Methods of water-lipid separation in cardiac magnetic resonance imaging
Through the Dixon two-point method and regional growth method triggered by electrocardioglycatch, the problem of limited application of water-lipid separation method in cardiac imaging is solved, high-quality dynamic cardiac water-lipid separation images are achieved, and the diagnostic effectiveness of heart disease is improved .
Patent Information
- Application Number
- CN202410488868.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-23
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-04-23
AI Technical Summary
The existing Dixon water-lipid separation method is limited in cardiac imaging, especially due to the influence of heartbeat and respiratory movement, resulting in low image quality and low diagnostic efficiency.
The Dixon two-point method, which is triggered by electrocardioglycensorship, collects echo images at two echo times by triggering excitation at the end of the cardiac diastolic period, and calculates the phase vector through image reconstruction and region growth methods to separate pure water and pure Image of fat.
It reduces the impact of cardiac on image quality, shortens acquisition time, improves image quality and water-lipid separation effect, and enhances the diagnostic efficacy of heart disease.
Smart Images

Figure CN118078249B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of magnetic resonance imaging, and more specifically, to a method for water-fat separation in cardiac magnetic resonance imaging. Background Art
[0002] Magnetic resonance imaging (MRI) is widely used in the precise diagnosis and treatment of diseases throughout the body because of its good soft tissue contrast and no ionizing radiation. MRI used in clinical examinations is to image the hydrogen protons in water molecules. Fat is the main interfering substance in MRI, because fat molecules have a resonance frequency similar to that of water molecules, but different T1 and T2 relaxation times can easily produce various artifacts on MRI images, reduce image quality, and thus affect disease diagnosis. Therefore, in MRI, the signal of fat is often suppressed. However, for heart-related diseases, fat is closely related to the differential diagnosis of many diseases. For example, fatty infiltration of the myocardium increases the risk of sudden death, which, if confirmed in histological examination, is a major criterion for the diagnosis of arrhythmic right ventricular cardiomyopathy (ARVC). Therefore, using magnetic resonance technology to distinguish the signal of fat from the signal of water in conventional tissues can help the precise diagnosis and treatment of heart-related diseases.
[0003] The Dixon method is a commonly used water-fat separation method in magnetic resonance imaging. The method was first proposed by Dixon in 1984. Its basic principle is to use magnetic resonance images with different echo times, and utilize the difference in resonance frequency caused by the chemical shift of protons in water and fat in the magnetic field to collect echo complex images of the two substances in the same phase (In phase) and opposite phase (Out phase), and then perform water-fat separation through image reconstruction methods to calculate the images of pure water and pure fat respectively.
[0004] The conventional Dixon water-fat separation method is mostly used for static organs such as musculoskeletal joints due to its relatively long acquisition time. For organs affected by respiratory movement, such as the liver, scanning and imaging are usually performed by holding the breath once. Because it is affected by both respiratory movement and heart beats, cardiac imaging is one of the most difficult organs to image with magnetic resonance imaging. Therefore, the application of the Dixon water-fat separation method on the heart needs to consider the effects of both breathing and heartbeat on imaging, and its overall application is relatively limited. Previously, some studies used the Dixon three-point method to achieve water-fat separation imaging of the heart. However, the Dixon three-point method requires the acquisition of corresponding images for three different echo times, and then uses the corresponding method to correct the corresponding phase distribution, and finally reconstructs images of pure water and pure fat. For cardiac imaging, it takes a long time to collect three echo images, so the subject has to hold his breath for a long time, which is a heavy burden for patients with heart disease and has a low success rate. Cardiac cine needs to collect cardiac images of the entire cardiac cycle. If the subject has an irregular heart rate, the imaging success rate is not high. Cine imaging uses a steady-state free precession (bSSFP) imaging sequence for data acquisition, which is prone to black band artifacts, reducing image quality and image disease diagnosis. Summary of the invention
[0005] In view of this, the present application aims to provide a new Dixon two-point cardiac magnetic resonance water-fat separation imaging method based on ECG gating triggering.
[0006] According to one aspect of the present application, a method for water-fat separation in cardiac magnetic resonance imaging is proposed, the method comprising:
[0007] Using ECG gating to trigger excitation at the end of diastole of a first cardiac cycle, and acquiring a first echo image at a first preset echo time, wherein the first preset echo time is the time when the magnetic resonance signals of water and fat evolve into anti-phase;
[0008] Using ECG gating to trigger excitation at the end of diastole of a second cardiac cycle, and acquiring a second echo image at a second preset echo time, wherein the second preset echo time is the time when the magnetic resonance signals of water and fat evolve into the same phase;
[0009] Reconstructing the first echo image and the second echo image to obtain a first complex image and a second complex image;
[0010] Based on the first complex image and the second complex image, a water-fat difference phase modulation signal of each voxel is obtained:
[0011] ,
[0012] in, represents the water-fat difference phase modulation signal, represents the corresponding voxel in the first complex image, representing a background phase in the second complex image;
[0013] Based on the phase information of the water-fat difference phase modulation signal, the phase vector calculation formula below is determined by the region growing method. The sign on the right side of the equal sign is positive or negative to calculate the phase vector of each voxel in the entire image. :
[0014] ;
[0015] According to the phase vector , separating a pure water image and a fat image from the first complex image and the second complex image.
[0016] In some embodiments, the first preset echo time is the time when the magnetic resonance signals of water and fat first evolve into anti-phase after the triggered excitation at the end of diastole of the first cardiac cycle, and the second preset echo time is the time when the magnetic resonance signals of water and fat first evolve into in-phase after the triggered excitation at the end of diastole of the second cardiac cycle.
[0017] In some embodiments, the first echo image and the second echo image are acquired using a fast radiofrequency spoiled gradient echo (fSPGR) sequence.
[0018] In some embodiments, based on the phase information of the water-fat difference phase modulation signal, the phase vector calculation formula below is determined by region growing method. The sign on the right side of the equal sign is positive or negative to calculate the phase vector of each voxel in the entire image. ,include:
[0019] Step 51, setting a traversal label Vst and a confirmation label Chk for each voxel, Vst and Chk of all voxels are initially set to 0, and creating multiple stacks, different stacks corresponding to different phase differences, and numbering the multiple stacks in ascending order according to the corresponding phase differences;
[0020] Step 52, select a voxel as the starting point, put it into the stack with the smallest number, and set the traversal label Vst of the voxel to 1, and the confirmation label Chk to 1;
[0021] Step 53, taking out the voxel that first enters the stack from the non-empty stack with the smallest number as the seed voxel;
[0022] Step 54, using the following formula to calculate the phase difference between the water-fat difference phase modulation signal of the seed voxel and the water-fat difference phase modulation signal of each voxel in the adjacent voxels of the seed voxel whose label Vst is 0 :
[0023] ,
[0024] in, represents the water-fat difference phase modulation signal of the seed voxel, represents the water-fat difference phase modulation signal of the voxels with label Vst of 0 in the neighboring voxels of the seed voxel, and the superscript * represents the complex conjugate. Express request and The phase of the product of Express request The absolute value of
[0025] According to the phase difference, each voxel whose traversal label Vst is 0 among the adjacent voxels of the seed voxel is placed into a corresponding stack, and the traversal label Vst of the voxel placed in the stack is set to 1;
[0026] Step 55: compare the sum of the water-fat difference phase modulation signals of all voxels with a confirmation label Chk of 1 in a preset area centered on the seed voxel. The angle difference between the phase modulation signal of the water-fat difference of the seed voxel D , and in When the phase vector calculation formula is determined, the sign on the right side of the equal sign is positive. When the phase vector calculation formula is negative, the phase vector of the seed voxel is calculated. , and setting Chk of the seed voxel to 1;
[0027] Step 56, determine whether there is a non-empty stack, if so, return to step 53, if not, end the region growing.
[0028] In some embodiments, the plurality of stacks uniformly cover the entire phase difference range.
[0029] In some implementations, in step 54, the neighboring voxels of the seed voxel include four voxels that are adjacent to the seed voxel in four directions: above, below, left, and right.
[0030] In some embodiments, the preset area centered on the seed voxel is a square area centered on the seed voxel.
[0031] In some embodiments, according to the phase vector , separating a pure water image and a fat image from the first complex image and the second complex image, comprising:
[0032] Calculated based on the following formula (WF) :
[0033] ,
[0034] in, An image representing pure water, An image representing fat, represents the corresponding voxel in the first complex image, representing a background phase in the second complex image;
[0035] Calculated based on the following formula (W+F) :
[0036] ,
[0037] in, represents a corresponding voxel in the second complex image;
[0038] According to the calculated (WF) and (W+F) , and get the image of pure water and fat images F .
[0039] The various aspects of the technical solution of this application have the following beneficial effects:
[0040] 1. ECG gating trigger is used to enable image acquisition at a relatively fixed cardiac phase time point, reducing the impact of cardiac movement on image quality and obtaining clear and stable cardiac MRI images;
[0041] 2. The Dixon two-point method, combined with image reconstruction, only requires two echo times, which shortens the acquisition time by 1 / 3 compared with the Dixon three-point method. Stable cardiac water-fat separation results can still be obtained under the condition of reducing the patient's breath-holding cooperation;
[0042] 3. Acquiring images only at the end of diastole, rather than during the entire cardiac cycle, can reduce the impact of irregular heartbeat on imaging quality.
[0043] 4. Use the fast radiofrequency spoiled gradient echo (fSPGR) sequence to acquire images, avoiding the black band artifact problem of the bSSFP sequence and improving image quality;
[0044] 5. The phase vector is extracted by the region growing method, which achieves a smooth transition between adjacent voxels and makes the water-fat separation effect more natural and reliable.
[0045] In summary, according to the technical solution of the present application, high-quality dynamic cardiac water-fat separation images can be quickly obtained within two cardiac cycles, thereby improving the detection rate and diagnostic efficiency of heart diseases.
[0046] The methods and devices of the present application have other characteristics and advantages, which will be apparent from the drawings and subsequent specific embodiments incorporated herein, or will be described in detail in the drawings and subsequent specific embodiments incorporated herein, which together are used to explain the specific principles of the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] The above and other objects, features and advantages of the present application will become more apparent through a more detailed description of exemplary embodiments of the present application in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the present application.
[0048] Figure 1 A flow chart of a method for water-fat separation in cardiac magnetic resonance imaging according to an embodiment of the present application is shown.
[0049] Figure 2 A schematic diagram of image acquisition of cardiac magnetic resonance water-fat separation imaging using ECG gating triggering according to an embodiment of the present application is shown.
[0050] Figure 3 A flow chart of calculating the phase vector of the entire image by using a region growing method according to an exemplary embodiment of the present application is shown.
[0051] FIG. 4 ( a ) to ( d ) show cardiac magnetic resonance water-fat separation imaging result images according to an exemplary embodiment of the present application.
[0052] FIG5 (a) to (d) show cardiac magnetic resonance water-fat separation imaging result images according to another exemplary embodiment of the present application. DETAILED DESCRIPTION
[0053] The preferred embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the preferred embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.
[0054] Figure 1 A flow chart of a method for water-fat separation in cardiac magnetic resonance imaging according to an embodiment of the present application is shown. As shown in the figure, the method includes steps 1 to 6.
[0055] Step 1, triggering excitation at the end of diastole of the first cardiac cycle by ECG gating, and acquiring a first echo image at a first preset echo time, wherein the first preset echo time is the time when the magnetic resonance signals of water and fat evolve into anti-phase.
[0056] Step 2, triggering excitation at the end of diastole of the second cardiac cycle by ECG gating, and acquiring a second echo image at a second preset echo time, wherein the second preset echo time is the time when the magnetic resonance signals of water and fat evolve into the same phase.
[0057] The present application adopts electrocardiogram gating (ECG gating) to utilize electrocardiogram signals to accurately control the triggering of magnetic resonance imaging sequences, thereby obtaining cardiac images in a relatively fixed cardiac phase state for diagnosis and analysis.
[0058] According to some embodiments of the present application, the echo image may be acquired by the following steps:
[0059] Before the MRI (Magnetic Resonance Imaging) scan begins, electrocardiogram electrodes are attached to the subjects to collect electrocardiogram (ECG) signals;
[0060] The acquired ECG signals are input into the physiology unit of the MRI scanner;
[0061] After the physiological unit identifies a specific cardiac phase (end diastole in this application), the scanning unit in the MRI scanner emits a radio frequency excitation of the MRI sequence and acquires an echo image after a set echo time;
[0062] This process is repeated for two consecutive cardiac cycles to obtain the first echo image and the second echo image, respectively.
[0063] It should be emphasized that the present application is not limited to the first cardiac cycle being in front and the second cardiac cycle being in the back in terms of time. The terms "first" and "second" are used only to distinguish two different cardiac cycles for collecting water-fat inverse phase echo images and water-fat in-phase echo images. It is possible to collect a water-fat inverse phase echo image (i.e., the first echo image) at the end of diastole of the previous cardiac cycle and collect a water-fat in-phase echo image (i.e., the second echo image) at the end of diastole of the next cardiac cycle. It is also possible to collect a water-fat in-phase echo image (i.e., the second echo image) at the end of diastole of the previous cardiac cycle and collect a water-fat in-phase echo image (i.e., the first echo image) at the end of diastole of the next cardiac cycle.
[0064] In some embodiments, the first preset echo time is the time when the magnetic resonance signals of water and fat first evolve into anti-phase after triggering excitation at the end of diastole of the first cardiac cycle, and the second preset echo time is the time when the magnetic resonance signals of water and fat first evolve into in-phase after triggering excitation at the end of diastole of the second cardiac cycle. By selecting the first anti-phase state point and the first in-phase state point of two consecutive cycles respectively, a water-fat signal with good quality, clear phase evolution and stable state can be obtained.
[0065] In a specific example, the two echo times are respectively 1.1 milliseconds (ms) for the first anti-phase time of water and fat under the 3T main magnetic field (i.e., the first preset echo time, also called TE1) and 2.2 milliseconds (ms) for the next phase time (i.e., the second preset echo time, also called TE2). Specifically, parallel imaging technology can be used to increase the imaging speed during the actual acquisition process.
[0066] Through ECG gating, the impact of cardiac motion on image quality can be greatly reduced, and clear cardiac MRI images can be obtained. Moreover, by triggering excitation at the end of diastole, images of myocardial tissue in a relatively static state can be obtained, thereby improving the water-lipid separation effect.
[0067] Figure 2 A schematic diagram of image acquisition using ECG-gated triggering of cardiac magnetic resonance water-fat separation imaging according to an embodiment of the present application is shown. The horizontal curve in the figure is the ECG signal, and the peak located by the vertical dotted line indicating the first excitation triggering moment is the R wave. The acquisition window of the end diastole can be determined based on the identified R wave. The fence-like vertical line sequence in the lower part of the figure is used to represent data acquisition.
[0068] In some embodiments, the first echo image and the second echo image may be acquired using a fast radiofrequency spoiled gradient echo (fSPGR) sequence. The use of fSPGR can effectively overcome the black band artifacts caused by bSSFP and significantly improve the quality of acquired images.
[0069] Step 3: reconstruct the first echo image and the second echo image to obtain a first complex image and a second complex image.
[0070] Through image reconstruction, the collected k-space echo raw data is converted into the final complex image. In some embodiments, FFT can be used for image reconstruction, which has the characteristics of small calculation amount and high efficiency, making the reconstruction of cardiac images faster and smoother, meeting the real-time requirements of dynamic cardiac imaging.
[0071] Through fast and stable image reconstruction, the water-fat separation effect is guaranteed, and the overall reconstruction time is optimized to meet the acquisition requirements under complex ECG signal coordinates.
[0072] Step 4, obtaining a water-fat difference phase modulation signal of each voxel based on the first complex image and the second complex image:
[0073] ,
[0074] in, represents the water-fat difference phase modulation signal, represents the corresponding voxel in the first complex image, represents the background phase in the second complex image.
[0075] Step 5: Based on the phase information of the water-fat difference phase modulation signal, the phase vector calculation formula below is determined by the region growing method. The sign on the right side of the equal sign is positive or negative to calculate the phase vector of each voxel in the entire image: :
[0076] .
[0077] According to the inventors' in-depth research, the choice of positive and negative signs in the above formula is determined by the contribution of water and fat in the imaging voxel to the voxel. If for a given voxel, the contribution of water signal is large, then the positive sign is taken, that is, the phase vector The phase modulation signal of the water-fat difference of the corresponding voxel In phase; if the contribution of fat is large, it takes a negative sign, that is, the phase vector The phase modulation signal of the water-fat difference of the corresponding voxel The inventor believes that one of the keys to solving the water-fat separation problem is how to obtain a high-quality, smooth spatial phase vector. , that is, properly determine the positive and negative signs in the above formula.
[0078] In this application, the region growing method is used to determine the positive and negative signs in the above formula. The region growing method is a pixel-based image segmentation method. Its basic idea is to use one or several seed pixels as the starting point, and gradually gather adjacent pixels with similar properties into the region to which the seed point belongs according to a predetermined growth rule, thereby achieving image segmentation. This application calculates the smooth spatial phase vector of the entire image through the region growing method, and determines whether it belongs to the same tissue type based on the phase information between the specific signals of the voxels, so that the calculation of the phase vector is smoother and continuous, which significantly improves the effect of water-fat separation, making it more accurate and natural.
[0079] Step 6: Separate the pure water image and the fat image from the first complex image and the second complex image according to the phase vector.
[0080] According to some embodiments, the pure water image and the fat image may be separated by the following steps:
[0081] Calculated based on the following formula (WF) :
[0082] ,
[0083] in, An image representing pure water, An image representing fat, represents the corresponding voxel in the first complex image, representing a background phase in the second complex image;
[0084] Calculated based on the following formula (W+F) :
[0085] ,
[0086] in, represents a corresponding voxel in the second complex image;
[0087] According to the calculated (WF) and (W+F) , and get the image of pure water and fat images F .
[0088] Background Phase It is usually a phase error caused by systemic reasons such as system RF and receiving coil during water-fat phase imaging.
[0089] The concept and design principle of the technical solution of this application are further explained and introduced below.
[0090] Through the above steps 1 to 3, the complex images of the water-fat in-phase and anti-phase echo times of the single-layer heart image are collected and reconstructed. For a voxel with a spatial coordinate position of (x, y), considering the influence of water and fat components at the same time, its magnetic resonance signal can be It is expressed as follows:
[0091] , [1]
[0092] In formula [1] W and F Respectively represent the contribution of water and fat in the voxel to the magnetic resonance signal, represents the phase accumulation due to the chemical shift of fat relative to water, It represents the phase error caused by systemic reasons such as system RF and receiving coil, while It represents the additional phase accumulated due to the inhomogeneity of the main magnetic field in the corresponding echo time.
[0093] For the water-lipid phase involved in this application ( ) and water-lipid inversion phase ( ) image, the complex images of the two echoes are derived from formula [1] as follows:
[0094] , [2]
[0095] , [3]
[0096] The subscript ip stands for in-phase, and the subscript op stands for out-of-phase.
[0097] in:
[0098] , [4]
[0099] It indicates the phase error caused by all factors when water and fat are in phase, also called background phase;
[0100] , [5]
[0101] It represents the phase accumulated due to the main magnetic field inhomogeneity during the echo time of two images.
[0102] The goal of this application is to obtain images of pure water and pure fat (i.e. W Images and F image), so it is necessary to eliminate the phase from equations [2] and [3] and The relevant terms of . From formula [2], we can get:
[0103] [6]
[0104] The background phase caused by system factors can be calculated when water and fat are in phase imaging The relevant terms of are as follows, which can be obtained from formula [6]:
[0105] , [7]
[0106] This eliminates the phase Next, focus on Related items.
[0107] The following expression can be derived from formula [3]:
[0108] , [8]
[0109] Can The water-fat difference phase modulation signal of the voxel is called.
[0110] In this way, the phase caused by the inhomogeneity of the main magnetic field can be obtained. The calculation formula of the related terms is:
[0111] [9]
[0112] At this point, the problem of separating the images is reduced to the problem of determining the sign on the right side of the equal sign of formula [9] (called the phase vector calculation formula in this scheme). As long as the positive and negative signs are determined, we can get .like Figure 1 In the embodiment shown, according to the present application, the phase information of the water-fat difference phase modulation signal is subjected to regional growth to determine the positive and negative signs on the right side of the equal sign of the phase vector calculation formula, thereby determining the phase vector of each voxel in the entire image. According to the above analysis, after eliminating the phase and After the relevant items are obtained, the image separation of pure water and fat can be achieved. The above is the principle analysis and design ideas of the technical solution proposed in this application.
[0113] According to the present embodiment, the Dixon two-point method is adopted to collect images at a fixed time period at the end of diastole, and combined with image reconstruction, it is possible to obtain stable cardiac water-fat separation results in a time that is 1 / 3 shorter than that of the Dixon three-point method. In addition, since images are collected only at the end of diastole instead of the entire cardiac cycle, the impact of arrhythmia on imaging quality can be reduced, which is beneficial to improving the detection rate and diagnostic efficiency of myocardial diseases.
[0114] Figure 3 A flow chart of calculating the phase vector of each voxel in the entire image by using a region growing method according to an exemplary embodiment of the present application is shown.
[0115] Step 51, set a traversal label Vst and a confirmation label Chk for each voxel, Vst and Chk of all voxels are initially set to 0, and create multiple stacks, different stacks correspond to different phase differences, and the multiple stacks are numbered in ascending order according to the corresponding phase differences.
[0116] Multiple stacks can evenly cover the entire phase difference range. In an example, the phase difference range is 0 to 90 degrees (see the related analysis below), and 9 empty stacks can be created with 10 degrees as one stack to evenly cover the entire phase difference range. A phase difference of 0 to 10 degrees corresponds to stack numbered 1, a phase difference of 10 to 20 degrees corresponds to stack numbered 2, ..., a phase difference of 80 to 90 degrees corresponds to stack numbered 9.
[0117] Step 52 , select a voxel as the starting point, put it into the stack with the smallest number, and set the traversal label Vst of the voxel to 1, and the confirmation label Chk to 1.
[0118] A voxel can be randomly selected as the starting point.
[0119] Step 53: Take out the voxel that first enters the stack from the non-empty stack with the smallest number as the seed voxel.
[0120] Step 54, using the following formula to calculate the phase difference between the water-fat difference phase modulation signal of the seed voxel and the water-fat difference phase modulation signal of each voxel in the adjacent voxels of the seed voxel whose label Vst is 0 :
[0121] ,
[10]
[0122] in, represents the water-fat difference phase modulation signal of the seed voxel, represents the water-fat difference phase modulation signal of the voxels with label Vst of 0 in the neighboring voxels of the seed voxel, and the superscript * represents the complex conjugate. Express request and The phase of the product of Express request The absolute value of the phase difference Each voxel whose traversal tag Vst is 0 among the adjacent voxels of the seed voxel is placed in a corresponding stack, and the traversal tag Vst of the voxel placed in the stack is set to 1.
[0123] It is easy to understand and can be determined without doubt by those skilled in the art that and , is a universal symbol for the water-fat difference phase modulation signal The subscript of adds distinguishing information to indicate the phase modulation signal of the water-fat difference of different voxels, making the relevant formula more accurate, clear and concise, rather than indicating that it is different from other signals.
[0124] In some examples, the neighboring voxels of the seed voxel include four voxels adjacent to the seed voxel in four directions: above, below, left, and right.
[0125] The result calculated by formula
[10] G It represents the absolute value of the phase difference between the water-fat difference phase modulation signal of the seed voxel and the water-fat difference phase modulation signal of the adjacent voxel, and its value is in the range of [0, , interval.
[0126] In the water-dominated region (i.e. W>F), the phase vector calculated by formula [9] is and are parallel, then according to formula
[10] we can get:
[0127] ;
[11]
[0128] In the fat-dominated region (i.e., W < F), the phase vector calculated by formula [9] and are in the reverse direction. Then, according to formula
[10] , we can get:
[0129] ;
[12]
[0130] If the region is at the water-fat boundary, assuming the seed voxel is in the water-dominated region and the adjacent voxel is in the fat-dominated region, then according to formula
[10] , we can get:
[0131] ,
[13]
[0132] Since G is in [0, , we can obtain
[0133] .
[14]
[0134] From formulas
[11] to
[13] , it can be obtained that when the seed voxel is in the region where either water or fat dominates, the phase difference between the seed voxel and the adjacent voxel is G ; when the seed voxel is in the water-fat boundary region, the phase difference between the seed voxel and the adjacent voxel is . Considering that the desired phase vector is spatially smooth, the following processing is performed on the phase difference G : when G > , G takes the value of . This can make the value of G fall within the interval [0, ], and use fewer stack numbers to complete region growing.
[0135] For the selected seed voxel, calculate the phase difference G of the water-fat difference phase modulation signal between it and the voxels with traversal label Vst = 0 among the four adjacent voxels above, below, left, and right. According to the phase difference G , divide the voxels with traversal label Vst = 0 among its four adjacent voxels above, below, left, and right into the corresponding stacks, and set their traversal label Vst to 1. For example, assume that the traversal label Vst of the adjacent voxel above the seed voxel is 0, and calculate the phase difference GIf the angle is 25 degrees, then the adjacent voxel above it is placed in the stack numbered 3, corresponding to 20 degrees to 30 degrees, and the traversal label Vst of the adjacent voxel is set to 1. For example, the traversal label Vst of the adjacent voxel above the seed voxel is set to 0, and the phase difference between the water-fat difference phase modulation signal of the seed voxel and the water-fat difference phase modulation signal of the adjacent voxel above it is calculated. G If it is 25 degrees, the adjacent voxel above it is placed in the stack numbered 3, corresponding to 20 degrees to 30 degrees, and the traversal label Vst of the adjacent voxel is set to 1.
[0136] Step 55: compare the sum of the water-fat difference phase modulation signals of all voxels with a confirmation label Chk of 1 in a preset area centered on the seed voxel. The angle difference between the phase modulation signal of the water-fat difference of the seed voxel D , and in When the phase vector calculation formula is determined, the sign on the right side of the equal sign is positive. When the phase vector calculation formula is negative, the phase vector of the seed voxel is calculated. , and set the Chk of the seed voxel to 1.
[0137] In a possible implementation manner, the preset area centered on the seed voxel is a square area centered on the seed voxel, such as a 7×7 square area.
[0138] Step 56, determine whether there is a non-empty stack, if so, return to step 53, if not, end the region growing.
[0139] After the above region growing method, the phase vector of the entire image can be obtained , which is the phase vector diagram.
[0140] As mentioned above, when the phase vector is known, we can calculate it by formula [8]: , combined with the value calculated by formula [6] , thus the W graph of pure water and the F graph of pure fat can be separated.
[0141] Figures 4 (a) to (d) show cardiac magnetic resonance water-fat separation imaging result images according to an exemplary embodiment of the present application, and Figures 5 (a) to (d) show cardiac magnetic resonance water-fat separation imaging result images according to another exemplary embodiment of the present application. Among them, Figures 4 (a) and 5 (a) are water-fat anti-phase images (i.e., the first complex image), Figures 4 (b) and 5 (b) are water-fat in-phase images (i.e., the second complex image), Figures 4 (c) and 5 (c) are separated pure water images, and Figures 4 (d) and 5 (d) are separated fat images.
[0142] As can be seen from the figures, the water-fat anti-phase image and the water-fat in-phase image show clear myocardial contours, with few obvious respiratory motion artifacts; the pure water image and fat image have a high signal-to-noise ratio, good image contrast between different tissues, and consistent phase information, and the water-fat boundary is clear and natural.
[0143] According to the above application examples, it can be seen that the technical solution according to the present application has excellent performance in terms of image quality, water-fat separation effect, etc.
[0144] It can be understood that the above-mentioned embodiments and implementation methods mentioned in this application can be combined with each other to form combined embodiments and implementation methods without violating the principle logic. Due to space limitations, this application will not go into details. Those skilled in the art can understand that in the above-mentioned method of the specific implementation method, the specific execution order of each step should be determined by its function and possible internal logic.
[0145] Note that, unless otherwise directly stated, all features disclosed in this specification (including any attached claims, abstracts and drawings) may be replaced by alternative features for achieving the same, equivalent or similar purposes. Therefore, unless otherwise explicitly stated, each feature disclosed is only an example of a group of equivalent or similar features. Where used, further, preferably, further and more preferably are simple beginnings for elaborating another embodiment based on the aforementioned embodiment, and the content of further, preferably, further or more preferably followed by the aforementioned embodiment is combined with the complete composition of another embodiment. Several further, preferably, further or more preferably settings following the same embodiment can be arbitrarily combined to form another embodiment.
[0146] It should be understood by those skilled in the art that the embodiments of the present application described above and shown in the accompanying drawings are only examples and do not limit the present application. The purpose of the present application has been fully and effectively achieved. The functions and structural principles of the present application have been demonstrated and explained in the embodiments, and the embodiments of the present application may be deformed or modified in any way without departing from the principles.
[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.
Claims
1. A method for water-fat separation in cardiac magnetic resonance imaging, characterized in that: The method comprises: Triggering excitation at the end of diastole of the first cardiac cycle by using an electrocardiogram gating method, and acquiring a first echo image at a first preset echo time, wherein the first preset echo time is the time when the magnetic resonance signals of water and fat first evolve into anti-phase after triggering excitation at the end of diastole of the first cardiac cycle; triggering the excitation at the end of diastole of the second cardiac cycle by using an electrocardiogram gating method, and acquiring a second echo image at a second preset echo time, wherein the second preset echo time is the time when the magnetic resonance signals of water and fat first evolve into the same phase after the excitation is triggered at the end of diastole of the second cardiac cycle; Reconstructing the first echo image and the second echo image to obtain a first complex image and a second complex image; Based on the first complex image and the second complex image, a water-fat difference phase modulation signal of each voxel is obtained: Among them, S1 represents the water-fat difference phase modulation signal, S op represents the corresponding voxel in the first complex image, representing a background phase in the second complex image; Based on the phase information of the water-fat difference phase modulation signal, the phase vector calculation formula below is determined by the region growing method. The sign on the right side of the equal sign is positive or negative to calculate the phase vector of each voxel in the entire image. According to the phase vector separating a pure water image and a fat image from the first complex images and the second complex images; in, Acquiring the first echo image and the second echo image using a fast radiofrequency spoiled gradient echo (fSPGR) sequence; According to the phase vector Separating a pure water image and a fat image from the first complex image and the second complex image, comprising: (WF) is calculated based on the following formula: Among them, W represents the image of pure water, F represents the image of fat, and S op represents the corresponding voxel in the first complex image, representing a background phase in the second complex image; (W+F) is calculated based on the following formula: Among them, S ip represents a corresponding voxel in the second complex image; Based on the calculated (WF) and (W+F), the pure water image W and the fat image F are obtained.
2. The method according to claim 1, characterized in that Based on the phase information of the water-fat difference phase modulation signal, the phase vector calculation formula below is determined by the region growing method. The sign on the right side of the equal sign is positive or negative to calculate the phase vector of each voxel in the entire image. include: Step 51, setting a traversal label Vst and a confirmation label Chk for each voxel, Vst and Chk of all voxels are initially set to 0, and creating multiple stacks, different stacks corresponding to different phase differences, and numbering the multiple stacks in ascending order according to the corresponding phase differences; Step 52, select a voxel as the starting point, put it into the stack with the smallest number, and set the traversal label Vst of the voxel to 1, and the confirmation label Chk to 1; Step 53, taking out the voxel that first enters the stack from the non-empty stack with the smallest number as the seed voxel; Step 54, using the following formula to calculate the phase difference G between the water-fat difference phase modulation signal of the seed voxel and the water-fat difference phase modulation signals of each voxel whose label Vst is 0 in the adjacent voxels of the seed voxel: Among them, S 1,s represents the water-fat difference phase modulation signal of the seed voxel, S 1,n represents the water-fat difference phase modulation signal of the voxels with label Vst of 0 in the neighboring voxels of the seed voxel, and the superscript * represents the complex conjugate. Indicates the request S 1,s and The phase of the product of Express request The absolute value of According to the phase difference G, each voxel whose traversal label Vst is 0 among the adjacent voxels of the seed voxel is placed into a corresponding stack, and the traversal label Vst of the voxel placed in the stack is set to 1; Step 55, compare the sum S of the water-fat difference phase modulation signals of all voxels with a confirmed label Chk of 1 in a preset area centered on the seed voxel. 1e The angle difference D between the phase modulation signal of the water-fat difference of the seed voxel and When the phase vector calculation formula is determined, the sign on the right side of the equal sign is positive. When the phase vector calculation formula is negative, the phase vector of the seed voxel is calculated. and setting Chk of the seed voxel to 1; Step 56, determine whether there is a non-empty stack, if so, return to step 53, if not, end the region growing.
3. The method according to claim 2, characterized in that The multiple stacks uniformly cover the entire phase difference range.
4. The method according to claim 2, characterized in that: In step 54, the neighboring voxels of the seed voxel include four voxels that are adjacent to the seed voxel in four directions: above, below, left, and right.
5. The method according to claim 2, characterized in that: The preset area centered on the seed voxel is a square area centered on the seed voxel.
Citation Information
Patent Citations
Procedure for heartbeat-dependent data acquisition in cardiac magnetic resonance tomography
DE102017107013A1
Magnetic resonance imaging apparatus
JP1990144037A
