The Impact of Sampling Scheme on Brain Sodium Magnetic Resonance Imaging (23Na-MRI) Concentration Mapping

Abstract

There is a wide range of literature reported brain tissue sodium concentration (TSC) values obtained from quantitative sodium magnetic resonance imaging (23Na-MRI). These reported values all fall within the accepted values of brain TSC. However, the variation within a given study tends to be quite narrow indicating either the study sample is not physiologically representative of the total population, or there are experimentally contributing factors influencing this range. Experimentally, there are many different k-space data sampling schemes used in 23Na-MRI. This work provides a comprehensive comparison of signal-to-noise ratio (SNR), contrast, and quantification accuracy between six of the most common sampling schemes used in 23 Na-MRI: Cartesian, density-adapted radial (DA-3DPR), constant-amplitude radial (CA-3DPR), Fermat looped, orthogonally encoded trajectories (FLORET), rotated spiral, and 3D cones. The sampling schemes were designed and optimized for brain tissue using clinically reasonable design constraints. When considering the scan duration and sampling point-spread function (PSF), the DA-3DPR sampling scheme provided the highest SNR without any substantial loss in contrast. Quantification accuracy was found to be dependent on the sampling scheme, with 3D cones providing the highest accuracy and DA-3DPR providing the lowest. Often sodium quantification studies rely on external references of known sodium concentration. The signal distributions within the external references used for sodium quantification were also sampling scheme dependent. In this work compelling evidence is provided that each sampling scheme has differences in sensitivity to field inhomogeneities. Overall, this work shows that the variation in the reported TSC values originates from the design and implementation of the k-space sampling scheme. The work promotes a deeper investigation of methodological factors and encourages more collaboration and transparency between research groups.