Multi-component magnetic resonance imaging (MRI) using steady-state acquisitions for practical relaxometry and myelin water fraction imaging

Abstract

A myelin water fraction (MWF) map image is produced using a magnetic resonance imaging (MRI) system. Here, a plurality of spoiled gradient recalled echo (SPGR) images of a specimen are acquired with a plurality of different flip angles (FAs). A plurality of balanced steady-state free precession (bSSFP) images of the specimen with a second plurality of different FAs and a plurality of different repetition times (TRs) are also acquired. Using driven equilibrium single pulse observation of longitudinal relaxation times (DESPOT1), a longitudinal relaxation time (T1) map from the SPGR images is generated. Using the T1 map and the bSSFP images, USA is applied and an off-resonance frequency (Δω) parameter map and a transverse relaxation time (T2) map are generated. Using Bayesian Monte-Carlo (BMC) analysis, the MWF map image is generated from the SPGR images, a subset of the bSSFP images, and the Δω parameter map.

Description

Leydig 775185HHSRef E-085-2024-0-PCT-011MULTI-COMPONENT MAGNETIC RESONANCE IMAGING (MRI) USING STEADYSTATE ACQUISITIONS FOR PRACTICAL RELAXOMETRY AND MYELIN WATER FRACTION IMAGINGCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to U. S. Provisional Patent Application No. 63 / 771,104, filed on March 13, 2025, which is incorporated herein by reference in its entirety.STATEMENT OF GOVERNMENT SUPPORT

[0002] The disclosed invention was made with Government support under project number ZIA AG000353-06 by the National Institutes of Health, National Institute on Aging. The United States Government has certain rights in the invention.FIELD

[0003] The present disclosure is related to the field of magnetic resonance imaging (MRI).BACKGROUND

[0004] Unlike computed tomography (CT) and positron emission tomography (PET), which both can provide images expressed in absolute units, magnetic resonance imaging (MRI) has conventionally been deployed as a highly sensitive qualitative modality. Thus, although conventional MRI has been a first line investigation for both radiological-diagnosis and treatmentmonitoring of neurological diseases, historically, many assessments have been based on images presented in arbitrary units.

[0005] In recent years, however, quantitative MRI has been a growing field of investigation for providing less subjective diagnosis. Key quantitative MRI parameters include tissue relaxation times — particularly longitudinal relaxation times (Ti) and transverse relaxation times (T2) that describe how fast the magnetization in the tissue molecules decays — and myelin water fraction (MWF), which is a surrogate of myelin content of a specimen tissue.

[0006] While a variety of innovative MRI approaches to measure relaxation times and MWF have been introduced, thus far, there remains a need for an approach that is capable of practical and widespread integration in MRI systems, particularly those configured for clinical settings.Leydig 775185HHSRef E-085-2024-0-PCT-012

[0007] For example, a standard technique for MWF imaging used in research settings — namely, the three-dimensional (3D) implementation of gradient and spin-echo (GRASE) — is essentially not available for use in clinical platforms. Similarly, the so-called multicomponent driven equilibrium single pulse observation of Ti and T2 (mcDESPOT) method — which is capable of providing a high resolution MWF map image — is essentially unavailable in most clinical settings.

[0008] According, despite the apparent existence of quantitative MRI methodologies, further solutions have been long sought by the MR relaxometry community (MRRC). Accordingly, there is a long-felt, but unmet need for an MRI methodology that is efficiently implementable in clinical practice to accurately measure relaxation times and MWF and output a high-resolution MWF MRI image.SUMMARY

[0009] Aspects of the present disclosure, therefore, provide improved MRI systems and methods that are efficiently implementable in clinical practice, while also enabling accurate measurements of relaxation times and MWF, as well as the generation of high-resolution MWF MRI images.

[0010] For example, as will be elucidated further below, aspects of the present disclosure provide technical improvements over the mcDESPOT method, and enable efficient production of high-resolution MWF map images and accurate measurements of Ti and T2 in clinical MRI systems.

[0011] The methodologies of the present disclosure utilize rapid gradient-echo MRI pulse sequences to create MRI data. Briefly, rapid gradient-echo MRI pulse sequences refer to MRI imaging sequencing techniques where the sequence repetition time (TR) is short compared with the Ti and T2 relaxation times.

[0012] For context, in clinical settings, a TR of less than 10 ms is typically used (depending on the specific scanner and sequence parameters), with the T2 relaxation time typically being greater than 10 ms, and the Ti relaxation time being greater than T2. In the context of myelin imaging, the signal from myelin water is of primary interest, which has been reported to have a T2 relaxation time of approximately 20 ms at 3T, and a Ti relaxation time of less than 500 ms atLeydig 775185HHSRef E-085-2024-0-PCT-0133T. Other tissue compositions, such as gray matter and cerebrospinal fluid, typically have longer Ti and T2 relaxation times than myelin water. While Ti and T2 relaxation times are fielddependent, the specification that TR be shorter than both Ti and T2 is a constraint used in achieving steady-state imaging, regardless of the field strength.

[0013] In general, there are a three categories of rapid gradient-echo MRI pulse sequencing: (1) radio frequency (RF) spoiled; (2) balanced steady-state free precession (SSFP); and (3) unbalanced gradient-spoiled SSFP. RF-spoiled sequences suppress residual transverse magnetization at the end of each sequence repetition through phase spoiling, optimizing contrast for Ti-weighted imaging. In balanced SSFP (bSSFP), all imaging gradients are balanced, meaning they have a net zero area over each repetition, which preserves transverse magnetization and provides high signal for tissues with large T2 / T1 ratios (such as fluids, particularly cerebrospinal fluid). Conversely, in unbalanced gradient-spoiled sequences, like SSFP-FID or SSFP-echo, the slice-select and readout gradients have a net non-zero area, resulting in reduced preservation of transverse magnetization.

[0014] The bSSFP MRI sequence is a component of quantitative MR imaging protocols (such as mcDESPOT), which is especially useful for the quantification of relaxation times, Ti and T2, and MWF.

[0015] Presently, such as with mcDESPOT, in order to use bSSFP MRI sequencing to produce high-quality MWF images, scans must be made at multiple RF phase angles (e.g., at 0 and TI phase angles). This so-called “phase cycling” is done to account for main magnetic field (Bo) inhomogeneity — which is a spatial variation in the main magnetic field that leads to image artifacts.

[0016] The mcDESPOT methodology, for example, uses bSSFP MRI sequencing with phase cycling to produce MWF images. Initially, mcDESPOT attracted great interest in the MRRC. The mcDESPOT methodology was initial perceived as interesting particularly for its capability of providing a rapid MWF map of an entire specimen (e.g., of a whole human brain) with an exquisitely high-resolution. The mcDESPOT methodology was also attractive because of its use of a relatively short echo time (TE), thereby allowing improved detection of the short-T2 component of the signal representing the MWF. Nevertheless, despite its outstanding promises for fast, high-resolution whole-specimen MWF imaging, mcDESPOT was unable to meet the long-felt need for an efficient MWF MR imaging methodology that is both easily implementable inLeydig 775185HHSRef E-085-2024-0-PCT-014clinical practice, while also providing high-resolution images and accurate measurements of relaxation times and MWF.

[0017] One reason for the shortcomings of mcDESPOT is inherent to its use of traditional bSSFP sequencing with the required phase cycling. Experience has shown that phase cycling creates significant technological complexities. For example, the implementation of phase cycling requires substantial specialized expertise to program the MRI sequencing, and also an MRI system that itself is specifically configured such that it is even capable of executing phase cycling. While some MRI systems configured for research are enabled for bSSFP MRI sequencing with phase cycling at multiple angles, most MRI systems implemented in clinical settings are not even enabled for quantitative MRI (qMRI), let alone for applying bSSFP sequencing with phase cycling to determine relaxation times and MWF mapping.

[0018] Aspects of the present disclosure address this technologically-based challenge by providing a methodology that compensates for Bo inhomogeneity without using phase cycling. Instead, embodiments of the present disclosure employ bSSFP MRI sequencing that adjusts the repetition time (TR) to compensate for Bo inhomogeneity. This change simplifies the process, making it more efficient and less complex. For example, application of the disclosed methodology eliminates the need for pulse programming to implement phase cycling and reduces the requirement for a deep understanding of the physics behind phase cycling. Adjusting the TR is a routine parameter change that radiologists and technicians commonly perform, and it is widely available for most pulse sequences from most manufacturers.

[0019] This advancement has far-reaching implications. It greatly enhances the quality and accuracy of bSSFP-based imaging, improving diagnoses and monitoring of various medical conditions, and allowing a truly applicable method for relaxometry and MWF imaging. For example, aspects of the present disclosure, which make it possible to estimate T2 and Bo maps by changing TR, improve quality and accuracy by correcting banding artifacts on bSSFP images. Also, the implications for MWF imaging are significant because MWF imaging has been shown to be useful in multiple sclerosis (primarily demyelinating disease), traumatic brain injury, and neurodegenerative diseases.

[0020] Ultimately, by streamlining the technique and reducing the need for specialized expertise and specially configured MRI systems, the improvements offered by the present disclosure can make high-quality MRI more accessible and efficient for healthcare providers,Leydig 775185HHSRef E-085-2024-0-PCT-015thereby benefiting patient care. Overall, the present disclosure marks a substantial improvement in medical imaging technology, offering the potential to efficiently improve diagnostic capabilities and patient outcomes.

[0021] In a first exemplary embodiment, the present disclosure provides a method for operating a magnetic resonance imaging (MRI) system to produce a myelin water fraction (MWF) map image. Here, a plurality of spoiled gradient recalled echo (SPGR) images of a specimen are acquired with a plurality of different flip angles (FAs). A plurality of balanced steady-state free precession (bSSFP) images of the specimen with a second plurality of different FAs and a plurality of different repetition times (TRs) are also acquired. Using driven equilibrium single pulse observation of longitudinal relaxation times (DESPOTI), a longitudinal relaxation time (Ti) map from the SPGR images is generated. Using the Ti map and the bSSFP images, an MRI analysis using steady-state acquisitions (USA) is applied and an off-resonance frequency (Aco) parameter map and a transverse relaxation time (T2) map are generated. Using Bayesian Monte-Carlo (BMC) analysis, the MWF map image is generated from the SPGR images, a subset of the bSSFP images, and the Aco parameter map.

[0022] According to a first aspect, in any implementation of the first exemplary embodiment, acquiring the SPGR images can include, operating the MRI system to apply and receive MRI signals to and from the specimen using SPGR pulse sequencing using the different FAs with a TR fixed at a first value.

[0023] According to a second aspect, in any implementation of the first exemplary embodiment, acquiring the bSSFP images can include: acquiring first bSSFP images at P different FAs and at a first TR; acquiring second bSSFP images at Q different FAs and at a second TR; and acquiring third bSSFP images at R different FAs and a third TR. The first bSSFP images are the subset of the bSSFP images used in the BMC analysis.

[0024] According to a third aspect, in any implementation of the second aspect of the first exemplary embodiment, acquiring the first bSSFP images includes operating the MRI system to apply and receive MRI signals to and from the specimen using bSSFP pulse sequencing using the P different FAs with a TR fixed at the first TR. Also, acquiring the second bSSFP images includes operating the MRI system to apply and receive MRI signals to and from the specimen using bSSFP pulse sequencing using the Q different FAs with the TR fixed at the second TR, and acquiring the third bSSFP images includes operating the MRI system to apply and receive MRILeydig 775185HHSRef E-085-2024-0-PCT-016signals to and from the specimen using bSSFP pulse sequencing using the R different FAs with the TR fixed at the third TR.

[0025] According to a fourth aspect, in any implementation of the second aspect of the first exemplary embodiment, the first TR is shorter than the second TR, and the second TR is shorter than the third TR.

[0026] According to a fifth aspect, in any implementation of the fourth aspect of the first exemplary embodiment, the second TR is twice the first TR and the third TR is three times the first TR.

[0027] According to a sixth aspect, in any implementation of the second aspect of the first exemplary embodiment, Q and R are 2, and P is between 2 and 10.

[0028] According to a seventh aspect, in any implementation of the first exemplary embodiment, the bSSFP images are acquired without phase cycling.

[0029] According to an eight aspect, in any implementation of the first exemplary embodiment either all of the bSSFP images are acquired with the RF phase set as 7t or all off the bSSFP images are acquired with RF phase set at 0.

[0030] According to a ninth aspect, in any implementation of the first exemplary embodiment, the method further includes outputting at least one of the MWF map image, the Ti map, the T2 map, or the Aco parameter map.

[0031] According to a tenth aspect, in any implementation of the ninth aspect of the first exemplary embodiment, the method can further include providing a quantified measurement of myelin content of the specimen base on the MWF map image.

[0032] According to an eleventh aspect, in any implementation of the ninth aspect of the first exemplary embodiment, the specimen is a human-patient’s brain, and the method further includes treating the patient for a neurological disorder based on the MWF map image.

[0033] According to a second exemplary embodiment, the present disclosure provides a magnetic resonance imaging (MRI) system. The MRI system includes one or more processors configured to control the MRI system to apply MRI signals to a specimen, to receive MRI data based on the applied MRI signals, and to generate MRI images according to a method. The method includes: acquiring spoiled gradient recalled echo (SPGR) images of the specimen with different flip angles (FAs); and acquiring balanced steady-state free precession (bSSFP) images of the specimen with different FAs and different repetition times (TRs). Also, using drivenLeydig 775185HHSRef E-085-2024-0-PCT-017equilibrium single pulse observation of longitudinal relaxation times (DESPOTI), a longitudinal relaxation time (Ti) map is generated from the SPGR images. Using the Ti map and the bSSFP images, an analysis using steady-state acquisitions (USA) is applied and an off-resonance frequency (Aco) parameter map and a transverse relaxation time (T2) map are generated. Using Bayesian Monte-Carlo (BMC) analysis, the MWF map image from the SPGR images, a subset of the bSSFP images, and the Aco parameter map are generated.

[0034] According to a third embodiment of the present disclosure, a non-transitory computer readable medium is provided that includes instructions, which upon execution by one or more processors, is configured to cause the one or more processors to execute a method that includes: acquiring spoiled gradient recalled echo (SPGR) images of the specimen with different flip angles (FAs); acquiring balanced steady-state free precession (bSSFP) images of the specimen with a different FAs and a different repetition times (TRs); using driven equilibrium single pulse observation of longitudinal relaxation times (DESPOTI), generating a longitudinal relaxation time (Ti) map from the SPGR images; using the Ti map and the bSSFP images, applying USA and generating an off-resonance frequency (Aco) parameter map and a transverse relaxation time (T2) map; and using Bayesian Monte-Carlo (BMC) analysis, generating the MWF map image from the SPGR images, a subset of the bSSFP images, and the Aco parameter map.

[0035] All examples and features mentioned above may be combined in any technically possible way. For example, any of the aspects described for the first embodiment may be applied to the second or third embodiments of the present disclosure.BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The present application will be described in even greater detail below based on the exemplary figures. The application is not limited to the examples described below. All features described and / or illustrated herein can be used alone or combined in different combinations in examples of the application. The features and advantages of various examples of the present application will become apparent by reading the following detailed description with reference to the attached drawings which illustrate the following:

[0037] FIG. 1 shows an overview of an exemplary mcDESPOT methodology.Leydig 775185HHSRef E-085-2024-0-PCT-018

[0038] FIG. 2 shows an overview of an exemplary mc-USA methodology implemented according to aspects of the present disclosure.

[0039] FIGS. 3a and 3b show result of simulation analysis for the determination of A« and T2 using USA from bSSFPit data acquired at either two different TRs (FIG. 3A) or three different TRs (FIG. 3B).

[0040] FIGS. 4A and 4B shown the results of simulation analysis for the determination of Aco and T2 using USA from bSSFP data generated at three different TRs.

[0041] FIG. 5 shows T2 and MWF parameter maps derived, respectively, using DESPOT2 and BMC-mcDESPOT, with or without Am correction.

[0042] FIG. 6 shows Am, T2 and MWF parameter maps derived using DESPOT2, USA, BMC-mcDESPOT, and mc-USA, with absolute difference maps also depicted.

[0043] FIG. 7 shows an exemplary magnetic resonance imaging (MRI) system and / or apparatus in accordance with one or more examples of the present disclosure.

[0044] FIG. 8 shows a representative computing and control environment in accordance with one or more examples of the present disclosure.

[0045] FIG. 9 shows an exemplary implementation of a method according to aspects of the present disclosure.DETAILED DESCRIPTION

[0046] Aspects of the present disclosure are directed to improving MRI methodologies and systems, including in relation to streamlining the production of high-quality MWF images. In particular, aspects of the present disclosure are directed to providing multi-component MRI using steady-state acquisitions (mc-USA) for practical and efficient relaxometry and myelin water fraction imaging.

[0047] The mc-USA methodology disclosed herein compensates for Bo inhomogeneity without using phase cycling, and thereby, avoids the shortcomings of the state of the art. Instead, the mc-USA methodology employs bSSFP MRI sequencing that adjusts the repetition time (TR) to compensate for Bo inhomogeneity. For example, different TRs may be used across multiple imaging acquisitions, rather than within a single acquisition. This deceptively-simple, but groundbreaking change effectively streamlines the MWF imaging process, making it usable for MRILeydig 775185HHSRef E-085-2024-0-PCT-019systems that are not enabled with complex phase shifting and / or avoiding the resource-intensive process of defining an MRI sequence that includes phase shifting. Also, as compared to the state of the art, the mc-USA methodology enhances the quality and accuracy of bSSFP-based imaging, thereby improving diagnoses and monitoring of various medical conditions, particularly those that correlate to myelin content. Furthermore, unlike previous approaches, quantitative MRI (qMRI) maps can be calculated from these multiple TR acquisitions, enhancing the utility of the methodology for generating accurate MWF images.

[0048] Myelin, it has been observed, is crucial for efficient signal transmission over long ranges in the nervous system because it increases the speed at which the impulses propagate along the axons. Axons are coated piecewise by multiple layers of phospholipid membranes (‘sheaths’) with embedded proteins produced by oligodendrocytes and Schwann cells in the central and peripheral nervous systems, respectively. Degradation of myelin impairs the signal transmission, and the nerve may eventually wither, leading to brain atrophy and brain dysfunction.

[0049] Knowledge of myelin content also supports the investigation of brain development. Whereas, accurate myelin measurements are valuable in studies of neurodegenerative diseases, such as multiple sclerosis (MS) and dementia. Thus, measurements and monitoring of myelin content would provide information for the diagnosis and prognosis in patients with suspected myelin degradation. For example, evaluation of myelin content is medically important because alterations in myelin content and white matter tissue have been associated with a myriad of nervous system diseases including aging, multiple sclerosis, epilepsy, Parkinson’s disease, Alzheimer's disease, as well as psychotic disorders. However, the etiology of cerebral myelin breakdown and white matter deterioration in all these conditions remains poorly understood. Hence, further investigations, including risk factors correlative studies, with longitudinal assessments, are still requited to elucidate the mechanisms underlying, or implicating, these myelin content-related changes and white matter damage.

[0050] Therefore, there is an urgent need for fast, reliable, and clinically applicable methods to measure surrogate biomarkers of cerebral tissue microstructure and composition, including relaxation times and myelin content. In particular, there is a need for fast, reliable, and clinically applicable methods for measuring a subject tissue’s MWF, which itself is a proxy of myelin content.Leydig 775185HHSRef E-085-2024-0-PCT-0110

[0051] In the context of the present disclosure, “fast” image acquisition refers to the ability to conduct scans within the typical timeframe of a clinical MRI session, which generally lasts about 20 to 60 minutes, including preparation and actual imaging time. The disclosed MWF imaging method is designed to fit within this clinical time window without extending the session duration.

[0052] “Reliability” in MRI means achieving high spatial resolution, appropriate contrast, consistency across different scans, and minimal artifacts. The disclosed MWF imaging method provides high spatial resolution at 1mm x 1mm x 1mm, with excellent contrast to clearly highlight demyelination. It ensures consistency across scans by quantitatively assessing tissue composition rather than relying on relative image intensities, and artifacts are minimized by compensating for Bo and Bi inhomogeneities.

[0053] “Clinically applicable” means that the technique can be used effectively in real-world medical settings. As mentioned, MWF imaging has been shown to be valuable in diagnosing and monitoring conditions like multiple sclerosis, traumatic brain injury, and neurodegenerative diseases, making it highly relevant for clinical use.

[0054] As touched on above, while a variety of MRI approaches have been introduced to evaluate MWF, their integration to clinical practice has been challenging due to significant technologically-driven shortcomings. Such shortcomings include the need for programming a phase cycling scheme into the pulse sequence, which requires extensive programming knowledge that most clinical practitioners do not possess. This complexity has hindered the integration of some advanced MRI techniques into clinical practice. Accordingly, the MR relaxometry community (MRRC) has been in need of truly clinically-integrable methods for MWF imaging. The mc-USA methodology of the present disclosure provides such a solution. For example, the mc-USA methodology is truly clinically-integrable because it only requires adjusting the TR, which is a routine, tunable parameter that clinicians are already familiar with and is widely available for modification in most MRI systems from manufacturers. This eliminates the need for complex programming or specialized knowledge, making it easily integrable into clinical workflows.

[0055] The improvements provided in the mc-USA methodology of the present disclosure may be elucidated by comparison to the existing mcDESPOT mythology.

[0056] FIG. 1 provides an overview of a process 100 implementing the mcDESPOT methodology.Leydig 775185HHSRef E-085-2024-0-PCT-0111

[0057] The mcDESPOT process 100 uses a set of MRI images 110 (acquired at different flip angles FAs and using phase shifting), and implements a set of rapid combined Ti and T2 mapping approaches, referred to as driven equilibrium single pulse observation of T1 (DESPOT1) and driven equilibrium single pulse observation of T2 (DESPOT2).

[0058] As shown in FIG. 1, the mcDESPOT process 100 includes operating an MRI system using spoiled gradient recalled echo (SPGR) imaging sequences to acquire MRI images (or MRI data) of a specimen (e.g., a human brain) at K different flip angles (FAs) 111. SPGR imaging is a type of RF spoiled rapid gradient-echo MRI pulse sequencing, which a person of ordinary skill in the art would readily understand how to implement.

[0059] Additionally, the mcDESPOT process 100 includes operating the MRI system using balanced steady state free precession (bSSFP) imaging sequences to acquire images of the specimen at L different FAs all at a first radio frequency (RF) phase (e.g., 0 radians) 112; and operating the MRI system using bSSFP imaging sequences to acquire images of the specimen at P different FAs at a second RF phase (e.g., 7t radians) 113. As such, mcDESPOT employs bSSFP imaging using phase switching. While bSSFP imaging with phase switching requires extensive technical knowledge to properly execute on specially-enabled MRI system, a skilled-artisan with the requisite advanced knowledge would be able to implement it accordingly.

[0060] The mcDESPOT process 100, then uses DESPOT1 120 and DESPOT2 125 to combine these SPGR-acquired images and bSSFP-acquired images to generate Ti map imagery 131, T2 map imagery 132, and off-resonance frequency map (AcoxTR map) imagery 133.

[0061] In more detail, first DESPOT1 120 is used to generate the T1 map 131 from the set of SPGR-acquired images 111, which were acquired over a range of K FAs at a constant repetition time (TR).

[0062] Then DESPOT2 is used to generate the T2 map 132 and the Aco x TR map 133 from: (1) the generated T1 map 131; (2) the first set of bSSFP-acquired images (bSSFP0) 112, which are those bSSFP images acquired at / . different FAs at the first RF phase (here, 0 radians); and (3) the second set of bSSFP-acquired images (bSSFPπ) 113, which are those bSSFP images acquired at P different FAs at the second RF phase (here, n radians).

[0063] The mcDESPOT process 100 then concludes with the creation of the MWF map image 151. As shown, the mcDESPOT process 100 uses a Bayesian Monte Carlo (BMC) analysis 140 to generate an MWF map image 151 from: the set of MRI images 110, and the Δω × TR map 133.Leydig 775185HHSRef E-085-2024-0-PCT-0112

[0064] To further elucidate the underlying functionality of mcDESPOT, consider the following discussion elaborating upon the bSSFP and SPGR signal models used and the assumptions that are applied in mcDESPOT that underlie mcDESPOT's capability of producing MWF imaging that is medically useful for characterizing myelin in a specimen. Through this discussion, a person of ordinary skill in the art would understand the process for implementing algorithms to capture the necessary MRI data, and perform the indicated DESPOT1, DESPOT2, and BMC operations to generate an MWF map.

[0065] As is known in the MRRC, certain MRI signal modeling for a nervous tissue specimen may assume a bicomponent system — which may correspond to a signal originating from the water trapped within myelin sheaths while a second signal from the intra- and extra-cellular water, as appropriate for white matter, or may assume a tricomponent system, which may incorporate an additional compartment to account for cerebrospinal fluid (CSF) water contamination at the vicinity of the cortical voxels. Under, such multi-component assumptions, multicomponent bSSFP and SPGR signal models have been described in the art.

[0066] For example, under conditions of steady-state for N exchanging species Sn, n e (1,..., N) and using loch-McConnell equations, the bSSFP magnetization, MbsSFP, at echo time TEbSSFP≅ 0.5 * TRbSSFP, where TRbsSFPis the repetition time that corresponds to the time between two successive radiofrequency (RF) pulses, is given by:^bSSFP = e ^bSSFp' ^bSSFp)^^ / J3W_ e^TR.sspp- TRbSSFp)^^ \ ^ATRbssFp- TRbSSFp)- I37v)ATRbssFPC+ (e (ATEbSSFPTEbSSFp) _ ATFbssFpC, [EQ 1]where I3jVis the 3A-by-3A identity matrix, the 3N-vector = [MxMyMz] with Mx= \ LM x™pSiFP... M% X>SS^SNFP] J,5M yv= \ LM^SSSFP... M y^,SS^SNFP] J and Mzz= \ LM%SSSFP... M%SSSFP] J, •> C =T / L2<bt 0 \ ^O[^ / T,2, S1— ® / T2, SN0 / ^2, Sj ■■■ 0 / T2 Snfs1 / T1s1■■■ fsN / ?i,sN\, At= — (|)tL20 I, with:\ 0 0 LjLeydig 775185HHSRef E-085-2024-0-PCT-0113ks?sAksNs, \ksNs2andks2sNT- c ^Tn^NkSNSmj ‘■SN ' 7, JO o o o " o &2(0 0 0 0 0 0, 0 0 0 *„('), with R(a) representing the rotation matrix of the RF pulse for flip angle (FA) a, taken about the / Iw0 ° \x-axis, so that R(<z) = I 0 cos(a) ■ Iwsin(a) ■ IwI where INis an A-by-A identity matrix.\ 0 — sin(a) ■ Iwcos(cr) ■ IwJ

[0067] Further, Morepresents the signal amplitude at TE = 0 ms and incorporates proton density and various machine factors, fSnis the fraction of the nthspecies so that with Σn=1fSn= 1, T2and T1are the transverse and longitudinal relaxation times, respectively, for a species as indicated by appropriate subscripting, kSnSmis the exchange rate from species Sn to species Sm with the steady-state assumption Σm≠nfSnkSnSm=fsmksmsn- Here= (θRF+ φSn) / t is the total phase offset, with θRFis the phase increment (i.e., phase cycle) of the RF excitation pulse andφSn= 2π · TRbSSFP· ΔωSnis the phase accumulation arising from the static field inhomogeneity and chemical shift (leading to off-resonance frequency inhomogeneities), ΔωSn, of the nthspecies, and a is the flip angle (FA) of the RF excitation pulse. Then the theoretical multicomponent bSSFP signal is given by:N NSbSSFP= |Σ MbSSFPx,Sn+ i Σ MbSSFPy,Sn| [EQ 2]+ 1 Z?iFFn=l n=l

[0068] Likewise, under the same conditions, the steady-state SPGR magnetization, MSPGR, at echo time TESPGR, is given by:Leydig 775185HHSRef E-085-2024-0-PCT-0114MSPGR= e^AspGR'TEspcR^R(,a') S (l3N- Rfa) e(AspcR'TRspcR)S)”1S ^e(AspcR'TRspcR)~ kN)AsPGR C) + S (e(ASPGfi™SPGR) _ I3W)A-IGRC, [EQ 3] / L20 0 \ / VoVo Vo\ where ASPGR= (L₂, 0, 0; 0, L₂, 0; 0, 0, L₁) and S = (V₀, V₀, V₀; V₀, V₀, V₀; V₀, V₀, IN), with V₀ an N-by-N null matrix.\ 0 0 k / \V0Vo Iw /

[0069] The theoretical multi-component SPGR signal is then given by:N N$SPGR £ MX8+ i [EQ 4]n=l n=l

[0070] For cerebral myelin imaging, a system of two water compartments (N= 2), consisting of a short and long Ti and T2 components, is generally adopted. This leads to reduced SPGR and bSSFP signal equations.

[0071] The short component corresponds to the signal of water trapped within the myelin sheets while the long component corresponds to intra / extra (I / E) cellular water. If exchange between water pools is neglected, this bicomponent model leads to a vector, X, of five unknown parameters given by: λ = [MWF T2,MWFT2,I / ET1,MWFT1,I / E], where MWF is the myelin water fraction with corresponding transverse and longitudinal relaxation times (T2,MWF, T1,MWFThe fraction of the I / E cellular water is defined as I / EF=(1-MWF) with corresponding transverse and longitudinal relaxation times (T2,I / E, T1,I / E).

[0072] These bSSFP and SPGR signal models may be further reduced by assuming a monocomponent system. In particular, assuming a single water compartment (N= 1), Equation 2 is reduced to:SbssFP,mono = \MO(M»SSEP+ iM»SSEP)\, [EQ 5]where:MbssFP=_ sincr sin y 7^(1 ~ k) _x(1 — Excos cr)(l — E2COS <p) — E2(E1— cos a)(k— cos<p)Leydig 775185HHSRef E-085-2024-0-PCT-0115and:MbssFP=_ sing V^(l - £1Xcos<p - £2) _y(1 — E1cosa)(l — E2cos<p)—E2(. E1— cos a)(E2— cos<p)

[0073] Similarly, Equation 4 is reduced to:sin a (1 — E.)^SPGR.mono — Mo~ j p ’ [EQ 6]1 — h1cos a

[0074] For Equations 5 and 6, the precession angle (p = 0RF+ 2TT ■ TRbsSFP■ Am, E1= exp(−TR / T1), and E2= exp(−TE / T2). Where, 0RFis the phase increment of the RF pulse, Aco is the off-resonance frequency with respect to the central water peak, and Mois a factor proportional to the equilibrium longitudinal magnetization. Also, as mentioned above, a is the flip angle (FA). In mcDESPOT,a B1⁺ map is used to account for potential deviation of the actual FFA, αact= B₁⁺α, from the nominal FA, α.

[0075] A B1⁺ map can be obtained from (additional) MRI scans acquiring B1. For example, a map for B1 (i.e., excitation radiofrequency inhomogeneity) can be acquired using the doubleangle method (DAM) by acquiring two fast spin-echo images with FAs of 45° and 90°, TE of 102 ms, TR of 3000 ms, and acquisition voxel size of 2.6 mm x 2.6 mm x 4 mm.

[0076] These simplified signal models can be used to derive the apparent longitudinal relaxation times Ti and transverse relaxation times T2, as well as the off-resonance frequency inhomogeneities Δω, which represents averaged values across water compartments within each voxel. Moreover, because the relaxation times Ti and T2 depend on water mobility as well as macromolecular tissue composition such as lipids and iron, they are very sensitive to capture differences in white matter integrity.

[0077] The mcDESPOT methodology employs the above-described signaling models in DESPOTi and DESPOT2, to generate longitudinal relaxation times Ti map imagery, transverse relaxation times T2 map imagery, and off-resonance frequency Δω map imagery.

[0078] In particular, DESPOT1 models T1 from SPGR data obtained at different FAs using Equation 6, while fixing the TR to a short value (e.g. less than 5 ms) for rapid scanning and incrementally increasing the FA. For example, TR may be fixed at the shortest possible valueLeydig 775185HHSRef E-085-2024-0-PCT-0116that the MRI system being employed is capable of using (considering the other conditions of the MRI sequencing). The minimum TR is constrained by the scanner's performance, often allowing for values as low as 1-2 ms in high-performance MRI systems. These short TRs are used to enable very fast image acquisition times, particularly in rapid gradient-echo sequences like SPGR. While there is no strict maximum TR value, it is uncommon to use long TRs in SPGR sequences, as the technique is designed for speed and efficiency with shorter TRs (i.e., typically below 5 ms). The FAs chosen may be based on the particular experimental conditions. Typically, numerical simulations are used to determine the most suitable FAs based on specific scanner parameters, tissue properties, and imaging goals, ensuring optimal signal-to-noise ratio and contrast. For example, FAs may be [2, 4, 6, 8, 10, 12, 14, 16, 18, 20] degrees, which can represent a set of FAs optimized for experimental conditions at 3T.

[0079] By holding TR constant and incrementally increasing the FA, a curve characterized by Ti is generated. The resulting data can be represented in the linear form (i.e., Y = mX + b) as:$SPGR= E1^ + MO(1 -E1), [EQ 6a]sin a tan a

[0080] Because the data may be presented in this linear representation, its slope (m) and the Y-intercept (b) can be estimated by linear regression. This means that Ti and Mo can be extracted as:-TRT = [EQ 6b]In (m)bM° = (1 — m)EQ 6c

[0081] In this way, therefore DESPOT1 can generate an image of the longitudinal relaxation times T1 map using SPGR images acquired with a short, fixed TR and at a plurality of FAs.

[0082] DESPOT2 uses Equation 5 and bSSFP data acquired at different FAs and a short fixed TR to model T1, T2, and Δω. For example, DESPOT2 may fix TR to the shortest possible value available in a particular MRI system for rapid scanning. It should be noted that DESPOT2Leydig 775185HHSRef E-085-2024-0-PCT-0117requires acquisition of the bSSFP images with, at least, two phase increments / cycling of the RF excitation pulse to correct for potential off-resonance frequency inhomogeneities Aco, in order to mitigate consequent spatial banding artifacts. The FAs use are different from FAs from SPGR. For example, FAs such as [2, 4, 7, 11, 16, 24, 32, 40, 50, 60] degrees are optimized for bSSFP acquisitions under experimental conditions at 3T. Nevertheless, it us well-understood in the art to use numerical simulations to determine the most suitable FAs based on specific imaging conditions, such as field strength, tissue properties, and hardware constraints, etc.

[0083] In particular implementation, using the expressions associated with Equation 5 mentioned above, DESPOT performs a three-parameter nonlinear fit for Mo, Aco x TR, and T? from sufficiently sampled bSSFP data. Because the location, extent, and severity of the banding artifact is influenced by 0RFAco, and a, data should be sampled over a range of RF phase increments and flip angles to ensure robust fitting.

[0084] For example, to perform the voxel-wise fitting in one implementation of DESPOT?, all bSSFP data is first normalized with respect to the mean value in order to eliminate the need to explicitly fit for Mo. Next, a stochastic region contraction technique is used. Briefly, to begin, Ni estimates of Aco x TR and T2 are randomly drawn from uniform distributions spanning 0 < Aco x TR < 2TT and 0 ms < T2 < 500 ms. For each combination, theoretical bSSFP data are generated and the sum-of-squares residuals between the theoretical signal points and normalized acquired bSSFP data are calculated. The Ni estimates are rank-sorted by increasing residual and the top N2 estimates are chosen. From these, the minimum and maximum AcoxTR and T2 estimates are determined. These values redefine the extent of the uniform AcoxTR and T2 distributions. The process then repeats, with Ni estimates drawn from these new distributions, etc. Convergence is achieved when the difference between the minimum and maximum of the N2 T2 estimates is less than a threshold value (e.g., 0.5%). Due to the periodic and repeating nature of sin(θRF+ Δω × TR) and cos(θRF+ Δω × TR), i.e., sin(θ) = sin(2π) and cos(θ) = cos(2π), Δω × TR need only be investigated between the limits of 0 and 2π. As the exact, absolute value of ΔωTR is not necessary, but only its relative value, the minimum of the N2 Δω × TR estimates can be chosen.

[0085] Accordingly, DESPOT2 can estimate T2 and Δω × TR from bSSFP data obtained at different FAs using Equation 5, with TR fixed to a short value (e.g., the shortest possible value for the MRI system for rapid scanning). It should be specifically noted as well that DESPOT2 requires acquisition of the bSSFP images with, at least, two phase increments / cycling of the RFLeydig 775185HHSRef E-085-2024-0-PCT-0118excitation pulse to correct for potential off-resonance frequency inhomogeneities in order to mitigate consequent spatial banding artifacts.

[0086] At this point in the mcDESPOT process, SPGR, bSSFP0, bSSFPπ images have been acquired at different FAs and with short TRs, and then with the use of DESPOTi and DESPOT2, maps for T1, T2, and Δω have been determined. As shown in FIG. 1, and indicated above, the next step is to generate the MWF map image from this data.

[0087] This next step of the mcDESPOT process proceeds under the assumption that ΔωMWF= ΔωI / E= Δω. The mcDESPOT process uses Bayesian Monte Carlo (BMC) formalism for MWF determination from the acquired MRI data and generated parameter maps. BMC has shown improved accuracy and precision in MWF determination as compared to the conventional stochastic region contraction algorithm.

[0088] Briefly, parameter estimate MWF, of MWF, belonging to the set of unknown parameters λ = [MWF T2,MWFT2,I / ET1,MWFT1,I / E] is given by:MWF̂ = ∫ MWF P(MWF | S) dMWF = ∫∫∫ MWF [P(λ*)L(S | λ) / P(S)] dλJ J JM1 V P(Xn)L(S I Kn)" MM / MWF™ r(5) - [^ 7]m=lwhere P(MWF | S) is the posterior distribution of MWF given S = (SSPGR, SbSSFP,n), representing the vector of measured signals, each of which is itself a vector describing the signal amplitude for each value of FA, m denotes one of a total of AT random sets of parameter combinations sampled from a grid defined over the integration parameter ranges and defines our Monte Carlo integration, λ* is equivalent to λ but excluding MWF being estimated and represents the vector of nuisance parameters in the determination of MWF, P(λ*) is the product of prior parameter distributions for the elements of λ* taken as noninformative Jeffreys priors, P(S) = ∫ P(λ*) L(S | λ) dλ* is a normalization constant, and L(S | λ) is the marginalized joint likelihood function of S given λ (25, 33).

[0089] For mcDESPOT, L(S | A) is given by:Leydig 775185HHSRef E-085-2024-0-PCT-0119K( ~ _ ~ _ j’X 2 / ~L(S | λ) = C × (SSPGR− MSPGR)(SSPGR− MSPGR) × (SbSSFP0_L− MbSSFP0)(SbSSFP0− MbSSFP0) × (SbSSFPπ_p~ _ y \ 2− MbSSFPπ)(SbSSFPπ− MbSSFPπ) [EQ 8]where C is a constant that depends only on K, L, and P which represent the total number of images acquired at different FAs for SPGR, bSSFP0 and bSSFPπ, respectively. S and M are respectively the experimental and theoretical signals (Equations 2 and 4) normalized by their respective mean values calculated over FAs.

[0090] Referring back to Equation 3 (the signal model for SPGR) and Equation 5 (the signal model for bSSFP), it can be seen that T1, T2, Δω, and α are all needed to calculate theoretical signals for images. Turning then to Equation 7, L(S | λ) is needed to calculate MWF. L(S | λ) is calculated by real signals and theoretical signals, which needs, T2, Δω, and α, etc.

[0091] The mcDESPOT methodology just described therefore is technically capable of providing an MWF map image of an entire specimen (e.g., the whole brain) with a high-resolution.

[0092] However, as indicated already, mcDESPOT has substantial technologically-induced shortcomings, particularly in relation to its use of traditional bSSFP imaging. As described above bSSFP imaging has notorious challenges with banding artifacts that originate from off-resonance frequency inhomogeneities, which if unaddressed leads to inaccuracies in T2 and MWF determinations. The mcDESPOT methodology accounts for these off-resonance frequency issues by acquiring bSSFP images with at least two different radiofrequency (RF) excitation pulse phase increments (a.k.a. phase cycling). This process of phase incrementation of the bSSFP sequence is experimentally challenging to achieve as it requires expertise in sequence development and MR physics to implement, especially in the type of MRI systems deployed in clinical settings. In practice, therefore, it has proven too impractical to implement mcDESPOT widely in a clinical setting due to the need to perform phase cycling. This has created long-felt issues because it hampers wide-spread integration of MWF imaging in clinical practice and research.Leydig 775185HHSRef E-085-2024-0-PCT-0120

[0093] To address the failures of mcDESPOT (and MWF imaging generally), the present disclosure introduces an improvement to the field of MWF imaging, which the present inventors have named: multi-component analysis using steady-state acquisitions (mc-USA). Unlike mcDESPOT, mc-USA does not require incrementation of the RF phase (i.e., no phase cycling), but instead estimates the off-resonance frequency inhomogeneities from bSSFP imaging data obtained at different repetition times (TRs), while keeping the RF phase the same (e.g., at the default value of the MRI system). As the present inventors have found, this change results in eliminates technical limitations associated with MRI systems needing to be enabled for phase cyling, thereby making it significantly easier to perform MWF imaging in clinical practice and research. Moreover, the mc-USA methodology simplifies the MRI process without any practical loss in image quality or an increase in scan time.

[0094] FIG. 2 provides an overview of an exemplary processes 200 implementing the mc-USA methodology.

[0095] The mc-USA process 200 uses a set of MRI images (acquired at different flip angles FAs and repetition times TRs, but without using phase shifting). The mc-USA process 200 also implements a rapid Ti mapping approach, referred to as driven equilibrium single pulse observation of Ti (DESPOTi), and a single phase bSSFP analysis (USA) to map T2 and Am.

[0096] As shown in FIG. 2, the mc-USA process 200 includes operating an MRI system using spoiled gradient recalled echo (SPGR) imaging sequences to acquire images (or MRI data) of a specimen (e.g., a human brain) at K different flip angles (FAs) 211. For example, K may be > 2, and in a particular embodiment K may be 10.

[0097] Additionally, the mc-USA process 200 includes: operating the MRI system using balanced steady state free precession (bSSFP) imaging sequences to acquire images of the specimen at: (212) P different FAs, all at a first radio frequency (RF) phase (e.g., the system-default π radians), and a first repetition time TRi; (213) Q different FAs, all at the same, first RF phase, and a second repetition time TR2; and R different FAs, all at the first RF phase, and a third repetition time TR3. Depending on the implemented embodiment, the number of P, Q, and R FAs may be different, as well as the underlying values of the FAs. For example, Q and R may be 2, while P may be 2 or more (e.g., 5 or 10).

[0098] In a particularly preferred embodiment, the RF phase is held constant at the default of the MRI machine, which is typically π radians. Regardless of the actual value used, a feature ofLeydig 775185HHSRef E-085-2024-0-PCT-0121mc-USA is that all bSSFP images are captured without phase switching. This provides a technological advantage because, in many MRI systems deployed in the clinical setting, the MRI systems implemented to not allow for the user to change the RF phase away from the default, thus making mc-USA easier to implement across more systems.

[0099] The mc-USA process 200, then uses DESPOTi 220 and USA 225 to combine these SPGR-acquired images and bSSFP-acquired images to generate: Ti map imagery 231, T2 map imagery 232, and off-resonance frequency map (AOJ x TR map) imagery 233.

[0100] In more detail, first DESPOTi 220 is used to generate the Ti map 231 from the series of SPGR-acquired images 211, which were acquired over a range of K FAs at a constant repetition time (TR). In a preferred embodiment, the mc-USA process 200 executes DESPOTi in accordance with how it is executed in mcDESPOT using Equation 6. This includes fixing the TR to a short time (relative to the limitations of the MRI machine being used). In a preferred embodiment, TR is set to the shortest possible value (typically < 10 ms) as allowed by the particular MRI machine for rapid scanning, e.g. 6 ms between successive pulse-sequences applied to the same slice.

[0101] In principle, with the presence of a Bp map to account for the deviation of the actual FA, aact= Bi a, from the nominal FA, a, only two SPGR images obtained at different FAs would be necessary for Ti determination. However, including more images can lead to higher accuracy and precision.

[0102] Next, the mc-USA process 200 executes USA 225 to generate T2 map imagery 232, and off-resonance frequency map (Aco x TR map) imagery 233 from the bSSFP images 212-214 captured with multiple FAs and TRs, but all with the same RF phase.

[0103] The present inventors have developed USA 225 on an insight that Bo inhomogeneity correction can be achieved for the bSSFP signal model (see Equation 5) by incrementing the TRbssFP instead of 9RF(i.e., 9RFis constant, preferably 9RF= 7r). This, in principle, will allow determination of Am since, as described above, the total phase offset is dependent of TRbsSFPthrough (p = 9RF-I- 2n ■ TRbsSFP■ Am.

[0104] The same process that was described above to derive the T2 and- Am maps may be used here. In this implementation, the phase offset (p is the variable that is modulated. The phase offset <p is defined as = 9RF+ 2TT ■ TRbsSFP■ Am. So instead of changing 9RFto modulate < >,Leydig 775185HHSRef E-085-2024-0-PCT-0122mc-USA changes TRbsSFPto achieve similar modulation. Common methods such as nonlinear least squares (NLLS), stochastic region contraction (SRC), and Bayesian analysis are typically used to process the underlying data, fit the model, and generate the T2 and Am maps from the fitted parameters.

[0105] For USA, with the presence of a B map, only two bSSFPjt images obtained at different FAs each at two different TRs (TR2 and TR3) would be necessary for T2 and Am determinations. Including more images is within the scope of the present disclosure, and may lead to higher accuracy and precision of the generated maps.

[0106] The mc-USA process 100 then concludes with the creation of the MWF map image 251.

[0107] As shown, the mc-USA process 200 uses a Bayesian Monte Carlo (BMC) analysis 240 to generate an MWF map 251 from: a sub set of MRI images; and the Acn x TR map 233. The subset of MRI images includes, in a preferred embodiment, the SPGR images acquired with A FAs 211 and the bSSFP images acquired at P FAs using the first repetition time TRi 212. The repetition time used for the SPGR-acquired images 211 may be the same as for the first set of bSSFP-acquired images 212, but is not required to be.

[0108] In more detail, applying Equation 7 for MWF determination, mc-USA makes use of the SPGR and bSSFPn images acquired at different FAs (with the later acquired also at different TRs), in combination with the Am map derived using USA, that is used as an input parameter under the assumption that AmMM / F= Am; / £= Am.

[0109] In mc-USA, a B* map is used to account for potential deviation of the actual FA, cract= Bi a, from the nominal FA, a. In the above fitting process, aactis the actual numerical FA value in the equation. Effectively, it is the nominal FA a (which is defined by the machine) that is corrected by B^.

[0110] Like mcDESPTO, mc-USA preferably uses a Bayesian Monte Carlo (BMC) analysis for determining MWF because BMC provides good accuracy and precision. However, other analyses, such as the conventional stochastic region contraction algorithm for determining MWF could also be used.Leydig 775185HHSRef E-085-2024-0-PCT-0123

[0111] The mc-USA methodology also makes use of the parameter estimate MWF, of MWF, modeled as belonging to the set of unknown parameters A = [MWF T2 MWF 2, I / ET'I. I / E is given by Equation 7 (repeated below for ease of understanding).> f f f P(A*)L(S I A) MWF = MWF P(MWF | S) dMWF = ••• MWF dXJ J JM1 V P(A^) US I Kn) ^ ZMWF- P(S) <[EQ 7]m=l

[0112] In mc-USA — in contrast to mcDESPOT (see Equation 8 above), L(S | A) is given by:KL(S | A) = C * I (SSPGR— MSPCR)(SSPGR— MSPGR) ) I (SbsSFPii_p_ ~ _ «p\ 2—MZjR5PPii)(S& SRFPit— J, [EQ 9]where C is a constant that depends only on K, L, and P which represent the total number of images acquired at different FAs for SPGR and bSSFPu, respectively. S and M are respectively the experimental and theoretical signals (Eqs. 2 and 4) normalized by their respective mean values calculated over FAs.

[0113] Note that bSSFP images obtained with the longest TRs are not necessary for the estimation of MWF using mc-USA and are only used to measure Am. Because Am is a field that changes smoothly spatially, these images can be acquired with limited number of FAs (e.g., 2) or lower spatial resolution (comparatively among the machine’s settings) to accelerate the total acquisition time.

[0114] With the MWF estimation complete, MWF map images are then generated. For example, through BMC analysis, unknown parameters can be estimated: A =[MWF T2 MWFTZJ / E Because the estimation is done voxel-wise, an MWF value will be obtained for each voxel, thus collectively providing the MWF map.Leydig 775185HHSRef E-085-2024-0-PCT-0124

[0115] As a result, mc-USA provides high quality Ti, T2, and MWF map images, which can then be used in in quantitative MRI diagnosis and treatment.

[0116] Extensive simulations analyses, as well as in-vivo data, have demonstrated and confirmed the feasibility and accuracy of mc-USA for relaxation times and MWF imaging within clinically acceptable acquisition times (e.g., on the order of 45 minutes). Additionally, mc-USA uniquely reduces the technical burden for implementation and opens the way to facilitate relaxation times and MWF imaging in clinical and research investigations. For example, mc-UAS eliminates the phase shifting process necessary for mc-DESPOT (and traditional bSSFP implementation), which has proven to be experimentally challenging to achieve for most users because it requires a relatively rare expertise in sequence development and MR physics to implement it, especially in clinical MRI machines. Indeed, the bSSFP implementation in typical clinical scanners adopts a phase increment of dRF— n to enhance signal-to-noise ratio (SNR) in the acquired bSSFP images. This phase increment is not generally changeable.

[0117] Methods

[0118] As explained above, T2 and MWF determinations require an accurate estimation of Am. Therefore, it is instructive to provide a framework for determining the minimal number of different repetition times (TRbssFp) as well as their optimal combination values for accurate Am determination using USA.

[0119] In principle, only two sets of bSSFPjt images acquired at different value of TRBSSFP would suffice for the estimation of Amusing USA. However, given the entanglement of TRbssrp and Am defining the phase accumulation, cp, it is likely that more than two TRBSSFP would be improve the accuracy of the determination of Am, beyond what is generally required in clinical and research settings. Therefore, the accuracy of the Am determination has been explored for the case of three different values of TRBSSFP.

[0120] In these analyses, bSSFP / c signals, assuming a monocomponent water compartment with representative values for white matter of T2 of 60 ms and Ti of 750 ms, were generated at FAs = [2 4 7 11 16 24 32 40 50 60]° and at different TRs comprised between 6 ms and 18 ms, bSSFPit signals were generated with phase accumulation, <p = 2n ■ TRbsSFP- Am, comprised between 0 and n with an increment of 0.01745 rad (i.e., 1°) covering the whole possible range. Of note, all these values are representative of the in-vivo acquisition as well as previously derivedLeydig 775185HHSRef E-085-2024-0-PCT-0125physiological values. Gaussian noise was then added to achieve a representative experimental SNR value of 750 defined as M0 / cr, where a is the standard deviation of the noise, as readily achieved in standard clinical mcDESPOT setting at 3T. For each parameters’ combination, USA was used to estimate A. While the focus is on Am determination, results of T2 estimation are also presented. This analysis and its results are discussed further below in connection with FIGS. 3A and 3B.

[0121] The analysis have shown that at least 3 bSSFPjt datasets obtained with different TRs will achieve accurate determination of Am using USA. For sake of reducing the total acquisition time, it was also investigated whether accurate determination of Am can be achieved when the number of the bSSFP^ datasets obtained at different FAs is limited to 2 for the longest TRs. Specifically, the bSSFPjt dataset obtained at the shortest TR of 6 ms was kept unchanged corresponding to FAs = [2 4 7 11 16 24 3240 50 60]°, while the bSSFPjt datasets of the longest TRs were fixed to FAs =

[0750] °. Of note, the choice of these two FAs provides the most accurate results based on an extensive numerical simulation all possible FAs combinations. Analysis was conducted for different input T2 and Ti values. All other experimental parameters were kept identical to the previous analysis. This analysis and its results are discussed further below in connection with FIGS. 4A and 4B.

[0122] In-vivo analyses

[0123] MRI scans were performed on a 3T whole-body Philips MRI system (Ingenia, Best, The Netherlands) using the internal quadrature body coil for transmission and a 32-channel phased-array head coil for signal acquisition. Under the approval of the Institutional Review Board, the following three sequences were acquired:• 3D SPGR images acquired with FAs of [2 4 6 8 10 12 14 16 18 20]°, echo time (TE) of 1.37 ms, and repetition time (TR) of 5 ms.• 3D bSSFP images acquired with FAs of [247 11 16243240 50 60]°, and TR of 6, 12, or 18 ms. These bSSFP images were acquired with RF excitation pulse phase cycling of it (bSSFPjt). TE was set to TR / 2.• 3D bSSFP images acquired with FAs of [247 11 16243240 50 60]°, and TR of 6 ms. These bSSFP images were acquired with RF excitation pulse phase cycling of 0 (bSSFPo). TE was set to TR / 2.Leydig 775185HHSRef E-085-2024-0-PCT-0126

[0124] All SPGR and bSSFP images were acquired with an acquisition matrix of 150 * 130 x 94, and a voxel size of 1.6 mm *1.6 mm x 1.6 mm. To correct for RF field inhomogeneity, B, the double-angle method (DAM) was used by acquiring two fast spin-echo images with FAs of 45° and 90°, TE of 102 ms, TR of 3000 ms, and acquisition voxel size of 2.6 mm x 2.6 mm x 4 mm (35, 36). All images were acquired with a field-of-view of 240 mm x 208 mm x 150 mm, and reconstructed to 1 mm xl mm * 1 mm.

[0125] Data Analysis

[0126] To demonstrate the importance of Aco correction in T2 and MWF mapping, the later were derived using DESPOT2 and BMC-mcDESPOT without accounting for Aco by using the SPGR and bSSFPji datasets acquired at different FAs and short TRs while assuming that Aco=O. Results were compared to those derived using SPGR, bSSFP7t and bSSFPO accounting for Aco. The later were considered as the reference maps. For each parameter, absolute difference (AE) maps were calculated.

[0127] Secondly, Aco, T2 and MWF maps were derived using the USA or mc-USA from the SPGR and bSSFPn: acquired at TRs of 6 ms, 12 ms and 18 ms and visually compared to the reference maps. AE maps were calculated and Bland-Altman analysis was also performed to quantitatively examine the agreement between regional Aco, T2 and MWF values, calculated in the whole brain white matter and major white matter lobes, derived from each method (DESPOT2 vs. USA, and BMC-mcDESPOT vs. mc-USA).

[0128] Thirdly, to accelerate the total acquisition time, Aco, T2 and MWF maps were derived using the same USA and mc-USA analyses, but limiting the bSSFPjt images acquired at the longest TRs of 12 ms and 18 ms to FAs =

[0750] °. Derived maps were visually compared to the maps derived using the full set of FAs. Bland-Altman analysis was also performed to examine the agreement between regional Aco, T2 and MWF values, calculated in the whole brain white matter and major white matter lobes, derived from the limited vs. the full FA datasets.

[0129] Results

[0130] FIGS. 3A and 3B show result of simulation analysis for the determination of Aco and T using USA from bSSFP^ data acquired at either two different TRs (FIG. 3A) or three different TRs (FIG. 3B). In both of FIGS. 3A and 3B, the left columns show the bSSFPjt signal as a function of Aco for each FA and TR. The middle columns show the estimated Am value from the bSSFPjcLeydig 775185HHSRef E-085-2024-0-PCT-0127signals for each input value of Am. The right columns show the estimated T2 value from the bSSFPic signals for each input value of Am.

[0131] For the left panel, four different scenarios were evaluated corresponding to TR1 / TR2 of 6 / 8 ms, 6 / 10 ms, 6 / 12 ms, or 6 / 18 ms. Similarly, for the right panel, four different scenarios were evaluated corresponding to TR1 / TR2 / TR3 of 6 / 7 / 8 ms, 6 / 8 / 10 ms, 6 / 10 / 12 ms, or 6 / 12 / 18 ms.

[0132] It is readily seen that derived Am and T2 values from bSSFPn: signals generated with two TRs deviate substantially from the input / true values, especially for any input Am value that leads to similar bSSPPn signal trends as a function of FAs despite differences in TRs. In this case, the nonlinear least-squares (NLLS) algorithm suffers from the limited signal dynamic to find the global minima as it becomes prone to local minima due to the sloppiness of the model. This issue is mitigated through use of bSSFIA datasets obtained at three different TRs. In this case, for each input / true Am value, there is at least two bSSFPjt signals that exhibit substantial differences leading to a sufficient dynamic and consequent stability of the NNLS algorithm to find the global solution. As expected, this stability increases with increasing the differences between the TRs, with ATR > 4 ms providing the most accurate estimates of Am and T2. However, incorporation of three TRs increases the total acquisition time. In an embodiment, TR2 = 2*TRi, and TR3 = 3*TRi.

[0133] FIGS. 4A and 4B shown the results of simulation analysis for the determination of Aco and T2 using USA from bSSFPjr data generated at three different TRs of 6 ms, 12 ms, and 18 ms for different input / true Ti and T2 combinations.

[0134] For FIGS. 4A and 4B, the bSSFP7t signals at TR = 6 ms were generated with FAs = [247 11 16243240 5060]°, while bSSFPji signals at TRs = 12 ms and 18 ms were generated with FAs =

[0750] °. The left columns show estimated Aco value from the bSSFPrt signals for each input / true value of Aco. The right columns show estimated T2 value from the bSSFP7t signals for each input / true value of Aco.

[0135] The results shown in FIGS. 4A and 4B demonstrate that limiting the bSSFPjt datasets to only two FAs for the largest TRs provides accurate Am and T2 determinations for all T2 and Ti input / true parameter combinations evaluated. Employing the strategy associated with this analysis will help maintain the total acquisition time within clinically acceptable duration for Ti, T2 and MWF determinations using, respectively, DESPOTi, USA and mc-USA.

[0136] FIG. 5 shows T2 and MWF parameter maps derived, respectively, using DESPOT2 and BMC-mcDESPOT with or without Am correction. T2 was derived using DESPOT2, and MWF wasLeydig 775185HHSRef E-085-2024-0-PCT-0128derived using BMC-mcDESPOT. The derivation was repeated twice with and without Aco correction. The absolute difference was comparing the effect of Aco correction on derived T2 and MWF maps.

[0137] Results are shown for three representative slices. Visual inspection shows that estimated T2 and MWF parameter values without Am correction deviate substantially from the reference values, that is, the values derived with Am correction. This deviation is readily seen in regions with strong off-resonance effects (i.e., high Am values). The absolute difference (AD) maps calculated between T2 or MWF parameter maps with and without Am correction provide a quantitative assessment of the extent of these regional deviations in estimated parameters supporting visual inspection.

[0138] FIG. 6 shows Am, T2 and MWF parameter maps derived using DESPOT2, USA, BMC-mcDESPOT, and mc-USA. Results are shown for three representative slices. For the images under USA / mc-USA, Aco (row 1) and T2 (row2) maps were derived using USA, and MWF maps (row3) were derived using mc-USA. For the images under DESPOT2 / BMC-mcDESPOT, Aco (row 1) and T2 (row2) maps were derived using DESPOT2, and MWF maps (row3) were derived using BMC-mcDESPOT. The absolute difference maps compare all images correspondingly.

[0139] Visual inspection shows that derived Am maps using USA or DESPOT2 exhibit similar regional patterns, with corresponding AE maps indicating minimal differences. Similarly, derived T2 maps using DESPOT2 or USA provides virtually similar estimated T2 values. Corresponding AE maps indicate that perfect agreement between the two methods is achieved especially in white matter regions while differences are seen only at the CSF regions. Finally, visual inspection of derived MWF maps using mc-USA or BMC-mcDESPOT are in close agreement with differences seen in at vicinity of CSF regions as indicated by the corresponding AE maps.

[0140] FIG. 7 shows an exemplary magnetic resonance imaging (MRI) system and / or apparatus in accordance with one or more examples of the present application. The system and / or apparatus 700 includes a controller / interface 702 that may be configured to apply selected magnetic fields, such as constant or pulsed field gradients, to a subject 730 or other specimen. An axial magnet controller 704 is in communication with an axial magnet 706 that is generally configured to produce a substantially constant magnetic field Bo. A gradient controller 708 is configured to apply a constant or time-varying magnetic field gradient in one or more selectedLeydig 775185HHSRef E-085-2024-0-PCT-0129directions or in a set of directions using magnet coils 710-712 to produce respective magnetic field gradient vector components Gx, Gy, Gz or combinations thereof to produce the gradients Gi, G2, G3 that are associated with b-matrices and / or tensors. A radiofrequency (RF) generator 714 is configured to deliver one or more RF pulses to a specimen using a transmitter coil 715. An RF receiver 716 is in communication with a receiver coil 718 and is configured to detect or measure net magnetization of spins. Typically, the RF receiver includes an amplifier 716A and an analog-to-digital convertor 716B that detect and digitized received signals to obtain the signals S(B). Slice selection gradients may be applied with the same hardware used to apply the diffusion gradients. The gradient controller 708 may be configured to produce pulses or other gradient fields along one or more axes as needed for a particular b-matrix. By selection of such gradients and other applied pulses, signals associated with a plurality of b-matrices are acquired.

[0141] For imaging, specimens are divided into volume elements (voxels) and MR signals for a plurality of gradient directions are acquired as discussed above, but signals may be acquired for one or only a few specimen voxels. In typical examples, signals are obtained for some or all voxels of interest.

[0142] A computer 724 or other processing system such as a personal computer, a workstation, a personal digital assistant, laptop computer, smart phone, and / or a networked computer may be provided for acquisition, control, and / or analysis of specimen data. The computer 724 generally includes a hard disk, a removable storage medium such as a floppy disk or CD-ROM, and / or other memory such as random access memory (RAM). Data may also be transmitted to and from a network using cloud-based processors and storage. Data may be uploaded to the Cloud or stored elsewhere. Computer-executable instructions for, e.g., data acquisition or control, b-matrix computations, such as random selected of directions and random selected of b-magnitudes as well as determining the associated b-matrices, and tensors and distribution estimations may be provided on a floppy disk or other storage medium such as a memory or delivered to the computer 724 via a local area network, the Internet, or other network. Signal acquisition, instrument control, and signal analysis may be performed locally or with distributed processing. For example, signal acquisition and signal analysis may be performed at different locations. Signal evaluation may be performed remotely from signal acquisition by communicating stored data to a remote processor. In general, control and data acquisition with the MRI system and / or apparatus may be provided with a local processor, or via instruction and data transmission via a network.Leydig 775185HHSRef E-085-2024-0-PCT-0130

[0143] FIG. 8 shows a representative computing and control environment in accordance with one or more examples of the present application. FIG. 8 and the following discussion are intended to provide a brief, general description of an exemplary computing / data acquisition environment in which the disclosed technology may be implemented. Although not required, the disclosed technology is described in the general context of computer executable instructions, such as program modules, being executed by a personal computer (PC), a mobile computing device, tablet computer, or other computational and / or control device, system, and / or apparatus. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform particular tasks or implement particular abstract data types. Moreover, the disclosed technology may be implemented with other computer system configurations, including, multiprocessor systems, network PCs, minicomputers, mainframe computers, and the like. The disclosed technology may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

[0144] With reference to FIG. 8, an exemplary system for implementing the disclosed technology includes a general purpose computing device in the form of an exemplary conventional PC 800, including one or more processing units 802, a system memory 804, and a system bus 806 that couples various system components including the system memory 804 to the one or more processing units 802. The system bus 806 may be any of several types of bus structures including a memory bus or memory controller, a peripheral bus, and a local bus using any of a variety of bus architectures. The exemplary system memory 804 includes read only memory (ROM) 808 and random access memory (RAM) 810. A basic input / output system (BIOS) 812, including the basic routines that help with the transfer of information between elements within the PC 800, is stored in ROM 808.

[0145] The exemplary PC 800 further includes one or more storage devices 810 (e.g., non-transitory computer-readable mediums) such as a hard disk drive for reading from and writing to a hard disk, a magnetic disk drive for reading from or writing to a removable magnetic disk, an optical disk drive for reading from or writing to a removable optical disk (such as a CD-ROM or other optical media), and a solid state drive. Such storage devices may be connected to the system bus 806 by a hard disk drive interface, a magnetic disk drive interface, an optical drive interface,Leydig 775185HHSRef E-085-2024-0-PCT-0131or a solid state drive interface, respectively. The drives and their associated computer readable media provide nonvolatile storage of computer-readable instructions, data structures, program modules, and other data for the PC 800. Other types of computer-readable media which may store data that is accessible by a PC, such as magnetic cassettes, flash memory cards, digital video disks, CDs, DVDs, RAMs, ROMs, and the like, may also be used in the exemplary operating environment.

[0146] A number of program modules may be stored in the storage devices 810 including an operating system, one or more application programs, other program modules, and program data. A user may enter commands and information into the PC 800 through one or more input devices 840 such as a keyboard and a pointing device such as a mouse. Other input devices may include a digital camera, microphone joystick, game pad, satellite dish, scanner, or the like. These and other input devices are often connected to the one or more processing units 802 through a serial port interface that is coupled to the system bus 806, but may be connected by other interfaces such as a parallel port, game port, or universal serial bus (USB). A monitor 846 or other type of display device is also connected to the system bus 806 via an interface, such as a video adapter. Other peripheral output devices, such as speakers and printers may be included. Additionally, and / or alternatively, one or more further output devices 845 may be also connected to the system bus 806. The output devices 845 may include one or more additional monitors, other visual output devices such as display devices, audio output devices, and / or other devices that communicate and / or provide information to the user.

[0147] The PC 800 may operate in a networked environment using logical connections to one or more remote computers, such as a remote computer 860. In some examples, one or more network or communication connections 850 are included. The remote computer 860 may be another PC, a server, a router, a network PC, or a peer device or other common network node, and typically includes many or all of the elements described above relative to the PC 800, although only a memory storage device 861 has been illustrated in FIG. 8. The personal computer 800 and / or the remote computer 860 can be connected to a logical a local area network (LAN) and a wide area network (WAN). Such networking environments are commonplace in offices, enterprise wide computer networks, intranets, and the Internet.

[0148] When used in a LAN networking environment, the PC 800 is connected to the LAN through a network interface. When used in a WAN networking environment, the PC 800 typicallyLeydig 775185HHSRef E-085-2024-0-PCT-0132includes a modem or other means for establishing communications over the WAN, such as the Internet. In a networked environment, program modules depicted relative to the personal computer 800, or portions thereof, may be stored in the remote memory storage device or other locations on the LAN or WAN. The network connections shown are exemplary, and other means of establishing a communications link between the computers may be used.

[0149] The memory 804 generally includes computer-executable instructions for performing one or more operations, functions, algorithms, and / or processes described herein. For example, memory portion 862 may store computer-executable instructions for determining gradient directions, and computer-executable instructions for b-matrix computation and random b-magnitude and direction selection may be stored at 873B or previously computed b-matrix specification can be stored in memory portion 873A. Computer-executable instructions for processing acquired signals (for example, determining mean diffusion tensor and covariance) may be stored in a memory portions 860, 871. Computer-executable instructions for data acquisition and control are stored in a memory portion 870. Acquired and processed data may be displayed using computer-executable instructions stored at memory portion 871. As noted above, data acquisition, processing, and instrument control may be provided at an MRI system 874 such as the MRI system described in FIG. 7 above, or distributed at one or more processing devices using a LAN or WAN.

[0150] FIG. 9 depicts an exemplary method 900 for producing a myelin water fraction (MWF) map image a using magnetic resonance imaging (MRI) system. The method 900 may be performed by a control system such as the computer 724 from FIG. 7, the PC 800 from FIG. 8, and / or other control systems (e.g., mobile computing devices, tablet computers, or other computational and / or control devices, systems, and / or apparatuses). For convenience, the embodiment depicts operations as blocks organized a particular order, yet, a person of ordinary skill in the art would recognize that other orders are suitable unless otherwise restricted.

[0151] The method 900 includes, in an operation 901, acquiring a plurality of spoiled gradient recalled echo (SPGR) images of a specimen with a plurality of different flip angles (FAs); and, in an operation 902, acquiring a plurality of balanced steady-state free precession (bSSFP) images of the specimen with a second plurality of different FAs and a plurality of different repetition times (TRs) 902.Leydig 775185HHSRef E-085-2024-0-PCT-0133

[0152] Using the SPGR images acquired in 901, the method 900 applies, in operation 903, driven equilibrium single pulse observation of longitudinal relaxation times (DESPOT1) to generate a longitudinal relaxation time (T1) map. Thus, a T1 map is produced from SPGR images at different FAs using DESPOT1.

[0153] Using the Ti map generated in operation 903 and the bSSFP images acquired in operation 902, in operation 904, the method executes the above-described USA methodology to generate an off-resonance frequency (Δω) parameter map and a transverse relaxation time (T2) map. USA utilizes bSSFP images at different repetition times in order to produce a T2 map and a Δω × TR1 map without needing bSSFP images acquired at different phase increments.

[0154] In operation 905, the method 900 uses Bayesian Monte-Carlo (BMC) analysis, to generate the MWF map from the SPGR images acquired in operation 901, a subset of the bSSFP images acquired in operation 902, and the Δω parameter map generated in 904.

[0155] In operation 906, the method 900 generates and outputs a human readable image of one or more of the MWF map, the T1 map, T2 map, and / or the Δω parameter map

[0156] Table I provides exemplary pseudocode for implementing the method of for producing a myelin water fraction (MWF) map image a using magnetic resonance imaging (MRI) system.Table Ibegin / / Define constants and parametersdefine imagingVolume as 3D region of interestdefine resolution as target spatial resolutiondefine TR_SPGR as repetition time for SPGRdefine TE SPGR as echo time for SPGRdefine SPGR_flipAngles as [FA1, FA2, ..., FAK] / / Array of flip angles for SPGRdefine TR1_bSSFP as repetition time for bSSFP at TR1define TR2_bSSFP as repetition time for bSSFP at TR2define TR3 bSSFP as repetition time for bSSFP at TR3Leydig 775185HHSRef E-085-2024-0-PCT-0134define TE_bSSFP as echo time for bSSFPdefine bSSFP_flipAnglesPhase Pi_TR1 as [FA1, FA2, ..., FAP] / / Array of flip angles for bSSFP at phase π with TR1_bSSFPdefine bSSFP_flipAnglesPhase Pi_TR2 as [FA1, FA2] / / Two flip angles for bSSFP at phase π with TR2_bSSFPdefine bSSFP_flipAnglesPhase Pi_TR3 as [FA1, FA2] / / Two flip angles for bSSFP at phase π with TR3_bSSFPdefine B1Map as 3D B1 field map / / Map for B1 inhomogeneity correctiondefine T1Map as 3D T1 relaxation time mapdefine numSamples as 20000 / / Number of Monte Carlo samples for integrationdefine noiseLevels as [sigma_SPGR, sigma_bSSFPpi] / / Noise standard deviationsdefine parameterBounds as {T2s: [2ms, 45ms], T2l: [45ms, 200ms], T1s: [100ms, 700ms], T1l: [700ms, 3000ms]} / / Parameter search bounds / / Initialize MRI scannerinitialize MRI_scanner with default parametersset gradientsystem for 3D acquisitionset RF pulse for SPGR and bSSFP / / Function to acquire 3D SPGR images at multiple flip anglesfunction acquireSPGRImages(imagingVolume, resolution,SPGR_flipAngles, TR_SPGR, TE_SPGR)configure MRI_scanner for SPGR_modefor each FA in SPGR_flipAnglesset flipAngle = FAset TR = TR_SPGRset TE = TE_SPGRLeydig 775185HHSRef E-085-2024-0-PCT-0135acquire 3D_volumeSignal for imagingVolumereconstruct 3D image from 3D volumesignalstore 3D image for flipAngle FAend forreturn all SPGR imagesend function / / Function to acquire 3D bSSFP images at phase π with different repititionsfunction acquirebSSFPImagesPhasePi(imagingVolume, resolution, bSSFP_flipAnglesPhase Pi_TR1, bSSFP_flipAnglesPhase Pi_TR2, bSSFP_flipAnglesPhase Pi_TR3, TR_bSSFP, TE_bSSFP)configure MRI_scanner for bSSFP_mode / / Acquire bSSFP images at phase π with TR1set phaseCycling = pifor each FA in bSSFP_flipAnglesPhase Pi_TR1set flipAngle = FAset TR = TR1_bSSFPset TE = TE_bSSFPacquire 3D_volumeSignal for imagingVolumereconstruct 3D image from 3D volumesignalstore 3D image for flipAngle FA and phase phaseCycling with TRIend for / / Acquire bSSFP images at phase π with TR2set phaseCycling = pifor each FA in bSSFP_flipAnglesPhase Pi_TR2set flipAngle = FAset TR = TR2_bSSFPLeydig 775185HHSRef E-085-2024-0-PCT-0136set TE = TE_bSSFPacquire 3D volumesignal for imagingVolumereconstruct 3D image from 3D volumesignalstore 3D image for flipAngle FA and phase phasecycling with TR2end for / / Acquire bSSFP images at phase π with TR3set phaseCycling = pifor each FA in bSSFP_flipAnglesPhase Pi_TR3set flipAngle = FAset TR = TR3_bSSFPset TE = TE_bSSFPacquire 3D volumesignal for imagingVolumereconstruct 3D_image from 3D_volumeSignalstore 3D image for flipAngle FA and phase phaseCycling with TR3end forreturn all bSSFP images phase piend function / / Function to acquire Bl mapfunction acquireB1Map(imagingVolume, resolution)configure MRI_scanner for B1_mapping_modeacquire 3D volumesignal for imagingVolumereconstruct BiMap from 3D volumesignalreturn BiMapend functionLeydig 775185HHSRef E-085-2024-0-PCT-0137 / / Function to generate T1 map using SPGR data and Bl mapfunction generateTIMap (SPGR images, BiMap, SPGR flipAngles ) initialize TIMap with zerosfor each voxel in imagingVolumeextract signalIntensity for voxel across SPGR_flipAnglescorrect signalIntensity using B1Mapfit single-component model to corrected signalIntensity and SPGR_flipAnglescalculate T1 value for voxelstore T1 value in TIMap at corresponding voxel location end forreturn TIMapend function / / Function to generate T2 and Δω * TR maps using mc-USAfunction generateT2DeltaMaps (bSSFP images, BIMap, TIMap,bSSFP f lipAnglesPhasePi TRI, bSSFP f lipAnglesPhasePi TR2, bSSFP_f lipAnglesPhasePi_TR3 )initialize T2Map and DeltaMap with zerosfor each voxel in imagingVolumeextract signallntensity for voxel acrossbSSFP f lipAnglesPhasePi TRIextract signallntensity for voxel acrossbSSFP f lipAnglesPhasePi TR2extract signallntensity for voxel acrossbSSFP f lipAnglesPhasePi TR3correct signalIntensity using B1Map and T1Mapfit multi-configuration model to corrected signallntensity and all bSSFP flip angle setscalculate T2 and Δω * TR values for voxelLeydig 775185HHSRef E-085-2024-0-PCT-0138store T2 value in T2Map and Δω * TR value in DeltaMap at corresponding voxel locationend forreturn T2Map, DeltaMapend function / / Function to estimate MWF using Bayesian Monte Carlo with updated bSSFP configurationsfunction estimateMWF(SPGR_images, bSSFP_images, B1Map, T1Map, T2Map, DeltaMap, noiseLevels, parameterBounds, numSamples)initialize priorDistributions for all parameters in parameterBoundsinitialize posteriorPDF as emptyfor each sample in numSamplessampleParameters randomly from priorDistributionscorrect SPGR_signals using B1Mapcorrect bSSFP_signals using B1Map, T1Map, and T2Mapcalculate SPGR_modelSignal using sampled parameterscalculate bSSFP_modelSignal for phase π using all bSSFP configurationscalculate likelihood for SPGR signal as GaussianLikelihood(SPGR_images, SPGR_modelSignal, noiseLevels[1])calculate likelihood for bSSFP signals as GaussianLikelihood(bSSFP_images, bSSFP_modelSignal, noiseLevels[2])update posteriorPDF with product of likelihoods and prior for current sampleend forLeydig 775185HHSRef E-085-2024-0-PCT-0139marginalize posteriorPDF over nuisance parameters to derive posterior distribution of MWFcalculate MWF estimate as mean of posteriorPDFreturn MWF estimateend function / / Main programbeginMainspgrImages = call acquireSPGRImages(imagingVolume, resolution, SPGR_flipAngles, TR_SPGR, TE_SPGR)bssfpImagesPhasePi = callacquirebSSFPImagesPhasePi ( imagingVolume, resolution,bSSFP f lipAnglesPhasePi TRI, bSSFP f lipAnglesPhasePi TR2, bSSFP_flipAnglesPhasePi_TR3, TR_bSSFP, TE_bSSFP)B1Map = call acquireB1Map(imagingVolume, resolution)TIMap = call generateTIMap (spgrlmages, BiMap, SPGR flipAngles ) T2Map, DeltaMap = call generateT2DeltaMaps (bssfpImagesPhasePi, BiMap, TIMap, bSSFP f lipAnglesPhasePi,bSSFP f lipAnglesPhasePi TRI, bSSFP f lipAnglesPhasePi TR2, bSSFP_f lipAnglesPhasePi_TR3 )noiseLevels = estimateNoiseLevels(spgrImages, bssfpImagesPhasePi)MWF = call estimateMWF(spgrImages, bssfpImagesPhasePi, B1Map, T1Map, T2Map, DeltaMap, noiseLevels, parameterBounds, numSamples)output MWF as 3D mapendMainend

[0157] Embodiments implemented according to the present disclosure represent marked-technological advancements out over current bSSFP MRI methodologies (particularly for myelinLeydig 775185HHSRef E-085-2024-0-PCT-0140content characterization). Significantly, the disclosed improvements circumvent the complexities of phase cycling for BO inhomogeneity correction, a process which involves great expertise and time investment in software programming and MRI physics. By shifting the focus to adjusting repetition time (TR), the disclosed improvements simplify operational aspects, making MWF imaging more accessible to a broader range of MRI technologists, clinicians, and researchers without the need for extensive training or specialized knowledge. The technological improvements described herein open the way for truly and easily integrable method for relaxometry and myelin imaging. Additionally, the disclosed improvements also negate the need for operating the MRI device to implement phase cycling, which not only simplifies the process but also negates the need for research licenses from vendors, often a prerequisite in implementing such phase cycling, thereby reducing administrative hurdles and associated costs.

[0157] Moreover, because the disclosed technology provides for a reduced-resource implementation of a Bo corrected bSSFP sequence, it can significantly enhance the quality of structural images for radiologists. In addition, the improvements provided in the present disclosure simplify the integration of advanced quantitative MRI (qMRI) measurements, such as myelin water fraction (MWF) mapping. As indicated above, MWF is a crucial in-vivo imaging biomarker for assessing cerebral myelin content, which is known to affect various physiological functions, including processing speed, executive function, and gait speed. A major hurdle in utilizing advanced qMRI techniques like MWF mapping has been their limited availability and applicability in clinical practice. MRI machines configured to implement the improvements described herein therefore provide clinically relevant technology with a greatly reduced resource requirement as compared to that presently available.

[0158] It will further be appreciated by those of skill in the art that the execution of the various machine-implemented processes and steps described herein may occur via the computerized execution of processor-executable instructions stored on a non-transitory computer-readable medium, e.g., random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), volatile, nonvolatile, or other electronic memory mechanism. Thus, for example, the operations described herein as being performed by computing devices and / or components thereof may be carried out by according to processor-executable instructions and / or installed applications corresponding to software, firmware, and / or computer hardware.Leydig 775185HHSRef E-085-2024-0-PCT-0141

[0159] The use of the term “at least one” followed by a list of one or more items (for example, “at least one of A and B”) is to be construed to mean one item selected from the listed items (A or B) or any combination of two or more of the listed items (A and B), unless otherwise indicated herein or clearly contradicted by context. The terms “comprising,” “having,” “including,” and “containing” are to be construed as open-ended terms (i.e., meaning “including, but not limited to,”) unless otherwise noted. Recitation of ranges of values herein are merely intended to serve as a shorthand method of referring individually to each separate value falling within the range, unless otherwise indicated herein, and each separate value is incorporated into the specification as if it were individually recited herein. All methods described herein can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The use of any and all examples, or exemplary language (e.g., “such as”) provided herein, is intended merely to better illuminate the application and does not pose a limitation on the scope of the application unless otherwise claimed. No language in the specification should be construed as indicating any non-claimed element as essential to the practice of the application.

[0160] It will be appreciated that the examples of the application described herein are merely exemplary. Variations of these examples may become apparent to those of ordinary skill in the art upon reading the foregoing description. The inventors expect skilled artisans to employ such variations as appropriate, and the inventors intend for the application to be practiced otherwise than as specifically described herein. Accordingly, this application includes all modifications and equivalents of the subject matter recited in the claims appended hereto as permitted by applicable law. Moreover, any combination of the above-described elements in all possible variations thereof is encompassed by the application unless otherwise indicated herein or otherwise clearly contradicted by context.

[0161] The following list of publications provide additional background details, and the entire content of each of the following are expressly incorporated by reference herein.1. Akhonda M, Faulkner ME, Gong Z, Laporte JP, Church S, D'Agostino J, et al. The effect of the human brainstem myelination on gait speed in normative aging. The journals of gerontology Series A, Biological sciences and medical sciences. 2023.Leydig 775185HHSRef E-085-2024-0-PCT-01422. Arshad M, Stanley JA, Raz N. Adult age differences in subcortical myelin content are consistent with protracted myelination and unrelated to diffusion tensor imaging indices.NeuroImage. 2016;143:26-39.3. Bouhrara M, Rejimon AC, Cortina LE, Khattar N, Bergeron CM, Ferrucci L, et al. Adult brain aging investigated using BMC-mcDESPOT based myelin water fraction imaging.Neurobiology of aging. 2019.4. Dvorak AV, Swift-LaPointe T, Vavasour IM, Lee LE, Abel S, Russell-Schulz B, et al. An atlas for human brain myelin content throughout the adult life span. Scientific Reports.2021;11(1):269.5. Kiely M, Triebswetter C, Cortina LE, Gong Z, Alsameen MH, Spencer RG, et al. Insights into human cerebral white matter maturation and degeneration across the adult lifespan.NeuroImage. 2022;247: 118727.6. Kolind S, Matthews L, Johansen-Berg H, Leite MI, Williams SC, Deoni S, et al. Myelin water imaging reflects clinical variability in multiple sclerosis. NeuroImage. 2012;60(1):263-70.7. Kolind S, Seddigh A, Combes A, Russell-Schulz B, Tam R, Yogendrakumar V, et al. Brain and cord myelin water imaging: a progressive multiple sclerosis biomarker. NeuroImage Clinical. 2015;9:574-80.8. Laule C, Vavasour IM, Moore GR, Oger J, Li DK, Paty DW, et al. Water content and myelin water fraction in multiple sclerosis. A T2 relaxation study. Journal of neurology.2004;251(3):284-93.9. Neeb H, Schenk J, Weber B. Multicentre absolute myelin water content mapping:Development of a whole brain atlas and application to low-grade multiple sclerosis.NeuroImage: Clinical. 2012;1(1):121-30.10. Neema M, Goldberg-Zimring D, Guss ZD, Healy BC, Guttmann CR, Houtchens MK, et al. 3 T MRI relaxometry detects T2 prolongation in the cerebral normal-appearing white matter in multiple sclerosis. NeuroImage. 2009;46(3):633-41.11. Oh J, Han ET, Lee MC, Nelson SJ, Pelletier D. Multislice brain myelin water fractions at 3T in multiple sclerosis. Journal of neuroimaging: official journal of the American Society of Neuroimaging. 2007;17(2):156-63.Leydig 775185HHSRef E-085-2024-0-PCT-014312. Vargas WS, Monohan E, Pandya S, Raj A, Vartanian T, Nguyen TD, et al. Measuring longitudinal myelin water fraction in new multiple sclerosis lesions. NeuroImage Clinical.2015;9:369-75.13. Moldovan K, Boxerman JL, O'Muircheartaigh J, Dean D, Eyerly-Webb S, Cosgrove GR, et al. Myelin water fraction changes in febrile seizures. Clin Neurol Neurosurg. 2018;175:61-7.14. Drenthen GS, Fonseca Wald ELA, Backes WH, Debeij-Van Hall M, Hendriksen JGM, Aldenkamp AP, et al. Lower myelin-water content of the frontal lobe in childhood absence epilepsy. Epilepsia. 2019;60(8): 1689-96.15. Dean DC, III, Sojkova J, Hurley S, Kecskemeti S, Okonkwo O, Bendlin BB, et al.Alterations of Myelin Content in Parkinson’s Disease: A Cross-Sectional Neuroimaging Study. PloS one. 2016;11(10):e0163774.16. Dean DC, 3rd, Hurley SA, Kecskemeti SR, O'Grady JP, Canda C, Davenport-Sis NJ, et al. Association of Amyloid Pathology With Myelin Alteration in Preclinical Alzheimer Disease. JAMA Neurol. 2017;74(1):41-9.17. Hirschfeld LR, Risacher SL, Nho K, Saykin AJ. Myelin repair in Alzheimer’s disease: a review of biological pathways and potential therapeutics. Translational Neurodegeneration. 2022;11(1):47.18. Bouhrara M, Reiter D, Bergeron C, Zukley L, Ferrucci L, Resnick S, et al. Evidence of demyelination in mild cognitive impairment and dementia using a direct and specific magnetic resonance imaging measure of myelin content. Alzheimer's & Dementia. 2018;14(8):998-1004.19. Balaji S, Johnson P, Dvorak AV, Kolind SH. Update on myelin imaging in neurological syndromes. Current opinion in neurology. 2022;35(4):467-74.20. MacKay AL, Laule C. Magnetic Resonance of Myelin Water: An in vivo Marker for Myelin. Brain Plasticity. 2016;2(l):71-91.21. Does MD. Inferring brain tissue composition and microstructure via MR relaxometry. Neuroimage. 2018;182:136-48.22. Alonso-Ortiz E, Levesque IR, Pike GB. MRI-based myelin water imaging: A technical review. Magn Reson Med. 2015;73(1):70-81.23. Lee J, Hyun J-W, Lee J, Choi E-J, Shin H-G, Min K, et al. So You Want to Image Myelin Using MRI: An Overview and Practical Guide for Myelin Water Imaging. Journal of Magnetic Resonance Imaging. 2021;53(2):360-73.Leydig 775185HHSRef E-085-2024-0-PCT-014424. Piredda GF, Hilbert T, Thiran J-P, Kober T. Probing myelin content of the human brain with MRI: A review. Magnetic Resonance in Medicine. 2021;85(2):627-52.25. Bouhrara M, Spencer RG. Improved determination of the myelin water fraction in human brain using magnetic resonance imaging through Bayesian analysis of mcDESPOT. NeuroImage.2016;127:456-71.26. Deoni SC. Correction of main and transmit magnetic field (BO and Bl) inhomogeneity effects in multicomponent-driven equilibrium single-pulse observation of T1 and T2. Magnetic resonance in medicine. 2011;65(4):1021-35.27. Deoni SC, Matthews L, Kolind SH. One component? Two components? Three? The effect of including a nonexchanging "free" water component in multicomponent driven equilibrium single pulse observation of T1 and T2. Magnetic resonance in medicine.2013;70(1):147-54.28. Deoni SC. Transverse relaxation time (T2) mapping in the brain with off-resonance correction using phase-cycled steady-state free precession imaging. Journal of magnetic resonance imaging: JMRI. 2009;30(2):411-7.29. Deoni SC, Peters TM, Rutt BK. High-resolution T1 and T2 mapping of the brain in a clinically acceptable time with DESPOT1 and DESPOT2. Magnetic resonance in medicine. 2005;53(1):237-41.30. Deoni SC, Rutt BK, Peters TM. Rapid combined T1 and T2 mapping using gradient recalled acquisition in the steady state. Magnetic resonance in medicine. 2003;49(3):515-26. 31. McConnell HM. Reaction Rates by Nuclear Magnetic Resonance. The Journal of chemical physics. 1958;28(3):430- 1.32. Bouhrara M, Spencer RG. Incorporation of nonzero echo times in the SPGR and bSSFP signal models used in mcDESPOT. Magnetic resonance in medicine. 2015;74(5): 1227-35.33. Bouhrara M, Spencer RG. Rapid simultaneous high-resolution mapping of myelin water fraction and relaxation times in human brain using BMC-mcDESPOT. NeuroImage.2017;147:800-11.34. Faulkner ME, Laporte JP, Gong Z, A. B. S. Akhonda M, Triebswetter C, Kiely M, et al. Lower myelin content is associated with lower gait speed in cognitively unimpaired adults. The Journals of Gerontology: Series A. 2023.Leydig 775185HHSRef E-085-2024-0-PCT-014535. Bouhrara M, Spencer RG. Steady state double angle method for rapid Bl mapping.Magnetic Resonance in Medicine 2018.36. Stollberger R, Wach P. Imaging of the active Bl field in vivo. Magnetic resonance in medicine. 1996;35(2):246-51.37. Gutenkunst RN, Waterfall JJ, Casey FP, Brown KS, Myers CR, Sethna JP. Universally Sloppy Parameter Sensitivities in Systems Biology Models. PLOS Computational Biology.2007;3(10):el89.38. Waterfall JJ, Casey FP, Gutenkunst RN, Brown KS, Myers CR, Brouwer PW, et al.Sloppy-model universality class and the Vandermonde matrix. Physical review letters.2006;97(15): 150601.39. Laule C, Vavasour IM, Kolind SH, et al. Magnetic resonance imaging ofmyelin. Neurotherapeutics 2007; 4(3): 460-48440. Labadie C, Lee JH, Rooney WD, et al. Myelin water mapping by spatially regularized longitudinal relaxographic imaging at high magnetic fields. Magn Reson Med 2014; 71(1): 375-387.41. Ozdemir, S., & Ider, Y. Z. (2019). bSSFP phase correction and its use in magnetic resonance electrical properties tomography. Magnetic Resonance in Medicine, 81(2), 934-946.42. Kitzler, H. H., Wahl, H., Eisele, J. C., Kuhn, M., Schmitz-Peiffer, H., Kern, S.,... & Linn, J. (2018). Multi-component relaxation in clinically isolated syndrome: lesion myelination may predict multiple sclerosis conversion. NeuroImage: Clinical, 20, 61-70.43. Kitzler, H. H., Wahl, H., Kuntke, P., Deoni, S. C., Ziemssen, T., Linn, J., & Kohler, C. (2022). Exploring in vivo lesion myelination dynamics: longitudinal myelin water imaging in early multiple sclerosis. NeuroImage: Clinical, 36, 103192.44. Jurick, S. M., Hoffman, S. N., Sorg, S., Keller, A. V., Evangelista, N. D., DeFord, N. E.,... & Jak, A. J. (2018). Pilot investigation of a novel white matter imaging technique in Veterans with and without history of mild traumatic brain injury. Brain injury, 32(10), 1255-1264.45. Spader, H. S., Dean, D. C., LaFrance, W. C., Raukar, N. P., Cosgrove, G. R., Eyerly-Webb, S. A.,... & Rogg, J. (2018). Prospective study of myelin water fraction changes after mild traumatic brain injury in collegiate contact sports. Journal of neurosurgery, 130(4), 1321-1329.46. Bouhrara, M., Reiter, D. A., Bergeron, C. M., Zukley, L. M., Ferrucci, L., Resnick, S. M., & Spencer, R. G. (2018). Evidence of demyelination in mild cognitive impairment and dementiaLeydig 775185HHSRef E-085-2024-0-PCT-0146using a direct and specific magnetic resonance imaging measure of myelin content. Alzheimer's & Dementia, 14(8), 998-1004.47. Gong, Z., Khattar, N., Kiely, M., Triebswetter, C., & Bouhrara, M. (2023). REUSED: A deep neural network method for rapid whole-brain high-resolution myelin water fraction mapping from extremely under-sampled MRI. Computerized Medical Imaging and Graphics, 108, 102282.

Claims

Leydig 775185HHSRef E-085-2024-0-PCT-0147CLAIMS:

1. A method for producing a myelin water fraction (MWF) map image using a magnetic resonance imaging (MRI) system, the method comprising:acquiring a plurality of spoiled gradient recalled echo (SPGR) images of a specimen with a plurality of different flip angles (FAs);acquiring a plurality of balanced steady-state free precession (bSSFP) images of the specimen with a second plurality of different FAs and a plurality of different repetition times (TRs);using driven equilibrium single pulse observation of longitudinal relaxation times (DESPOT1), generating a longitudinal relaxation time (T1) map from the SPGR images;using the T1 map and the bSSFP images, applying USA and generating an off-resonance frequency (Δω) parameter map and a transverse relaxation time (T2) map; and using Bayesian Monte-Carlo (BMC) analysis, generating the MWF map image from the SPGR images, a subset of the bSSFP images, and the Δω parameter map.

2. The method of claim 1, wherein acquiring the SPGR images comprises, operating the MRI system to apply and receive MRI signals to and from the specimen using SPGR pulse sequencing using the different FAs with a TR fixed at a first value.

3. The method of claim 1, wherein acquiring the bSSFP images comprises:acquiring first bSSFP images at A different FAs and at a first TR;acquiring second bSSFP images at Q different FAs and at a second TR; and acquiring third bSSFP images at A different FAs and a third TR, and wherein the first bSSFP images are the subset of the bSSFP images used in the BMC analysis.

4. The method of claim 3,wherein acquiring the first bSSFP images comprises operating the MRI system to apply and receive MRI signals to and from the specimen using bSSFP pulse sequencing using the P different FAs with a TR fixed at the first TR,Leydig 775185HHSRef E-085-2024-0-PCT-0148wherein acquiring the second bSSFP images comprises operating the MRI system to apply and receive MRI signals to and from the specimen using bSSFP pulse sequencing using the Q different FAs with the TR fixed at the second TR, and wherein acquiring the third bSSFP images comprises operating the MRI system to apply and receive MRI signals to and from the specimen using bSSFP pulse sequencing using the R different FAs with the TR fixed at the third TR.

5. The method of claim 3, wherein the first TR is shorter than the second TR, and the second TR is shorter than the third TR.

6. The method of claim 5, wherein the second TR is twice the first TR and the third TR is three times the first TR.

7. The method of claim 3, wherein Q and R are 2, and P is between 2 and 10.

8. The method of claim 1, wherein the bSSFP images are acquired without phase cycling.

9. The method of claim 1, wherein all of the bSSFP images are acquired with the RF phase set as i.

10. The method of claim 1, the method further comprising outputting at least one of the the MWF map image, the Ti map, the T2 map, or the Am parameter map.

11. The method of claim 10, the method further comprising providing a quantified measurement of myelin content of the specimen base on the MWF map image.

12. The method of claim 10, wherein the specimen is a human-patient’s brain, and the method further comprises treating the patient for a neurological disorder based on the MWF map image.

13. A magnetic resonance imaging (MRI) system, the MRI system comprising one or more processors configured to control the MRI system to apply MRI signals to a specimen, to receive MRI data based on the applied MRI signals, and to generate MRI images according to a method comprising:Leydig 775185HHSRef E-085-2024-0-PCT-0149acquiring a plurality of spoiled gradient recalled echo (SPGR) images of the specimen with a plurality of different flip angles (FAs);acquiring a plurality of balanced steady-state free precession (bSSFP) images of the specimen with a second plurality of different FAs and a plurality of different repetition times (TRs);using driven equilibrium single pulse observation of longitudinal relaxation times (DESPOT1), generating a longitudinal relaxation time (T1) map from the SPGR images;using the T1 map and the bSSFP images, applying USA and generating an off-resonance frequency (Δω) parameter map and a transverse relaxation time (T2) map; and using Bayesian Monte-Carlo (BMC) analysis, generating the MWF map image from the SPGR images, a subset of the bSSFP images, and the Δω parameter map.

14. The MRI system of claim 13, wherein the MRI system is not enabled to apply bSSFP sequencing with phase cycling.

15. A non-transitory computer readable medium comprising instructions, which upon execution by one or more processors, is configured to cause the one or more processors to execute a method comprising:acquiring a plurality of spoiled gradient recalled echo (SPGR) images of the specimen with a plurality of different flip angles (FAs);acquiring a plurality of balanced steady-state free precession (bSSFP) images of the specimen with a second plurality of different FAs and a plurality of different repetition times (TRs);using driven equilibrium single pulse observation of longitudinal relaxation times (DESPOT1), generating a longitudinal relaxation time (T1) map from the SPGR images;using the T1 map and the bSSFP images, applying USA and generating an off-resonance frequency (Δω) parameter map and a transverse relaxation time (T2) map; and using Bayesian Monte-Carlo (BMC) analysis, generating the MWF map image from the SPGR images, a subset of the bSSFP images, and the Δω parameter map.

Images