JP2009273530A - Magnetic resonance imaging apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To acquire stable images which are free from streak artifacts or image distortion in EPI imaging or echo-planar imaging. <P>SOLUTION: A pre-measured echo measured by shortest TE (echo time) imaging without applying phase encoding gradient pulse is separated into an odd echo and an even echo after synthesizing between adjacent slices. In order to smoothly connect the phase variations in an echo train direction, linear fitting of phase variations in a frequency direction is performed with respect to pre-measured data between the odd echoes by using the number of echoes specified from reference echoes to obtain the echo variations (242). The phase variations are accumulated in the echo train direction by using the fitting result, and the result is taken as an odd echo after the correction (243). The same fitting in the frequency direction and phase accumulation in the echo train direction are also performed for differences between the odd echo and the even echo. The phases of the data between the odd echoes after the correction are added, and the result is taken as an even echo after the correction. A peak position in a kx direction of main measurement data is moved to the center by using the odd echo and the even echo after the correction (220). <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a magnetic resonance imaging (MRI) apparatus that measures nuclear magnetic resonance (hereinafter referred to as “NMR”) signals from hydrogen, phosphorus, etc. in a subject and visualizes nuclear density distribution, relaxation time distribution, and the like. In particular, the present invention relates to an MRI apparatus capable of effectively suppressing streak artifacts caused by a local phase change of an echo signal in imaging using a multi-echo pulse sequence.

  As an imaging method using an MRI apparatus, there is a multi-echo imaging method such as a single-shot or multi-shot echo planar imaging (EPI) method that acquires an echo signal necessary for image reconstruction by one or several RF irradiations. is there. In multi-echo imaging, multiple echo signals are acquired in time series while inverting the frequency encoding gradient magnetic field, so in the measurement space (k space), even-numbered echoes (hereinafter referred to as even-numbered echoes) and odd-numbered ones are acquired. The direction of the frequency direction is arranged opposite to the echo acquired in (1) (hereinafter referred to as odd-numbered echo). Normally, when sampling each echo to make time-series data, the sampling time is determined so that the center and the peak of the echo coincide with each other, but the local inhomogeneity of the static magnetic field and the incomplete gradient magnetic field are incomplete. If there is a characteristic, the center of the sampling time and the peak of the echo do not match. In that case, the peak of the echo in the even-numbered echo and the odd-numbered echo is located on the opposite side of the frequency encoding 0 axis (ky axis) in the measurement space.

  Such a peak shift between the even echo and the odd echo causes an artifact called an N / 2 artifact.

  The peak shift between the even echo and the odd echo should appear symmetrically about the ky axis, but when a phase change occurs in the phase encoding direction in the measurement space, that is, an echo train (within one excitation) There may be a case where the phase difference differs depending on the eddy current of the gradient magnetic field in a plurality of echoes measured continuously. Further, when the offset in the phase encoding direction is not optimally adjusted, a noise-like change may be mixed in the phase of correction data obtained by one-dimensional Fourier transform of the correction echo signal. Streak artifacts occur in an image corrected using such correction data. Further, since a secondary change is mixed in the phase of the correction data, distortion may occur in the corrected image.

  A method for performing N / 2 artifact correction including such phase change in the phase encoding direction and phase noise mixed secondarily has been proposed (Patent Document 1). In this method, the number of pre-measurement echoes corresponding to this measurement echo is obtained, the phase change (phase difference) is calculated for each pre-measurement echo that is temporally adjacent, and the phase change in the phase encoding direction of this measurement echo is calculated. Create a reflected correction phase map. At this time, the phase difference is averaged in order to increase the accuracy of the phase difference.

By correcting this measurement echo using such a correction phase map, N / 2 artifacts can be removed with high accuracy.
JP 2001-112735 A

  By using the method described in Patent Document 1, N / 2 artifacts including phase errors in the phase encoding direction can be removed with high accuracy. That is, the phase difference between the odd-numbered echo and the even-numbered echo is obtained, but the change in the phase difference between the odd-numbered echoes or between the even-numbered echoes is not taken into consideration. Therefore, accurate correction cannot be performed for image distortion caused by a phase change when the offset in the phase encoding direction varies with time.

  In addition, the phase noise that is secondarily mixed in the correction data is averaged, but when the signal intensity at the position in the frequency encoding direction becomes low, the phase information at that position is sufficiently obtained. Therefore, accurate phase noise removal cannot be performed. For this reason, even when correcting the change in the phase difference between the odd and even echoes, the peak position deviation cannot be corrected sufficiently. When such pre-measurement data is used to correct the echo peak position as it is, streak artifacts occur at the frequency position.

  Therefore, the present invention suppresses N / 2 artifacts accurately even when there is a time variation in the offset in the phase encoding direction, and an echo signal having a low signal intensity at a specific position in the frequency encoding direction in the correction data. It is an object of the present invention to suppress streak artifacts with high accuracy even when there is a mixture.

  In order to solve the above-described problem, the MRI apparatus of the present invention corrects the main measurement echo using the pre-measurement data measured without applying the phase encode gradient pulse, and the echo (odd number) acquired the odd-numbered pre-measurement echo. Echo) and even-numbered echoes (even-numbered echoes), and the phase difference in the frequency direction is fitted with a straight line for the phase difference between odd-numbered echoes, between even-numbered echoes, and between odd-numbered and even-numbered echoes. Is used to accumulate phase changes in the echo train direction and smoothly connect the phase changes in the echo train direction. By subtracting the phase of the corresponding previous measurement data from the main measurement echo using the corrected previous measurement data, the peak position in the kx direction of the main measurement echo is moved to the center.

  That is, the MRI apparatus of the present invention comprises a static magnetic field generating means for generating a static magnetic field, a high frequency pulse irradiating means for irradiating a subject placed in a static magnetic field generated by the static magnetic field generating means, and a frequency direction And a gradient magnetic field applying means for applying a gradient magnetic field pulse in the phase encoding direction, a receiving means for detecting an echo signal generated from the subject, a signal processing means for processing the echo signal and creating an image, and once A multi-echo pulse sequence for obtaining a plurality of echo signals by applying a plurality of gradient magnetic field pulses after irradiation of the high-frequency pulse, and in accordance with the multi-echo pulse sequence, high-frequency pulse irradiation means, gradient magnetic field applying means, receiving means and signal processing means Control means for controlling the main measurement for image acquisition and a gradient magnetic field in the phase encoding direction. Except for not applying a pulse, pre-measurement using the same pulse sequence as the main measurement is performed, and the signal processing means is acquired in the main measurement using pre-measurement data including a plurality of echoes acquired in the previous measurement. Correction means for correcting the main measurement echo, wherein the correction means separates the previous measurement data into odd-numbered echoes and even-numbered echoes, and between even-numbered echoes, even-numbered echoes. The phase difference is calculated for each of the acquired echoes and between the odd-numbered acquired echo and the even-numbered acquired echo, and the phase difference is fitted in the frequency direction and then accumulated in the echo train direction. Pre-measurement data correction means for correcting measurement data is provided, and the main measurement echo is corrected using the pre-measurement data after correction. .

  In the MRI apparatus of the present invention, it is preferable that the pre-measurement data is acquired in a neighboring slice among a plurality of slices, which is performed under the shortest TE imaging condition in which the time from the radio frequency pulse irradiation to the echo acquisition is set to the shortest. The signal strength of the pre-measurement data is increased by synthesizing the pre-measurement echoes. Further, the correction means performs frequency direction fitting using only echoes whose signal value is equal to or greater than a preset threshold value, and increases the accuracy of subsequent phase accumulation.

  According to the MRI apparatus of the present invention, the phase error included in the previous measurement data is corrected by dividing the previous measurement data into odd-numbered echoes and even-numbered echoes and performing frequency direction fitting and echo train direction phase accumulation. The influence on the result can be eliminated, and an image with reduced artifacts can be obtained.

Hereinafter, preferred embodiments of the MRI apparatus of the present invention will be described with reference to the accompanying drawings.
FIG. 1 is a block diagram showing the overall configuration of an MRI apparatus to which the present invention is applied. This MRI apparatus includes a static magnetic field generation system 11, a gradient magnetic field generation system 12, a transmission system 13, a reception system 14, a sequencer 15, a reconstruction calculation unit (signal processing system) 16, a control system 17, and a display unit (display) 18. And an operation unit (not shown).

  The static magnetic field generation system 11 is composed of a permanent magnet, a normal conducting magnet, or a superconducting magnet, and generates a uniform static magnetic field in the space where the subject 10 is placed. The direction of the static magnetic field is usually the body axis direction of the subject 10 or the direction orthogonal to the body axis of the subject 10.

  The gradient magnetic field generation system 12 includes a gradient magnetic field coil 121 wound in three axial directions of X, Y, and Z, and a gradient magnetic field power source 122 that magnetizes each of these coils. By magnetizing each coil of the power supply 122, gradient magnetic fields Gs, Gp, and Gf in three axial directions of X, Y, and Z are applied to the subject 10. The imaging cross section of the subject 10 is set depending on how the gradient magnetic field is applied.

  The transmission system 13 includes a high-frequency oscillator 131, a modulator 132, a high-frequency amplifier 133, and a high-frequency irradiation coil 134, and performs nuclear magnetic resonance on atomic nuclei constituting the cross section of the subject 10 set in the gradient magnetic field generation system 12. In order to cause this, the RF pulse output from the high-frequency oscillator 131 is amplified by the high-frequency amplifier 133 and then supplied to the high-frequency irradiation coil 134 installed in the vicinity of the subject 10 to irradiate the subject 10.

  The receiving system 14 includes a high-frequency receiving coil 141, a receiving circuit 142, and an analog / digital (hereinafter referred to as “A / D”) converter 143. The receiving system 14 includes an electromagnetic wave irradiated from the high-frequency irradiation coil 134 of the transmitting system 13. An echo signal, which is an NMR signal by nuclear magnetic resonance of an atomic nucleus, is detected by a high-frequency receiving coil 141 disposed in the vicinity of the subject 10 and input to an A / D converter 143 via a receiving circuit 142, and a digital signal And is sent to the reconstruction calculation unit 16 as data sampled at a timing according to a command from the sequencer 15 (hereinafter referred to as “sampling data”).

The control system 17 sends a command to the sequencer 15 and controls the gradient magnetic field generation system 12, the transmission system 13 and the reception system 14 according to a predetermined pulse sequence set in the sequencer, and the operation of the reconstruction calculation unit 16 To control.
The reconstruction calculation unit 16 performs various calculations such as Fourier transform on the sampling data, and calculates a phase change and correction data, which will be described later, and correction using the correction data. The image reconstructed by the reconstruction computation unit 16 is displayed on a display unit (display) 18.

  Next, the operation of the MRI apparatus having the above configuration will be described. FIG. 2 shows an operation procedure.

  The MRI apparatus of the present embodiment executes a pulse sequence by multi-echo imaging as the main measurement (step 210) for measuring the echo signal to be imaged, and the echo signal (main measurement echo) 20 obtained by the main measurement is obtained. Then, correction is performed using the echo signal (pre-measurement echo) 30 obtained by the pre-measurement (step 230) performed separately from the main measurement (step 220), and imaged (step 225). The main measurement echo 20 is corrected by subjecting the main measurement echo 20 and the previous measurement echo 30 to Fourier transform (steps 215 and 235) in the frequency direction to obtain x-ky data and making a difference.

  When the pre-measurement echo 30 is subjected to correction as correction data, it is divided into odd-numbered echoes 31 and even-numbered echoes 32, the respective differences (phase differences) are fitted in the frequency direction, and the phases are accumulated in the echo train direction (steps 240 and 250). Thus, phase noise secondarily mixed in the previous measurement echo is removed, and correction data reflecting a phase change including a time variation is obtained. In addition, although the measurement for acquiring the correction data is referred to as “pre-measurement”, the timing for executing the pre-measurement may be before or after the main measurement.

Details of each step shown in FIG. 2 will be described below.
<Steps 210 and 230>
In the main measurement (step 210) and the previous measurement (step 230), a plurality of echo signals are measured with one excitation by a single shot or multi-shot EPI sequence. 3 and 4 show an example of a pulse sequence used in the main measurement and the previous measurement. These pulse sequences are known GrE EPI sequences and will be described briefly.

  First, an RF pulse 301 is irradiated to a subject including magnetization to be detected, and simultaneously, a gradient magnetic field pulse 302 for selecting a slice is applied to select a slice to be imaged. Next, after applying a pulse 304 that gives an offset of the frequency encode gradient magnetic field, a frequency encode gradient magnetic field pulse 306 that is continuously inverted is applied, and the echo signals 120 and 130 are generated within each period of the frequency encode gradient magnetic field 306 to be inverted. Since it is generated in series, this is sampled for each time range 307 to obtain time series data. The intensity, application timing, and time range 307 of the frequency encoding gradient magnetic field 306 to be reversed are the same as those in the main measurement sequence and the previous measurement. However, in this measurement sequence shown in FIG. 3, gradient magnetic field pulses 303 and 305 for phase encoding are applied to each echo signal 120, whereas in the previous measurement sequence shown in FIG. 4, phase encoding is not used.

  In the case of a single shot sequence, the repeating unit 308 shown in the figure is performed once, and in the case of a multi-shot sequence, the repeating unit 308 is repeated a plurality of times to obtain data that respectively fills the entire k space. A plurality of echoes acquired in one repeat unit is called an echo train, an odd-numbered echo is called an odd-numbered echo, and an even-numbered echo is called an even-numbered echo.

  Usually, such measurement is performed by changing the slice gradient magnetic field 302 to acquire data of a plurality of slices. Although the number of slices is not particularly limited, in the present embodiment, it is desirable to acquire data of a plurality of slices at least in the previous measurement. By using data of a plurality of slices, correction data with higher accuracy can be created.

  In the pre-measurement, the echo time is set so that the time (T) from the RF pulse 301 to the measurement of the echo signal with the maximum signal intensity (referred to as the maximum echo) is the shortest. The maximum echo is an echo that occurs at TE (echo time) in the case of an SE-based EPI sequence, and is the first echo in the case of a GE-based EPI sequence. In general, in the EPI sequence, as shown in FIG. 5, the signal intensity decreases due to T2 or T2 * attenuation as time elapses after irradiation with the RF pulse, but the time (501) until the maximum echo is obtained should be minimized. Thus, the echo 503 can be acquired with a higher signal intensity in the previous measurement set to the shortest TE compared to the echo 504 obtained when the measurement is performed in the normal TE setting time (502).

<Frequency direction Fourier transform steps 215 and 235>
The echo signals 20 and 30 obtained in the main measurement and the previous measurement are each Fourier-transformed in the frequency direction to obtain x-ky data. This is because the peak shift of the echo signal in the measurement space corresponds to a phase change in the space after the Fourier transform (x-ky space), and thus the data is corrected by phase correction.

Here, errors included in the main measurement echo will be described with reference to FIGS.
As already described, in this measurement echo, when the peak position of the echo does not coincide with the center of the sampling time 307, the peak position of the odd echo and the peak position of the even echo in the k space are respectively Shifts in the opposite direction. This is shown in FIG. In the k space 600 shown in FIG. 6A, the center dotted line is the axis 610 of frequency encode 0, and odd-numbered echoes and even-numbered echoes are arranged in order from the top. The odd-echo peaks 621, 622,... Are shifted to the right with respect to the axis 610, and the even-echo peaks 631, 632,. When such a main measurement echo is reconstructed as it is, an N / 2 artifact as shown in FIG. 6B is obtained.

  Further, when there is an error in the offset in the phase encoding direction, as shown in FIG. 7A, the deviation of the echo peak from the axis 610 increases as the echo number increases (that is, the echo train advances). As shown in FIG. 7B, when an offset error 702 in the phase encode direction exists in the frequency encode gradient magnetic field pulse (dotted line) 701, the peak position of the echo corresponds to the offset error in the positive direction and the negative direction. This is because the minute 703 shifts in the opposite direction. That is, the positive and negative frequency encoding amounts are biased, and as a result, the amount of frequency encoding that cannot be canceled is accumulated as the echo train progresses, and the peak moves away from the center of the measurement space.

  When the error in the application timing of the frequency encoding gradient magnetic field and the sampling time and the offset error in the phase encoding direction are combined, the peak position changes between echo trains as shown in FIG. In the ky space, even at the same frequency position, the phase is disturbed depending on the position of the echo train. If image reconstruction is performed as it is, image distortion and N / 2 artifacts occur as shown in FIG.

  The error included in the main measurement echo described above is corrected by moving around the peak position of the main measurement data in the correction step.

<Correction step>
In the correction, (1) the peak position in the frequency direction of the measurement data appears at a position equal to the peak position of the previous measurement data, and (2) the peak deviation of the echo signal in the measurement space is the space after Fourier transform ( This is performed by subtracting the phase of the previous measurement data from the main measurement data in the x-ky space by utilizing the fact that the phase changes in the x-ky space) (FIG. 2: step 220).

  However, when the signal intensity at the frequency position of the previous measurement data is low and phase disturbance occurs, if correction is performed using the data, N / 2 artifacts are generated at the frequency position. Therefore, in the present embodiment, the phase disturbance at the frequency position of the previous measurement data is corrected (steps 236, 240, 250) before the x-ky correction process (step 220).

<< Signal Synthesis Step 236 >>
In this step, in order to increase the signal intensity of the previous measurement data, the signal of the neighboring slice is synthesized with respect to the previous measurement echo signal 30 after Fourier transform in the frequency direction. This utilizes the fact that there is little phase change due to the physical properties of the imaging target between neighboring slices.

  The number of slices to be combined (number of combined images) is not particularly limited, but is about 5 for example. When adjacent slices such as the first and last slices of all slices are insufficient with respect to the number of composites, the shortage is left as it is and the composition process is performed. FIG. 9 shows a combination of slices used for signal synthesis when the number of synthesis is five. As shown in the figure, when the first slice is the first slice, the first through third slices are used, and when the last slice (the eighth slice in the figure) is the sixth slice, the eighth slice is used. Perform synthesis using the eyes.

  However, there is a phase difference between the slices due to the RF pulse and the slice selection magnetic field, and it cannot be synthesized as it is between neighboring slices. Therefore, the signal is synthesized after removing the phase difference between neighboring multi-slices. Specifically, a phase difference is made between the reference echo of the slice to be synthesized and the reference echo of each slice synthesized therewith, and the result is phase-differenced to all echoes of each slice. An echo having a maximum signal value is used as the reference echo. When the pre-measurement data of the slice z imaged under the shortest TE measurement condition is S (x, B, z), the pre-measurement after multi-slice synthesis is performed by calculating the equations (1) and (2). Data S ′ (x, ky, z) is calculated (step 236).

In the equation, S _Dif (x, B, z, i) is a phase difference in the reference echo B (B-th echo) between the slice z to be synthesized and each neighboring slice i, and n is the number of slices to be synthesized. With respect to the combined signal value, the signal value is divided by the combined number n to match the signal value with each neighboring slice.

  In this way, the pre-measurement data after synthesizing the neighboring slices is separated into odd-numbered echoes 31 and even-numbered echoes 32, and frequency fitting and echo train direction accumulation described below are performed for each.

<< Steps 241 to 243 >>
Steps 241 to 243 are performed in order to correct the phase disturbance at the frequency position for the previous measurement data. These processes are performed between odd-numbered echoes and even-numbered echoes. However, when correction is performed between odd-numbered echoes and between even-numbered echoes, an error between the application timing of the frequency encoding gradient magnetic field and the sampling time (the original cause of N / 2 artifacts) remains uncorrected. Therefore, one of the odd-numbered echo and the even-numbered echo (even-numbered echo 32 in the illustrated example) is phase-differed with the other (odd-numbered echo 31) (step 251), and the same processing is performed on the difference data 33. The details of steps 241 to 243 will be described below by taking the case of an odd-numbered echo as an example.

  In steps 241 to 243, the phase difference data between odd-numbered echoes is fitted with a straight line in the frequency direction (step 242), and the phase change is accumulated in the echo train direction using the fitting result (step 243). Prior to frequency direction fitting, threshold processing is performed on echoes used for the fitting (step 241). Step 242 of the frequency direction fitting is a correction of phase accumulation when phase disturbance in the frequency position occurs in the reference echo that is a reference for phase accumulation in the echo train direction in the next phase accumulation in the echo train direction (step 243). Since the accuracy is lowered, it is necessary to carry out prior to the phase accumulation.

The frequency direction fitting (step 242) is for smoothing the phase change in the frequency direction with respect to the phase change between odd-numbered echoes or even-numbered echoes. Among these, linear fitting is performed by the least square method using data for a predetermined number (m) of echoes from the reference echo B. At this time, in order not to use a signal having a low signal strength for the frequency direction fitting, signals having a ratio equal to or less than the set ratio are excluded from the maximum signal strength (step 241). The threshold value Th for exclusion is set by the following equation (3).
In the equation, the maximum signal intensity in the fitting target echo is MaxSig, and the threshold setting parameter is ThRatio.

  When the phase difference between odd-numbered echoes or even-numbered echoes is taken, the error between the application timing of the frequency encoding gradient magnetic field and the sampling time is canceled, so only the offset error in the phase encoding direction remains, and the phase change is in the frequency direction. It becomes a straight line. FIG. 10 shows the relationship between the shift of the peak of the echo signal from the center of the measurement space and the phase change. As shown in the figure, when the peak of the echo signal 711 is at the center 710 of the measurement space, the phase change in the space obtained by one-dimensional Fourier transform of the echo is 0 as shown by the straight line 721, but the echo signal 712 When the peak deviates from the center 710 (deviation amount 703), the phase changes linearly as indicated by a straight line 722 in a space where the echo signal is one-dimensional Fourier transformed. In the frequency direction fitting in step 242, the phase difference between odd-numbered echoes is fitted to this straight line 722 to calculate the slope and intercept of the phase change. Specifically, as described above, the least square method is used by using data for a plurality of echoes (m) from the reference echo B. A specific calculation method is represented by the following equation (4).

In Expression (4), Dif _O is a phase change between odd-numbered echoes, and a and b are the average of slope and intercept (average of m pieces) obtained by the method of least squares.

Next, an average of the slope and intercept obtained by the above equation (4) is accumulated in the ecotrain direction, that is, the phase encoding direction, and the phase disturbance of the frequency position in the echo train direction is corrected. FIG. 11 shows the phase accumulation. FIG. 11 shows the peak positions in the measurement space of the pre-measurement echoes (odd echoes and even-number echoes). As shown in FIG. 11A, the peak positions of the pre-measurement echoes are shifted due to the phase disturbance 1101. There is an echo. In the phase accumulation step 243, the phase is accumulated in the echo train direction (1102, 1103) using the phase change (Dif _O ) calculated using the phase of the echo having a high signal intensity in the frequency fitting step 242 (1102, 1103), Complements the phase information of echoes with low signal strength. As a result, the phase disturbance 1101 is corrected. A specific calculation method is represented by the following equation (5).

P ′ (x, 2 × i + 1) in Expression (5) is a phase change of the (2 × i + 1) th pre-measurement echo after the phase accumulation, and represents the odd and even echoes together. Is.

  Thus, by performing the processing from step 241 to step 243 on the previous measurement data between the odd-numbered echoes, the phase disturbance at the frequency position can be corrected.

  For pre-measurement data between even-numbered echoes, as described above, the same processing from step 241 to step 243 is performed on the data after the phase difference between even-numbered and odd-numbered echoes. Simultaneously with the correction of the phase disturbance, the error between the application timing of the frequency encoding gradient magnetic field and the sampling time is corrected.

<< Step 252 >>
At the end of the correction step, the phase change between the odd echoes corrected in steps 241 to 243 is added to the phase difference data between the even echoes and odd echoes corrected in steps 241 to 243 (step 252). Specifically, the following equations (6) and (7) are calculated.
In equation (6), Dif _OE is the phase change between even and odd echoes, O (x, B _o ) and E (x, B _e ) are the phase of the reference echo in the odd echo and the reference in the even echo, respectively. The phase of the echo, [O (x, B _o ) −E (x, B _e )] represents the phase difference result with respect to the reference echo. c and d are a coefficient and an intercept obtained by frequency direction fitting, and are equivalent to a result obtained by adding a fitting result with respect to a phase difference of odd-numbered echoes and a fitting result with respect to a phase difference of even-numbered echoes. In addition, P ′ (x, ky) in the equation (7) is a phase change of the previous measurement data after the phase accumulation.

  As a result, the processing from step 241 to step 243 is equivalent to the processing performed between even-numbered echoes.

  Thus, the phase of the main measurement echo 20 is corrected using the previous measurement data corrected by the correction processing 240 and 250 of the previous measurement data (step 220). As shown in FIG. 12, by subtracting the phase of the previous measurement data (b) from the main measurement data (a), the centers of the peaks of the echoes are aligned in the entire measurement space (c). Finally, the corrected main measurement data (c) is Fourier transformed in the phase direction to perform image reconstruction (step 225). In the image 40 created by the above processing, N / 2 artifacts are greatly reduced.

  According to the present embodiment, the streak artifact caused by the phase disturbance included in the previous measurement data can be significantly reduced by performing the frequency direction fitting on the previous measurement data. In particular, the streak artifact can be eliminated by adding a process for increasing the signal intensity, such as obtaining the pre-measurement data using the shortest TE imaging, or performing signal synthesis between neighboring slices.

  FIG. 13 and FIG. 14 show the results of confirming the effect of this embodiment by imaging using a phantom. FIG. 13 shows an example of an image in which streak artifact is suppressed, and FIG. 14 shows an example of an image in which distortion is suppressed. In the figure, (a) shows the case where no correction is performed, (b) shows the case where only the x-ky correction of step 220 is performed, and (c) shows the case where the correction according to the present embodiment is performed.

Comparing (a) and (b) with respect to FIG. 13, it can be seen that the N / 2 artifact generated in the uncorrected image is greatly reduced by the x-ky correction. In the image of (b), Streak artifacts still remain. In the image of (c), it can be seen that the artifact is completely eliminated by performing the processing of the present embodiment.
Also in FIG. 14, it can be seen that artifacts that are not corrected only by the x-ky correction are greatly reduced.

  The embodiments of the present invention have been described above. In this embodiment, the odd inter-echo data and the phase difference data between the even echo and the odd echo are respectively fitted and phase accumulated, and the phase of the data between the odd echoes is added to the phase difference data between the even echo and the odd echo. Addition to correct even-numbered echoes, but fitting and phase accumulation are performed for the data between even-numbered echoes and the phase difference data between odd-numbered and even-numbered echoes, and the phase difference between even-numbered and odd-numbered echoes. The phase of the data between even echoes may be added to the data to obtain a corrected odd echo. Further, the processing of the present invention can be applied not only to multi-shots, but also to one-shot type sequences, 3D measurement, and the like.

The block diagram which shows the whole structure of the MRI apparatus to which this invention is applied Flow chart showing one embodiment of operation of MRI apparatus Diagram showing an example of the pulse sequence used in this measurement Diagram showing an example of pulse sequence used in pre-measurement The figure explaining the shortest TE imaging condition in the pre-measurement Diagram explaining N / 2 artifact Diagram explaining offset error in phase encoding direction The figure explaining the artifact when the deviation between the frequency encoding gradient magnetic field pulse and the sampling time and the offset error in the phase encoding direction are combined. Diagram explaining signal synthesis step The figure explaining the phase change fitting in the frequency direction Diagram explaining phase accumulation in echo train direction Diagram explaining correction of main measurement data using previous measurement data The figure which shows the example of an image to which embodiment of this invention is applied The figure which shows the example of an image to which embodiment of this invention is applied

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 ... Static magnetic field generation system, 12 ... Gradient magnetic field generation system, 13 ... Transmission system, 14 ... Reception system, 15 ... Sequencer, 16 ... Reconstruction calculating part (Signal processing system) ), 17... Control system, 18... Display unit (display).

Claims (6)

Static magnetic field generating means for generating a static magnetic field, high-frequency pulse irradiating means for irradiating a subject placed in the static magnetic field generated by the static magnetic field generating means, and gradient magnetic field pulses in the frequency direction and the phase encoding direction Gradient magnetic field applying means for applying a signal, receiving means for detecting an echo signal generated from the subject, signal processing means for processing the echo signal and creating an image, and a plurality of gradients after one high-frequency pulse irradiation A multi-echo pulse sequence for obtaining a plurality of echo signals by applying a magnetic field pulse, and a control unit for controlling a high-frequency pulse irradiation unit, a gradient magnetic field application unit, a reception unit and a signal processing unit according to the multi-echo pulse sequence A magnetic resonance imaging apparatus,
The control means performs main measurement for image acquisition and pre-measurement using the same pulse sequence as the main measurement except that a gradient magnetic field pulse in the phase encoding direction is not applied,
The signal processing means includes a correcting means for correcting the main measurement echo acquired in the main measurement using pre-measurement data including a plurality of echoes acquired in the previous measurement,
The correction unit separates the previous measurement data into odd-numbered echoes and even-numbered echoes, and between odd-numbered echoes, between even-numbered echoes and odd-numbered echoes, A pre-measurement data correction unit that calculates a phase difference for each of even-numbered echoes, fits the phase difference in the frequency direction, accumulates in the echo train direction, and corrects the pre-measurement data. ,
A magnetic resonance imaging apparatus, wherein the main measurement echo is corrected using pre-measurement data after correction. The magnetic resonance imaging apparatus according to claim 1.
The correction means includes Fourier transform means for Fourier transforming the echo signal in the frequency direction to create x-ky space data,
A magnetic resonance imaging apparatus, wherein the phase difference calculation, fitting, accumulation, and correction of the main measurement echo are performed on the x-ky space data. The magnetic resonance imaging apparatus according to claim 1.
The control means measures the pre-measurement echo for a plurality of slices,
The magnetic resonance imaging apparatus according to claim 1, wherein the pre-measurement echo correcting unit performs phase difference after synthesizing pre-measurement echoes obtained in neighboring slices among the plurality of slices. The magnetic resonance imaging apparatus according to claim 1.
The magnetic resonance imaging apparatus according to claim 1, wherein the pre-measurement echo correcting means performs frequency direction fitting using only the echo whose signal value is not less than a preset threshold value among the pre-measurement data. The magnetic resonance imaging apparatus according to claim 1.
In the pre-measurement, the control means sets the time from radio frequency pulse irradiation to echo acquisition to the shortest, and acquires pre-measurement data. The magnetic resonance imaging apparatus according to claim 2.
The correction means, based on the information of the phase change corresponding to the main measurement echo signal arranged in time series on the measurement space, the previous measurement echo signal of the even number acquisition from the phase of the even measurement main measurement echo signal And subtracting the phase of the previous measurement echo signal acquired in odd number from the phase of the main measurement echo signal acquired in odd number.