Magnetic resonance CEST imaging pre-saturation method and equipment based on amplitude and off-resonance frequency modulation

Abstract

The invention discloses a magnetic resonance CEST imaging pre-saturation method and equipment based on amplitude and off-resonance frequency modulation, and belongs to the field of magnetic resonance imaging. According to the method, a Block-McConnell mathematical simulation model is utilized, and the amplitude of a CEST pre-saturation sub-pulse and an off-resonance frequency parameter are used as adjustable parameters for nonlinear optimization so as to minimize the variation degree of a CEST effect value when the CEST effect value experiences non-uniform intensity of a radio frequency field, so that the influence of the non-uniformity of the radio frequency field on CEST imaging is reduced. The method does not need to depend on additional hardware equipment and a radio frequency field shimming process, does not need to collect additional data for post-processing correction, and does not increase additional collection time while ensuring convenient implementation.

Description

Technical Field

[0001] The present invention belongs to the field of magnetic resonance imaging, and particularly relates to the optimization of presaturation pulses for magnetic resonance chemical exchange saturation transfer imaging. Background Art

[0002] Magnetic resonance chemical exchange saturation transfer imaging (CEST) is a novel molecular imaging technique that utilizes the chemical exchange process between hydrogen protons in trace macromolecular proteins and hydrogen protons in free water in the human body to indirectly detect information on trace macromolecular proteins in the human body, thereby detecting the metabolic process of the human body.

[0003] In order to obtain higher spectral resolution and signal-to-noise ratio, people have gradually extended CEST imaging to high fields and explored its applications at high fields. As the field strength increases, the wavelength of the radiofrequency pulse becomes shorter, resulting in the formation of standing waves in tissues, that is, constructive interference is formed near the central region of the tissue, while destructive interference occurs at a position one-quarter wavelength away from the center, manifested as spatial inhomogeneity of the radiofrequency field (B1). Since both the presaturation and imaging modules of CEST imaging require radiofrequency excitation, the quality of CEST imaging is greatly affected by the inhomogeneity of the B1 field.

[0004] In the past few decades, people have invented various methods to reduce the influence of B1 field inhomogeneity on magnetic resonance imaging. In the 1980s, with the development of spectrometers and radiofrequency pulse technology, people proposed pulses based on amplitude and frequency modulation, called adiabatic pulses, which can achieve uniform excitation flip angles under inhomogeneous B1 fields. However, the adiabatic pulse theory needs to ignore the influence of relaxation effects, and no study has explored the feasibility of adiabatic pulses in scenarios such as CEST presaturation that require the application of long pulses. Parallel excitation is another method proposed in recent years, which reduces the influence of B1 field inhomogeneity through modulation between multiple excitation coils. Parallel excitation modulates the amplitude and phase between each excitation channel to achieve uniform B1 field intensity in a local space. However, this method requires additional hardware and more complex system calibration processes, so it has not been widely used in clinics. In addition to improvements at the data acquisition end, people have also invented many data correction algorithms to correct the influence of B1 field inhomogeneity at the post-processing end, including methods using the relationship between steady-state CEST signals and the B1 field, methods using polynomial fitting models to fit the correction curve of B1 field inhomogeneity, and correction methods using Omega maps, etc. However, these algorithms all rely on collecting data with multiple different presaturation powers for B1 field inhomogeneity correction, so the acquisition time will increase exponentially. Summary of the Invention

[0005] The object of the present invention is to provide a magnetic resonance CEST presaturation method and device that do not rely on additional hardware and do not require collecting additional data, and have high clinical and commercial value, so as to solve the problem that CEST imaging is affected by the inhomogeneity of the radiofrequency field intensity.

[0006] The specific technical solution adopted by the present invention is as follows:

[0007] In a first aspect, the present invention provides a method for optimizing the presaturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation, which includes the following steps:

[0008] S1. For a target object to be subjected to magnetic resonance CEST imaging, use a multi-pool Bloch-McConnell mathematical simulation model to perform simulation according to the target CEST presaturation power. During the simulation process, simulate the change in the radiofrequency field intensity caused by the radiofrequency field inhomogeneity during the pulsed CEST presaturation process by adjusting the radiofrequency field inhomogeneity factor of the CEST presaturation sub-pulse, obtain the z-spectrum under different CEST presaturation sub-pulse modulation parameters, and calculate the CEST effect value at the frequency of interest according to the z-spectrum; the CEST presaturation sub-pulse modulation parameters are composed of the amplitude and off-resonance frequency of each sub-pulse in the CEST presaturation module;

[0009] S2. Use the CEST presaturation sub-pulse modulation parameters and the corresponding CEST effect values calculated by the simulation as sample data, and take the minimum variation degree of the CEST effect value under the change of the radiofrequency field intensity during the CEST presaturation process as the optimization goal, and perform non-linear optimization on the amplitude and off-resonance frequency of each sub-pulse in the CEST presaturation module to obtain the optimal modulation parameters of the CEST presaturation sub-pulse with the least influence of the radiofrequency field intensity change under the target CEST presaturation power.

[0010] As a preference of the above first aspect, the multi-pool Bloch-McConnell mathematical simulation model is represented by the following formula:

[0011]

[0012] where expm represents the exponential function with the natural constant e as the base, Δt represents the increment of time t, respectively represent the magnetization vectors at times t + Δt and t and matrix n is the number of solute pools, n ≥ 1; respectively represent the magnetization vector components of the free water pool in the x, y, and z directions, respectively represent the magnetization vector components of n solute groups, where the magnetization vector components of the i-th solute group in the x, y, and z directions are represents the magnetization vector component of the semi-solid magnetization transfer group in the z direction, while the magnetization vector components of the semi-solid magnetization transfer group in the x and y directions are negligible; the coefficient matrix where:

[0013]

[0014] where Δω w represents the difference between the off-resonance frequency ω of the applied CEST pre-saturation sub-pulse and the resonance frequency of the free water group, and Δω i represents the difference between the off-resonance frequency ω of the applied CEST pre-saturation sub-pulse and the resonance frequency of the i-th solute group; respectively represent the longitudinal relaxation rate and the transverse relaxation rate of the free water group, respectively represent the longitudinal relaxation rate and the transverse relaxation rate of the i-th solute group, and respectively represent the longitudinal relaxation rate and the transverse relaxation rate of the semi-solid magnetization transfer group; respectively represent the steady-state magnetization vectors of the free water group, the i-th solute group, and the semi-solid magnetization transfer group; k w_si , k si_w respectively represent the hydrogen proton exchange rates from the free water group to the i-th solute group and from the i-th solute group to the free water group. There is the following correlation between the exchange rate and the steady-state magnetization vector: k w_MT , k MT_w respectively represent the hydrogen proton exchange rates from the free water group to the semi-solid magnetization transfer group and from the semi-solid magnetization transfer group to the free water group. There is the following correlation between the exchange rate and the steady-state magnetization vector: ω1 represents the angular frequency of the applied CEST pre-saturation sub-pulse, which is determined by the gyromagnetic ratio γ and the sub-pulse amplitude B1, ω1 = γ·B1, and the CEST pre-saturation sub-pulse amplitude B1 = B 1_weight ·B 1_modulation ·B 1_desired , where B 1_desired represents the amplitude of the desired CEST pre-saturation sub-pulse, which is determined by the target CEST pre-saturation power corresponding to the current simulation; B 1_weight is a radio frequency field intensity inhomogeneity factor within the range of [1 - ε, 1 + ε], which is used to simulate the change in the radio frequency field intensity caused by the radio frequency field inhomogeneity, and ε is the preset change amplitude; B 1_modulation is the amplitude adjustment factor of the CEST pre-saturation sub-pulse to be optimized.

[0015] As a preference of the first aspect above, during the simulation process, for all sub-pulses in the CEST presaturation module, it is necessary to determine B when the target CEST presaturation power corresponding to the current simulation is satisfied. 1_desired , and at the same time sample different values from the change range of B 1_weight and assign them to all sub-pulses to simulate the change of the RF field intensity caused by the RF field inhomogeneity. Then, different B 1_modulation and off-resonance frequencies ω are assigned to each sub-pulse to form different CEST presaturation sub-pulse modulation parameters. Finally, using the multi-population-based Bloch-McConnell mathematical simulation model, the excitation process of each sub-pulse in the CEST presaturation module in the time domain is simulated under different CEST presaturation sub-pulse modulation parameters to obtain CEST imaging data and calculate the z-spectrum.

[0016] As a preference of the first aspect above, during the simulation process, the change range of the RF field intensity inhomogeneity factor B 1_weight is 0.8 to 1.2, and the change step of B 1_weight is 0.05.

[0017] As a preference of the first aspect above, among the CEST presaturation sub-pulse modulation parameters, the amplitude of the sub-pulse needs to be set within the optimized upper and lower bounds to ensure that the SAR value of the sequence does not exceed the system limit. The off-resonance frequency of the sub-pulse needs to be superimposed with a modulation frequency on the basis of the original CEST polarization frequency, and the modulation frequency needs to be set within the optimized upper and lower bounds to ensure the frequency selectivity of CEST imaging.

[0018] As a preference of the first aspect above, when performing non-linear optimization on the amplitude and off-resonance frequency of each sub-pulse in the CEST presaturation module, a non-linear optimization process based on sequential quadratic programming is adopted, and the objective function used is as follows:

[0019]

[0020] Where: coef pulsei represents the CEST presaturation sub-pulse modulation parameters, including the amplitude B1 and off-resonance frequency ω of each sub-pulse in the CEST presaturation module; C(CEST(B1,ω target )) is the loss function, which is used to calculate the variation degree of the CEST effect value at the interested frequency ω 1_weight when B target changes in [1-ε,1+ε]. C(CEST(B1,ω target )) is defined as:

[0021]

[0022] where: CEST(B1, ω target ) represents the CEST effect value at the interested frequency ω when the amplitude of the CEST presaturation sub-pulse is B1 target , and CEST(B 1_desired , ω target ) represents the CEST effect value at the interested frequency ω when the amplitude of the CEST presaturation sub-pulse is B 1_desired ; target ;

[0023] Penalty term is defined as:

[0024]

[0025] In the formula: represents the CEST effect value at the interested frequency ω when performing Bloch-McConnell simulation using only free water groups (i.e., without considering solute groups and semi-solid magnetization transfer groups in the Bloch-McConnell mathematical simulation model) when the amplitude of the CEST presaturation sub-pulse is B1, which is used to constrain the distortion and central drift of the z-spectrum caused by off-resonance frequency modulation, so as to minimize the CEST effect value additionally introduced by off-resonance frequency modulation; α is a weight parameter used to balance the penalty term and the loss function. target ;

[0026] In a second aspect, the present invention provides a presaturation method for magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation, which includes:

[0027] Obtain a pre-constructed parameter dictionary; the construction method of the parameter dictionary is: for different CEST presaturation powers, respectively according to the magnetic resonance CEST imaging presaturation parameter optimization method described in any item of the first aspect above, after obtaining the optimal modulation parameters of the CEST presaturation sub-pulse with the least influence by the radiofrequency field intensity change at each CEST presaturation power, associate and store the CEST presaturation power and the optimal modulation parameters of the CEST presaturation sub-pulse in the parameter dictionary;

[0028] Before performing CEST imaging presaturation, according to the specified CEST presaturation power to be adopted, look up the corresponding optimal modulation parameters of the CEST presaturation sub-pulse in the parameter dictionary, and after correspondingly setting the amplitudes and off-resonance frequencies of each sub-pulse in the CEST presaturation module, execute the CEST presaturation module to collect the CEST image with the least influence by the radiofrequency field intensity change.

[0029] In a third aspect, the present invention provides a computer-readable storage medium, characterized in that a computer program is stored on the storage medium, and when the computer program is executed by a processor, the method for optimizing the presaturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation as described in any one of the above first aspects is implemented, or the presaturation method of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation as described in the above second aspect is implemented.

[0030] In a fourth aspect, the present invention provides a computer electronic device, characterized by comprising a memory and a processor;

[0031] The memory is used for storing a computer program; the processor is used for implementing the method for optimizing the presaturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation as described in any one of the above first aspects, or implementing the presaturation method of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation as described in the above second aspect when executing the computer program.

[0032] In a fifth aspect, the present invention provides a magnetic resonance imaging device, characterized by comprising a magnetic resonance scanner and a control unit, and a computer program is stored in the control unit, and when the computer program is executed, it is used to control the magnetic resonance scanner to acquire a CEST image with the least influence by the change of the radiofrequency field intensity according to the presaturation method of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation as described in the above second aspect.

[0033] The present invention has the following beneficial effects compared with the prior art:

[0034] (1) The method provided by the present invention is convenient to implement, without the need to use an additional multi-channel coil and without a complex system calibration process, and only the amplitude and off-resonance frequency parameters of the presaturation pulse need to be adjusted according to the desired presaturation pulse power.

[0035] (2) The method provided by the present invention relies on the prior knowledge provided by simulation, rather than relying on any post-processing calibration algorithm. Therefore, there is no need to additionally collect data with multiple different presaturation powers, which greatly improves the acquisition efficiency and reduces the acquisition time compared with the existing post-processing calibration methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a contrast of the differences between the traditional CEST presaturation method and the CEST presaturation method proposed by the present invention in the time domain and the frequency domain;

[0037] Figure 2Scanning results of the phantom embodiments. (A): CEST images of the phantom at different radiofrequency field strengths; (B): Scanning results of the radiofrequency field strength coefficient quantification map, where the variation range of the radiofrequency field strength in the phantom experiment is between 0.8 and 1.2; (C): MTR calculated using two different presaturation methods in two regions of interest with the maximum and minimum 10% of the radiofrequency field strength asym Signal distribution map.

[0038] Figure 3 In the phantom experiment, the average z-spectrum (first and second rows) and average MTR calculated in two regions of interest with the maximum and minimum 10% of the radiofrequency field strength asym Spectrum (third and fourth rows).

[0039] Figure 4 Scanning results of a healthy volunteer embodiment. (A): CEST images of the healthy volunteer at different radiofrequency field strengths; (B): CEST source image at 15 ppm applied to white matter segmentation; (C): Scanning results of the radiofrequency field strength coefficient quantification map, where the variation range of the radiofrequency field strength in the healthy volunteer experiment is between 0.85 and 1.15; (D): MTR calculated using two different presaturation methods in two regions of interest with the maximum and minimum 10% of the radiofrequency field strength in the region where the white matter volume fraction of the healthy volunteer is greater than 90% asym Signal distribution map.

[0040] Figure 5 In the healthy volunteer experiment, the average z-spectrum (first and second rows) and average MTR calculated in two regions of interest with the maximum and minimum 10% of the radiofrequency field strength in the region where the white matter volume fraction of the healthy volunteer is greater than 90% asym Spectrum (third and fourth rows). Detailed implementation manners

[0041] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific implementation manners of the present invention will be given in conjunction with the accompanying drawings. Many specific details are set forth in the following description to fully understand the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below. The technical features in each embodiment of the present invention can be combined correspondingly without conflict.

[0042] The present invention provides an optimization method for pre-saturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation. This method requires simulation using the Bloch-McConnell equation. The conventional three-group exchange model of water-solute-semi-solid magnetization transfer effect (MT) can be expressed by the following formula. Since the MT group has an ultra-short T2 time, it is considered that the transverse magnetization vector of the MT group reaches a steady state instantaneously. Therefore, the change in the transverse magnetization vector of the MT group is ignored:

[0043]

[0044] where Δω w represents the difference between the off-resonance frequency pulse applied externally and the resonance frequency of the free water group, and Δω s represents the difference between the off-resonance frequency pulse applied externally and the resonance frequency of the solute group, and Δω MT represents the difference between the off-resonance frequency pulse applied externally and the resonance frequency of the MT group; represent the longitudinal relaxation rate and the transverse relaxation rate of the free water group respectively, represent the longitudinal relaxation rate and the transverse relaxation rate of the solute group respectively, and R rfMT represents the attenuation rate of the longitudinal magnetization vector of the MT group affected by the saturation pulse, represents the transverse relaxation rate of the MT group; represent the steady-state magnetization vectors of the free water group, the solute group and the MT group respectively; k w_s represents the hydrogen proton exchange rate from the free water group to the solute group, and k s_w represents the hydrogen proton exchange rate from the solute group to the free water group. The exchange rate and the steady-state magnetization vector are related as follows: k w_MT represents the hydrogen proton exchange rate from the free water group to the MT group, and k MT_w represents the hydrogen proton exchange rate from the MT group to the free water group. And the two exchange rates and the steady-state magnetization vector are related as follows: ω1 represents the angular frequency of the applied radio frequency pulse, which is determined by the gyromagnetic ratio γ and the amplitude B1 of the radio frequency pulse, ω1 = γ·B1. The above Bloch-McConnell equation simulation can jointly simulate multiple different solute groups and be rewritten in the following matrix form to represent the multi-group based Bloch-McConnell mathematical simulation model for iterative solution:

[0045]

[0046] where expm represents the exponential function with the natural constant e as the base, and Δt represents the increment at time t. represent the magnetization vectors at times t + Δt and t respectively where matrix n is the number of solute clusters, n ≥ 1; represent the magnetization vector components of the free water cluster in the x, y, and z directions respectively represent the magnetization vector components of the n solute clusters respectively. The magnetization vector components of the i-th solute cluster in the x, y, and z directions are respectively i = 1, 2…n; represents the magnetization vector component of the semi-solid magnetization transfer cluster (MT) in the z direction, while the magnetization vector components of this cluster in the x and y directions are negligible due to its extremely short T2 time. Coefficient matrix where:

[0047]

[0048] where Δω i represents the difference between the off-resonance frequency ω of the externally applied off-resonance frequency pulse and the resonance frequency of the i-th solute cluster; represent the longitudinal relaxation rate and the transverse relaxation rate of the i-th solute cluster respectively represent the T1 relaxation time and the T2 relaxation time of the i-th solute cluster respectively; represents the steady-state magnetization vector of the i-th solute cluster; k w_si represents the hydrogen proton exchange rate from the free water cluster to the i-th solute cluster, k si_w represents the hydrogen proton exchange rate from the i-th solute cluster to the free water cluster, and there is the following correlation between the two exchange rates and the steady-state magnetization vector: ω1 represents the angular frequency of the applied CEST pre-saturation sub-pulse, which is determined by the gyromagnetic ratio γ and the amplitude B1 of the CEST pre-saturation sub-pulse, ω1 = γ·B1, and the amplitude B1 of the CEST pre-saturation sub-pulse is B 1_weight ·B 1_modulation ·B 1_desired where B 1_desired represents the desired amplitude of the CEST pre-saturation sub-pulse, which is determined by the target CEST pre-saturation power corresponding to the current simulation. In practical applications, the amplitude B of a single sub-pulse is calculated based on this target CEST pre-saturation power and the preset pulse waveform (such as a Gaussian pulse) 1_desired ; B 1_weightis a radiofrequency field intensity non-uniformity factor within the range of [1-ε, 1+ε]. During the simulation process, the change of B 1_weight can be used to simulate the change of radiofrequency field intensity caused by radiofrequency field non-uniformity. ε is the preset change amplitude. The present invention only considers the radiofrequency field intensity non-uniformity in space and ignores the radiofrequency field intensity fluctuation in the time dimension. Therefore, the same B 1_weight is used between each sub-pulse. The change of the sub-pulse amplitude caused by the radiofrequency field spatial non-uniformity is calculated from B 1_weight ·B 1_desired ; B 1_modulation is the amplitude adjustment factor of the CEST pre-saturation sub-pulse to be optimized. B 1_weight and B 1_modulation are both dimensionless ratio values and can be superimposed on B 1_desired to change the actual amplitude of the CEST pre-saturation sub-pulse. And B 1_modulation is the parameter that can be actively adjusted in the present invention, that is, in the CEST pre-saturation sub-pulse modulation parameters, the amplitude of each sub-pulse in the CEST pre-saturation module is actually controlled by B 1_modulation .

[0049] Therefore, in the above-mentioned multi-population-based Bloch-McConnell mathematical simulation model, the CEST pre-saturation sub-pulse modulation parameters are composed of the amplitude B1 of each sub-pulse in the CEST pre-saturation module (actually controlled by B 1_modulation ) and the off-resonance frequency ω.

[0050] Therefore, in the embodiments of the present invention, based on the above-mentioned multi-population-based Bloch-McConnell mathematical simulation model, a method for optimizing the pre-saturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation is provided, which includes the following steps:

[0051] S1. For the target object to be subjected to magnetic resonance CEST imaging, use the multi-population-based Bloch-McConnell mathematical simulation model to perform simulation according to the target CEST pre-saturation power. During the simulation process, simulate the change of radiofrequency field intensity caused by radiofrequency field non-uniformity during the pulsed CEST pre-saturation process of the target object by adjusting the radiofrequency field intensity non-uniformity factor of the CEST pre-saturation sub-pulse, obtain the z-spectrum under different CEST pre-saturation sub-pulse modulation parameters, and calculate the CEST effect value at the frequency of interest according to the z-spectrum; the CEST pre-saturation sub-pulse modulation parameters are composed of the amplitude of each sub-pulse in the CEST pre-saturation module and the off-resonance frequency,

[0052] S2. Using the CEST presaturation sub-pulse modulation parameters obtained from simulation calculations and the corresponding CEST effect values as sample data, and taking the minimum variation degree of the CEST effect values under the change of the radiofrequency field intensity during the CEST presaturation process as the optimization objective, nonlinearly optimize the amplitudes and off-resonance frequencies of each sub-pulse in the CEST presaturation module to obtain the optimal modulation parameters of the CEST presaturation sub-pulse with the least influence of the radiofrequency field intensity change under the target CEST presaturation power.

[0053] During the simulation process, since the sub-pulses in the CEST presaturation module are applied continuously, time-domain simulation calculations also need to be carried out according to the specific excitation times of each sub-pulse. For all sub-pulses in the CEST presaturation module, it is necessary to determine the B when meeting the target CEST presaturation power corresponding to the current simulation 1_desired , and at the same time sample different values from the change range of B 1_weight and assign them to all sub-pulses to simulate the change of the radiofrequency field intensity caused by the radiofrequency field inhomogeneity. Then, assign different B 1_modulation and off-resonance frequency ω to each sub-pulse to form different CEST presaturation sub-pulse modulation parameters. Finally, use the multi-population-based Bloch-McConnell mathematical simulation model to simulate the excitation process of each sub-pulse in the CEST presaturation module in the time domain under different CEST presaturation sub-pulse modulation parameters, obtain the CEST imaging data through the readout module and calculate the z-spectrum. The signal value of each voxel in the CEST imaging data is the aforementioned parameter The method of calculating the z-spectrum from the CEST imaging data belongs to the prior art and will not be elaborated here.

[0054] In addition, it should be noted that the CEST effect value at the interested frequency ω target of the present invention is calculated according to the z-spectrum, and the specific interested frequency and CEST effect value can be selected according to the actual situation. In the embodiment of the present invention, the selected CEST effect value is the APT signal with ω target = 3.5 ppm, which is obtained by subtracting the signal value at -3.5 ppm from the signal value at +3.5 ppm on the z-spectrum.

[0055] In the embodiment of the present invention, in order to reduce the influence of the radiofrequency field inhomogeneity on the CEST image, the variation degree of the CEST effect value when B 1_weight varies between 0.8 and 1.2 (the change step is 0.05) is simulated, and the amplitudes and off-resonance frequencies of the presaturation sub-pulses are used as adjustable parameters for non-linear optimization. The optimization process of the present invention uses the variation degree of the CEST effect value at the interested frequency ω target as the loss function to calculate when B 1_weightThe degree of variation of the CEST effect value at the interesting frequency ω caused by the variation in [1 - ε, 1 + ε], C(CEST(B1, ω target )) is defined as: target )) is defined as:

[0056]

[0057] where: CEST(B1, ω target ) represents the CEST effect value at the interesting frequency ω when the amplitude of the CEST pre-saturation sub-pulse is B1, and CEST(B target , ω 1_desired ) represents the CEST effect value at the interesting frequency ω when the amplitude of the CEST pre-saturation sub-pulse is B target . In this embodiment, ε is taken as 0.2, B1 is sampled at a step size of 0.05 within the range of 0.8 to 1.2, and then C(CEST(B1, ω 1_desired )) is actually obtained by summing 9 (ΔCEST) target when B target = 0.80, 0.85, 0.90, 0.95, 1.00, 1.05, 1.10, 1.15, 1.20. 1_weight = 0.80, 0.85, 0.90, 0.95, 1.00, 1.05, 1.10, 1.15, 1.20. 2 is obtained by summation.

[0058] Since the present invention uses off-resonance frequency modulation, to a certain extent, this modulation method will cause the center drift and distortion of the z-spectrum, resulting in abnormal signal peaks when analyzing the asymmetry of the z-spectrum. Therefore, an additional penalty term needs to be added during the optimization process. The penalty term is defined as:

[0059]

[0060] In the formula: represents the CEST effect value at the interesting frequency ω when performing Bloch-McConnell simulation using only the free water group (i.e., without considering the solute group and the semi-solid magnetization transfer group in the Bloch-McConnell mathematical simulation model) when the amplitude of the CEST pre-saturation sub-pulse is B1. It is used to constrain the distortion and center drift of the z-spectrum caused by using polarization frequency modulation. This penalty term simulates the response of the optimized pulse under the condition of only direct water saturation effect, so that target reaches the minimum at each off-resonance frequency, to suppress the center drift and distortion of the z-spectrum caused by off-resonance frequency modulation, thereby minimizing the additional CEST effect value introduced by off-resonance frequency modulation. reaches the minimum at each off-resonance frequency, to suppress the center drift and distortion of the z-spectrum caused by off-resonance frequency modulation, thereby minimizing the additional CEST effect value introduced by off-resonance frequency modulation.

[0061] Based on the penalty term P and the loss function C, a weight parameter α can be used to sum the two, and then the target function used is obtained as follows:

[0062]

[0063] where: coef pulsei represents the CEST presaturation sub-pulse modulation parameter, which includes the amplitude B1 and the off-resonance frequency ω of each sub-pulse in the CEST presaturation module.

[0064] In an embodiment of the present invention, when non-linearly optimizing the amplitude and off-resonance frequency of each sub-pulse in the CEST presaturation module, a non-linear optimization process based on sequential quadratic programming is used to perform the optimization solution of coef pulsei

[0065] In addition, when optimizing the CEST presaturation sub-pulse modulation parameter, necessary constraint conditions need to be set for the amplitude and off-resonance frequency of each sub-pulse in the CEST presaturation module. The off-resonance frequency of the sub-pulse in the present invention can be regarded as an additional superimposed modulation frequency on the basis of the original CEST polarization frequency (determined by the original design of the CEST presaturation module). In order to ensure that the SAR value of the invented method based on amplitude and off-resonance frequency modulation does not exceed the system limit, and at the same time to ensure the frequency selectivity of the CEST imaging signal, upper and lower limits need to be set for the amplitude modulation and off-resonance frequency modulation of the sub-pulse.

[0066] In addition, the present invention can further provide a magnetic resonance CEST imaging presaturation method based on amplitude and off-resonance frequency modulation, which includes the following steps:

[0067] Obtain a pre-constructed parameter dictionary; the construction method of the parameter dictionary is: for different CEST presaturation powers, respectively according to the above-mentioned magnetic resonance CEST imaging presaturation parameter optimization method, after obtaining the optimal modulation parameters of the CEST presaturation sub-pulse with the least influence of the radiofrequency field intensity change under each CEST presaturation power, associate and store the CEST presaturation power and the optimal modulation parameters of the CEST presaturation sub-pulse in the parameter dictionary.

[0068] Before performing CEST imaging presaturation, according to the specified CEST presaturation power adopted, find the corresponding optimal modulation parameters of the CEST presaturation sub-pulse from the parameter dictionary, and after correspondingly setting the amplitude and off-resonance frequency of each sub-pulse in the CEST presaturation module, execute the CEST presaturation module to collect the CEST image with the least influence of the radiofrequency field intensity change.

[0069] The magnetic resonance CEST imaging presaturation method based on amplitude and off-resonance frequency modulation of the present invention is compared with the traditional presaturation method​Figure 1 As shown, in the traditional presaturation module, the amplitudes of all sub-pulses are the same, and the corresponding off-resonance frequencies are also the same. However, in the presaturation method of the present invention, after optimization, there will be differences in the amplitude and off-resonance frequency of each sub-pulse.

[0070] It should be noted that the form of the above parameter dictionary is not limited, as long as it can establish a mapping relationship between the corresponding CEST presaturation power and the optimal modulation parameters of the CEST presaturation sub-pulses. Since the CEST presaturation power determined by simulation is generally discrete, in the CEST imaging experiment, according to the actually expected applied CEST presaturation power, the optimal modulation parameters of the CEST presaturation sub-pulses corresponding to the CEST presaturation power closest to it are searched in the parameter dictionary, and the CEST image with the least influence of the radiofrequency field intensity change is acquired based on this.

[0071] Next, an embodiment will be used to further demonstrate the aforementioned method for optimizing the presaturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation of the present invention, as well as the technical effects that the magnetic resonance CEST imaging presaturation method can achieve, so that those skilled in the art can better understand the essence of the present invention.

[0072] Embodiment

[0073] In this embodiment, the CEST scanning parameters are set as follows:

[0074] The scanning in the embodiment of the present invention is based on a Siemens 3T scanner (MAGNETOM Prisma, Siemens Healthcare, Erlangen, Germany), using a 20-channel receive head coil, and using a CEST imaging sequence based on fast spin echo readout (TSE). The CEST presaturation pulse uses a B375 Gaussian pulse provided by the Siemens pulse tool, with a single pulse duration of 100 ms, a total of 10 saturation pulses are applied, and the total saturation time is 1 s; the pulse interval time is 10 ms, and a spoiling gradient is applied during the interval time to remove the transverse magnetization vector and ensure the smoothness of the z-spectrum. The readout uses a single-layer TSE sequence with a slice thickness of 5 mm, and the FOV is 212*185.5 mm 2 , and the acquisition matrix size is 96*84, and the in-plane voxel size is 2.2*2.2 mm 2, TR=3000ms, TE=7.2ms, echo chain length is 48, and Fat Sat is used for fat signal suppression. In the embodiment, a total of 41 polarization frequencies (Inf, ±6, ±5, ±4.5, ±4, ±3.75, ±3.5, ±3.25, ±3, ±2.5, ±2, ±1.5, ±1, ±0.75, ±0.5, ±0.25, 0, 10, 15, 20, 30, 40, 50, 60, 70, 80ppm) are applied to the CEST saturation pulse, and the total scanning time is 4 minutes and 11 seconds. The water saturation offset reference method (WASSR) is used to correct the inhomogeneity of the main magnetic field. The WASSR sequence collects a total of 26 frequencies. Except for Inf, the remaining 25 frequencies are evenly distributed between -1.5ppm and +1.5ppm, and the sequence acquisition time is 55 seconds. In addition, a quantitative sequence of radio frequency field intensity coefficient provided by the manufacturer was scanned, with an acquisition time of 10 seconds. The spatial positioning, FOV, acquisition matrix, phase encoding direction, and uniform field mode of the quantitative sequence of radio frequency field intensity coefficient and the WASSR sequence are consistent with the CEST imaging sequence. All scanning sequences use coil sensitivity maps for channel merging and normalization to avoid erroneous conclusions due to the inhomogeneity of the receiving coil.

[0075] 1.MRI phantom experiment:

[0076] The phantom experiment was configured with a 100mM / L nicotinamide phantom for CEST experiments, and manganese chloride and agar were used to regulate the T1 and T2 values of the phantom. The manganese chloride concentration used in the phantom was 0.1653mM / L, the agar concentration was 2.55w / v, and a cylindrical sealed container with a waist circumference of 470mm and a total volume of 5L was selected. The final scan of the phantom had a T1 of 1150ms and a T2 of 98ms, which is consistent with the T1 and T2 value range of normal human cranial tissue.

[0077] The phantom experiment was simulated using the Bloch-McConnell equation with 3 groups (the number of solute groups n = 2) to obtain the optimal pulse parameters. Specific simulation parameters: free water (T1 = 1150ms, T2 = 98ms, M0 = 74.59), amide protons (T1 = 1000ms, T2 = 10ms, M0 = 0.0745, k s_w =50Hz,Δω=3.5ppm), amine protons (T1=1000ms, T2=10ms, M0=0.03, k s_w =2000Hz,Δω=3.0ppm).

[0078] Four different CEST presaturation pulse powers (1.5 μT, 2.0 μT, 2.5 μT, 3.0 μT) were scanned in the phantom experiment. Therefore, the optimal pulse parameters need to be optimized for these four different presaturation pulse powers respectively.

[0079] 2. MRI healthy volunteer experiment:

[0080] The healthy volunteer experiments in the examples were all approved by the local ethics review board. A total of 4 volunteers participated in the experiment (two males and two females, aged 22 - 24).

[0081] The phantom experiment used the Bloch - McConnell equation of 7 pools (the number of solute pools n = 6) for simulation to obtain the optimal pulse parameters. Specific simulation parameters: free water (T1 = 1000 ms, T2 = 100 ms, M0 = 74.59), amide proton (T1 = 1000 ms, T2 = 10 ms, M0 = 0.0745, k s_w = 50 Hz, Δω = 3.5 ppm), amine proton (T1 = 1000 ms, T2 = 10 ms, M0 = 0.03, k s_w = 2000 Hz, Δω = 3.0 ppm), semi - solid magnetization transfer proton (T1 = 1000 ms, T2 = 0.01 ms, M0 = 6.0, k s_w = 45 Hz, Δω = 0 ppm), nuclear Overhauser effect proton (T1 = 1000 ms, T2 = 1 ms, M0 = 0.5, k s_w = 16 Hz, Δω = - 3.5 ppm), creatine proton (T1 = 1000 ms, T2 = 10 ms, M0 = 0.04, k s_w = 500 Hz, Δω = 1.8 ppm), hydroxyl proton (T1 = 1000 ms, T2 = 55 ms, M0 = 0.1, k s_w = 2000 Hz, Δω = 0.9 ppm).

[0082] Five different CEST presaturation pulse powers (1.0 μT, 1.5 μT, 2.0 μT, 2.5 μT, 3.0 μT) were scanned in the healthy volunteer experiment. The optimal pulse parameters need to be optimized for these five different presaturation pulse powers respectively.

[0083] The data processing and analysis methods of the examples are as follows:

[0084] In the phantom experiment, according to the acquired radiofrequency field intensity coefficient quantitative map, two regions of interest (ROIs) with the largest and smallest 10% of the radiofrequency field intensity in the phantom were extracted respectively, and the average z - spectrum and MTR within the two ROI regions were calculated. asymSpectrum. In the experiment of healthy volunteers, first, a CEST source image with a presaturation frequency of 15 ppm was selected and the data was segmented. The white matter region obtained by segmentation was selected, and voxels with a white matter volume fraction greater than 90% were extracted for analysis. According to the quantitatively mapped radiofrequency field intensity coefficient collected, two ROI regions with the maximum and minimum 10% radiofrequency field intensities in the selected white matter region were selected, and the average CEST z-spectrum and MTR asym spectra within the two ROIs were calculated respectively, and the signal distributions within the two ROIs were analyzed.

[0085] The result analysis of this embodiment is as follows:

[0086] Figure 2 As shown in A, the experimental results on the phantom using the traditional presaturation pulse (the first row) and the proposed pulse based on amplitude and off-resonance frequency modulation (the second row) are presented. Experiments with presaturation powers of 1.5 uT, 2.0 uT, 2.5 uT, and 3.0 uT were carried out in the embodiment respectively. Figure 2 B is the quantitatively mapped radiofrequency field intensity coefficient diagram of the phantom, which conforms to the mode where the radiofrequency field intensity at the center of the phantom is higher and the radiofrequency field intensity at the edge is lower under the circularly polarized excitation mode. Opposite CEST contrast inhomogeneous forms were observed in the results of 1.5 uT and 3.0 uT. At 1.5 uT, relatively higher CEST signals were presented in the region with higher radiofrequency field intensity, while at 3.0 uT, relatively lower CEST signals were presented in the region with higher radiofrequency field intensity. Figure 2 The second row in A shows the results of the modulation-based presaturation pulse invented by the present invention. Compared with the traditional saturation method ( Figure 2 the first row in A), it can reduce the influence of radiofrequency field intensity inhomogeneity and obtain a more uniform CEST image. This result can be further corroborated by Figure 2 the box plot in C. When the presaturation power is 1.5 uT, 2.5 uT, and 3.0 uT, there are significant signal distribution differences between the two ROI regions using the traditional saturation method, while using the proposed saturation method can greatly reduce the signal distribution differences between the two ROIs.

[0087] Figure 3 The average z-spectrum (the first and second rows) and MTR asym spectra within the ROI with the maximum and minimum 10% radiofrequency field intensities in the phantom are shown. The results show that there is no obvious change in the z-spectrum before and after the optimization of the presaturation pulse, while the average MTR asym spectral lines within the two ROI regions after optimization are closer to each other near 3.5 ppm compared with before optimization.

[0088] As Figure 4Shown are the experimental results on a healthy volunteer using the conventional saturation method and the invented pre-saturation method based on amplitude and off-resonance frequency modulation. As Figure 4 shown in A, in the region with a lower radiofrequency intensity in the prefrontal lobe, the conventional saturation method ( Figure 4 first row of A) shows an abnormal region with low CEST signal in the prefrontal lobe (such as Figure 4 the position indicated by the red arrow in A), while the invented saturation method ( Figure 4 second row of A) can reduce the influence of radiofrequency field intensity variation at the corresponding position without showing low CEST signal. Figure 4 The box plot of D shows that in the healthy volunteer experiment, there are significant signal distribution differences in the ROI regions with two different radiofrequency field intensities using the conventional saturation method, while using the invented saturation method can reduce the signal distribution difference between the two ROIs.

[0089] Figure 5 Shown are the average z-spectra ( Figure 5 first and second rows) and MTR asym spectra ( Figure 5 third and fourth rows) in two regions of interest (ROIs) with the maximum and minimum 10% radiofrequency field intensities in the region where the white matter volume fraction of the healthy volunteer is greater than 90%. The results show that there is no obvious change in the z-spectrum before and after the optimization of the pre-saturation pulse, while the optimized MTR asym spectral lines are closer in the two ROI regions near 3.5 ppm.

Claims

1. A method for optimizing pre-saturation parameters in magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation, characterized in that, Including the following steps: S1. For a target object to be subjected to magnetic resonance CEST imaging, use a multi-pool Bloch-McConnell mathematical simulation model to perform simulation according to the target CEST presaturation power. During the simulation process, simulate the change in the radiofrequency field intensity caused by the radiofrequency field inhomogeneity during the pulsed CEST presaturation process by adjusting the radiofrequency field intensity inhomogeneity factor of the CEST presaturation sub-pulse, obtain the z-spectrum under different CEST presaturation sub-pulse modulation parameters, and calculate the CEST effect value at the frequency of interest according to the z-spectrum; the CEST presaturation sub-pulse modulation parameters are composed of the amplitude and off-resonance frequency of each sub-pulse in the CEST presaturation module; S2. Using the CEST presaturation sub-pulse modulation parameters and the corresponding CEST effect values calculated by the simulation as sample data, and taking the minimum variation degree of the CEST effect value under the change of the radiofrequency field intensity during the CEST presaturation process as the optimization goal, perform non-linear optimization on the amplitude and off-resonance frequency of each sub-pulse in the CEST presaturation module to obtain the optimal modulation parameters of the CEST presaturation sub-pulse with the least influence of the radiofrequency field intensity change under the target CEST presaturation power.

2. The method for optimizing the pre-saturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation according to claim 1, wherein The multi-pool Bloch-McConnell mathematical simulation model is represented by the following formula: where expm represents the exponential function with the base of the natural constant e, and Δt represents the increment at time t, represent the magnetization vectors at times t + Δt and t respectively and matrix n is the number of solute clusters, n ≥ 1; represent the magnetization vector components of the free water clusters in the x, y, and z directions respectively, represent the magnetization vector components of the n solute clusters respectively. The magnetization vector components of the i-th solute cluster in the x, y, and z directions are respectively represents the magnetization vector component of the semi-solid magnetization transfer cluster in the z direction, while the magnetization vector components of the semi-solid magnetization transfer cluster in the x and y directions are negligible; the coefficient matrix where: where Δω w represents the difference between the off-resonance frequency ω of the applied CEST pre-saturation sub-pulse and the resonance frequency of the free water pool, and Δω i represents the difference between the off-resonance frequency ω of the applied CEST pre-saturation sub-pulse and the resonance frequency of the i-th solute pool; respectively represent the longitudinal relaxation rate and the transverse relaxation rate of the free water pool, respectively represent the longitudinal relaxation rate and the transverse relaxation rate of the i-th solute pool, respectively represent the longitudinal relaxation rate and the transverse relaxation rate of the semi-solid magnetization transfer pool; respectively represent the steady-state magnetization vectors of the free water pool, the i-th solute pool, and the semi-solid magnetization transfer pool; respectively represent the hydrogen proton exchange rates from the free water pool to the i-th solute pool and from the i-th solute pool to the free water pool; ω1 represents the angular frequency of the applied CEST pre-saturation sub-pulse, which is determined by the gyromagnetic ratio γ and the sub-pulse amplitude B1, ω1 = γ·B1, and the CEST pre-saturation sub-pulse amplitude B1 = B 1_weight ·B 1_modulation ·B 1_desired where B 1_desired represents the amplitude of the CEST pre-saturation sub-pulse to be applied, which is determined by the target CEST pre-saturation power corresponding to the current simulation; B 1_weight is a radio frequency field intensity non-uniformity factor within the range of [1 - ε, 1 + ε], which is used to simulate the change in the radio frequency field intensity caused by the radio frequency field non-uniformity, and ε is the preset change amplitude; B 1_modulation is the adjustment factor for the amplitude of the CEST pre-saturation sub-pulse to be optimized.

3. The method for optimizing the pre-saturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation according to claim 2, wherein, During the simulation process, for all sub-pulses in the CEST presaturation module, it is necessary to determine the B when the target CEST presaturation power corresponding to the current simulation is satisfied. 1_desired , and at the same time sample different values from the range of variation of B 1_weight and assign them to all sub-pulses to simulate the change in the RF field intensity caused by the RF field inhomogeneity. Then, different B 1_modulation and off-resonance frequencies ω are assigned to each sub-pulse to form different CEST presaturation sub-pulse modulation parameters. Finally, using the multi-population Bloch-McConnell mathematical simulation model, the excitation process of each sub-pulse in the CEST presaturation module in the time domain is simulated under different CEST presaturation sub-pulse modulation parameters to obtain CEST imaging data and calculate the z-spectrum.

4. The method for optimizing pre-saturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation according to claim 2, characterized in that During the simulation process, the inhomogeneity factor B of the radio frequency field intensity 1_weight varies in the range of 0.8 to 1.2, and the change step is 0.

05.

5. The method for optimizing pre-saturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation according to claim 1, wherein Among the CEST presaturation sub-pulse modulation parameters, the amplitude of the sub-pulse needs to set the optimized upper and lower bound ranges to ensure that the SAR value of the sequence does not exceed the system limit, and the off-resonance frequency of the sub-pulse needs to superimpose a modulation frequency on the basis of the original CEST polarization frequency, and the modulation frequency needs to set the optimized upper and lower bound ranges to ensure the frequency selectivity of CEST imaging.

6. The method for optimizing the pre-saturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation according to claim 1, wherein When performing non-linear optimization on the amplitude and off-resonance frequency of each sub-pulse in the CEST presaturation module, use a non-linear optimization process based on sequential quadratic programming, and the objective function used is as follows: where: coef pulsei represents the CEST pre-saturation sub-pulse modulation parameter, including the amplitude B1 and the off-resonance frequency ω of each sub-pulse in the CEST pre-saturation module; C(CEST(B1, ω target )) is the loss function, which is used to calculate the variation degree of the CEST effect value at the interested frequency ω 1_weight when B varies in [1 - ε, 1 + ε], and C(CEST(B1, ω target )) is defined as: target ​ ΔCEST = CEST(B1, ω target ) - CEST(B 1_desired , ω target ) where: CEST(B1, ω target ) represents the CEST effect value at the interested frequency ω target when the amplitude of the CEST pre-saturation sub-pulse is B1, and CEST(B 1_desired , ω target ) represents the CEST effect value at the interested frequency ω 1_desired when the amplitude of the CEST pre-saturation sub-pulse is B target ; Penalty term is defined as: Wherein: denotes the CEST effect value at the interested frequency ω when performing Bloch-McConnell simulation only using the free water pool and the amplitude of the CEST pre-saturation sub-pulse is B1; α is a weight parameter used to balance the penalty term and the loss function. target ​ 7. A pre-saturation method for magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation, characterized in that, Including: Obtain a pre-constructed parameter dictionary; the construction method of the parameter dictionary is: for different CEST presaturation powers, respectively according to the magnetic resonance CEST imaging presaturation parameter optimization method described in any one of claims 1 to 6, after obtaining the optimal modulation parameters of the CEST presaturation sub-pulse with the least influence of the radiofrequency field intensity change under each CEST presaturation power, associate and store the CEST presaturation power and the optimal modulation parameters of the CEST presaturation sub-pulse in the parameter dictionary; Before performing CEST imaging presaturation, according to the specified CEST presaturation power to be used, look up the corresponding optimal modulation parameters of the CEST presaturation sub-pulse in the parameter dictionary, and correspondingly set the amplitude and off-resonance frequency of each sub-pulse in the CEST presaturation module, and then execute the CEST presaturation module to acquire a CEST image with the least influence of the radiofrequency field intensity change.

8. A computer-readable storage medium, characterized in that, A computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the method for optimizing pre-saturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation according to any one of claims 1 to 6, or implements the pre-saturation method of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation according to claim 7.

9. A computer electronic device, characterized in that, It includes a memory and a processor; The memory is used to store a computer program; the processor is used to implement the method for optimizing pre-saturation parameters of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation according to any one of claims 1 to 6, or implement the pre-saturation method of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation according to claim 7 when executing the computer program.

10. A magnetic resonance imaging device, characterized in that, It includes a magnetic resonance scanner and a control unit. A computer program is stored in the control unit. When the computer program is executed, it is used to control the magnetic resonance scanner to acquire CEST images with the least influence of radiofrequency field intensity change according to the pre-saturation method of magnetic resonance CEST imaging based on amplitude and off-resonance frequency modulation according to claim 7.

Images