Method for optimizing flip angles in magnetic resonance variable flip angle sequence, CEST imaging method, medium, and device

By optimizing the flip angle in whole-brain CEST imaging, using extended phase map and gradient descent algorithm, the problem of low signal-to-noise ratio is solved, the imaging quality is improved, and the image blur caused by noise amplification is avoided, and more efficient signal reading is achieved out.

WO2024235015A1PCT designated stage expired Publication Date: 2024-11-21ZHEJIANG UNIV
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
PCT/CN2024/090877
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-15
Filing Date
2024-04-30
Publication Date
2024-11-21

AI Technical Summary

Technical Problem

In the existing whole-brain CEST imaging methods, the flip angle is not optimized to cause the imaging signal-to-noise ratio (SNR) to be low, and traditional methods are prone to amplification of noise when increasing the signal, affecting the imaging quality.

Method used

By modeling the signal-to-noise ratio maximization problem as a flip angle optimization problem, the objective function is represented by an extended phase map (EPG), and the flip angle is optimized through iterative solution and gradient descent algorithm, the noise amplification caused by signal filtering is avoided, combined with Resolution penalty items to take into account both signal-to-noise ratio and resolution.

Benefits of technology

The signal-to-noise ratio of magnetic resonance imaging is improved, image blur caused by noise amplification is avoided, and more efficient signal reading is achieved. At the same time, no manual preset window function is required, which improves imaging quality.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN2024090877_21112024_PF_FP_ABST
    Figure CN2024090877_21112024_PF_FP_ABST
Patent Text Reader

Abstract

Disclosed in the present invention are a method for optimizing flip angles in a magnetic resonance variable flip angle sequence, a CEST imaging method, a medium, and a device. According to the present invention, a signal-to-noise ratio (SNR) enhancement problem is modeled as a flip angle optimization problem, and a total objective function composed of an SNR maximization objective term and a resolution penalty term is constructed, wherein the total objective function can be solved for an optimal solution by finding derivatives with respect to flip angles, so as to obtain an optimal flip angle capable of maximizing SNR. In the present invention, the objective function comprises the resolution penalty term, so that resolution can also considered in variable flip angle CEST when the SNR is optimized. Compared with a conventional filtering method, the present invention avoids noise amplification caused by filtering, and does not require manual presetting of a window function.
Need to check novelty before this filing date? Find Prior Art

Description

Magnetic resonance variable flip angle sequence flip angle optimization method, CEST imaging method, medium and equipment Technical Field

[0001] The present invention belongs to the field of magnetic resonance technology, and in particular relates to a magnetic resonance variable flip angle sequence flip angle optimization method, a CEST imaging method, a medium and a device. Background Art

[0002] Chemical exchange saturation transfer (CEST) imaging is a magnetic resonance molecular imaging technique that can detect a variety of metabolites in the human body, such as glucose, free proteins, and peptides. The signal-to-noise ratio (SNR) of CEST imaging is crucial to image quality. Because CEST imaging relies on the attenuation of water signals, its inherent SNR is low. In typical CEST imaging applications, such as brain tumors, the contrast between normal tissue and lesions is only 2% to 3%. This low SNR further reduces the contrast between the two.

[0003] Furthermore, for CEST imaging to be clinically valuable, its imaging speed and spatial coverage need to be close to those of existing structural imaging techniques. Gradient echo-based sequences, such as gradient echo (GRE) and echo planar imaging (EPI), have fast imaging speeds but are susceptible to artifacts at the tissue-air interface due to magnetic susceptibility. Researchers have proposed a whole-brain CEST imaging sequence with variable flip angles (SPACE). This sequence, SPACE-CEST, is based on turbo spin echo (TSE) and is unaffected by magnetic susceptibility. By reducing the flip angle (FA), it effectively extends the echo train length (ETL), shortening acquisition time while simultaneously covering the entire brain. However, the constant flip angle 120° (CFA120°) readout mode used in the SPACE-CEST sequence has a non-optimized flip angle and low SNR efficiency. In addition, the higher flip angle it uses results in a high specific absorption rate (SAR).

[0004] In the prior art, some researchers have combined variable flip angles with k-space filtering to improve the signal-to-noise ratio. This method uses a preset target window, defines the signal-to-noise ratio as a function of the flip angle and the target window, and optimizes the flip angle through iterative optimization to maximize the signal-to-noise ratio. However, this method involves filtering the k-space signal. For low signals later in the echo chain, the filtering multiplies them by a large coefficient, which amplifies the noise while improving the signal. Furthermore, the choice of target window will also affect the SNR.

[0005] Summary of the Invention

[0006] The purpose of the present invention is to solve the problem that the flip angle used in the existing whole-brain CEST imaging method with variable flip angle CEST is not optimized, resulting in low imaging SNR, and to provide a flip angle optimization method for a magnetic resonance imaging sequence with variable flip angle, a CEST imaging method, a medium and a device.

[0007] The specific technical solutions adopted in the present invention are as follows:

[0008] In a first aspect, the present invention provides a method for optimizing the flip angle of a magnetic resonance variable flip angle sequence, comprising:

[0009] S1. For a magnetic resonance variable flip angle sequence with an echo train length of N-1, the flip angle optimization problem of its N-1 refocusing pulses is modeled as the following objective function:

[0010] st0≤FA≤180°

[0011] Where: s i is the signal amplitude without attenuation after the i-th refocusing pulse is applied (which is caused only by the flip angle and relaxation); w i is the signal attenuation coefficient of the i-th refocusing pulse (attenuation due to dephasing, phase encoding, etc.); FAs = [α1, α2, ..., α N-1 ] is the set of flip angles to be optimized, FA represents the flip angle α of any i-th refocusing pulse i , its optimization range constraint is [0°, 180°]; a and b are two weight coefficients in the objective function, and both a and b are greater than 0;

[0012] S2. Expressing the objective function with an extended phase graph (EPG) and then performing iterative solution to obtain the optimal solution of the optimized flip angle set as the flip angle setting value of N-1 refocusing pulses in the magnetic resonance variable flip angle sequence.

[0013] As a preferred embodiment of the first aspect, each round of iterative process for solving the objective function is as follows:

[0014] S21, the objective function The three parts are split and represented by extended phase diagrams, where:

[0015] Expressed as g using the expanded phase diagram:

[0016] Expressed as h using the expanded phase diagram:

[0017] Expressed as m using the expanded phase diagram:

[0018] Where: C is a slave EPG state f i The matrix from which the signal is extracted is a matrix where only the second row and the second column are 1 and the rest of the elements are 0; D i For a slave EPG state f i Extract the signal matrix, only the second row and second column in the matrix is ​​w i 2 The rest of the elements are 0;

[0019] S22. Derivatives of the function parts represented by each extended phase diagram are obtained for N-1 flip angles using the adjoint state method, where for n=1, 2, ..., N-1, the general formulas of the derivatives of the three function parts are as follows:

[0020] g for the nth flip angle α n The derivative of is:

[0021] Where: n = N-1 When n<N-1

[0022] h for the nth flip angle α n The derivative of is:

[0023] Where: n = N-1, μ n =f n+1 H C, n<N-1 μ n =μ n+1 P n+1 +f n+1 H C;

[0024] m for the nth flip angle α n The derivative of is:

[0025] Where: n = N-1, ν n =f n+1 H Dn+1 , when n<N-1 n =ν n+1 P n+1 +f n+1 H D n+1 ;

[0026] Among the derivatives of g, h, and m, P n represents the extended phase diagram state f n Transfer to f n+1 The state transition operator is calculated as follows: n =R(α n ,φ n )E(τ,T1,T2)S

[0027] Where S represents the effect of the dispersion relative to the state transfer; R(α n ,φ n ) represents the effect of the refocusing pulse on the transfer, and the calculation formula is:

[0028] Where: i is the imaginary unit; φ n is the phase of the nth refocusing pulse;

[0029] E represents the effect of relaxation on state transition, and the calculation formula is:

[0030] Where: T1 is the longitudinal relaxation time, T2 is the transverse relaxation time, and τ is the echo interval;

[0031] S23, according to the algebraic relationship of the three function parts in the objective function, the derivatives of the three function parts with respect to the flip angle are recombined to form the overall derivative of the objective function with respect to N-1 flip angles, and the flip angle set FAs = [α1, α2, ..., α N-1 ] for optimization.

[0032] As a preferred embodiment of the first aspect, in S23, when optimizing the flip angle set to be optimized based on the global derivative, a gradient descent algorithm or a conjugate gradient descent algorithm is adopted.

[0033] As a preferred embodiment of the first aspect, the signal attenuation coefficient w of the i-th refocusing pulse is i The determination is made by preliminary experiments. The determination method is as follows: the water model is measured in advance by experiments when the flip angle is FAs=[α1,α2,...,α i ,...,α N-1 The k-space signals [K1, K2, ..., K i ,...,K N-1], and then measure the longitudinal relaxation time T1 and transverse duration T2 of the water model; based on the longitudinal relaxation time T1, transverse relaxation time T2, flip angle FAs = [α1, α2, ..., α i ,...,α N-1 ] and echo interval τ, the signal amplitude without attenuation [s1,s2,...,s i ,...,s N-1 ], with K i With s i The ratio of is taken as the signal attenuation coefficient w of the i-th refocusing pulse i .

[0034] As a preference in the first aspect, the ratio of the weight coefficients a / b is 100 to 2000.

[0035] In a second aspect, the present invention provides a magnetic resonance CEST imaging method, wherein the imaging sequence used in the method includes a CEST saturation module, a fat suppression module and a variable flip angle readout module; wherein the echo chain length of the variable flip angle sequence in the variable flip angle readout module is N-1, and the flip angles of the N-1 refocusing pulses are obtained according to the magnetic resonance variable flip angle sequence flip angle optimization method according to any one of claims 1 to 5.

[0036] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for optimizing the flip angle of a magnetic resonance variable flip angle sequence as described in any one of the schemes in the first aspect is implemented.

[0037] In a third aspect, the present invention provides a computer electronic device comprising a memory and a processor;

[0038] The memory is used to store computer programs;

[0039] The processor is configured to implement the magnetic resonance variable flip angle sequence flip angle optimization method as described in any one of the solutions in the first aspect when executing the computer program.

[0040] In a fourth aspect, the present invention provides a magnetic resonance imaging device with a variable flip angle, comprising a magnetic resonance scanner and a control unit, wherein the control unit stores a computer program, and when the computer program is executed, it is used to implement the magnetic resonance variable flip angle sequence flip angle optimization method as described in any one of the schemes in the first aspect above; the magnetic resonance scanner is used to perform magnetic resonance CEST imaging and obtain CEST images according to the optimized flip angle sequence.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] This paper models the SNR enhancement problem in magnetic resonance imaging (MRI) procedures as a flip-angle optimization problem for a variable flip-angle sequence. This optimization problem has analytical derivatives with respect to the flip angle, enabling efficient solution. Furthermore, the optimization problem incorporates a resolution penalty term, allowing the variable flip-angle CEST method to optimize SNR while also taking resolution into account. Compared to traditional filtering methods, this method avoids noise amplification caused by filtering and eliminates the need for manually presetting window functions. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 shows the timing diagram of an SNR-boosted variable flip angle CEST imaging sequence, where the echo train length ETL is N-1, and the N-1 variable flip angle readouts are the parts that need to be optimized.

[0044] Figure 2 is a schematic diagram of the EPG status;

[0045] Figure 3 shows the flip angle, signal curve, and point spread function obtained from optimization under different settings. (A-C) consider only SNR, (D-F) only consider resolution, and (G-I) consider both SNR and resolution.

[0046] Figure 4 shows a local magnification of a layer of an image acquired with an optimized flip angle considering only resolution, a local magnification of a layer of an image acquired with a 120° CFA, and the quantitative blur of different layers of the image. DETAILED DESCRIPTION

[0047] The present invention will be further described and illustrated below with reference to the accompanying drawings and specific embodiments. The technical features of each embodiment of the present invention may be combined accordingly, provided that there is no conflict between them.

[0048] In order to make the above-mentioned objects, features and advantages of the present invention more clearly understood, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways than those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below. The technical features in the various embodiments of the present invention can be combined accordingly without conflicting with each other.

[0049] The inventive concept of the present invention is to model the signal-to-noise ratio (SNR) enhancement problem as a flip angle optimization problem, and construct an overall objective function consisting of a signal-to-noise ratio maximization target term and a resolution penalty term, wherein the resolution penalty term takes the form of a function of two terms divided by each other. The overall objective function can use a method of solving the derivative of the flip angle to solve the optimal solution, thereby obtaining the optimal flip angle solution that can maximize the SNR.

[0050] For ease of understanding, the principle of the present invention will be described in detail before describing the specific technical solutions of the present invention.

[0051] 1 SNR enhancement modeled as an optimization problem

[0052] The present invention provides a SNR-enhanced variable flip angle CEST imaging method, termed SNR-boosted variable flip angle CEST. This method models the SNR as a function of flip angle and optimizes this function to maximize the SNR while simultaneously constraining the signal to reduce image blur caused by signal modulation. The timing diagram of the SNR-boosted variable flip angle CEST sequence is shown in Figure 1. It includes three components: CEST saturation, fat suppression, and signal readout. The present invention optimizes the signal readout component.

[0053] For a uniform object, its i-th echo in k-space can be expressed as K i =s i *w i (1)

[0054] where s i is the signal amplitude related to the flip angle and relaxation, w i The signal attenuation is caused by factors such as gradient, magnetic field inhomogeneity, signal scaling, etc. In TSE imaging, when a series of refocusing pulses with flip angles of all 180° are used, s i =exp(-i*TE / T2); when other flip angles are used, s i is a function of T1, T2, and flip angle, and its value can be obtained through Bloch simulation or extended phase graph (EPG) simulation. When the k-space center is acquired and the magnetic field is uniform, without considering global signal scaling, w i The value of is 1; usually, due to phase separation, w i The value of is less than 1. The attenuation coefficient w i It can be calculated by the following formula

[0055] where K i is the experimentally measured k-space signal, s i is the signal amplitude obtained through EPG simulation.

[0056] According to Parseval's theorem, the energy of image space and k space is equal, that is,

[0057] Among them I iis the signal value in the image space. Assuming that the variance of the noise is a constant, to improve the signal-to-noise ratio in the image space, the signal value in the image space needs to be increased, that is, the signal value in the k-space is increased. Therefore, the signal-to-noise ratio maximization problem can be expressed as

[0058] Where N-1 is the number of echoes in each echo train, that is, the echo train length ETL. i Greater than w at the edge of k-space i , the optimization result of the above problem will make the k-space center signal too large and the edge signal too low, resulting in image blur. Therefore, it is necessary to introduce a resolution penalty term

[0059] The coefficients a and b control the weighting of SNR and resolution, with the ratio of a / b preferably ranging from 100 to 2000. This resolution penalty ensures that signals are close to each other while also maximizing the signal value. Optimizing only the resolution yields a constant signal, equivalent to an ideal point spread function (PSF).

[0060] 2. Solution of optimization problem

[0061] The following focuses on using EPG to solve the above optimization problem.

[0062] 2.1 Introduction to the EPG Framework

[0063] For a variable flip angle sequence with ETL N-1, let the flip angle of the excitation pulse be α0 = 90°, and the flip angles of the N-1 refocusing pulses be [α1, ..., α N-1 ], the EPG states corresponding to the excitation pulse and N-1 refocusing pulses are [f1,f2,...,f N-1 ,f N ], each EPG state f i is a column vector, the second component of which represents the signal value s i As shown in FIG2, f1 is the state just after applying the excitation pulse α0 = 90°, f2 is the state just after applying the first refocusing pulse α1, and f n is the state immediately after the n-1th refocusing pulse is applied, and so on. EPG state f n to f n+1 The transition is determined by the state transition operator P n To describe

[0064] where R represents the effect of the refocusing pulse on the transfer, φ nis the phase of the nth refocusing pulse; E represents the effect of relaxation on state transfer; S represents the effect of dispersion on state transfer.

[0065] The effect R of the refocusing pulse on the state transition is a function related to the flip angle α and the phase φ of the refocusing pulse, and the expression is:

[0066] 2.2 Optimization Problem Decomposition and Solving Subproblem Derivatives

[0067] st0≤FA≤180°

[0068] Split the optimization problem (5) into three parts: For these three parts, we calculate their derivatives with respect to the refocusing pulse flip angle. First, we consider Change it to EPG status

[0069] Where C(2,2)=1, and all other elements of the C matrix are 0. The C matrix is ​​used to extract the signal from the EPG state, such as s i 2 =f i H Cf i .

[0070] Next, we consider using the adjoint state method (ASM) to find the last flip angle α of g N-1 The derivative of :

[0071] in

[0072] Consider the penultimate flip angle α of g N-2 The derivative of

[0073] Split the right side of formula (11) into three parts, defined as E=0.5*{(f N-1 -f2) H C(f N-1 -f2)+(f N-1 -f3) H C(f N-1 -f3)+....+(f N-1 -f N ) H C(f N-1 -f N )} (13)

[0074] F=0.5*{(f N -f2) H C(f N -f2)+(f N -f3) H C(f N -f3)+...+(f N -f N ) H C(f N -f N )} (14)

[0075] These three parts are related to α N-2 The derivatives of

[0076] Combining equations (15), (16), and (17), we can obtain

[0077] in

[0078] Similarly, for any nth (n<N-1) refocusing pulse, the flip angle α n , we can get g for α n The derivative formula is:

[0079] in

[0080] So far, the analytical derivatives of g for all flip angles to be optimized have been obtained.

[0081] Similarly, we can deduce and The derivatives of N-1 flip angles are as follows:

[0082] for Change it to EPG status:

[0083] First consider the derivative of h with respect to the last flip angle, and we can get it by the adjoint state method:

[0084] Where: μ N-1 =f N H C (22)

[0085] Similarly, for any nth (n<N-1) refocusing pulse, the flip angle α n , we can get h for α n The derivative formula is:

[0086] in:

[0087] μ n =μ n+1 P n+1 +f n+1 H C (24)

[0088] for Change it to EPG status

[0089] Among them, D i (2,2)=w i 2 , D i The other elements of the matrix are all 0. First consider the derivative of m with respect to the last flip angle:

[0090] Where: ν N-1 =f N H D N (27)

[0091] Similarly, for any nth (n<N-1) refocusing pulse, the flip angle α n , we can get m for α n The derivative formula is:

[0092] Where: ν n =ν n+1 P n+1 +f n+1 H D n+1 (29)

[0093] Therefore, problem (5) has an analytical derivative form and can be solved using algorithms such as gradient descent and conjugate gradient descent.

[0094] Based on the above theoretical discussion, the present invention will now illustrate the specific implementation and technical effects of the magnetic resonance variable flip angle sequence flip angle optimization method through specific embodiments.

[0095] Example

[0096] In this embodiment, the specific steps of the magnetic resonance variable flip angle sequence flip angle optimization method are as follows:

[0097] S1. For a magnetic resonance variable flip angle sequence with an echo train length of N-1, the flip angle optimization problem of its N-1 refocusing pulses is modeled as the following objective function:

[0098] st0≤FA≤180°

[0099] Where: s i is the signal amplitude without attenuation after the i-th refocusing pulse is applied, which is caused only by the flip angle and relaxation; w i is the attenuation coefficient of the signal of the i-th refocusing pulse due to dephasing, phase encoding and other factors; FAs = [α1, α2, ..., α N-1 ] is the set of flip angles to be optimized, FA represents the flip angle α of any i-th refocusing pulse i , and its optimization range constraint is [0°, 180°]; a and b are two weight coefficients in the objective function, and both a and b are greater than 0.

[0100] In this embodiment, the signal attenuation coefficient w of the i-th refocusing pulse is i Determined through preliminary experiments, the determination method is:

[0101] Through experiments in advance, the water model is measured when the flip angle is FAs=[α1,α2,...,α i ,...,α N-1 The k-space signals [K1, K2, ..., K i ,...,K N-1 ], and then measure the longitudinal relaxation time T1 and transverse duration T2 of the water model; based on the longitudinal relaxation time T1, transverse relaxation time T2, flip angle FAs = [α1, α2, ..., α i ,...,α N-1 ] and echo interval τ, the signal amplitude without attenuation [s1,s2,...,s i ,...,s N-1 ], with K i With s i The ratio of is taken as the signal attenuation coefficient w of the i-th refocusing pulse i .

[0102] The flip angle α applied by each of the N-1 refocusing pulses during the experiment can be selected from a set of commonly used values ​​based on relevant literature.

[0103] S2. The objective function is represented by an extended phase diagram and then iteratively solved to obtain the optimal solution of the optimized flip angle set, which is used as the flip angle setting value of N-1 refocusing pulses in the magnetic resonance variable flip angle sequence.

[0104] In the above process of solving the objective function, each round of iteration process is as follows:

[0105] S21, the objective function The three parts are split and represented by extended phase diagrams, where:

[0106] Expressed as g using the expanded phase diagram:

[0107] Expressed as h using the expanded phase diagram:

[0108] Expressed as m using the expanded phase diagram:

[0109] Where: C is a slave EPG state f i The matrix from which the signal is extracted is a matrix where only the second row and the second column are 1 and the rest of the elements are 0; D i For a slave EPG state f i Extract the signal matrix, only the second row and second column in the matrix is ​​w i 2 The rest of the elements are 0;

[0110] S22. Derivatives of the function parts represented by each extended phase diagram are obtained for N-1 flip angles using the adjoint state method, where for n=1, 2, ..., N-1, the general formulas of the derivatives of the three function parts are as follows:

[0111] g for the nth flip angle α n The derivative of is:

[0112] Where: n = N-1 When n<N-1

[0113] h for the nth flip angle α n The derivative of is:

[0114] Where: n = N-1, μ n =f n+1 H C, n<N-1 μ n =μ n+1 P n+1 +f n+1 H C;

[0115] m for the nth flip angle α n The derivative of is:

[0116] Where: n = N-1, ν n =f n+1 H Dn+1 , when n<N-1 n =ν n+1 P n+1 +f n+1 H D n+1 ;

[0117] In the above derivatives of g, h, and m, P n represents the extended phase diagram state f n Transfer to f n+1 The state transition operator is calculated as follows: n =R(α n ,φ n )E(τ,T1,T2)S

[0118] Where S represents the effect of the dispersion relative to the state transfer; R(α n ,φ n ) represents the effect of the refocusing pulse on the transfer, and the calculation formula is:

[0119] Where: i is the imaginary unit; φ n is the phase of the nth refocusing pulse;

[0120] E represents the effect of relaxation on state transition, and the calculation formula is:

[0121] Where: T1 is the longitudinal relaxation time, T2 is the transverse relaxation time, and τ is the echo interval;

[0122] S23, according to the algebraic relationship of the three function parts in the objective function, the derivatives of the three function parts with respect to the flip angle are recombined to form the overall derivative of the objective function with respect to N-1 flip angles, and the flip angle set FAs = [α1, α2, ..., α N-1 ] for optimization.

[0123] In this embodiment, when optimizing the flip angle set to be optimized based on the global derivative, a gradient descent algorithm or a conjugate gradient descent algorithm may be used.

[0124] The optimal solution of the flip angle set obtained based on the above optimization can be used as the flip angle setting value of the N-1 refocusing pulses of the variable flip angle sequence in the imaging sequence. In this way, an imaging sequence for magnetic resonance CEST imaging can be formed for performing magnetic resonance imaging. As shown in Figure 1, the imaging sequence includes a CEST saturation module, a fat suppression module and a variable flip angle readout module, wherein the echo chain length of the variable flip angle sequence in the variable flip angle readout module is N-1, and the flip angles of the N-1 refocusing pulses are obtained according to the above magnetic resonance variable flip angle sequence flip angle optimization method. The CEST saturation module and fat suppression module in the sequence belong to the prior art. In this embodiment, please refer to Figure 1 for details, and no further details will be given.

[0125] To demonstrate the effects of the two optimization items in the objective function of the present invention, this embodiment optimizes the flip angle for three cases: (1) optimizing only for SNR, i.e., setting coefficient a to 1 and coefficient b to 0; (2) optimizing only for resolution, i.e., setting coefficient a to 0 and coefficient b to 1; and (3) optimizing both SNR and resolution, i.e., setting coefficient a to 1 and coefficient b to 0.001.

[0126] During the optimization process of this embodiment, the physiological parameters are set as: T1 = 1000 ms, T2 = 100 ms; the sequence parameters are: echo spacing (tesp) = 4.06 ms, echo train length ETL = 140; the initial value and upper and lower limits of the flip angle to be optimized are: the initial value of all flip angles in the entire echo train is 10°, the upper limit of all flip angles is 180°, and the lower limit of all flip angles is 0°.

[0127] The results of optimizing only for SNR are shown in Figures 3A, 3B, and 3C. The optimized signal curve is similar to that of the CFA 120°, but the flip angle used is much smaller. The results of optimizing only for resolution are shown in Figures 3D, 3E, and 3F. The optimized signal curve remains almost constant throughout the entire echo chain. Correspondingly, the resulting point spread function almost coincides with the ideal point spread function. The results of optimizing for both SNR and resolution are shown in Figures 3G, 3H, and 3I. The optimized signal increases by 4.7% compared to the CFA 120° signal, and the full width half maximum (FWHM) of the resulting point spread function is 6.1% smaller than that of the CFA 120°.

[0128] FIG4B shows an image acquired with only the optimized flip angle, which has no obvious blur in the phase encoding direction compared to the image acquired with CFA 120° in FIG4A . The quantitative blur in FIG4C further shows that CFA 120° has a higher blur and a more blurred image.

[0129] It should be noted that the magnetic resonance variable flip angle sequence flip angle optimization method provided in the above embodiment can essentially exist in the form of a program processing flow. Therefore, similarly, based on the same inventive concept, another preferred embodiment of the present invention also provides a computer electronic device corresponding to the magnetic resonance variable flip angle sequence flip angle optimization method provided in the above embodiment, which includes a memory and a processor;

[0130] The memory is used to store computer programs;

[0131] The processor is configured to implement the aforementioned method for optimizing the flip angle of a magnetic resonance variable flip angle sequence when executing the computer program.

[0132] Similarly, based on the same inventive concept, another preferred embodiment of the present invention also provides a computer-readable storage medium corresponding to the magnetic resonance variable flip angle sequence flip angle optimization method provided in the above embodiment, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the magnetic resonance variable flip angle sequence flip angle optimization method as described above is implemented.

[0133] It is understood that the above-mentioned memory and storage medium may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage. At the same time, the storage medium may also be various media that can store program code, such as a USB flash drive, a mobile hard drive, a magnetic disk, or an optical disk. Of course, with the widespread use of cloud servers, the above-mentioned software programs can also be installed on the cloud platform to provide corresponding services, so computer-readable storage media are not limited to the form of local hardware.

[0134] It is understandable that the above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components.

[0135] It should also be noted that those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here. In the various embodiments provided in this application, the division of steps or modules in the device and method is only a logical function division. In actual implementation, there may be other division methods, for example, multiple modules or steps can be combined or integrated together, and a module or step can also be split.

[0136] Furthermore, the logic instructions in the aforementioned memory can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention.

[0137] Similarly, based on the same inventive concept, another preferred embodiment of the present invention further provides a magnetic resonance imaging device with a variable flip angle corresponding to the magnetic resonance variable flip angle sequence flip angle optimization method provided in the above embodiment, which includes a magnetic resonance scanner and a control unit. In particular:

[0138] The control unit stores a computer program, which, when executed, is used to implement the method for optimizing the flip angle of a magnetic resonance variable flip angle sequence according to any one of the solutions of the first aspect;

[0139] The magnetic resonance scanner is used to perform magnetic resonance CEST imaging and acquire CEST images according to the optimized flip angle sequence.

[0140] It should be noted that the MRI scanner can be any MRI scanner capable of performing CEST imaging. Its structure is conventional, and a mature commercial product can be used, with no specific limitation on the model. Furthermore, in addition to storing the aforementioned computer program, the control unit of the MRI device should also contain the imaging sequences and other software programs necessary to perform CEST imaging.

[0141] Of course, the above-mentioned control unit can be an independent control unit or a control unit provided in the magnetic resonance scanner. That is, the above-mentioned magnetic resonance variable flip angle sequence flip angle optimization method can be integrated into the control unit of the magnetic resonance imaging device in the form of a data processing program, so that the magnetic resonance scanner can directly optimize the variable flip angle sequence in the imaging sequence online without the need for an additional control unit and perform magnetic resonance CEST imaging based on the optimized flip angle sequence.

[0142] The embodiment described above is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Persons skilled in the art may make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, any technical solution obtained by equivalent substitution or equivalent transformation falls within the scope of protection of the present invention.

Claims

1. A method for optimizing the flip angle of a magnetic resonance variable flip angle sequence, characterized in that: include: S1. For a magnetic resonance variable flip angle sequence with an echo train length of N-1, the flip angle optimization problem of its N-1 refocusing pulses is modeled as the following objective function: st0≤FA≤180° Where: s i is the signal amplitude without attenuation after the i-th refocusing pulse is applied; w i is the signal attenuation coefficient of the i-th refocusing pulse; FAs = [α1, α2, ..., α N-1 ] is the set of flip angles to be optimized, FA represents the flip angle α of any i-th refocusing pulse i , the optimization range constraint is [0°, 180°]; a and b are two weight coefficients in the objective function, and both a and b are greater than 0; S2. The objective function is represented by an extended phase graph (EPG) and then iteratively solved to obtain the optimal solution of the optimized flip angle set as the flip angle setting value of N-1 refocusing pulses in the magnetic resonance variable flip angle sequence.

2. The method for optimizing the flip angle of a magnetic resonance variable flip angle sequence according to claim 1, characterized in that: The process of each iteration in solving the objective function is as follows: S21, the objective function The three parts are split and represented by extended phase diagrams, where: Expressed as g using the expanded phase diagram: Expressed as h using the expanded phase diagram: Expressed as m using an extended phase diagram: Where: C is a state from EPG f i The matrix of the signal is extracted from the matrix, in which only the second row and the second column are 1 and the rest are 0; D i For a state from EPG i The matrix of the signal is extracted from the matrix, in which only the second row and the second column are w i 2 The rest of the elements are 0; S22. Derivatives are obtained for each function part represented by the extended phase diagram using the adjoint state method for N-1 flip angles, wherein for n=1, 2, ..., N-1, the general formulas of the derivatives of the three function parts are as follows: g for the nth flip angle α n The derivative of is: Where: n = N-1 When n<N-1 h for the nth flip angle α n The derivative of is: Where: n = N - 1, μ n = f n+1 H C, when n < N - 1, μ n = μ n+1 P n+1 + f n+1 H C; m for the nth flip angle α n The derivative of is: Where: n=N-1, n n =f n+1 H D n+1 ,n<N-1 timen n =n n+1 P n+1 +f n+1 H D n+1 ; In the derivatives of g, h, and m, P n represents the extended phase diagram state f n Transfer to f n+1 The state transition operator is calculated as follows: P n =R(α n ,f n )E(τ,T1,T2)S Where S represents the effect of dispersion on state transfer; R(α n ,φ n ) represents the effect of the refocusing pulse on the transfer, The calculation formula is: Where: i is the imaginary unit; φ n is the phase of the nth refocusing pulse; E represents the effect of relaxation on state transition, and the calculation formula is: Where: T1 is the longitudinal relaxation time, T2 is the transverse relaxation time, and τ is the echo interval; S23, according to the algebraic relationship of the three function parts in the objective function, the derivatives of the three function parts with respect to the flip angle are recombined to form the overall derivative of the objective function with respect to N-1 flip angles, and the flip angle set FAs=[α1, α2, ..., α N-1 ] for optimization.

3. The method for optimizing the flip angle of a magnetic resonance variable flip angle sequence according to claim 1, characterized in that: In S23, when optimizing the flip angle set to be optimized based on the global derivative, a gradient descent algorithm or a conjugate gradient descent algorithm is used.

4. The method for optimizing the flip angle of a magnetic resonance variable flip angle sequence according to claim 1, characterized in that: The signal attenuation coefficient w of the i-th refocusing pulse i Determined through preliminary experiments, the determination method is: Through experiments in advance, the water model was measured when the flip angle was FAs=[α1,α2,...,α i ,...,α N-1 The k-space signals [K1, K2, ..., K i ,...,K N-1 ], and then measure the longitudinal relaxation time T1 and transverse duration T2 of the water model; based on the longitudinal relaxation time T1, transverse relaxation time T2, flip angle FAs = [α1, α2, ..., α i ,...,α N-1 ] and echo interval τ, the signal amplitude without attenuation [s1,s2,...,s i ,...,s N-1 ], with K i With s i The ratio of is taken as the signal attenuation coefficient w of the i-th refocusing pulse i .

5. The method for optimizing the flip angle of a magnetic resonance variable flip angle sequence according to claim 1, characterized in that: The ratio of the weight coefficient a / b is 100 to 2000.

6. A magnetic resonance CEST imaging method, characterized in that: The imaging sequence used includes a CEST saturation module, a fat suppression module and a variable flip angle readout module; wherein the echo chain length of the variable flip angle sequence in the variable flip angle readout module is N-1, and the flip angles of N-1 refocusing pulses are obtained according to the magnetic resonance variable flip angle sequence flip angle optimization method according to any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the processor, the method for optimizing the flip angle of a magnetic resonance variable flip angle sequence according to any one of claims 1 to 5 is implemented.

8. A computer electronic device, characterized in that: including memory and processor; The memory is used to store computer programs; The processor is used to implement the magnetic resonance variable flip angle sequence flip angle optimization method according to any one of claims 1 to 5 when executing the computer program.

9. A magnetic resonance imaging device with a variable flip angle, characterized in that: The invention comprises a magnetic resonance scanner and a control unit, wherein a computer program is stored in the control unit, and when the computer program is executed, it is used to implement the magnetic resonance variable flip angle sequence flip angle optimization method as described in any one of claims 1 to 5; the magnetic resonance scanner is used to perform magnetic resonance CEST imaging and obtain CEST images according to the optimized flip angle sequence.

Citation Information

Patent Citations

  • Magnetic resonance variable flip angle sequence flip angle optimization method, CEST imaging method, medium and equipment

    CN116577711A

  • System and Method for Improved Spin-Echo-Based Magnetic Resonance Imaging

    US20200041598A1

  • Method and apparatus for processing magnetic resonance data

    US20200072931A1

  • Magnetic Resonance Imaging Method and System with Optimal Variable Flip Angles

    US20200256939A1

  • Ratiometric pulsed CEST imaging

    US20200379070A1