Systems and methods for phase reversed interleaved multi-echo (PRIME) acquisition for distortion-free MRI
The PRIME acquisition method addresses dMRI's low resolution and distortion issues by using phase-reversed echoes for high-fidelity field map estimation, achieving high-resolution and distortion-free dMRI with efficient scan times.
Patent Information
- Application Number
- PCT/US2026/027648
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-09
- Filing Date
- 2026-05-11
- Publication Date
- 2026-11-12
AI Technical Summary
Diffusion magnetic resonance imaging (dMRI) suffers from low spatial resolution, low signal-to-noise ratio (SNR), and geometric distortion, limiting its clinical and research applications, particularly in high-resolution neuroimaging.
The Phase Reversed Interleaved Multi-Echo (PRIME) acquisition method involves acquiring multiple echoes with opposite phase encoding polarities to estimate high-fidelity field maps and correct geometric distortions, allowing for high-resolution and distortion-free dMRI by utilizing additional echoes without increasing scan time.
PRIME achieves highly accelerated, high-resolution dMRI with reduced geometric distortion, enabling mesoscale imaging at 490 µm isotropic resolution and improved diffusion relaxometry parameter estimation.
Smart Images

Figure US2026027648_12112026_PF_FP_ABST
Abstract
Description
Atty. Dkt. No. 125141.05023. MGH2024-029SYSTEMS AND METHODS FOR PHASE REVERSED INTERLEAVED MULTI-ECHO (PRIME) ACQUISITION FOR DISTORTION-FREE MRICROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is based on, claims priority to, and incorporates herein by reference for all purposes, U. S. Provisional Patent Application No. 63 / 803,522 filed on May 9, 2025.STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH
[0002] This invention was made with government support under EB032378 awarded by the National Institutes of Health. The government has certain rights in the invention.BACKGROUND
[0003] Diffusion magnetic resonance imaging (dMRI), which may also be referred to as diffusion weighted imaging (D Wl), is a commonly used magnetic resonance imaging (MRI) technique that produces contrast based on the diffusion of water molecules within the imaged tissue, giving insight to tissue microstructure. However, dMRI often suffers from low spatial resolution, low signal to noise ratio (SNR), and geometric distortion and blurring, limiting the clinical and research applications, especially in the context of high-resolution neuroimaging. Thus, novel systems and methods for improving dMRI image quality and resolution are still needed.SUMMARY OF THE DISCLOSURE
[0004] According to an aspect of the present disclosure, a method for magnetic resonance imaging (MRI) is provided. The method includes acquiring a first shot of echo planar imaging (EPI) data with a first phase encoding polarity using an MRI system. The method also includes acquiring a second shot of EPI data with a second phase encodingAtty. Dkt. No. 125141.05023. MGH2024-029polarity that is opposite the first phase encoding polarity using the MRI system. Acquiring each of the first shot and second shot includes acquiring a first echo at a first resolution and acquiring a second echo at a second resolution that is lower than the first resolution. The method further includes estimating a field map from the second echo of the first shot and the second echo of the second shot. The method also includes reconstructing images with reduced distortion from the first echo of the first shot and the first echo of the second shot using the estimated field map.
[0005] According to another aspect of the present disclosure, a system for generating MRI images with reduced distortion is provided. The system includes a computer processor configured to control an MRI system to acquire a first shot of EPI data with a first phase encoding polarity. The computer processor is also configured to control the MRI system to acquire a second shot of EPI data with a second phase encoding polarity that is opposite the first phase encoding polarity. To acquire the first shot and second shot of EPI data, the computer processor is configured to control the MRI system to acquire a first echo at a first resolution and acquire a second echo at a second resolution that is lower than the first resolution. The computer processor is further configured to estimate a field map from the second echo of the first shot and the second echo of the second shot. The computer processor is also configured to reconstruct MRI images with reduced distortion from the first echo of the first shot and the first echo of the second shot using the estimated field map.
[0006] According to another aspect of the present disclosure, a method for magnetic resonance imaging is provided. The method includes acquiring a first shot of diffusion-weighted EPI MRI data with a first phase encoding polarity using an MRI system. The method also includes acquiring a second shot of diffusion-weighted EPI data with a second phase encoding polarity that is opposite the first phase encoding polarityAtty. Dkt. No. 125141.05023. MGH2024-029using the MRI system. Acquiring each of the first shot and second shot includes applying a radiofrequency slab encoding pattern configured to facilitate super-resolution in a slice direction, acquiring a first echo having a first contrast, and acquiring a second echo having a second contrast during a dead time period configured to mitigate spin history effects. The method further includes estimating one or more field maps from at least one of the first echo of the first shot and the first echo of the second shot or the second echo of the first shot and the second echo of the second shot. The method also includes reconstructing a first set of images with reduced distortion from the first echo of the first shot and the first echo of the second shot using at least one of the one or more estimated field maps, where the first set of images has the first contrast. The method further includes reconstructing a second set of images with reduced distortion from the second echo of the first shot and the second echo of the second shot using at least one of the one or more estimated field maps, where the second set of images has the second contrast. The method also includes generating a diffusion relaxometry parameter map based on the first set of images and the second set of images.
[0007] These are but a few, non-limiting examples of aspects of the present disclosures. Other features, aspects and implementation details will be described hereinafter.BRIEF DESCRIPTION OF THE DRAWINGS
[0008] Various objects, features, and advantages of the disclosed subject matter can be more fully appreciated with reference to the following detailed description of the disclosed subject matter when considered in connection with the following drawings, in which like reference numerals identify like elements.
[0009] FIG. 1 provides a flowchart setting forth steps of a non-limiting exampleAtty. Dkt. No. 125141.05023. MGH2024-029process for generating magnetic resonance imaging (MRI) images.
[0010] FIG. 2A provides a pulse sequence diagram illustrating a non-limiting example implementation of the described methods.
[0011] FIG. 2B provides another pulse sequence diagram illustrating a nonlimiting example implementation of the described methods.
[0012] FIG. 3A illustrates k-space traversal for non-limiting example implementations of the described methods.
[0013] FIG. 3B illustrates k-space traversal for another non-limiting example implementation of the described methods, applied in the context of diffusion relaxometry.
[0014] FIG. 4 shows experimental data comparing field maps and image quality achieved using a conventional method and an example implementation of the method described herein.
[0015] FIG. 5 shows experimental data comparing images generated using conventional methods and an example implementation of the method described herein.
[0016] FIG. 6 shows experimental data comparing field maps generated using a conventional method and an example implementation of the method described herein.
[0017] FIG. 7 shows additional experimental data demonstrating brain diffusion weighted images generated using an example implementation of the methods described herein.
[0018] FIG. 8 shows additional experimental data demonstrating brain diffusion weighted images and diffusion relaxometry parameter maps generated using another example implementation of the methods described herein.
[0019] FIG. 9 shows additional experimental data demonstrating mesoscale brain diffusion weighted images and fractional anisotropy maps.Atty. Dkt. No. 125141.05023. MGH2024-029
[0020] FIG. 10 provides a pulse sequence diagram illustrating non-limiting implementations of the PRIME pulse sequence with three echoes acquired in each shot.
[0021] FIG. 11 shows experimental data demonstrating the improved quality and reduced error achieved using the described methods.
[0022] FIG. 12 shows additional experimental data demonstrating high-quality and high-resolution D WI achieved using the described methods in which the second echo provided high-quality field maps and phase navigator data.
[0023] FIG. 13 shows experimental data acquired with three echoes, which can be used to provide diffusion relaxometry data in accordance with the present disclosure.
[0024] FIG. 14 shows additional experimental data in which mesoscale DWI was achieved according to the present disclosure.
[0025] FIG. 15 is a block diagram of an example magnetic resonance imaging (" MRI”) system that can implement the methods described in the present disclosure.
[0026] FIG. 16 is a block diagram of an example imaging system that can implement the methods of the present disclosure.
[0027] FIG. 17 is a block diagram of example components that can implement the system of FIG. 16.DETAILED DESCRIPTION
[0028] Before any aspects of the present disclosure are explained in detail, it is to be understood that the invention is not limited in its application to the details of construction and the arrangement of components set forth in the following description or illustrated in the following drawings. The invention is capable of other embodiments and of being practiced or of being carried out in various ways. Also, it is to be understood that the phraseology and terminology used herein is for the purpose of description andAtty. Dkt. No. 125141.05023. MGH2024-029should not be regarded as limiting. The use of "including,” "comprising,” or "having” and variations thereof herein is meant to encompass the items listed thereafter and equivalents thereof as well as additional items. Unless specified or limited otherwise, the terms "mounted," "connected,” "supported," and "coupled” and variations thereof are used broadly and encompass both direct and indirect mountings, connections, supports, and couplings. Further, "connected” and "coupled” are not restricted to physical or mechanical connections or couplings.
[0029] The following discussion is presented to enable a person skilled in the art to make and use embodiments of the invention. Various modifications to the illustrated embodiments will be readily apparent to those skilled in the art, and the generic principles herein can be applied to other embodiments and applications without departing from embodiments of the invention. Thus, embodiments of the invention are not intended to be limited to embodiments shown but are to be accorded the widest scope consistent with the principles and features disclosed herein. The following detailed description is to be read with reference to the figures, in which like elements in different figures have like reference numerals. The figures, which are not necessarily to scale, depict selected embodiments and are not intended to limit the scope of embodiments of the invention. Skilled artisans will recognize the examples provided herein have many useful alternatives and fall within the scope of embodiments of the invention.
[0030] Diffusion magnetic resonance imaging (dMRI) has been widely used in neuroscience studies for analyzing in vivo brain tissue microstructure and neurological or neurodegenerative disorders, including stroke, white matter (WM) disorders, Alzheimer’s disease, demyelination, tumors, and Parkinson’s disease. However, in clinical settings, dMRI is typically acquired with relatively lower resolution (e.g., about 2 mm isotropic resolution) compared to structural contrast-weighted images, which limitsAtty. Dkt. No. 125141.05023. MGH2024-029analyzing the detailed brain structures such as gray-WM boundary and white-matter fascicles in small cortical regions. Moreover, microstructure features of the human brain, including cell bodies or neurites, can be extracted from dMRI using diffusion models; however, the dMRI resolutions for the models are mostly limited to 1.5~2.5 mm isotropic resolution for in vivo analysis, which inevitably requires indirect estimation of microanatomical information beyond the voxel resolution and may depend on model assumptions and parameter settings. However, increasing the resolution of dMRI is particularly challenging due to 1) low signal-to-noise (SNR) due to decreased voxel size and T2 signal decay occurring over long echo times (TE), 2) T2 and T2* voxel blurring caused by lengthy echo spacing (ESP) and echo-planar imaging (EPI) k-space readouts, and 3) Bo inhomogeneity and eddy current-induced geometric distortion.
[0031] Undersampling the k-space acquisition of dMRI with parallel imaging can mitigate signal decay, voxel blurring, and geometric distortion; however, it suffers from residual aliasing artifacts and noise amplification with high in-plane acceleration factors. To overcome the limitations of the single-shot EPl for dMRI acquisitions, multi-shot EPl has become a popular approach, which enables the reduction of the T2 / T2* voxel blurring and geometric distortion by reducing the readout duration in each shot. However, combining multiple shots is difficult due to shot-to-shot phase variations caused by physiological noise. Several methods, including navigator-based and navigator-free approaches (e.g., multiplexed sensitivity encoding (MUSE), multi-shot sensitivity-encoded diffusion data recovery using structured low-rank matrix completion (MUSSELS), and low-rank modeling of local k-space neighborhoods (LORAKS)), have been proposed to combine multiple shots considering phase variations between the shots. Moreover, to overcome the SNR limitation of 2D EPI-based acquisitions, 3D EPI techniques for volumetric acquisition, such as 3D multi-slab acquisitions, have beenAtty. Dkt. No. 125141.05023. MGH2024-029proposed. Generalized SLIce Dithered Enhanced Resolution Simultaneous Multi-Slab (gSlider-SMS) is one of the acquisition techniques that uses a slab radiofrequency (RF) encoding to acquire multiple volumetric slabs simultaneously.
[0032] Geometric distortions of dMRI due to Bo inhomogeneity and eddy currents can be corrected by obtaining additional shots with reversed phase-encoding and estimating field maps using software, such as FSL TOPUP. Hybrid-space SENSE has been proposed for joint reconstruction of interleaved blip-up and blip-down images with incorporated field maps estimated from FSL TOPUP, which showed improved g-factor compared to standard parallel imaging reconstruction for each of the blip-up and blipdown shots separately. Recently, another approach, Echo-Planar Time-Resolving Imaging (EPTI), was developed for distortion-free dMRI along with relaxometry imaging using spatiotemporal encoding and reconstruction schemes.
[0033] Blip-up and -down acquisition (BUDA)-EPI, which uses joint parallel imaging reconstruction with a structured low-rank constraint framework, enables high-resolution dMRI, and combining BUDA with a gSlider acquisition technique further improved fidelity; BUDA-circular-EPI further pushed the resolution by reducing the readout length and TE with circular EPI acquisition. However, there are limitations of BUDA-EPI. First, the achievable acceleration per shot is limited since residual artifacts in the interim SENSE reconstructed images can be propagated to the estimated field maps (e.g., using FSL TOPUP). This practically constrains BUDA-EPI to use limited in-plane acceleration factors (Rtn-piane) (e.g., typically Rtn-piane of 4-fold or lower). Second, long repetition time (TR) are needed for gSlider RF encoding to mitigate spin history-related slab boundary artifacts and slab-cross talk effects, which leaves significant dead time in the sequence, especially when simultaneous multi-slice (SMS) is employed.
[0034] The systems and methods described herein provide a novel pulse sequenceAtty. Dkt. No. 125141.05023. MGH2024-029for highly accelerated dMRI with reduced geometric distortion. The pulse sequence method may be referred to as " Phase Reversed Interleaved Multi-Echo (PRIME)” acquisition. PRIME benefits from inserting one or more additional echoes, which advantageously do not incur an increased scan time penalty when gSlider is used. The additional echo(es) can be utilized to generate high-fidelity field maps estimation, provide a phase navigator for shot-to shot phase correction, provide a motion navigator for motion estimation across diffusion direction, or to provide multi-echo DWI acquisitions for varying contrast (e.g., diffusion relaxometry acquisition). As will be shown herein, the PRIME sequence can achieve highly accelerated, high-resolution, and distortion-free dMRI on clinical and research scanners. In some non-limiting examples, mesoscale dMRI at 490 pm isotropic resolution can be achieved by utilizing high-performance gradient systems.
[0035] In some implementations, high in-plane acceleration (e.g., Rm-piane of 5 -fold or higher) with 2D partial Fourier (pF) for both phase encoding and frequency encoding directions, can be achieved using the high-fidelity field maps estimated from the second echo, which is made at a lower resolution and acceleration to increase its SNR while matching the effective ESP of the first readout. This ensures that both readouts have identical geometric distortion but also allows the second readout to use lower Rin-piane so that the field maps are estimated from higher-quality interim reconstructions. High acceleration achieved in the first echo helped reduce relaxation-related voxel blurring with reduced ESP and echo time (TE), while joint reconstruction of blip-up and blip-down shots using low-rank constraint with phase prior (S-LORAKS) showed improved reconstruction performance.
[0036] In some implementations, high-resolution diffusion relaxometry parameters can be estimated from dual-echo PRIME data. In this case, the second echoAtty. Dkt. No. 125141.05023. MGH2024-029can have the same resolution, ESP, and Rm-piane as the first echo to match the voxel size and distortion between echoes. This also benefits from long TR when gSlider RF encoding is used with SMS, where one or more additional echoes can be acquired for diffusion relaxometry estimation without prolonging scan time.
[0037] In some implementations, a triple-echo PRIME acquisition can be used for diffusion relaxometry, where all three echoes have matched resolution, effective echo spacing (ESP), and in-plane acceleration factor (Rin-piane). In one non-limiting example, the three echoes can be acquired with echo times of approximately 40, 84, and 126 ms. When using moderate acceleration (e.g., Rin-piane of 4-fold or lower), field maps can be estimated directly from the first echo rather than requiring the second echo. The moderate acceleration (e.g., Rin-piane of 3-fold) can be selected to ensure the echo train spans the necessary temporal range (e.g., TE of approximately 40 to 120 ms) required to robustly separate intra-axonal and extra-cellular T2 components during diffusion relaxometry fitting (e.g., MTE-SMT fitting).
[0038] Referring now to FIG. 1, a flowchart setting forth steps of a process 100 for MRI is provided. The process 100 includes using an MRI system to acquire a first shot of MRI data with a first phase encoding polarity, as in process block 102, and a second shot of MRI data with a second phase encoding polarity, as in process block 108. The MRI data may include data acquired with an EPI readout, such as spin echo EPI.
[0039] The MRI data may also be acquired as diffusion-weighted data, in which diffusion weighting gradients may or may not be applied. For example, when acquiring diffusion-weighted data, one or more diffusion gradients can be applied prior to acquiring the data to produce data with diffusion weighting (e.g., b > 0 s / mm2), or time can be left in the pulse sequence for diffusion gradients without applying the gradients to achieve baseline diffusion weighted data (e.g., b = 0 s / mm2). The acquisition may be repeatedAtty. Dkt. No. 125141.05023. MGH2024-029(e.g., process blocks 102 and 108) for various diffusion weights, gradient directions, slices, slab encodings, and so forth.
[0040] Each shot is characterized by a phase encoding polarity that describes the direction that k-space is traversed in the phase encoding (PE) direction. This k-space trajectory direction corresponds to the polarity of the phase encoding blips used to move between PE lines. Thus, the shots may be acquired with opposite phase encoding polarities and referred to as "blip-up" and "blip-down" shots, respectively.
[0041] In some implementations, the MRI data can be acquired to facilitate superresolution reconstruction in the slice direction. For example, the gSlider method may be employed. In this way, both the first and second shots are acquired with radiofrequency (RF) slab encoding in which the slab is volumetrically encoded using multiple RF encoding patterns. The RF encoding scheme can be designed with bases that are highly independent and provide high image signal-to-noise ratio in each slab acquisition, enabling self-navigation of the diffusion's phase corruption. When reconstructing the data, the slices can be reconstructed to resolve slices with higher slice resolution (e.g., smaller slice thickness) than that of the acquired slab, based on the slab-encoding acquisition scheme. When using gSlider or another super-resolution method, the pulse sequence may also be designed with dead time that mitigates spin history-related effects (e.g., slab boundary artifacts) between shots. In some implementations, a triple-echo acquisition may be used, where for diffusion relaxometry applications, all three echoes may have matched resolution, effective echo spacing, and in-plane acceleration factor to ensure consistent voxel size and geometric distortion across echoes.
[0042] In some implementations, the number of gSlider RF encoding patterns can be varied based on the target resolution and SNR requirements. For example, five gSlider RF encoding pulses may be deployed, resulting in ten shots per slab (five gSlider RFAtty. Dkt. No. 125141.05023. MGH2024-029encodings for each of blip-up and blip-down shots). In other implementations, seven gSlider RF encoding pulses maybe deployed for mesoscale dMRI acquisition, resulting in fourteen shots per slab. Increasing the number of RF encodings can boost SNR, for example by a factor of approximately the square root of the ratio of the number of encodings (e.g., increasing from five to seven encodings provides an SNR boost of approximately 1.18). For a fixed target resolution, increasing the number of RF encodings results in a thicker excitation slab, and consequently fewer thick slabs are needed to cover the same anatomical field of view, which can provide flexibility to reduce TR while maintaining the same spatial coverage.
[0043] In some implementations, the data may be acquired with partial Fourier (pF) used in one or two dimensions to reduce the echo time (TE), echo spacing, and corresponding relaxation-related blurring.
[0044] The data may also be acquired with simultaneous multislice (SMS) encoding, which may be referred to as "multiband” in which multiple slices or slabs (e.g., MB = 2, 3, 4, etc.) are simultaneously acquired.
[0045] As illustrated in process blocks 102 and 108, acquisition of each shot includes acquiring a first echo, as shown in process block 104, and acquiring a second echo, as shown in process block 106. In some implementations, additional echoes may also be acquired. For example, process block 106 may be repeated to acquire a third and fourth echo.
[0046] As will be described further below, the data acquired during the second echo can be used to estimate a field map that characterizes Bo inhomogeneity. Such field map can be applied in the reconstruction of the data acquired during the first echo to correct for spatial distortion induced by Bo inhomogeneity. In some implementations, the data acquired during the second echo can be used to estimate motion across diffusionAtty. Dkt. No. 125141.05023. MGH2024-029directions or to estimate shot-to-shot phase differences. For example, the data acquired during the second echo can be used as a phase navigator to estimate shot-to-shot phase variations or as a motion navigator to estimate motion across diffusion directions. When used as a motion navigator, the magnitude images of the second echo can be used for rigid-body registration between diffusion directions, and the resulting registration parameters can be applied to the high-resolution target images from the first echo. Despite the lower resolution of the second echo (e.g., 2 x 2 mm2in-plane resolution), the high SNR of the second echo can allow for robust rigid-body registration. In some implementations, the data acquired during the second echo can be acquired with a different contrast than the first (e.g., echo time) to estimate relaxation parameters.
[0047] Each echo has an effective echo spacing (ESP) that is based on the corresponding nominal ESP (e.g., time between the center of each k-space line) and the corresponding in-plane acceleration factor. For example, the effective ESP for each echo can be defined as ESPeffective=ESP-~l.in— plane
[0048] In some implementations, the effective echo spacing of the first and second echo can be matched such that the images generated from the first echo and the second echo have identical geometric distortion. In this way, the first and second echo may not be required to have matched nominal echo spacing, spatial resolution, or in-plane acceleration, but still have matched geometric distortion in the resulting images by having matched effective echo spacings. Similarly, optional subsequent echoes (e.g., third, fourth, and so forth) may have matched effective echo spacing as the first echo. The use of matched effective echo spacing can especially improve the accuracy of shot-to-shot phase estimation, field map estimation, and subsequent combination of shots with different phase encoding directions. In other implementations, the effective echoAtty. Dkt. No. 125141.05023. MGH2024-029spacings of the first and subsequent echoes do not need to be equal.
[0049] In some implementations, the first echo can be acquired at a first resolution in process block 104, while the second echo is acquired ata second resolution in process block 106. For example, the second echo can be acquired at a lower spatial resolution than that of the first. In this way, the second echo may advantageously have higher SNR than that of the first, to provide high-fidelity field maps or navigator data, as will be described.
[0050] In some implementations, the first and second echoes can also be acquired with unequal in-plane acceleration factors (Rin-piane). For example, the second echo can be acquired with a lower acceleration factor than the first to reduce errors in the field map, phase differences, or motion estimated based on the data acquired during the second echoes. The use of a lower acceleration factor can advantageously reduce errors that can be propagated through the reconstruction pipeline when the field map is applied to correct for distortion effects or when the navigator data is applied to correct differences across shots or diffusion acquisitions.
[0051] In other implementations, the first and second echoes can be acquired with matched resolution, acceleration, or both but different contrasts, to facilitate estimation of relaxometry parameters (e.g., T2, T2*, Ti) or diffusion relaxometry parameters (e.g., intra-axonal signal fraction, intra-axonal T2, extra-cellular T2, parallel diffusivity, extracellular perpendicular diffusivity, and extra-cellular signal fraction. In some implementations, a high time-bandwidth product RF encoding (e.g., gSlider RF encoding) and a non-Carr-Purcell-Meiboom-Gill (non-CPMG) condition with orthogonal crusher gradients can be used to maintain signal levels across echoes, achieving comparable signal levels between the second echo and single-echo acquisitions acquired separately with the same echo time.Atty. Dkt. No. 125141.05023. MGH2024-029
[0052] In some implementations, the second echo and optional additional echoes can be acquired during the dead time included in the sequence to mitigate spin history effects between shots, as previously described. Such dead time is often advantageous for gSlider, gSlider-SMS, or other super-resolution acquisition methods. In this way, the second echo and optional additional echoes can advantageously be acquired to provide estimation of high-fidelity field maps, high-quality phase or motion navigator data, or multi-echo data (e.g., relaxometry data) without incurring a scan time penalty.
[0053] The data are processed in process block 110 to estimate a field map that describes the Bo field inhomogeneity. The field map can be estimated using blip-up / blip-down methods such as FSL’s TOPUP, BUDA, and so forth based on data acquired with opposite PE polarities. The strategy for field map estimation can be tailored to the specific acquisition protocol. For example, in some implementations using high in-plane acceleration (e.g., Rin. Pia„e of 5 -fold), images generated from the second echo of the blip-up shot can be compared or otherwise processed with images generated from the second echo of the blip-down shot to determine the Bo field map or distortion map. The second echo, acquired at lower resolution and acceleration, provides higher SNR and mitigates artifacts, enabling higher fidelity field map estimation. In other implementations using moderate acceleration (e.g.,of 4-fold or lower), images generated from the first echo of the blip-up shot can be compared or otherwise processed with images generated from the first echo of the blip-down shot to determine the Bo field map or distortion map, since the moderate acceleration allows for robust self-navigation from the first echo.
[0054] Once a field map is estimated, the images are reconstructed in process block 112. Process block 112 may include reconstructing images from the first echoes of the blip-up and blip-down shots, from the second echoes of the blip-down shots, or both. The field map can be applied to first echo data, the second echo data, or both to reduceAtty. Dkt. No. 125141.05023. MGH2024-029geometric distortion in the resulting images.
[0055] As a non-limiting example, the first echoes of the blip-up and blip-down shots can be jointly reconstructed using a model-based approach (e.g., BUDA, hybrid-SENSE, S-LORAKS) in which the estimated field map is used to correct geometric distortion and a low-rank constraint, which enforces low rankness across shots, is applied. Alternatively, when the second echo is used as a phase navigator, rapid Hybrid-SENSE reconstruction with explicit shot-to-shot phase correction can be employed, which can achieve a substantial reduction in computation time (e.g., approximately 100-fold reduction) compared to iterative low-rank reconstruction methods while maintaining comparable image quality.
[0056] In some implementations, the second and optional additional echoes can be used to estimate and correct phase differences between shots. For example, the second echo can serve as a phase navigator to estimate shot-to-shot phase variations within a single diffusion direction. This phase navigator information can enable rapid Hybrid-SENSE reconstruction with explicit shot-to-shot phase correction, bypassing the need for computationally intensive iterative low-rank multi-shot reconstruction methods (e.g., S-LORAKS). In some implementations, the second and optional additional echoes can be used to estimate motion between acquisitions (e.g., diffusion weights and directions) that can be applied to correct the images generated using data acquired during the first echo. In this way, the reconstruction may include correction of shot-to-shot phase differences or motion artifacts in the lst-echo data.
[0057] In some implementations, the reconstruction may also include in-plane unaliasing reconstruction (e.g., SENSE, GRAPPA, and so forth), slice unaliasing (e.g., SMS, slice-GRAPPA, and so forth), super-resolution reconstruction (e.g., gSLIDER), or a combination thereof. The reconstruction may also include other processing steps, suchAtty. Dkt. No. 125141.05023. MGH2024-029as partial Fourier reconstruction, denoising, filtering, and so forth.
[0058] In some implementations, the reconstruction may also include processing the multi-contrast data to estimate relaxation parameters or relaxometry maps. For example, the relaxation parameters can be estimated based on the signal decay between data acquired during the first and second echoes. Reconstruction may also include processing the diffusion weighted images to generate diffusion weighted data (e.g., apparent diffusion coefficients, fractional anisotropy, mean diffusivity, radial diffusivity, and so forth) based on various diffusion models (e.g., diffusion tensor imaging (DTI), diffusion weighted imaging (DWI), neurite orientation dispersion and density imaging (NODDI), soma and neurite density imaging (SANDI), neurite exchange imaging (NEXI), and so forth). Reconstruction may also include processing the diffusion relaxometry data using a diffusion relaxometry model, such as a white matter diffusion model (e.g., MTE-SMT) to provide diffusion relaxometry parameter maps (e.g., intra-axonal signal fraction, intra-axonal T2, extra-cellular T2, parallel diffusivity, extra-cellular perpendicular diffusivity, and extra-cellular signal fraction). In some implementations, the parallel diffusivity estimation may assume that parallel diffusivities are the same in intra-axonal and extra-cellular spaces. Reconstruction may also include diffusion fractography analysis to generate fiber tract maps, including analysis of short-range association fibers such as U-fibers for connectivity analysis. Advantageously, the reconstruction may also include methods for gray matter (GM) analysis that are historically limited by the low-resolution images achievable with current imaging methods. In this way, the improved resolution and quality provided by the presently described methods advantageously improves ability to characterize gray matter.
[0059] The reconstructed images, which may include diffusion weighted images, diffusion tensor images, diffusion data, relaxation parameter maps, diffusion relaxometryAtty. Dkt. No. 125141.05023. MGH2024-029parameter maps, diffusion tractography maps, and so forth, can be displayed or stored for future use in process block 114.
[0060] Now referring to FIGS. 2A and 2B, pulse sequence diagrams are provided that illustrate non-limiting example implementations of the described method. As shown in FIG. 2 A, a 90° excitation pulse 202 is applied with MB factor of 2 and an RF slab encoding scheme. Bipolar diffusion gradients 204 are applied around a 180° refocusing pulse 206 to produce a first echo 208. The first echo is acquired with high resolution and an effective echo spacing defined as ESPe^ective l=ESPno”imal>1Another 180° refocusing pulse 210 is applied to produce a second echo 212. The second echo 212 is acquired with an effective echo spacing defined as ESPeffective 2=ESPno^-mal’2, which matches the effective echo spacing of the first echo 208 ESPeffective l= ESPe^ective 2In one nonlimiting example implementation, the second echo 212 can be acquired with lower resolution than that of the first echo 208 to produce a high-quality field map that is applied to correct images reconstructed from data acquired during the first echo 208, which is referred to as "target DWI acquisition” in FIG. 2A. In another non-limiting example implementation, the second echo 212 can be acquired with high resolution to provide additional target DWI acquisition data (e.g., a second image with a second contrast). Both the first and second echoes 208 and 212 are first acquired in a first shot 214 with a first PE gradient direction (e.g., blip-up). The acquisition block is repeated in a second shot 216 with a second PE gradient direction that is opposite that of the first shot 214 (e.g., blip-down). In this example implementation, the pulse sequence also includes optional fat saturation 218 applied at the beginning of the sequence block.
[0061] FIG. 2B illustrates how each shot of the conventional BUDA gSlider acquisition includes a period of dead time 250 that is inserted into the sequence to reduceAtty. Dkt. No. 125141.05023. MGH2024-029spin-history effects between shots. The proposed methods take advantage of this dead time to produce a second echo and acquire a second readout 252 in each shot, as previously described. This additional data can be used to improve the quality of the field maps estimated from paired blip-up and blip-down data or to provide additional information (e.g., diffusion relaxometry data).
[0062] Examples
[0063] Example 1
[0064] The example described below provides non-limiting example implementations of the described phase-reversed interleaved multi-echo acquisition (PRIME) sequence that enables highly acceleration, high-resolution, and distortion-free dMRI using an additional echo without prolonging scan time when gSlider encoding is utilized. As previously described, the PRIME acquisition includes a first echo for target diffusion-weighted imaging (DWI) acquisition with high-resolution and a second echo acquired with either 1) lower-resolution for high-fidelity field map estimation, or 2) matching resolution to enable efficient diffusion relaxometry acquisitions.
[0065] In this example study, the sequence was evaluated on in vivo data acquired from healthy volunteers on clinical and Connectome 2.0 scanners. In vivo experiments demonstrated that 1) high in-plane acceleration (Rin-piane of 5-fold with 2D partial Fourier) was achieved using the high-fidelity field maps estimated from the second echo, which was made at a lower resolution and acceleration to increase its SNR while matching the effective echo spacing of the first readout, 2) high-resolution diffusion relaxometry parameters were estimated from dual-echo PRIME data using a white matter model of multi-TE spherical mean technique (MTE-SMT), and 3) high-fidelity mesoscale DWI at 550 pm isotropic resolution could be obtained in vivo by capitalizing on the high-performance gradients of the Connectome 2.0 scanner.Atty. Dkt. No. 125141.05023. MGH2024-029
[0066] Methods
[0067] FIG. 2A illustrates the pulse sequence diagram of PRIME. The sequence includes two spin echoes where the first echo is for target diffusion-weighted imaging (DWI) acquisition with high-resolution and the second echo can be acquired with either 1) low-resolution for field map estimation, or 2) high-resolution for the second echo of target DWI acquisition. The effective ESP (defined as nominal ESP divided by Rin -plane) of the first and second echo readouts was matched to have the same geometric distortion. gSlider RF encoding was used to acquire multiple volumetric slabs and perform superresolution reconstruction to resolve thin slices for high isotropic resolution dMRI, which was combined with blipped-controlled aliasing in parallel imaging (blipped-CAIPI) SMS technique. Blip-up and -down shots with reversed phase-encoding directions were acquired for each diffusion direction to correct geometric distortions of dMRI caused by Bo inhomogeneity and eddy currents. Five gSlider RF encoding pulses were designed based on the gSlider RF encoding Toolbox, where 10 shots (5 gSlider RF encodings x blip-up / -down shots) were acquired per volume. Variable-rate selective excitation (VERSE) technique was used to mitigate specific absorption rate (SAR) constraints. The open-source MR sequence development framework, Pulseq, was used for the sequence implementation.
[0068] FIG. 2B shows the sequence timing comparison between the conventional BUDA-EP1 and PRIME. With Rtn-piane of 5-fold at 1 x 1 mm2, ESP = 0.70 ms, multi-band (MB) factor of 2, and 2D pF of 6 / 8 x 6 / 8 on a clinical scanner, single-echo acquisition of BUDA-EPI with gSlider RF encoding has long dead time in each TR (e.g., around 182 ms out of TRsiice = 270 ms in this case), to mitigate spin history-related slab boundary artifacts. PRIME utilizes this time by inserting additional RF encoding and readout without prolonging TR.Atty. Dkt. No. 125141.05023. MGH2024-029
[0069] To demonstrate the signal level ofthe second echo of PRIME; a comparison between the second echo images of PRIME and the single-echo images separately acquired with the same TE was conducted. DW1 data were reconstructed for b = 0 and 1,000 s / mm2with Rin-piane x MB = 4 x 2 and 2D pF (6 / 8 x 6 / 8] at 1 x 1 x 5 mm3slab, where TE = 118 ms. The signal norm ofthe second echo image compared to the singleecho image was 0.97 and 1.01 for b = 0 and 1,000 s / mm2, respectively, which demonstrates that signal levels of PRIME were comparable with those of single-echo images that were separately acquired with another scan.
[0070] The representative schematic k-space trajectories of PRIME are presented in FIGS. 3A-3B. For high-fidelity field maps estimation (FIG. 3A), the first echo was acquired with Rin-piane of 5 -fold at 1 x 1 mm2and D pF on a clinical scanner, whereas the second echo was acquired with the same effective ESP as the first echo for Rin-piane of 4-fold at 2 x 2 mm2and ID pF. High in-plane acceleration with 2D pF can mitigate relaxation-related voxel blurring with reduced ESP and TE. The blip-up and -down shots cover complementary k-space positions, and all quadrants of k-space can be covered with the aid of the virtual coil constraints in S-LORAKS.
[0071] For high -resolution diffusion relaxometry (FIG. 3B], both echoes had the same resolution of 1.1 x 1.1 mm2and Rin-piane of 4-fold with 2D pF to match the voxel size and distortion. With the high gradient strength ofthe Connectome 2.0 scanner, mesoscale diffusion imaging of 550 x 550 pm2resolution Rin-piane of 4-fold with 2D pF was achieved, where the second echo has the resolution of 1.38 x 1.38 mm2and Rin-piane of 3-fold with 2D pF.
[0072] The top panel of FIG. 4 illustrates the standard single-echo BUDA-EPI reconstruction pipeline at Rin-piane x MB = 5 x 2 with 2D pF of 6 / 8 x 6 / 8 at 1 x 1 x 5 mm3slab resolution, where interim SENSE reconstructed blip-up and -down images wereAtty. Dkt. No. 125141.05023. MGH2024-029fed into FSL TOPUP to estimate the field maps. However, due to potential aliasing and g-factor loss at 10-fold total acceleration, the estimated field maps suffered from artifacts, as indicated by the arrows. The bottom panel of FIG. 4 shows that PRIME, which has a dual-echo reconstruction pipeline, utilized the second echo acquired at a lower resolution and in-plane acceleration (Rm-piane x MB = 4 x 2 with 2D pF of 6 / 8 x 6 / 8) to boost SNR and mitigate artifacts, enabling higher fidelity of field maps estimation. The two high- resolution shots from the first echo were then reconstructed using S-LORAKS with the estimated field maps. The difference images between BUDA-EPI and PRIME show that residual aliasing artifacts in the estimated field maps from BUDA-EPI can hamper distortion correction in the final reconstructed images.
[0073] The blip-up and -down shots of the first echo image were jointly reconstructed using a model-based approach with low-rank-based regularization (S-LORAKS) along with phase prior, where the estimated field maps are incorporated in the forward operator. Each gSlider RF-encoded slab was reconstructed based on S-LORAKS, and the volume images with the isotropic resolution were reconstructed by resolving RF-encoded slabs with Bi+and Ti corrections.
[0074] To mitigate the motion between the volume images across diffusion directions, especially for mesoscale in vivo dMRl of 550 pm isotropic resolution, FSL MCFLIRT was used for registration of the diffusion volume images. Marchenko-Pastur principal component analysis (MPPCA) algorithm was applied using MRtrix3 software to denoise the mesoscale data with a patch size of 5. Fractional anisotropy (FA) maps were obtained using FSL DTIF1T.
[0075] To validate the DWI reconstruction performance of PRIME, we compared it with SENSE reconstructed blip-up and -down images, hybrid SENSE, and BUDA-EPI. Hybrid SENSE used the field map and shot-to-shot phase difference estimated fromAtty. Dkt. No. 125141.05023. MGH2024-029SENSE reconstructions of blip-up and -down shots. Hybrid SENSE and BUDA-EPI utilized the field maps for their reconstruction, which were estimated from the first echo images, whereas PRIME used the field maps estimated from the second echo images.
[0076] In addition, to evaluate the reconstructed images with a quantitative metric, the 'reference DWIs’ were acquired using Rin-plane = 5 (echo #1] with 5 shots and 3 (echo #2) with 3 shots for each blip-up and -down acquisitions, and they were reconstructed using S-LORAKS with the estimated field maps from the second echo. A root mean square error (RMSE) metric was used for the evaluation of the reconstructed DWIs using Rin-piane = 5 (echo #1) with 2 shots.
[0077] To demonstrate the PRIME sequence used for high-resolution diffusion relaxometry mapping, we utilized a two-compartment tissue model with the spherical mean technique (SMT), while extending the model by accounting for the compartmental T2 values. This multi-echo SMT (MTE-SMT) model is applicable to study WM microstructure and composition in clinical settings. The MTE-SMT model is included in the package, Microstructure.jl, which was developed in Julia programming language and designed for fast and probabilistic microstructure imaging. Modeling and fitting assumptions for parameter estimation can be flexibly adjusted in Microstructure.jl. In this 2 2 2example, we fixed parallel diffusivities in both tissue compartments to 1.7 pm2 / ms for in vivo imaging in the WM and estimated tissue parameters of interest, including the intra-axonal signal fraction (fta°), intra-axonal T2 (T2ia), extra -cellular T2 (T2ec), extra-cellular perpendicular diffusivity (D^J; the extra-cellular signal fraction (fee0] was derived as 1 — f£i.°a-
[0078] The experiments were conducted with the approval of the Institutional Review Board. Five healthy volunteers were scanned on 3T MAGNETOM Prisma (Siemens Healthineers, Erlangen, Germany) scanner with a 32ch head receive array andAtty. Dkt. No. 125141.05023. MGH2024-029Connectome 2.0 (Siemens Healthineers, Erlangen, Germany) scanner with a 72 ch head receive array. The imaging parameters for the PRIME sequence are presented in Table 1, and Bi+maps were separately acquired usingaturbo-FLASH sequence with matched FOV at the resolution of 3.44 x 3.44 x 5.5 mm3to correct the imperfect gSlider RF slab profile due to Bi+inhomogeneity. For the MTE-SMT diffusion model, additional b = 0 s / mm2images were acquired with different TEs - 40, 50, 70, 80, 90, 100, 140, and 160 ms.
[0079] Table 1Atty. Dkt. No. 125141.05023. MGH2024-029
[0080] Results
[0081] FIG. 5 shows the reconstructed DWIs for b = 1,000 s / mm2with Rin-plane — 5 and 1D pF (6 / 8) at 1 x 1 x 4 mm3resolution. SENSE reconstructed blip-up and -down images show residual aliasing artifacts and amplified noise, andhybrid-SENSE still suffers from noise implications. BUDA-EPI shows better performance than those SENSE-based methods, whereas PRIME outperforms BUDA-EPI and other methods, as shown in the difference images, and has the lowest RMSE values. The estimated field maps using BUDA-EPI and PRIME are presented in FIG. 6. The reference field maps were estimated from the blip-up and -down images acquired using Rin-piane = 3 (echo #2) with 3 shots. PRIME enabled higher fidelity of field maps estimation by utilizing the second echo acquired at a lower resolution and in-plane acceleration of Rin-piane = 3 to boost SNR and mitigate artifacts.
[0082] Similar results were observed with gSlider encoding and b = 1,000 s / mm2with Rm-plane x MB = 5 x 2 and 2D pF (6 / 8 x 6 / 8) at 1 x 1 x 5 mm3slab. SENSE reconstructions on the individual shots suffered from noise amplification and artifacts due to the 10-fold acceleration. Hybrid-SENSE had residual artifacts at this acceleration. While BUDA-EPI and PRIME showed improved image quality, the imperfect field mapsAtty. Dkt. No. 125141.05023. MGH2024-029estimation of BUDA-EPI from the SENSE images that have residual aliasing artifacts lead to distortion differences between BUDA-EPI and PRIME.
[0083] The reconstructed dual-echo DW'Is, for b = 0 and 1,000 s / mm2with Rin-piane x MB = 4 x 2 and 2D pF (6 / 8 x 6 / 8) at 1 x 1 x 5 mm3slab, are presented in FIG. 7. The first echo was acquired with TEi = 53 ms and ESPi = 0.70 ms, whereas the second echo was acquired with TEz = 106 ms and ESPz = 0.70 ms. Both b = 0 and b = 1,000 s / mm2DWIs for both echoes show high geometric fidelity and SNR with high in-plane resolution.
[0084] The high-resolution diffusion relaxometry imaging obtained using the MTE- SMT diffusion model from dual-echo DWIs, which were acquired with 1.1 mm isotropic resolution with Rin-piane x MB = 4 x 2 and 2D pF (6 / 8 x 6 / 8) for 2-shell acquisitions of b = 0, 750, and 2,000 s / mm2, is demonstrated in FIG. 8. With the dual-echo data from PRIME, MTE-SMT provided compartment-specific T2 and non-Tz -weighted compartmental signal fraction maps from the intra-axonal and extra-cellular tissue compartments in the WM. This demonstrated the ability to use PRIME for efficient and high-resolution combined diffusion relaxometry mapping.
[0085] FIG. 9 shows the mesoscale averaged DWIs and FA maps for b = 1,000 s / mm2with Rin-piane x MB = 4 x 2 and 2D pF (6 / 8 x 6 / 8) at 550 pm isotropic resolution. By capitalizing on the Connectome 2.0 gradients, TEi = 43 ms could be achieved, and second echo was obtained with TE2 = 112 ms using Rin-piane x MB = 3 x 2 and 2D pF (6 / 8 x 6 / 8) at 1.38 x 1.38 x 0.55 mm3resolution, which were utilized for high-fidelity field maps estimation. The high-fidelity mesoscale dMRI could be achieved with a single PRIME data acquisition (NEX = 1) of 90 directions and MPPCA denoising. A comparison between the acquired and denoised single DWIs for mesoscale data showed that MPPCA improved the SNR of a single DWI while preserving structure information.
[0086] This study developed and tested a new highly accelerated distortion-freeAtty. Dkt. No. 125141.05023. MGH2024-029dMRI sequence, PRIME, which benefits from incorporating an additional echo without prolonging TR, especially when gSlider RF encoding was used. Through in vivo experiments, we demonstrated that PRIME could be utilized for the following cases as non-limiting examples: 1) high-fidelity DWIs acquisition with a high acceleration factor, 2) high-resolution diffusion relaxometry imaging using MTE-SMT, and 3) mesoscale DWIS of 550 pm isotropic resolution for in vivo data.
[0087] While BUDA-EPI with gSlider RF encoding has demonstrated that high- fidelity of high-resolution dMRI could be achieved, it had limited in-plane acceleration factor per shot and substantially long dead time due to long TR used to accommodate gSlider RF encoding. In particular, residual artifacts in the SENSE reconstructed images were propagated to the estimated field maps when high acceleration per shot was used [e.g,, Rm-piaae x MB = 5 x 2), while PRIME produced high-fidelity field maps from the additional low-resolution second echo images without incurring a scan time penalty by using the dead time in each TR. By using VERSE technique, dual-echo PRIME was achieved without SAR constraints where a VERSE factor of 2.2 was used for MB = 2. PRIME may also employ more echoes (e.g., triple-echo PRIME), which may have implications on SAR levels.
[0088] Multi-shot EPI acquisition reduces voxel blurring and geometric distortion by reducing the number of k-space readouts in each shot; however, shot-to-shot phase variations due to physiological noise make it challenging to combine the shots. One of the navigator-free methods, S-LORAKS, which uses low-rank modeling of local k-space neighborhoods with smooth phase prior, could reconstruct the high-fidelity DWIs by combining two shots of blip-up and -down images with the incorporated field maps obtained from the second echo. While S-LORAKS could provide high-fidelity images, it required substantial reconstruction time, which took about 1 h per volume on MATLABAtty. Dkt. No. 125141.05023. MGH2024-029using parallel cluster computations. Utilizing the second echo data as a navigator for phase correction between shots can potentially make the reconstruction faster by avoiding intensive low-rank-based computations.
[0089] PRIME showed comparable signal levels of the second echo images with those of single-echo images that were separately acquired with another scan. Using a high time-bandwidth product of the gSlider RF encoding and non-Carr-Purcell-Meiboom-Gill (CPMG) condition with orthogonal crusher gradients could achieve comparable signal levels of the second of PRIME with those of single-echo. This enabled high-resolution dualecho acquisition of PRIME, which could be utilized for diffusion relaxometry mapping with diffusion models.
[0090] In PRIME experiments, the effective ESP, the first and second echo readouts was matched to have the same geometric distortion. The effective ESP of the second echo does not necessarily have to be the same as the first echo as long as high fidelity field maps can be estimated. Matching the effective ESP of the first and second echo readouts limited the in-plane acceleration of the second echo on the clinical scanner. For instance, the Rm- plane for the second echo was constrained to be 4-fold with ID pF on the clinical scanner when the Rin-plane of 5-fold was used for the first echo to match the effective ESP. Using arbitrary ESP for the second echo under further validation ofthe effective ESP for the field map estimation can enable a more flexible combination of R -piane for the first and second echo readouts.
[0091] Recent advances in MRI hardware, such as Connectome 2.0 and MAGNUS scanners, enable the design of more advanced dMRI sequences using strong gradient coils. By capitalizing on the Connectome 2.0 gradients that have a maximum gradient strength of 500 mT / in and a maximum slew rate of 600 T / m / s, we could achieve 8 / A = 2.2 / 17,3 ms for b = 1,000 s / mm2, TEi = 43 ms, TEz = 112 ms, and effective ESP = 0.14 msAtty. Dkt. No. 125141.05023. MGH2024-029for mesoscale 550 pm isotropic resolution. To achieve the same resolution on a clinical scanner (e.g., a 3T MAGNETOM Prisma scanner, which has a maximum gradient strength of 80 mT / m and a maximum slew rate of 200 T / m / s), we need 8 / A = 10.0 / 35.3 ms for b = 1,000 s / mm2, TEi = 79 ms, TE2 = 189 ms, and effective ESP = 0.24 ms for mesoscale 550 pm isotropic resolution. Thus, the Connectome 2.0 scanner allowed us to reduce TE1 / TE2 by about 45,6 / 40.7% and effective ESP by about 41.7%. For a voxel that has T2 = 65 ms, for instance, this corresponds to a 74.0% SNR gain thanks to the reduced TE and 30.9% SNR loss due to the increased bandwidth, thereby resulting in a net 32.9% SNR gain. With short TE and ESP, high-fidelity mesoscale DWIs that had high SNR and less geometric distortion could be obtained usingthe PRIME sequence and reconstruction scheme. Here, we used Rin-piane of 4-fold instead of 5-fold to retain SNR for mesoscale diffusion imaging. Increasing the in-plane acceleration can reduce TE by reducing the number of frequency -encoding lines while it reduces Fourier encoding as well; thus, there is a trade-off between TE reduction and decreased noise averaging due to Fourier encoding.
[0092] Though diffusion models can extract microstructure features of the human brain from dMRI, the resolutions of in vivo data are mostly limited to 1.5~ 2.5 mm isotropic resolution. Here, we demonstrated that the PRIME sequence enabled high- resolution (i.e., 1.1 mm isotropic resolution with a clinical scanner) diffusion relaxometiy using the WM diffusion model, MTE-SMT, which provided diffusion relaxometry parameters, including intra-axonal signal fraction, intra-axonal T2, extra-cellular T2, extra-cellular perpendicular diffusivity, and extra-cellular signal fraction. In order to retain SNR for diffusion relaxometry mapping, especially for b = 2,000 s / mm2, we demonstrated the results on 1,1 mm isotropic resolution, which has around 1.33 larger voxel size (i.e., around 1.77 signal averages in terms of SNR) than 1 mm isotropic resolution. PRIME lends itself to more advanced diffusion models, such as neuriteAtty. Dkt. No. 125141.05023. MGH2024-029orientation dispersion and density imaging (N0DD1), soma and neurite density imaging (SANDI), neurite exchange imaging (NEXI), and their multi-TE extensions. In particular, PRIME may enable advanced gray matter (GM) analysis, where current GM models use around 2 mm isotropic resolutions, which are too coarse to image the cortex (1—4 mm thickness) due to severe partial volume effects.
[0093] PRIME took advantage of long TR, which was needed for gSlider RF encoding to mitigate spin history-related slab boundary artifacts. Though Bi+and Ti corrections could reduce the slab boundary artifacts, PRIME showed residual artifacts with TR of around 3.5-4 s. Increasing TR may mitigate the slab boundary artifacts while reducing the scan efficiency and allowing more dead time. For mesoscale dMRI, long scan time mightmake itvulnerable to motion across the diffusion directions, and volume image registration may be advantageous, especially for temporal and occipital lobes which are sensitive to head motion. Reducing the number of diffusion directions using gSlider-spherical ridgelets (gSlider-SR) can be another way to reduce the total scan time while maintaining SNR and angular information. Super-resolution techniques with rotating views can be alternative strategies to achieve high-resolution dMRI while minimizing slab boundary artifacts.
[0094] Example 2
[0095] The following non-limiting example demonstrates how the described PRIME acquisition method benefits from inserting additional echoes without incurring a scan time penalty when gSlider RF encoding is used, as shown in FIG. 10. The additional second (and optional third or more) echoes can be utilized in several ways, including: 1) high-fidelity field maps estimation, 2) phase navigator for shot-to-shot phase correction, 3) motion navigator for motion estimation across diffusion directions, and 4) multi-echo DWIs acquisition. In this example, we demonstrated that highly accelerated, high-Atty. Dkt. No. 125141.05023. MGH2024-029resolution, and distortion-free dMRI was achieved using the PRIME sequence on clinical and Connectome 2.0 scanners.
[0096] Methods
[0097] FIG. 10 shows the pulse sequence diagram of the example implementation of PRIME. PRIME has two or three spin echoes, where the first echo provides target DWI acquisition with high-resolution and the second / third echoes are acquired to provide reference correction data (e.g., field or distortion maps, phase navigator, motion navigator) or to provide multi-echo target DWI acquisition. The effective ESP (defined as nominal ESP / Rin-plane) of each echo’s readout is matched to have the same geometric distortion. gSlider RF encoding is used to acquire multiple volumetric slabs and perform super-resolution reconstruction to resolve thin slices for high isotropic resolution dMRI, combined with blipped-CAIPI SMS. Blip-up / -down shots with reversed phase-encoding directions are acquired for each diffusion direction to correct geometric distortions of dMRI. VERSE is used to mitigate SAR constraints. The open-source framework, Pulseq, is used for the sequence implementation.
[0098] In FIG. 10, the bottom panel demonstrates a sequence timing comparison between conventional BUDA-EPI and PRIME. Single-echo BUDA-EPI with gSlider RF encoding has a long dead time in each TR to mitigate spin history-related slab boundary artifacts, whereas PRIME utilizes this time by inserting additional RF encoding and readout without prolonging TR.
[0099] PRIME utilizes the second echo acquired ata lower resolution and in-plane acceleration to boost SNR and mitigate artifacts, enabling higher fidelity of field maps along with shot-to-shot phase difference estimation. The two high-resolution shots from the first echo are then reconstructed using hybrid SENSE with the estimated field maps and phase differences. The second echo can also be used for motion estimation acrossAtty. Dkt. No. 125141.05023. MGH2024-029diffusion directions, which then can be applied to the first echo correction. If second / third echoes have the same resolution, e.g., diffusion relaxometry acquisition, as the first echo, each echo was reconstructed using S-LORAKS with the estimated field maps from the first echo.
[0100] 3T Siemens Prisma: a healthy volunteer was scanned using 32ch head receive array: FOV = 220 x 220 x 130 mm3, TR / TE1 / TE2 = 4850 / 51 / 102 ms, 1 x 1 x4mm3resolution, b = 1,000 s / mm2, 32 directions. Reference DWIs were acquired using Rin-plane= 5 (echo#1, 5-shots) and 3 (echo#2, 3-shots) for each blip-up / -down acquisition, while only 2 -shots were used for the actual reconstruction for each echo.
[0101] 3T Siemens Connectome 2.0: two healthy volunteers were scanned using 72ch head receive array, where Gmax = 500 mT / m and slew-rate = 600 T / m / s were used. For triple-echo DWIS, FOV = 220 x 220 x 130 mm3, TR / TE1 / TE2 / TE3 = 3500 / 40 / 83 / 124 ms, ESP1 / ESP2 / ESP3 = 0.41 ms, Rin-pianeX MB = 3 x 2 with 2D pF (6 / 8 x 6 / 8), 1 mm isotropic resolution using gSlider, b = 750 / 1,200 / 2,000 s / mm2, 16 / 31 / 47 directions for each shell. For mesoscale DWIs, FOV = 220 x 220 x 113 mm3, TR / TE1 / TE2 = 4500 / 48 / 128 ms, ESP1 / ESP2 = 0.60 / 0.30 ms, Rin-piane x MB = 4 x 2 with 2D pF (6 / 8 x 6 / 8), 490pm isotropic resolution using gSlider, b = 1,000 s / mm2, 90 directions. Similar parameters were used for 680 pm resolution.
[0102] Results
[0103] FIG. 11 demonstrates SENSE reconstructed images show residual aliasing artifacts and amplified noise. BUDA-EPI shows better performance than those SENSE-based methods, whereas PRIME outperforms BUDA-EPI and others and has the lowest RMSE values. PRIME enabled higher fidelity in field maps estimation by utilizing the second echo acquired at a lower resolution and in-plane acceleration to boost SNR and mitigate artifacts. In particular, FIG. 11 shows reconstructed DWIs and field maps usingAtty. Dkt. No. 125141.05023. MGH2024-0292 shots for b = 1,000 s / mm2with Rm-piane = 5 and 1D pF (6 / 8) at 1 x 1 x 4 mm3resolution using SENSE, blip-up and -down (BUDA)-EPI, and PRIME. The reference DWIs and field maps were acquired using Rin-piane = 5 (echo #1) with 5 shots and 3 (echo #2) with 3 shots for each blip-up and -down acquisitions, and they were reconstructed using S-LORAKS with the estimated field maps from the second echo. An RMSE metric was used to evaluate the reconstructed DWIs and field maps using Rin-plane = 5 (echo #1) with 2 shots.
[0104] FIG. 12 demonstrates a single-direction DWI for b = 2,000 s / mm2on a Connectome 2.0 scanner. The estimation of field maps and phase correction from the second echo enabled high geometric fidelity and SNR. In particular, FIG. 12 shows reconstructed DWIs, for b = 1,000 s / mm2with Rin-piane * MB = 4 x 2 and 2D pF (6 / 8 x 6 / 8) at 0.68 x 0.68 x 3.44 mm3slab. The second echo acquired with Rin-piane x MB = 2 x 2 and 2D pF (6 / 8 x 6 / 8) at 2 x 2 x 3.44 mm3slab was used for field maps estimation and phase correction.
[0105] FIG. 13 shows the reconstructed triple-echo DWIs, for 3-shell acquisition at 1 mm isotropic resolution with TEs = 40 / 83 / 124 ms, which can be applied to high-resolution diffusion relaxometry imaging. In particular, FIG. 13 shows reconstructed triple-echo DWIs, for 3-shell acquisition (b = 0, 750, 1,200, and 2,000 s / mm2) with Rin-piane x MB = 4 x 2 and 2D pF (6 / 8 x 6 / 8) at 1 mm isotropic resolution.
[0106] FIG. 14 depicts the mesoscale averaged DWIs for b = 1,000 s / mm2at 490 pm isotropic resolution. Using PRIME and capitalizing on the Connectome 2.0 gradients, TE1 / TE2 = 48 / 128 ms could be achieved with mesoscale resolution. In particular, FIG. 14 shows mesoscale averaged DWIs for b = 1,000 s / mm2with Rin-piane x MB = 4 x 2 and 2D pF (6 / 8 x 6 / 8) at 490 pm isotropic resolution.
[0107] We developed a highly accelerated distortion -free dMRI sequence, PRIME, which benefits from incorporating additional echoes without prolonging TR, especiallyAtty. Dkt. No. 125141.05023. MGH2024-029when gSlider RF encoding was used. We demonstrated that PRIME could be utilized for at least the following cases: 1) high-fidelity DWIs acquisition with field maps estimation and phase / motion correction using additional echo, 2 high-resolution multi-echo DWIs, and 3) mesoscale DWIs of 490 pm isotropic resolution for in vivo data.
[0108] Example System
[0109] Referring particularly now to FIG. 15, an example of an MRI system 1500 that can implement the methods described herein is illustrated. The MRI system 1500 includes an operator workstation 1502 that may include a display 1504, one or more input devices 1506 (e.g., a keyboard, a mouse), and a processor 1508. The processor 1508 may include a commercially available programmable machine running a commercially available operating system. The operator workstation 1502 provides an operator interface that facilitates entering scan parameters into the MRI system 1500. The operator workstation 1502 may be coupled to different servers, including, for example, a pulse sequence server 1510, a data acquisition server 1512, a data processing server 1514, and a data store server 1516. The operator workstation 1502 and the servers 1510, 1512, 1514, and 1516 may be connected via a communication system 1540, which may include wired or wireless network connections.
[0110] The MRI system 1500 also includes a magnet assembly 1524 that includes a polarizing magnet 1526, which maybe a high-field magnet. The polarizing magnet 1526 produces a main magnetic field (Bo). The magnet assembly 1524 may also include shimming coils that can shape Bo. The MRI system 1500 may optionally include a wholebody RF coil 1528 and a gradient system 1518 that controls a gradient coil assembly 1522.
[0111] The pulse sequence server 1510 functions in response to instructions provided by the operator workstation 1502 to operate a gradient system 1518 and aAtty. Dkt. No. 125141.05023. MGH2024-029radiofrequency (" RF") system 1520. Gradient waveforms for performing a prescribed scan are produced and applied to the gradient system 1518, which then excited gradient coils in an assembly 1522 to produce the magnetic field gradients (e.g., Gx, Gy, and Gz) that can be used for spatially encoding magnetic resonance signals. The gradient coil assembly 1522 forms part of a magnet assembly 1524 that includes a polarizing magnet 1526 and a whole-body RF coil 1528.
[0112] RF waveforms are applied by the RF system 1520 to the RF coil 1528, or a separate local coil to perform the prescribed magnetic resonance pulse sequence. Responsive magnetic resonance signals detected by the RF coil 1528, or a separate local coil, are received by the RF system 1520. The responsive magnetic resonance signals may be amplified, demodulated, filtered, and digitized under direction of commands produced by the pulse sequence server 1510. The RF system 1520 includes an RF transmitter for producing a wide variety of RF pulses used in MRI pulse sequences. The RF transmitter is responsive to the prescribed scan and direction from the pulse sequence server 1510 to produce RF pulses of the desired frequency, phase, and pulse amplitude waveform. The generated RF pulses may be applied to the whole-body RF coil 1528 or to one or more local coils or coil arrays.
[0113] The RF system 1520 also includes one or more RF receiver channels. An RF receiver channel includes an RF preamplifier that amplifies the magnetic resonance signal received by the coil 1528 to which it is connected, and a detector that detects and digitizes the I and Q quadrature components of the received magnetic resonance signal. The magnitude of the received magnetic resonance signal may, therefore, be determined at a sampled point by the square root of the sum of the squares of the I and Q components:M = √(I2+ Q2)Atty. Dkt. No. 125141.05023. MGH2024-029
[0114] and the phase of the received magnetic resonance signal may also be determined according to the following relationship:
[0115] The pulse sequence server 1510 may receive patient data from a physiological acquisition controller 1530. By way of example, the physiological acquisition controller 1530 may receive signals from a number of different sensors connected to the patient, including electrocardiograph (" ECG” signals from electrodes, or respiratory signals from a respiratory bellows or other respiratory monitoring devices. These signals maybe used by the pulse sequence server 1510 to synchronize, or "gate,” the performance of the scan with the subject’s heartbeat or respiration.
[0116] The pulse sequence server 1510 may also connect to a scan room interface circuit 1532 that receives signals from various sensors associated with the condition of the patient and the magnet system. Through the scan room interface circuit 1532, a patient positioning system 1534 can receive commands to move the patient to desired positions during the scan.
[0117] The digitized magnetic resonance signal samples produced by the RF system 1520 are received by the data acquisition server 1512. The data acquisition server 1512 operates in response to instructions downloaded from the operator workstation 1502 to receive the real-time magnetic resonance data and provide buffer storage, so that data are not lost by data overrun. In some scans, the data acquisition server 1512 passes the acquired magnetic resonance data to the data processor server 1514. In scans that require information derived from acquired magnetic resonance data to control the further performance of the scan, the data acquisition server 1512 may be programmed to produce such information and convey it to the pulse sequence server 1510. ForAtty. Dkt. No. 125141.05023. MGH2024-029example, during pre-scans, magnetic resonance data may be acquired and used to calibrate the pulse sequence performed by the pulse sequence server 1510. As another example, navigator signals may be acquired and used to adjust the operating parameters of the RF system 1520 or the gradient system 1518, or to control the view order in which k-space is sampled. In still another example, the data acquisition server 1512 may also process magnetic resonance signals used to detect the arrival of a contrast agent in a magnetic resonance angiography (" MRA") scan. For example, the data acquisition server 1512 may acquire magnetic resonance data and processes it in real-time to produce information that is used to control the scan.
[0118] The data processing server 1514 receives magnetic resonance data from the data acquisition server 1512 and processes the magnetic resonance data in accordance with instructions provided by the operator workstation 1502. Such processing may include, for example, reconstructing two-dimensional or three-dimensional images by performing a Fourier transformation of raw k-space data, performing other image reconstruction algorithms (e.g., iterative or backprojection reconstruction algorithms), applying filters to raw k-space data or to reconstructed images, generating functional magnetic resonance images, or calculating motion or flow images.
[0119] Images reconstructed by the data processing server 1514 are conveyed back to the operator workstation 1502 for storage. Real-time images may be stored in a data base memory cache, from which they may be output to operator display 1502 or a display 1536. Batch mode images or selected real time images may be stored in a host database on disc storage 1538. When such images have been reconstructed and transferred to storage, the data processing server 1514 may notify the data store server 1516 on the operator workstation 1502. The operator workstation 1502 may be used byAtty. Dkt. No. 125141.05023. MGH2024-029an operator to archive the images, produce films, or send the images via a network to other facilities.
[0120] The MRI system 1500 may also include one or more networked workstations 1542. For example, a networked workstation 1542 may include a display 1544, one or more input devices 1546 (e.g., a keyboard, a mouse), and a processor 1548. The networked workstation 1542 may be located within the same facility as the operator workstation 1502, or in a different facility, such as a different healthcare institution or clinic.
[0121] The networked workstation 1542 may gain remote access to the data processing server 1514 or data store server 1516 via the communication system 1540. Accordingly, multiple networked workstations 1542 may have access to the data processing server 1514 and the data store server 1516. In this manner, magnetic resonance data, reconstructed images, or other data may be exchanged between the data processing server 1514 or the data store server 1516 and the networked workstations 1542, such that the data or images may be remotely processed by a networked workstation 1542.
[0122] Referring now to FIG. 16, an example of an MRI system 1600 is shown, which may be used in accordance with some aspects of the systems and methods described in the present disclosure. As shown in FIG. 16, a computing device 1650 can receive one or more types of data (e.g., signal evolution data, k-space data, receiver coil sensitivity data) from data source 1602. In some configurations, computing device 1650 can execute at least a portion of an imaging system 1604 to reconstruct images from magnetic resonance data (e.g., k-space data) acquired using a blip-up and blip-down (e.g., PRIME) technique. In some configurations, the imaging system 1604 can implement an automated pipeline to provide MRI images, diffusion weighted MRI images, diffusionAtty. Dkt. No. 125141.05023. MGH2024-029parameter maps, relaxation parameter maps, diffusion relaxometry parameter maps, and so forth.
[0123] Additionally or alternatively, in some configurations, the computing device 1650 can communicate information about data received from the data source 1602 to a server 1652 over a communication network 1654, which can execute at least a portion of the imaging system 1604. In such configurations, the server 1652 can return information to the computing device 1650 (and / or any other suitable computing device) indicative of an output of the imaging system 1604.
[0124] In some configurations, computing device 1650 and / or server 1652 can be any suitable computing device or combination of devices, such as a desktop computer, a laptop computer, a smartphone, a tablet computer, a wearable computer, a server computer, a virtual machine being executed by a physical computing device, and so on. The computing device 1650 and / or server 1652 can also reconstruct images from the data.
[0125] In some configurations, data source 1602 can be any suitable source of data (e.g., measurement data, images reconstructed from measurement data, processed image data), such as an MRI system, another computing device (e.g., a server storing measurement data, images reconstructed from measurement data, processed image data), and so on. In some configurations, data source 1602 can be local to computing device 1650. For example, data source 1602 can be incorporated with computing device 1650 (e.g., computing device 1650 can be configured as part of a device for measuring, recording, estimating, acquiring, or otherwise collecting or storing data). As another example, datasource 1602 can be connected to computing device 1650 by a cable, a direct wireless link, and so on. Additionally or alternatively, in some configurations, data source 1602 can be located locally and / or remotely from computing device 1650, and canAtty. Dkt. No. 125141.05023. MGH2024-029communicate data to computing device 1650 (and / or server 1652) via a communication network (e.g., communication network 1654).
[0126] In some configurations, communication network 1654 can be any suitable communication network or combination of communication networks. For example, communication network 1654 can include a Wi-Fi network (which can include one or more wireless routers, one or more switches, etc.), a peer-to-peer network (e.g., a Bluetooth network), a cellular network (e.g., a 3G network, a 4G network, etc., complying with any suitable standard, such as CDMA, GSM, LTE, LTE Advanced, WiMAX, etc.), other types of wireless network, a wired network, and so on. In some configurations, communication network 1654 can be a local area network, a wide area network, a public network (e.g., the Internet), a private or semi-private network (e.g., a corporate or university intranet), any other suitable type of network, or any suitable combination of networks. Communications links shown in FIG. 16 can each be any suitable communications link or combination of communications links, such as wired links, fiber optic links, Wi-Fi links, Bluetooth links, cellular links, and so on.
[0127] Referring now to FIG. 17, an example of hardware 1700 that can be used to implement data source 1602, computing device 1650, and server 1652 in accordance with some configurations of the systems and methods described in the present disclosure is shown.
[0128] As shown in FIG. 17, in some configurations, computing device 1650 can include a processor 1702, a display 1704, one or more inputs 1706, one or more communication systems 1708, and / or memory 1710. In some configurations, processor 1702 can be any suitable hardware processor or combination of processors, such as a central processing unit (" CPU”), a graphics processing unit (" GPU”), and so on. In some configurations, display 1704 can include any suitable display devices, such as a liquidAtty. Dkt. No. 125141.05023. MGH2024-029crystal display (" LCD”) screen, a light-emitting diode (" LED”) display, an organic LED (" OLED”) display, an electrophoretic display (e.g., an "e-ink" display), a computer monitor, a touchscreen, a television, and so on. In some configurations, inputs 1706 can include any suitable input devices and / or sensors that can be used to receive user input, such as a keyboard, a mouse, a touchscreen, a microphone, and so on.
[0129] In some configurations, communications systems 1708 can include any suitable hardware, firmware, and / or software for communicating information over communication network 1654 and / or any other suitable communication networks. For example, communications systems 1708 can include one or more transceivers, one or more communication chips and / or chip sets, and so on. In a more particular example, communications systems 1708 can include hardware, firmware, and / or software that can be used to establish a Wi-Fi connection, a Bluetooth connection, a cellular connection, an Ethernet connection, and so on.
[0130] In some configurations, memory 1710 can include any suitable storage device or devices that can be used to store instructions, values, data, or the like, that can be used, for example, by processor 1702 to present content using display 1704, to communicate with server 1652 via communications system(s) 1708, and so on. Memory 1710 can include any suitable volatile memory, non-volatile memory, storage, or any suitable combination thereof. For example, memory 1710 can include random-access memory (" RAM”), read-only memory (" ROM”), electrically programmable ROM (" EPROM”), electrically erasable ROM (" EEPROM”), other forms ofvolatile memory, other forms of non-volatile memory, one or more forms of semi-volatile memory, one or more flash drives, one or more hard disks, one or more solid state drives, one or more optical drives, and so on. In some configurations, memory 1710 can have encoded thereon, or otherwise stored therein, a computer program for controlling operation of computingAtty. Dkt. No. 125141.05023. MGH2024-029device 1650. In such configurations, processor 1702 can execute at least a portion of the computer program to present content (e.g., images, user interfaces, graphics, tables), receive content from server 1652, transmit information to server 1652, and so on. For example, the processor 1702 and the memory 1710 can be configured to perform the methods described herein.
[0131] In some configurations, server 1652 can include a processor 1712, a display 1714, one or more inputs 1716, one or more communications systems 1718, and / or memory 1720. In some configurations, processor 1712 can be any suitable hardware processor or combination of processors, such as a CPU, a GPU, and so on. In some configurations, display 1714 can include any suitable display devices, such as an LCD screen, LED display, OLED display, electrophoretic display, a computer monitor, a touchscreen, a television, and so on. In some configurations, inputs 1716 can include any suitable input devices and / or sensors that can be used to receive user input, such as a keyboard, a mouse, a touchscreen, a microphone, and so on.
[0132] In some configurations, communications systems 1718 can include any suitable hardware, firmware, and / or software for communicating information over communication network 1654 and / or any other suitable communication networks. For example, communications systems 1718 can include one or more transceivers, one or more communication chips and / or chip sets, and so on. In a more particular example, communications systems 1718 can include hardware, firmware, and / or software that can be used to establish a Wi-Fi connection, a Bluetooth connection, a cellular connection, an Ethernet connection, and so on.
[0133] In some configurations, memory 1720 can include any suitable storage device or devices that can be used to store instructions, values, data, or the like, that can be used, for example, by processor 1712 to present content using display 1714, toAtty. Dkt. No. 125141.05023. MGH2024-029communicate with one or more computing devices 1650, and so on. Memory 1720 can include any suitable volatile memory, non-volatile memory, storage, or any suitable combination thereof. For example, memory 1720 can include RAM, ROM, EPROM, EEPROM, other types of volatile memory, other types of non-volatile memory, one or more types of semi-volatile memory, one or more flash drives, one or more hard disks, one or more solid state drives, one or more optical drives, and so on. In some configurations, memory 1720 can have encoded thereon a server program for controlling operation of server 1652. In such configurations, processor 1712 can execute at least a portion of the server program to transmit information and / or content (e.g., data, images, a user interface] to one or more computing devices 1650, receive information and / or content from one or more computing devices 1650, receive instructions from one or more devices (e.g., a personal computer, a laptop computer, a tablet computer, a smartphone], and so on.
[0134] In some configurations, the server 1652 is configured to perform the methods described in the present disclosure. For example, the processor 1712 and memory 1720 can be configured to perform the methods described herein.
[0135] In some configurations, data source 1602 can include a processor 1722, one or more data acquisition systems 1724, one or more communications systems 1726, and / or memory 1728. In some configurations, processor 1722 can be any suitable hardware processor or combination of processors, such as a CPU, a GPU, and so on. In some configurations, the one or more data acquisition systems 1724 are generally configured to acquire data, images, or both, and can include an MRI system. Additionally or alternatively, in some configurations, the one or more data acquisition systems 1724 can include any suitable hardware, firmware, and / or software for coupling to and / or controlling operations of an MRI system. In some configurations, one or more portions ofAtty. Dkt. No. 125141.05023. MGH2024-029the data acquisition system(s) 1724 can be removable and / or replaceable.
[0136] Note that, although not shown, data source 1602 can include any suitable inputs and / or outputs. For example, data source 1602 can include input devices and / or sensors that can be used to receive user input, such as a keyboard, a mouse, a touchscreen, a microphone, a trackpad, a trackball, and so on. As another example, data source 1602 can include any suitable display devices, such as an LCD screen, an LED display, an OLED display, an electrophoretic display, a computer monitor, a touchscreen, a television, etc., one or more speakers, and so on.
[0137] In some configurations, communications systems 1726 can include any suitable hardware, firmware, and / or software for communicating information to computing device 1650 [and, in some configurations, over communication network 1654 and / or any other suitable communication networks). For example, communications systems 1726 can include one or more transceivers, one or more communication chips and / or chip sets, and so on. In a more particular example, communications systems 1726 can include hardware, firmware, and / or software that can be used to establish a wired connection using any suitable port and / or communication standard [e.g., VGA, DVI video, USB, RS-232, etc.), Wi-Fi connection, a Bluetooth connection, a cellular connection, an Ethernet connection, and so on.
[0138] In some configurations, memory 1728 can include any suitable storage device or devices that can be used to store instructions, values, data, or the like, that can be used, for example, by processor 1722 to control the one or more data acquisition systems 1724, and / or receive data from the one or more data acquisition systems 1724; to generate images from data; present content (e.g., data, images, a user interface) using a display; communicate with one or more computing devices 1650; and so on. Memory 1728 can include any suitable volatile memory, non-volatile memory, storage, or anyAtty. Dkt. No. 125141.05023. MGH2024-029suitable combination thereof. For example, memory 1728 can include RAM, ROM, EPROM, EEPROM, other types of volatile memory, other types of non-volatile memory, one or more types of semi-volatile memory, one or more flash drives, one or more hard disks, one or more solid state drives, one or more optical drives, and so on. In some configurations, memory 1728 can have encoded thereon, or otherwise stored therein, a program for controlling operation of medical image data source 1602. In such configurations, processor 1722 can execute at least a portion of the program to generate images, transmit information and / or content (e.g., data, images, a user interface) to one or more computing devices 1650, receive information and / or content from one or more computing devices 1650, receive instructions from one or more devices (e.g., a personal computer, a laptop computer, a tablet computer, a smartphone, etc.), and so on.
[0139] In some configurations, any suitable computer-readable media can be used for storing instructions for performing the functions and / or processes described herein. For example, in some configurations, computer-readable media can be transitory or non-transitory. For example, non-transitory computer-readable media can include media such as magnetic media (e.g., hard disks, floppy disks), optical media (e.g., compact discs, digital video discs, Blu-ray discs), semiconductor media (e.g., RAM, flash memory, EPROM, EEPROM), any suitable media that is not fleeting or devoid of any semblance of permanence during transmission, and / or any suitable tangible media. As another example, transitory computer-readable media can include signals on networks, in wires, conductors, optical fibers, circuits, or any suitable media that is fleeting and devoid of any semblance of permanence during transmission, and / or any suitable intangible media.
[0140] As used herein in the context of computer implementation, unless otherwise specified or limited, the terms "component," "system," "module," "controller," "framework," and the like are intended to encompass part or all of computer-relatedAtty. Dkt. No. 125141.05023. MGH2024-029systems that include hardware, software, a combination of hardware and software, or software in execution. For example, a component may be, but is not limited to being, a processor device, a process being executed (or executable) by a processor device, an object, an executable, a thread of execution, a computer program, or a computer. By way of illustration, both an application running on a computer and the computer can be a component. One or more components (or system, module, and so on) may reside within a process or thread of execution, may be localized on one computer, may be distributed between two or more computers or other processor devices, or may be included within another component (or system, module, and so on).
[0141] In some implementations, devices or systems disclosed herein can be utilized or installed using methods embodying aspects of the disclosure. Correspondingly, description herein of particular features, capabilities, or intended purposes of a device or system is generally intended to inherently include disclosure of a method of using such features for the intended purposes, a method of implementing such capabilities, and a method of installing disclosed (or otherwise known) components to support these purposes or capabilities. Similarly, unless otherwise indicated or limited, discussion herein of any method of manufacturing or using a particular device or system, including installing the device or system, is intended to inherently include disclosure, as embodiments of the disclosure, of the utilized features and implemented capabilities of such device or system.
[0142] As used herein, the phrase "at least one of A, B, and C" means at least one of A, at least one of B, and / or at least one of C, or any one of A, B, or C or combination of A, B, or C. A, B, and C are elements of a list, and A, B, and C may be anything contained in the Specification.
[0143] The present disclosure has described one or more preferred embodiments,Atty. Dkt. No. 125141.05023. MGH2024-029and it should be appreciated that many equivalents, alternatives, variations, and modifications, aside from those expressly stated, are possible and within the scope of the invention.
Claims
Atty. Dkt. No. 125141.05023. MGH2024-029CLAIMS1. A method for magnetic resonance imaging [MRI], comprising: acquiring a first shot of echo planar imaging (EPI) data with a first phase encoding polarity using an MRI system;acquiring a second shot of EPI data with a second phase encoding polarity that is opposite the first phase encoding polarity using the MRI system;wherein acquiring each of the first shot and second shot comprises: acquiring a first echo at a first resolution; andacquiring a second echo at a second resolution that is lower than the first resolution;estimating a field map from the second echo of the first shot and the second echo of the second shot; andreconstructing images with reduced distortion from the first echo of the first shot and the first echo of the second shot using the estimated field map.
2. The method of claim 1, wherein an effective echo spacing of the first echo and an effective echo spacing of the second echo are matched such that the first echo and the second echo have identical geometric distortion.
3. The method of claim 2, wherein the first echo is acquired with a first acceleration factor and the second echo is acquired with a second acceleration factor, wherein the effective echo spacing of the first echo is based on the first acceleration factor and the effective echo spacing of the second echo is based on the second acceleration factor.Atty. Dkt. No. 125141.05023. MGH2024-0294. The method of claim 3, wherein the second acceleration factor is lower than the first acceleration factor.
5. The method of claim 1, wherein reconstructing the images comprises jointly reconstructing the images from the first echo of the first shot and the first echo of the second shot using the estimated field map while applying a low-rank constraint.
6. The method of claim 1, wherein the first echo of the first shot and the first echo of the second shot are each acquired with diffusion-weighted encoding, and wherein the reconstructed images comprise diffusion-weighted images.
7. The method of claim 1, wherein acquiring each of the first shot and the second shot comprise encoding the first echo and the second echo using radiofrequency slab encoding to facilitate super-resolution in a slice direction.
8. The method of any one of claims 1 or 7, wherein the second echo of the first shot and the second echo of the second shot are each acquired during a dead time period configured to mitigate spin history effects.
9. The method of claim 1, wherein each of the first echoes is a spin echo and each of the second echoes is a spin echo.
10. The method of claim 1, further comprising using the second echo of the first shot and the second echo of the second shot to estimate a phase difference betweenAtty. Dkt. No. 125141.05023. MGH2024-029the first shot and the second shot; and wherein reconstructing the images comprises applying a correction based on the estimated phase difference.
11. The method of claim 1, further comprising using the second echo of the first shot and the second echo of the second shot to estimate a motion, and wherein reconstructing the images further comprises correcting the images based on the estimated motion.
12. The method of claim 1, wherein the second echo of the first shot and the second echo of the second shot are used as a phase navigator to estimate shot-to-shot phase variations.
13. The method of claim 12, wherein reconstructing the images comprises applying a Hybrid-SENSE reconstruction with shot-to-shot phase correction based on the estimated phase variations.
14. The method of claim 1, further comprising generating diffusion fractography maps from the reconstructed images.
15. A system for generating magnetic resonance imaging (MRI) images with reduced distortion, the system comprising:a computer processor configured to:control an MRI system to acquire a first shot of echo planar imaging (EPI) data with a first phase encoding polarity;Atty. Dkt. No. 125141.05023. MGH2024-029control the MRI system to acquire a second shot of EPI data with a second phase encoding polarity that is opposite the first phase encoding polarity;wherein to acquire the first shot and second shot of EPI data, the computer processor is configured to control the MRI system to:acquire a first echo at a first resolution; andacquire a second echo at a second resolution that is lower than the first resolution;estimate a field map from the second echo of the first shot and the second echo of the second shot; andreconstruct MRI images with reduced distortion from the first echo of the first shot and the first echo of the second shot using the estimated field map.
16. The system of claim 15, wherein each of the first echoes has a first acceleration factor and a first effective echo spacing based on the first acceleration factor and each of the second echoes has a second acceleration factor and a second effective echo spacing based on the second acceleration factor; and wherein the first effective echo spacing and the second effective echo spacing are matched such that the first echo and the second echo have identical geometric distortion.
17. The system of claim 16, wherein the first acceleration factor and the second acceleration factor are unequal.
19. The system of claim 15, wherein to reconstruct the MRI images with reduced distortion, the computer processor is configured to jointly reconstruct theAtty. Dkt. No. 125141.05023. MGH2024-029images from the first echo of the first shot and the first echo of the second shot using the estimated field map while applying a low-rank constraint.
19. The system of claim 15, wherein each of the first echoes and each of the second echoes are spin echoes acquired with diffusion-weighted encoding, and wherein the reconstructed images comprise diffusion-weighted images.
20. The system of claim 15, wherein the computer processor is further configured to control the MRI system to encode the first echo and the second echo using radiofrequency slab encoding to facilitate super-resolution in a slice direction.
21. The system of any one of claims 15 or 20, wherein the second echo of the first shot and the second echo of the second shot are each acquired during a dead time period configured to mitigate spin history effects.
22. A method for magnetic resonance imaging [MRI], comprising: acquiring a first shot of diffusion-weighted echo-planar imaging (EPI ) MRI data with a first phase encoding polarity using an MRI system;acquiring a second shot of diffusion-weighted EPI data with a second phase encoding polarity that is opposite the first phase encoding polarity using the MRI system;wherein acquiring each of the first shot and second shot comprises: applying a radiofrequency slab encoding pattern configured to facilitate superresolution in a slice direction;acquiring a first echo having a first contrast; andAtty. Dkt. No. 125141.05023. MGH2024-029acquiring a second echo having a second contrast during a dead time period configured to mitigate spin history effects;estimating one or more field maps from at least one of the first echo of the first shot and the first echo of the second shot or the second echo of the first shot and the second echo of the second shot;reconstructing a first set of images with reduced distortion from the first echo of the first shot and the first echo of the second shot using at least one of the one or more estimated field maps, the first set of images having the first contrast;reconstructing a second set of images with reduced distortion from the second echo of the first shot and the second echo of the second shot using at least one of the one or more estimated field maps, the second set of images having the second contrast; and generating a diffusion relaxometry parameter map based on the first set of images and the second set of images.
23. The method of claim 22, wherein reconstructing the first set of images comprises jointly reconstructing the images from the first echo of the first shot and the first echo of the second shot using at least one of the estimated field maps while applying a low-rank constraint; and wherein reconstructing the second set of images comprises jointly reconstructing the images from the second echo of the first shot and the second echo of the second shot using at least one of the estimated field maps while applying a low-rank constraint.
24. The method of claim 22, wherein acquiring each of the first shot and second shot further comprises acquiring a third echo having a third contrast during the dead time period configured to mitigate spin history effects.
