ProFit: two‐dimensional prior‐knowledge fitting of J‐resolved spectra

Abstract

A two-dimensional fitting procedure is introduced, capable of extracting the full amount of information from 2D J-resolved magnetic resonance spectroscopic data. The fitting procedure uses a linear combination of 2D model spectra. For reducing the degrees of freedom and increasing robustness, it is divided into a non-linear outer loop and an inner linear least-squares fit for the concentrations. In vitro and in vivo experiments on a whole-body 3 T MR scanner show the detectability of a wide range of metabolites in the human brain, namely total creatine, N-acetylaspartate, N-acetylaspartylglutamate, choline-containing compounds, glutamate, myo-inositol, glutathione, scyllo-inositol, gamma-aminobutyric acid, alanine and ascorbic acid.

Magnetic resonance spectroscopy reveals abnormalities of glucose metabolism in the Alzheimer's brain

“…NAA is typically diminished in AD,17, 18 and this pattern held for this study ( F [2,75] = 7.231, P = 0.001; NAA was lower in the AD than in the younger CN ( P = 0.0003) and older CN groups ( P = 0.022). Metabolite ratios in the normal groups were consistent with those reported earlier in ProFit literature 24. PRESS Creatine levels were not significantly different between groups ( F [2, 43] = 1.106, P = 0.340), suggesting that between‐group differences in the ratios of other metabolites to creatine are not merely attributable to differences in creatine.…”

“…This post hoc analysis retained significance for these comparisons. A pertinent limitation of the J‐PRESS method is that it may overestimate the Glc signal, possibly due to macromolecule contamination 24. However, this would also be true for the two control groups, and there is at the moment no clear evidence of macromolecule differences causing MRS signal differences in AD.…”

“…Metabolite measurements were excluded from analysis if their CRLB exceeded a threshold of 20% 25. Aspartate (Asp), glycine (Gly), and phosphorylethanolamine (PE) were preemptively excluded from the analysis due to known in vivo overestimations or signal contamination 24. See Figure 1B and C for a representative J‐PRESS ProFit sample spectrum and select basis sets.…”

Proton MR spectroscopy of the brain at 3 T: an update

“…Beside glutamate, it shows the detectability of a wide range of metabolites, including myo-inositol, glutathione, scyllo-inositol, GABA, alanine and ascorbic ac [62].…”

“…A number of spectral editing techniques have been proposed in an attempt to resolve GABA. The techniques successfully implemented at 3 T include: single quantum difference spectroscopy (SQDS) [64], STEAM with optimized sequence times [65], maximum-echo sampling [48], double-quantum filtering with selective J rewinding [66] and ProFit of J-resolved spectra [62]. The selective homonuclear Hartmann-Hahn transfer method, recently proposed by Choi et al [67], provides a spectral selectivity and a sensitivity greater than those reported by previous GABA-editing methods.…”