Multi-dimensional spectroscopic NMR and MRI using marginal distributions
Inventors
Basser, Peter J. • Benjamini, Dan H.
Assignees
US Department of Health and Human Services
Publication Number
US-11846690-B2
Publication Date
2023-12-19
Expiration Date
2037-08-11
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
Multi-dimensional spectra associated with a specimen are reconstructed using lower dimensional spectra as constraints. For example, a two-dimensional spectrum associated with diffusivity and spin-lattice relaxation time is obtained using one-dimensional spectra associated with diffusivity and spin-lattice relaxation time, respectively, as constraints. Data for a full two dimensional spectrum are not acquired, leading to significantly reduced data acquisition times.
Core Innovation
The invention relates to novel magnetic resonance (MR) methods and systems for reconstructing multi-dimensional spectra using lower dimensional spectra as constraints. A key innovation is accelerating acquisition and reconstruction of two-dimensional (2D) spectra, such as those associated with diffusivity and spin-lattice relaxation time (T1), by employing one-dimensional (1D) marginal distributions obtained from corresponding 1D measurements as constraints. This approach reduces data acquisition times significantly by avoiding the need to acquire the full 2D data set.
The problem being solved is the long acquisition times and significant computational resources required to process large data sets in conventional multidimensional NMR and MRI relaxometry. Traditional methods require dense sampling of 2D parameter spaces and inversion of Fredholm integrals of the first kind, an ill-conditioned problem that necessitates large amounts of data for stable solutions, making certain in vivo preclinical and clinical applications infeasible. Limitations include safety constraints on pulse application and excessively long scan times.
The disclosed marginal distributions constrained optimization (MADCO) methods use 1D projections or marginal distributions as equality constraints to stabilize and reduce the number of measurements required. By first estimating the 1D marginal distributions independently, these are then used to constrain reconstruction of the 2D (or higher dimensional) spectra, leading to enhanced accuracy and stability with far fewer acquisitions. MADCO methods can be applied to various 2D MRI experiments (e.g., D-T2, T1-T2), and can extend to higher dimensional spectra by utilizing lower order multi-dimensional spectra as constraints.
Claims Coverage
The patent claims cover methods and systems for reconstructing multi-dimensional magnetic resonance spectra using marginal distributions as constraints, comprising several inventive features related to acquisition parameters, specimen characteristics, and data processing.
Reconstruction of multi-dimensional MR spectra using marginal distributions as constraints
A method of acquiring MR specimen data by varying first and second acquisition parameters associated with specimen characteristics, determining first and second marginal distributions dependent on these parameters, and reconstructing an (n+m)-dimensional spectrum using these marginal distributions as constraints.
Flexible dimensionality in marginal distribution constraints
Method wherein the first marginal distribution can be one-dimensional and the second two-dimensional, allowing reconstruction of three-dimensional spectra from combined marginal distributions.
Inclusion of two-dimensional data sets for spectrum reconstruction
Incorporation of acquiring a two-dimensional specimen data set associated with both acquisition parameters to assist in reconstructing the multi-dimensional spectrum.
Specified specimen characteristics and acquisition parameters
Use of specimen characteristics such as diffusivity (D), spin-lattice (T1) and spin-spin relaxation times (T2), and acquisition parameters like b-value and inversion time τ, to define the MR signal variations and reconstruction process.
MR system operable for multi-dimensional spectrum reconstruction with marginal distribution constraints
An MR system comprising an acquisition system and a data processor configured to reconstruct multi-dimensional spectra using first and second marginal distributions as constraints, including hardware elements like magnets and RF generators to produce static fields and induce spin rotations.
The claims collectively define a method and system for acquiring and processing MR data to reconstruct multi-dimensional spectra constrained by lower dimensional marginal distributions, enabling reduced acquisition times and enhanced stability in spectrum estimation.
Stated Advantages
Significantly reduced data acquisition times by using marginal distributions as constraints, requiring an order of magnitude less data than conventional approaches.
Improved stability and accuracy of reconstructed multi-dimensional spectra even with substantially fewer acquisitions.
Greater robustness and reduced sensitivity to experimental parameters compared to conventional unconstrained methods.
Compatibility with various 2D MRI experiments including D-T1, D-T2, T1-T2, and potential extension to higher dimensions.
Potential application in biological, preclinical, and clinical settings due to feasible scan times and clinically relevant pulse sequences.
Capability to probe molecular exchange dynamics in a model-free manner, improving the measurement of cell membrane permeability and other dynamic parameters.
Documented Applications
Use in biological, preclinical, and clinical MRI to obtain multi-dimensional relaxation and diffusion spectra, including applications in brain tissue characterization, nerve microstructure analysis, and microdynamic MRI scans.
Study of neurological conditions, inflammation, cancer, stroke, neuroplasticity, and normal and injured ex vivo tissue samples with high resolution and quality.
Whole-body imaging applications including scanning of heart, placenta, liver, kidneys, spleen, colon, prostate, muscles, and peripheral nerves.
Use in genotype/phenotype studies and analysis of non-biological materials such as polymers, gels, food products, chemical engineering separation systems, soil, clay, rock, and other porous and non-porous media.
Ex vivo and in vitro evaluation of animals, plants, micro-organisms, organs, or other biological specimens for development, disease, injury, and wound healing monitoring.
Interested in licensing this patent?