31P-MRS of healthy human brain: revealing the hidden PME signals under phosphoethanolamine and phosphocholine resonances at 7T

PurposeFor decades, it has been common practice to quantify brain phosphomonoester (PME) 31P signals by a two-component model composed of phosphoethanoamine (PE) and phosphcholine (PC). This study presented the evidence of hidden PME (h-PME) signals and their selective detections toward accurate quantification of PE and PC. MethodsGaussian lineshape analysis was modeled to reveal h-PME. Inversion-recovery (IR) sequence was employed to null the PE and PC resonances for selective detection of h-PME. The fully-relaxed 31P spectra after h-PME correction were used to quantify PE, PC and other brain metabolites in a group of 16 healthy subjects. ResultsSpectral lineshape analysis and IR modulation revealed previously overlooked h-PME signals underlying sharp PE and PC resonances. The h-PME signals appeared as a broad "bump" (LW1/2: 105 {+/-} 25 Hz, N = 16), leading to poor spectral resolution between PE and PC. Fast relaxing h-PME signals, tentatively assigned to blood 2,3-DPG, were selectively detectable using IR sequence at an optimal inversion delay of 5.8 s. In fully relaxed 31P spectra, h-PME measured 44 {+/-} 9 % of the total PME signal, equivalent to 1.36 {+/-} 0.39 mM in single phosphoryl unit, compared to 1.41 {+/-} 0.23 mM for PE and 0.31{+/-} 0.10 mM for PC. ConclusionHidden PME signals are a significant constitute of PME signals in human brain 31P spectra, and need to be taken into account for quantifying PE and PC as biomarkers of altered phospholipid metabolism in brain pathologies.