A general solution to the problem of localizing nuclear magneticresonance spectra for biomedical research and clinicalapplications is being extended and perfected. The method, spectrallocalization by imaging (SLIM) and its generalized extension(GSLIM), can use anatomical information, as obtained routinely bystandard MRI methods, to define any number of volume elements ofany shape. This information may be combined mathematically withspectroscopic data, yielding metabolic concentrations in allanatomical regions simultaneously, but in a much shorter time thanrequired for conventional chemical shift imaging. GSLIM alsoreveals concentration or other variations within regions. RESEARCH HIGHLIGHTS:Thorough theoretical analyses and computer simulations of themethod have been carried out. Our recent results confirm that SLIMand GSLIM are reliable and efficient methods for measuring averagemetabolic concentrations from volume elements of any shape underpractical conditions (noise and inhomogeneities of any kind insidethe volume elements). These results have been further verified bya new complex-phantom study. We have also applied the method toobtain localized 31P from models derived from human images. We areable to derive good quality localized spectra from regions ofcerebellum, cerebrum, gray and white matter. We have now usedthese methods for quantitative analysis, and carried out furtherextensive verification studies.
Showing the most recent 10 out of 55 publications
