CEST data fitting method, device and medium based on extrapolated magnetization transfer signal

The NEMR method addresses interference effects in CEST imaging by using a two-pool model to fit BM equation parameters, improving lesion detection in CEST imaging.

JP7595379B2Active Publication Date: 2024-12-06ZHEJIANG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
JP2023566736
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Priority Date
2021-04-29
Filing Date
2021-12-22
Publication Date
2024-12-06
Estimated Expiration
2041-12-22

AI Technical Summary

Technical Problem

Existing CEST imaging methods struggle to accurately remove interference effects such as polymer magnetization transfer (MT) effects, as they assume Lorentzian curves that do not accurately represent these effects, and fail to account for unsteady system states.

Method used

A CEST data fitting method based on extrapolated magnetization transfer signals (NEMR) that incorporates MT linearity into the Bloch-McConnell equation, using a two-pool model to fit parameters and simulate the BM equation, allowing for accurate removal of interference effects even in unsteady states.

Benefits of technology

The NEMR method effectively removes interference signals by simulating the BM equation, enabling accurate restoration of physiological parameters and enhancing the visibility of lesions in CEST imaging, particularly in ischemic stroke patients.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007595379000103
    Figure 0007595379000103
  • Figure 0007595379000104
    Figure 0007595379000104
  • Figure 0007595379000105
    Figure 0007595379000105
Patent Text Reader

Abstract

The present invention provides a method, apparatus and medium for CEST data fitting based on extrapolated magnetization transfer signals in the field of magnetic resonance imaging. [Solution] The CEST data fitting method based on the extrapolated magnetization transfer signal of the present invention can obtain the MT effect by fitting the model represented by the Bloch-McConnell equation with the linearity of the MT effect, and can eliminate the interference effect by calculating the difference between the fitted curve and the curve collected from the experiment. The fitting method based on Bloch-McConnell of the present invention can take into account the RF pulse waveform actually used in the experiment and can be used in the non-steady state of the system.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical Field]

[0001] The present invention relates to the technical field of magnetic resonance imaging, and more particularly to a method, apparatus and medium for fitting CEST data based on extrapolated magnetization transfer signals. [Background technology]

[0002] Chemical Exchange Saturation Transfer (CEST) imaging is a magnetic resonance molecular imaging technique that can detect proteins and polypeptides released in living organisms. CEST imaging has already been applied in several clinical fields, such as brain tumor grading, monitoring early treatment response to radiochemotherapy, tumor marker prediction, and ischemic stroke detection. However, CEST signals are susceptible to many effects, and researchers both in China and abroad have proposed many CEST signal calculation methods to eliminate these interference effects. These methods can basically be divided into two types: obtaining the final signal by calculating the difference between the reference signal and the marker signal, and obtaining the final signal by model fitting. Asymmetric Magnetization Transfer (MTR) asym The MTR (Magnetization transfer ration asymmetry) assumes that (1) the reference signal does not contain the CEST effect, (2) the marker signal contains the CEST effect, and (3) other effects are equal in the reference and marker signals. Then, subtracting the experimental signal from the reference signal removes the other effects and retains only the CEST effect. However, other interfering effects, such as polymer magnetization transfer (MT) effects, are not all equal in the reference and marker signals, so the MTR aysmThe signals obtained by this method include other effects. Multi-pool Lorentzian fitting assumes that each effect is represented by a single Lorentzian curve, and these Lorentzian curves are superimposed to form the experimentally collected curve. The values ​​of each effect can be obtained by fitting the curves collected from the experiment. However, the curves of other effects are not all Lorentzian; for example, MT, whose curve is hyper-Lorentzian. The curve fitted from the extrapolated magnetization transfer signal (EMR, Extrapolated Semisolid Magnetization Transfer Reference) represents effects other than the CEST effect. The CEST effect can be obtained by calculating the difference between the curve fitted from EMR and the experimental curve. EMR assumes that the saturation pulse used in the experiment is a square wave and that the system reaches a steady state, but these two conditions are not always met. Summary of the Invention [Problem to be solved by the invention]

[0003] Although the model fitting methods in the prior art approximate experimental conditions, such approximations are not valid under many experimental conditions. Therefore, the object of the present invention is to solve the problems of the prior art and to provide a CEST numerical fit method based on extrapolated magnetization transfer signals (NEMR, Numerical Fit of Extrapolated Semisolid Magnetization Transfer Reference Signal) that is robust and suitable for multiple experimental conditions in terms of eliminating interference effects in CEST imaging. [Means for solving the problem]

[0004] The specific technical solutions used in the present invention are as follows:

[0005] In a first aspect, the present invention provides a CEST data fitting method based on extrapolated magnetization transfer signals, which sequentially performs S1 to S3 for each voxel in a CEST image to obtain a CEST effect value from which background interference signals of each voxel have been removed, and for S1 to S3, In S1, the original Z spectrum is subjected to main magnetic field frequency B0 offset correction to obtain a corrected Z spectrum; In S2, a part of the data points including only the MT effect in the corrected Z spectrum is taken as fitting sample data, and a two-pool model in which the initial value and upper and lower limits of the parameter to be determined are preset is fitted to obtain the fitting value of the parameter to be determined; The two-pool model is shown in Equation 8 below:

number

[0006] In a preferred example of the first aspect, in S2, fitting of the parameters to be determined in the two-pool model is performed twice. During the first fitting, parameter optimization is performed using the fitting sample data for each of the five parameters to be determined within their predetermined upper and lower limit ranges to obtain first fitting values. During the second fitting, upper and lower limits narrower than the upper and lower limit range of the first fitting are selected, with the first fitting value as the center, and the same fitting sample data as in the first fitting are used again to perform a second, more accurate fitting of the two-pool model to obtain final fitting values ​​for the parameters to be determined.

[0007] More preferably, in the second fitting, only some parameters are selected from the five parameters to be determined of the two-pool model and their upper and lower limits are adjusted, and the upper and lower limits of the remaining parameters are maintained to be consistent with those of the first fitting; More preferably, when fitting for the second time, the upper and lower limit parameter combinations are TIFF0007595379000018.tif829 or TIFF0007595379000019.tif716 or It needs to be changed to TIFF0007595379000020.tif821.

[0008] As a preferred example of the two-time fitting scheme in the first embodiment, in S2, the data points containing only the MT effect are data points whose frequency is within the range of 80 to 6 ppm, preferably data points whose frequency is within the range of 80 to 20 ppm.

[0009] As a preferred example of the two-time fitting scheme in the first embodiment, in the step S2, when fitting for the first time, the parameters to be determined are TIFF0007595379000021.tif849 has a preset upper limit of [1.5,0.25,12e -6,0.16,40], and the preset lower bound is [0.6,0.065,8e -6 ,0.001,20].

[0010] As a preferred example of the two-fitting scheme in the first embodiment, in S2, when fitting for the second time, for a parameter whose upper and lower limits should be changed, the upper limit value is changed to 120% to 150% of the first fitting value of the parameter, and the lower limit value is changed to 80% to 50% of the first fitting value of the parameter, and further, it is preferable that the upper limit value is changed to 120% of the first fitting value of the parameter, and the lower limit value is changed to 80% of the first fitting value of the parameter.

[0011] In a preferred example of the two-step fitting scheme in the first embodiment, in S2, the absorption rate of the radiofrequency pulse by the MT population is TIFF0007595379000022.tif955, TIFF0007595379000023.tif715 is a linear function, including Lorentzian, Gaussian, or hyper-Lorentzian.

[0012] More preferably, in step S3, the frequency of interest is 3.5 ppm or -3.5 ppm, and the CEST effect value corresponding to the frequency of interest of 3.5 ppm is APT # The CEST effect value corresponding to the frequency of interest of -3.5 ppm is the NOE # is.

[0013] In a second aspect, the present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the method for fitting CEST data based on extrapolated magnetization transfer signals according to any one of the first aspects.

[0014] In a third aspect, the present invention provides an apparatus for fitting CEST data based on a numerical value of an extrapolated magnetization transfer signal, the apparatus comprising a processor and a storage medium, wherein a computer program is stored on the storage medium, and when the computer program is executed by the processor, the method for fitting CEST data based on an extrapolated magnetization transfer signal according to any one of the first aspects is performed.

[0015] In a fourth aspect, the present invention provides a magnetic resonance imaging apparatus for removing CEST imaging interference effects, comprising a magnetic resonance scanner and a control unit, the magnetic resonance scanner being for acquiring a CEST image by magnetic resonance CEST imaging, the control unit being capable of acquiring the CEST image, and a computer program being stored in the control unit, the computer program being configured to, when executed, perform the CEST data fitting method based on the extrapolated magnetization transfer signal according to any one of the first aspect, and output a CEST effect value from which background interference signals of each voxel have been removed. [Effects of the Invention]

[0016] Compared with the prior art, the present invention has the following beneficial effects:

[0017] The fitting method of the present invention essentially solves the BM equation step by step, and this method has two advantages: First, because NEMR repeats in the time dimension, the waveform of the actual high-frequency pulse used can be incorporated into the solving process rather than assuming a square wave; Second, the NEMR method explains the CEST effect by simulating the BM equation rather than using a simplified analytical expression, and can be applied even in the non-steady state of the system. The BM equation is the most basic equation that explains the CEST effect, and physiological parameter values ​​can be accurately restored by performing simulations and iterations based on this. [Brief explanation of the drawings]

[0018] [Figure 1] FIG. 1 is a flowchart of the fitting method of the present invention. [Figure 2] FIG. 2 is a diagram of one curve collected from an experiment in the Examples and its fitting curve. [Figure 3] FIG. 3 shows the T2 weighting map, diffusion weighting map, APT# map and NOE# map of one ischemic stroke patient in the example. DETAILED DESCRIPTION OF THE INVENTION

[0019] The present invention will be further discussed and explained below by means of specific embodiments with reference to the drawings.

[0020] The method for eliminating CEST imaging interference effects according to the present invention is called Numerical Fit of Extrapolated Semisolid Magnetization Transfer Reference Signal (NEMR). This CEST data fitting method based on the extrapolated magnetization transfer signal incorporates the MT effect line shape into the Bloch-McConnell (BM) equation, and the MT effect can be obtained by fitting the model represented by this equation. The interference effects can be eliminated by calculating the difference between the fitted curve and the experimental curve.

[0021] The purpose of this invention is to incorporate the MT line shape into the BM equation by reasonably approximating the BM equation, and form a two-pool model for subsequent fitting. First, the principle of the two-pool model is explained in detail below.

[0022] The magnetic resonance imaging process may be described by the Bloch equations in Equation 1 below:

number

number

[0023] Since the T2 of the MT population is in the microsecond range, M xb , M yb will quickly reach a steady state, i.e.,

number

number

number

number

number

number

[0024] The solution of the homogeneous differential equation is the two-pool model of the present invention, which has five unknowns: TIFF0007595379000046.tif748, and the remaining parameters are all known quantities. By fitting this model, a fitting curve including the MT effect can be obtained. By calculating the difference between this fitting curve and the Z spectrum curve after main magnetic field frequency B0 offset correction, a clearer CEST effect can be obtained.

[0025] Although the present invention may be used to obtain different CEST effect values, since different CEST effect values ​​correspond to different frequencies in the curve, for ease of illustration, in the present invention, a frequency that can reflect a certain CEST effect in the Z spectrum is referred to as a frequency of interest. For example, if the CEST effect value to be calculated is APT # (Amide proton transfer signal from NEMR analysis, amide proton transfer signal obtained by analysis of extrapolated magnetization transfer signal) value, the corresponding frequency of interest is 3.5 ppm, and the CEST effect value to be calculated is NOE #(Numerical Fit of Extrapolated Semisolid Magnetization Transfer Reference Signal, a nuclear Overhauser effect signal obtained by analyzing the extrapolated magnetization transfer signal), the corresponding frequency of interest is -3.5 ppm. Of course, if there is another frequency in the Z spectrum that can also reflect a certain CEST effect, that frequency can also be used as the frequency of interest. By calculating the difference at the frequency of interest in the fitting curve and the corrected Z spectrum, a CEST effect value can be obtained from which the background interference signal corresponding to the frequency of interest has been removed. In the process of implementing the present invention, one CEST effect value may be calculated at a time, or multiple CEST effect values ​​may be calculated simultaneously; this is not a limitation.

[0026] Based on the above principle, the present invention provides two methods for fitting CEST data based on extrapolated magnetization transfer signals. As shown in Figures 1(a) and 1(b), the difference between the two methods is the parameter fitting method of the two-pool model. The first method is a single-step fitting, while the second method is a double-step fitting. The remaining methods are the same. The two methods are described in detail below with examples.

[0027] Example 1 In this embodiment, as shown in Figure 1(a), the process is performed by the NEMR fitting method using the first type of fitting method. This method first acquires a CEST image, then sequentially performs S1 to S3 for each voxel in the CEST image to obtain the APT of each voxel. # Values ​​and NOEs # This will get the value of APT # Map and NOE # A map needs to be constructed. The specific execution process of S1 to S3 is explained in detail below.

[0028] In step S1, the acquired original Z spectrum is subjected to main magnetic field frequency B0 offset correction to obtain a corrected Z spectrum. The main magnetic field frequency B0 offset correction is a conventional technique, and a detailed description thereof will be omitted.

[0029] In S2, a portion of the data points in the corrected Z spectrum that contain only the MT effect are taken as fitting sample data, and a two-pool model with preset initial values ​​and upper and lower limits for the parameters to be determined is fitted to obtain the fitting values ​​for the parameters to be determined.

[0030] As described above, the two-pool model in the present invention is represented by the following Equation 8:

number

[0031] In the two-pool model, five parameters are determined. Subsequent parameter fitting can be performed on TIFF0007595379000065.tif749. Some data points containing only MT effects may be analytically determined based on the actual original Z spectrum. In this example, data points within the frequency range of 80 to 20 ppm are preferred.

[0032] In the first method, the five parameters to be determined are directly determined by fitting sample data. TIFF0007595379000066.tif849 may be fitted, and fitting values ​​of the five parameters to be determined that maximize the fitting degree of the entire two-pool model equation may be obtained from the upper and lower limit ranges of each of the five parameters to be determined. Specific fitting methods may be performed using software such as MATLAB, SPSS, or other conventional techniques, and the fitting degree may be expressed as the minimum mean square error.

[0033] When fitting, the parameters to be determined The preset upper and lower limits and initial values ​​of TIFF0007595379000067.tif749 can be adjusted according to the actual situation, as long as they are in accordance with the normal value range of each parameter. The preset upper limit of TIFF0007595379000068.tif749 is [1.5,0.25,12e -6 ,0.16,40], and the preset lower bound is [0.6,0.065,8e -6 , 0.001, 20], and the initial values ​​are set as the respective lower limits.

[0034] In S3, a fitting curve including the MT effect can be obtained based on the two-pool model by substituting five fitting values. The frequency ω in the two-pool model is the abscissa of the curve, and M in the two-pool model za is the ordinate of the curve. By subtracting the Z spectrum after main magnetic field B frequency correction from the fitting curve, a difference curve can be obtained. In this difference curve, the ordinate values ​​corresponding to 3.5 ppm and -3.5 ppm are taken as the abscissa and -3.5 ppm, respectively, to determine the APT # Values ​​and NOEs # Of course, for simplicity, it is not necessary to obtain a difference curve, and if the difference between the ordinate value corresponding to 3.5 ppm in the fitting curve and the ordinate value corresponding to 3.5 ppm in the Z spectrum after main magnetic field B0 frequency correction can be calculated, the APT of the current voxel can be calculated. #The NOE of the current voxel can be calculated by calculating the difference between the ordinate value corresponding to -3.5 ppm in the fitting curve and the ordinate value corresponding to -3.5 ppm in the Z spectrum after main magnetic field B frequency correction. # It can be a value.

[0035] Each voxel in the entire CEST image corresponds to an APT # Values ​​and NOEs # If you get the value, one APT # Map and one NOE # Maps can be created and APT # The value at each location in the map corresponds to the voxel at that location. # value, and NOE # The value at each location in the map corresponds to the voxel at that location. # value.

[0036] To further demonstrate the technical effect of the method, the NEMR fitting method is tested on a cerebral CEST image of an ischemic stroke patient as an example. In this example, the CEST imaging sequence used to acquire the image includes a CEST saturation module with 10 Gaussian waveform saturation pulses, each with a pulse duration of approximately 100 ms and an amplitude of 1 µT, a spectral pre-saturation inversion recovery fat suppression module, and a fast spin echo acquisition module.

[0037] As shown in Figure 1(a), first, the main magnetic field frequency B0 offset correction is performed on the collected Z spectrum data (frequency range 80 to -6 ppm as shown in the experimental curve in Figure 2), and then the above-mentioned NEMR fitting is performed. In the first fitting, the data in the frequency range 80 to 20 ppm is used as fitting sample data (the points used for fitting are shown in Figure 2), and the unknown parameters TIFF0007595379000069.tif748 has an upper limit of [1.5,0.25,12e -6 ,0.16,40], and the lower bound is [0.6,0.065,8e -6,0.001,20], where the initial value is equal to the lower limit. The final fitting curve is shown in Figure 2.

[0038] The experimental results of this example are shown in Figure 3, which shows the APT fitted by the NEMR method of one-step fitting. # Map and NOE # The map not only clearly displays lesions visible in other modes (lesions in T2-weighted images, white solid arrows), but also lesions invisible in other modes (NOE # It can also be seen that the lesions on the map (white dotted arrows) can be displayed.

[0039] Example 2 In this embodiment, as shown in Figure 1(b), the process is performed by the NEMR method using the second type of fitting method. Similarly, the method first acquires a CEST image, and then sequentially performs S1 to S3 for each voxel in the CEST image to obtain the APT# value and NOE# value for each voxel, thereby obtaining the APT # Map and NOE # You need to configure the map.

[0040] Steps S1 and S3 of this embodiment are the same as those of the first embodiment, with the only difference being that fitting of the parameters to be determined in the two-pool model is performed twice in step S2. The first fitting is the same as in Example 1. First, the five parameters to be determined are optimized within the preset upper and lower limit ranges using the fitting sample data to obtain the first fitting values. The initial values ​​of the five parameters to be determined are also set as their respective lower limits.

[0041] However, multi-parameter fitting is heavily affected by the upper and lower limits of variables, resulting in some parameters being equal to the given limit values. Therefore, this embodiment must alleviate this problem by configuring a second step of fitting. During the second fitting, the first fitting value can be used as the center, and upper and lower limits narrower than the upper and lower limit range of the first fitting can be selected. The same fitting sample data as in the first fitting can be used again to perform a second, more accurate fitting of the two-pool model to obtain the final fitting value of the target parameter. In a specific implementation process, it is not necessary to adjust the upper and lower limits of all five target parameters of the two-pool model. The upper and lower limits of some of the five target parameters of the two-pool model can be further expanded, while the upper and lower limits of the remaining parameters can still be maintained consistent with the first fitting. During the second fitting, this embodiment uses a parameter combination of upper and lower limits. It needs to be changed to TIFF0007595379000070.tif729.

[0042] In this embodiment, some parameters (e.g., TIFF0007595379000071.tif728) to the neighborhood of the first result allows these parameters to be varied within a narrower range, thereby reducing the impact that improper setting of the upper and lower limits had on the final results.

[0043] In addition, the degree of change of the upper and lower limit values ​​of the first fitting value, which should be changed, should not be too large or too small. For TIFF0007595379000072.tif728, it is preferable to change the upper limit of the parameter used for the second fitting to 120% of the first fitting value of that parameter, and the lower limit of the parameter used for the second fitting to 80% of the first fitting value of that parameter, thereby further optimizing these parameters to be closer to the first fitting value.

[0044] To further demonstrate the technical effect of the above method, the above NEMR fitting method is tested on the cerebral CEST images of an ischemic stroke patient as an example. In this example, the CEST imaging sequence used to acquire the images is the same as in Example 1.

[0045] As shown in Figure 1(b), first, the main magnetic field frequency B0 offset correction is performed on the collected data, and then NEMR fitting is performed. In the first fitting, data from frequencies 80 to 20 ppm are used as fitting sample data, and the unknown fitting target parameters are calculated. TIFF0007595379000073.tif848 has an upper limit of [1.5, 0.25, 12e-6, 0.16, 40] and a lower limit of [0.6, 0.065, 8e-6, 0.001, 20], and the initial value is equal to the lower limit. When fitting for the second time, use the same fitting sample data as the first fitting, The upper limit of TIFF0007595379000074.tif729 was obtained the first time 120% of TIFF0007595379000075.tif729, and the lower limit is the first acquired TIFF0007595379000076.tif829, and keep the upper and lower limits of the remaining parameters consistent with the first fitting, and set the initial values ​​of all parameters as the lower limits of the parameter values.

[0046] The experimental results of this example are shown in Figure 3, which shows the APT fitted by the two-step NEMR method. # Map and NOE # The map not only clearly displays lesions visible in other modes (lesions in T2-weighted images, white solid arrows), but also lesions invisible in other modes (NOE # It is also possible to display the lesions on the map (white dotted arrows), and it can be seen that the display effect is superior to that of Example 1.

[0047] It should be noted that the above two embodiments are merely two examples of the present invention, and are not limited thereto. For example, the selected frequency range of the fitting sample data in S2 of the present invention may be other ranges, such as 80 to 20 ppm, 80 to 6 ppm, etc. Similarly, when performing parameter fitting, the upper and lower limit restrictions of the parameters may be adjusted according to the actual situation, and the upper limit value may be changed to 120% to 150% of the first fitting value of the parameter obtained in the first fitting, and the lower limit value may be changed to 80% to 50% of the first fitting value of the parameter. Furthermore, when fitting for the second time, there may be various parameter combinations for which the upper and lower limit values ​​should be changed, such as: In addition to TIFF0007595379000077.tif829, TIFF0007595379000078.tif815 or TIFF0007595379000079.tif821, etc. may be selected. When fitting, the initial values ​​of the parameters may be set to reasonable values ​​within the upper and lower limit ranges, and several different candidate values ​​may be set randomly or uniformly within the upper and lower limit ranges, and the candidate value with the smallest obtained fitting error may be selected as the final initial value.

[0048] In another preferred embodiment of the present invention, the CEST data fitting method based on the extrapolated magnetization transfer signal may be stored in a computer-readable storage medium in the form of a computer program, which, when invoked and executed by a processor, can execute the CEST data fitting method based on the extrapolated magnetization transfer signal according to the method.

[0049] Similarly, in another preferred embodiment of the present invention, there is further provided a CEST data fitting apparatus for fitting based on a numerical value of an extrapolated magnetization transfer signal, the apparatus comprising a processor and a storage medium, wherein a computer program is stored in the storage medium, and when the computer program is executed by the processor, the apparatus performs the CEST data fitting method based on the extrapolated magnetization transfer signal.

[0050] The computer-readable storage medium is generally provided in the form of memory hardware, which may include random access memory (RAM) and non-volatile memory (NVM), such as at least one magnetic disk memory.

[0051] The processor that processes the above-mentioned programs may be a general-purpose processor including a central processing unit (CPU) or a network processor (NP), or may be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component.

[0052] Of course, with the widespread application of cloud servers, the above software programs may be installed on a cloud platform to provide corresponding services, and therefore the computer-readable storage medium is not limited to the form of local hardware.

[0053] In another preferred embodiment, the CEST data fitting method based on the extrapolated magnetization transfer signal may be integrated into the control unit of a magnetic resonance imaging apparatus in the form of a program. The magnetic resonance imaging apparatus should include a conventional magnetic resonance scanner and a control unit. The magnetic resonance scanner may be implemented according to conventional technology and is a mature commercial product, so detailed description is omitted. The control unit should store the computer program and also include imaging sequences and other software programs required to perform CEST imaging. The magnetic resonance scanner is used to acquire CEST images by magnetic resonance CEST imaging, while the control unit can acquire CEST images acquired by the magnetic resonance scanner. The control unit also stores the computer program, which, when executed, executes the CEST data fitting method based on the extrapolated magnetization transfer signal to output a CEST effect value with the background interference signal of each voxel removed.

[0054] Those skilled in the art will understand that each module and function of the present invention can be implemented by a circuit, other hardware, or executable program code, as long as the corresponding function is performed. When code is used, the code may be stored in a storage device and executed by a corresponding element of a computing device. Implementation of the present invention is not limited to any specific combination of hardware and software. The hardware models used in the present invention may be commercially available products and may be selected according to the actual user needs. Of course, the magnetic resonance imaging sequence and device may also require other necessary hardware or software, and a detailed description thereof will be omitted here.

[0055] The above-described embodiments are merely preferred solutions of the present invention and are not intended to limit the present invention. Those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Therefore, any technical solutions obtained by equivalent substitution or equivalent conversion are included in the scope of protection of the present invention.

Claims

1. 1. A method for fitting CEST data based on an extrapolated magnetization transfer signal, comprising: Steps S1 to S3 are sequentially performed for each voxel in the CEST image to obtain a CEST effect value for each voxel after removing the background interference signal, and the steps S1 to S3 are as follows: In the step S1, a main magnetic field frequency B0 offset correction is performed on the acquired original Z spectrum to acquire a corrected Z spectrum; In step S2, a part of the data points including only the MT effect in the corrected Z spectrum is taken as fitting sample data, and a two-pool model in which the initial value and upper and lower limits of the parameter to be determined are preset is fitted to obtain a fitting value of the parameter to be determined; The two-pool model is shown in Equation 8 below: [0080] In the formula, exp represents an exponential function with the natural constant e as the base, 【number】 indicates the increment of time t, 【number】 are respectively 【number】 Magnetization vector at time t 【number】 where: 【number】 and the matrix 【number】 are the x-direction magnetization intensity component of the water group, the y-direction magnetization intensity component of the water group, the z-direction magnetization intensity component of the water group, and the z-direction magnetization intensity component of the MT group, respectively, and the coefficient matrix is 【number】 and the matrix is 【number】 and the vector is 【number】 and 【number】 is the difference between the frequency of the externally applied high-frequency pulse and the frequency of water, and ω 1 indicates the intensity of the applied high frequency pulse, and k ab represents the rate of conversion from the water population to the MT population, and k ba indicates the rate of conversion from the MT population to the water population, and R rfb is the rate of absorption of the radiofrequency pulse by the MT population, 【number】 represents the difference between the frequency of the externally applied high frequency pulse and the frequency of the MT, 【number】 and 【number】 are the longitudinal and transverse relaxation rates of the water ensemble, respectively, 【number】 and 【number】 are the longitudinal and transverse relaxation rates of the MT population, respectively, and M 0a and M. 0b are the steady-state magnetization intensities of the water and MT ensembles, respectively, The parameter to be determined in the two-pool model is 【number】 and In step S3, a fitting curve including the MT effect is obtained by the two-pool model into which the fitting value is substituted, and the difference between the fitting curve and the corrected Z spectrum at the frequency of interest is set as a CEST effect value from which background interference signals have been removed.

13. A method for fitting CEST data based on an extrapolated magnetization transfer signal, comprising:

2. In step S2, fitting of the parameters to be determined in the two-pool model is performed twice. In the first fitting, parameter optimization is performed for the five parameters to be determined within the preset upper and lower limit ranges using the fitting sample data to obtain a first fitting value. In the second fitting, upper and lower limits narrower than the upper and lower limit range of the first fitting are selected with the first fitting value as the center, and a second fitting with higher accuracy is performed on the two-pool model using the same fitting sample data as in the first fitting to obtain a final fitting value of the parameters to be determined.

2. The method for fitting CEST data based on extrapolated magnetization transfer signals according to claim 1.

3. In the step S2, when fitting for the second time, only some of the five parameters to be determined in the two-pool model are selected and their upper and lower limits are adjusted, and the upper and lower limits of the remaining parameters are maintained so as to match those of the first fitting.

3. The method for fitting CEST data based on extrapolated magnetization transfer signals according to claim 2.

4. In the step S2, when fitting for the second time, the upper and lower limit parameter combinations are or or needs to be changed to 4. The method for fitting CEST data based on extrapolated magnetization transfer signals according to claim 3.

5. In step S2, the data points containing only the MT effect are those whose frequencies are within the range of 80 to 6 ppm.

2. The method for fitting CEST data based on extrapolated magnetization transfer signals according to claim 1.

6. In step S2, the data points including only the MT effect are data points whose frequencies are within the range of 80 to 20 ppm.

6. The method for fitting CEST data based on extrapolated magnetization transfer signals according to claim 5.

7. In step S2, when fitting for the first time, the parameters to be determined are The upper limit is set to [1.5, 0.25, 12e -6 , 0.16, 40], and the preset lower limit is [0.6, 0.065, 8e -6 , 0.001, 20] 3. The method for fitting CEST data based on extrapolated magnetization transfer signals according to claim 2.

8. In step S2, when fitting for the second time, for parameters whose upper and lower limits should be changed, the upper limit value is changed to 120% to 150% of the first fitting value of the parameter, and the lower limit value is changed to 80% to 50% of the first fitting value of the parameter.

5. The method for fitting CEST data based on extrapolated magnetization transfer signals according to claim 4.

9. In step S2, the absorption rate of the radio frequency pulse by the MT population is and is a linear function, including Lorentzian, Gaussian, or super-Lorentzian 3. The method for fitting CEST data based on extrapolated magnetization transfer signals according to claim 2.

10. In step S3, the frequency of interest is 3.5 ppm or -3.5 ppm, and the CEST effect value corresponding to the frequency of interest of 3.5 ppm is APT # value, and the CEST effect value corresponding to the frequency of interest of -3.5 ppm is the NOE # is The method for fitting CEST data based on extrapolated magnetization transfer signals according to claim 9.

11. 1. A computer-readable storage medium, comprising: When a computer program is stored in the storage medium and executed by a processor, the method for fitting CEST data based on an extrapolated magnetization transfer signal according to any one of claims 1 to 10 is executed. A computer-readable storage medium comprising:

12. A CEST data fitting device that performs fitting based on the numerical value of an extrapolated magnetization transfer signal, The method for fitting CEST data based on extrapolated magnetization transfer signals according to any one of claims 1 to 10 is performed by a processor and a storage medium, the storage medium storing a computer program, and the computer program being executed by the processor. A CEST data fitting apparatus for fitting based on the numerical value of an extrapolated magnetization transfer signal.

13. 1. A magnetic resonance imaging apparatus for eliminating CEST imaging interference effects, comprising: A magnetic resonance scanner and a control unit are provided, the magnetic resonance scanner is for acquiring a CEST image by magnetic resonance CEST imaging, the control unit is capable of acquiring the CEST image, and a computer program is stored in the control unit, and when the computer program is executed, the CEST data fitting method based on the extrapolated magnetization transfer signal according to any one of claims 1 to 10 is performed, and a CEST effect value from which a background interference signal of each voxel has been removed is output.

2. A magnetic resonance imaging apparatus for eliminating CEST imaging interference effects.