Quantification of J-resolved proton spectra in two-dimensions with LCModel using GAMMA-simulated basis sets at 4 Tesla

Abstract

A two-dimensional, J-resolved magnetic resonance spectroscopic extraction approach was developed employing GAMMA-simulated, LCModel basis-sets. In this approach, a two-dimensional J-resolved (2D-JPRESS) dataset was resolved into a series of one-dimensional spectra where each spectrum was modeled and fitted with its theoretically customized LCModel template. Metabolite levels were derived from the total integral across the J-series of spectra for each metabolite. Phantoms containing physiologic concentrations of the major brain chemicals were used for validation. Varying concentrations of glutamate and glutamine were evaluated at and around their accepted in vivo concentrations in order to compare the accuracy and precision of our method with 30 ms PRESS. We also assessed 2D-JPRESS and 30 ms PRESS in vivo, in a single voxel within the parieto-occipital cortex by scanning ten healthy volunteers once and a single healthy volunteer over nine repeated measures. Phantom studies demonstrated that serial fitting of 2D-JPRESS spectra with simulated LCModel basis sets provided accurate concentration estimates for common metabolites including glutamate and glutamine. Our in vivo results using 2D-JPRESS suggested superior reproducibility in measuring glutamine and glutamate relative to 30 ms PRESS. These novel methods have clear implications for clinical and research studies seeking to understand neurochemical dysfunction.