Magnetic resonance chemical exchange saturation transfer imaging method, device, equipment, product and storage medium

Abstract

The invention is suitable for the technical field of magnetic resonance chemical exchange saturation transfer imaging, and provides a magnetic resonance chemical exchange saturation transfer imaging method, device, equipment, product and storage medium. A gradient echo sequence is adopted to collect saturated water molecule signals, and the collected water molecule signals are stored in a K space; and wave control coding is introduced in the acquisition process, and a CEST image is reconstructed based on a water molecule signal of a K space. The introduction of wave control coding can reduce the scanning time. Due to the fact that the wave control coding sampling efficiency is high, a larger matrix size can be supported within the same time, high-resolution imaging is achieved, the partial volume effect can be relieved, and CEST quantitative accuracy is improved.

Description

Technical Field

[0001] This application relates to the field of magnetic resonance chemical exchange saturation transfer imaging technology, and in particular to a magnetic resonance chemical exchange saturation transfer imaging method, apparatus, equipment, product and storage medium. Background Technology

[0002] Chemical Exchange Saturation Transfer (CEST) is a molecular imaging technique based on magnetic resonance imaging. It involves selectively saturating (or canceling out) the magnetization of exchangeable protons (such as -OH, -NH, -NH2) in specific molecules (e.g., proteins, metabolites) within the body. This "saturation" state is then transferred to the free water through chemical exchange between these protons and surrounding free water molecules, thus indirectly detecting information about these low-concentration macromolecules or metabolites.

[0003] However, in scenarios such as three-dimensional (3D) or high-resolution imaging, a large number of signals often need to be acquired, resulting in excessively long scan times and affecting clinical applications. Due to the influence of radiofrequency saturation pulses, the signal-to-noise ratio of CEST images is limited, resulting in low image resolution and susceptibility to volume effects that affect quantitative accuracy, leading to incorrect estimations of lesion extent or activity. Summary of the Invention

[0004] This application provides a magnetic resonance chemical exchange saturation transfer imaging method, apparatus, device, product, and storage medium, which can reduce the data acquisition time in K-space and improve the image resolution.

[0005] To achieve the above objectives, this application adopts the following technical solution: Firstly, a chemical exchange saturation transfer imaging method is provided, the method comprising: The saturated water molecule signal is acquired using a gradient echo sequence and stored in the K space; wherein the readout gradient in the readout direction GRO, the phase encoding gradient in the layer selection direction GSPE, and the wave-controlled encoding gradient in the intralayer phase encoding GPE are all enabled. CEST images are reconstructed based on water molecule signals in the K-space.

[0006] Optionally, the filling method of the K space includes: the number of times the water molecule signal corresponding to each frequency offset in the K space is collected is greater than or equal to 2, and in each collection process, the K space is filled sequentially from the center point to the periphery of the K space.

[0007] Optionally, the number of samplings can be determined based on the size of the K-space.

[0008] Optionally, during each acquisition process, the water molecule signal corresponding to each frequency offset in the K space is acquired through a keyhole partial acquisition method.

[0009] Optionally, the CEST image reconstruction based on K-space water molecule signals includes: reconstructing the CEST image based on K-space water molecule signals using compressed sensing.

[0010] Optionally, the method of reconstructing the CEST image based on water molecule signals in K-space using compressed sensing includes: CEST images are reconstructed based on water molecule signals in K-space using compressed sensing with L1 regularization constraints.

[0011] Optionally, the GSPE and the GPE vary according to a sine wave waveform.

[0012] Optionally, the GSPE lasts for 0.5 cycles longer than the GPE; the GSPE starts 0.25 cycles earlier than the GPE; and the GPE starts when the GRO reaches its plateau.

[0013] In a second aspect, a chemical exchange saturation transfer imaging apparatus is provided, comprising at least one module for performing the method described in any embodiment of the first aspect.

[0014] Thirdly, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the chemical exchange saturation transfer imaging method as in any optional implementation of the first aspect.

[0015] Fourthly, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the chemical exchange saturation transfer imaging method as described in any of the first aspects.

[0016] Fifthly, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the steps of the chemical exchange saturation transfer imaging method as described in any of the first aspects.

[0017] The magnetic resonance chemical exchange saturation transfer imaging method provided in this application has the following effective effects: This application uses gradient echo sequences to acquire saturated water molecule signals and stores the acquired water molecule signals in K-space. Beam-controlled coding is introduced during the acquisition process, and CEST images are reconstructed based on the water molecule signals in K-space. The introduction of beam-controlled coding, during GRE readout, enables intra-layer phase coding (GPE) and layer-selective orientation phase coding (GSPE), so that the sampling points in K-space are no longer limited to a straight line, but move along non-Cartesian trajectories such as three-dimensional spirals or radial lines. This improves the coil sensitivity dealiasing performance in parallel magnetic resonance imaging, thereby increasing the downsampling factor and reducing scan time. Due to the high sampling efficiency, a larger matrix size can be supported in the same time, thus achieving high-resolution imaging, helping to mitigate some volume effects and improve the quantitative accuracy of CEST. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating a CEST imaging process according to an embodiment of this application; Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application; Figure 3 A schematic flowchart of a magnetic resonance chemical exchange saturation transfer imaging method provided in an embodiment of this application; Figure 4 A schematic diagram of the sequence timing acquired during CEST imaging, provided as an embodiment of this application; Figure 5 A schematic diagram illustrating the time required for a complete data acquisition process as provided in this application embodiment; Figure 6 This is a schematic diagram of a magnetic resonance chemical exchange saturation transfer imaging device provided in an embodiment of this application. Detailed Implementation

[0019] It should be noted that the terminology used in the embodiments of this application is only for explaining specific embodiments of this application and is not intended to limit this application. In the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more, "at least one" or "one or more" means one, two or more. The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

[0020] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0021] Chemical exchange saturation transfer (CEST) imaging is a highly sensitive molecular imaging technique capable of detecting endogenous metabolites / compounds (such as glutamate, glycogen, proteins / peptides, etc.) and exogenous CEST contrast agents in the tissue microenvironment. This technique selectively saturates exchangeable protons in target molecules, and chemical exchange information of the target molecules can be obtained by measuring changes in free water signals. For example, glutamate CEST (gluCEST) imaging can image glutamate by detecting the chemical exchange between amine protons and free water, and has significant potential applications in the diagnosis and assessment of stroke, tumors, and central nervous system diseases. Amide proton transfer (APT) imaging can obtain protein / peptide information by detecting the chemical exchange process between amide protons and water molecules, and has significant potential applications in brain tumor grading and activity assessment, early diagnosis of cerebral infarction, identification of pseudoprogression after radiotherapy, and pathological assessment of other central nervous system diseases.

[0022] Please see Figure 1 , Figure 1 This is a schematic diagram of a CEST imaging process.

[0023] First, a saturated radio frequency pulse is applied to the exchangeable protons of the solute molecules at a specific frequency offset by a scanner. This selectively saturates the exchangeable protons of the solute molecules, which then transfer the saturated state to the surrounding water molecules through chemical exchange, thereby reducing the water signal.

[0024] Secondly, after turning off the saturated radio frequency pulse, the saturated water molecule signal is acquired and filled into the K space.

[0025] Then, wait for the water proton magnetization vector to recover to its initial state through relaxation, in preparation for the next excitation.

[0026] The above three steps complete one signal acquisition. To obtain a complete image, the above steps need to be repeated at multiple frequency shifts.

[0027] In scenarios such as three-dimensional (3D) or high-resolution imaging, a large number of signals often need to be acquired, resulting in excessively long scan times and affecting clinical applications. Due to the influence of radiofrequency saturation pulses, the signal-to-noise ratio of CEST images is limited, resulting in low image resolution and susceptibility to volume effects that affect quantitative accuracy, leading to incorrect estimations of lesion extent or activity.

[0028] To address the aforementioned issues, this application provides a chemical exchange saturation transfer imaging method. This method employs gradient echo sequences to acquire saturated water molecule signals, stores these signals in K-space, and introduces waveguide coding during the acquisition process. CEST images are then reconstructed based on the water molecule signals in K-space. The introduction of waveguide coding, during GRE readout, enables intra-layer phase coding (GPE) and layer-selective orientation phase coding (GSPE), allowing sampling points in K-space to move along non-Cartesian trajectories such as three-dimensional spirals or radial lines, rather than being confined to a straight line. This improves the coil sensitivity dealiasing performance in parallel magnetic resonance imaging, thereby increasing the downsampling factor and reducing scan time. Due to the high sampling efficiency, a larger matrix size can be supported within the same time frame, achieving high-resolution imaging, helping to mitigate some volumetric effects, and improving the quantitative accuracy of CEST.

[0029] In this application embodiment, the execution subject of magnetic resonance chemical exchange saturation transfer imaging can be an electronic device.

[0030] Figure 2 This is a schematic diagram of the structure of an electronic device shown in an embodiment of this application, such as... Figure 2 As shown in the embodiment of this application, an electronic device 20 is provided, which may be a mobile terminal or a server.

[0031] If the electronic device is a mobile terminal, then the electronic device can be a device such as a smartphone, tablet computer, or personal computer.

[0032] If the electronic device is a server, then the electronic device can be a standalone server, a server cluster, a cloud server, or other similar devices.

[0033] like Figure 2 As shown, the electronic device may include a processor 201, a memory 202, a communication bus 203, and a communication interface 204. The processor 201 is connected to the memory 202 and the communication interface 204 via the communication bus 203. It should be noted that... Figure 2 The illustrated device structure does not constitute a limitation on the electronic device 20. The electronic device 20 may include more or fewer components than illustrated, or combine certain components, or have different component arrangements. This application embodiment does not limit this. The following, in conjunction with... Figure 2A detailed description of each component of the electronic device 20 is provided.

[0034] The processor 201 can be a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, used to execute relevant programs to perform the steps executed by the electronic device in the computational graph optimization method described in the embodiments of this application. The processor 201 can also be an integrated circuit chip with signal processing capabilities. In implementation, each step of the computational graph optimization method of this application can be completed by the integrated logic circuits in the hardware of the processor 201 or by instructions in software form. The processor 201 described above can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the method disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or can be executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory 202. Processor 201 collects information from memory 202 and, in conjunction with its hardware, executes the steps performed by the electronic device in the computational graph optimization method of the embodiments of this application. Memory 202 can be read-only memory (ROM), static storage medium, dynamic storage medium, or random access memory (RAM). Memory 202 can store programs and data, such as the program for the computational graph optimization method in the embodiments of this application. When the program stored in memory 202 is executed by processor 201, processor 201 executes the various steps performed by the electronic device in the computational graph optimization method of the embodiments of this application. Communication interface 204 uses transceiver devices, such as, but not limited to, transceivers, to realize communication between electronic device 20 and other devices or communication networks.

[0035] Please see Figure 3 , Figure 3This application provides a flowchart of a chemical exchange saturation transfer imaging method. The following describes the application of this chemical exchange saturation transfer imaging method. Figure 2 Taking the electronic device shown as an example, this chemical exchange saturation transfer imaging method includes the following steps: S301. Acquire saturated water molecule signals using gradient echo sequences and store the acquired water molecule signals in K space; wherein, within the data readout window where the readout gradient is maintained, both the layer selection direction phase encoding direction and the intra-layer phase encoding direction are enabled.

[0036] In the gradient echo, among the three components—Readout Gradient (GRO), Layer Selected Oriented Phase Encoding (GSPE), and Intra-Layer Phase Encoding (GPE)—the Readout Gradient is activated during signal acquisition to determine the left-right (or head-to-toe) position of the human body. By applying a gradient field, water protons at different locations have different precession frequencies. The signal acquired by the coil is the sum of signals from water protons of different frequencies; the frequency is decomposed using Fourier transform to determine the signal's position. The Intra-Layer Phase Encoding gradient is used to determine the front-back (or another dimension) position of the human body. The Layer Selected Oriented Phase Encoding is used to determine the vertical (layer-to-layer) position of the human body.

[0037] In this application, during the readout or acquisition of saturated water molecule signals, three gradients—GRO, GSPE, and GPE—are simultaneously activated. When these three gradients change simultaneously during readout, the trajectory of the sampling point is jointly determined by the gradient encoding in these three directions. By applying wavecontrol encoding to the readout data, the readout gradient GRO remains activated simultaneously with the data. The same wavecontrol encoding gradient is activated on each gradient echo in the echo chain, ensuring that consistent wavecontrol encoding information is recorded for each echo.

[0038] In one example, GSPE and GPE change according to a sinusoidal waveform, and GSPE and GPE in K-space... and The two generate orthogonal periodic oscillations in the direction, which drive the sampling points in... Draw a circular or elliptical spiral trajectory in a plane. Responsible The sampling points are advanced forward in the direction of the process, and the three elements combine to form a three-dimensional spiral that moves forward in K-space. This method covers most of the K-space with high efficiency, and the amount of data collected in a single scan far exceeds that of traditional progressive scanning, directly improving the scanning rate.

[0039] If GSPE and GPE change according to a cosine waveform, the gradient rise needs to first rise to the peak of the cosine wave before continuing downwards. This could lead to problems such as eddy currents due to excessively long rise times and high rise rates. This application, by changing according to a sine waveform, can effectively reduce eddy currents induced in the conductive structure, and the continuous waveform change generates less gradient noise, which helps improve patient comfort in clinical applications.

[0040] As an example, the GRO can vary according to a target waveform, which can be a linearly increasing waveform or a specific gradient waveform.

[0041] In one example, GSPE has 0.5 more cycles than GPE; GSPE starts 0.25 cycles earlier than GPE, while GPE... It will be activated when the plateau period is reached.

[0042] Among them, GSPE and GPE drive respectively and When sampling in a specific direction, this frequency difference will cause the sampling point to be in... The trajectory in the plane is no longer a simple circle or ellipse, but a closed or slowly precessing Lissajous curve. It can cover a region of K-space more uniformly and densely.

[0043] Wherein, 0.25 periods correspond to a 90-degree phase difference, at the initial moment of signal acquisition ( If GPE starts from zero (sin 0 = 0), and GSPE has reached its peak (cos 0 = 1), the sampling point will be directly from... A point on the axis begins to move. This ensures that the starting direction of the sampling trajectory in the central region of K-space is determined, which helps to achieve symmetrical and uniform coverage.

[0044] in, The plateau period is the gradient at which the readout direction begins during data readout. GPE is activated during this plateau period to avoid phase encoding in the nonlinear regions where the gradient is rising or falling, thereby ensuring... Encoding accuracy.

[0045] In one example, S301 is used to implement the filling process of the K-space, or the data acquisition process of the K-space. For example... Figure 1 As shown, the acquisition process for saturated water molecule signals occurs after each saturation pulse execution process. For example, by applying a saturated radio frequency pulse to the exchangeable protons of solute molecules at a specific frequency offset using a scanner, the exchangeable protons of solute molecules are selectively saturated, and then the protons transfer the saturated state to the surrounding water molecules through chemical exchange, and then S301 is executed.

[0046] Gradient Recalled Echo (GRE) is a fast imaging sequence in magnetic resonance imaging (MRI). It generates echo signals by reversing the direction of the gradient magnetic field. The signal acquisition process is as follows: First, protons are excited using a small-angle radio frequency pulse, causing a transverse component in the macroscopic magnetization vector and initiating precession. Then, a phase-dissociated gradient (usually negative) is applied in the readout direction to accelerate the phase divergence of the protons, resulting in rapid signal decay. Next, a refocusing gradient of opposite polarity and equal duration is applied, causing the phases diverged by the gradient field to begin converging again. At a specific time point (i.e., echo time TE), the phase difference caused by the gradient field is eliminated, and the transverse magnetization vector re-achieves macroscopic coherence under the influence of the gradient field, forming a signal peak, i.e., the gradient echo. The receiving coil is activated at this moment to acquire this signal, completing the water molecule signal readout.

[0047] Wavelength-controlled coding is a coding technique that uses a pre-designed, continuously varying gradient waveform over time to control the trajectory and distribution of K-space sampling points. During the acquisition of saturated water molecule signals, in-layer phase coding (GPE) and layer-selective directional phase coding (GSPE) are continuously activated according to preset waveforms, guiding the sampling points to move along a smooth, continuous curve (such as a spiral) in K-space. Compared to traditional line-by-line acquisition, this trajectory method can acquire K-space data from a larger area at a time.

[0048] Furthermore, due to the continuous change in gradient, waveguide coding introduces a specific phase for each acquired data point. This coding information helps the spatial sensitivity of the multi-channel coil capture originally aliased pixels, providing easier mathematical conditions for reconstruction in parallel imaging or compressed sensing reconstruction algorithms. This allows for higher downsampling factors while maintaining image quality when combined with advanced reconstruction algorithms, thereby further improving the scanning rate.

[0049] S302. Reconstruction of CEST images based on water molecule signals in K-space.

[0050] One example is a CEST image, which is a CEST parameter map, such as a CEST parameter map that can be used for clinical diagnosis.

[0051] In one example, before S302, the acquisition process ends, resulting in a filled K-space where each point in the matrix represents a specific amplitude and phase.

[0052] In one example, after obtaining a filled K-space, an inverse Fourier transform is first performed on the K-space data acquired at each frequency offset point to convert the signal from the frequency domain (K-space) to the spatial domain (image space), generating a grayscale image at the corresponding frequency. The grayscale image includes a non-saturated pulse image (S0) and a saturated pulse image (Ssat). Among them, the unsaturated pulse image (S0) reflects the signal strength of the tissue without saturation interference, while the saturated image (Ssat) is the image after the water signal is saturated to different degrees at different frequency shifts.

[0053] If multi-channel coil acquisition is used, each coil will independently reconstruct an image. Multiple images can be combined into a final, high signal-to-noise ratio image using methods such as sum of squares (SoS).

[0054] Secondly, molecular information is extracted through Z-spectrum analysis and quantitative calculations. Specifically, a region of interest (such as a tumor region or a normal brain tissue region) is selected on the image, and the signal intensity of that region at each frequency offset point is extracted. A curve is plotted with the saturation frequency offset (ppm) as the horizontal axis and the normalized signal intensity as the vertical axis; this is the Z-spectrum. The CEST effect is then calculated. Finally, the quantitative calculation results are applied to each pixel of the image to generate a color parametric map.

[0055] Thus, gradient echo sequences are used to acquire saturated water molecule signals, which are then stored in K-space. Beam-controlled coding is introduced during the acquisition process, and CEST images are reconstructed based on the water molecule signals in K-space. The introduction of beam-controlled coding enables the beam-controlled coding gradients of the readout gradient (GRO), layer-selective directional phase coding (GSPE), and intralayer phase coding (GPE), allowing sampling points in K-space to move along non-Cartesian trajectories such as three-dimensional spirals or radial lines, rather than being confined to a straight line. This allows for the coverage of a larger area of ​​K-space in a single scan, thereby reducing scanning time. Due to the high sampling efficiency, a larger matrix size can be supported within the same time frame, achieving high-resolution imaging, which helps mitigate some volumetric effects and improves the quantitative accuracy of CEST.

[0056] In one example, S301 is the process of acquiring water molecule signals, which is also the process of filling the K-space. The filling order of the K-space can be controlled, starting from the center point of the K-space and proceeding outwards. That is, a center-to-outward spatial filling order is adopted, which is a sampling strategy that prioritizes the acquisition of low-frequency central regions in the K-space.

[0057] In magnetic resonance imaging, data points in K-space represent spatial frequency information. The center (origin) of K-space contains low-frequency components, which are used to determine the image's contrast, signal-to-noise ratio, and overall contour. The periphery of K-space contains high-frequency components, which are used to determine the image's resolution and edge details.

[0058] By filling the K-space from the center outwards, contrast information is prioritized after excitation, which reduces the likelihood of rescanning. During the center-to-outward acquisition process, motion-induced artifacts typically present as diffuse blurring, having minimal impact on diagnosis. Sampling the central region of the K-space provides reliable low-frequency constraints for subsequent reconstruction algorithms, supporting higher speedups.

[0059] In one example, data corresponding to each frequency offset in K-space can be obtained by repeatedly acquiring saturated water molecule signals. That is, the number of acquisitions is greater than or equal to 2, and only a part of the data corresponding to each frequency offset in K-space is acquired each time. The total data corresponding to each frequency offset in K-space is obtained by accumulating the data multiple times. In each acquisition process, the data is filled from the center point of K-space outwards.

[0060] One example is a complete data acquisition process that includes... Figure 1 The three processes are shown.

[0061] Among them, as the number of acquisitions increases, the amount of data acquired in each excitation process decreases, which greatly shortens the acquisition time. As the readout time progresses, the signal will naturally decay due to T2* relaxation. By shortening the acquisition time, the acquisition can be completed before the signal decays severely, thus improving the signal-to-noise ratio.

[0062] One example is that the amount of data collected each time can be preset, and after obtaining the total amount of data corresponding to each frequency offset in the K space, the number of collections can be determined based on the amount of data collected each time and the total amount of data.

[0063] For example, the number of data collections = total data volume / data volume collected each time.

[0064] The amount of data collected each time is determined based on a signal-to-noise ratio (SNR) threshold, which is used to indicate the required image quality.

[0065] Increasing the number of acquisitions reduces the acquisition time per acquisition, resulting in sharper images and a higher signal-to-noise ratio. However, this can also lead to longer scan times and increased radio frequency energy deposition. The number of acquisitions corresponding to each frequency offset in the K-space can be adjusted based on actual needs to ensure that scan time, signal-to-noise ratio, and image sharpness all meet user requirements. For CEST imaging with weak signal changes, image quality can be improved by increasing the number of acquisitions corresponding to each frequency offset in the K-space.

[0066] In one example, during each acquisition process, the water molecule signal corresponding to each frequency offset in the K-space can be acquired using a keyhole partial acquisition method.

[0067] Keyhole imaging is a magnetic resonance imaging acceleration technique based on partial K-space acquisition. It is suitable for dynamic or multi-dimensional imaging requiring high temporal resolution (such as CEST imaging). Utilizing the property that the center of K-space determines image contrast, it primarily updates the central region of K-space, while other regions are acquired through other means (such as reference images), thus significantly reducing scan time while maintaining image quality.

[0068] One example involves acquiring the water molecule signal corresponding to each frequency offset in the K-space using a keyhole-based acquisition method, including: First, an image containing the entire K-space is acquired as a reference image; this reference image can be an image without saturation pulses. During data acquisition, the central portion of the K-space (i.e., the keyhole region) is acquired, eliminating the need to acquire data from the periphery of the K-space. By using the keyhole region acquisition method, the amount of data acquired at each frequency offset point is reduced, significantly decreasing the data acquisition time per session.

[0069] Secondly, once data acquisition is complete, the acquired low-resolution keyhole portion data can be merged with the peripheral K-space data of the reference image to obtain a complete K-space. Then, a CEST image can be reconstructed based on the completed K-space.

[0070] In magnetic resonance imaging (MRI), to shorten scan time, the amount of data sampled in the K-space can be reduced, resulting in undersampling (e.g., acquiring only 20% of the data). However, undersampling produces severe aliasing artifacts. Compressed sensing (CS) can mathematically reconstruct a clear image from a small amount of aliased data.

[0071] One example is the reconstruction of CEST images based on water molecule signals in the K-space using compressed sensing with regular constraints.

[0072] In compressed sensing, undersampling results in fewer equations than unknowns (mathematically known as "ill-posed"), leading to non-unique solutions. Regularization constraints, by introducing prior knowledge (i.e. our expectations of image features) among countless possible solutions, constrain the form of the solutions, forcing the algorithm to select the optimal solution that both conforms to the collected data and has the expected features.

[0073] One example is the L1 regularization constraint, which refers to constraining the sparsity of an image using the L1 norm.

[0074] For example, CEST images need to meet two conditions: after undersampling Fourier transform, the image must match the acquired K-space data. The consistency and image are sparse in a certain transform domain (such as the wavelet domain) (i.e., most coefficients are zero or close to zero), and a mathematical model is first established based on this condition; The mathematical model is decomposed into several easily solvable subproblems using the Alternating Direction Multiplier Method (ADMM). After iterative convergence, the solution obtained is the desired CEST image.

[0075] Among them, wave control coding endows the data with rich phase information at the acquisition end, while L1 regularized constraint compressed sensing makes full use of phase information to solve sparse constraints at the image reconstruction level. The beneficial effect of the combination of the two is far greater than the simple addition of the two.

[0076] Specifically, the introduction of waveguide coding continuously changes the phase coding gradient (GPE and GSPE) during readout, causing sampling points to move along non-Cartesian trajectories such as spirals. This improves acquisition efficiency and introduces a unique phase history for each data point. In L1 regularized compressed sensing, the spatial frequency coordinates and phase information of each sampling point are crucial input information during image reconstruction. The rich phase history provided by waveguide coding is equivalent to providing additional spatial constraints for the reconstruction algorithm. The phase information from waveguide coding allows L1 regularization to more accurately distinguish between real signals and aliasing artifacts during iterative solving, thus reconstructing aliasing-free, high signal-to-noise ratio images even at extremely high speedups.

[0077] Furthermore, beta coding alters the distribution characteristics of K-space data. Because the sampling trajectory spirals along the readout direction (GRO), each beta-coded readout data vector exhibits a spiral-shaped sparse coverage in K-space. This characteristic introduces a displacement of voxels along the readout direction, spreading them over a wider area during downsampling and accelerated scanning, which helps to de-aliasing in parallel imaging. The reconstructed image from K-space data after de-aliasing is more easily identified by L1 regularization as noise components requiring removal in the sparse transform domain, significantly enhancing the overall imaging process's tolerance to slight patient movements, respiratory movements, and magnetic field drift. In addition, the technology of this application, targeting the CEST imaging characteristics, adopts a K-space filling method that fills the K-space from the center outward, so that after each excitation, the low-frequency information that determines the contrast is acquired first.

[0078] Specifically, in K-space, all sampling points to be acquired are determined on the layer-selective phase encoding (SPE)-intra-layer phase encoding (PE) plane. After determining the index of each sampling point, the index matrix is ​​rearranged in polar coordinates (transforming the coordinate system of the SPE-PE index into a coordinate system of polar angle-polar radius), and all sampling points to be acquired are equally divided into multiple sectors of the same size according to the polar angle. Within each sector, the SPE-PE index point list within the sector is rearranged in ascending order of distance from each point to the center of the plane. Finally, for the gradient echo chain within each excitation, the phase encoding gradient magnitude in the corresponding SPE and PE directions of each gradient echo is automatically calculated based on the rearranged index list within the corresponding sector, and the corresponding echo data is read into the K-space data matrix.

[0079] Even though the saturation information gradually decays over time after a single excitation, the acquired K-space center data still maintains image contrast, while the K-space periphery provides high-frequency information such as the edges of physiological tissues in the image. This improvement enables wave-controlled encoded gradient echo sequences to be effectively adapted to magnetic resonance imaging scanning tasks like CEST, which involve large signal fluctuations.

[0080] Furthermore, addressing the issue of low signal-to-noise ratio in CEST imaging signals after saturation, this application further utilizes keyhole technology to improve the sampling strategy. First, a K-space image is acquired as a reference signal during a scan with extreme frequency offset (also known as a reference image scan) in an unsaturated / saturated image, according to predetermined acceleration factors, matrix size, and other parameters. Next, for the signals to be scanned in the saturated image, based on the K-space filled by the reference signal, a central portion (i.e., the keyhole region) is selected on the phase encoding plane of the SPE-PE, omitting the remaining peripheral portions, serving as a sampling mask corresponding to the saturated image. Finally, the designed center-outward K-space filling strategy is applied to the keyhole sampling mask to obtain the sampling order of the signals corresponding to the saturated image. This keyhole-based partial acquisition method reduces the amount of data acquired at each frequency offset point, significantly reducing the single data acquisition time, while omitting excessive high-frequency K-space information and suppressing the influence of noise.

[0081] Furthermore, the introduction of beam-controlled coding allows for the simultaneous activation of beam-controlled gradients for both GPE and GSPE during readout, resulting in each acquired data point possessing a unique coded phase determined by the gradient waveform. This phase information complements the spatial sensitivity information of the multi-coil system. L1 canonical compressed sensing can be naturally integrated into the parallel imaging reconstruction framework. When solving the objective function, both the multi-coil sensitivity information and the phase information from the beam-controlled coding can be used as prior constraints. This provides the reconstruction algorithm with dual constraints when dealiasing signals, leading to higher dealiasing accuracy. This means that with the same number of coils, higher speedup factors can be supported, or lower g-factor noise amplification can be achieved at the same speedup factor.

[0082] The ultimate goal of CEST imaging is to obtain accurate parametric maps, prioritizing the acquisition of K-space centers to precisely capture CEST contrast (minor signal changes). Aliasing artifacts and noise are removed, improving the image signal-to-noise ratio and resolution, and mitigating some volumetric effects. High-quality S0 and Ssat images ensure the reliability of subsequent quantitative calculations. Images reconstructed through the combined effects of wave-controlled coding and compressed sensing have pixel values ​​closer to the true tissue signal intensity, resulting in calculated CEST effect values ​​with smaller biases, higher repeatability, and a more accurate reflection of the concentration and distribution of target molecules.

[0083] Please see Figure 4 and Figure 5 , Figure 4 and Figure 5 This is a schematic diagram of the sequence timing acquired during a CEST imaging process, as provided in an embodiment of this application.

[0084] The K-space of this application includes multiple frequency offsets, each offset corresponding to multiple trains, such as... Figure 4 The data corresponding to each offset in the K space, from Train1 to TrainM, needs to be obtained through multiple acquisitions, with each acquisition yielding one Train.

[0085] in, Figure 1 This is a complete data collection process.

[0086] Figure 5 The time of a complete acquisition process (shot TR) includes: relaxation recovery time ( ), saturation pulse time ( ) and data collection time ( ).in, The echo train length represents the number of lines in the K-space acquired in a single acquisition process.

[0087] Figure 5Three modules are illustrated: Saturation, Data Acquisition, and Delay. Initially, the saturation module generates a narrow pulse, the excitation pulse. Then, the data readout module uses a gradient echo sequence to acquire the saturated water molecule signal and stores it in the K-space. During the data readout time window, while the readout gradient GRO remains active, both the layer-selective phase encoding (GSPE) direction and the intra-layer phase encoding (GPE) direction are active. After data readout, the relaxation recovery module waits for the water proton magnetization vector to relax to its equilibrium state for a time T1. This provides consistent initial conditions for the next excitation, ensuring the stability of signal intensity and contrast.

[0088] Among them, in each saturation pulse time window ( (Inside): Apply a continuous-wave saturating radio frequency pulse to selectively saturate exchangeable protons at the target resonant frequency.

[0089] In each data collection time window ( Within the first stage, an SLR pulse (an excitation pulse designed based on the SLR algorithm) is applied to excite the water proton generation signal. Subsequently, gradient channels in three directions (layer-selected directional phase encoding) are used. Intra-layer phase coding Reading the gradient Simultaneously, it is activated and continuously changes according to a preset waveform. Through the continuously changing gradient waveform, the sampling point moves in the K-space along a non-Cartesian trajectory (such as a spiral), which significantly improves the data acquisition efficiency of a single excitation.

[0090] After the gradient waveform stabilizes, the analog-to-digital converter (ADC) is turned on to acquire water molecule signals and fill the data into the K space.

[0091] See Figure 6 , Figure 6 This is a schematic diagram of a magnetic resonance chemical exchange saturation transfer imaging device provided in an embodiment of this application. This magnetic resonance chemical exchange saturation transfer imaging device can be deployed in the electronic equipment described in the foregoing embodiments. Figure 6 As shown, the magnetic resonance chemical exchange saturation transfer imaging device includes an acquisition module and a reconstruction module.

[0092] As an example, a module in a magnetic resonance chemical exchange saturation transfer imaging device can be used to perform... Figure 4 The steps of the magnetic resonance chemical exchange saturation transfer imaging method include, for example, the acquisition module performing S301 in the aforementioned embodiment, and the reconstruction module performing S302 in the aforementioned embodiment.

[0093] The following is a detailed explanation: Optionally, the acquisition module is used to: acquire saturated water molecule signals using a gradient echo sequence with wave control coding, and store the acquired water molecule signals in the K space; wherein, within the data readout time window during which the readout gradient GRO is kept on, both the layer selection direction phase coding GSPE direction and the intra-layer phase coding GPE direction are on. The reconstruction module is used to reconstruct the CEST image based on the water molecule signals in the K space.

[0094] Optionally, the filling method of the K space includes: the number of times the water molecule signal corresponding to each frequency offset in the K space is collected is greater than or equal to 2, and in each collection process, the K space is filled sequentially from the center point to the periphery of the K space.

[0095] Optionally, the acquisition module is also used to: determine the number of acquisitions based on the size of the K-space.

[0096] Optionally, during each acquisition process, the water molecule signal corresponding to each frequency offset in the K space is acquired through a keyhole partial acquisition method.

[0097] Optionally, the CEST image reconstruction based on K-space water molecule signals includes: reconstructing the CEST image based on K-space water molecule signals using compressed sensing.

[0098] Optionally, the method of reconstructing the CEST image based on water molecule signals in K-space using compressed sensing includes: CEST images are reconstructed based on water molecule signals in K-space using compressed sensing with L1 regularization constraints.

[0099] Optionally, the GSPE and the GPE vary according to a sine wave waveform.

[0100] Optionally, the GSPE lasts for 0.5 cycles longer than the GPE; the GSPE starts 0.25 cycles earlier than the GPE; and the GPE starts when the GRO reaches its plateau.

[0101] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units is merely an example. In practical applications, the above functions can be assigned to different functional units as needed, that is, the internal structure of the power supply control device can be divided into different functional units to complete all or part of the functions described above. The functional units in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0102] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps in the various method embodiments described above.

[0103] This application provides a computer program product that, when run on a testing device, enables the testing device to perform the steps described in the various method embodiments above.

[0104] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, refer to the relevant descriptions of other embodiments.

[0105] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0106] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

[0107] Finally, it should be noted that the above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A magnetic resonance chemical exchange saturation transfer imaging method, characterized in that, The method includes: The saturated water molecule signal is acquired using a gradient echo sequence and stored in the K space; wherein, during the data readout time window in which the readout gradient GRO is kept on, both the layer selection direction phase encoding GSPE direction and the intra-layer phase encoding GPE direction are on. CEST images are reconstructed based on water molecule signals in the K-space.

2. The method according to claim 1, characterized in that, The filling method of the K space includes: the number of times the water molecule signal corresponding to each frequency offset in the K space is collected is greater than or equal to 2, and in each collection process, the K space is filled sequentially from the center point to the periphery of the K space.

3. The method according to claim 1 or 2, characterized in that, During each acquisition process, the water molecule signal corresponding to each frequency offset in the K space is acquired through a keyhole acquisition method.

4. The method according to claim 1 or 2, characterized in that, The CEST image reconstructed based on K-space water molecule signals includes: CEST images were reconstructed using compressed sensing based on water molecule signals in K-space.

5. The method according to claim 1, characterized in that, The GSPE and the GPE change according to a sine wave waveform.

6. The method according to claim 5, characterized in that, The GSPE lasts for 0.5 cycles longer than the GPE; the GSPE starts 0.25 cycles earlier than the GPE; and the GPE starts when the GRO reaches its plateau.

7. A magnetic resonance chemical exchange saturation transfer imaging device, characterized in that, include: At least one module that performs the method as described in any one of claims 1 to 6.

8. An electronic device, characterized in that, Includes at least one processor; and A memory storing a computer program executable on the processor, wherein the processor executes the computer program to perform the method as described in any one of claims 1 to 6.

9. A program product, characterized in that, When the program product is run on a computing node, it causes the at least one computing node to perform the method of any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, Includes programs or instructions that, when executed by the processor of a computing node, implement the method as described in any one of claims 1 to 6.

Images