Ultra-high field T1, T2 and T based on magnetic resonance fingerprint 1ρ Quantitative measurement methods

Abstract

The present invention discloses ultra-high field T1, T2 and T based on magnetic resonance fingerprint 1ρ Quantitative measurement method, which can achieve rapid quantitative analysis of tissue T1, T2 and T under ultra-high field conditions 1ρ The present invention introduces T1, T2 and T in the signal evolution by adding a preparation module containing adiabatic pulses to the traditional MRF sequence. 1ρ Weight. Combined with the EPG method, the evolution of the magnetization vector was simulated, and a magnetic resonance fingerprint dictionary with different relaxation time value combinations was established. The reconstructed image acquired and reconstructed was matched with the magnetic resonance fingerprint dictionary to obtain T1, T2 and T 1ρ Quantitative graph. Compared with the traditional method, the method of the present invention does not need to collect multiple groups of signals separately for exponential fitting, which greatly shortens the scanning time.

Description

Technical Field

[0001] The present invention belongs to the field of quantitative imaging methods, and in particular to the field of ultra-high field T1, T2 and T based magnetic resonance fingerprint. 1ρ Quantitative measurement methods. Background Art

[0002] Magnetic resonance imaging (MRI) is a commonly used medical imaging method. It has the advantages of being non-invasive, non-ionizing radiation, and having no tissue penetration depth limitation. It can provide quantitative information about tissue properties and microstructure.

[0003] Magnetic resonance fingerprinting (MRF) was originally proposed by Ma.D et al. [MaD, Gulani V, Seiberlich N, et al. Magnetic resonance fingerprinting. Nature. 2013; 495: 187–192]. It is an innovative technology for medical imaging that combines the principles of magnetic resonance imaging and pattern recognition to obtain richer tissue information and provide more accurate analysis. The core idea of ​​MRF is to obtain a unique "fingerprint" by applying a specific sequence to the magnetic resonance signal in time and parameter space, matching it with the dictionary obtained by Bloch equation simulation, finding the best matching curve, and thus obtaining a multi-quantitative parameter map.

[0004] Most quantitative MRI methods allow for the measurement of tissue properties. Methods for quantifying a single property involve measuring the change in signal after varying a single acquisition parameter while holding other parameters constant, but are relatively slow and typically only provide one property at a time. Faster alternatives have been developed, including some commonly used techniques, but these are often affected by other unwanted tissue or system properties, such as B1, diffusion, and magnetization transfer, which can introduce errors. Magnetic resonance fingerprinting is a framework that uses different modalities to achieve the goal of rapid multi-attribute quantitative imaging. In this framework, all desired tissue property information is probed simultaneously to improve the signal-to-noise ratio and scanning efficiency of the entire experiment.

[0005] T 1ρIt is the spin-lattice relaxation time in a rotating coordinate system. Due to its sensitivity to the slow motion process of macromolecules, it is a valuable biomarker for assessing tissue properties. Yang.Y et al. [Yang.Y, Wang.C et al. Arobust adiabatic constant amplitude spin-lock preparation module formyocardial T1ρ quantification at 3T. NMR Biomed. 2023, 36(2), e4830] used a composite adiabatic pulse for the T of tissues. 1ρ Improved robustness to B0 and B1 field defects, showing fewer artifacts. 1ρ Quantitative calculation requires sampling the time of different spin-locking pulses and obtaining T through exponential fitting. 1ρ Quantitative information requires a lot of time and cost. And the method can only measure the T of the tissue 1ρ Relaxation time cannot provide more quantitative information.

[0006] In vivo, the traditional T1, T2 and T 1ρ Relaxation imaging methods require the acquisition of images at different TI, TE or TSL, resulting in a long scanning time. The subject's movement can seriously affect the imaging quality, especially in abdominal imaging.

[0007] In view of the above problems, the present invention proposes a method based on the ultra-high field T1, T2 and T 1ρ Quantitative measurement method, capable of simultaneously quantifying T1, T2 and T 1ρ The images are all inherently co-registered, without image distortion, and are sampled quickly and robust to motion. This method designs a new MRI pulse sequence that converts the T1, T2, and T 1ρ It provides a new platform for quantitative analysis. Summary of the invention

[0008] In order to solve the above problems existing in the existing quantitative methods, the present invention provides a method based on the ultra-high field T1, T2 and T 1ρ Quantitative methods.

[0009] To achieve the above object, the present invention specifically proposes the following solutions:

[0010] Ultra-high field T1, T2 and T based on magnetic resonance fingerprint 1ρ The quantitative measurement method includes the following steps:

[0011] Step 1, designing a magnetic resonance pulse sequence, which includes randomly alternating T1 preparation modules, T2 preparation modules and T 1ρPreparation module, where T 1ρ The preparation module uses adiabatic half-pass RF pulses, and both the T2 preparation module and the T1 preparation module use adiabatic pulses;

[0012] Step 2, using the extended phase map (EPG) method combined with the magnetic resonance pulse sequence designed in step 1 to simulate the magnetic resonance signal evolution process and calculate the magnetic resonance fingerprint dictionary;

[0013] Step 3, using the magnetic resonance pulse sequence designed in step 1, sampling to obtain magnetic resonance k-space data;

[0014] Step 4, reconstructing the magnetic resonance k-space data to obtain a reconstructed image;

[0015] Step 5: Match and identify the reconstructed image with the magnetic resonance fingerprint dictionary to obtain T1, T2 and T 1ρ Multi-parameter quantitative diagram, where T1 refers to the spin-lattice relaxation time, T2 refers to the spin-spin relaxation time, and T 1ρ Refers to the spin-lattice relaxation time in the rotating coordinate system.

[0016] As mentioned in step 1:

[0017] T 1ρ The preparation module starts with a 90° adiabatic half-pass RF pulse along the x-axis to flip the magnetization vector to the XY plane; subsequently, a continuous spin-locking pulse is applied along the y-axis to lock the transverse magnetization; during the spin-locking time TSL, the magnetization is adjusted according to the spin-lattice relaxation time T in the rotating coordinate system. 1ρ decay; then applying a second 90° adiabatic half-pass RF pulse along the -x axis will return the magnetization vector to the longitudinal direction;

[0018] The T1 preparation module starts with a 180° adiabatic pulse along the x-axis, flipping the magnetization vector from the z-axis to the -z-axis;

[0019] The T2 preparation module starts with a 90° rectangular RF pulse along the x-axis to flip the magnetization vector to the XY plane; after the spin-spin relaxation time T2, two 180° adiabatic pulses are applied to reduce the influence of the inhomogeneity of the main magnetic field B0 and the RF field B1; then after the spin-spin relaxation time T2, a second 90° rectangular RF pulse is applied along the -x axis to return the magnetization vector to the longitudinal direction.

[0020] As mentioned above, T1 preparation module, T2 preparation module and T 1ρ After the preparation module is completed, a damage gradient is applied in the Z-axis direction.

[0021] After each spoiled gradient as described above, a FISP (fast imaging in steady state free precession) acquisition module is first applied, and a waiting time is applied after the FISP acquisition module ends.

[0022] In the above-mentioned T1 preparation module, T2 preparation module and T 1ρ The same repetition time TR, the same echo time TE and different flip angles FA are used in each pulse excitation of the preparation module.

[0023] As described above, in step 2, the magnetic resonance pulse sequence designed in step 1 is combined with the T1 preparation module, the T2 preparation module and the T 1ρ Preparation module, the simulated magnetic resonance signal evolution is as follows:

[0024]

[0025] Where S0 represents the MRI signal strength before applying the preparation module, S1 represents the MRI signal strength after the T1 preparation module, S2 represents the MRI signal strength after the T2 preparation module, and S3 represents the MRI signal strength after the T 1ρ The magnetic resonance signal intensity after preparing the module; T1 refers to the spin-lattice relaxation time, T2 refers to the spin-spin relaxation time, T 1ρ refers to the spin-lattice relaxation time in the rotating coordinate system, TI is the inversion recovery time, T 2prep T2 is the preparation time for the module, and TSL is the spin lock time.

[0026] The specific steps of reconstructing the magnetic resonance k-space data in step 4 to obtain the reconstructed image are as follows:

[0027] (1) Calculate the density compensation function of the magnetic resonance k-space data using a spiral sampling trajectory;

[0028] (2) The magnetic resonance k-space data is transformed into a reconstructed image based on the density compensation function and the non-uniform Fourier transform method.

[0029] As described above, the specific method for matching and identifying the reconstructed image with the magnetic resonance fingerprint dictionary in step 5 is: performing dot product calculation on the reconstructed image and the magnetic resonance fingerprint dictionary for matching.

[0030] Compared with the existing methods, the present invention has the following beneficial effects:

[0031] 1. The present invention introduces T1 preparation module, T2 preparation module and T 1ρ Prepare the module so that it can obtain T1, T2, and T in a single scan 1ρ Three parameters can provide more comprehensive organizational information.

[0032] 2. The preparation module in the pulse sequence uses adiabatic pulses, which greatly reduces the artifacts caused by B0 and B1 field defects in ultra-high fields.

[0033] 3. In the traditional method, multiple data need to be collected for fitting to obtain a quantitative parameter map, which takes a long time. This method does not need to collect multiple data for fitting. A single acquisition can complete multi-parameter imaging in a relatively short time, reducing scanning time and reducing the impact on the subject's discomfort and movement.

[0034] 4. This method is based on signal pattern matching rather than a specific image reconstruction method, which makes it more robust to motion artifacts. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flow chart of the method of the present invention;

[0036] Figure 2 The magnetic resonance pulse sequence diagram of the present invention and its related parameter settings, wherein (a) is the magnetic resonance pulse sequence diagram of the present invention, (b) is the amplitude and phase of the pulse in the T1 preparation module of the present invention respectively changing with time, (c) is the amplitude and phase of the pulse in the T2 preparation module of the present invention respectively changing with time, (d) is the T 1ρ The amplitude and phase of the pulse in the preparation module change with time, where w1 represents the amplitude, w represents the phase, and t represents the time; (e) is a list of pseudo-randomly changing RF pulse flip angles, and (f) is a spiral sampling trajectory with variable density, where kx and ky represent the x-axis and y-axis of the magnetic resonance k-space, respectively;

[0037] Figure 3 The results of MRF and traditional MRI were obtained by adding different concentrations of MnCl2 to glucose solution, where (ac) are T1map, T2 map and T 1ρ map, (df) are T1map, T2 map and T 1ρ In the traditional method, the saturation recovery method is used to measure the T1 map, the MSME sequence is used to measure the T2 map, and the images collected at different TSLs are combined to obtain the T 1ρmap; In a 10 mM glucose solution, the relaxation time of the solution is adjusted by applying different concentrations of MnCl2, where the role of glucose is to introduce chemical exchange components into the solution; (gi) shows the linear fitting diagram between the relaxation time measured by the traditional MRI method and the relaxation time measured by MRF, (g) is the linear fitting diagram between the T1 relaxation time measured by the traditional MRI method and the T1 relaxation time measured by MRF; (h) is the linear fitting diagram between the T2 relaxation time measured by the traditional MRI method and the T2 relaxation time measured by MRF; (i) is the T measured by the traditional MRI method 1ρ Relaxation time and T measured by MRF 1ρ Linear fitting diagram between relaxation times; the fitting equation and determination coefficient are shown in the upper left corner, where x corresponds to the relaxation time measured by the conventional method, y corresponds to the relaxation time measured by MRF, and R 2 is the coefficient of determination;

[0038] Figure 4 The results are compared between the quantitative parameter map of mouse brain MRF obtained by the present invention and the quantitative parameter map of mouse brain obtained by the traditional method. DETAILED DESCRIPTION

[0039] In order to facilitate those skilled in the art to understand and implement the present invention, the present invention is further described in detail below in conjunction with embodiments. It should be understood that the implementation examples described herein are only used to illustrate and explain the present invention and are not used to limit the present invention.

[0040] Embodiment 1:

[0041] Ultra-high field T1, T2 and T based on magnetic resonance fingerprint 1ρ Quantitative measurement method, introducing T1 weight, T2 weight and T 1ρ Weight, where T1 refers to the spin-lattice relaxation time, T2 refers to the spin-spin relaxation time, and T 1ρ refers to the spin-lattice relaxation time in the rotating coordinate system, and designs the pseudo-randomly changing sequence parameter TI (inversion recovery time), T 2prep (T2 preparation module preparation time), TSL (spin lock time) and FA (flip angle), while using constant TR (repetition time) and TE (echo time), the magnetic resonance fingerprint dictionary of different relaxation time combinations is calculated by the extended phase map method (EPG). The collected data is matched and identified with the magnetic resonance fingerprint dictionary to obtain T1, T2 and T 1ρ Multi-parameter quantitative graph. The specific steps are as follows:

[0042] Step 1, designing a magnetic resonance pulse sequence;

[0043] The magnetic resonance pulse sequence includes a randomly alternating arrangement of a T1 preparation module, a T2 preparation module and a T 1ρ Preparation module, that is, the magnetic resonance pulse sequence includes T1 preparation module, T2 preparation module and T 1ρ Preparation module, T1 preparation module, T2 preparation module and T 1ρ The order and number of times of implementation of the preparation modules are random, so that the evolution of the magnetic resonance signal contains T1 weight, T2 weight and T 1ρ Weight, T 1ρ The preparation module uses adiabatic half-pass RF pulses, and both the T1 preparation module and the T2 preparation module use adiabatic pulses. The same repetition time TR, the same echo time TE, and different flip angles FA are used in each pulse excitation. After each preparation module, a damage gradient is applied in the Z-axis direction to disperse the residual magnetization vector in the XY plane to avoid affecting subsequent imaging. After each damage gradient, the FISP (fast imaging in steady-state free precession) acquisition module is first applied. After the FISP acquisition module ends, a waiting time is applied to partially restore the magnetization vector intensity and enhance T1 encoding.

[0044] In this embodiment, T 1ρ The preparation module encoding is performed by adding a spin lock module and a change in the flip angle FA in the pulse sequence. The T1 preparation module and the T2 preparation module encoding are respectively added to the pulse sequence by adding a single and two adiabatic pulses, and adding T1 weights and T2 weights introduced by the pseudo-random flip angle FA. In this embodiment, the adiabatic pulses added by the T1 preparation module and the T2 preparation module are adiabatic refocusing pulses;

[0045] T 1ρ The preparation module starts with a 90° adiabatic half-pass RF pulse along the x-axis to flip the magnetization vector to the XY plane; subsequently, a continuous spin-locking pulse is applied along the y-axis to lock the transverse magnetization; during the spin-locking time TSL, the magnetization is adjusted according to the spin-lattice relaxation time T in the rotating coordinate system. 1ρ decay; then a second 90° adiabatic half-pass RF pulse is applied along the -x axis to return the magnetization vector to the longitudinal (i.e., z-axis) direction; T 1ρ After the preparation module, a destruction gradient is applied to break the remaining magnetization vector in the XY plane;

[0046] The T1 preparation module starts with a 180° adiabatic pulse along the x-axis to flip the magnetization vector from the z-axis to the -z-axis. After the T1 preparation module, a damaging gradient is applied to destroy the remaining magnetization vector in the XY plane.

[0047] The T2 preparation module starts with a 90° rectangular RF pulse along the x-axis to flip the magnetization vector to the XY plane; after the spin-spin relaxation time T2, two 180° adiabatic pulses are applied to reduce the influence of the inhomogeneity of the main magnetic field B0 and the RF field B1; then after the spin-spin relaxation time T2, a second 90° rectangular RF pulse is applied along the -x axis to return the magnetization vector to the longitudinal (i.e., z-axis) direction; after the T2 preparation module is completed, a damage gradient is applied to destroy the remaining magnetization vector in the XY plane.

[0048] Step 2, using the EPG method to combine the magnetic resonance pulse sequence designed in step 1 with each preparation module (i.e., T1 preparation module, T2 preparation module, and T 1ρ Preparation module), simulating the evolution process of magnetic resonance signals and calculating the magnetic resonance fingerprint dictionary;

[0049] In the process of calculating the magnetic resonance fingerprint dictionary, it is necessary to combine the evolution of the magnetic resonance signals corresponding to each preparation module. 1ρ Before and after the preparation module, the magnetic resonance signal intensity evolves according to the following formula:

[0050]

[0051] Where S0 represents the MRI signal strength before applying the preparation module, S1 represents the MRI signal strength after the T1 preparation module, S2 represents the MRI signal strength after the T2 preparation module, and S3 represents the MRI signal strength after the T 1ρ Formula 1 represents the change in magnetic resonance signal intensity before and after applying the T1 preparation module. Formula 2 represents the change in magnetic resonance signal intensity before and after applying the T2 preparation module. Formula 3 represents the change in magnetic resonance signal intensity before and after applying the T 1ρ Changes in magnetic resonance signal intensity before and after preparation of the module. T1 refers to the spin-lattice relaxation time, T2 refers to the spin-spin relaxation time, T 1ρ refers to the spin-lattice relaxation time in the rotating coordinate system, TI is the inversion recovery time, T 2prep T2 is the preparation time for the module, and TSL is the spin lock time.

[0052] Step 3: Use the magnetic resonance pulse sequence designed in step 1 to sample and obtain magnetic resonance k-space data.

[0053] Step 4: reconstruct the magnetic resonance k-space data to obtain a reconstructed image.

[0054] The image reconstruction steps are summarized as follows:

[0055] (1) Obtaining density compensation function: Calculating the density compensation function of the magnetic resonance k-space data using a spiral sampling trajectory;

[0056] (2) Obtaining a reconstructed image: The magnetic resonance k-space data is transformed into image domain data, i.e., the reconstructed image, based on the density compensation function and the non-uniform Fourier transform method.

[0057] Step 5: Match and identify the reconstructed image with the magnetic resonance fingerprint dictionary to obtain T1, T2 and T 1ρ Multi-parameter quantitative plots;

[0058] Using pattern recognition, a multi-parameter quantitative map is obtained from the reconstructed image in step 4, specifically by matching the image domain data (i.e., the reconstructed image) with the magnetic resonance fingerprint dictionary by dot product calculation to obtain T1, T2, and T 1ρ Multi-parameter quantitative plots.

[0059] This method does not need to collect multiple data separately for fitting. A single acquisition can complete multi-parameter imaging in a relatively short time, reducing scanning time and reducing the impact of discomfort and movement on the subject. In addition, this method is based on pattern matching rather than specific image reconstruction, and has good robustness to motion artifacts.

[0060] Embodiment 2:

[0061] Using the ultra-high field T1, T2 and T based on magnetic resonance fingerprint described in Example 1 1ρ Quantitative measurement method, ultra-high field T1, T2 and T 1ρ Quantitative measurement.

[0062] This embodiment was carried out on a Bruker 9.4T / 30cm magnetic resonance imager. The in vitro experimental subject was an aqueous glucose solution (10mM), to which different concentrations of MnCl2 (10-800μM) were added to adjust the relaxation time. The in vivo experimental subject was a healthy C57BL / 6 mouse weighing 35g, and brain imaging was performed. Anesthesia of the mice was induced using a mixed gas of about 3% isoflurane, and anesthesia of the mice was maintained using a mixed gas of 1-2% isoflurane. The body temperature of the mice was maintained stable using a circulating water bath device, and respiration was monitored to keep its respiration at about 45 times / minute. Sampling was performed using a birdcage coil excitation and a four-channel array coil reception method.

[0063] Magnetic resonance fingerprinting: Use Figure 2 The magnetic resonance pulse sequence shown in the figure acquires magnetic resonance k-space data. In the magnetic resonance pulse sequence, the T1 preparation module, the T2 preparation module and the T 1ρ The same repetition time TR (10.9 ms), the same echo time TE (1.6 ms) and different flip angles FA are used in each pulse excitation of the preparation module. The flip angle FA ranges from 5 degrees to 50 degrees. Figure 2(e) As shown in Figure 5, T1 preparation module, T2 preparation module and T 1ρ Prepare modules such as Figure 2 As shown in (a), the magnetic resonance signals collected at different time points include T1, T2 and T 1ρ After each FISP acquisition module, a waiting time of 610ms is applied to partially restore the magnetization vector intensity and enhance T1 encoding. Adiabatic pulses are used in all three preparation modules to increase the robustness of the magnetic resonance pulse sequence to the inhomogeneities of the main magnetic field B0 and the radio frequency field B1. After each preparation module, a damage gradient is applied along the z-axis to disperse the residual magnetization vector in the XY plane to avoid affecting subsequent imaging. The variable density spiral acquisition includes 24 spiral lines that have been compensated for the 0th order gradient, and the number of sampling points of each spiral line is 938. One spiral line is collected in each excitation, and each spiral line is rotated 15 degrees compared with the trajectory of the previous excitation. Therefore, 24 spiral lines can cover the entire k-space. Each MRF unit (i.e., the complete magnetic resonance pulse sequence) uses 1800 excitations, and each MRF unit is repeated 24 times. The interval between each repetition of the MRF unit is set to 6s to allow the spin to return to the initial state and reduce the temperature of the gradient coil. In the case of full sampling, the sequence scan time is 14.1 minutes, and in the case of 24 times undersampling, the sequence scan time is 29.5 seconds.

[0064] The gold standard of T1 was measured by the saturation recovery method, where the recovery time was 100, 300, 600, 1000, 2000, 4000, 8000 and 12000 ms. The gold standard of T2 was acquired by the MSME (multi-layer multi-echo) sequence, using 100 echo times for fitting (10-1000 ms). 1ρ The gold standard is obtained by the RARE (rapid acquisition with relaxation enhancement) sequence, which is obtained by adding T 1ρ Prepare the module and fit it using 8 different TSL values ​​(5, 10, 20, 40, 60, 80, 120 and 200 ms).

[0065] The magnetic resonance fingerprint dictionary is based on Figure 2 The MRI pulse sequence timing diagram in the above method is obtained by combining the MRI signal evolution process brought by each preparation module with the EPG method and the MRI pulse sequence calculation. The 1800 reconstructed images at different time points acquired and reconstructed by the above method are matched with the MRI fingerprint dictionary using the dot product method, and the multi-parameter quantitative maps of the water model and the mouse brain are calculated respectively (the multi-parameter quantitative maps include T1map, T2map and T 1ρ map). The simulation water model results are as follows Figure 3As shown, the relaxation time measured by MRF has a good linear correlation with the gold standard. Figure 4 The results of in vivo experiments are shown. The T1map, T2map and T2map of the mouse brain were obtained by conventional methods and the present invention. 1ρ map parameter maps, there is no significant difference between them. The results show that this method can obtain better multi-parameter maps.

[0066] It is worth noting that the specific embodiments described are only examples of the present invention. Personnel in the technical field of the present invention may make various details supplement, modify or replace the described embodiments without departing from the spirit of the present invention or exceeding the scope defined by the appended claims.

Claims

1. Based on magnetic resonance fingerprint ultra-high field T1, T2 and T 1ρ A quantitative measurement method, characterized in that The steps include: Step 1, designing a magnetic resonance pulse sequence, which includes randomly alternating T1 preparation modules, T2 preparation modules and T 1ρ Preparation module, where T 1ρ The preparation module uses adiabatic half-pass RF pulses, and both the T2 preparation module and the T1 preparation module use adiabatic pulses; Step 2, using the extended phase map (EPG) method combined with the magnetic resonance pulse sequence designed in step 1 to simulate the magnetic resonance signal evolution process and calculate the magnetic resonance fingerprint dictionary; Step 3, using the magnetic resonance pulse sequence designed in step 1, sampling to obtain magnetic resonance k-space data; Step 4, reconstructing the magnetic resonance k-space data to obtain a reconstructed image; Step 5: Match and identify the reconstructed image with the magnetic resonance fingerprint dictionary to obtain T1, T2 and T 1ρ Multi-parameter quantitative diagram, where T1 refers to the spin-lattice relaxation time, T2 refers to the spin-spin relaxation time, and T 1ρ Refers to the spin-lattice relaxation time in the rotating coordinate system; In step 1: T 1ρ The preparation module starts with a 90° adiabatic half-pass RF pulse along the x-axis to flip the magnetization vector to the XY plane; Subsequently, continuous spin-locking pulses are applied along the y-axis to lock the transverse magnetization; During the spin locking time TSL, the magnetization intensity changes according to the spin-lattice relaxation time T in the rotating coordinate system. 1ρ decay; then applying a second 90° adiabatic half-pass RF pulse along the -x axis will return the magnetization vector to the longitudinal direction; The T1 preparation module starts with a 180° adiabatic pulse along the x-axis, flipping the magnetization vector from the z-axis to the -z-axis; The T2 preparation module starts with a 90° rectangular RF pulse along the x-axis to flip the magnetization vector to the XY plane; after the spin-spin relaxation time T2, two 180° adiabatic pulses are applied to reduce the influence of the inhomogeneity of the main magnetic field B0 and the RF field B1; then after the spin-spin relaxation time T2, a second 90° rectangular RF pulse is applied along the -x axis to return the magnetization vector to the longitudinal direction.

2. The ultra-high field T1, T2 and T based on magnetic resonance fingerprint according to claim 1 1ρ A quantitative measurement method, characterized in that The T1 preparation module, T2 preparation module and T 1ρ After the preparation module is completed, a damage gradient is applied in the Z-axis direction.

3. The ultra-high field T1, T2 and T based on magnetic resonance fingerprint according to claim 2 1ρ A quantitative measurement method, characterized in that A FISP acquisition module is first applied after each of the damage gradients, and a waiting time is applied after the FISP acquisition module ends.

4. The ultra-high field T1, T2 and T based on magnetic resonance fingerprint according to claim 3 1ρ A quantitative measurement method, characterized in that In the T1 preparation module, T2 preparation module and T 1ρ The same repetition time TR, the same echo time TE and different flip angles FA are used in each pulse excitation of the preparation module.

5. The ultra-high field T1, T2 and T based on magnetic resonance fingerprint according to claim 4 1ρ A quantitative measurement method, characterized in that In step 2, the magnetic resonance pulse sequence designed in step 1 is combined with the T1 preparation module, the T2 preparation module and the T 1ρ Preparation module, the simulated magnetic resonance signal evolution is as follows: Where S0 represents the MRI signal strength before applying the preparation module, S1 represents the MRI signal strength after the T1 preparation module, S2 represents the MRI signal strength after the T2 preparation module, and S3 represents the MRI signal strength after the T 1ρ The magnetic resonance signal intensity after preparing the module; T1 refers to the spin-lattice relaxation time, T2 refers to the spin-spin relaxation time, T 1ρ refers to the spin-lattice relaxation time in the rotating coordinate system, TI is the inversion recovery time, T 2prep T2 is the preparation time for the module, and TSL is the spin lock time.

6. The ultra-high field T1, T2 and T based on magnetic resonance fingerprint according to claim 1 1ρ A quantitative measurement method, characterized in that The specific steps of reconstructing the magnetic resonance k-space data in step 4 to obtain the reconstructed image are as follows: The density compensation function of the magnetic resonance k-space data is calculated using a spiral sampling trajectory; The magnetic resonance k-space data are transformed into a reconstructed image according to a density compensation function and a non-uniform Fourier transform method.

7. The ultra-high field T1, T2 and T based on magnetic resonance fingerprint according to claim 1 1ρ A quantitative measurement method, characterized in that The specific method for matching and identifying the reconstructed image with the magnetic resonance fingerprint dictionary in step 5 is: performing dot product calculation on the reconstructed image and the magnetic resonance fingerprint dictionary for matching.

Images