Method and apparatus for measuring brain free water content, and magnetic resonance imaging system

Abstract

The embodiment of the present application discloses a brain free water content measurement method, device and magnetic resonance system. The method comprises the following steps: in response to the radio frequency excitation field generated according to the magnetic resonance fingerprint sequence applied to the brain, the balance magnetization miscellaneous M0 signal is collected from the radiation emitted by each voxel of the brain, and the M0 value of each voxel of the brain is obtained; the receiving coil sensitivity RP value of each voxel of the brain is obtained; the M0 value of each voxel of the brain is divided by the RP value of the corresponding voxel respectively, and the proton density PD value of each voxel of the brain is obtained; the PD value of the cerebrospinal fluid is taken as a reference PD value; the PD value of each voxel of the brain is divided by the reference PD value respectively, and the free water content of each voxel of the brain is obtained. The embodiment of the present application improves the speed and accuracy of the brain free water content measurement.

Description

Technical Field

[0001] This invention relates to the field of MRI (Magnetic Resonance Imaging) technology, and in particular to methods, devices and MRI systems for measuring the free water content in the brain. Background Technology

[0002] Proton density (PD) is a fundamental magnetic resonance imaging (MRI) measurement that represents the characteristic concentration of free water protons in each voxel. Many neurological diseases are accompanied by increased local or global free water content; for example, stroke, brain tumors, and hepatic encephalopathy are often associated with cerebral edema. While conventional MRI can localize edema, non-quantitative imaging methods lack specificity for normal brain parenchyma, tumor tissue, and edema. Understanding the correct water content is crucial for monitoring treatment response in these diseases. Recent studies have also shown that free water content in the substantia nigra is a promising biomarker for monitoring Parkinson's disease progression, while conventional MRI provides only limited and valuable diagnostic information. Figure 1 These are standard MRI images of Parkinson's disease patients. A is a T1-weighted image (longitudinal relaxation time), B is a T2-weighted image (lateral relaxation time), and C is a T1 Flair (fluid attenuation inversion recovery) image. The white circles pointed to by the white arrows in A, B, and C represent the substantia nigra. Therefore, accurately and conveniently quantifying the free water content in the human brain is of great significance for the study of neurological diseases.

[0003] To date, methods for quantitative imaging of brain free water content or PD based on magnetic resonance imaging have been proposed. These methods can be broadly classified into several categories:

[0004] The first type of approach derives proton density from the T1 plot by assuming a formulaic relationship between PD and T1. However, the application of this type of method is limited because the water content in different regions of healthy brain tissue has different dependencies on T1, and these dependencies are even more pronounced under pathological conditions.

[0005] The second type of method directly determines the MO (equilibrium magnetization miscellaneous) parameter by combining multiple sequences and then determines the water content by correlating the MO in the tissue with the corresponding value. This type of method can achieve high accuracy by considering multiple correction factors, such as the T2* effect, the emitter field inhomogeneity B1+, and the receiver coil inhomogeneity. However, this type of method has limitations such as long acquisition time, the need for sequence customization, and complex post-processing, which hinders its clinical application.

[0006] The third category attempts to obtain multiple biophysical parameters, such as T1, T2, and PD, based on a single MR sequence acquisition. In this category, a technique called "MRF (MR Fingerprinting)" can generate accurate T1, T2, and MO maps within a short scan time based on pattern recognition properties. However, since MO is the product of PD and RP (Receiver Coil Profile), it cannot directly quantify the brain's free water content. Summary of the Invention

[0007] In view of this, the present invention provides, on the one hand, a method for measuring brain free water content, and on the other hand, a device and MRI system for measuring brain free water content, so as to improve the speed and accuracy of brain free water content measurement.

[0008] A method for measuring the free water content in the brain, comprising:

[0009] In response to a radio frequency excitation field generated based on a magnetic resonance fingerprint sequence applied to the brain, balanced magnetization miscellaneous MO signals are acquired from the radiation emitted by each excited voxel of the brain to obtain the MO value of each voxel of the brain.

[0010] Acquire the receiver coil sensitivity RP value for each voxel in the brain;

[0011] Divide the MO value of each voxel in the brain by the RP value of the corresponding voxel to obtain the proton density PD value of each voxel in the brain.

[0012] Use the PD value of cerebrospinal fluid as the reference PD value;

[0013] The free water content of each voxel in the brain is obtained by dividing the PD value of each voxel by the reference PD value.

[0014] Using the PD value of cerebrospinal fluid as a reference PD value includes:

[0015] The mean PD value of all voxels in the cerebrospinal fluid region of the contralateral ventricle is calculated and used as the reference PD value.

[0016] The acquisition of the RP value of each voxel in the brain includes:

[0017] In response to a radio frequency excitation field applied to the brain, generated according to a sequence for acquiring RP, RP signals are acquired from radiation emitted by each voxel of the brain that is excited, and the RP value of each voxel of the brain is obtained.

[0018] The acquisition of MO signals from radiation emitted by each voxel stimulated by the brain further includes:

[0019] The longitudinal relaxation time T1 signal is acquired by collecting the radiation emitted by each excited voxel in the brain, and the T1 value of each voxel in the brain is obtained.

[0020] The acquisition of the RP value of each voxel in the brain includes:

[0021] The longitudinal relaxation rate R1 of each voxel in the brain is obtained by taking the reciprocal of the T1 value.

[0022] An R1 map of the brain is generated based on the measured R1 values ​​of all voxels in the brain.

[0023] A unified segmentation algorithm, specifically designed for estimating the multiplicative deviation between the measured R1 value and the actual R1 value of each voxel in the brain, is used to uniformly segment the R1 map of the brain, thereby obtaining the multiplicative deviation between the measured R1 value and the actual R1 value of each voxel in the brain. The multiplicative deviation corresponding to each voxel in the brain is then used as the RP of each voxel in the brain.

[0024] A device for measuring brain free water content, the device comprising:

[0025] The signal acquisition module is used to acquire balanced magnetized miscellaneous MO signals from the radiation emitted by each voxel of the brain in response to a radio frequency excitation field generated based on the magnetic resonance fingerprint sequence applied to the brain, so as to obtain the MO value of each voxel of the brain.

[0026] The RP acquisition module is used to acquire the receiver coil sensitivity RP value for each voxel in the brain;

[0027] The PD acquisition module is used to divide the MO value of each voxel in the brain by the RP value of the corresponding voxel to obtain the proton density (PD) value of each voxel in the brain; the PD value of the cerebrospinal fluid is used as the reference PD value.

[0028] The free water content acquisition module is used to divide the PD value of each voxel in the brain by the reference PD value to obtain the free water content of each voxel in the brain.

[0029] The RP acquisition module acquires the RP of each voxel in the brain, including:

[0030] In response to a radio frequency excitation field applied to the brain, generated according to a sequence for acquiring RP, RP signals are acquired from radiation emitted by each voxel of the brain that is excited, and the RP value of each voxel of the brain is obtained.

[0031] The signal acquisition module is further used for,

[0032] The T1 signal is collected from the radiation emitted by each voxel of the brain that is stimulated, and the T1 value of each voxel of the brain is obtained.

[0033] Furthermore, the RP acquisition module acquires the RP of each voxel in the brain, including:

[0034] The reciprocal of the T1 value of each voxel in the brain is taken to obtain the measured R1 value of each voxel in the brain; an R1 map of the brain is generated based on the measured R1 values ​​of all voxels in the brain; a unified segmentation algorithm specifically used to estimate the multiplicative deviation between the measured R1 value and the actual R1 value of each voxel in the brain is used to uniformly segment the R1 map of the brain to obtain the multiplicative deviation between the measured R1 value and the actual R1 value of each voxel in the brain, and the multiplicative deviation corresponding to each voxel in the brain is used as the RP of each voxel in the brain.

[0035] A magnetic resonance imaging system includes any of the devices described above.

[0036] In the above embodiments, by applying a radio frequency excitation field generated according to the MRF sequence to the brain, the MO value of each voxel in the brain is collected, and the PR value of each voxel in the brain is obtained. According to MO = PD * RP, the PD value of each voxel in the brain is obtained. Then, the PD value of the cerebrospinal fluid is used as the reference PD value. The PD value of each voxel in the brain is divided by the reference PD value to obtain the free water content of each voxel in the brain. Thus, the acquisition time is short, the sequence does not need to be customized, and the post-processing is simple, which improves the speed of brain free water content measurement and realizes the quantitative measurement of brain free water content, thereby improving the accuracy of brain free water content measurement. Attached Figure Description

[0037] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which will make the above and other features and advantages of the present invention more apparent to those skilled in the art. In the drawings:

[0038] Figure 1 A standard MRI image of a patient with Parkinson's disease;

[0039] Figure 2 A flowchart of a method for measuring brain free water content provided in the first embodiment of the present invention;

[0040] Figure 3 This is a flowchart of the free water content measurement method provided in the second embodiment of the present invention;

[0041] Figure 4 In this embodiment of the invention, T1 and MO images of the brain were acquired using MRF sequences;

[0042] Figure 5 This is a map of the free water content in the brain calculated using the method provided in this embodiment of the invention.

[0043] Figure 6This is a schematic diagram of the brain free water content measuring device provided in an embodiment of the present invention.

[0044] The accompanying figure is labeled as follows:

[0045] 201～205 step 301～307 step 41 T1 scan of the brain 42 MO images of the brain 61 Signal acquisition module 62 RP Acquisition Module 63 PD acquisition module 64 Free water content acquisition module Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the present invention clearer, the following embodiments are provided to further illustrate the present invention in detail.

[0047] Figure 2 The flowchart of the brain free water content measurement method provided in the first embodiment of the present invention is as follows:

[0048] Step 201: In response to the radio frequency excitation field generated based on the magnetic resonance fingerprint (MRF) sequence applied to the brain, the MO signal is acquired from the radiation emitted by each voxel of the brain that is excited, and the MO value of each voxel of the brain is obtained.

[0049] Step 202: Obtain the RP value of each voxel in the brain.

[0050] Step 203: Divide the MO value of each voxel in the brain by the RP value of the corresponding voxel to obtain the PD value of each voxel in the brain.

[0051] Step 204: Use the PD value of cerebrospinal fluid as the reference PD value.

[0052] In one optional embodiment, the average PD value of all voxels in the cerebrospinal fluid region of the lateral ventricle is calculated, and this average value is used as the reference PD value.

[0053] Step 205: Divide the PD value of each voxel in the brain by the reference PD value to obtain the free water content of each voxel in the brain.

[0054] In the above embodiments, by applying a radio frequency excitation field generated according to the MRF sequence to the brain, the MO value of each voxel in the brain is collected, and the PR value of each voxel in the brain is obtained. According to MO = PD * RP, the PD value of each voxel in the brain is obtained. Then, the PD value of the cerebrospinal fluid is used as the reference PD value. The PD value of each voxel in the brain is divided by the reference PD value to obtain the free water content of each voxel in the brain. Thus, the acquisition time is short, the sequence does not need to be customized, and the post-processing is simple, which improves the speed of brain free water content measurement and realizes the quantitative measurement of brain free water content, thereby improving the accuracy of brain free water content measurement.

[0055] In one optional embodiment, in step 202, obtaining the RP value of each voxel in the brain can be achieved in one of the following two ways:

[0056] Method 1: In response to a radio frequency excitation field generated according to a sequence for acquiring RP applied to the brain, RP signals are acquired from the radiation emitted by each voxel of the brain that is excited, and the RP value of each voxel of the brain is obtained.

[0057] The sequences used to obtain the RP include FLASH (Fast Low Angle SHot) and others. Method Two: First, in step 201, while acquiring the MO signal from the radiation emitted by each excited voxel in the brain, it further includes: acquiring the T1 signal from the radiation emitted by each excited voxel in the brain to obtain the T1 value of each voxel in the brain;

[0058] Then, the reciprocal of the T1 value for each voxel in the brain is taken to obtain the measured R1 (longitudinal relaxation rate) value for each voxel in the brain;

[0059] Then, an R1 map of the brain is generated based on the measured R1 values ​​of all voxels in the brain.

[0060] Finally, a unified segmentation algorithm specifically designed for estimating the multiplicative deviation between the measured R1 value and the actual R1 value of each voxel in the brain is used to uniformly segment the R1 map of the brain, thereby obtaining the multiplicative deviation between the measured R1 value and the actual R1 value of each voxel in the brain. The multiplicative deviation corresponding to each voxel in the brain is then used as the RP of each voxel in the brain.

[0061] Figure 3 The flowchart of the free water content measurement method provided in the second embodiment of the present invention is as follows:

[0062] Step 301: Apply a radio frequency excitation field generated according to the MRF sequence to the brain to excite the brain nuclei.

[0063] The earliest known MRF technology comes from a 2013 paper in Nature entitled "Magnetic Resonance Fingerprinting" (http: / / scihub.cc / 10.1038 / nature11971), with Dan Ma as the first author.

[0064] Compared to traditional MRI weighted imaging (such as T1-weighted imaging, T2-weighted imaging, and PD-weighted imaging), it possesses unique signals, hence the name "magnetic resonance fingerprint," similar to a human fingerprint. MRF technology is a novel method of quantitative MRI that can rapidly and simultaneously measure multiple tissue properties, such as T1, T2, and MO, in a single acquisition.

[0065] Step 302: Collect T1 and MO signals from radiation emitted by stimulated nuclei in the brain to obtain T1 and MO maps of the brain.

[0066] The T1 chart and the MO chart have exactly the same dimensions (i.e., width and height).

[0067] Figure 4 In this embodiment of the invention, T1 and MO images of the brain were acquired using MRF sequences, where 41 is the T1 image and 42 is the MO image.

[0068] Step 303: According to R1 i，j =1 / T1 i，j The R1 map of the brain is generated from the T1 map of the brain.

[0069] Where (i, j) are the position coordinates of any pixel in the T1 and R1 images of the brain, i is the coordinate in the width direction, j is the coordinate in the height direction, 0≤i≤M, 0≤j≤N, M is the width of the T1 and R1 images of the brain, and N is the height of the T1 and R1 images of the brain. i，j T1 is the measured R1 value of pixel (i, j) in the R1 map of the brain. i，j This represents the T1 value of pixel (i, j) in the T1 map of the brain. The unit of T1 is usually milliseconds.

[0070] That is, the R1 map of the brain is exactly the same size as the T1 map of the brain (i.e., the width and height are the same), and for any pixel (i, j) in the R1 map of the brain, its measured R1 value is R1. i，j Satisfy: R1 i，j =1 / T1 i，j .

[0071] Step 304: Using a unified segmentation algorithm specifically designed to estimate the multiplicative deviation between the measured R1 value and the actual R1 value of each voxel in the brain, the R1 map of the brain is uniformly segmented to obtain the multiplicative deviation between the measured R1 value and the actual R1 value of each voxel in the brain. The multiplicative deviation corresponding to each voxel in the brain is used as the RP value of each pixel in the RP map of the brain, thereby generating the RP map of the brain.

[0072] The unified segmentation algorithm was proposed by Weiskopf N, Lutti A, Helms G, et al., in their paper "Unified segmentation based correction of R1 brainmaps for RF transmit field inhomogeneities," published in Neuroimage (pages 2116-2124) in 2011. Details of the article are as follows:

[0073] Weiskopf N, Lutti A, Helms G, et al. Unified segmentation based correction of R1 brain maps for RF transmit field inhomogeneities (UNICORT) [J]. Neuroimage, 2011, 54(3): 2116-2124.

[0074] The unified segmentation algorithm is used to estimate the multiplicative bias of any type of data with multiplicative bias. In the brain R1 map obtained in step 303, the R1 value of each pixel is the measured R1 value. There is a multiplicative bias between the measured R1 value and the actual R1 value, i.e., measured R1 value = actual R1 value * multiplicative bias. By using the unified segmentation algorithm to uniformly segment the brain R1 map obtained in step 303, the multiplicative bias between the measured R1 value and the actual R1 value of each pixel can be obtained.

[0075] Step 305: Based on PD i，j =MO i，j / RP i，j A PD map of the brain is generated from the MO map and RP map of the brain.

[0076] Where (i, j) are the position coordinates of any pixel in the PD, MO, and RP maps of the brain, i is the coordinate in the width direction, j is the coordinate in the height direction, 0≤i≤M, 0≤j≤N, M is the width of the PD, MO, and RP maps of the brain, and N is the height of the PD, MO, and RP maps of the brain. i，j RP is the MO value of pixel (i, j) in the MO map of the brain. i，j PD is the RP value of pixel (i, j) in the RP map of the brain. i，j denoted as PD value of pixel (i, j) in the PD map of the brain.

[0077] That is, the PD map of the brain is exactly the same size as the MO map and RP map (i.e., the width and height are the same), and for any pixel (i, j) in the PD map of the brain, its PD value is PD. i，j Satisfy: PD i，j =MO i，j / RP i，j .

[0078] Step 306: Use the mean PD value of the lateral ventricle cerebrospinal fluid (CSF, CerebroSpina1F1uid) region in the PD map of the brain as the reference PD value: PD ref .

[0079] That is, first, locate the cerebrospinal fluid region of the lateral ventricle in the PD image, then calculate the average PD value of all pixels in that region, and use this average value as the PD value. ref .

[0080] Step 307: According to FWCnor i，j =PD i，j / PDref gives the free water content of each voxel in the brain.

[0081] Among them, FWCnor i，j Let be the free water content of the voxel corresponding to pixel (i, j).

[0082] Figure 5 This is a graph showing the free water content in the brain calculated using the method provided in this embodiment of the invention.

[0083] The beneficial technical effects of the above embodiments are as follows:

[0084] 1. Using only one sequence: the MRF sequence, the signal acquisition time is short, and it avoids the errors caused by image registration. Furthermore, the post-processing is simple, so it can obtain the free water content of the brain more quickly and is more efficient. This method is more suitable for clinical applications.

[0085] Second, because the dependence of water content on T1 differs between pathological conditions and healthy controls, many existing techniques have limitations in clinical application. This method, however, uses a unified segmentation algorithm to estimate the multiplicative bias of the R1 plot, thereby obtaining the RP. This algorithm is effective in many neurodegenerative diseases, thus ensuring the clinical applicability of this method.

[0086] Figure 6 This is a schematic diagram of the structure of a brain free water content measuring device provided in an embodiment of the present invention. The device mainly includes:

[0087] The signal acquisition module 61 is used to acquire MO signals from the radiation emitted by each voxel of the brain in response to a radio frequency excitation field generated according to the MRF sequence applied to the brain, and to obtain the MO value of each voxel of the brain.

[0088] RP acquisition module 62 is used to acquire the RP value of each voxel in the brain.

[0089] The PD acquisition module 63 is used to divide the MO value of each voxel in the brain obtained by the signal acquisition module 61 by the RP value of the corresponding voxel obtained by the RP acquisition module 62 to obtain the PD value of each voxel in the brain; and to use the PD value of the cerebrospinal fluid as the reference PD value.

[0090] The free water content acquisition module 64 is used to divide the PD value of each voxel in the brain obtained by the PD acquisition module 63 by the reference PD value obtained by the PD acquisition module 63 to obtain the free water content of each voxel in the brain.

[0091] In one optional embodiment, the PD acquisition module 63 uses the PD value of cerebrospinal fluid as a reference PD value, including: averaging the PD values ​​of all voxels in the cerebrospinal fluid region of the contralateral ventricle, and using the average value as the reference PD value.

[0092] In one optional embodiment, the RP acquisition module 62 acquires the RP of each voxel in the brain, including: in response to a radio frequency excitation field generated according to a sequence for acquiring the RP applied to the brain, acquiring the RP signal from radiation emitted by each excited voxel in the brain, to obtain the RP value of each voxel in the brain.

[0093] In an optional embodiment, the signal acquisition module 61 is further configured to acquire the T1 signal from the radiation emitted by each voxel of the brain that is stimulated, and obtain the T1 value of each voxel of the brain.

[0094] Furthermore, the RP acquisition module 62 acquires the RP of each voxel in the brain, including: taking the reciprocal of the T1 value of each voxel in the brain obtained by the signal acquisition module 61 to obtain the measured R1 value of each voxel in the brain; generating an R1 map of the brain based on the measured R1 values ​​of all voxels in the brain; uniformly segmenting the R1 map of the brain using a unified segmentation algorithm specifically used to estimate the multiplicative deviation between the measured R1 value and the actual R1 value of each voxel in the brain, obtaining the multiplicative deviation between the measured R1 value and the actual R1 value of each voxel in the brain, and using the multiplicative deviation corresponding to each voxel in the brain as the RP of each voxel in the brain.

[0095] The magnetic resonance imaging system proposed in this embodiment of the invention may include the free water content measuring device provided in the first or second embodiment above.

[0096] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for measuring the free water content in the brain, characterized in that, include: In response to a radio frequency excitation field generated based on a magnetic resonance fingerprint sequence applied to the brain, the balanced magnetization miscellaneous M0 signal and longitudinal relaxation duration T1 signal are acquired from the radiation emitted by each excited voxel of the brain to obtain the M0 value and T1 value of each voxel of the brain. Acquiring the receiver coil sensitivity RP value for each voxel in the brain, wherein acquiring the receiver coil sensitivity RP value for each voxel in the brain includes: The longitudinal relaxation rate R1 of each voxel in the brain is obtained by taking the reciprocal of the T1 value. An R1 map of the brain is generated based on the measured R1 values ​​of all voxels in the brain. Obtain the multiplicative deviation between the measured R1 value and the actual R1 value for each voxel in the brain; Based on the multiplicative bias corresponding to each voxel in the brain, the receiving coil sensitivity RP value of each voxel in the brain is determined; Divide the M0 value of each voxel in the brain by the RP value of the corresponding voxel to obtain the proton density PD value of each voxel in the brain. Use the PD value of cerebrospinal fluid as the reference PD value; The free water content of each voxel in the brain is obtained by dividing the PD value of each voxel by the reference PD value.

2. The method according to claim 1, characterized in that, The multiplicative deviation between the measured R1 value and the actual R1 value for each voxel in the brain includes: The R1 map of the brain is uniformly segmented using a unified segmentation algorithm based on multiplicative bias, and the multiplicative bias between the measured R1 value and the actual R1 value of each voxel in the brain is obtained.

3. The method according to claim 1, characterized in that, Using the PD value of cerebrospinal fluid as a reference PD value includes: The mean PD value of all voxels in the cerebrospinal fluid region of the contralateral ventricle is calculated and used as the reference PD value.

4. The method according to claim 1, characterized in that, The acquisition of the RP value of each voxel in the brain includes: In response to a radio frequency excitation field applied to the brain, generated according to a sequence for acquiring RP, RP signals are acquired from radiation emitted by each voxel of the brain that is excited, and the RP value of each voxel of the brain is obtained.

5. A device for measuring the free water content in the brain, characterized in that, The device includes: The signal acquisition module (61) is used to acquire the balanced magnetization miscellaneous M0 signal and the longitudinal relaxation duration T1 signal from the radiation emitted by each voxel of the brain in response to the radio frequency excitation field generated according to the magnetic resonance fingerprint sequence applied to the brain, so as to obtain the M0 value and T1 value of each voxel of the brain. The RP acquisition module (62) is used to acquire the receiver coil sensitivity RP value of each voxel in the brain, wherein acquiring the receiver coil sensitivity RP value of each voxel in the brain includes: The longitudinal relaxation rate R1 of each voxel in the brain is obtained by taking the reciprocal of the T1 value. An R1 map of the brain is generated based on the measured R1 values ​​of all voxels in the brain. Obtain the multiplicative deviation between the measured R1 value and the actual R1 value for each voxel in the brain; Based on the multiplicative bias corresponding to each voxel in the brain, the receiving coil sensitivity RP value of each voxel in the brain is determined; The PD acquisition module (63) is used to divide the M0 value of each voxel in the brain by the RP value of the corresponding voxel to obtain the proton density PD value of each voxel in the brain; and to use the PD value of the cerebrospinal fluid as the reference PD value. The free water content acquisition module (64) is used to divide the PD value of each voxel in the brain by the reference PD value to obtain the free water content of each voxel in the brain.

6. The apparatus according to claim 5, characterized in that, The RP acquisition module (62) acquires the RP of each voxel in the brain, including: In response to a radio frequency excitation field applied to the brain, generated according to a sequence for acquiring RP, RP signals are acquired from radiation emitted by each voxel of the brain that is excited, and the RP value of each voxel of the brain is obtained.

7. A magnetic resonance imaging system, characterized in that, Includes the apparatus as described in any one of claims 5 or 6.

Images