WO2016186644A1

WO2016186644A1 - System and method of obtaining spatially-encoded nmr parameters from arbitrarily-shaped compartments and linear algebraic modeling

Info

Publication number: WO2016186644A1
Application number: PCT/US2015/031406
Authority: WO
Inventors: Yi Zhang; Jinyuan Zhou; Paul A. Bottomley
Original assignee: The Johns Hopkins University
Priority date: 2015-05-18
Filing date: 2015-05-18
Publication date: 2016-11-24

Abstract

An NMR method and system for acquiring and reconstructing a value of an NMR parameter spatially localized to a compartment of interest including performing a first MRI of a portion of a sample with a first MRI pulse sequence using the NMR system and using a set of k-space spatial encoding gradients or coil sensitivity encoding maps to obtain a first magnetic resonance image to identify a compartment of interest; generating a second MRI pulse sequence that encodes the NMR parameter with a subset of the set of k-space spatial encoding gradients or the coil sensitivity encoding maps; applying the second MRI pulse sequence using the NMR system to acquire spatial information relating to the NMR parameter from the compartment of interest; segmenting the first magnetic resonance image into a plurality of compartments that includes the compartment of interest; and reconstructing a value of the NMR parameter in the compartment.

Description

SYSTEM AND METHOD OF OBTAINING SPATIALLY-ENCODED NMR PARAMETERS FROM ARBITRARILY-SHAPED COMPARTMENTS AND LINEAR

ALGEBRAIC MODELING

[0001] This invention was made with U.S. Government support of Grant Nos.

EB007829, CA166171, EB009731, and NS083435 awarded by The Department of Health and Human Services, The National Institutes of Health (NIH). The U.S. Government has certain rights in this invention.

BACKGROUND

1. Field of Invention

[0002] The field of the currently claimed embodiments of this invention relates to systems and methods of obtaining spatially localized nuclear magnetic resonance parameters.

2. Discussion of Related Art

[0003] There are many parameters that can be measured by nuclear magnetic resonance

(NMR) and magnetic resonance imaging (MRI) including nuclear spin density, longitudinal relaxation time (Ti) (1), longitudinal relaxation time in the rotating frame (Tip) (2), transverse relaxation time (T₂) (1), the inhomogeneously broadened-T¾T₂*, relaxation time (3), apparent diffusion coefficients (ADC) (4), perfusion measures (5), functional MRI (fMRI) measures (6), spectral parameters (7), chemical reaction rates (8), magnetization transfer ratios (9), and chemical exchange saturation transfer (CEST) indices (10), etc, using hydrogen (¾), carbon (¹³C), fluorine (¹⁹F), sodium (²³Na) and phosphorus (³¹P) nuclei to name a few. These measures can derive from endogenous compounds present in biological systems, or from exogenous substances or tracers introduced into the system for the purpose of providing diagnostic, prognostic and/or therapeutic information.

[0004] In heterogeneous systems, such as the human body, animals or experimental animal models, these parameters are nonuniformly distributed, and therefore it is desirable to obtain measures that are localized to smaller partitions or regions of the system, such as an organ, a lesion, a pathology, or a region from which NMR or MRI information is being sought, within the organ. The standard way to achieve this is MRI which can provide a nearly continuous distribution of the parameters throughout the sample, limited only by the voxel resolution of the image and/or the time required for the scan. However, often there is inadequate time to provide images of the desired parameter at high image resolution due to insufficient signal-to-noise ratio (SNR), insufficient scan time when other clinical information must be acquired during an exam, for example, in the case of patient studies etc. In addition, image voxels acquired by MRI are rectangular and do not conveniently conform to arbitrarily-shaped compartments that coincide with the morphology of an organ or pathology of interest, such that simply increasing the image voxel size does not provide adequate coverage of the desired volume from which the parametric information is being sought, due to overlap of the rectangular voxel with other tissues, for example.

[0005] Consequently, there remains a need for systems and methods that can obtain

NMR or MRI parameters such as those noted above for at least compartments of interest much more rapidly than measures obtained with conventional NMR and MRI systems and methods.

SUMMARY

[0006] A method of localizing a nuclear magnetic resonance (NMR) parameter to at least one compartment of interest of a sample undergoing an NMR examination, comprising: placing the sample in an NMR system; performing a first magnetic resonance imaging (MRI) of at least a portion of the sample with a first MRI pulse sequence using the NMR system and using at least one of a set of k-space spatial encoding gradients or coil sensitivity encoding maps to obtain a first magnetic resonance image to identify at least one compartment of interest; generating a second MRI pulse sequence that encodes the NMR parameter with a subset of the at least one of the set of k-space spatial encoding gradients or the coil sensitivity encoding maps; applying the second MRI pulse sequence using the NMR system to acquire spatial information relating to the NMR parameter from the at least one compartment of interest; segmenting the first magnetic resonance image into a plurality of compartments that includes the at least one compartment of interest; and reconstructing a value of the NMR parameter in the at least one compartment of interest. [0007] An NMR system for acquiring and reconstructing a value of an NMR parameter spatially localized to at least one compartment of interest, the NMR system employing an applied static magnetic field Bo, and NMR excitation and detection systems, and comprising an MRI apparatus configured to accommodate a sample of interest, wherein the MRI apparatus is configured to: perform a first magnetic resonance imaging (MRI) of at least a portion of the sample with a first MRI pulse sequence using the NMR system and using at least one of a set of k-space spatial encoding gradients or coil sensitivity encoding maps to obtain a first magnetic resonance image to identify at least one compartment of interest; generate a second MRI pulse sequence that encodes the NMR parameter with a subset of the at least one of the set of k-space spatial encoding gradients or the coil sensitivity encoding maps; apply the second MRI pulse sequence using the NMR system to acquire spatial information relating to the NMR parameter from the at least one compartment of interest; segment the first magnetic resonance image into a plurality of compartments that includes the at least one compartment of interest; and reconstruct a value of the NMR parameter in the at least one compartment of interest.

[0008] These and other features and advantages will be apparent from a reading of the following detailed description and a review of the associated drawings. It is to be understood that both the foregoing general description and the following detailed description are explanatory only and are not restrictive of aspects as claimed.

BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Further objectives and advantages will become apparent from a consideration of the description, drawings, and examples.

[0010] FIG. 1 is a schematic illustration of nuclear magnetic resonance system according to an embodiment of the current invention.

[0011] FIG. 2A is a flow chart to explain an embodiment for implementing SLAM.

[0012] FIG. 2B is a flow chart to explain an embodiment of a protocol for using the

SLAM method to measure a compartmental average MRI parameter "x". [0013] FIGS. 3A-3E show five pulse sequences in accordance with embodiments of the present invention [a: inversion recovery; b: partial saturation; c: spin echo; d: multi-echo Carr- Purcell (CP) or CP-Meiboom-Gill (CPMG); e: gradient refocused echo] for measuring relaxation times. RF refers to the radiofrequency excitation pulse, PE stands for phase encoding, FE stands for frequency encoding, TE refers to echo time, TI is inversion delay time, TR is repetition time, AQ stands for acquisition, and Gx, Gy and Gz refer to the linear gradient fields in x, y, and z directions, respectively.

[0014] FIGS. 4A-4B show pseudo-first-order forward rates, kf, measured from 16-step

FT, compared with results from four-step "SLAM" method combined with the "TRiST" method (19); (a) and eight-step TRiST SLAM (b), in accordance with embodiments of the present invention. Coefficients of determination (r²) are 0.71 (a) and 0.90 (b). The solid line is the identity line.

[0015] FIG. 5 shows schematic steps of incorporating linear interpolation and Bo correction into SLAM. s₁ ... s_M denote standard &-space raw data acquired at M saturation frequencies. s[ ... s^f _N refer to interpolated SLAM hybrid A space data after Fourier transformation (FT) in the saturation frequency dimension, corresponding to s^! _M,„_N in Eq. [7]. p[ ... p^f _N represent intermediate SLAM results after correction for nonuniformity in the main magnetic field, Bo, corresponding to p^f _c,_N in Eq. [7]. p^r _c,_N is the inverse FT of p^f _c,_N in the saturation frequency offset dimension and gives the SLAM compartmental z-spectra.

[0016] FIGS. 6A-6D show an anatomical image, and SLAM overlaid with conventional

FT z-spectra for a patient with a low-grade oligodendroglioma, a: Segmentation of the gadolinium-enhanced Ti-weighted (Tlw) image showing tumor (1), contralateral brain (2), "rest of the brain" (3), scalp (4) and background (5) compartments, b-d: SLAM z-spectra (black) in accordance with embodiments of the present invention, overlaid onto conventional FT (gray) z- spectra for acceleration factors R = 4 (top), 9 (middle) and 45 (bottom) for the tumor (b), contralateral brain (c), and "rest of the brain" compartments (d). The horizontal axis indicates saturation frequency offsets and the vertical axis indicates the saturated CEST signal intensity normalized by the unsaturated signal intensity expressed as a percentage. [0017] FIGS. 7A-7D show an anatomical image, and SLAM overlaid with conventional

FT z-spectra for a patient with a high-grade anaplastic astrocytoma, a: Segmentation of the corresponding gadolinium-enhanced Tlw anatomical image showing tumor (1), contralateral brain (2), "rest of the brain" (3), scalp (4) and background (5). b-d: SLAM z-spectra (black) acquired using an embodiment of the present invention, overlaid onto conventional FT (gray) z- spectra for acceleration factors, R, of 4 (top), 9 (middle) and 45 (bottom), for the tumor (b), the contralateral brain (c), and the "rest of the brain" compartments (d). The horizontal axis indicates saturation frequency offsets and the vertical axis indicates the saturated CEST signal intensity normalized by the unsaturated signal intensity expressed as a percentage.

[0018] FIGS. 8A-8B show error analysis of SLAM CEST data acquired using an embodiment of the present invention, vs. conventional FT z-spectra at all frequency offsets, a: Mean ± SD of error percentage (SLAM/FT - 1) for all brain compartments from 8 subjects, with 12 different accelerations factors (R, horizontal axes), b: Pearson's correlation coefficients (mean ± standard deviation) comparing SLAM with conventional FT spectra for all brain compartments, data sets and acceleration factors.

DETAILED DESCRIPTION

[0019] Some embodiments of the current invention are discussed in detail below. In describing embodiments, specific terminology is employed for the sake of clarity. However, the invention is not intended to be limited to the specific terminology so selected. A person skilled in the relevant art will recognize that other equivalent components can be employed and other methods developed without departing from the broad concepts of the current invention. All references cited anywhere in this specification, including the Background and Detailed Description sections, are incorporated by reference as if each had been individually incorporated.

[0020] Recently, inventors of the current application described a new method of spatially localizing NMR spectroscopy (MRS) signals to whole user-defined compartments using linear algebraic modeling (SLAM) (11, 12). The method provides highly accelerated acquisitions that generate almost the maximum SNR that is achievable from the whole compartment. The central idea of SLAM is to group voxels defined by scout MRI into compartments, and reduce the number of phase-encoding gradient steps to a small subset of the original ones. This subset is comprised of phase-encoding steps typically chosen from the center of image &-space (the "k- space" of a two-dimensional [2D] image is equal to the 2D inverse Fourier transform [FT] of the image). The reduction is possible because the number of unknowns is significantly reduced from the number of voxels, to the number of compartments. The compartmental spectra are solved or reconstructed by incorporating the compartment segmentation information into the standard FT reconstruction model. SLAM has been applied to human cardiac ³¹P MRS (11) and brain Ή MRS (12), achieving 4- to 120-fold acceleration factors compared to conventional patient acquisitions, or up to 3 -fold higher SNR in heart or tumor compartments compared to summing the constituent MRS voxels post-acquisition, if the scan-time is held constant.

[0021] The term "sample" is intended to have a broad meaning. It can be, for example, an inorganic or organic material, a living or non-living organism, and/or a portion of an organism. In some cases, it can be a living human or animal.

[0022] The term "compartment" of a sample refers to a subsection of the sample. It could be, without limitation, an organ, or a section of an organ, of a living animal or person, for example. In another example, it could be a tumor or a portion of a tumor in an animal or a person, for example. Further, although a portion of a sample imaged in a voxel is also considered to be a "compartment", the term compartment is not limited to that example. A compartment can be a more general segment of the sample that could take many conventional voxels to represent an image of the compartment. The compartment can be defined in one- two- or three-dimensions (ID, 2D or 3D), for example, and can be of an arbitrary shape as defined by the user (including irregular shaped compartments that are not rectangular circular or spherical but have convex or concave edges, or that contain or surround other compartments, and or include a plurality of regions that have similar properties but which are physically isolated from each other).

[0023] Some embodiments of the current invention provide a novel extension of the

SLAM method to the localization and measurement of spatial averages from user-defined compartments, for a wide range of NMR parameters so they can be obtained at highly accelerated speeds, essentially permitting up to real time measurements. Embodiments of the current invention can be applied to any or all of the above mentioned parameters, including but not limited to— nuclear spin density, relaxation times (Ti, Tip, T₂, T₂*), diffusion coefficients, perfusion measures, functional measures, spectral parameters, chemical reaction rates, and chemical exchange saturation transfer (CEST) indices, using hydrogen ('Η), carbon (¹³C), fluorine (¹⁹F), sodium (²³Na) and phosphorus (³¹P) nuclei. (The U.S. App. Pub. No. 14/006,069, "SYSTEM AND METHOD OF PERFORMING MAGNETIC RESONANCE SPECTROSCOPIC IMAGING" assigned to the same assignee as the current application is hereby incorporated herein by reference.) Some methods and systems according to embodiments of the current invention can be especially useful for providing such measures in animal or patient MRI protocols in which such parametric information would otherwise not be obtainable due to a limited exam time or low sensitivity or SNR or the low concentration of metabolite or compound being interrogated by the NMR/MRI procedure. The fact that results can be provided essentially as fast as real time during an imaging procedure can provide capabilities that would otherwise not be possible. For example, a user could select portions of a scout image to obtain NMR parameter measurements of interest while a person is still present undergoing the imaging procedure. Post-processing analysis often results in the impractically of obtaining such information since the person would have to return for another imaging session after each postprocessing analysis.

[0024] Ref (11) described the application of the SLAM technique to MRS spatially in

ID. Ref (12) described the application of the SLAM technique to MRS spatially in 2D and 3D with sensitivity encoding (13). According to an embodiment of the current invention, we shall describe first how to extend SLAM techniques to localize any single one of the aforementioned parameters, in any of one to three spatial dimensions, including in combination with sensitivity encoding, which uses the NMR MRI coil sensitivity profiles to provide spatial encoding information. We will then present examples of applications in which embodiments of the current invention are used to measure relaxation times, diffusion coefficients, perfusion measures, chemical reaction rates, and the measurement of CEST indices.

[0025] Figure 1 is a schematic illustration of a nuclear magnetic resonance (NMR) system 100 according to an embodiment of the current invention. The NMR system 100 includes a magnetic resonance imaging (MRI) apparatus that is configured to accommodate a sample of interest. The MRI apparatus can include a magnetic resonance scanner 101, a data storage unit 108, and a data processing unit 109. Magnetic resonance scanner 101 has a main magnet 105 providing a substantially uniform main magnetic field Bo for a sample (subject or object) 102 under observation on scanner bed 103, a gradient system 106 providing a perturbation of the main magnetic field Bo to encode spatial information of the constituent molecules of subject 102 under observation, and a radio-frequency (RF) coil system 107 to transmit electromagnetic waves and to receive magnetic resonance signals from subject 102. The receive coil system may employ one or more receive coil elements, wherein use of a plurality or receive elements permits their use for spatial sensitivity ("SENSE") encoding.

[0026] Data storage unit 108 may be, for example, a hard disk drive, a network area storage (NAS) device, a redundant array of independent disks (RAID), a flash drive, an optical disk, a magnetic tape, a magneto-optical disk, etc. However, the data storage unit 108 is not limited to these particular examples. It can include other existing or future developed data storage devices without departing from the scope of the current invention.

[0027] The data processing system 109 is in communication with magnetic resonance scanner 101 to receive magnetic resonance signals for forming magnetic resonance images of sample 102. Data processing system 109 may be partially or totally incorporated within a structure housing magnetic resonance scanner 101. Data processing system 109 may be partially or totally incorporated in a workstation that is structurally separate from and in communication with magnetic resonance scanner 101. Data processing system 109 may be incorporated in a workstation that is structurally separate from and in communication with magnetic resonance scanner 101. An operator 113 may interact with the NMR system 100 with input/output device 112. (For a fuller description of a typical MRI/NMR system see also, for example, P. A. Bottomley et al, "High-field nuclear magnetic resonance imaging/spectroscopy system" US patent 4,689,563 Aug 25 1987)

[0028] The NMR system 100 can be further configured to perform a first magnetic resonance imaging (MRI) of at least a portion of the sample with a first MRI pulse sequence using the NMR system 100 and using at least one of a set of &-space spatial encoding gradients or coil sensitivity encoding ("SENSE") maps to obtain a first magnetic resonance image to identify at least one compartment of interest; generate a second MRI pulse sequence that encodes the NMR parameter with a subset of the at least one of the set of A:-space spatial encoding gradients or the coil sensitivity encoding maps; apply the second MRI pulse sequence using the NMR system to acquire spatial information relating to the NMR parameter from the at least one compartment of interest; segment the first magnetic resonance image into a plurality of compartments that includes the at least one compartment of interest; and reconstruct a value of the NMR parameter in the at least one compartment of interest. Methods:

Imaging localization with SLAM

[0029] In accordance with the conventions used in Ref. (12), the image reconstruction problem involves solving the linear equation,

for the image-space data, p , where s is the vectorized £-space raw data corresponding to the parameter of interest for which the compartment average measure is being sought (eg, Ti, T_lp, T₂, etc), E is the combined encoding matrix, M' is the total number of known A space data and M is the number of unknown image space voxels. M' is the product of the number of phase- encoding steps, the number of frequency-encoding steps and the number of coil elements used for sensitivity encoding (Nc).

[0030] For multi-receive acquisition (Nc > 1), E can be constructed as,

where PE is the phase-encoding operator, FE is the frequency-encoding operator which now corresponds to the read-out gradient, SE is the sensitivity encoding matrix for each coil element (1, ..., Nc), ® is the Kronecker operator (14), and M' ' (= M'INc) is the product of the number of phase-encoding steps (Μ" ') and the number of frequency-encoded points (M" I M'"). When frequency-encoding or sensitivity-encoding is not used, for example, in one dimensional MRS, the FE and SE are simply 1.

[0031] Typically for MRI however, because the frequency-encoding dimension is fully- sampled and 2D FT is separable in both phase- and frequency-encoding dimensions, a first FT is applied in the frequency-encoding (read-out) dimension before constructing Eqs. [1-2]. FE in Eq. [2] is then replaced by an identity matrix, and s is replaced by a hybrid -space vector of the FT of the raw data in the frequency-encoding dimension. Furthermore, a noise covariance matrix can be included in both the s hybrid matrix and the E operator using a pre-whitening strategy (15). For simplicity, the same notations are used after these preprocessing steps. Thus, for a single-channel receive acquisition (Nc = 1), E is simply (PE ⁽S> FE)_M„_M , where FE is an identity matrix and s is the FT of &-space raw data in the frequency-encoding dimension.

[0032] The SLAM method incorporates compartment segmentation and correction information (12) into Eq. [1] as: s_/w ~ E_M.._M x ^,^ x b_M,_M x b_M._M x Α_Μ,_Μ x p_M . [3]

[0033] Here the auxiliary b matrix contains the spatial localization information relating to the segmented compartments (11, 12), and A is a correction matrix accounting for spatial and temporal main magnetic field (Bo) and RF excitation/detection field (Bi) inhomogeneity. Dimensionality reduction is performed whereby only the compartmental average values of the signal are retained, leading to,

^SM_' ⁼ ^M'*M ^x A_M._M x b_M,_c x p_c , [4] where b^r preserves the C columns in b^~1 corresponding to the C compartmental average values in p^r . Two possible solutions to Eq. [4] follow:

Pc ⁼ (E_/W_'*M ^x A/ Hw x b_M,_c) x s_M, [5] and p^r _c = (b^c )⁺ x A_M¾J x (E_M»_M )⁺ x s_M, , [6] where "+" denotes the Moore-Penrose pseudo-inverse. A truncated singular value decomposition (TSVD) (16) strategy is used for computing the pseudo-inverses by discarding singular values that are less than 2% of the maximum ones. The second solution, Eq. [6], was shown to be more compatible with receiver coil sensitivity encoding (SENSE), than Eq. [5] (12), and thus is chosen when SENSE is used. M' for SLAM can be much smaller than that used for conventional SENSE MRI (13) involving FT followed by SENSE unfolding, because only the C unknown compartmental values are solved. [0034] The steps of implementing SLAM are shown in Fig. 2A. Each application begins with the acquisition of a simple "scout MRI", which is subsequently used to segment out the compartments from which the desired compartment-average measurements are to be obtained. Note that the term "scout image" or "scout MRI" means any MRI acquired during the scan that is used for the purpose of segmenting out compartments, and need not be an image acquired specifically for this purpose. For example, it could be an image acquired for other diagnostic purposes, just as long as it provides sufficient anatomic detail to permit segmentation. After the scout image is obtained, a regular MRI sequence that incorporates encoding for the parameter of interest is applied (Ti, T¾ etc), with the critical change that the number of spatial encoding steps for the MRI sequence are drastically reduced, in accordance with the SLAM method, and the parametric information reconstructed as described above, and exemplified below.

Application of SLAM to measure relaxation times

[0035] Contrast information predominantly attributed to the distinct relaxation times (Ti,

Ti_p, T¾ and T₂*) of different tissues or compartments is the pillar upon which the widespread use of MRI in the clinic and in research stands. However, the durations required for quantitative relaxation time mapping are long, resulting in the compromise use of qualitative relaxation time weighted-MRI in many clinical studies, which limits inter- and intra-institutional reproducibility and comparisons. This sacrifices the potentially useful diagnostic or prognostic information provided by quantitative relaxation times: specific relaxation times are often not even measured, with "Ti-weighted" (Tlw), "T₂-weighted" (T2W), "diffusion-weighted" (Dw) and the like, images routinely used in their stead. However, SLAM can be extended from the original spectroscopy application, to fundamentally alter this scan time problem by greatly speeding up the scan-time required for compartment average NMR relaxation time measurements in accordance with the current invention.

[0036] Specifically, in a first embodiment of the SLAM method designed to obtain quantitative compartment-average Ti measurements: after scout MRI is performed, an inversion recovery MRI sequence (Fig. 3A) comprised of an inversion MRI pulse with a flip angle (FA) sufficient to invert the nuclear magnetization (e.g., FA ~ 180°) followed by an MRI excitation pulse (FA < 90°) is applied at an inversion delay time (TI) later with a reduced number of phase encoding steps, M" chosen from central A space in accordance with the SLAM method as described above. The inversion recovery sequence is used to generate a set of Ti "parameter- sensitive" MRI data, by repeatedly applying the sequence with different values of TI (e.g., 6 values) that span a range of time of approximately 1 to 3 times the Ti values present in the sample/subject being investigated, as is known by those skilled in the art; and repeated with the different spatial phase encoding steps to complete the reduced subset of M" ' phase encoding steps. Then, for each TI, an average value of the Ti -weighted signal from each compartment, p^r , is reconstructed with Eq. [5] or [6]. With multiple TI and p^r values available, compartmental-average Ti values are solved by fitting a 3-parameter exponential function, p^r = u-v · exp(-TI/T,) to the set of p^r data acquired at the different TI values, where u, v, and Ti are unknown constants to be determined.

[0037] In a second embodiment of the SLAM method designed to obtain quantitative compartment-average Ti measurements: after scout MRI, a partial saturation MRI sequence (Fig. 3B) comprised of a series of equally spaced MRI excitation pulses with the same FA (typically, FA < 90°) separated in time by the same repetition period, TR, is applied with a reduced number of phase encoding steps M' ' ' chosen from central &-space in accordance with the SLAM method as described above. The sequence is repeated with different values of TR (e.g., 6 values) that span a range of time of approximately 1 to 3 times the Ti values present in the sample/subject being investigated, as is known by those skilled in the art to generate a set of Ti -sensitive MRI data; and repeated with the different phase encoding steps to complete the reduced subset of M' " steps. Then, for each TR value, an average value of the Ti -weighted signal from each compartment, p^r , is reconstructed using Eq. [5] or [6]. With multiple TR and p^r values available, compartmental-average Ti values are the solved by fitting a 3-parameter exponential function, p^r = u-v · exp(-TR/T.,) to the set of p^r data acquired at the different TR values, where u, v, and Ti are (generally different) unknown constants to be determined.

[0038] In a third embodiment of the SLAM method designed to obtain quantitative compartment-average T₂ measurements, after scout MRI, a single echo time (TE) MRI sequence (Fig. 3C) is used with SLAM. The sequence is comprised of an MRI excitation pulse (typical FA < 90°) followed by an echo-producing pulse with FA substantially equal to 180° at time TE/2. The sequence is repeated for a range of different TE values (e.g. ~6) that span a range of time of approximately 1 to 3 times the T₂ values present in the sample/subject being investigated to generate a set of T₂ parameter-sensitive MRI data, as is known by those skilled in the art; and repeated as well with the different phase encoding steps to complete the full complement of the SLAM M'" phase encoding steps. For each T₂ weighted signal recorded at each TE value, a compartmental-average MRI signal value, p^r , is reconstructed using Eq. [5] or [6]. Then, the known TE and p^r values are fitted to a 2-parameter exponential function, p^r = w · exp(-TE/T₂) where w is an unknown constant, and solved to obtain the T₂ of each compartment.

[0039] In a fourth embodiment of the SLAM method designed to obtain quantitative compartment-average T₂ measurements, after scout MRI, a multiple echo time (TE) MRI sequence (Fig. 3D) is used with SLAM. The sequence is comprised of an MRI excitation pulse (typical FA < 90°) followed by scries of n_e echo-producing pulses with FA substantially equal to 180°, repeated at intervals of TE, starting at time TE/2 after the excitation pulse, such that the full range of n_e echoes span a range of time of approximately 1 to 3 times the T₂ values present in the sample/subject being investigated, as is known by those skilled in the art, and the set of echo data acquired following the n_e pulses comprises a set of T₂ parameter-sensitive MRI data. The entire sequence is repeated with the different phase encoding steps so as to complete acquisition of the full complement of the SLAM M' " spatial encoding steps. For each T₂- weighted signal recorded at each TE value, a compartmental-average MRI signal value, p^r , is reconstructed using Eq. [5] or [6]. Then, the Icnown TE and p^r values are fitted to a 2-parameter exponential function, p^r = w · exp(-TE/T₂) where w is an unknown constant, and solved to obtain the T₂ of each compartment.

[0040] In a fifth embodiment of the SLAM method designed to obtain quantitative compartment-average T₂* measurements: after scout MRI, a gradient refocused echo MRI sequence (Fig. 3E) comprised of a series of equally spaced MRI excitation pulses is applied with the same FA (typically, FA < 90°) to generate a gradient refocused echo at time TE later. The sequence is repeated with different values of TE that approximately span the range of T₂* values in the sample as is known by those skilled in the art, and with the reduced number of phase encoding steps M' " chosen from central A space in accordance with the SLAM method as described above. Then, for each TE value, an average value of the T₂*-weighted signal from each compartment, p^r , is reconstructed using Eq. [5] or [6]. With multiple TR and p^r values available, compartmental-average T₂* values are solved by fitting a 2-parameter exponential function, p^r = w · exp(-TE/T₂*) to the set of p^r data acquired at the different TE values where w and T₂* are unknown constants to be determined.

[0041] Tip can be solved similarly in further embodiments of this invention wherein the

Ti and T₂ sensitive MRI sequences are replaced by a T_lp MRI sequence with a reduced number of spatial encoding steps whereupon the SLAM reconstruction method is used to provide a dramatic reduction in scan time for the compartment average measurement, as compared to a fully resolved image.

Application of SLAM to measure diffusion and perfusion indices

[0042] In a sixth embodiment of the SLAM method designed to obtain quantitative compartment-average ADCs, after scout MRI, a diffusion-weighted MRI sequence is used with SLAM. The sequence is comprised of an MRI excitation pulse (typical FA < 90°) followed by an echo-producing pulse with FA substantially equal to 180° and with a pair of diffusion- weighting gradient pulses with one gradient pulse inserted before and one inserted after the 180° pulse. During the gradient pulse, the gradient strength can be static (in accordance with the pulse gradient spin echo, [PGSE] method) or oscillate (in accordance with the oscillating gradient spin echo, [OGSE] method) as is known to those skilled in the art. The sequence is repeated for a range of different diffusion weighting values (b, in s/mm²) that match the range of the ADC value (in mm²/s) present in the sample/subject being investigated, as is known by those skilled in the art; and repeated as well, with the different phase encoding steps to complete the full complement of the SLAM M' " phase encoding steps, and to generate thereby, a set of diffusion parameter-sensitive MRI data. For each diffusion weighted signal recorded at each b value, a compartmental-average MRI signal value, p^r , is reconstructed using Eq. [5] or [6]. Then, the known b and p^r values are fitted to a 2-parameter exponential function, p^r = W · exp(-b · ADC) where w is an unknown constant, and solved to obtain the ADC of each compartment.

[0043] In a seventh embodiment of the SLAM method designed to obtain quantitative compartment-average perfusion measures, after scout MRI, an arterial spin labeling MRI sequence is used with SLAM. In prior art, arterial spin labeling (ASL) MRI is used to measure tissue perfusion by deploying an MRI protocol wherein a first MRI is acquired with a spatially selective pulse turned on (the "labeling" pulse) and applied on the inflowing blood side of the tissue of interest. Then a second MRI is acquired either with the labeling pulse turned off or applied equidistant from the tissue of interest as the labeling pulse, but on the outflowing blood side (the "control"). The labeling pulse is applied at a location from the tissue of interest commensurate with the range of influx rates of the arterial blood that infuses it. Subtraction of the label image from the control image provides a tissue perfusion measure. For the new embodiment wherein SLAM is used to obtain quantitative compartment-average perfusion measurements, the ASL sequence is implemented with spatial selective labeling and control pulses, but with the reduced number of spatial encoding steps in accordance with the SLAM method described above. Specifically, the ASL sequence with labeling and control pulses comprise a set of perfusion parameter-sensitive MRI data: each are repeated with the M' ' ' phase encoding steps to complete the full complement of the SLAM encoding set. For each of the label and control acquisitions, a compartmental-average MRI signal value, p^r , is reconstructed using Eq. [5] or [6]. Then, the subtraction of the two p^r values gives compartmental-average perfusion measurement.

Application of SLAM to measure creatine kinase reaction rates

[0044] Adenosine triphosphate (ATP) is the most important energy source in the human body, which can be generated along with free creatine (Cr) through the following creatine kinase (CK) reaction from phosphocreatine (PCr) and adenosine diphosphate (ADP).

PCr + ADP < ^kf

k_r > ATP+Cr where k_f and k_r are the pseudo-first-order forward and reverse reaction rates, and specifically k_f refers to the percentage of PCr that is converted to ATP per second. The forward CK ATP flux in heart failure patients was found to be reduced significantly compared to that in normal subjects (17). Central to measuring CK ATP flux, the forward reaction rate, k_f , is currently most rapidly measured by the "FAST" (18) or "TRiST" (19) methods which measure PCr from ³¹P MRS spectra repetitively using different TR values during which time the ATP peak is saturated with chemical selective NMR excitation or saturation signal, and with a control chemically-selective NMR excitation signal spectrally directed at the opposite side of PCr compared to ATP. However, even these methods require ~40 min for cardiac applications, which may be intolerable for some patients, or which limit acquisition of other useful cardiac MRI information, and which make such measurements susceptible to motion artifacts. The scan time can be significantly reduced if the MRS spectra used in TRiST or FAST are directly measured by SLAM.

[0045] In an eighth embodiment, Eq. [5] with FE and SE set to 1 is used to reconstruct

MRS spectra directly with SLAM using the reduced £-space data sets based on TRiST or FAST encoding (which form a set of reaction rate-sensitive NMR data), which are thereby accelerated by a factor proportional to the reduction in A-space encodings. For example, results of SLAM compared with the standard FT method are shown in Figs. 4A-4B, acquired with two different acceleration factors, R = 2, 4 corresponding to 2- and 4-fold reductions in the number of &-space encodings, to achieve 2- and 4-fold reductions in scan-time. The data in Figs. 4A-4B illustrates measures of the cardiac-compartment average forward CK reaction rate, k_f , obtained from 10 normal volunteers and 9 heart patients. Adaptation of SLAM to CEST

[0046] In CEST MRI, images are typically acquired repetitively using a gradient refocused echo or spin-echo sequence (e.g., in Figs 3C-3E), except that the sequence is preceded with a frequency (or chemically-) selective saturation NMR pulse. The frequency of the saturation is advanced in repeat applications of the sequence to generate a set of CEST parameter-sensitive MRI data from which images are obtained. The variation of the signal in each pixel as a function of saturation frequency is known as the "z-spectrum": asymmetry in the z-spectra about the water resonance has been shown to have potential diagnostic value (10).

[0047] In the CEST experiment, correcting the nominal saturation frequency offset for the main magnetic field (Bo) inhomogeneity is crucial for generating meaningful CEST maps or z-spectra in the saturation frequency offset domain (20,21). In conventional CEST MRI, z- spectra in image space are first linearly interpolated to a finer resolution along the saturation frequency offset dimension and then shifted based on a separately acquired Bo map on a voxel- by-voxel basis (22). [0048] In a ninth exemplary embodiment wherein SLAM is applied to CEST MRI to yield compartment average CEST measurements, this Bo correction is accommodated in SLAM, which operates in £-space instead of image space, with the performance of two steps. First, a linear interpolation is applied along the saturation frequency offset dimension, to CEST A space raw data that is acquired with the SLAM-reduced spatial phase encoding. This interpolation is equivalent to interpolation in image space, of the frequency offset domain. Second, a Bo correction is implemented by incorporating a regular Bo map that is ordinarily acquired for shimming purposes as part of an MRI exam, into the A term in Eq. [6], after stacking the p^r _c solutions for all offset frequencies. The FT of the stack in the offset dimension is then performed, yielding

Pc_*w ⁼ (k_/w_*c) ^x A_M._M x (E_M„_M) x s_M„_w , [7] where s^f _M„_N is the FT of stacked s_M,»_w in the frequency offset dimension; N is the number of interpolated frequency offsets; and p^f _c,_N is the FT of stacked p^r _c._N in the offset dimension. In

Eq. [7], A = e-'^{'2 *8}^™ i_s different for each frequency offset index, j, with -N/2 <j < Nil, and B₀(x,y,z) is the standard Bo spatial correction map. Finally, an inverse FT of p^f _c,_N leads to the compartmental-average z-spectrum in image space, p^._w . Schematic steps for incorporating interpolation and shifting into SLAM are shown in Fig. 5.

[0049] Another important adaptation that is not yet included in Eq. [7] is the normalization of z-spectra, wherein the saturated image is divided by the unsaturated reference image. The unsaturated reference image is not subjected to any interpolation or shifting operations. To accommodate normalization, Eq. [7] can be modified to:

PC*A/ = K*cY X ^AM*M {[(^{E »}M)^{+ S}L^»J^■> '*M)^{+ S}M"J}> [⁸] where . / denotes element-wise division of two matrices and s°_M,._N is the stacked &-space data from the unsaturated scan obtained by replicating acquired raw data, s°_M, , N times in the frequency offset domain. All other processing steps are the same as shown in Fig. 5, with the data sets comprising a set of CEST-sensitive MRI data. [0050] Figs. 6A-6D compares SLAM z-spectra (black) with conventional FT z-spectra

(gray) in a patient with a low-grade oligodendroglioma, at three acceleration factors (top to bottom: R = 4, 9, and 45) corresponding to up to 45-fold reductions in scan-time. SLAM z- spectra from the tumor (Fig. 6B), contralateral (Fig. 6C) and "rest of the brain" (Fig. 6D) compartments are essentially coincident with the conventional FT spectra. Note that, with R = 45, SLAM used only a single phase-encoding A:-space line: the rest of the spatial information derives from the sensitivity encoding using the multi-element head detector coil. Figs. 7A-7D compares SLAM and conventional FT z-spectra for another patient with a high-grade anaplastic astrocytoma. Again the z-spectra from the three compartments are highly coincident, especially with an R = 4 (top) acceleration factor.

[0051] Figs. 8A-8B plots the error and Pearson's correlation coefficients (r) for the entire z-spectra obtained from SLAM in 8 subjects, as compared to those from conventional FT. The percentage error (Fig. 8A) increases as the acceleration factor grows, with a mean value of < 3.0% and SD of < 8.3%, comparable to those measured at frequency offsets of ±3.5 ppm. The Pearson's correlation coefficients remain high for all R, dropping to 0.9995 ± 0.0009% (mean ± SD) at R = 45, reflecting the high coincidence of the SLAM and conventional FT z-spectra overall (Figs. 6A-7D).

[0052] SLAM is a new method to accelerate the acquisition of NMR and MRI measurements that are spatially localized to arbitrarily shaped compartments, many-fold. If compartment-average indices can suffice, SLAM can speed up clinical or animal studies by up to scores-fold and generate quantitatively equivalent results to compartment average measures obtain by conventional (FT) MRI methods. SLAM's ability to yield MRI measures within scan times of a minute or less should facilitate its application in many clinical settings where scan time is at a premium.

REFERENCES

1. Bottomley PA, Foster TH, Argersinger RE, Pfeifer LM. A review of normal tissue

hydrogen NMR relaxation times and relaxation mechanisms from 1-100 MHz:

dependence on tissue type, NMR frequency, temperature, species, excision, and age. Med Phys 1984; 11 :425. Borthakur A, Mellon E, Niyogi S, Witschey W, Kneeland JB, Reddy R. Sodium and Tip MRI for molecular and diagnostic imaging of articular cartilage. NMR Biomed

2006;19(7):781-821.

Chavhan GB, Babyn PS, Thomas B, Shroff MM, Haacke EM. Principles, Techniques, and Applications of T2*-based MR Imaging and Its Special Applications 1.

Radiographics 2009;29(5): 1433-1449.

Basser PJ, Mattiello J, LeBihan D. MR diffusion tensor spectroscopy and imaging. Biophys J 1994;66(1):259.

Detre JA, Leigh JS, Williams DS, Koretsky AP. Perfusion imaging. Magn Reson Med 1992;23(l):37-45.

Biswal B, Zerrin Yetkin F, Haughton VM, Hyde JS. Functional connectivity in the motor cortex of resting human brain using echo-planar mri. Magn Reson Med 1995;34(4):537- 541.

Posse S, Otazo R, Dager SR, Alger J. MR spectroscopic imaging: principles and recent advances. J Magn Reson Imaging 2013;37(6): 1301-1325i

Forsen S, Hoffman RA. Study of moderately rapid chemical exchange reactions by means of nuclear magnetic double resonance. The Journal of Chemical Physics

1963;39(11):2892-2901.

Wolff SD, Balaban RS. Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo. Magn Reson Med 1989;10(1):135-144.

Ward K, Aletras A, Balaban R. A new class of contrast agents for MRI based on proton chemical exchange dependent saturation transfer (CEST). J Magn Reson

2000;143(l):79-87.

Zhang Y, Gabr RE, Schar M, Weiss RG, Bottomley PA. Magnetic resonance

Spectroscopy with Linear Algebraic Modeling (SLAM) for higher speed and sensitivity. J Magn Reson 2012;218:66-76.

Zhang Y, Gabr RE, Zhou J, Weiss RG, Bottomley PA. Highly-accelerated quantitative 2D and 3D localized spectroscopy with linear algebraic modeling (SLAM) and sensitivity encoding. J Magn Reson 2013;237: 125-138.

Pruessmann KP, Weiger M, Scheidegger MB, Boesiger P. SENSE: sensitivity encoding for fast MRI. Magn Reson Med 1999;42(5):952-962. 14. Loan CFV. The ubiquitous Kronecker product. J Comput Appl Math 2000;123(1):85- 100.

15. Pruessmann KP, Weiger M, Bornert P, Boesiger P. Advances in sensitivity encoding with arbitrary k-space trajectories. Magn Reson Med 2001;46(4):638-651.

16. Hoge WS, Brooks DH, Madore B, Kyriakos WE. A tour of accelerated parallel MR

imaging from a linear systems perspective. Concepts Magn Reson Part A 2005;27(1): 17- 37.

17. Weiss RG, Gerstenblith G, Bottomley PA. ATP flux through creatine kinase in the

normal, stressed, and failing human heart. Proc Natl Acad Sci U S A 2005;102(3):808- 813.

18. Bottomley PA, Ouwerkerk R, Lee RF, Weiss RG. Four-angle saturation transfer (FAST) method for measuring creatine kinase reaction rates in vivo. Magn Reson Med

2002;47(5):850-863.

19. Schar M, El-Sharkawy AMM, Weiss RG, Bottomley PA. Triple repetition time

saturation transfer (TRiST) ³¹P spectroscopy for measuring human creatine kinase reaction kinetics. Magn Reson Med 2010;63(6): 1493-1501.

20. Zhou J, Blakeley JO, Hua J, Kim M, Laterra J, Pomper MG, van Zijl P. Practical data acquisition method for human brain tumor amide proton transfer (APT) imaging. Magn Reson Med 2008;60(4):842-849.

21. Kim M, Gillen J, Landman BA, Zhou J, van Zijl P. Water saturation shift referencing (WASSR) for chemical exchange saturation transfer (CEST) experiments. Magn Reson Med 2009;61(6): 1441-1450.

22. Zhou J, Zhu H, Lim M, Blair L, Quinones - Hinojosa A, Messina SA, Eberhart CG,

Pomper MG, Laterra J, Barker PB. Three-dimensional amide proton transfer MR imaging of gliomas: Initial experience and comparison with gadolinium enhancement. J Magn Reson Imaging 2013;38(5):1 1 19-1 128.

[0053] The embodiments illustrated and discussed in this specification are intended only to teach those skilled in the art how to make and use the invention. In describing embodiments of the invention, specific terminology is employed for the sake of clarity. However, the invention is not intended to be limited to the specific terminology so selected. The above- described embodiments of the invention may be modified or varied, without departing from the invention, as appreciated by those skilled in the art in light of the above teachings. It is therefore to be understood that, within the scope of the claims and their equivalents, the invention may be practiced otherwise than as specifically described.

Claims

WE CLAIM:

1. A method of localizing a nuclear magnetic resonance (NMR) parameter to at least one compartment of interest of a sample undergoing an NMR examination, comprising: placing said sample in an NMR system; performing a first magnetic resonance imaging (MRI) of at least a portion of said sample with a first MRI pulse sequence using said NMR system and using at least one of a set of k- space spatial encoding gradients or coil sensitivity encoding maps to obtain a first magnetic resonance image to identify at least one compartment of interest; generating a second MRI pulse sequence that encodes said NMR parameter with a subset of said at least one of said set of k-space spatial encoding gradients or said coil sensitivity encoding maps; applying said second MRI pulse sequence using said NMR system to acquire spatial information relating to said NMR parameter from said at least one compartment of interest; segmenting said first magnetic resonance image into a plurality of compartments that includes said at least one compartment of interest; and reconstructing a value of said NMR parameter in said at least one compartment of interest.

2. The method of claim 1, wherein said at least one of said set of k-space spatial encoding gradients or coil sensitivity encoding maps is said coil sensitivity encoding maps.

3. The method of claim 1, wherein said subset of said set of k-space spatial encoding gradients are chosen from a center of k-space.

4. The method of claim 1, wherein said NMR parameter is one of a nuclear spin density, a relaxation time Ti, Tip, T₂, or T₂*, a diffusion coefficient, a perfusion measure, a functional measure, a spectral parameter, a chemical reaction rate, or a chemical exchange saturation transfer (CEST) index.

5. The method of claim 1, wherein said value of said NMR parameter obtained from said at least one compartment of interest, substantially represents an average value of said NMR parameter in said at least one compartment of interest.

6. The method of claim 1, wherein said at least one compartment of interest has an irregular shape.

7. The method of claim 4, further comprising applying said second MRI pulse sequence a plurality of times, wherein each of said plurality of applications of said second MRI pulse sequence is encoded to cause a corresponding MRI signal to be sensitive to one parameter selected from a nuclear spin density, a relaxation time Ti, Tip, T₂, or T₂*, a diffusion coefficient, a perfusion measure, a functional measure, a spectral parameter, a chemical reaction rate, or a chemical exchange saturation transfer (CEST) index, said corresponding MRI signal thereby being a parameter-sensitive MRI signal.

8. The method of claim 7, further comprising reconstructing for each of said plurality of applications of said second MRI pulse sequence a corresponding compartment average measure of said parameter-sensitive MRI signal, wherein each said compartment average measure of the said parameter-sensitive MRI signal is used to determine a compartment average measure of one of a nuclear spin density, a relaxation time Ti, Ti_p, T₂, or T₂*, a diffusion coefficient, a perfusion measure, a functional measure, a spectral parameter, a chemical reaction rate, or a chemical exchange saturation transfer (CEST) index.

9. The method of claim 7, wherein an agent by which said MRI signal from said second MRI pulse sequence is rendered sensitive to a relaxation time is one of a pulse sequence timing parameter.

10. The method of claim 9, wherein said pulse sequence timing parameter is one of a pulse sequence repetition period, an inversion time, or an echo time.

11. The method of claim 7, wherein an agent by which said MRI signal from said second MRI pulse sequence is rendered sensitive to a diffusion coefficient is a pulsed magnetic field gradient or oscillating magnetic field gradient.

12. The method of claim 7, wherein an agent by which said MRI signal from said second MRI pulse sequence is rendered sensitive to perfusion is a slice selective MRI pulse applied to one of inflowing or outflowing blood.

13. The method of claim 7, wherein an agent by which said MRI signal from said second MRI pulse sequence is rendered sensitive to a chemical reaction rate is a chemical selective irradiation pulse applied to a chemical specie participating in a corresponding chemical reaction.

14. The method of claim 7, wherein an agent by which said MRI signal from said second MRI pulse sequence is rendered sensitive to a CEST index is a frequency selective saturation NMR pulse with an offset frequency.

15. The method of claim 14, wherein said MRI signal is corrected for inhomogeneity in a main magnetic field (Bo) by a correction process, said correction process comprising: stacking CEST parameter sensitive MRI data for each offset frequency such that corresponding data sets are aligned in a saturation frequency dimension; linearly interpolating image &-space data in a saturation frequency domain; incorporating a regular Bo map into a correction matrix that accounts for spatial and temporal variations in applied magnetic fields; and reconstructing a compartment-average CEST sensitive parameter for each Bo-corrected offset frequency.

16. The method of claim 14, wherein an additional MRI data set is acquired in which no frequency-selective saturation is applied, and wherein a CEST sensitive MRI signal at each offset frequency is normalized by dividing by a signal acquired without frequency-selective saturation for said at least one compartment.

17. An NMR system for acquiring and reconstructing a value of an NMR parameter spatially localized to at least one compartment of interest, said NMR system employing an applied static magnetic field Bo, and NMR excitation and detection systems, and comprising an MRJ apparatus configured to accommodate a sample of interest, wherein said MRI apparatus is configured to: perform a first magnetic resonance imaging (MRI) of at least a portion of said sample with a first MRI pulse sequence using said NMR system and using at least one of a set of k- space spatial encoding gradients or coil sensitivity encoding maps to obtain a first magnetic resonance image to identify at least one compartment of interest; generate a second MRI pulse sequence that encodes said NMR parameter with a subset of said at least one of said set of k-space spatial encoding gradients or said coil sensitivity encoding maps; apply said second MRI pulse sequence using said NMR system to acquire spatial information relating to said NMR parameter from said at least one compartment of interest; segment said first magnetic resonance image into a plurality of compartments that includes said at least one compartment of interest; and reconstruct a value of said NMR parameter in said at least one compartment of interest.

18. The NMR system of claim 17, wherein said at least one of said set of k-space spatial encoding gradients or coil sensitivity encoding maps is said coil sensitivity encoding maps.

19. The NMR system of claim 17, wherein said subset of said set of k-space spatial encoding gradients are chosen from a center of k-space.

20. The NMR system of claim 17, wherein said NMR parameter is one of a nuclear spin density, a relaxation time Ti, Tip, T₂, or T₂*, a diffusion coefficient, a perfusion measure, a functional measure, a spectral parameter, a chemical reaction rate, or a chemical exchange saturation transfer (CEST) index.

21. The NMR system of claim 17, wherein said value of said NMR parameter obtained from said at least one compartment of interest, substantially represents an average value of said NMR parameter in said at least one compartment of interest.

22. The NMR system of claim 17, wherein said at least one compartment of interest has an irregular shape.

23. The NMR system of claim 20, further comprising applying said second MRI pulse sequence a plurality of times, wherein each of said plurality of applications of said second MRI pulse sequence is encoded to cause a corresponding MRI signal to be sensitive to one parameter selected from a nuclear spin density, a relaxation time Ti, Tip, T₂, or T₂*, a diffusion coefficient, a perfusion measure, a functional measure, a spectral parameter, a chemical reaction rate, or a chemical exchange saturation transfer (CEST) index, said corresponding MRI signal thereby being a parameter-sensitive MRI signal.

24. The NMR system of claim 23, further comprising reconstructing for each of said plurality of applications of said second MRI pulse sequence a corresponding compartment average measure of said parameter-sensitive MRI signal, wherein each said compartment average measure of the said parameter-sensitive MRI signal is used to determine a compartment average measure of one of a nuclear spin density, a relaxation time Ti, Tip, T₂, or T₂*, a diffusion coefficient, a perfusion measure, a functional measure, a spectral parameter, a chemical reaction rate, or a chemical exchange saturation transfer (CEST) index.

25. The NMR system of claim 24, wherein an agent by which said MRI signal from said second MRI pulse sequence is rendered sensitive to a relaxation time is one of a pulse sequence timing parameter.

26. The NMR system of claim 25, wherein said pulse sequence timing parameter is one of a pulse sequence repetition period, an inversion time, or an echo time.

27. The NMR system of claim 24, wherein an agent by which said MRI signal from said second MRI pulse sequence is rendered sensitive to a diffusion coefficient is a pulsed magnetic field gradient or oscillating magnetic field gradient.

28. The NMR system of claim 24, wherein an agent by which said MRI signal from said second MRI pulse sequence is rendered sensitive to perfusion is a slice selective MRI pulse applied to one of inflowing or outflowing blood.

29. The NMR system of claim 24, wherein an agent by which said MRI signal from said second MRI pulse sequence is rendered sensitive to a chemical reaction rate is a chemical selective irradiation pulse applied to a chemical specie participating in a corresponding chemical reaction.

30. The NMR system of claim 24, wherein an agent by which said MRI signal from said second MRI pulse sequence is rendered sensitive to a CEST index is a frequency selective saturation NMR pulse with an offset frequency.

31. The NMR system of claim 30, wherein said MRI signal is corrected for inhomogeneity in a main magnetic field (Bo) by a correction process, said correction process comprising: stacking CEST parameter sensitive MRI data for each offset frequency such that corresponding data sets are aligned in a saturation frequency dimension; linearly interpolating image &-space data in a saturation frequency domain; incorporating a regular Bo map into a correction matrix that accounts for spatial and temporal variations in the applied magnetic fields; and reconstructing a compartment-average CEST sensitive parameter for each Bo-corrected offset frequency.

32. The NMR system of claim 30, wherein an additional MRI data set is acquired in which no frequency-selective saturation is applied, and wherein a CEST sensitive MRI signal at each offset frequency is normalized by dividing by a signal acquired without frequency-selective saturation for said at least one compartment.