JP6632905B2 - Magnetic resonance imaging equipment - Google Patents

Description

  Embodiments of the present invention relate to a magnetic resonance imaging apparatus.

  The magnetic resonance imaging apparatus excites a nuclear spin of a subject placed in a static magnetic field with a radio frequency (RF) signal of a Larmor frequency, and reconstructs a magnetic resonance signal generated from the subject with the excitation. An imaging device that generates an image by using

  In the field of magnetic resonance imaging, there is an imaging method called diffusion weighted imaging (DWI). In an image taken by diffusion-weighted imaging, that is, a diffusion-weighted image, a contrast corresponding to an apparent diffusion coefficient (ADC) of water molecules in a tissue can be obtained, and thus diagnosis of cerebral infarction, tumor, and the like can be obtained. Many are used. An ADC calculated for each pixel and arranged corresponding to the pixel position of the diffusion-weighted image is called an ADC map or an ADC image.

  In diffusion weighted imaging, a pulse sequence in which two gradient magnetic field pulses called MPG (Motion Probing Gradient) pulses are applied before a pulse sequence for data collection such as EPI (Echo Planer Imaging). An index defined by the magnitude and pulse length of an MPG pulse and the interval between two MPG pulses is called a b-value.

  The above ADC can be calculated from the signal intensity and the b value of each pixel of the diffusion weighted image. At this time, a diffusion weighted image captured with at least two different b values is required.

  On the other hand, in order to increase the measurement accuracy of the ADC, it is necessary to further increase the number of b values. For example, the number of b values may need to be 10 or more. However, it is necessary to perform diffusion weighted imaging every time the b value is changed. Therefore, if the number of b values is increased, the entire imaging time becomes extremely long. In addition, when the number of b values is increased, there is a problem that an image is distorted in accordance with the value of b value, in addition to a longer imaging time.

  On the other hand, various high-speed imaging methods for shortening the imaging time have been conventionally studied. For example, a high-speed imaging method called k-tBLAST or k-tSENSE has been proposed.

US patent 7,005,853 B2

J. Tsao et.al, “k-t BLAST and k-t SENSE: Dynamic MRI With High Frame Rate Exploiting Spatiotemporal Correlations”, Mag. Res. In Medicine 50: 1031-1042 (2003)

  The problem to be solved by the present invention is to provide a magnetic resonance imaging apparatus capable of shortening the imaging time of a diffusion weighted image corresponding to a plurality of b values and acquiring a diffusion weighted image with less distortion. .

  The magnetic resonance imaging apparatus according to the present embodiment includes an estimating unit that estimates distortion of a diffusion-weighted image caused by application of MPG and data for generating a plurality of diffusion-weighted images corresponding to a plurality of different b values. And a collection unit that collects reconstruction data thinned out in accordance with a predetermined speed-up rate in the phase encoding direction and expansion data, and uses at least the expansion data and the estimated distortion. A generation unit configured to generate the diffusion-weighted image in which the aliasing has been removed and the distortion has been corrected for each of the plurality of different b-values from the data for reconstruction, for each of the b-values. Prepare.

FIG. 1 is a configuration diagram illustrating an example of the entire configuration of a magnetic resonance imaging apparatus according to an embodiment. FIG. 2 is a functional block diagram of the first embodiment. FIG. 3 is a diagram illustrating an example of a pulse sequence of diffusion weighted imaging. The figure which showed and compared Mono-Exponential Model and Bi-Exponential Model. The figure which shows the example of the pulse sequence set to a different b value. FIG. 7 is a diagram showing a comparison between a scan without thinning and a scan with thinning. FIG. 4 is a diagram schematically illustrating how aliasing occurs. The figure which shows the relationship between a training scan and a main scan. 9 is a flowchart illustrating an example of a pre-scan process. FIG. 7 is a diagram for explaining the concept of processing for obtaining correction data from a reconstructed image. 9 is a flowchart illustrating a processing example of a training scanless type main scan. FIG. 4 is a view for explaining the concept of processing for extracting data for development from data of a main scan. 9 is a flowchart illustrating a processing example of generating data for development. The figure which shows the concept of the process which produces | generates the data for expansion | deployment. 9 is a flowchart illustrating a processing example of a reconfiguration process including an expansion process. FIG. 3 is a first diagram illustrating the concept of a reconstruction process including an expansion process. FIG. 3 is a first diagram illustrating the concept of a reconstruction process including an expansion process. FIG. 9 is a functional block diagram of the magnetic resonance imaging apparatus according to the second embodiment. 9 is a flowchart illustrating a processing example of a reconfiguration process including an expansion process. The figure which shows the example in which the interval of a b value is an irregular interval.

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the following embodiments, portions denoted by the same reference numerals perform the same operation, and duplicate description will be omitted as appropriate.

(First embodiment)
FIG. 1 is a block diagram illustrating the overall configuration of the magnetic resonance imaging apparatus 1 according to the first embodiment. The magnetic resonance imaging apparatus 1 includes a magnet gantry 100, a bed 500, a control cabinet 300, a console 400, and an RF (Radio Frequency) coil 20.

  The magnet mount 100 has a static magnetic field magnet 10, a gradient magnetic field coil 11, and a WB (Whole Body) coil 12, and these components are housed in a cylindrical housing. The bed 500 has a bed main body 50 and a top plate 51.

  The control cabinet 300 includes a static magnetic field power supply 30, a gradient magnetic field power supply 31 (31x for the X axis, 31y for the Y axis, 31z for the Z axis), an RF receiver 32, an RF transmitter 33, and a sequence controller 34. .

  The console 400 includes a processing circuit 40, a storage circuit 41, a display 42, and an input device 43. The console 400 functions as a host computer.

  The static magnetic field magnet 10 of the magnet mount 100 has a substantially cylindrical shape, and generates a static magnetic field in a bore in which a subject, for example, a patient is transported. The bore is a space inside the cylinder of the magnet mount 100. The static magnetic field magnet 10 has a built-in superconducting coil, and the superconducting coil is cooled to extremely low temperature by liquid helium. The static magnetic field magnet 10 generates a static magnetic field by applying a current supplied from the static magnetic field power supply 30 to the superconducting coil in the excitation mode. Thereafter, when the mode shifts to the permanent current mode, the static magnetic field power supply 30 is disconnected. Once in the permanent current mode, the static magnetic field magnet 10 continues to generate a large static magnetic field for a long time, for example, over one year. A black circle on the chest of the subject indicates the center of the magnetic field.

  The gradient magnetic field coil 11 also has a substantially cylindrical shape, and is fixed inside the static magnetic field magnet 10. The gradient magnetic field coil 11 applies a gradient magnetic field to the subject in the X-axis, Y-axis, and Z-axis directions by a current supplied from a gradient magnetic field power supply (31x, 31y, 31z).

  The bed main body 50 of the bed 500 can move the top board 51 in the vertical and horizontal directions. The subject placed on the table 51 is moved to a predetermined height before imaging. Thereafter, at the time of imaging, the subject is moved into the bore by moving the top board 51 in the horizontal direction.

  The WB coil 12 is also called a whole-body coil, and is fixed in a substantially cylindrical shape inside the gradient magnetic field coil 11 so as to surround the subject. The WB coil 12 transmits an RF pulse transmitted from the RF transmitter 33 toward the subject, and also transmits a magnetic resonance signal emitted from the subject by excitation of hydrogen nuclei, that is, an MR (Magnetic Resonance) signal. Receive.

  The magnetic resonance imaging apparatus 1 includes an RF coil 20 as shown in FIG. The RF coil 20 is a coil placed close to the body surface of the subject. The RF coil 20 includes a plurality of coil elements 200. Since the plurality of coil elements 200 are arranged in an array inside the RF coil 20, they may be called a PAC (Phased Array Coil).

  In the high-speed imaging method referred to as parallel imaging, an imaging process is performed by thinning out images in the phase encoding direction to increase the speed of imaging, and expansion processing for removing aliasing caused by the thinning-out using the sensitivity map of each element coil 200 is performed. Is being done. Also, in the imaging method called ktSENSE disclosed in Non-Patent Document 1, and in the imaging according to the second embodiment described later, expansion processing for removing aliasing using the sensitivity map of each element coil 200 is also performed. Is going.

  The RF transmitter 33 generates an RF pulse based on an instruction from the sequence controller 34. The generated RF pulse is transmitted to the WB coil 12 and applied to the subject. An MR signal is generated from the subject by the application of the RF pulse. The RF signal 20 or the WB coil 11 receives this MR signal.

  The MR signal received by the RF coil 20, more specifically, the MR signal received by each coil element in the RF coil 20 is transmitted to the RF receiver 32 via a cable provided on the top board 51 and the bed main body 50. Is entered. The RF receiver 32 performs AD (Analog to Digital) conversion of the MR signal and outputs the result to the sequence controller 34. The digitally changed MR signal is sometimes called raw data. The AD conversion may be performed inside the RF coil 20 or in the coil selection circuit 36.

  The sequence controller 34 scans the subject under the control of the console 400 by driving the gradient magnetic field power supply 31, the RF transmitter 33, and the RF receiver 32, respectively. Upon receiving raw data from the RF receiver 32 by scanning, the sequence controller 34 transmits the raw data to the console 400.

  The sequence controller 34 includes a processing circuit (not shown). This processing circuit is configured by, for example, a processor that executes a predetermined program, or hardware such as an FPGA (Field Programmable Gate Array) and an ASIC (Application Specific Integrated Circuit).

  The console 400 includes a storage circuit 41, a display 42, an input device 43, and a processing circuit 40. The storage circuit 41 is a storage medium including an external storage device such as a hard disk drive (HDD) and an optical disk device in addition to a read only memory (ROM) and a random access memory (RAM). The storage circuit 41 stores various information and data, and also stores various programs executed by a processor included in the processing circuit 40.

  The input device 43 is, for example, a mouse, a keyboard, a trackball, a touch panel, or the like, and includes various devices for an operator to input various information and data. The display 42 is a display device such as a liquid crystal display panel, a plasma display panel, and an organic EL panel.

  The processing circuit 40 is a circuit including, for example, a CPU and a dedicated or general-purpose processor. The processor realizes various functions to be described later by executing various programs stored in the storage circuit 41. The processing circuit 40 may be configured by hardware such as an FPGA or an ASIC. Various functions described later can also be realized by these hardware. Further, the processing circuit 40 can also realize various functions by combining software processing by a processor and a program with hardware processing.

  FIG. 2 is a block diagram illustrating functions realized particularly by the console 400 in the magnetic resonance imaging apparatus 1 according to the first embodiment. As shown in FIG. 2, the processing circuit 40 of the console 400 includes a distortion estimation data collection function 402, a thinned data collection function 403, a development data collection function 404, a distortion estimation function 405, a development processing function with correction 406, and a display control function. 407 is realized. Each of these functions is realized, for example, by a processor included in the processing circuit 40 executing a predetermined program stored in the storage circuit 41.

  The distortion estimation data collection function 402 collects distortion estimation data for estimating distortion of a diffusion weighted image caused by application of MPG (Motion Probing Gradient). The distortion estimation data is collected by, for example, a pre-scan performed at the time of installation of the apparatus. The pre-scan is performed on an imaging target other than the patient, for example, a phantom having a predetermined shape. The distortion estimation data collection function 402 sets the imaging conditions for the pre-scan in the sequence controller 402 and executes the pre-scan. Then, the distortion estimation data collection function 402 collects the MR signals collected by the pre-scan, and sends the collected MR signals to the distortion estimation function 405 as distortion estimation data.

  The distortion estimation function 405 estimates the distortion of the diffusion-weighted image due to the application of the MPG from the distortion estimation data, and stores information on the estimated distortion in the storage circuit 41 as distortion data.

  The thinning data collection function 403 performs a main scan and a training scan on a subject (for example, a patient), and collects data for generating a plurality of diffusion-weighted images corresponding to a plurality of different b values. The data collected in the main scan is sent to the expansion processing function with correction 406 as thinned data.

  The expansion data collection function 404 extracts data collected by the training scan from the data collected by the thinned data collection function 403 and sends the data to the expansion processing function 406 as expansion data. As will be described later, the main scan can be configured so that data to be collected in the training scan is collected in the main scan. Such a main scan is called a training scanless main scan.

  The expansion processing function with correction 406 generates a diffusion-weighted image in which aliasing has been removed and distortion has been corrected by the expansion processing using the above-described expansion data and distortion data, for each b value from the thinned data. I do.

  The display control function 407 performs control and processing for displaying the diffusion-weighted image generated by the expansion processing function with correction 406 on the display 42.

  Before giving a detailed description of each function described above, a pulse sequence and an imaging method used in the first embodiment will be described. Note that the same pulse sequence and imaging method are used in other embodiments.

  FIG. 3 shows a sequence in which a diffusion-weighted MPG pulse is incorporated into an SE (spin echo) type EPI (echo planer imaging) sequence as an example of a diffusion-weighted imaging pulse sequence commonly used in each embodiment. FIG. In this sequence, a refocus pulse having a flip angle of 180 ° is applied TE / 2 after an excitation pulse having a flip angle of 90 °, and data collection of an EPI sequence is started approximately TE / 2 after the refocus pulse. You. Here, TE represents the effective echo time. Strictly speaking, the time from the excitation pulse to the acquisition of data at the center of the k-space (phase encoding direction) of the EPI sequence is defined as the effective echo time TE.

  The MPG pulse is a gradient magnetic field pulse applied to enhance diffusion, and one MPG pulse is applied between the excitation pulse and the refocus pulse and between the refocus pulse and the data acquisition.

As an index indicating the magnitude of the effect of the MPG pulse on the ADC (apparent diffusion coefficient) of the tissue, a b-factor (b-factor) is usually used. The value of the b value is expressed by the following (Equation 1).
b = γG ² τ ² (T− (τ / 3)) (Equation 1)
Here, γ is the magnetic rotation ratio, G is the magnitude of the gradient magnetic field of the MPG pulse, τ is the pulse length of the MPG pulse, and T is the leading edge of the first MPG pulse to the leading edge of the subsequent MPG pulse. Is the interval.

In the single shot SE-EPI, all the k-space data for one slice is collected by one data acquisition accompanying one excitation pulse. On the other hand, in the multi-shot SE-EPI sequence, k-space data for one slice is collected in a plurality of times. The signal strength S (signal strength S at the center of k-space) of the MR signal collected by the pulse sequence of FIG. 2 is represented by the following (Equation 2).
S (b) = S ₀ * exp [−b * ADC] (Equation 2)
Here, S ₀ is a constant, b is the b value of (Equation 1), and ADC is the diffusion coefficient ADC.

(Equation 2) is a model (Mono-Exponential Model) in which the signal intensity can be represented by one exponential function. In this model, the logarithm of the signal strength changes linearly with the b value as in the following (Equation 3).
ln (S / S0) =-b * ADC (Equation 3)
As can be seen from (Equation 3), if at least two b values and the signal strength corresponding to each b value are obtained, the diffusion coefficient ADC can be calculated. However, by increasing the number of b values, The calculation accuracy of the diffusion coefficient ADC can be improved.

  On the other hand, when the b value is high or low, it is known that the change in the signal strength according to the b value deviates from the Mono-Exponential Model and approaches the Bi-Exponential Model.

FIG. 4 is a diagram showing the Mono-Exponential Model and the Bi-Exponential Model in comparison. In liver tests and the like, an analysis called IVIM (Intra Voxel Incoherent Motion) is performed. In IVIM, in order to analyze both the perfusion component and the diffusion component of a tissue, the diffusion coefficient (ADC) _P due to perfusion and the diffusion coefficient (ADC) _D due to diffusion are For example, it is calculated based on a Bi-Exponential Model represented by the following (Equation 4).
S (b) = S ₀ * [f * exp (−b * (ADC) _P ) + (1−f) * exp (−b * (ADC) _D ] (Equation 4)
F in (Equation 4) represents the ratio of the perfusion component to the whole.

As can be seen from the graph of the Bi-Exponential Model shown in FIG. 4, the change in signal intensity expressed in logarithm changes linearly in a region where the b value is large, but changes greatly in a curve in a region where the b value is small. Varies with the rate. In the IVIM, a signal value obtained by a plurality of types of imaging using a plurality of types of b-values is curve-fitted by a Bi-Exponential Model to thereby obtain a diffusion coefficient (ADC) _P and a diffusion coefficient (diffusion). ADC) _D is required. In order to obtain the diffusion coefficient (ADC) _P and the diffusion coefficient (ADC) _D by diffusion with high accuracy, it is desirable to take an image with a large number of types, for example, 10 or more b values, and the imaging time is long. Become.
Further, when obtaining diffusion coefficients in a plurality of directions, the imaging time becomes longer according to the number of directions. Therefore, each embodiment described below provides a technique capable of shortening the imaging time even when imaging is performed with a plurality of b values.

  In the case of imaging with a plurality of b values, it is necessary to consider the problem of image distortion caused by an eddy current magnetic field and the like in addition to an increase in imaging time. FIG. 5 shows an example of a pulse sequence of diffusion weighted imaging in which the b value is set to a different type, that is, a different value. In the example shown in FIG. 5, τ is set to different values for the pulse length τ and the pulse interval T of the MPG pulse to obtain different b values. An eddy current magnetic field is generated by the application of the MPG pulse, and a phase error occurs in the MR signal due to the eddy current magnetic field. This phase error causes a position error in the reconstructed image. In the case of imaging with a plurality of b values, as shown in FIG. 5, since the position and shape of the MPG pulse are different for each b value, the state of generation of the eddy current magnetic field is different for each b value. As a result, the position error in the reconstructed image shows a different value for each b value. Therefore, in each of the embodiments described below, in addition to the technology for shortening the imaging time, a technology capable of correcting a position error even when imaging with a plurality of b values is provided.

  FIGS. 6A and 6B are diagrams showing the b value as the horizontal axis and the position of the phase encoding on the k space as the vertical axis. The phase encoding direction and the lead-out direction can be in any two directions orthogonal to each other, but in the following description, the phase encoding direction is assumed to be the y direction (ky direction in k space) and the lead-out direction is assumed to be the x direction. . 6A and 6B, the lead-out direction is a direction perpendicular to the paper surface.

  FIG. 6A is a diagram showing a conventional scanning method without thinning, and the intervals of phase encoding are Nyquist intervals at each b value. The Nyquist interval means an interval at which aliasing does not occur in the phase encoding direction, that is, an interval satisfying the Nyquist theorem. The width in the phase encoding direction is a width at which a desired resolution can be obtained. On the other hand, FIG. 6B is a diagram showing a scanning method in which phase encoding is thinned out at a predetermined speed-up rate. In the example of FIG. 6B, the speed-up rate is set to 4, and the phase encoding interval is four times the Nyquist interval. Note that the number of b values is the same number of 10 in FIGS. 6A and 6B.

  In the scanning method shown in FIG. 6B, the number of phase encodes is 1/4 that of the scanning method shown in FIG. 6A, so that the imaging time is also reduced to 1/4. However, in the scanning method shown in FIG. 6B, since the Nyquist theorem is not satisfied, aliasing occurs in the phase encoding direction.

  FIG. 7 is a diagram schematically showing how aliasing occurs. In each of the two diagrams in FIG. 7, the vertical axis is the real space y, and the ky space is obtained by performing an inverse Fourier transform. The abscissas of the two figures are obtained by performing inverse Fourier transform on the b-value space, and this space is hereinafter referred to as Fb space. The two-dimensional space represented by the y space and the Fb space is called a y-Fb space.

The diagram on the left side of FIG. 7 is a diagram showing signals corresponding to the scan without thinning-out, and no aliasing has occurred. On the other hand, the diagram on the right side of FIG. 7 is a diagram illustrating signals corresponding to scans with thinning. The hatched signal regions near the center of the left-hand diagram overlap a plurality of times, indicating that aliasing has occurred. For example, four pixel values ρ ₁ ~ρ ₄ in the four pixel positions on the left side of the figure, the right side of FIG overlaps the position of the [rho _alias. Here, the pixel position or the pixel value means a position in the y-Fb space or a signal value.

The expansion process described below is a process of obtaining _four pixel values ρ _{1 to} ρ 4 from a pixel value ρ _alias at a certain pixel position in the right diagram and arranging them at corresponding pixel position values in the left diagram, that is, 7 is performed on each pixel shown in the right-hand side of FIG. 7 in the opposite direction to remove aliasing.

  In this development processing, a scan called a training scan is performed separately from the main scan, and the above-mentioned aliasing removal processing is performed using the development data collected in the training scan.

  FIG. 8A is a diagram illustrating a relationship between the training scan and the main scan. The training scan is also performed on a plurality of b values in the same manner as the main scan, and each b value in the training scan corresponds to each b value in the main scan. However, the phase scan direction differs between the training scan and the main scan. The phase encoding interval in the training scan is the Nyquist interval, and aliasing does not occur in an image generated from data collected in the training scan.

  On the other hand, the entire range of the phase encoding in the training scan is smaller than the entire range of the main scan. Specifically, in the training scan, only the low frequency region of the ky space is collected, and the high frequency region is not collected. Therefore, the resolution of the image generated from the training scan is lower than the resolution of the image generated from the main scan.

  However, the data for decompression generated from the training scan is only used for the decompression process to the last, and the finally generated image depends on the resolution of the image generated from the main scan. Therefore, even if the resolution of the image generated from the training scan is low, it does not matter much.

  The training scan may be performed before the main scan, or may be performed after the main scan. However, the training scan and the main scan need to be performed on the same subject.

  Although the total imaging time of the main scan and the training scan is slightly longer than that of the main scan alone (FIG. 6B), it is shorter than that of the conventional scan without thinning (FIG. 6A). , Greatly reduced.

  FIG. 8B is a diagram illustrating a scanning method that can further reduce the imaging time. In the scan shown in FIG. 8B, data for development is collected during the main scan. Therefore, it is not necessary to perform a training scan separately from a main scan. That is, the scan illustrated in FIG. 8B can be called a training scanless type main scan. In the main scan of the training scanless type, the phase encoding interval is set to the Nyquist interval only in the low frequency region of the ky space, and data is collected at thinned intervals in other regions. That is, in the low-frequency region of the ky space, the scan is used for both the main scan and the training scan.

  Next, more specific methods of the pre-scan, the training scan-less main scan, and the expansion processing will be sequentially described.

  FIG. 9 is a flowchart illustrating an example of a pre-scan process. As described above, the pre-scan is a scan for collecting distortion estimation data for estimating distortion of a diffusion-weighted image caused by application of an MPG pulse. The pre-scan is performed, for example, when the apparatus is installed. The imaging target of the pre-scan is, for example, a phantom having a predetermined shape.

  Steps ST100 to ST104 are steps corresponding to the distortion estimation data collection function 402. Steps ST105 to ST108 are steps corresponding to the distortion estimation function 405.

  In step ST100, imaging conditions for pre-scan based on diffusion weighted imaging are set in the sequence controller. The b value in the pre-scan does not necessarily need to be the same as the b value used in the main scan. However, it is preferable to cover the range of the b value assumed to be used in the main scan, and it is preferable to perform the pre-scan for a plurality of b values within the range.

  In step ST101, an initial value of the b value is set, and then diffusion-weighted imaging is performed on the phantom according to the set imaging conditions (step ST102). When the imaging for one b value is completed, the b value is updated to a different value, and the same imaging is repeated until the diffusion weighted imaging for all b values is completed (steps ST103 and ST104). The MR signals collected in the pre-scan are sent to the distortion estimation function 405.

  The distortion estimation function 405 generates a reconstructed image for each b value (step ST105). Thereafter, the distortion estimation function 405 calculates the distortion for each b value by comparing the phantom shape of the reconstructed image with the actual phantom shape, that is, the shape without distortion (step ST106). Then, the correction data is calculated from the distortion (step ST107), and the calculated correction data, that is, the distortion estimation data is stored in the storage circuit 41 (step ST108).

FIG. 10 is a diagram for explaining the concept of processing for obtaining correction data from a reconstructed image. Now, it is assumed that the actual shape of the phantom is a rectangle indicated by a broken line in FIG. 10 in the xy plane. It is assumed that the shape of the phantom on the reconstructed image is the shape shown by the solid line in FIG. In such a case, in step ST106, the distortion estimation function 405 determines the actual position of the phantom (the position indicated by the small broken circle in FIG. 10) and the corresponding reconstruction at a plurality of predetermined positions of the phantom shape. The position of the phantom on the image (the position indicated by a small solid circle in FIG. 10) is compared, and the difference between the two is determined as the distortion Δy _{x, y} in the y direction at each position (x, y).

In FIG. 10, the distortion Δy _{x, y} is obtained at the six positions at the end of the phantom. However, the positions and the number of the distortions to be obtained are not limited to these, and the distortion Δy _{x, y} is obtained in a wider range and at more positions. You may ask. Also, the distortion Δy _{x, y} is obtained for each b value from the reconstructed image corresponding to each b value.

Although the distortion Δy _{x, y} may be stored as correction data as it is, a method of obtaining it as a function of the b value is also conceivable. For example, the distortion Δy _{x, y} is represented by a quadratic function such as the following (Equation 5), and the coefficients (α _{x, y} , β _{x, y} , γ _{x, y} ) of the respective orders are calculated as correction data. A method is conceivable.
Δy _{x, y} = α _{x, y} * b ² + β _{x, y} * b + γ _{x, y} (Equation 5)
The coefficients (α _{x, y} , β _{x, y} , γ _{x, y} ) are calculated based on the distortion Δy _{x, y} at each position (x, y) at which the distortion Δy _{x, y} was obtained, and a plurality of b used in the pre-scan. It can be calculated by a technique such as curve fitting of the plotted points of the values with a quadratic function.

If the correction data is stored in the storage circuit 41 as the above-described coefficients (α _{x, y} , β _{x, y} , γ _{x, y} ), the distortion Δy _x for any b value used in the main scan _{, y} . Also, the strain [Delta] y _{x, y} obtained position, i.e., the strain [Delta] y _x, and measuring the position of _y, if the position to be corrected in this scan are different, strain [Delta] y x at a plurality of measurement _positions, from _y, The distortion at the position to be corrected may be obtained by an interpolation process or the like.

Although the above-described distortion Δy _{x, y} is a distortion in the real space, it may be calculated as a phase distortion in the k space instead.

  Since the distortion mainly occurs in the phase encoding direction (in this example, the y direction), it is not necessary to pay much attention to the distortion in the readout direction (x direction).

  In the pre-scan, since the phase encoding direction is scanned without thinning, the imaging time may be longer than in the main scan described below. However, the pre-scan may be performed at the time of installation or the like, and there is no problem even if the imaging time is slightly longer. In addition, since the pre-scan is performed on the phantom, for example, it is not necessary to consider the physical and mental burden on the patient.

  Next, the training scanless type main scan will be specifically described. The main scan is a scan performed on a subject, for example, a patient.

  FIG. 11 is a flowchart illustrating a processing example of the training scanless type main scan. Steps ST200 to ST205 are steps corresponding to the thinned data collection function 403, and step ST206 is a step corresponding to the development data collection function 404.

  In step ST200, the imaging conditions of the main scan based on the diffusion weighted imaging are set in the sequence controller. In the main scan, as shown in FIG. 8B, an imaging condition for collecting data at a phase encoding interval is set to a Nyquist interval only in a low frequency region of the ky space, and at a thinned interval in other regions. You.

  In step ST201, an initial value of the b value is set, and thereafter, diffusion-weighted imaging is performed on the subject according to the set imaging conditions (step ST202). When imaging for one b-value is completed, the b-value is updated to a different value, and the same imaging is repeated until diffusion-weighted imaging for all b-values is completed (steps ST203 and ST204).

  FIG. 12 is a diagram illustrating the concept of the processing of step ST205 and step ST206. In step ST205, thinning data of the entire k-space is extracted from the collected data (shown on the left side of FIG. 12) and stored in the storage circuit 41 for each b value as k-space data for generating a reconstructed image. save. The reconstructed image here means, for example, a reconstructed image for diagnosing a subject.

  On the other hand, in step ST206, non-decimated data in the k-space low-frequency region is extracted from the collected data, and is stored in the storage circuit 41 for each b value as k-space data for generating expansion data.

  Next, a specific procedure of the expansion processing will be described. The expansion process includes two phases: a phase for generating expansion data, and a phase of a reconstruction process including the expansion process.

  FIG. 13 is a flowchart illustrating a processing example of a phase for generating data for development. FIG. 14 is a diagram illustrating the concept of the process of the phase for generating the data for development.

  Steps ST300 to ST305 are steps corresponding to the expansion processing function with correction 406.

  In step ST300, the expansion processing function with correction 406 acquires, from the storage circuit 41, k-space data (no thinning) for generating expansion data corresponding to all b values. This data is data in a kx-ky space as shown in FIG. However, the data in the ky direction is data only in the low frequency region (see the lower right diagram in FIG. 12).

  In step ST301, inverse Fourier processing is performed in k-space to generate real space data (xy space data) for each b value, as illustrated in FIG. 14B. In FIG. 14B, the shape of the subject is simulated by a circle. Aliasing does not occur in the k-space data for generating the development data because the data is not thinned in the phase encoding direction.

Next, the expansion processing function with correction 406 reads out the correction data (α _{x, y} , β _{x, y} , γ _{x, y} ) stored in the storage circuit 41 and calculates the b value based on (Equation 5). Each time, the distortion of the real space data (xy space data) is corrected. Since the correction data (α _{x, y} , β _{x, y} , γ _{x, y} ) does not correspond to all positions of the real space data, interpolation processing of the correction data is performed as necessary, Perform distortion correction.

  In step ST303, the y-direction data and the b-direction data are two-dimensionally arrayed to generate yb space data as illustrated in FIG. Specifically, y-direction data corresponding to a specific x position of each xy space data is extracted from the xy space data and arranged in the b direction. Therefore, a plurality of yb space data are actually generated corresponding to each x position. However, in FIG. 14C, in order to avoid complexity, yb space data corresponding to one x position is generated. -B Only spatial data is shown.

  Next, in step ST304, the yb space data is subjected to an inverse Fourier transform in the b direction to generate y-Fb space data as illustrated in FIG.

  In step ST305, expansion data defined in the y-Fb space is generated from the generated y-Fb space data. The expansion data generated here is, for example, data corresponding to the covariance matrix M defined in Non-Patent Document 1.

  In the above description, distortion correction using correction data is performed on xy space data for each b value (step ST302). Alternatively, the distortion correction using the correction data may be performed after the yb space is generated in step ST303.

  Next, the phase of the reconstruction processing including the expansion processing will be described. FIG. 15 is a flowchart illustrating a processing example of the phase of the reconstruction processing including the expansion processing. FIGS. 16 and 17 are diagrams showing the concept of the processing in the phase of the reconstruction processing including the expansion processing.

  Steps ST400 to ST405 in FIG. 15 are steps corresponding to the expansion processing function with correction 406.

  In step ST400, the expansion processing function with correction 406 acquires from the storage circuit 41 k-space data (with thinning) for generating a reconstructed image corresponding to all b values. This data is data in a kx-ky space as shown in FIG. The data in the kx-ky space includes the entire area in the ky direction (see the upper right diagram in FIG. 12).

  In step ST401, inverse Fourier processing is performed in k-space to generate real space data (xy space data) for each b value as illustrated in FIG. Also in FIG. 16B, the shape of the subject is simulated by a circle. However, since the k-space data for generating the reconstructed image is thinned out in the phase encoding direction, aliasing occurs in the y direction, and the circle is folded in the y direction.

  In step ST402, the y-direction data and the b-direction data are two-dimensionally arrayed to generate yb space data as illustrated in FIG. Specifically, similarly to step ST303, y-direction data corresponding to a specific x position of each xy space data is extracted from the xy space data and arranged in the b direction. Therefore, a plurality of yb space data are actually generated corresponding to each x position. However, in FIG. 16C, to avoid complexity, yb space data corresponding to one x position is generated. -B Only spatial data is shown.

  Next, in step ST403, the yb space data is subjected to an inverse Fourier transform in the b direction to generate y-Fb space data as illustrated in FIGS. 16D and 17A. FIG. 16D and FIG. 17A are the same figures. There are a plurality of y-Fb space data corresponding to each x position.

  In step ST404, expansion data is applied to the y-Fb space data to perform an expansion operation, thereby generating y-Fb space data from which aliasing has been removed. FIG. 17B illustrates the y-Fb spatial data from which aliasing has been removed. The expansion operation here is, for example, an operation corresponding to Expression [5] in Non-Patent Document 1, and the like.

  Next, in step ST405, y-Fb space data is subjected to Fourier transform in the b direction to generate yb space data. FIG. 17C is a diagram illustrating the yb space data generated in step ST405.

In step ST406, the correction data (α _{x, y} , β _{x, y} , γ _{x, y} ) stored in the storage circuit 41 is read, and based on (Equation 5), the correction data in the y-direction in the y-b space data is read. Correct the distortion. Similar to step ST302, distortion correction is performed by performing interpolation processing of correction data as necessary.

  In addition, a plurality of y-Fb space data illustrated in FIGS. 17A and 17B and a plurality of yb space data illustrated in FIG. 17C exist for each x position.

  Finally, in step ST407, the x direction data is combined for each b value, and xy space data for each b value is generated as illustrated in FIG. 17D. The xy space data for each b value is a reconstructed image for each b value from which aliasing has been removed.

  In the above description, the distortion correction is performed on the yb space data (step ST406). Alternatively, after step ST407, distortion correction may be performed on the xy space data for each b value.

  As described above, in the magnetic resonance imaging apparatus 1 according to the first embodiment, even when diffusion-weighted imaging is performed using a large number of b values, imaging is performed by thinning out the phase encoding direction of each main scan. The time can be reduced, and the physical and mental burden on the subject can be reduced. In addition, aliasing caused by thinning can be appropriately removed by a development process. The expansion processing may be performed after the imaging, and does not affect the imaging time of the subject.

  The image distortion caused by the application of the MPG pulse differs depending on the value of the b value. On the other hand, in the magnetic resonance imaging apparatus 1 according to the first embodiment, since image distortion can be appropriately corrected even for a plurality of different b values, deterioration in image quality can be suppressed.

(Second embodiment)
Next, a magnetic resonance imaging apparatus 1 according to a second embodiment will be described. In the second embodiment, data of a plurality of channels collected by a plurality of RF coils or a plurality of element coils is added to a plurality of RF coils or a plurality of element coils in addition to the deployment data used in the first embodiment. Expansion processing is performed using a plurality of corresponding sensitivity maps.

  FIG. 18 is a functional block diagram of the magnetic resonance imaging apparatus 1 according to the second embodiment. The main difference from the first embodiment (FIG. 2) is that a sensitivity map collection function 408 is provided.

  The sensitivity map collection function 408 is also a function realized by the processing circuit 40. A scan for collecting a sensitivity map is executed by the sensitivity map collection function 408, and a plurality of sensitivity maps corresponding to a plurality of RF coils or a plurality of element coils are generated from data collected by the scan. The scan for collecting the sensitivity map is performed on the same subject as the main scan. The scan for collecting the sensitivity map may be performed before the main scan or may be performed after the main scan. The generated sensitivity map data is sent to the expansion processing function with correction function 406.

  FIG. 19 is a flowchart illustrating an example of a process of generating a reconstructed image including a development process using a sensitivity map according to the second embodiment.

  The processing of steps ST500 to ST503 is basically the same as the processing of steps ST400 to ST403 in the first embodiment. However, the second embodiment is different in that the processes of steps ST500 to ST503 are performed on data of a plurality of channels collected by a plurality of RF coils or a plurality of element coils.

  Therefore, the y-Fb space data generated in step ST503 is data for a plurality of channels.

  In step ST504, expansion processing is performed on the y-Fb spatial data for a plurality of channels by applying the expansion data and the sensitivity map to generate y-Fb spatial data from which aliasing has been removed. The expansion operation here is, for example, an operation corresponding to Expression [7] in Non-Patent Document 1, and the like.

  The processing in steps ST505 to ST507 is substantially the same as the processing in steps ST405 to ST407 in the first embodiment, and a description thereof will be omitted.

  In the second embodiment, data collected by a plurality of RF coils or a plurality of element coils is used. That is, even though the number of images finally obtained is the same, more data is used to generate this image. Therefore, in addition to the effects obtained in the first embodiment, it is expected that a higher SNR and a higher speed-up rate can be realized.

(Other embodiments)
In the embodiments described above, the intervals of the b value are assumed to be equal as illustrated in FIG. 4 and the like. However, a reconstructed image can be generated even if the intervals of the b value are unequal. For example, as shown in FIG. 20, it is conceivable that the interval of the b value is set to be short in a region where the change of the signal intensity is sharp, and the interval of the b value is set to be long in the region where the change of the signal intensity is small. Also, from the viewpoint of image quality, it is better to set the b value so that the logarithm of the b value is at equal intervals.

  In the above-described embodiment, the distortion correction is performed in the real space (for example, the y space). However, the present invention is not limited to this. The correction data of the distortion in the real space generated from the collected data of the pre-scan may be converted into the correction data of the phase distortion, and the phase distortion may be corrected in the k-space in the image reconstruction processing of the main scan.

  According to the magnetic resonance imaging apparatus of at least one embodiment described above, it is possible to shorten the imaging time of the diffusion weighted image corresponding to a plurality of b values and acquire a diffusion weighted image with less distortion.

(Correspondence with the description of the claims)
The distortion estimation data collection function and the distortion estimation function in the description of the embodiment are examples of the estimation unit in the claims. The thinned data collection function and the development data collection function in the description of the embodiment are examples of the collection unit in the claims. The expansion processing function with correction in the description of the embodiment is an example of the generation unit in the description of the claims. The thinning data in the description of the embodiment is an example of the reconstruction data in the description of the claims.

  Although several embodiments of the present invention have been described, these embodiments are provided by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in other various forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and equivalents thereof.

Reference Signs List 1 magnetic resonance imaging apparatus 20 RF coil 40 processing circuit 41 storage circuit 42 display 43 input device 400 console 402 distortion estimation data collection function 403 thinning data collection function 404 expansion data collection function 405 distortion estimation function 406 expansion processing function with correction

Claims (8)

An estimator for estimating distortion of the diffusion-weighted image caused by application of MPG (Motion Probing Gradient);
Data for generating a plurality of diffusion-weighted images corresponding to a plurality of different b-values, wherein reconstruction data and decompression data thinned out in a phase encoding direction according to a predetermined speed-up rate are A collection unit to collect,
At least the data for decompression and the decompression process using the estimated distortion removes aliasing and corrects the diffusion-weighted image in which the distortion is corrected for each of the plurality of different b values. A generating unit that generates from the data for each of the b values;
A magnetic resonance imaging apparatus comprising: The expansion data is data for expanding the thinned reconstruction data.
The magnetic resonance imaging apparatus according to claim 1. Further comprising a plurality of RF coils,
The generation unit, by the expansion processing using the expansion data and a plurality of sensitivity maps corresponding to the plurality of RF coils, the diffusion weighted image from which aliasing has been removed, from the reconstruction data, generated for each b value,
The magnetic resonance imaging apparatus according to claim 1. The estimating unit calculates the distortion as a distortion in a real space in a phase encoding direction caused by the application of the MPG,
The generation unit performs correction using the distortion in the real space after the expansion processing, to generate the diffusion-weighted image in which the distortion is corrected,
The magnetic resonance imaging apparatus according to claim 1. The distortion is calculated for each imaging condition including at least the b value.
The magnetic resonance imaging apparatus according to claim 1. The collecting unit collects the decompression data such that a range in the phase encoding direction in the k space is smaller than a range in the phase encoding direction of the reconstruction data.
The magnetic resonance imaging apparatus according to claim 1. The collection unit, when a scan for collecting the reconstruction data is a main scan, the training scan performed before the main scan, collects the deployment data,
The magnetic resonance imaging apparatus according to claim 1. The collecting unit collects the data for reconstruction and the data for decompression by a main scan, and in the main scan, the data in the phase encoding direction is not thinned out in a low frequency region of k space, while the low frequency Areas other than the area are scans that have been thinned out.
The magnetic resonance imaging apparatus according to claim 1.