A method for magnetic resonance metabolic imaging acquisition and reconstruction with high spatio-temporal resolution

Abstract

The present invention discloses a method for magnetic resonance metabolic imaging acquisition and reconstruction with high spatio-temporal resolution. A pulse sequence based on prior knowledge is constructed, and the pulse sequence based on prior knowledge is used for image acquisition, reconstruction, and compound concentration correction; the pulse sequence based on prior knowledge includes a plurality of balanced steady-state free precession (bSSFP) sequences. A pulse phase difference is added between the radiofrequency pulses corresponding to each step of phase encoding in each balanced steady-state free precession bSSFP sequence, and the numerical values of the pulse phase differences corresponding to different balanced steady-state free precession bSSFP sequences should be different. The present invention adopts a magnetic resonance imaging acquisition mode, and more efficient frequency encoding is used for spatial encoding in the readout direction Gx, improving the acquisition efficiency, having a higher image signal-to-noise ratio, and being less susceptible to magnetic field inhomogeneity; the present invention converts the signal intensity value into the concentration of metabolic compounds, realizing the absolute quantification of the concentration of metabolic compounds.

Description

Technical Field

[0001] The present invention relates to the technical field of magnetic resonance metabolic imaging, and particularly relates to a method for collecting and reconstructing magnetic resonance metabolic imaging with high spatio-temporal resolution. Background Art

[0002] Magnetic resonance imaging (MRI) signals reflect water and its properties and do not provide metabolic information. Although magnetic resonance metabolic imaging can directly detect the characteristic spectral peaks of metabolites, magnetic resonance metabolic imaging still faces many challenges technically. One is that the concentration of metabolic molecules is low. The concentration of water in the body is about 50 moles, while the concentration of metabolic molecules is usually less than 10 mmol / L. Secondly, in conventional chemical shift imaging, since chemical shift information needs to be obtained, frequency encoding cannot be used and only phase encoding can be used for spatial encoding, resulting in low encoding efficiency during the chemical shift imaging process.

[0003] Balanced steady-state free precession (bSSFP) technology is a fast and efficient magnetic resonance imaging technology, characterized in that within each repetition time ( TR ), the gradients in all directions are completely balanced, which enables the magnetization vector to form a steady state during each repetition time TR . In this steady state, both the longitudinal and transverse magnetization vectors contribute to the formation of the signal, resulting in a high signal-to-noise ratio. For a compound with a longitudinal relaxation time of and a transverse relaxation time of , the formula for the steady-state transverse magnetization vector of balanced steady-state free precession bSSFP can be expressed as Formula 1: (Formula 1)

[0004] Wherein, represents the steady-state transverse magnetization vector, is the spin density, is the flip angle of the radio frequency pulse, the intermediate quantity , and the intermediate quantity . It can be seen that the steady-state transverse magnetization vector of balanced steady-state free precession bSSFP depends not only on sequence parameters such as the repetition time TR and the flip angle of the radio frequency pulse, but also on the properties of the metabolic compound itself. Especially for substances with a low ratio, the signal-to-noise ratio can be significantly improved, making it an attractive option for metabolic imaging.

[0005] Balanced steady-state free precession bSSFP can significantly improve the signal-to-noise ratio, especially for low Deuterated metabolites and hyperpolarization 13 C metabolites. However, the chemical shift equilibrium steady-state free precession imaging CSI-bSSFP technology is affected by the low spectral resolution, resulting in compound image aliasing, and the multi-echo equilibrium steady-state free precession ME-bSSFP technology is easily affected by magnetic field inhomogeneity to form dark bands. Therefore, there is no relevant literature report on the high-temporal and spatial resolution magnetic resonance metabolic imaging acquisition and reconstruction technology that is not easily affected by compound image aliasing and magnetic field inhomogeneity at the same time. This is a technical blank and a key problem that needs to be solved in the field of magnetic resonance metabolic imaging technology. Summary of the invention

[0006] In view of the above problems existing in the prior art, the present invention provides a method for magnetic resonance metabolic imaging acquisition and reconstruction with high temporal and spatial resolution. The magnetic resonance imaging acquisition method is used during data acquisition to greatly improve the data acquisition efficiency, and finally metabolite image separation is performed using prior knowledge based on chemical shift.

[0007] The above-mentioned object of the present invention is achieved by the following technical solutions:

[0008] A method for acquiring and reconstructing magnetic resonance metabolic imaging with high temporal and spatial resolution comprises the following steps:

[0009] Step 1, constructing a pulse sequence based on prior knowledge, wherein the pulse sequence based on prior knowledge includes a plurality of balanced steady-state free precession bSSFP sequences in sequence, and in each balanced steady-state free precession bSSFP sequence, a pulse phase difference is added between the radio frequency pulses corresponding to each step of phase encoding;

[0010] According to the relationship between the initial signal intensity and the pulse phase difference corresponding to the steady-state transverse magnetization vector and the preset value of the initial signal intensity simulation, the signal simulation curve of the pulse phase difference-steady-state transverse magnetization vector corresponding to each metabolic compound in the metabolites to be separated is simulated, and the value of each pulse phase difference in the pulse sequence based on prior knowledge is determined according to the signal simulation curve of the pulse phase difference-steady-state transverse magnetization vector of each metabolic compound;

[0011] The total number of equilibrium steady-state free precession bSSFP sequences is determined according to the number of metabolic compounds in the metabolites to be separated;

[0012] The metabolites to be separated include metabolites of known concentrations and metabolites of unknown concentrations;

[0013] Step 2, based on the pulse sequence based on prior knowledge constructed in step 1, corresponding to each equilibrium steady-state free precession bSSFP sequence, collect magnetic resonance images corresponding to the metabolites to be separated;

[0014] Step 3: Reconstruct the magnetic resonance images collected in Step 2 according to the relationship between the steady-state transverse magnetization vector of each metabolite and the initial signal intensity and the pulse phase difference, to obtain the initial signal separation images corresponding to each metabolite;

[0015] Step 4: Based on the mean value of the initial signal intensity of the metabolite with known concentration and the mean value of the concentration of the metabolite with known concentration, obtain the concentration conversion factor; use the concentration conversion factor to convert the initial signal separation images corresponding to each metabolite into the corresponding compound concentration images.

[0016] Between adjacent balanced steady-state free precession bSSFP sequences as described above, a first spoiling gradient G1 is increased in the slice direction Gz, a second spoiling gradient G2 is increased in the phase encoding direction Gy, and a third spoiling gradient G3 is increased in the readout direction Gx; frequency encoding is used for spatial encoding in the readout direction Gx.

[0017] As described in Step 1 above, according to the relationship between the steady-state transverse magnetization vector and the initial signal intensity and the pulse phase difference, as well as the preset value of the initial signal intensity simulation, simulate the signal simulation curve of the pulse phase difference - steady-state transverse magnetization vector corresponding to each metabolite in the metabolite to be separated, specifically including the following steps:

[0018] Simulate the signal simulation curve of the pulse phase difference - steady-state transverse magnetization vector based on the following formula: ,

[0019] Wherein, represents the simulated value of the steady-state transverse magnetization vector of the th metabolite; represents the preset value of the initial signal intensity simulation of the th metabolite, represents the preset range value of the pulse phase difference; is the serial number of the metabolite; is the imaginary unit, the intermediate quantity , the intermediate quantity , the intermediate quantity , the intermediate quantity , the intermediate quantity , the intermediate quantity ; is the longitudinal relaxation time; is the transverse relaxation time; is the flip angle of the radiofrequency pulse; is the frequency offset caused by magnetic field inhomogeneity; is the frequency offset caused by the chemical shift of the th metabolite, denoted as the frequency offset caused by chemical shift ; is the repetition time, is the echo time;

[0020] The steady-state transverse magnetization vector is decomposed into a signal intensity and a corresponding phase value, so that the signal simulation curve of the pulse phase difference - steady-state transverse magnetization vector is decomposed into the signal intensity simulation curve of the pulse phase difference - steady-state transverse magnetization vector and the phase value simulation curve of the pulse phase difference - steady-state transverse magnetization vector;

[0021] The signal intensity simulation curves of the pulse phase difference - steady-state transverse magnetization vector of each metabolite are placed in the same coordinate system; the phase value simulation curves of the pulse phase difference - steady-state transverse magnetization vector of each metabolite are placed in another coordinate system.

[0022] Determine the value of the pulse phase difference according to the following rules:

[0023] For each metabolite, the phase values of the simulated values of the steady-state transverse magnetization vectors corresponding to the determined values of the pulse phase differences in the phase value simulation curve of the pulse phase difference - steady-state transverse magnetization vector are different;

[0024] In the same balanced steady-state free precession bSSFP sequence, the values of the pulse phase differences corresponding to all phase encodings are the same;

[0025] The determined value of the pulse phase difference is not in the bSSFP signal dark band region of the signal intensity simulation curve of the pulse phase difference - steady-state transverse magnetization vector of each metabolite.

[0026] The total number of the balanced steady-state free precession bSSFP sequences as described above is greater than or equal to the number of metabolite types.

[0027] Step 2 as described above includes the following steps:

[0028] Step 2.1, perform regional shimming on the region of interest;

[0029] Step 2.2, acquire a non-localized spectrum, and determine the frequency shift caused by the actual chemical shift of a metabolite in the metabolites to be separated through the non-localized spectrum. If the frequency shift caused by the actual chemical shift of the metabolite is different from the frequency shift caused by the chemical shift during the simulation in Step 1, then adjust the center frequency of the pulse sequence based on prior knowledge until the frequency shift caused by the actual chemical shift of the metabolite is the same as the frequency shift caused by the chemical shift during the simulation in Step 1 is the same;

[0030] Step 2.3: Based on the pulse sequence based on prior knowledge obtained in Step 2.2, magnetic resonance images of the metabolites to be separated are collected, and one magnetic resonance image is collected for each balanced steady-state free precession bSSFP sequence.

[0031] As described above, Step 3 includes the following steps:

[0032] The initial signal intensity corresponding to each metabolite compound is obtained by inverse calculation using the reconstruction algorithm based on echo asymmetry iterative decomposition and least squares estimation according to the following formula: ,

[0033] where, represents the total steady-state transverse magnetization vector signal of all metabolite compounds in the magnetic resonance image with the serial number equal to ; represents the serial number of the balanced steady-state free precession bSSFP sequence, ; N is the total number of balanced steady-state free precession bSSFP sequences; is the initial signal intensity of the th metabolite compound, denoted as the initial signal intensity ; M represents the number of types of metabolite compounds in the metabolites to be separated; is the serial number of the metabolite compound, ; is the coefficient matrix, and the elements in the coefficient matrix satisfy the following conditions: ,

[0034] represents the pulse phase difference of the th balanced steady-state free precession bSSFP sequence; represents the element in the n th row and m th column of the coefficient matrix;

[0035] The initial signal intensity corresponding to the metabolite compound is presented in the form of an image to obtain the initial signal separation image of each metabolite compound.

[0036] As described above, Step 4 includes the following steps:

[0037] Step 4.1: Calculate the concentration conversion factor according to the following formula: ,

[0038] is the average initial signal intensity of the metabolite compound with known concentration, is the average concentration of the metabolite compound with known concentration;

[0039] Step 4.2. Calculate the concentration of each metabolite compound by the following formula: ,

[0040] where is the concentration of the metabolite compound with the serial number , is the number of molecules contributing to the signal of the metabolite compound with known concentration, is the number of molecules contributing to the signal of the metabolite compound;

[0041] Present the concentration of the metabolite compound in the form of an image to obtain the corresponding compound concentration image.

[0042] A computer device includes a memory and a processor. The memory stores a computer program. When the processor executes the computer program, it implements step 1 and steps 3 - 4 of the magnetic resonance metabolic imaging acquisition and reconstruction method described above.

[0043] A computer - readable storage medium stores a computer program. When the computer program is executed by a processor, it implements step 1 and steps 3 - 4 of the magnetic resonance metabolic imaging acquisition and reconstruction method described above.

[0044] The advantages and positive effects of the present invention are:

[0045] 1. The present invention utilizes the frequency response characteristics of balanced steady - state free precession bSSFP, adopts a magnetic resonance imaging acquisition mode, and uses a more efficient frequency encoding for spatial encoding in the readout direction Gx. Compared with the traditional chemical shift imaging that uses inefficient phase encoding for spatial encoding in both the readout direction Gx and the phase - encoding direction Gy, the present invention greatly improves the data acquisition efficiency.

[0046] 2. The pulse sequence based on prior knowledge designed by the present invention according to prior knowledge (i.e., the chemical shift of the compound) can adjust the number of magnetic resonance image acquisitions according to the number of types of metabolite compounds in the metabolite to be separated, and can meet the needs of observing different metabolic pathways (i.e., separating different numbers of types of metabolite compounds), especially in the case of generating multiple metabolites.

[0047] 3. The repetition time TR of the pulse sequence based on prior knowledge designed by the present invention is shorter, and the signal enhancement effect is better. Compared with other imaging methods based on balanced steady - state free precession bSSFP technology, the signal - to - noise ratio of the obtained image is higher. Secondly, the proportion of the dark band area of the bSSFP signal is less, and it is not easily affected by magnetic field inhomogeneity.

[0048] 4. The reconstruction algorithm based on echo asymmetry iterative decomposition and least squares estimation adopted by the present invention iteratively decomposes the magnetic field inhomogeneity, which can further weaken the influence of the magnetic field inhomogeneity.

[0049] 5. The algorithm adopted by the present invention converts the signal intensity value into the concentration of metabolic compounds, which can achieve the absolute quantification of the concentration of metabolic compounds and provide very important guiding significance for the metabolic analysis of living bodies. Brief Description of the Drawings

[0050] Figure 1 is the implementation flowchart of a high spatio-temporal resolution magnetic resonance metabolic imaging acquisition and reconstruction method of the present invention.

[0051] Figure 2 is the schematic diagram of the pulse sequence based on prior knowledge of the present invention;

[0052] represents the number of phase encoding steps of the balanced steady-state free precession bSSFP sequence; TR is the repetition time, TE is the echo time; N is the total number of the balanced steady-state free precession bSSFP sequence;

[0053] Gz is the slice direction, Gy is the phase encoding direction, and Gx is the readout direction;

[0054] G1 is the first spoiling gradient, G2 is the second spoiling gradient, and G3 is the third spoiling gradient;

[0055] is the pulse phase difference;

[0056] represents the first radio frequency pulse corresponding to each step of phase encoding in the first balanced steady-state free precession bSSFP sequence, represents the first radio frequency pulse corresponding to each step of phase encoding in the second balanced steady-state free precession bSSFP sequence, represents the N th first radio frequency pulse corresponding to each step of phase encoding in the

[0057] represents the second radio frequency pulse corresponding to each step of phase encoding in the first balanced steady-state free precession bSSFP sequence, represents the second radio frequency pulse corresponding to each step of phase encoding in the second balanced steady-state free precession bSSFP sequence, represents the N th second radio frequency pulse corresponding to each step of phase encoding in the

[0058] Figure 3a It is the signal intensity simulation curve of pulse phase difference - steady - state transverse magnetization vector;

[0059] Figure 3b It is the phase value simulation curve of pulse phase difference - steady - state transverse magnetization vector;

[0060] Figure 4a It is the intensity map obtained by decomposing the magnetic resonance image acquired using the pulse sequence based on prior knowledge of the present invention;

[0061] Figure 4b It is the phase map obtained by decomposing the magnetic resonance image acquired using the pulse sequence based on prior knowledge of the present invention;

[0062] Among them, - 240°, - 135°, 0°, 135° and 240° are the numerical values of the pulse phase difference selected in the embodiments of the present invention.

[0063] Figure 5 It is the finally obtained compound concentration image. Detailed implementation manners

[0064] For the convenience of those of ordinary skill in the art to understand and implement the present invention, the present invention will be further described in detail below in conjunction with embodiments. It should be understood that the embodiments described herein are only used to illustrate and explain the present invention and are not used to limit the present invention.

[0065] Embodiment 1

[0066] A method for magnetic resonance metabolic imaging acquisition and reconstruction with high spatio - temporal resolution. Load a pulse sequence based on prior knowledge on a magnetic resonance imaging system supporting sequence editing. Obtain 5 pulse phase differences for separating 3 deuterated compounds (water, glucose, and lactate) through simulation. Acquire 5 magnetic resonance images using the pulse sequence based on prior knowledge, and then separate the compound concentration images corresponding to the 3 deuterated compounds through a reconstruction algorithm.

[0067] A method for magnetic resonance metabolic imaging acquisition and reconstruction with high spatio - temporal resolution, the process is as Figure 1 shown, and includes the following steps:

[0068] Step 1: Construct a pulse sequence based on prior knowledge required for the experiment on a pulse sequence implementation platform and determine the sequence parameters. Specifically, it includes the following steps:

[0069] Step 1.1: Construct a pulse sequence based on prior knowledge. The pulse sequence based on prior knowledge sequentially includes a plurality of balanced steady - state free precession bSSFP sequences. In each balanced steady - state free precession bSSFP sequence, a pulse phase difference is added between the radio - frequency pulses corresponding to each step of phase encoding:

[0070] To obtain multiple magnetic resonance images required for metabolites to be separated, the designed pulse sequence based on prior knowledge successively includes multiple balanced steady-state free precession bSSFP sequences;

[0071] To make the balanced steady-state free precession bSSFP sequences acquired each time obtain different magnetic resonance images, a pulse phase difference is added between the radio frequency pulses corresponding to each step of phase encoding in each balanced steady-state free precession bSSFP sequence. The numerical values of the pulse phase differences corresponding to all phase encodings in the same balanced steady-state free precession bSSFP sequence are the same, and the numerical values of the pulse phase differences corresponding to different balanced steady-state free precession bSSFP sequences are different. As Figure 2 shown, the number of phase encoding steps of each balanced steady-state free precession bSSFP sequence is , and each step of phase encoding corresponds to 2 radio frequency pulses (i.e., the first radio frequency pulse and the second radio frequency pulse). Denote the pulse phase difference of the th balanced steady-state free precession bSSFP sequence as pulse phase difference , represents the serial number of the balanced steady-state free precession bSSFP sequence, , N is the total number of balanced steady-state free precession bSSFP sequences. In this embodiment, a pulse phase difference is added between the first radio frequency pulse and the second radio frequency pulse corresponding to each step of phase encoding in the 1st balanced steady-state free precession bSSFP sequence. A pulse phase difference is added between the first radio frequency pulse and the second radio frequency pulse corresponding to each step of phase encoding in the 2nd balanced steady-state free precession bSSFP sequence, and so on. A pulse phase difference is added between the first radio frequency pulse and the second radio frequency pulse corresponding to each step of phase encoding in the th balanced steady-state free precession bSSFP sequence. A pulse phase difference N is added between the first radio frequency pulse and the second radio frequency pulse corresponding to each step of phase encoding in the th balanced steady-state free precession bSSFP sequence.

[0072] To eliminate the signal residue from the previous acquisition, between adjacent balanced steady-state free precession bSSFP sequences, a first spoiling gradient G1 is added in the slice direction Gz, a second spoiling gradient G2 is added in the phase encoding direction Gy, and a third spoiling gradient G3 is added in the readout direction Gx; among them, frequency encoding is used for spatial encoding in the readout direction Gx.

[0073] Step 1.2: Determine the numerical values of the phase differences of each pulse in the pulse sequence based on prior knowledge and the total number of balanced steady-state free precession (bSSFP) sequences N :

[0074] According to the relationship between the steady-state transverse magnetization vector corresponding to the initial signal intensity and the pulse phase difference, and the preset simulated value of the initial signal intensity, simulate the signal simulation curves of the pulse phase difference - steady-state transverse magnetization vector corresponding to each metabolite in the metabolite to be separated. According to the signal simulation curves of the pulse phase difference - steady-state transverse magnetization vector of each metabolite, determine the numerical values of the phase differences of each pulse in the pulse sequence based on prior knowledge; determine the total number of balanced steady-state free precession (bSSFP) sequences according to the number of types of metabolites in the metabolite to be separated N . Specifically, it includes the following steps:

[0075] Step 1.2.1: The metabolite to be separated includes metabolites with known concentrations and metabolites with unknown concentrations. According to the relationship between the steady-state transverse magnetization vector corresponding to the initial signal intensity and the pulse phase difference, and the preset simulated value of the initial signal intensity, simulate the signal simulation curves of the pulse phase difference - steady-state transverse magnetization vector corresponding to each metabolite in the metabolite to be separated. The steady-state transverse magnetization vector is in complex form and can be decomposed into a signal intensity and a corresponding phase value. Therefore, the signal simulation curve of the pulse phase difference - steady-state transverse magnetization vector can be decomposed into the signal intensity simulation curve of the pulse phase difference - steady-state transverse magnetization vector and the phase value simulation curve of the pulse phase difference - steady-state transverse magnetization vector; place the signal intensity simulation curves of the pulse phase difference - steady-state transverse magnetization vector of each metabolite in the same coordinate system; place the phase value simulation curves of the pulse phase difference - steady-state transverse magnetization vector of each metabolite in another coordinate system

[0076] For each metabolite, in each balanced steady-state free precession (bSSFP) sequence, the relationship between the steady-state transverse magnetization vector of the corresponding balanced steady-state free precession (bSSFP) and the initial signal intensity and the pulse phase difference satisfies the following formula: (Formula 2)

[0077] Where represents the steady-state transverse magnetization vector of the th metabolite, is the serial number of the metabolite; is the imaginary unit, intermediate quantity , intermediate quantity , intermediate quantity , intermediate quantity , intermediate quantity , intermediate quantity ; is the longitudinal relaxation time; is the transverse relaxation time; is the flip angle of the radiofrequency pulse; is the frequency offset caused by magnetic field inhomogeneity; is the frequency offset caused by the chemical shift of the th metabolite compound, denoted as the frequency offset caused by chemical shift ; is the repetition time, is the echo time; is the initial signal intensity of the th metabolite compound, denoted as the initial signal intensity , and the initial signal intensity

[0078] In step 1.2.1, only the change trend of the steady-state transverse magnetization vector needs to be obtained. Therefore, for each metabolite compound, the signal simulation curve of the pulse phase difference - steady-state transverse magnetization vector is simulated based on the following formula: (Formula 3)

[0079] represents the simulated value of the steady-state transverse magnetization vector of the th metabolite compound; represents the preset simulated value of the initial signal intensity of the th metabolite compound, represents the preset range value of the pulse phase difference. In this embodiment, is taken; In this embodiment, the metabolites to be separated include water (the serial number of the metabolite compound corresponding to water ), glucose (the serial number of the metabolite compound corresponding to glucose ), and lactate (the serial number of the metabolite compound corresponding to lactate ) 3 metabolite compounds. Water is a metabolite compound with a known concentration, glucose and lactate are metabolite compounds with unknown concentrations. The longitudinal relaxation time of water is 320 milliseconds, and the transverse relaxation time of water is 12 milliseconds; The longitudinal relaxation time of glucose is 64 milliseconds, and the transverse relaxation time of glucose is 32 milliseconds; The longitudinal relaxation time of lactate is 297 milliseconds, and the transverse relaxation time of lactate is 61 milliseconds; The flip angle of the radiofrequency pulse is 40°; The frequency offset caused by magnetic field inhomogeneity can be obtained from the main magnetic field distribution map; the frequency shift caused by the chemical shift of water in this example is 0 Hz, and the frequency shift caused by the chemical shift of glucose is -68 Hz, and the frequency shift caused by the chemical shift of lactate is -215 Hz. The signal intensity simulation curves of the pulsed phase difference - steady-state transverse magnetization vector corresponding to water, glucose, and lactate are as Figure 3a shown, and the phase value simulation curves of the pulsed phase difference - steady-state transverse magnetization vector corresponding to water, glucose, and lactate are as Figure 3b shown.

[0080] Step 1.2.2. Determine the pulsed phase differences in the pulse sequence based on prior knowledge according to the signal intensity simulation curves of the pulsed phase difference - steady-state transverse magnetization vector of each metabolite, the phase value simulation curves of the pulsed phase difference - steady-state transverse magnetization vector, and the following principles of the numerical values and the total number of balanced steady-state free precession bSSFP sequences N :

[0081] To ensure the separation effect of the magnetic resonance images of metabolites during reconstruction, when determining the numerical values of the pulsed phase differences , for the magnetic resonance images acquired at each pulsed phase difference , the phase values of the steady-state transverse magnetization vectors of the corresponding metabolites should change, that is, for each metabolite, the numerical values of the determined pulsed phase differences correspond to different phase values of the simulated values of the steady-state transverse magnetization vectors in the phase value simulation curve of the pulsed phase difference - steady-state transverse magnetization vector;

[0082] Therefore, the numerical values of the pulsed phase differences corresponding to different balanced steady-state free precession bSSFP sequences should be different, and the numerical values of the pulsed phase differences corresponding to all phase encodings in the same balanced steady-state free precession bSSFP sequence are the same, that is, there are pulsed phase differences .

[0083] To ensure a high signal-to-noise ratio of the magnetic resonance images corresponding to each metabolite after separation and reconstruction, the determined numerical values of the pulsed phase differences are not in the bSSFP signal dark band regions in the signal intensity simulation curves of the pulsed phase difference - steady-state transverse magnetization vector of each metabolite;

[0084] The total number of balanced steady-state free precession bSSFP sequences N is greater than or equal to the number of metabolite types. Since the pulsed phase difference equals the same value, so the pulse phase difference The number of values equals the total number of balanced steady-state free precession bSSFP sequences N , in order to ensure the separation effect of metabolite compounds in subsequent reconstruction, the number of values of the pulse phase difference needs to follow being greater than or equal to the number of types of metabolite compounds, that is, the total number of balanced steady-state free precession bSSFP sequences N is greater than or equal to the number of types of metabolite compounds. In this embodiment, the total number of balanced steady-state free precession bSSFP sequences N is set to be equal to the number of metabolite compounds plus 2, then the number of values of the pulse phase difference is 5 (i.e., the total number of sequences N= 5), and the 5 values of the pulse phase difference are -240°, -135°, 0°, 135°, 240° respectively;

[0085] Step 2. According to the pulse sequence based on prior knowledge constructed in Step 1, for each balanced steady-state free precession bSSFP sequence, perform magnetic resonance image acquisition; specifically including the following steps:

[0086] Step 2.1. Before performing magnetic resonance image acquisition using the pulse sequence based on prior knowledge constructed in Step 1, perform regional shimming on the region of interest in advance to make the field in the region of interest more uniform, so as to obtain a better reconstruction effect. The region of interest includes the scanning region of the metabolite to be separated.

[0087] Step 2.2. Before performing image acquisition after shimming, a non-localized spectrum (Non-locolized spectrocopy) should be acquired. Non-localized refers to any region in space, and the non-localized spectrum is a one-dimensional single-pulse spectrum;

[0088] Determine the frequency shift caused by the actual chemical shift of a metabolite compound in the metabolite to be separated through the non-localized spectrum. If the frequency shift caused by the actual chemical shift of the metabolite compound is different from the frequency shift caused by the chemical shift during the simulation in Step 1.2 , then adjust the center frequency of the pulse sequence based on prior knowledge until the frequency shift caused by the actual chemical shift of the compound is the same as the frequency shift caused by the chemical shift during the simulation in Step 1.2 . In this embodiment, the frequency shift caused by the actual chemical shift is different from the frequency shift caused by the chemical shift during the simulation by -2.1 ppm (-129 Hz), so the center frequency of the pulse sequence based on prior knowledge is set to +129 Hz.

[0089] Step 2.3: Based on the pulse sequence with prior knowledge obtained after adjustment in Step 2.2, perform magnetic resonance imaging (MRI) acquisition on the metabolites to be separated. One balanced steady-state free precession (bSSFP) sequence corresponds to one acquired MRI image. Therefore, the MRI image with the serial number equal to corresponds to the pulse phase difference . Each MRI image is in complex form and can be separated into a corresponding intensity map (as shown in Figure 4a ) and a phase map (as shown in Figure 4b ).

[0090] Step 3: Based on the relationship between the steady-state transverse magnetization vector of each metabolite compound and the initial signal intensity and pulse phase difference, reconstruct the MRI images acquired in Step 2 to obtain the initial signal separation images corresponding to each metabolite compound. Specifically, it includes the following steps:

[0091] For the MRI images acquired based on each pulse phase difference, the total steady-state transverse magnetization vector corresponding to all metabolite compounds satisfies the following condition: (Equation 4)

[0092] represents the total steady-state transverse magnetization vector signal of all metabolite compounds in the MRI image with the serial number equal to , denoted as the total steady-state transverse magnetization vector , and the total steady-state transverse magnetization vector belongs to the measured value; represents the number of types of metabolite compounds in the metabolites to be separated, and the serial number of the metabolite compound is ; represents the theoretical value of the steady-state transverse magnetization vector of the th metabolite compound corresponding to the balanced steady-state free precession bSSFP sequence with the serial number equal to , and satisfies the following formula: (Equation 5)

[0093] represents the pulse phase difference of the th balanced steady-state free precession bSSFP sequence.

[0094] Then the total steady-state transverse magnetization vector satisfies the following condition: (Equation 6)

[0095] Separate different metabolite compounds: For the N acquired MRI images, rearrange Equation 6 into matrix form: (Equation 7)

[0096] Among them, is the coefficient matrix, and the coefficient matrix can be calculated by Formula 6, that is, the elements in the coefficient matrix satisfy the following conditions: (Formula 8)

[0097] represents the element in the n th row and m th column of the coefficient matrix.

[0098] Then, use the reconstruction algorithm based on echo asymmetry iterative decomposition and least squares estimation to inversely solve Formula 7 to obtain the initial signal intensity corresponding to each metabolite. The initial signal intensity is data in matrix form. The initial signal intensity corresponding to the metabolite is presented in the form of an image to obtain the initial signal separation image of each metabolite;

[0099] Step 4: For the initial signal separation image corresponding to each metabolite, based on the concentration mean of the known concentration metabolites and the initial signal intensity mean of the known concentration metabolites, obtain the concentration conversion factor, and finally use the concentration conversion factor to convert the initial signal separation image corresponding to each metabolite into the corresponding compound concentration image. Specifically, it includes the following steps:

[0100] Step 4.1: Calculate the concentration conversion factor according to the following formula : (Formula 9)

[0101] is the initial signal intensity mean of the known concentration metabolites, is the concentration mean of the known concentration metabolites. In this embodiment, water is the known concentration metabolite, then the initial signal intensity mean (dimensionless) corresponding to water, and the concentration mean corresponding to water is 20 mmol / L, so the concentration conversion factor is

[0102] Step 4.2: Calculate the concentration of each metabolite by the following formula: (Formula 10)

[0103] Among them, is the serial number of The concentration of the metabolic compound, denoted as the concentration of the metabolic compound ; is the number of molecules contributing to the signal by the metabolic compound with a known concentration; is the number of molecules contributing to the signal by the metabolic compound (the metabolic compound is a metabolic compound with a known concentration or an unknown concentration).

[0104] In this embodiment, taking the maximum value of the initial signal intensity as an example, the number of molecules contributing to the signal by water is 2, that is, the number of molecules contributing to the signal by the metabolic compound with a known concentration is 2, the maximum value of the initial signal intensity of glucose is 10 9 , the number of molecules contributing to the signal by glucose is 2, so the maximum value of the concentration of glucose is 10 millimoles; the maximum value of the initial signal intensity of lactic acid is 9 10 8 , the number of molecules contributing to the signal by lactic acid is 3, so the maximum value of the concentration of lactic acid is 6 millimoles per liter.

[0105] The concentration of the metabolic compound is data in matrix form. The concentration of the metabolic compound is presented in the form of an image to obtain the corresponding compound concentration image. In this embodiment, the compound concentration images corresponding to glucose, water, and lactic acid are as Figure 5 shown.

[0106] The present invention utilizes the frequency response characteristics of balanced steady-state free precession bSSFP, adopts a magnetic resonance imaging acquisition mode, and uses more efficient frequency encoding for spatial encoding in the readout direction Gx. While traditional chemical shift imaging uses inefficient phase encoding for spatial encoding in both the readout direction Gx and the phase encoding direction Gy, greatly improving the data acquisition efficiency. In this embodiment, when separating 3 compounds and the image matrix size is 32×32, the scanning time of our method is 32×repetition time TR ×5, while the scanning time of traditional chemical shift imaging is 32×32×repetition time TR .

[0107] Example 2

[0108] A magnetic resonance metabolic imaging acquisition and reconstruction device with high spatio-temporal resolution, based on a magnetic resonance metabolic imaging acquisition and reconstruction method of Example 1, includes:

[0109] A sequence construction module for implementing step 1 described in Example 1;

[0110] A magnetic resonance image acquisition module for acquiring magnetic resonance images corresponding to metabolites to be separated;

[0111] An initial signal separation image acquisition module, configured to implement step 3 described in Embodiment 1;

[0112] A compound concentration image acquisition module, configured to implement step 4 described in Embodiment 1.

[0113] Embodiment 3

[0114] A computer device includes a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, it implements step 1 and steps 3 - 4 in the above-mentioned Embodiment 1.

[0115] Embodiment 4

[0116] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, it implements step 1 and steps 3 - 4 in the above-mentioned Embodiment 1.

[0117] Embodiment 5

[0118] A computer program product includes a computer program. When the computer program is executed by a processor, it implements step 1 and steps 3 - 4 in the above-mentioned Embodiment 1.

[0119] The content not described in detail in the present invention belongs to the prior art well-known to those skilled in the art.

[0120] The specific embodiments described herein are merely illustrative of the spirit of the present invention. Those skilled in the art to which the present invention pertains can make various modifications or supplements to the described specific embodiments or use similar ways to substitute them, but will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.

Claims

1. A method for magnetic resonance metabolic imaging acquisition and reconstruction with high spatio-temporal resolution, characterized in that, It includes the following steps: Step 1: Construct a pulse sequence based on prior knowledge. The pulse sequence based on prior knowledge sequentially includes a plurality of balanced steady-state free precession (bSSFP) sequences. In each balanced steady-state free precession (bSSFP) sequence, a pulse phase difference is added between the radiofrequency pulses corresponding to each step of phase encoding. According to the relationship between the steady-state transverse magnetization vector corresponding to the initial signal intensity and the pulse phase difference, and the simulated preset value of the initial signal intensity, simulate the signal simulation curves of the pulse phase difference - steady-state transverse magnetization vector corresponding to each metabolite in the metabolite to be separated. Determine the numerical values of each pulse phase difference in the pulse sequence based on prior knowledge according to the signal simulation curves of the pulse phase difference - steady-state transverse magnetization vector of each metabolite. Determine the total number of balanced steady-state free precession (bSSFP) sequences according to the number of types of metabolites in the metabolite to be separated. The metabolite to be separated includes metabolites with known concentrations and metabolites with unknown concentrations. Step 2: Based on the pulse sequence based on prior knowledge constructed in Step 1, acquire the magnetic resonance image corresponding to the metabolite to be separated for each balanced steady-state free precession (bSSFP) sequence. Step 3: According to the relationship between the steady-state transverse magnetization vector corresponding to each metabolite and the initial signal intensity and the pulse phase difference, reconstruct the magnetic resonance image acquired in Step 2 to obtain the initial signal separation images corresponding to each metabolite. Step 4: Based on the mean value of the initial signal intensities of the metabolites with known concentrations and the mean value of the concentrations of the metabolites with known concentrations, obtain the concentration conversion factor. Use the concentration conversion factor to convert the initial signal separation images corresponding to each metabolite into the corresponding compound concentration images.

2. The method for magnetic resonance metabolic imaging acquisition and reconstruction with high spatio-temporal resolution according to claim 1, wherein Between adjacent balanced steady-state free precession (bSSFP) sequences, increase the first spoiling gradient G1 in the slice direction Gz, increase the second spoiling gradient G2 in the phase encoding direction Gy, and increase the third spoiling gradient G3 in the readout direction Gx; perform spatial encoding using frequency encoding in the readout direction Gx.

3. A method for collecting and reconstructing magnetic resonance metabolic imaging with high spatio-temporal resolution according to claim 1, characterized in that, In Step 1, according to the relationship between the steady-state transverse magnetization vector corresponding to the initial signal intensity and the pulse phase difference, and the simulated preset value of the initial signal intensity, simulate the signal simulation curves of the pulse phase difference - steady-state transverse magnetization vector corresponding to each metabolite in the metabolite to be separated, which specifically includes the following steps: Simulate the signal simulation curves of the pulse phase difference - steady-state transverse magnetization vector based on the following formula: ， Among them, represents the simulated value of the steady-state transverse magnetization vector of the th metabolite compound; represents the preset simulated value of the initial signal intensity of the th metabolite compound, represents the preset range value of the pulse phase difference; is the serial number of the metabolite compound; is the imaginary unit, intermediate quantity , intermediate quantity , intermediate quantity , intermediate quantity , intermediate quantity , intermediate quantity ; is the longitudinal relaxation time; Transverse relaxation time; is the flip angle of the radio frequency pulse; is the frequency offset caused by magnetic field inhomogeneity; is the frequency offset caused by the chemical shift of the th metabolite compound, denoted as the frequency offset caused by chemical shift ; is the repetition time, is the echo time; Decompose the steady-state transverse magnetization vector into a signal intensity and the corresponding phase value, so that the signal simulation curve of the pulse phase difference - steady-state transverse magnetization vector is decomposed into the signal intensity simulation curve of the pulse phase difference - steady-state transverse magnetization vector and the phase value simulation curve of the pulse phase difference - steady-state transverse magnetization vector. Place the signal intensity simulation curves of the pulse phase difference - steady-state transverse magnetization vector of each metabolite in the same coordinate system; place the phase value simulation curves of the pulse phase difference - steady-state transverse magnetization vector of each metabolite in another coordinate system.

4. A method for magnetic resonance metabolic imaging acquisition and reconstruction with high spatio-temporal resolution according to claim 3, characterized in that, Determine the numerical value of the pulse phase difference according to the following rules: For each metabolic compound, the numerical values of the determined pulse phase differences are different from the phase values of the simulated values of the steady-state transverse magnetization vectors corresponding to them in the simulated curve of the pulse phase difference - steady-state transverse magnetization vector phase value; In the same balanced steady-state free precession bSSFP sequence, the numerical values of the pulse phase differences corresponding to all phase encodings are the same; The numerical values of the determined pulse phase differences are not in the bSSFP signal dark band region of the simulated curve of the pulse phase difference - steady-state transverse magnetization vector signal intensity for each metabolic compound.

5. A method for magnetic resonance metabolic imaging acquisition and reconstruction with high spatio-temporal resolution according to claim 1, characterized in that The total number of the balanced steady-state free precession bSSFP sequences is greater than or equal to the number of types of metabolic compounds.

6. A method for collecting and reconstructing magnetic resonance metabolic imaging with high spatio-temporal resolution according to claim 3, characterized in that Step 2 includes the following steps: Step 2.1, perform regional shimming on the region of interest; Step 2.2: Collect a non-localized spectrum, and determine the frequency shift caused by the actual chemical shift of a metabolic compound in the metabolites to be separated through the non-localized spectrum. If the frequency shift caused by the actual chemical shift of the metabolic compound is different from that caused by the chemical shift during the simulation in Step 1, adjust the center frequency of the pulse sequence based on prior knowledge until the frequency shift caused by the actual chemical shift of the metabolic compound is the same as that caused by the chemical shift during the simulation in Step 1; Step 2.3, according to the pulse sequence based on prior knowledge obtained in Step 2.2, perform magnetic resonance image acquisition on the metabolite to be separated, and one magnetic resonance image is acquired corresponding to each balanced steady-state free precession bSSFP sequence.

7. A method for magnetic resonance metabolic imaging acquisition and reconstruction with high spatio-temporal resolution according to claim 3, characterized in that, Step 3 includes the following steps: The reconstruction algorithm based on echo asymmetry iterative decomposition and least squares estimation performs inverse calculation as follows, and the initial signal intensity corresponding to each metabolite is obtained from the following formula: , Among them, represents the total steady-state transverse magnetization vector signal of all metabolic compounds in the magnetic resonance image with the serial number equal to ; represents the serial number of the balanced steady-state free precession bSSFP sequence, ; N is the total number of the balanced steady-state free precession bSSFP sequences; is the initial signal intensity of the th metabolic compound, denoted as the initial signal intensity ; M represents the number of types of metabolic compounds in the metabolite to be separated; is the serial number of the metabolic compound, ; is the coefficient matrix, and the elements in the coefficient matrix satisfy the following conditions: , represents the pulse phase difference of the < n th row and the m th column of the coefficient matrix; The initial signal intensity corresponding to the metabolic compound Presented in the form of an image to obtain the initial signal separation image of each metabolic compound.

8. A method for magnetic resonance metabolic imaging acquisition and reconstruction with high spatio-temporal resolution according to claim 7, characterized in that, Step 4 includes the following steps: Step 4.

1. Calculate the concentration conversion factor according to the following formula :[[]] ​ is the mean initial signal intensity of the known concentration metabolite compound, is the mean concentration of the known concentration metabolite compound; Step 4.

2. Calculate the concentrations of each metabolite compound using the following formula: , Among them, is the concentration of the metabolite compound with the serial number ; is the number of molecules of the metabolite compound with a known concentration contributing to the signal, is the number of molecules of the metabolite compound contributing to the signal; The concentration of the metabolic compound is presented in the form of an image to obtain a corresponding compound concentration image.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements Step 1 and Steps 3 - 4 of the magnetic resonance metabolic imaging acquisition and reconstruction method described in any one of Claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements Step 1 and Steps 3 - 4 of the magnetic resonance metabolic imaging acquisition and reconstruction method described in any one of Claims 1 to 8.

Images