JPH0284934A - Chemical shift image preparing method by nuclear magnetic resonance image diagnostic device - Google Patents

Abstract

PURPOSE:To correct the discontinuation of phases and to prevent an error to be caused by phase jump by separating an image area into the suitable number of voxcels, detecting an area to have possibility to generate the phase discontinuation, executing phase shift to the whole areas and executing reverse phase shift to this detected area. CONSTITUTION:In a chemical shift image preparing method to obtain a chemical shift image, for which the phase error of a Dixon method to separate and display the images of water and fat is corrected, the image area is divided into 32X32 voxcels. When the image area is defined as (256)<2> pixels, the condition of 1 voxcel=64 pixels is obtained. This division is executed with a purpose to suppress the influence of noise, to decrease resolution so that the correction can not be missed by the slight change of the phase and to execute average operation. The voxcel to include the phase of the pixel Pij>alpha or Pij<-alpha by >=6 pixcels is extracted as a phase jump voxcel. In such a condition, alpha=160 deg.. The phase jump voxcels are linked and a phase jump area is limited. For this limitation, phase jump voxcels 22 in an image 21 are linked and a closed area 23 is provided in an outside rather than a linking line and defined as the phase jump area.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention uses a nuclear magnetic resonance imaging diagnostic system to produce a chemical shift image corrected for the phase error of the Dixon method, which separates and displays images of water and fat. Regarding a shift image creation method.

(Conventional technology) Nuclear magnetic resonance imaging device! Imaging performed by MRI (hereinafter referred to as MRI) mainly uses protons, which are hydrogen nuclei. However, even if the same proton is present, the resonance frequency of hydrogen nuclei in different compounds, such as water and fat, changes sensitively depending on the chemical structure. This shift in resonance frequency is called chemical shift.

The resonant frequency of protons in the case of water and fat is f = 3.5
There is a difference in ppm. Several methods have been proposed that utilize this difference to obtain separated images of water and fat. Among these, the Dixon method is often used in MRI with a relatively low magnetic field.

The Dixon method will be explained with reference to FIG. Figure 2 is 3
This is a part of the pulse sequence of the gradient magnetic field applied to the axis, and only the radio frequency (hereinafter referred to as RF) axis and the signal axis are shown.

In the figure, (a) is a diagram of scan 1 in which the time interval between excitation pulse 1 and inversion pulse 2 is selected to be Ta=TE/2. Here, TE is the time interval between excitation pulse 1 and SE signal 3. At this time, the magnetization vectors of the proton components of water and fat are in phase, as shown in Figure 3 (a). (b) The figure is a diagram of scan 2 when the time interval TB between the excitation pulse 1 and the inversion pulse 2 is set as To = (TE/2)-ε. At this time, the chemical shift of water and fat (
A phase difference corresponding to 3.5 ppm) is given to the chemical vectors of water and fat, resulting in opposite phases as shown in FIG. 3 (b). In the figure, 4 is the magnetization vector of water and 5 is the magnetization vector of fat.

This Dixon method has the disadvantage that if there is non-uniformity in the static field, a phase error will occur corresponding to the non-uniformity, and therefore a correct separated image cannot be obtained.

Shizuka V above! A method of correcting errors due to iS non-uniformity is the method shown in FIG. 4, which is an improved version of the Dixon method.

(A) The figure shows Ta=T after applying inversion pulse 2 and excitation pulse 1.
Figure (B) is a diagram of scan 1 applied at E/2, (B) is a diagram of scan 2 where excitation pulse 2 was applied at Ta = (TE/2>-ε, /2)
FIG. 3 is a diagram of scan 3 applied with +ε. In the figure, the same reference numerals are used for parts equivalent to those in FIG. 3. When a scan as shown in FIG. 4 is performed, the phase relationship between water and fat is as shown in FIG. 5. (A) The figure shows the case of scan 1 with TB=TE/2, and water and fat are in the same phase.

(mouth) The figure shows scan 2 with TB = (TE/2) - ε
In the figure, water and fat are in opposite phases, but the phase is shifted by "+θ" due to the non-uniformity of the static magnetic field. (C) The figure shows the case of scan 3 where TB = (TE/2) + ε. Water and fat are also in opposite phases, but in scan 3, the phase is "-θ" due to the non-uniformity of the static magnetic field. It's off by just that.

The image data at scan 1°2.3 is now Sz, 8
2, Ss, water amplitude is W, fat amplitude is F9, phase offset is α, image data 81, S2
, Ss is as shown in the following equation.

Sl= (W+F)6 eXp(j(Z)...(1) 82 = (W-F) ・eXD (j θ)exp
(jα) ... (2) Ss = (W-f)-eXp (-j θ) ・
exp (J α) (3). From these equations, calculate the phase error 〇 due to static magnetic field inhomogeneity.

Equation (3) is divided by Equation (2) to obtain Ss/52=exo(-j2θ) In this way, the static magnetic field non-uniformity error can be eliminated.

With this improved Dickson method, the static magnetic field inhomogeneity error can be determined as described above, but the ΔE of the magnetic field inhomogeneity
If it exceeds the chemical shift of water and fat by 3.5 ppm, the separation of water and fat components will be reversed. FIG. 6 is a diagram showing how the correction value tjB differs depending on the magnitude of static magnetic field inhomogeneity. (A) The figure is for the case of 2θ minus π, and the correction value is at the x1 position on the positive side of the X axis. (b) The figure is 2θ>
In the π diagram, the correction value is at the x2 position on the negative side of the X axis. Therefore, when trying to correct using the modified Dixon method, the measurement range of non-uniformity is determined by the hanging of ε in Figure 4,
If the phase exceeds π, measurements cannot be performed correctly in the region beyond. In this way, when 20>π, a phase jump occurs and the relationship between Xεθ becomes discontinuous as shown in FIG. In the figure, point a is in phase I! At the point where 10 exceeds π, the phase line 1
0 causes a phase jump in phase 1111.

As a means of correction when a phase jump occurs in this way, correction is made by adding 2π to the phase of each point indicated by the phase 1i111. Similarly in the case of 2D, the X direction.

The phase is searched on each one-dimensional axis in the y direction, and if there is any discontinuity, it is corrected by either adding or subtracting 2π.

(Problems to be Solved by the Invention) However, this method of finding discontinuous points in phase and adding or subtracting 2π when there is discontinuity has the following drawbacks.

1. Easily affected by noise. If there is noise at even one point, it will be affected by the noise and accurate correction will not be possible.

2. In the case of two dimensions, correction may depend on whether the search is performed from the X direction or from the Y direction.

3. If the phase discontinuity line is complicated as shown in FIG. 8, the phase may be corrected twice at point a and point b.

The present invention has been made in view of the above points, and its purpose is to correct this phase discontinuity when there is a possibility that a phase jump of ±π has occurred in the phase distribution of MRI, and to correct the phase jump. An object of the present invention is to realize a method for creating a chemical shift image using MRI that does not cause errors caused by.

(Means for Solving the Problems) The present invention for solving the above problems uses a nuclear magnetic resonance imaging diagnostic apparatus that obtains chemical shift images corrected for phase errors of the Dixon method, which separates and displays images of water and fat. A chemical shift image creation method includes a method of separating an image area into an appropriate number of voxels, a method of detecting a region where phase discontinuity may occur in each voxel, and a method of applying a phase shift to the entire region. The present invention is characterized by comprising a method and a method of applying a reverse phase shift to the region subjected to the phase shift.

(Operation) Calculate the phase error due to static magnetic field inhomogeneity from the image data that has been reconstructed and configured, create voxels consisting of multiple pixels, find phase jumps in voxel units, and perform phase shifts. The shift direction of the area to which the phase shift has been performed is calculated and determined, and based on the result, a reverse phase shift is applied as necessary to correct non-uniform errors including phase jumps.

(Example) Hereinafter, an example of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a flowchart of the procedure for carrying out the method of the invention. The chemical shift image creation method will be explained according to the flowchart.

Step 1 Since the data obtained by scanning is complex data having a phase angle θ shift, a complex image reconstruction calculation is performed. (Equation (1), Equation (2), Equation (3)) Divide the data S3 of scan 3 by the data S2 of scan 2 to output the phase error exp (-j2θ) (Equation (4)), and calculate the non-uniform phase error. Create a phase distribution map using

Slab 3 The noise level is determined from the average (mO) and variance (σ) of the absolute values obtained by one-dimensional Fourier transform of each 3 to 5 lines at both ends in the warp direction using the raw data of scan 1. Generally, the noise level N is N=m, )+3.5σ. A value of N or less is set to '0' and a value of N or more is set to '1'.This processing is performed to omit noise pixels in the phase processing to be performed later and to suppress the influence of noise.

Step 4 Divide the image area into 32x32 voxels. If the image area is (256>2 pixels, then 1 voxel =
It is 64 pixels. This division is performed for the purpose of suppressing the influence of noise and performing average operation by lowering the resolution so as not to erroneously correct correction due to a slight change in phase.

Step 5 Voxels whose pixel phases include Pij>+α or Pij<−α for 6 or more pixels are depicted as phase-jump voxels. α=about 160”.P
ij indicates the phase of the point at coordinates (i, j).

Step 6 Check whether there are phase jump voxels, and if so, proceed to step 7.

Step 7 The phase jump region is limited by connecting the phase jump voxels.

This limitation is achieved by connecting the phase jump voxels 22 in the image 21 as shown in FIG.
3 is set as the phase jump region.

Step 8 A phase shift of π (180') is performed on the voxels within the closed region 23 designated as the phase jump region.

Step 9 Detect whether a new phase jump occurs due to the phase shift of the phase jump voxel in the region where the phase shift was performed in Step 8.

After detection, return to step 6.

Step 6: Check for the presence or absence of phase jump voxels over the entire image area, and if there are none, proceed to Step 10.

Step 10 Reduce the phase (2θ) of the phase map to 1/2, and change the phase (θ)
Let's find out.

Step 11 Voxel Vs in the phase shift region at the boundary of the region subjected to phase shift in Step 8 and voxel V in the normal region
At N, the next calculation is performed to determine the phase shift direction.

Average phase difference Δα between Vs and VN! , Δα2.

When 1Δα11...VsXeXp (-jπ/2), Δα21 ·=Vs

If Δα11〉1Δα2I, phase difference PS=-π If Δα11〉1Δα21, phase difference SP=+π Step 12 (Exo (-j (1/2)
2θ) and the phase shift correction exp(−jPs).

84 =82 ・eXp(J (IV2)2θ)・ex
p(-jPs) (5) Step 13 Obtain a separated image of water and fat.

W-(1/2)(31+84) =W-exp (j (Z)=Wr+jWi... (
6) F-(1/2> (81S4) = F-exo (jl-Fr+jFi... (7) Here, since exp (-jPs) in equation (5) is phase jump correction, equation (6), It does not appear in equation (7).

Step 14 From equation (6), W = r + A ■ 1”... (8) From equation (7), F = CF1] Kogo... (9) From equations (8) and (9), the static magnetic field is The respective absolute values of water and fat with the uniform error phase offset error removed are obtained.

As explained above, according to the method of this embodiment, since the image is divided into voxels and the averaging operation is performed, it is less susceptible to the influence of noise. In addition, the phase shift amount is corrected in step 1.
Since steps 1 and 12 are performed, even if a normal voxel is mistakenly treated as a phase jump region, it will be returned to normal and no problem will occur. Therefore, since the possibility of causing errors in phase jump correction is less than 6, it is possible to perform phase jump correction more accurately and with fewer errors than conventional methods.

Note that the present invention is not limited to the above embodiments.

In order to minimize phase jumps, zero-order correction may be first performed on the phase map so that the phase near the center of the image is set to O.

Although the phase shift amount is described as ±π, other phase amounts may be used.

When measuring nonuniformity using a homogeneous phantom, instead of correcting the phase shift, a corresponding uniform amount correction may be performed.

(Effects of the Invention) As described above in detail, the present invention is applicable to a chemical shift image creation method in which phase jump may occur. It is possible to correct phase discontinuity and eliminate errors caused by phase jumps, which has a great practical effect.

[Brief explanation of the drawing]

FIG. 1 is a 70-diagram for carrying out the method of one embodiment of the present invention; FIG. 2 is a diagram of the pulse sequence of the Dixon method;
Figure 3 is a diagram showing the phase relationship of magnetization vectors of water and fat components by the Dixon method, Figure 4 is a diagram of the RF pulse sequence of the improved Dixon method, and Figure 5 is a diagram showing the phase relationship of the magnetization vectors of water and fat components in the improved Dixon method. A diagram of the phase relationship between water and fat, Figure 6 is a diagram showing the correction value depending on the magnitude of static magnetic field inhomogeneity, Figure 7 is an explanation of discontinuity caused by phase jump, and Figure 8 is a diagram of phase continuity. FIG. 9, which is a diagram showing the case where the lines are complicated, is a diagram showing the setting of the phase jump region. 1...Excitation pulse 2...Inversion pulse 3...
・SE signal 4...Water 5...Fat
10.11... Phase line 12... Phase discontinuity line 21... Image 22... Phase jump voxel 23... Closed region

Claims (1)

[Claims] In a method for creating a chemical shift image using a nuclear magnetic resonance imaging diagnostic apparatus, which obtains a chemical shift image corrected for the phase error of the Dixon method that separates and displays images of water and fat, there is a method of separating the image area into an appropriate number of voxels. , a method of detecting an area where phase discontinuity may occur in this voxel unit, a method of applying a phase shift to the entire area, and a method of applying an opposite phase shift to the area where the phase shift has been performed. A method for creating a chemical shift image using a nuclear magnetic resonance imaging diagnostic apparatus, comprising: