Although multidimensional diffusion/relaxation NMR experiments are widely used in materials sciences and engineering applications, preclinical and clinical MRI applications of these techniques were not feasible. Moreover, higher-field MRI scanners posed another obstacle to translation of this NMR method. Their specific absorption rate (SAR) limits the use of multi-echo or CPMG pulse trains, so that the large amounts of data required by these methods cannot be collected in vivo due to exceedingly long scan times. Therefore, the primary challenges this invention overcomes are the migration of NMR methods to MRI and to vastly shorten scan times.
The “Marginal Distributions Constrained Optimization (MADCO)” methodology provides a novel framework to accelerate and improve the reconstruction of multidimensional NMR relaxation/diffusion spectra suitable for use in MRI applications on a voxel-by-voxel basis. This approach uses 1D acquired spectra as a priori information to estimate a 2D (or higher dimensional) spectrum. 1D marginal distributions are used as constraints when the 2D spectra are reconstructed.
The MADCO methodology was demonstrated with polyvinylpyrrolidone-water solution phantoms, which contains 3 species of diffusivities and T1 relaxation times (i.e., 3 peaks in the D-T1 space). With a reasonably accurate estimate of the 1D marginal distributions, MADCO accelerated the acquisition of the 2D spectra by more than a factor of 200 (i.e., two orders of magnitude) with a high level of accuracy, compared to conventional, unconstrained methods.