JP2002000579A - Mr diffusion image imaging instrument - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize an MR diffusion image imaging instrument capable of decreasing the influence of body movement in a diffusion image measurement without using a single shot sequence and a navigation echo. SOLUTION: After magnetization is excited in a single-dimensional direction like a wavelet function by a combination of an RF pulse 11 and an inclined magnetic field 16, homopolar diffusion detection inclined magnetic fields 16 and 17 are arranged with a π pulse 12 sandwiched, and a readout inclined magnetic field 18 is applied in an x direction to measure a spin echo signal 19. The scale and shift quantity of the wavelet function are varied to execute a series of measurements, and the measurement data is reverse-wavelet transformed to re-form the image. The (y) direction is subjected to a wavelet encoding instead of a normal phase encoding, and the (x) direction is subjected to a frequency encoding. A slice selection is performed in a (z) inclined magnetic field at the 12 o'clock of a refocus πpulse. Although an arcifact is distributed on the entire image in a Fourier encoding, in a wavelet encoding, even if an object moves temporarily, an excited image is controlled so that a part of the image is dim at moving.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a magnetic resonance diagnostic apparatus (MRI), and more particularly to an MR diffusion imaging apparatus for medical magnetic resonance diagnosis.

[0002]

2. Description of the Related Art A diffusion image measurement method using a medical magnetic resonance diagnostic apparatus (MRI) is a method of forming a contrast based on the diffusion coefficient of water molecules in a living body. This diffusion image measurement method can detect cerebral infarction in the acute stage of about 3 hours after the onset, and enables rapid treatment immediately after the onset, so that its clinical application is rapidly expanding.

[0003] The diffuse image is measured by the method of Stejskal and Tanner. This involves exciting magnetization, applying a first diffusion gradient magnetic field to disperse the spin phase at the sub-voxel level, applying a π pulse, and further applying a second diffusion gradient magnetic field to apply the spin diffusion phase. Is to be undone.

The phase of a stationary spin returns to its original state, but the phase of the spin whose coordinates have moved due to molecular diffusion between two gradient magnetic fields does not return to its original state. Is reduced.

[0005] The decrease in the signal of the diffused spin due to the diffusion gradient magnetic field is expressed by the following equation (1) (for example, D. Le Biha
n, et al. , “MR Imaging of Intravoxel Incoher
entMotions: Application to Diffusion and Perfus
ion in Neurologic Disorders ”, Radioology, 161, 4
01-407 (1986)).

(Equation 1) Here, γ is the gyromagnetic ratio, G is the intensity of the diffusion gradient magnetic field pulse, and D is the apparent diffusion coefficient, which includes not only pure diffusion but also effects such as microcirculation in a living body. Also, δ is the application time of the diffusion gradient magnetic field pulse, △ is the interval between two diffusion gradient magnetic field pulses, and S ₀ is the signal intensity when there is no diffusion.

[0006] The molecular diffusion of water is due to Brownian motion and originally has no direction, but in a living body, it becomes directional due to cell membrane or microscopic motion. In this case, the diffusion coefficient D is a symmetric second-order tensor amount.

By changing the axis to which the diffusion gradient magnetic field is applied and performing measurement, it is also possible to image the directionality of nerve fibers and the like.

[0008] The above-mentioned weak points of the diffusion image measurement method are as follows.
Since a high-intensity diffusion gradient magnetic field is used, it is easily affected by the movement of the subject. This effect causes large image degradation even if the body motion during the phase encoding cycle is 1 mm or less.

There are two methods for taking measures against this image deterioration.
The first method combines high-speed imaging, which acquires all image data with a single excitation of magnetization, with diffusion image measurement. Diffusion image measurement is based on Single shot Echo Planar Imagin.
It is executed in a sequence based on g (SS-EPI) or Single shot Fast SE (SS-FSE). The single shot sequence is particularly effective for patients who show seizures or who have difficulty resting.

Disadvantages of the first method, the high-speed photographing method, are that the S / N and resolution of the image are sacrificed, and that the hardware is limited.

The second method is a method of detecting information of a body motion generated during a phase encoding cycle by a navigation echo and removing the influence of the body motion by correcting the phase of the image data.

The advantage of this second method is that it is not limited by hardware. The disadvantages of this second method, on the other hand, are that it can only be corrected for certain movements of the subject, often for rigid translational or rotational movements.

When the movement of the subject is not a simple movement in a slice but a three-dimensional translation or rotation, three navigation echoes corresponding to three axes must be generated, and the sequence becomes long. And the amount of data processing increases, making real-time processing difficult.

As another countermeasure, heartbeat synchronization is also used to exclude the influence of heartbeat movement.

[0015]

By the way, a single-shot sequence such as SS-EPI or SS-FSE can be used for diffusion image measurement with a high magnetic field machine. Single shot sequences are difficult due to hardware limitations.

[0016] Therefore, Multi-shot EPI and ordinary FSE
It is practical to use the k-space division type sequence and the navigation echo together.

However, as described above, the correction by the navigation echo has a problem that accurate correction cannot be performed if the body motion is not a simple rigid body motion.

An object of the present invention is to realize an MR diffusion image photographing apparatus capable of reducing the influence of body movement in diffusion image measurement without relying on a single shot sequence and a navigation echo.

[0019]

In order to achieve the above object, the present invention is configured as follows. (1) Static magnetic field, gradient magnetic field, high frequency magnetic field, magnetic resonance signal detecting means, image reconstructing means, image displaying means,
In the MR diffusion imaging apparatus equipped with these control means, a first step of simultaneously applying a high-frequency magnetic field and a gradient magnetic field to a predetermined space region including the subject to excite the spatial coordinates in a wavelet function is performed. , Followed by 1-3
A second application of an axial first diffusion gradient magnetic field to the region;
Performing a third step of applying a refocusing π pulse to the region following the diffusion gradient magnetic field;
A fourth step of applying a second diffusion gradient magnetic field to the region following the refocusing π pulse is performed, and a fifth step of detecting a signal by applying a readout gradient magnetic field to the diffusion gradient magnetic field is performed. Then, the first to fifth steps are repeated while changing the scale and the shift amount of the wavelet function, and then an image is formed by using the inverse wavelet transform.

(2) Preferably, in the above (1),
One direction in space is encoded with a wavelet excitation function, and the remaining one or two directions are Fourier-encoded.

(3) Preferably, the above (1),
In (2), the refocusing π pulse in the third step has slice selectivity.

(4) Preferably, the above (1),
In (2), a first step of simultaneously applying the high-frequency magnetic field and the gradient magnetic field to excite spatial coordinates in a wavelet function form, and a fifth step of applying a read-out gradient magnetic field to detect a signal. At any time, phase encoding in the slice direction is performed to obtain three-dimensional image data.

(5) Preferably, the above (1),
In (2), (3) and (4), after completing the first to fifth steps of one cycle, the unexcited region is excited in a wavelet function without waiting for the longitudinal relaxation time, and the next cycle The first to fifth steps are performed, and similarly, the entire imaging time is reduced by exciting the magnetization of the unexcited portion without waiting for longitudinal relaxation.

By combining wavelet encoding and diffusion measurement, it is possible to measure a diffused image in which the image quality is less deteriorated due to the movement of the subject. In wavelet encoding, each time one magnetization is excited by RF irradiation, only a one-dimensional part of space is excited to measure a signal.

Therefore, even if the subject moves temporarily, the movement only causes blurring of a portion in the excited image at the time of the movement.

On the other hand, in Fourier encoding,
Fourier component (spatial frequency component) measured at the time of movement
Are affected by motion, the artifacts are distributed throughout the image.

Since the wavelet encoding only excites a part of the longitudinal magnetization, the excitation efficiency of the magnetization is lower than that of the Fourier encoding.

However, since there is unexcited magnetization, the next translation or scale wavelet excitation can be performed continuously without providing a relaxation time for longitudinal magnetization, and the overall excitation efficiency is equivalent to that of Fourier encoding. Can be.

[0029]

Embodiments of the present invention will be described below in detail with reference to the accompanying drawings. First, the principle of the present invention will be described. The combination of the RF pulse and the gradient magnetic field causes the magnetization in the spatial one-dimensional direction (referred to as the y direction) to be spatially excited as a wavelet function in a predetermined spatial region including the subject, and then the Stejskal-Tanner According to the sequence, π
A diffusion detection gradient magnetic field of the same polarity is arranged across the pulse, and x
A spin echo is measured by applying a directional readout gradient magnetic field.

A series of measurements are executed by changing the scale and shift amount of the wavelet function, and the measured data is converted into an inverse wavelet.
The image is reconstructed by the conversion.

In the y direction, wavelet encoding is performed instead of ordinary phase encoding, and in the x direction, frequency encoding is performed. Slice selection is performed by the z gradient magnetic field at the time of the refocusing π pulse. The application of the diffusion gradient magnetic field is performed on the x-axis and y-axis.
An axis and a z-axis of 1 to 3 axes are possible and can be selected according to the purpose.

The wavelet function is a function obtained by reducing or enlarging a localized basic function and translating it. These functions are similar to each other and form an orthogonal function system. Such a wa
There are many types of velet basis functions.

FIG. 3 shows a Haar wavelet as an example.
The mathematical expression is the following expression (2).

(Equation 2) Here, j is a scale parameter (j = 0, 1, 2,...)
.., Log2N-1), N is the number of data points, and k is a translation parameter (k = 1, 2, 4,..., 2 ^j ).

The basic Haar function ψ _0,0 is reduced, and ψ
_{Get 1,0} , ... Function [psi ₁₁ is obtained by the function [psi _{1, 0} is translated.

The encoding method of the MR signal using the wavelet function is described in, for example, L. P. Panych and F. A-Jol
Esz's paper “A Dynamically Adaptive Imaging Al
gorithm for Wavelet-Encoded MRI ”, Magnetic R
esonance in Medicine, 32, p738-p748, (1994).

Due to the excitation in the form of a wavelet function, the measurement signal is expressed as the following equation (3) as a function of y.

(Equation 3) Here, ρ (y) is the density of magnetization, and ρ (j, k) is ρ
(Y) represents a wavelet transform. The density of magnetization ρ (y) is
It is obtained from ρ (j, k) by the inverse wavelet transform shown in the following equation (4).

(Equation 4) Instead of the usual Fourier transform phase encoding,
Encoding and decoding of position coordinates in the y direction are performed by wavelet excitation / inverse wavelet transform. The encoding in the x direction is frequency encoding as in the Fourier transform method.

The wavelet excitation shape is localized in the space y direction, and a region where magnetization is not excited remains on the y axis.

Accordingly, the wavelet of the next translation or scale can be obtained without providing the relaxation time of the longitudinal magnetization.
Excitation can be performed. For the optimal excitation order of the wavelet function for this purpose, see J. Mol. B. Weaver, Y. Paper "Wavelet-encoded MR Imaging" by Xu et al., Magneti
c Resonance in Medicine, 24, p275-p287 (1992)
Is described in detail.

FIG. 5 shows an example. In FIG. 5, reference numeral 51 denotes a subject, for example, a plurality of types of excitation site widths in the y direction of the subject 51. TR is the repetition period of the wavelet excitation, and is shorter than the longitudinal relaxation time Tl. Tmi
n is a time interval in which the same spin is excited, and is limited by the longitudinal relaxation time Tl.

As shown in FIG. 5, most wavelet excitations can be repeated in less than Tmin.

With the wavelet excitation pulse, it is difficult to select a slice at the time of excitation, and the slice selection is performed by using a refocus π pulse for spin echo formation as slice selectivity and limiting the signal generation slice.

By adding phase encoding in the slice direction, it is possible to obtain a three-dimensional diffused image. The repetition of the Z-phase encoding requires a waiting time TR of 1 to 5 seconds in consideration of the vertical relaxation time. From the inside, the iteration loop is an x-frequency encoding (readout), a y-wavelet loop, and a z-phase encoding loop.

In the wavelet encoding, each time one magnetization is excited by RF irradiation, only one one-dimensional space is excited to measure a signal.

Thus, even if the subject 51 moves temporarily, the movement only causes a blur in a part of the image excited at the time of the movement.

Therefore, by combining the wavelet encoding and the diffusion measurement method, it is possible to measure a diffusion image in which the image quality is less deteriorated due to the movement of the subject 51.

On the other hand, in the Fourier encoding method, since the Fourier component (spatial frequency component) measured at the time of movement is affected by the movement, artifacts are distributed over the entire image, and the image quality of the entire image deteriorates. .

This is because the region to be photographed is a region that does not involve much movement, but if there is a region near the region that is included in the photographing region, the region with the movement causes The image quality will be degraded.

On the other hand, if the method combines the wavelet encoding and the diffusion measurement method, as described above,
Even if a part of the image temporarily moves, the movement only causes a blur in a part of the image excited at the time of the movement, and does not cause a deterioration in the image quality of a region to be photographed.

Since the wavelet encoding only excites a part of the longitudinal magnetization, the excitation efficiency of the magnetization is lower than that of the Fourier encoding.

However, the presence of unexcited magnetization allows continuous wavelet excitation of the next translation or scale without providing a relaxation time for longitudinal magnetization, and the overall excitation efficiency is equivalent to that of Fourier encoding. Can be.

FIG. 4 is a schematic configuration diagram of a nuclear magnetic resonance diagnostic apparatus to which the present invention is applied. Referring to FIG.
Is an electromagnet or a permanent magnet for generating a uniform static magnetic field B0 inside the subject 401, 414a is a transmitting coil for generating a high-frequency magnetic field (RF magnetic field), and 414b is for detecting a nuclear magnetic resonance signal generated from the subject. Detection coil.

Reference numeral 409 denotes gradient magnetic fields Gx, Gy and Gz whose intensities linearly change in three orthogonal directions x, y and z.
Is a power supply for supplying a current to the gradient magnetic field.

Reference numeral 408 denotes a computer, and 406 denotes a signal processing system.
4, a RAM 425, a magnetic disk 426, a magneto-optical disk 427, and a display 428.

An operation section 421 has a keyboard 422 and a mouse 423.

Next, the outline of the operation of the nuclear magnetic resonance diagnostic apparatus shown in FIG. 4 will be described. Synthesizer 4 of transmission system 404
11 is modulated by the modulator 412 and amplified by the power amplifier 413, and the high-frequency coil 414a
To supply. Thus, a high-frequency magnetic field is generated inside the subject 401 to excite nuclear spins. Usually, the target is ¹ H, but the target may be other nuclei having nuclear spin, such as ³¹ P and ¹² C.

The nuclear magnetic resonance signal emitted from the subject 401 is received by the high-frequency coil 414b,
After 15, the signal is subjected to quadrature phase detection by the quadrature phase detector 416, A / D converted by the A / D converter 417, and input to the computer 408. The high-frequency coil 414
a and 414b may be used for both transmission and reception, or may be separate.

After the signal processing, the computer 408 causes the CRT display 428 to display an image corresponding to the nuclear spin density distribution, relaxation time distribution, spectrum distribution, and the like. The RAM 425 stores data that is being calculated by the computer 408 or final data.

The gradient magnetic field generation system 403, the transmission system 404, and the detection system 405 are all controlled by a sequencer 407, and the sequencer 407 is controlled by a computer 408. The computer 408 is connected to the operation unit 42
1 is controlled by a command.

The characteristic operation of the present invention in such an apparatus will be described in detail below. Here, it is assumed that a one-dimensional wavelet function-like excitation is performed in the y-axis direction on a slice (xy plane) orthogonal to the z-axis.

FIG. 1 is a diagram showing a sequence of a typical single slice measurement example. In FIG. 1, π / 2
-The RF pulse 11 and the y gradient magnetic field 14 are simultaneously applied to excite the y-axis direction in a wavelet function (first step).

The y gradient magnetic field 14 has a negative polarity portion for returning the phase of the transverse magnetization.

In a second step following the first step, a first diffusion gradient magnetic field 16 in three axial directions is applied. Next, in the third step, the refocusing π pulse 12 is applied together with the slice selection z gradient magnetic field 13 to refocus only the magnetization inside the imaging slice plane.

In the fourth step following the refocusing π pulse 12, a second diffusion gradient magnetic field 17 is applied. The phases of spins that do not undergo molecular diffusion are restored by the first and second diffusion gradient magnetic fields 16 and 17, but the phases of spins that have undergone molecular diffusion do not recover and their contribution to echo signals is reduced.

Next, in a fifth step, a readout gradient magnetic field 18 is applied, and an echo signal 19 is sampled in a sampling period 101.

The first to fifth steps are repeated while changing the scale and shift amount of the wavelet function, and all the wavelet
After measuring the signal for the function, reverse for the y-axis direction
Wavelet inverse transform is performed, and in the x direction, inverse Fourier transform is performed to reconstruct a two-dimensional image.

The wavelet function includes various functions other than the Haar function shown in FIG. 3 (the number on the left side of the subscript of the function ψ indicates the scale, and the number on the right side indicates the position). FIG. 6 shows the Battle-Lemarie function as an example. Battle
Since the Lemarie function has a gentler shape than the Haar function, the excitation RF pulse can be shortened. However, on the other hand, since the skirt of the waveform shape is long, overlap occurs between the wavelet function waveforms.

The diffusion gradient magnetic fields 16 and 17 may be from one axis to three axes.
Can be applied to the shaft. When applied to a specific axis, a diffusion contrast is formed by the diffusion in the specific axis direction. When applied in three axial directions, an average diffusion contrast is obtained in each direction of the space.

When the diffusion is not spatially isotropic, the diffusion coefficient becomes a second-order symmetric tensor, and there are six independent components. For example, the application axis of the diffusion gradient magnetic field is x-axis, y-axis, z-axis.
The tensor component can be calculated from six measurements using the axis, the x-axis and the y-axis, the y-axis and the z-axis, and the z-axis and the x-axis. Then, the main direction of diffusion can be obtained for each pixel from the calculated tensor, and this is applied to depiction of nerve running and the like.

FIG. 5 shows the excitation order of the wavelet function. The example shown in FIG. 5 is a case where the field of view is covered by 16 pixels. In this case, 16 wavelet excitations are required, and the point where the number of matrices matches the number of excitations is the same as in phase encoding. Excitation of the same spin occurs at intervals of Tmin in FIG. The wavelet excitation can be repeated with a shorter period TR.

As described above, the overall measurement time can be reduced by assembling the wavelet excitation order.

Next, an example of three-dimensional measurement will be described with reference to FIG. In FIG. 2, the π / 2-RF pulse 11 and the first
A z-direction phase encoding gradient magnetic field 21 is applied between the diffusion gradient magnetic field 16 of FIG. During the application of the refocusing π pulse 12, the slice selection z gradient magnetic field is not applied. In a fourth step following the refocusing π pulse 12, a second diffusion gradient magnetic field 17 is applied.

In the fifth step, the readout gradient magnetic field 18 is applied, and the echo signal 19 is sampled in the sampling period 101.

The first to fifth steps are repeated while changing the scale and the shift amount of the wavelet function.
After measuring the signal for the t-function, the wavelet excitation is repeated with varying steps of z-phase encoding.

After collecting all data, the reverse wave is applied to the y-axis direction.
Let inverse transformation is performed, and in the x and z directions, inverse Fourier transformation is performed to reconstruct a three-dimensional image. The repetition rate of the z-phase encoding step is limited by the longitudinal relaxation time T1.

The z-direction phase encoding gradient magnetic field 21 is not limited to the position shown in FIG. 2, but may be placed at any position between the excitation of magnetization and the readout (in principle, RF
And cannot be applied at the same time as the diffusion gradient magnetic field).

As described above, according to the embodiment of the present invention, since the measurement is performed by combining the wavelet encoding and the diffusion measurement method, even if the subject moves temporarily, the movement does not change. Only the blurring of a part of the image excited at the time when the image is excited can be prevented, and the movement can prevent the entire image from being blurred.

Therefore, it is possible to realize an MR diffusion image photographing apparatus capable of reducing the influence of body movement in diffusion image measurement without relying on a single shot sequence and a navigation echo.

In the above-described example, the case where only one dimension of the space is wavelet-encoded has been described. However, it is also possible to perform a two-dimensional wavelet function-like excitation by a space selective excitation pulse, and to perform body movement. Can be further reduced.

References relating to the design of the space-selective excitation pulse include J. Pauly, D. Nishimura and A. Maco
Paper "Ak-Space Analysis of Small" by vski
-Tip-Angle Excitation ”, Journal of Magnetic
Resonance, 81, p43-p56 (1989).

[0080]

As described above, according to the present invention,
It is possible to realize an MR diffusion imaging apparatus capable of reducing the influence of body movement in diffusion image measurement without relying on a single shot sequence and a navigation echo.

In addition, it is possible to perform diffusion image measurement by utilizing the low sensitivity of the wavelet encoding to body movement, and to obtain a diffusion image with less body movement artifact even with a low magnetic field MRI magnet that cannot perform a single shot sequence. Can be.

In the measurement of the diffusion anisotropy such as a tensor, it is necessary to repeatedly measure the diffusion contrast image several times. However, in the present invention, it is possible to obtain a diffusion image with less body motion artifact even in such a case. Can be.

[Brief description of the drawings]

FIG. 1 is a diagram showing a sequence of the present invention.

FIG. 2 is a diagram showing another sequence of the present invention.

FIG. 3 is a diagram showing a shape of a wavelet function.

FIG. 4 is a diagram showing the overall configuration of a magnetic resonance imaging apparatus to which the present invention is applied.

FIG. 5 is a diagram showing a sequence of excitation of a series of wavelet functions.

FIG. 6 is a diagram showing the shape of another wavelet function.

[Explanation of symbols]

 11 wavelet excitation π / 2-RF pulse 12 refocus π-RF pulse 13 slice selection gradient magnetic field pulse 14 y direction wavelet encoding gradient magnetic field pulse 15 x direction warp gradient magnetic field pulse 16, 17 diffusion gradient magnetic field pulse 18 readout gradient magnetic field pulse 19 echo signal 51 subject 101 sampling section 401 subject 402 static magnetic field generation magnetic circuit 403 gradient magnetic field generation system 404 transmission system 405 detection system 406 signal processing system 407 sequencer 408 computer 409 gradient magnetic field coil 410 gradient magnetic field power supply 411 synthesizer 412 modulator 413 RF Amplifier 414a Transmission RF coil 414b Detection RF coil 415 Preamplifier 416 Quadrature phase detector 417 A / D converter 421 Operation unit

Claims (1)

[Claims] 1. A static magnetic field, a gradient magnetic field, a high-frequency magnetic field,
In an MR diffusion imaging apparatus including a magnetic resonance signal detecting unit, an image reconstructing unit, an image displaying unit, and these control units, an RF magnetic field and a gradient magnetic field are simultaneously applied to a predetermined spatial region including a subject. To excite spatial coordinates in the form of a wavelet function
Performing a second step of applying a first diffusion gradient magnetic field in one to three axes to the region following the excitation, and performing a refocusing π pulse following the diffusion gradient magnetic field. Performing a third step of applying a second diffusion gradient magnetic field to the region following the refocusing π pulse; applying a readout gradient magnetic field to the diffusion gradient magnetic field; A fifth step of applying a signal to detect a signal is performed, and the first to fifth steps are repeated while changing the scale and the shift amount of the wavelet function.
An MR diffusion imaging apparatus characterized in that an image is formed using t-transform.