JPS6264345A - Nuclear magnetic resonance image pickup apparatus - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

Detailed Description of the Invention (Industrial Application Field) The present invention relates to a nuclear magnetic resonance imaging apparatus (hereinafter referred to as nuclear magnetic resonance).
In particular, the present invention relates to improvements in means for obtaining calculation images of relaxation time T++T2 and proton density ρ.

(Prior Art) Conventionally, images measured by NMRfi imaging devices have been considered to be medically useful! There are techniques for obtaining images related to relaxation time Tt values (T+brass) and images related to transverse relaxation time T2 values (T2 images).

(Problems to be Solved by the Invention) However, the T1 image and the T2 image have been obtained by different methods as follows.

■T+(i is calculated as follows, for example. Inversion recovery method (l nvers
The IR8E method is a combination of the ion recov cry method (hereinafter abbreviated as IR method) and the spin echo method (abbreviated as spin echo method 2 or less SE method), and the saturation recovery method (hereinafter abbreviated as IR method) as shown in Figure 6.
Covcry (hereinafter abbreviated as SR method) and SE method are used to obtain one original image each, and calculation is performed using these two images and an approximate expression for signal strength.

As shown in Figure 6, the 5R8E method is a pulse sequence in which an echo signal is obtained by applying a 180° pulse after applying a 90' pulse, and the time from the 90' pulse to the center of the echo signal is Ts, 90" The time from pulse application to 90' pulse application in the next view is defined as Tr.

In addition, as shown in Fig. 5, the IR8E method uses 5R8 in Fig. 6.
In the pulse sequence in which 1808 pulses for inversion/recovery are applied before each 90° pulse of E method, the time from application of 180° pulse for inversion/recovery to application of 90° pulse is Ta. , 90
The time from the ' pulse to the center of the echo signal is defined as the time from the application of the 180' pulse for TsN impurity recovery to the application of the 1806 pulse in the next view.

Dim-eXp (-Thi/Ding, )) It is.

This I! Il! For the theoretical formula, if Tr>T+ and exp (-Tr/T+)-0, then the engineering name is 10.・4) tP(-τs/Tλ)IH: ]-0・-etP
(-Ts/ TL)l l-2・e-Xi'(-T, 7
．． /T+)1, therefore FL/I = 1-1-θP(-74/TOT, =
T shim 12 engineering whale / (cho 5 to 1 ee R) 1 From this formula T
Find the 1 value.

■If you want to obtain a T2 image, please refer to the publication ``Video Information (
M) 4 June 1984 (VOl, 16 No.

CP described on pages 570 to 576 of 11)
TI+ρ is removed from a plurality of echo data using the MG method, and the T2 value is determined using the least squares method.

Note that in the CP method, which allows multiple pieces of echo data to be taken out continuously in one data collection to obtain T2(a), if the length of the applied pulse is incomplete, the error will increase as the echoes are obtained. However, a pulse sequence called the CPMG method has solved this problem, and as shown in Figure 7, 180" pulses are applied n times in 900 pulses. This is a pulse sequence that is repeatedly applied to generate n echoes.

Conventional methods using such techniques have the following drawbacks.

(1) The T+ image and the T2 image are obtained separately, and the calculated image of T+ N T2 Nρ cannot be obtained at the same time.

(2) An accurate value cannot be determined because an approximate formula is used.

(3) Due to the condition Tr>T, Tr must be changed, and the total scan time is long.

(4) The scan parameters of the original image are not optimized, and the best calculated image cannot be found under the given conditions.

(5) There is a systematic error due to the influence of the slice shape.

In view of these points, it is an object of the present invention to obtain highly accurate T,* from a plurality of original images in a short time in an NMR prefectural imaging device.
NMR that can simultaneously obtain T 2 eρ calculation images
An object of the present invention is to provide an imaging device.

(Means for solving the problem) In order to achieve such an objective, in this research, we devised a new pulse sequence for calculation images in nuclear magnetic resonance l1i)e equipment, and applied it to the IRnSE method and FR imaging E Law (FR
is Fast Recovery method (7) 18) k:J
: Take an image, obtain (n0m) images from this image,
A T I * T 2 gρ image is calculated from this image, and each scan parameter in the IRnSE method and FRmSE method in this case is T 1 in the range of interest.
+ T 21ρ It is characterized in that the evaluation function of the calculated image is selected to be the best.

<Example> The present invention will be described in detail below using the drawings. FIG. 1 is a block diagram of essential parts of an embodiment of an NMR1i imager according to the present invention. In the figure, 1 is a magnet assembly;
A space (hole) for inserting the object is provided inside, and a main magnetic field coil 2 surrounds this space to apply a uniform static magnetic field Ho to the object, and generates a gradient magnetic field. Gradient magnetic field coil 3 (X gradient magnetic field coil configured to be able to individually generate a gradient magnetic field,
y slope! (consisting of a 1s coil and 2 gradient magnetic field coils), an RF transmitter coil 4 that provides a high-frequency pulse to excite the spin of the atomic nucleus within the object, and an
A receiving coil 5W- for detecting MR signals is arranged.

The main magnetic field coil is a static magnetic field 11. Ij 'wJ circuit 15,
Gx.

Each Gy and Gz gradient magnetic field coil is connected to a gradient magnetic field dust control circuit 14.
Then, the RF transmitter coil is connected to the power amplifier 18, and the NMR
The signal receiving coils are each connected to a preamplifier 19.

A controller 13 controls the sequence of generation of gradient magnetic fields and high-frequency magnetic fields, and performs necessary control to capture the obtained NMR signal into the waveform memory 21.

17 is a gate modulation circuit, and 16 is a high frequency oscillator that generates a high frequency signal. The gate modulation circuit 17 appropriately modulates the high frequency signal output from the high frequency oscillator 16 using a control signal from the controller 13 to generate a high frequency pulse with a predetermined phase. This high frequency pulse is RFit force amplification @1
8 to the RF transmitting coil 4.

One is a preamplifier that amplifies the NMR signal obtained from the detection coil 5, 20 is a phase detection circuit that phase-detects the NMR signal by referring to the output signal of the high-frequency oscillator, and 21 stores the phase-detected waveform signal from the preamplifier. A waveform memory containing an A/D converter.

A computer 11 receives the signal from the waveform memory 21 and performs predetermined signal processing to obtain an l1liI image, and a display 12 such as a television monitor displays the obtained tomographic image.

Reference numeral 30 denotes an operation console, which is connected to the computer 11 and is an input means for inputting various information necessary for the apparatus.

Next, the procedure for creating a calculated image in such a configuration will be explained.

Here, the novel IRnSE method according to the present invention and FRmS
An example will be explained in which the TINT2 and ρ images are calculated from a total of seven images, three images obtained by the IR3SE method and four images obtained by the FR4SE method, using the pulse sequence of the E method.

Note that the lRa5E method uses IR8E as shown in Figure 2.
This is a pulse sequence based on the law, but in one view, 180''' (2) pulses are repeatedly applied three times to obtain three echo signals.

Furthermore, the FR4SE method is a method that employs a pulse sequence as shown in FIG. That is, after obtaining the last (in this case, the fourth) echo signal in the 5R8E method, 1800 pulses are applied, and then 90 pulses are applied.
A multi-echo method that obtains echo signals using a pulse sequence that forcibly directs magnetization toward the main magnetic field direction by applying ° pulses and 180 ° pulses to shorten the time for magnetization to recover to an equilibrium state. be.

(1) Pulse Sequence The pulse sequences shown in FIGS. 2 and 3 will be explained in more detail as follows.

Distance between the 90° pulse and the center of the first echo signal T B
I, s2 below the center interval of each echo signal after the first echo signal, 9 given to the center of the fourth echo signal and after it
Interval to 0° pulse T B 3 , first 180° pulse (180° pulse for inversion recovery)
The interval Td from to the application of 901 pulses, IT at the time of repetition
Each digit can be selected arbitrarily. These time managements are performed by the controller 13, and the time settings can be performed using the operating vehicle 30.

180° [1 pulse is a 180° pulse for spin inversion, 1'80° (ko) is a 180@ pulse for spin echo, and both are 90 ''-4 to reduce pulse error.
5. Composite pulse with 270° start and 90″ is used.

In the lRa5E method, the number of 180° pulses for each view is an even number.

Note that the suffix value attached to the frequency of each pulse represents the phase difference with the 90° excitation pulse, and these pulses are non-selected pulses.

The excitation pulse 90' is a selection pulse and is Gaussian modulated.

Application of such 90° pulses to 1800 pulses is performed as follows. In other words, controller 13
A predetermined 90° pulse or 180° pulse signal obtained through the gate modulation circuit 17 under the control of the power amplifier 18
to the RF transmitting coil 4 and applied to the target object.
Generates an F magnetic field.

On the other hand, the gradient magnetic field is as follows. The gradient magnetic field Gx in the X direction is a projection gradient, and a is 180°
This is a spoiler for eliminating noise outside the slice plane due to pulses.

The two-direction gradient magnetic field Gz is a slice gradient, and the y-direction gradient magnetic field G
y is a warb slope, and b is a spoiler for eliminating artifacts due to 180° pulse error. Also,
C is a spoiler to remove correlation between views.

The application of each gradient m field is controlled by a controller 13.

Each echo signal generated by the above pulse sequence is detected by the receiving coil 5. The spin echo signal detected by the receiving coil is stored in a waveform memory 21 via a preamplifier 191 and a phase detection circuit 20.

(2) About the signal strength formula 13 Tz/2T+ ) -2-e
x P (-Ti-7 to 10T5+/ton+ +Tsz
/Z T +) +26p' (-71, /7 (As, the first echo is 1o ・eXl] (Ts + /T2) The second echo is 1o ・eXr) (-Ts + /T2-TI2 /
T2) The third echo is 1o -eXI) (Ts + /T2-27!B 2
/T2).

Here, CTFZ3 is a function representing the influence of the slice shape, and is determined as follows.

2) Since C- is a function only of T+/Tr, the required T+
/ T r by numerical integration, and from this value C0U can be determined as a polynomial of T + / T r.

From the above, the signal strength formula F including the influence of slice shape
s (TrzTs, T+2TZ/P) is found as the product of Fn and the coefficient 00 run representing the influence of the slice shape.

Fs(T?z ding'hl'Ti/Tz/P)
-Fn ("'5-Ts, TuTg-)CotLd
(T+/Tr), where Coad
For example, 0.2<'T + / T r <10.0
In the case of C, Mu −8.1537 E −6(T+ /Tr
)'-2,95088E-4(T+/Tr)
'+4.27675E-3(T+/Tr)'
-3,17902E -2(T+/Tr)'
+ 1.29262E -j (T+ /Tr) 2
-2.85542-1 (TI/T,?)+1
．． 0557 Pulse sequence consists of one 90” pulse and an even number F n
(Tr, Ts>T1, TL) #).

B> The signal strength when the angle of magnetization is C1 is that the pulse sequence is one 90° pulse and an odd number of 180° pulses.
If it is composed of pulses, then

) If a Gaussian 90° pulse is used, α at a distance of 2 from the center of the slice is calculated as follows:
・Becomes mouth (Otsu).

c) By integrating equation (1) by 2 using equation (2), the signal intensity including the influence of the slice shape can be found, resulting in the following equation.

...(3) Since the integral in equation (3) is a function of only (T+/Tr), this value is written as C0u(T+/Tr).

If it is composed of 180° pulses of
If the signal intensity formula that does not include the slice-shaped Shosuga is Fn, then the signal intensity when the angle at which the magnetization falls is C0 is calculated as below, and can be calculated in the same manner as in the above case.

For example, if a Gaussian 90''' pulse is used, 0
．． 2<TI/T2<10.0, the coefficient C4V4N representing the influence of slice shape is C4vea--2,4203E-5(T+/Tr
)'+5.6861 E-4(T+/Tr)
'-3,6523E-3(T+/Tτ)81-1
．． 0071 E-2<T+/Tr)'+3.
2162 E-1(T+/Tr)+0.91
It is 78.

The above is the calculation when using a Gaussian 90° pulse, but the same applies when using other 90" pulses as long as the angle α at which the magnetization falls due to the 90° pulse at a point 2 from the center of the slice is found. It can be calculated as follows.

■ The signal strength formula for the FR4SE method is
-Tsi)/T+Lori/[1-σi'(-(Th-Tgi1-3cho≦2-Tst)/T+-(Ts?Tsz+"
Th3')/T4, the first echo is No -eXp (-Ts + /T2) and the second echo is Io -eXll (-Tg + /T2 Tg
2 /T2) The third echo is io -f3XI) (Ts + /T2-2T!!
2 /T2) The fourth echo is Io-eXp(-Ts + /T2 2T!$ 2 /
T2).

(3) Optimization of scan parameters for human body TIIT
21ρ Calculation Scan parameters that give the best image evaluation function are calculated using the W4 difference propagation law. The signal strength formula in (2) above is used as the signal strength formula.

Here, a method for determining scan parameters that optimize the evaluation function of the calculated image iii from the theoretical formula of the signal strength and the determined T1, Tg, and ρ values will be described. Here, the sum of standard deviations normalized as an evaluation function, that is, ση/T ++σ7z / T 2+σ4/ρ, however, 61. σT2, σ, is T I * Tg *
Use the standard deviation of I).

Scan parameters of 7 images fl/Pzz・
・°/PTs signal strength formula as F I * F 2 * −
−, r F T, then f
T+ calculated by ff.

The covariance matrix of the value of T2eρ 'T+ Tz P is
P -(A" ■I□3 A>-'However, v123
is the covariance matrix of the original image, and the variance σ2 of the original image is σ2ocn, where n is the average value and Ta is the sampling time.
-'Ta-'t' Expressed and organized.

Therefore, the variance of the value of N T I + 72 *ρ is V
It is found as the diagonal element of TITZp.

From the above, the evaluation function of the calculated image is PI /Pz t...
・f'7 /T(Ll/TlLX degree-/ding'cbqr%
It is found as a function of +i'r t + LL / LY.

Based on this principle, appropriate scan parameters are determined by the following procedure.

■Define the theoretical formula for signal strength.

(2) Find the evaluation function of the calculated image as a function of the scan parameters from the theoretical formula, the range of T+ and Tg-ρ to be measured, and the variance of the original image.

■In the above ■, the evaluation function of the calculated image was found as a multivariable function of the scan parameters, so the i!lF
Find the scan parameters that give the best value function.

An example of the scan parameters obtained in this manner is as follows.

■When scanning with a total scan time of 600 seconds, in the lRa5E method, TT-2, 1 second Td = 0.457 seconds T, , = 0.02 seconds Ts 2 = 0.02 seconds Average number of times (AVE) - 1 In the FR4SE method, Tr = 1.3 seconds Ts I = 0.02 seconds T, , = 0.073 seconds Ts 3-0.02 seconds average number of times (AVE) -2 ■If the total scan time is 300 seconds, IR3SE
In the method, Trx 1,28 seconds 7d = 0.366 seconds Ts + = 0.02 seconds Ts 2-0.02 seconds Average number of times (AVE) = 1 In the FR4SE method, Tr = 1.1 seconds Ts + = 0.02 seconds T, 2 - 0.073 seconds Ts 3 "" 0.02 seconds average number of times (AVE) -2 (4) Imaging is performed using the scan parameters in (3) above.

That is, in the IR3SE method, a warp gradient (gradient meter 11 iGy) is scanned over different predetermined views (the number of views is, for example, 127) with the above parameters, and an echo signal is measured. The measured echo signals are divided into first, second, and third echo signal groups, and each is reconstructed into two-dimensional images using the computer 11 to obtain 1q of three original images.

Next, in the FR4SE method, the warp gradient (gradient magnetic JiG
y) different predetermined views (the number of views is, for example, 1
21) with the above parameters, and similarly measure the echo signal and store it in the waveform memory 21. The obtained data is divided into the first, second, third and fourth echo signal groups, and similarly, the computer 11 is used to divide the data into two echo signal groups.
Reconstruct into dimensional images to obtain four original images.

(5) Using the seven original images obtained in (4) above,
T I T T 21 ρ (calculated image) is determined by the nonlinear least squares method (calculated by the computer 11).

In the above method, the gradient 1111Gx, Gy,
The relationship between G2, slice, projection, and warp is arbitrary.

Through the above procedure, the calculated image of TI*72*ρ can be obtained accurately and simultaneously.

(Other Examples) Note that the present invention is not limited to the above-mentioned embodiments, but can be implemented as follows.

(1) As a pulse sequence, in the embodiment, the interval from the 90° pulse to the first echo signal is T! i1, the interval between each echo signal after the first echo signal is Ts2, and the interval from the 90° pulse to the first echo signal and the interval from the first echo signal to the second echo signal are Ts, respectively.
I sThe interval from the second echo signal to the third echo signal and the interval from the third echo signal to the fourth echo signal are T
s2, from the subsequent (2n-2)th echo signal to the (2nd
n-1>interval to echo signal and interval from (2n-1)th echo signal to 2nth echo signal (however, n≧3
) is Tsn. In this case, the 180° pulse error can be removed by using an even number of echoes.

Furthermore, the interval between each echo signal may be completely arbitrary.

(2) Gaussian modulated 90° pulse, composite 1
The 80° pulse is not limited to the embodiment, and other 90° pulses and 180'' pulses may be used, and calculated images can be obtained in the same way.

(3) In the examples, the 1R380 method and the FR4SE method are used! The case where seven images are obtained from the fi image and each calculation image of T1゜T2+ρ is calculated based on these has been shown, but without being limited to this, in general, (n+m) images are obtained from m images by the IRnSE method and the FRmSE method. It is also possible to obtain an image of , and obtain each calculation image of T I + 72 *ρ based on this image.

(Effects of the Invention) As explained above, the present invention has the following effects.

■Accurate values can be obtained because no approximation formulas are used.

(2) By using scan parameters that minimize the variance of the calculated image, the best calculated image can be found in a predetermined imaging time.

(2) By setting the number of 180° pulses in the IRnSE method to an even number, the standard deviation of the calculated image can be reduced. FIG. 4 shows the sum of 12 normalized standard differences at representative points of the human body's TI and T2 values for various pulse sequence combinations.

As is clear from this graph, when the number of 180° pulses in the IRnSE method (n is an integer) is an even number, the standard deviation is small. In other words, the IR3S method is better than the IR45E method.
It turns out that it is better to use the E method.

(2) By using the FR method, the standard deviation can be made smaller than other methods. In FIG. 4, lRa
The 5E method and the FR4SE method are the best.

■Also, by using the FR method, T1. Dispersion can be reduced over a wide range of T2.

■Composite pulse 90' to 180° (2) pulse
-45270° Accurate ・By using 9G”-45 and setting the spacing between each echo signal so that two adjacent echo signals are equal, 180° pulse error can be removed with even echo signals. can.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the main part of an embodiment of the NMRI imager according to the present invention, FIG. 2 is a diagram showing a pulse sequence of the IR3SE method, FIG. 3 is a diagram showing a pulse sequence of the FR4SE method, and FIG. 4 is a diagram showing a pulse sequence of the FR4SE method. The figure shows the sum of normalized standard deviations at representative points of TI+72M of the human body for various combinations of pulse sequences, and FIGS. 5 to 7 are diagrams showing examples of conventional pulse sequences. DESCRIPTION OF SYMBOLS 1... Magnet assembly, 2... Main magnetic field coil, 3... Gradient magnetic field coil, 4... RF transmission coil,
5...Receiving coil, 11...5 viewer, 12
...Display device, 13... Controller, 14... Gradient magnetic field $ control circuit, 15... Static magnetic field control circuit, 16.
...High frequency oscillator, 17...Gate modulation circuit, 18.
... Power amplifier, 19 ... Preamplifier, 20 ... Phase detection circuit, 21 ... Waveform memory, 30 ... Operation console.

Claims (1)

[Claims] 1) A nuclear magnetic resonance imaging device that applies high-frequency pulses and a magnetic field to an object to generate a nuclear magnetic resonance signal, and uses this signal to obtain an image of the tissue of the object, It is characterized by comprising a control/calculation means having the following functions (a) to (c) for obtaining a calculated image regarding at least one of relaxation time (T_1, T_2) or proton density (ρ). Nuclear magnetic resonance imaging device. Note (a) IRnSE method (n is an integer) and FRm that provide the best evaluation function for T_1, T_2, and ρ calculation images in the range of interest
Find scan parameters for the SE method (m is an integer). (b) Imaging is performed using the scan parameters determined in (b) above using a pulse sequence using the IRnSE method and the FRmSE method. (c) Calculate T_1, T_2, and ρ images from (n+m) images obtained from imaging using the IRnSE method and the FRmSE method. 2) The control calculation means is configured such that a 180° pulse is applied an even number of times in the pulse sequence in the IRnSE method.
The nuclear magnetic resonance imaging device described in Section 1. 3) The control/calculation means uses the IR3SE method and the FR4SE method.
2. The nuclear magnetic resonance imaging apparatus according to claim 1, wherein the nuclear magnetic resonance imaging apparatus is configured to perform imaging according to the method to obtain seven images, and calculate T_1, T_2, and ρ images from these seven images. 4) The control/calculation means uses the IR3SE method and the FR4SE method.
According to the first aspect of the present invention, the method is configured such that seven images are obtained by imaging according to the method, T_1, T_2, and ρ images are calculated from these seven images, and the following values are used as scan parameters. The nuclear magnetic resonance imaging device described in Section 1. (1) In the IR3SE method, a predetermined number of views are scanned with T_r=2.1 seconds T_d=0.46 seconds T_s_1=0.02 seconds T_s_2=0.02 seconds. (2) In the FR4SE method, a predetermined number of views are scanned with T_r=1.3 seconds T_s_1=0.02 seconds T_s_2=0.073 seconds T_s_3=0.02 seconds. (3) The seven images to be obtained are the IR3SE method and the FR4SE method.
The modulus is calculated by averaging the ratio of 1 to 2. 5) The control/calculation means uses the IR3SE method and the FR4SE method.
According to the first aspect of the present invention, the method is configured such that seven images are obtained by imaging according to the method, T_1, T_2, and ρ images are calculated from these seven images, and the following values are used as scan parameters. The nuclear magnetic resonance imaging device described in Section 1. (1) In the IR3SE method, a predetermined number of views are scanned with T_r=1.28 seconds T_d=0.37 seconds T_s_1=0.02 seconds T_s_2=0.02 seconds. (2) In the FR4SE method, a predetermined number of views are scanned with T_r=1.1 seconds T_s_1=0.02 seconds T_s_2=0.073 seconds T_s_3=0.02 seconds. (3) The seven images to be obtained are the IR3SE method and the FR4SE method.
Average the modulus in a 1:1 ratio.