JPS63311943A - Transversal relax time image calculation system - Google Patents

Abstract

PURPOSE:To remove the effect of the non-uniformity of a static magnetic field and the imperfection of a high frequency magnetic field and to calculate a transversal relax time image with high accuracy, by calculating the transversal relax time image by operating the data obtained after performing the mutual division of the images due to a plurality of spin echo methods made different in a peak time interval. CONSTITUTION:When imaging is performed on the basis of two sequences, wherein the time intervals TE of the mutual peaks of echo signals are set to TE1, TE2, like sequences 301, 302, the image signal intensities S1n, S2n obtained from the n-th echo signals of said sequences are represented by S1n=rho.exp(-nTE1/T2).fn(DELTAH) and S2n=rho.exp(-nTE2/T2).fn(DELTAH) when hydrogen atom density is set to SIGMAr, the time interval between the peaks of the echo signals is set to TEn and a transversal relax time is set to T2. Therefore, S1n/S2n=exp{-n(TE1-TE2)/T2} is formed and the effect of the non-uniformity DELTAH of a static magnetic field is removed. Since TE1-TE2 is a known value set from the outside, S1n/S2n is arranged in a time series manner and the same means as the calculation method due to a multiecho method is used to make it possible to calculate a T2-value without receiving the effect of DELTAH.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a transverse relaxation time image calculation method, and in particular to a transverse relaxation time suitable for obtaining a highly accurate transverse relaxation time image in a diagnostic magnetic resonance apparatus. This relates to an image calculation method.

[Conventional technology]

Conventionally, when obtaining a transverse relaxation time image using an MHI device, a method of calculating from multiple images obtained by a multi-echo sequence has been most often applied.・T2・Using・Multiple Echo・MHI・Technique・Magnetic Resonance in Medicine (Erro
rs in the Measurements of
T, Using Multiple-Echo M
RI Techniques MagMtic Re5
Onance in Medicine 3pp.

562-574 (1986)), when considering an ideal state, the signal strength Sn of the image obtained from the n-th echo in the multi-echo method is Sn = p ・5xp (-n T E / T2)
(1) ρ: Hydrogen atom density TE: Time interval between echo peaks T2: Transverse relaxation time, so by giving an appropriate TE and obtaining at least two echo images, ρ and T2 can be calculated for each pixel. Since simultaneous equations are obtained, these values can be calculated. A method of improving reliability by using a nonlinear least squares estimation method using three or more echo images is also often applied.

[Problem that the invention seeks to solve]

The above-mentioned conventional technology has a problem in that the accuracy of the calculated transverse relaxation time image decreases due to the influence of imperfections in the high-frequency magnetic field and non-uniformity in the static magnetic field on the measurement signal. This is reported in the above literature. For example, if static magnetic field inhomogeneity ΔH exists during the measurement time, the observed signal will be affected by this, so equation (1) listed on the previous page no longer holds true, and instead, Sn= p ・exp (-n T E / Ts)・f
The strength of the observation signal is expressed by n(ΔH) (1)'. For example, according to the analysis results, t"z(ΔH)=1/2(cos"φ(1-cosθ)
) (2,1)f, (ΔH) = 1/2(cos
'φ(1-cosθ)) (2,2). However, here, θ=+Ryo1 [7 "Ding Fθ.

φ= tan"" (ΔH/H,) H□: High frequency magnetic field strength θ. :Ideal flip angle (π)T with 180 pulses
y: If ΔH, which is the transverse relaxation time, is 0, then fn(ΔH)≠1, so in the above equation (1)′, the signal strength Sn is the time constant T2
The effect is not only the attenuation due to the above, but also the influence of fn(ΔH) superimposed thereon. Therefore, if the T2 value is calculated assuming the above equation (1) as in the past, the accuracy will be reduced due to this influence. The same holds true for flip angle deviations due to imperfections in the high frequency magnetic field.

The purpose of the present invention is to solve these conventional problems, eliminate the effects of static magnetic field inhomogeneity or high-frequency magnetic field imperfections, and improve transverse relaxation that allows calculation of transverse relaxation time images with higher accuracy. The object of the present invention is to provide a time image calculation method.

[Means for solving problems]

In order to solve the above-mentioned problems, the present invention includes means for generating a static magnetic field, a gradient magnetic field, and a high-frequency magnetic field, a detecting means for extracting magnetic resonance signals from an object to be inspected, and image reconstruction for the detected signals. In a diagnostic magnetic resonance apparatus that has means for performing various calculations, each image is captured using a multi-echo pulse sequence using a plurality of spin echo methods in which the time intervals of the peaks of magnetic resonance signals (echo signals) are different. A feature of this method is that the transverse relaxation time image is calculated by calculating the information obtained after dividing images obtained from corresponding echoes.

[Effect]

In the present invention, first, one image is photographed using a plurality of multi-echo sequences in which the time interval TE between the peaks of echo signals is changed. For example, sequence 30 in FIG.
1,302, when imaging is performed using two sequences in which TE is set as TE□ and TE, the image signal strength S obtained from each n-th echo signal is
1n l S in is 5in= p ・expCnT Ex/Tz)・fn(
ΔH) (3,1)Szn” ρ・exp(nT E
x/T2)'fn(ΔH) (3,2).

Here, if we divide equation (3, 1) by equation (3, 2), we get St
n/5an=exp(n(TE□'rE2)/Ti)(
In equation (4), the influence of ΔH is removed, and -TE is a known value set from the outside. Therefore, the T2 value can be calculated without being affected by ΔH by arranging Sin/S2n in time series and using the same means as the conventional multi-echo method for calculating the T2 value.

As mentioned above, by dividing the corresponding echo images of multiple multi-echo sequences with different time intervals between the peaks of the echo signals, it is possible to obtain information in which the inhomogeneity of the static magnetic field has been canceled. From these, it is possible to calculate a more accurate transverse relaxation time image using the same calculation method as before.

Becomes Noh.

〔Example〕

An embodiment of the present invention will be described below in detail with reference to the drawings.

FIG. 2 is a configuration diagram of a diagnostic magnetic resonance apparatus to which the present invention is applied.

A sequence control unit 201 that controls various pulses and magnetic fields generated to detect a magnetic resonance signal (hereinafter referred to as MR multiplied signal) from a subject (not shown) is used to cause a specific nuclide in the subject to resonate. A transmitter 202 for generating high-frequency pulses, a magnetic field control unit 203 for generating a gradient magnetic field whose strength and direction can be controlled arbitrarily, and an MR-multiplied signal resonant frequency, a static magnetic field, and an MR-multiplied signal generated from the subject. Receiver 2 that performs measurement after detection
05, the processing device 20G performs image reconstruction and calculation of a calculation image as in the present invention based on the measurement signal taken in from the receiver 205, and the finally obtained image is displayed on a display (for example, CRT display) 207. The magnetic field drive unit 204 generates a magnetic field necessary for measurement based on the control signal output from the magnetic field control unit 203.

An embodiment of the present invention with the above configuration is shown in FIGS. 1 and 3.
This will be explained using FIG.

FIG. 1 is a flowchart showing an outline of processing in an embodiment of the present invention.

Step 1: A control signal is output from the contrast control unit 201, and images are taken using a plurality of multi-echo sequences with different peak time intervals of echo signals. Here, two sequences 30 as shown in FIG.
1,302, consider the case where the first echo to the Nth echo are observed.

Step 102: Reconstruct an image from all echo signals obtained by each sequence.

Step 103: Perform division between two images obtained from corresponding echoes of the two sequences.

Do this for all echoes from 1st to nth\
conduct. The information obtained in this way is expressed as in equation (4). Note that when the pixel value on the dividing side is 0 or a value close to 0, division is not performed and O is substituted for the result.

Data string Sin/Szn (n = 1yL"'*
N) are arranged into echo classes, and the nonlinear least squares estimation method is applied to (4
) is applied to determine the T2 value at each pixel.

The processing in steps 102 to 104 above is performed in the processing device 206 in FIG.

FIG. 4 is a flowchart showing details of the process of the least squares estimation method in step 104 of FIG. below,
The detailed processing will be explained according to the flowchart in FIG.

Note that the following processing is performed pixel by pixel.

Step 401: First, the first
Echo image S. , S2X is applied to equation (4), and the first of T2
Find an approximate value T2°. From equation (4), T2゜=-(T
E i-T Ex ) / n n (s 1z
/S 2

Step 402: Let fn = Szn/San, and at this time, T, =T, in equation (4). The value fno of fn at the time of closing is calculated for each echo, and the sum of the squared errors with the observed value fh', E=Σ(fn'-fna)'', is calculated.

Step 403: If E is smaller than a predetermined threshold, the process proceeds to step 406; otherwise, the process proceeds to step 404.

Step 404: 72=T in equation (4). +ΔT
2, the tiller is expanded for 8T2 and the first order is obtained. If we take the term, fn=fna+(δfn/aΔTzbzmTzo・ΔT
It becomes 2. Here, let the squared error E be E = Σ(fn' - fn)2, and calculate 8T2. that minimizes E. That is, aE/aΔT2
Calculate ΔT2 such that =0.

Step 405: T. By replacing T2°+ΔT2, the process returns to step 402.

Step 406: Determine that the series of processing has converged,
The current T-2° value is finally calculated as 2
As a value, the process ends.

As mentioned above, the T2 value found at that point is (
4) The value of T2 is corrected by repeating the process until the sum of all echoes of the square error between the value on the right side when substituted into the equation and the measured value of the image signal becomes less than a certain threshold value. By doing this, the final transverse relaxation time image is obtained.

In addition, in this example, a method for calculating a transverse relaxation time image from two different sequences was described, but the above-mentioned two
Images are taken not from a regular sequence, but from three or more sequences such that the time interval between the peaks of each echo signal is an arithmetic progression, and then divided by the corresponding echo images of adjacent sequences in the arithmetic progression. It is also possible to improve the S/N by averaging all the information obtained for each echo, and from these, calculate the transverse relaxation time image in the same manner as above.

〔Effect of the invention〕

As explained above, according to the present invention, from information in which the effects of non-uniformity of the static magnetic field and imperfections of the high-frequency magnetic field have been removed,
Since a transverse relaxation time image can be obtained, a more accurate transverse relaxation time calculation image can be displayed.

[Brief explanation of the drawing]

Fig. 1 is a flowchart showing a processing overview of an embodiment of the present invention, Fig. 2 is a configuration diagram of a diagnostic magnetic resonance apparatus to which the present invention is applied, and Fig. 3 shows two types of multi-echo methods applied in the present invention. FIG. 4 is a flowchart showing the processing procedure of the least squares estimation method applied when calculating the transverse relaxation time image. 201: controller, 202: transmitter, 203:
Magnetic field control unit, 2o4: Magnetic field drive unit, 205: Receiver, 2
06: Processing device, 207: Display. Figure 1

Claims (1)

[Claims] 1. A means for generating a static magnetic field, a gradient magnetic field, and a high-frequency magnetic field, a detecting means for extracting a magnetic resonance signal from an object to be inspected, and performing various calculations including image reconstruction on the detected signal. In a diagnostic magnetic resonance apparatus having a magnetic resonance system, each image is captured using a multi-echo pulse sequence using a plurality of spin echo methods in which the time intervals of the peaks of magnetic resonance signals are different, and the images obtained from the corresponding echoes of each sequence are compared with each other. A transverse relaxation time image calculation method characterized in that a transverse relaxation time image is calculated by calculation between information obtained after dividing by . 2. In the transverse relaxation time image calculation method according to claim 1, images are taken using two multi-echo pulse sequences with different peak time intervals of the magnetic resonance signals, and the images of each sequence are A transverse relaxation time image calculation method characterized by arranging information obtained after performing division between images obtained from corresponding echoes in time series and calculating a transverse relaxation time image as one multi-echo image data string. . 3. In the transverse relaxation time image calculation method according to claim 1, imaging is performed such that the time intervals of the peaks of the magnetic resonance signals between the plurality of multi-echo pulse sequences are an arithmetic progression, and The information obtained after performing division between the images obtained from the corresponding echoes for each combination of two adjacent sequences of the difference sequence is further averaged between all the corresponding echoes, and the information obtained in this way is A transverse relaxation time image calculation method characterized by arranging the obtained information in time series and calculating a transverse relaxation time image as one multi-echo image data string.