
9
Year of publication
1997
Pages
1057 - 1066
Database
ISI
SICI code
0730-725X(1997)15:9<1057:NOMIIH>2.0.ZU;2-K
Abstract
Single slice proton magnetic resonance spectroscopic images were taken from a volume of interest (VOI), which was localised using the point resolved spectroscopy (PRESS) sequence, Non-uniformities in the excita tion profile of the VOI within the image plane were accounted for usin g a uniformity correction image, Metabolite images were divided on a v oxel-by-voxel basis with a water image obtained from a VOI of the same spatial dimensions from a phantom acquisition, Normalisation procedur es carried out on phantom metabolite images were needed to account for the spatial shift in the excitation profile of the VOI, which varies linearly with the chemical shift of the metabolite, Patient metabolite images were then normalised using a VOI water image from a phantom ac quisition, which was spatially shifted according to the metabolite che mical shift, In a study of 14 stroke patients, metabolite maps of the distribution of choline, creatine and N-acetyl aspartate (NAA) were co nstructed, The spectral signal intensities from these metabolites in t he infarcted and in the contralateral region were calculated using the raw maps and secondly with the same maps after the B-1 normalisation process, In the initial analysis only the metabolite NAA showed a sign ificant decrease, whilst all three metabolites showed a significant re duction in the normalised analysis, These results confirm that normali sation of PRESS localised images does have a significant bearing upon any quantitative measurements made using such images, (C) 1997 Elsevie r Science Inc.