JP4823257B2 - Fiber rendering method and fiber rendering device - Google Patents

Description

  The present invention relates to a fiber rendering method and an MRI (Magnetic Resonance Imaging) apparatus, and more particularly, to a method and an MRI apparatus for suitably rendering brain white matter fibers obtained by diffusion tensor imaging.

FIG. 21 is a flowchart showing a conventional fiber rendering method.
In Step P1, an MR image of an axial plane or an oblique plane is generated and displayed from three-dimensional image data collected by a diffusion tensor method or other imaging method (T1-weighted, T2-weighted, etc.) with an MRI apparatus. To do.
In step P2, as shown in FIG. 22, the operator sets a two-dimensional region of interest R1 (or a three-dimensional region of interest) on the displayed MR image G1.
In step P3 ′, as shown in FIG. 23, regular lattice points are generated in the region of interest R1 (or volume of interest) and set as tracking start points S1, S2, S3,.

In step P5, one of the tracking start points is selected.
In step P6 ′, the diffusion tensor analysis is performed at the selected tracking start point in the three-dimensional image data collected by the diffusion tensor method with the MRI apparatus to obtain the direction of the principal axis vector, that is, the first eigenvector.

In step P7, if the position of the unit distance along the direction of the principal axis vector is inside the three-dimensional image data space, the position is set as an adjacent point, and the process proceeds to step P8 ′. If there is no outside the three-dimensional image data space, the process proceeds to step P11. .
In step P8 ′, data at adjacent points is created by interpolation of three-dimensional image data, etc., and diffusion tensor analysis is performed to determine the direction of the principal axis vector and the FA (Fractional Anisotropy) value.
In step P9, if the FA value is equal to or greater than the threshold, the fiber tracking has not reached the end of the brain white matter fiber, so the fiber tracking is continued to return to step P7. If the FA value is less than the threshold, the end of the brain white matter fiber has been reached. Proceed to step P11 to end fiber tracking.
In this way, steps P7 to P9 are repeated until there is no 3D image data or fiber tracking reaches the end of the brain white matter fiber. For example, as shown in FIG. 24, the tracking start point S1 to the adjacent points N1, N2 , N3, and so on. At that time, connectivity is determined by using an inner product of vectors.

In step P11, the region from the tracking start point to the last adjacent point is stored as one brain white matter fiber.
In Step P12, if there is still a tracking start point that has not been selected in Step P5, the process returns to Step P5, and if not, the process proceeds to Step P14 ′.
In step P14 ′, for example, as shown in FIG. 25, an image is generated and displayed as if the stored brain white matter fibers are viewed from a desired observation direction.

  The explanation of the diffusion tensor and the traveling direction of the nerve fiber can be found in, for example, “Microstructural and Physiological Features of Tissues and Elucidated by Quantitative by Quantitative and Quantitatively Quantified by Quantitative Evidence by Quantitatively Quantified by Quantitative and Physician. 1996) "and" Diffusion Anisotropy-2D and 3D Images of Brain White Matter Fibers-: Department of Radiology, Kyoto Prefectural University of Medicine Yasuomi Kinosada: The 30th MR Imaging Society, September 4, 1998 (Sapporo) " It is described in.

When the lattice points regularly generated in step P3 in FIG. 21 are set as the tracking start points, as shown in FIG. 26, when the observation direction is parallel to the direction in which the lattice points are arranged, the tracking start alignment in the observation direction is started. Since the nerve fibers passing through the dots appear to overlap, the fiber density seems to have suddenly decreased, and there is a problem of giving an unnatural impression.
Accordingly, a first object of the present invention is to provide a fiber rendering method capable of preventing the fiber density from appearing to suddenly decrease in a specific observation direction.

When the threshold value in step P9 in FIG. 21 is reduced, a portion where the FA value is considerably low, that is, a portion where the fiber tracking reliability is considerably low is rendered. However, since a part with considerably low reliability of fiber tracking is rendered in the same display manner as a part with high reliability, there is a problem that it cannot be distinguished and hinders accurate diagnosis.
On the other hand, when the threshold value in step P9 of FIG. 21 is increased, the fiber tracking is terminated before reaching the end of the brain white matter fiber, so that there is a problem that the fiber cannot be drawn sufficiently.
Accordingly, a second object of the present invention is to provide a fiber rendering method that can render fibers in a display mode reflecting the reliability of fiber tracking.

Conventionally, since the eigenvalue of the diffusion tensor has not been reflected in the tracked fiber display, there is a problem that the change in the eigenvalue of the diffusion tensor cannot be visually recognized even when the drawn fiber is viewed.
Accordingly, a third object of the present invention is to provide a fiber rendering method that can render fibers in a display mode reflecting changes in eigenvalues of diffusion tensors.

As shown in FIG. 27, nerve fibers having different connection directions intersect at the nerve fiber intersection C. However, conventionally, since the tracking is performed using only the direction of the principal axis vector at the selected adjacent point, it is not possible to distinguish the fibers intersecting at the nerve fiber intersection, and the problem is that the tracking direction is wrong as shown in FIG. There is a point.
Accordingly, a fourth object of the present invention is to provide a fiber rendering method that can trace a direction without mistake even in a portion where nerve fibers having different connection directions intersect.

In the diagnosis of cerebral white matter degenerative disease and the like, it is useful information to know whether or not the nerve fiber connection between the two sites is broken.
Accordingly, a fifth object of the present invention is to provide a fiber rendering method in which the connection state of nerve fibers between two parts designated by an operator can be visually recognized.

In a first aspect, the present invention sets a region of interest or a volume of interest in three-dimensional image data collected by a diffusion tensor method with an MRI apparatus, and regular grid points in the region of interest or volume of interest. Then, the point where the grid point is randomly moved two-dimensionally or three-dimensionally is used as the tracking start point, and the direction of the principal axis vector is obtained by performing diffusion tensor analysis at each tracking start point in the three-dimensional image data, Select adjacent points along the direction of the principal axis vector, perform diffusion tensor analysis at the adjacent points, repeat the determination of the direction of the principal axis vector, track the fibers, and view each tracked fiber from the desired observation direction. Provided is a fiber rendering method characterized by generating and displaying an image.
In the fiber rendering method according to the first aspect described above, the number of tracking start points that are lined up is the same in any observation direction. Therefore, it is possible to prevent the fiber density from appearing to suddenly decrease in a specific observation direction. Note that, when viewed in the entire region of interest or the volume region of interest, the tracking start points have a uniform density and do not cause coarseness.

In a second aspect, the present invention sets a tracking start point in 3D image data collected by a diffusion tensor method using an MRI apparatus, and performs a diffusion tensor analysis at each tracking start point in the 3D image data. Obtain the direction of the principal axis and the diffusion anisotropy value, select adjacent points along the direction of the principal axis vector, perform diffusion tensor analysis at the adjacent points, and repeatedly obtain the direction of the principal axis vector and the diffusion anisotropy value The fiber is tracked, and an image is generated as if each tracked fiber was viewed from the desired observation direction, and displayed with opacity reflecting the diffusion anisotropy value at each tracking start point and each adjacent point. A fiber rendering method is provided.
In the fiber rendering method according to the second aspect, the transparency of the rendered fiber is changed according to the diffusion anisotropy value. Accordingly, the reliability of fiber tracking can be visually recognized from the transparency of the drawn fiber.

In a third aspect, the present invention provides a fiber rendering method characterized by using an FA value as a diffusion anisotropy value in the fiber rendering method configured as described above.
In the fiber rendering method according to the third aspect, the transparency of the rendered fiber can be changed by an FA value that takes a value between “0” and “1” according to the diffusion anisotropy.

According to a fourth aspect, in the fiber rendering method having the above-described configuration, the opacity at a certain adjacent point is X _{n + 1,} and the FA value at the previous adjacent point or the tracking start point is FA _n. When the opacity is X _n ,
X _{n + 1} = FA _n · X _n
A fiber rendering method is provided.
In the fiber rendering method according to the fourth aspect, the transparency can be gradually increased from the tracking start point to the end portion, and the transparency can be rapidly increased at the end portion.

In a fifth aspect, the present invention sets a tracking start point in 3D image data collected by a diffusion tensor method using an MRI apparatus, and performs a diffusion tensor analysis at each tracking start point in the 3D image data. Determine the direction of the principal axis vector and the eigenvalue of the diffusion tensor, select an adjacent point along the direction of the principal axis vector, perform diffusion tensor analysis at that adjacent point, and repeatedly obtain the direction of the principal axis vector and the eigenvalue of the diffusion tensor. A fiber characterized in that an image as if each tracked fiber was viewed from a desired observation direction was generated and displayed in a display color reflecting the eigenvalues of the diffusion tensor at each tracking start point and each adjacent point Provide a drawing method.
In the fiber rendering method according to the fifth aspect, the display color of the rendered fiber is changed according to the eigenvalue of the diffusion tensor. Therefore, the change in the eigenvalue of the change diffusion tensor of the displayed color of the drawn fiber can be visually recognized.

In a sixth aspect, the present invention relates to the fiber rendering method having the above-described configuration, where the display colors (R, G, B) are set when the eigenvalues of the diffusion tensor are λ1, λ2, and λ3.
R: G: B = 1: λ2 / λ1: λ3 / λ1
A fiber rendering method is provided.
In the fiber rendering method according to the sixth aspect, it can be seen that the closer the display color is to “white”, the more isotropic, and the closer the display color is to “red”, the more anisotropic the diffusion is.

In a seventh aspect, the present invention sets a tracking start point in 3D image data collected by a diffusion tensor method with an MRI apparatus, and performs a diffusion tensor analysis at each tracking start point in the 3D image data. Obtain the direction of the principal axis vector, set the direction of the principal axis vector as the tracking direction vector, select an adjacent point along the tracking direction vector, perform diffusion tensor analysis at that adjacent point, obtain diffusion tensor information, and obtain the diffusion tensor information and A fiber rendering characterized in that a fiber is tracked by repeatedly obtaining a tracking direction vector from at least one previous tracking direction vector, and an image as if each tracked fiber was viewed from a desired observation direction is generated and displayed. Provide a method.
In the fiber rendering method according to the seventh aspect, since a new tracking direction vector is obtained from the diffusion tensor information of a certain neighboring point and at least the previous tracking direction vector, even in a portion where nerve fibers having different connection directions intersect. Based on the connection direction so far, it is possible to distinguish nerve fibers having different connection directions, and to track each nerve fiber without mistaking the direction.

In an eighth aspect, the present invention provides the fiber rendering method having the above-described configuration, wherein eigenvalues of diffusion tensors at certain neighboring points are λ1, λ2, λ3, eigenvectors are e1, e2, e3, and tracking direction vectors are d _{i. +1} and when the tracking direction vector at the previous adjacent point or tracking start point is d _i ,
d _{i + 1} = {λ1 (e ₁ · d _i ) e ₁ + λ ₂ (e ₂ · d _i ) e ₂ + λ 3 (e ₃ · d _i ) e ₃ } / | λ1 (e ₁ · d _i ) e ₁ + λ 2 (E ₂ · d _i ) e ₂ + λ3 (e ₃ · d _i ) e ₃ |
A fiber rendering method is provided.
In the fiber rendering method according to the eighth aspect, the tracking direction vector d _i , the eigenvalues λ1, λ2, λ3 of the diffusion tensor at a certain adjacent point, and the eigenvectors e1, e2, e3 are used. d _{i + 1} can be obtained.

In a ninth aspect, the present invention sets a start-side region of interest and an end-side region of interest or a start-side region of interest region and an end-side region of interest volume region in three-dimensional image data collected by a diffusion tensor method using an MRI apparatus. A tracking start point is set in the start side region of interest or start side region of interest, and a fiber is tracked by performing a diffusion tensor analysis from each tracking start point in the three-dimensional image data. Providing a fiber rendering method characterized by determining whether or not to pass through a side region of interest or an end side region of interest and generating and displaying an image as if only the fibers determined to pass are viewed from a desired observation direction To do.
In the fiber rendering method according to the ninth aspect, since only the nerve fibers passing through the two parts are depicted, the connection state of the nerve fibers between the two parts can be visually recognized.

In a tenth aspect, the present invention provides a fiber rendering method having the above-described configuration, wherein the first eigenvalue of the diffusion tensor relating to the fiber determined to pass is λ1, the FA value is FA, and the total length of the fiber is L. Then, it is the sum of all the determined fibers,
M_Value = Σλ1 · FA / L
A fiber rendering method characterized by calculating and displaying a value.
In the fiber rendering method according to the tenth aspect, quantitative evaluation is possible by using M_Value as an index of the strength with which nerve fibers connect two sites.

In an eleventh aspect, the present invention relates to means for setting a region of interest or a volume region of interest in three-dimensional image data collected by a diffusion tensor method in an MRI apparatus, and regular in the region of interest or volume of interest. Means for setting grid points, means for randomly moving grid points in 2D or 3D as tracking start points, and performing diffusion tensor analysis at each tracking start point in 3D image data A means for obtaining the direction of the vector, a means for tracking the fiber by repeatedly selecting a neighboring point along the direction of the principal axis vector, performing a diffusion tensor analysis at the neighboring point and obtaining the direction of the principal axis vector, and each tracked There is provided a fiber rendering device comprising means for generating and displaying an image as if a fiber is viewed from a desired observation direction.
In the fiber rendering device according to the eleventh aspect, the fiber rendering method according to the first aspect can be suitably implemented.

In a twelfth aspect, the present invention relates to a means for setting a tracking start point in 3D image data collected by a diffusion tensor method using an MRI apparatus, and a diffusion tensor analysis at each tracking start point in the 3D image data. To determine the direction of the main axis vector and the diffusion anisotropy value, and select adjacent points along the direction of the main axis vector and perform diffusion tensor analysis at the adjacent points to determine the direction of the main axis vector and the diffusion anisotropy value. A means of tracking the fibers by repeating the search, and an opacity that reflects the diffusion anisotropy value at each tracking start point and each adjacent point by generating an image as if each tracked fiber was viewed from the desired observation direction. There is provided a fiber rendering device comprising a display means.
In the fiber rendering device according to the twelfth aspect, the fiber rendering method according to the second aspect can be suitably implemented.

In a thirteenth aspect, the present invention provides a fiber rendering device characterized by using an FA value as a diffusion anisotropy value in the fiber rendering device configured as described above.
In the fiber rendering device according to the thirteenth aspect, the fiber rendering method according to the third aspect can be suitably implemented.

In a fourteenth aspect, in the fiber rendering device having the above-described configuration, the opacity at a certain adjacent point is X _{n + 1,} and the FA value at the previous adjacent point or the tracking start point is FA _n. When the opacity is X _n ,
X _{n + 1} = FA _n · X _n
A fiber rendering device is provided.
In the fiber rendering device according to the fourteenth aspect, the fiber rendering method according to the fourth aspect can be suitably implemented.

In a fifteenth aspect, the present invention relates to a means for setting a tracking start point in 3D image data collected by a diffusion tensor method using an MRI apparatus, and a diffusion tensor analysis at each tracking start point in the 3D image data. To determine the direction of the principal axis vector and the eigenvalue of the diffusion tensor, and to select an adjacent point along the direction of the principal axis vector and perform a diffusion tensor analysis at that adjacent point to obtain the direction of the principal axis vector and the eigenvalue of the diffusion tensor To track the fibers by repeating the above, and to generate an image as if each tracked fiber was viewed from the desired observation direction and display it in a display color reflecting the eigenvalues of the diffusion tensor at each tracking start point and each adjacent point A fiber rendering device characterized by comprising:
In the fiber rendering device according to the fifteenth aspect, the fiber rendering method according to the fifth aspect can be suitably implemented.

In a sixteenth aspect, the present invention provides the fiber rendering device according to claim 15, wherein the display colors (R, G, B) are set when the eigenvalues of the diffusion tensors are λ1, λ2, and λ3.
R: G: B = 1: λ2 / λ1: λ3 / λ1
A fiber rendering device is provided.
In the fiber rendering device according to the sixteenth aspect, the fiber rendering method according to the sixth aspect can be suitably implemented.

In the seventeenth aspect, the present invention relates to a means for setting a tracking start point in 3D image data collected by a diffusion tensor method using an MRI apparatus, and a diffusion tensor analysis at each tracking start point in the 3D image data. To determine the direction of the principal axis vector and to set the direction of the principal axis vector as the tracking direction vector, and to select an adjacent point along the tracking direction vector and perform diffusion tensor analysis at that adjacent point to obtain diffusion tensor information and diffuse it Means for tracking a fiber by repeatedly obtaining a tracking direction vector from tensor information and at least one previous tracking direction vector; and means for generating and displaying an image of each tracked fiber viewed from a desired observation direction. Provided is a fiber rendering device characterized by comprising.
In the fiber rendering device according to the seventeenth aspect, the fiber rendering method according to the seventh aspect can be suitably implemented.

In an eighteenth aspect, the present invention provides the fiber rendering device having the above-described configuration, wherein the eigenvalues of the diffusion tensors at certain neighboring points are λ1, λ2, λ3, the eigenvectors are e1, e2, e3, and the tracking direction vector is d _{i. +1} and when the tracking direction vector at the previous adjacent point or tracking start point is d _i ,
d _{i + 1} = {λ1 (e ₁ · d _i ) e ₁ + λ ₂ (e ₂ · d _i ) e ₂ + λ 3 (e ₃ · d _i ) e ₃ } / | λ1 (e ₁ · d _i ) e ₁ + λ 2 (E ₂ · d _i ) e ₂ + λ3 (e ₃ · d _i ) e ₃ |
A fiber rendering device is provided.
In the fiber rendering device according to the eighteenth aspect, the fiber rendering method according to the eighth aspect can be suitably implemented.

In a nineteenth aspect, the present invention sets a start-side region of interest and an end-side region of interest or a start-side region of interest region and an end-side region of interest volume region in three-dimensional image data collected by a diffusion tensor method using an MRI apparatus. Means for setting a tracking start point in the start side region of interest or start side volume of interest region, means for tracking a fiber by performing diffusion tensor analysis from each tracking start point in the three-dimensional image data, and tracking Means for determining whether each of the fibers passed through the end-side region of interest or the end-side region of interest volume, and means for generating and displaying an image of only the fibers determined to pass through as viewed from a desired observation direction; A fiber rendering device characterized by comprising:
In the fiber rendering device according to the nineteenth aspect, the fiber rendering method according to the ninth aspect can be suitably implemented.

In a twentieth aspect, the present invention provides a fiber rendering device having the above-described configuration, wherein the first eigenvalue of the diffusion tensor relating to the fiber determined to pass is λ1, the FA value is FA, and the total length of the fiber is L. Then, it is the sum of all the determined fibers,
M_Value = Σλ1 · FA / L
There is provided a fiber rendering device characterized by comprising means for calculating and displaying.
In the fiber rendering device according to the twentieth aspect, the fiber rendering method according to the tenth aspect can be suitably implemented.

According to the fiber rendering method and fiber rendering device of the present invention, the following effects can be obtained.
(1) Even if the observation direction is changed, the number of tracking start points arranged in an overlapping manner is approximately the same. For this reason, it is possible to prevent the fiber density from appearing to suddenly decrease in a specific observation direction. Note that, when viewed in the entire region of interest or the volume region of interest, the tracking start points have a uniform density and do not cause coarseness.
(2) It can be seen that the portion of the rendered fiber with low transparency has high fiber tracking reliability, and the portion with high transparency has low fiber tracking reliability. Therefore, even if a portion having a considerably low fiber tracking reliability is depicted, a portion having a very low fiber tracking reliability can be distinguished from a portion having a high reliability, and there is no problem in making an accurate diagnosis.
(3) Whether the diffusion is isotropic or anisotropic can be visually recognized by the displayed color of the drawn fiber.
(4) Even in a portion where nerve fibers having different connection directions intersect, nerve fibers having different connection directions can be distinguished on the basis of the connection direction so far, and each nerve fiber can be traced without mistaking the direction.
(5) Since only the nerve fibers passing through the two parts can be depicted, the connection state of the nerve fibers between the two parts can be visually recognized.
(6) Quantitative evaluation of the strength with which nerve fibers connect two sites is possible.

  Hereinafter, the present invention will be described in more detail with reference to embodiments shown in the drawings. Note that the present invention is not limited thereby.

-First embodiment-
FIG. 1 is a block diagram showing an MRI apparatus according to an embodiment of the present invention.
In this MRI apparatus 100, the magnet assembly 1 has a bore (space portion) for inserting a subject therein, and forms a gradient magnetic field so as to surround the bore (the gradient coil is an X-axis). , Y-axis, and Z-axis coils, the combination of which determines the slice axis, the warp axis, and the lead axis) 1G, and transmission for applying RF pulses to excite spins of nuclei in the subject It comprises a coil 1T, a receiving coil 1R that detects an NMR signal from the subject, a static magnetic field power source 2 that forms a static magnetic field, and a static magnetic field coil 1C.
A permanent magnet may be used instead of the static magnetic field power supply 2 and the static magnetic field coil 1C (superconducting magnet).

  The gradient coil 1G is connected to the gradient coil drive circuit 3. Further, the transmission coil 1T is connected to the RF power amplifier 4. The receiving coil 1R is connected to the preamplifier 5.

  The sequence storage circuit 8 operates the gradient coil drive circuit 3 based on the stored pulse sequence in accordance with a command from the computer 7 to form a gradient magnetic field by the gradient coil 1G, and operates the gate modulation circuit 9 Then, the high frequency output signal from the RF oscillation circuit 10 is modulated into a pulse signal having a predetermined timing and a predetermined envelope, and this is added as an excitation pulse to the RF power amplifier 4 and amplified by the RF power amplifier 4. Is applied to the transmission coil 1T, and an RF pulse is transmitted.

  The preamplifier 5 amplifies the NMR signal from the subject detected by the receiving coil 1 </ b> R of the magnet assembly 1 and inputs it to the phase detector 12. The phase detector 12 uses the output of the RF oscillation circuit 10 as a reference signal, phase-detects the NMR signal from the preamplifier 5, and provides it to the A / D converter 11. The A / D converter 11 converts the analog signal after the phase detection into MR data of a digital signal and inputs it to the computer 7.

The computer 7 reads MR data from the A / D converter 11 and performs image reconstruction calculation to generate an MR image. The computer 7 is also responsible for overall control such as receiving information input from the console 13. Further, the computer 7 performs a fiber rendering process which will be described later with reference to FIG.
The display device 6 displays MR images and fiber images described later.

FIG. 2 is a flowchart showing fiber rendering processing by the MRI apparatus 100.
In Step P1, the MR image of the axial plane or oblique plane is generated and displayed from the three-dimensional image data collected by the MRI apparatus 100 by the diffusion tensor method or other imaging methods (T1-weighted, T2-weighted, etc.).
In step P2, as shown in FIG. 4, the operator sets a two-dimensional region of interest R1 (or a three-dimensional region of interest) on the displayed MR image G1.
In step P3, as shown in FIG. 5, regular lattice points g1, g2, g3,... Are generated in the region of interest R1 (or volume of interest).
In step P4, as shown in FIG. 6, the points where the lattice points g1, g2, g3,... Are randomly moved in two dimensions (or three dimensions) are set as tracking start points S1, S2, S3,. Here, as a random number for random movement, for example, a distribution function such as a Gaussian distribution or a uniform distribution may be used. Further, the moving range may be such that most of the moving range is within the interval range of the lattice points g1, g2, g3,.

In step P5, one of the tracking start points is selected.
In step P6, the diffusion tensor analysis is performed at the selected tracking start point in the three-dimensional image data collected by the diffusion tensor method in the MRI apparatus 100 to obtain the direction of the principal axis vector, the FA value, and the eigenvalue.

In step P7, if the position of the unit distance along the direction of the principal axis vector is inside the three-dimensional image data space, the position is set as an adjacent point, and the process proceeds to step P8. If the position is outside the three-dimensional image data space, the process proceeds to step P11.
In step P8, data at adjacent points is created by interpolation of three-dimensional image data, etc., and diffusion tensor analysis is performed to determine the direction of the principal axis vector, the FA value, and the eigenvalue. In step P9, if the FA value is equal to or greater than the threshold, the fiber tracking has not reached the end of the brain white matter fiber, so the fiber tracking is continued to return to step P7. If the FA value is less than the threshold, the end of the brain white matter fiber has been reached. Proceed to step P11 to end fiber tracking.
In this way, steps P7 to P9 are repeated until there is no 3D image data or fiber tracking reaches the end of the brain white matter fiber. For example, as shown in FIG. 7, from tracking start point S1 to adjacent points N1, N2 , N3, and so on. At that time, connectivity is determined by using an inner product of vectors.

In step P11, the region from the tracking start point to the last adjacent point is stored as one brain white matter fiber.
In Step P12, if there is still a tracking start point that has not been selected in Step P5, the process returns to Step P5, and if not, the process proceeds to Step P14 in FIG.

  In step P14 of FIG. 3, for example, as shown in FIG. 8, an image is generated such that the stored brain white matter fibers are viewed from a desired observation direction.

In step P15, the opacity at the tracking start point and X _o. When the opacity at a certain adjacent point is X _{n + 1} , the FA value at the previous adjacent point or the tracking start point is FA _n , and the opacity is X _n ,
X _{n + 1} = FA _n · X _n
And

In step P16, when the eigenvalues of the diffusion tensor are λ1, λ2, and λ3, the display color (R, G, B) is changed to
R: G: B = 1: λ2 / λ1: λ3 / λ1
And

  In Step P17, an image of the fiber is displayed using the opacity X and the display colors (R, G, B).

According to the MRI apparatus 100 of the first embodiment, the following effects can be obtained.
(1) As shown in FIG. 8 and FIG. 9, even if the observation direction is changed, the number of tracking start points arranged in an overlapping manner is approximately the same. For this reason, it is possible to prevent the fiber density from appearing to suddenly decrease in a specific observation direction. Note that, when viewed in the entire region of interest or the volume region of interest, the tracking start points have a uniform density and do not cause coarseness.
(2) It can be seen that the portion of the rendered fiber with low transparency has high fiber tracking reliability, and the portion with high transparency has low fiber tracking reliability. Therefore, even if the threshold value in Step P9 in FIG. 2 is reduced to draw a portion where the fiber tracking reliability is considerably low, the portion where the fiber tracking reliability is considerably low can be distinguished from the portion where the reliability is high. It will not interfere with the diagnosis.
(3) It is understood that the closer the display color of the drawn fiber is to “white”, the more isotropic diffusion is, and the closer the display color is to “red”, the more anisotropic the diffusion is.

In addition, you may deform | transform as follows.
(1) The opacity X may be calculated based on another index that reflects the diffusion anisotropy (for example, eigenvalue ratios λ2 / λ1, λ3 / λ1, relative anisotropy, volume ratio).
(2) The display colors (R, G, B) may be determined as R: G: B = λ1 / (λ1 + λ2 + λ3): λ2 / (λ1 + λ2 + λ3): λ3 / (λ1 + λ2 + λ3).

-Second Embodiment-
FIG. 10 is a flowchart illustrating fiber rendering processing by the MRI apparatus according to the second embodiment.
In step Q1, an MR image of an axial surface or oblique surface is generated and displayed from three-dimensional image data collected by a diffusion tensor method or other imaging method (T1-weighted, T2-weighted, etc.) with an MRI apparatus.
In step Q2, as shown in FIG. 4, the operator sets a two-dimensional region of interest R1 (or three-dimensional region of interest) on the displayed MR image G1.
In step Q3, as shown in FIG. 5, regular lattice points g1, g2, g3,... Are generated in the region of interest R1 (or volume of interest).
In step Q4, as shown in FIG. 6, the points where the lattice points g1, g2, g3,... Are randomly moved two-dimensionally (or three-dimensionally) are set as tracking start points S1, S2, S3,. Here, as a random number for random movement, for example, a distribution function such as a Gaussian distribution or a uniform distribution may be used. Then, the process proceeds to Step Q5 in FIG.

In step Q5 in FIG. 11, one of the tracking start points is selected.
At step Q6, the diffusion tensor analysis is performed at the selected tracking start point in the three-dimensional image data collected by the diffusion tensor method with the MRI apparatus to obtain the direction of the principal axis vector, the FA value, and the eigenvalue, and the principal axis vector is traced. Let it be a vector.

In step Q7, if there is 3D image data corresponding to the position of the unit distance along the direction of the tracking direction vector, the process proceeds to step Q8 using that as an adjacent point, and corresponds to the position of the unit distance along the direction of the main axis vector. If there is no 3D image data, the process proceeds to step Q11.
In step Q8, data at adjacent points is created by interpolation of three-dimensional image data, etc., and diffusion tensor analysis is performed to obtain eigenvectors, FA values, and eigenvalues.
In step Q9, if the FA value is equal to or greater than the threshold, the fiber tracking has not reached the end of the brain white matter fiber, so the process proceeds to step Q10 to continue the fiber tracking. If the FA value is less than the threshold, the end of the brain white matter fiber has been reached. Proceed to step Q11 to end fiber tracking.

In step Q10, eigenvalues of diffusion tensors at adjacent points are set to λ1, λ2, and λ3, eigenvectors are set to e1, e2, and e3, a tracking direction vector is set to d _{i + 1,} and the previous adjacent point or tracking start point is set. Let d _i be the tracking direction vector of
d _{i + 1} = {λ1 (e ₁ · d _i ) e ₁ + λ ₂ (e ₂ · d _i ) e ₂ + λ 3 (e ₃ · d _i ) e ₃ } / | λ1 (e ₁ · d _i ) e ₁ + λ 2 (E ₂ · d _i ) e ₂ + λ3 (e ₃ · d _i ) e ₃ |
And
FIG. 13 is a conceptual diagram showing the tracking direction vector d _{i + 1} .
Then, the process returns to step Q7.

  In this way, steps Q7 to Q10 are repeated until there is no 3D image data or fiber tracking reaches the end of the brain white matter fiber. For example, as shown in FIG. 7, from tracking start point S1 to adjacent points N1, N2 , N3, and so on. At that time, connectivity is determined by using an inner product of vectors.

In step Q11, the region from the tracking start point to the last adjacent point is stored as one brain white matter fiber.
In step Q12, if there is still a tracking start point that has not been selected in step Q5, the process returns to step Q5, and if not, the process proceeds to step Q14 in FIG.

  In step Q14 of FIG. 12, for example, as shown in FIG. 8, an image is generated such that the stored brain white matter fibers are viewed from a desired observation direction.

In step Q15, the opacity at the tracking start point and X _o. When the opacity at a certain adjacent point is X _{n + 1} , the FA value at the previous adjacent point or the tracking start point is FA _n , and the opacity is X _n ,
X _{n + 1} = FA _n · X _n
And

In step Q16, when the eigenvalues of the diffusion tensor are λ1, λ2, and λ3, the display color (R, G, B) is changed to
R: G: B = 1: λ2 / λ1: λ3 / λ1
And

  In step Q17, a fiber image is displayed using the opacity X and the display colors (R, G, B).

According to the MRI apparatus of the second embodiment, the following effects can be obtained in addition to the effects of the first embodiment.
(4) As shown in FIG. 14, if the previous tracking direction vectors di and dj are different, even if the adjacent points N _{i + 1} and N _{j + 1} match or are close to each other, the tracking direction vector d _{i + 1} and d _{j + 1} are different. Therefore, as shown in FIG. 15, even in a nerve fiber intersection C where nerve fibers with different connection directions intersect, nerve fibers with different connection directions can be distinguished based on the connection directions so far, and the directions are not mistaken. Each nerve fiber can be tracked.

  In order to determine the tracking direction vector, an appropriate number N may be given and an average vector of N tracking direction vectors may be used.

-Third embodiment-
FIG. 16 is a flowchart showing fiber rendering processing by the MRI apparatus of the third embodiment.
In step Q1, an MR image of an axial surface or oblique surface is generated and displayed from three-dimensional image data collected by a diffusion tensor method or other imaging method (T1-weighted, T2-weighted, etc.) with an MRI apparatus.
In step Q2 ′, as shown in FIG. 19, a two-dimensional start region of interest R1 (or a three-dimensional start region of interest volume region) and a two-dimensional end side region of interest R2 (or two) are displayed on the displayed MR image G1. The operator sets a three-dimensional end side volume of interest).
In step Q3, as shown in FIG. 5, regular lattice points g1, g2, g3,... Are generated in the start side region of interest R1 (or volume of interest region).
In step Q4, as shown in FIG. 6, the points where the lattice points g1, g2, g3,... Are randomly moved two-dimensionally (or three-dimensionally) are set as tracking start points S1, S2, S3,. Here, as a random number for random movement, for example, a distribution function such as a Gaussian distribution or a uniform distribution may be used. Then, the process proceeds to Step Q5 in FIG.

In step Q5 in FIG. 17, one of the tracking start points is selected.
At step Q6, the diffusion tensor analysis is performed at the selected tracking start point in the three-dimensional image data collected by the diffusion tensor method with the MRI apparatus to obtain the direction of the principal axis vector, the FA value, and the eigenvalue, and the principal axis vector is traced. Let it be a vector.

In step Q7, if the position of the unit distance along the direction of the tracking direction vector is inside the three-dimensional image data space, the position is set as an adjacent point, and the process proceeds to step Q8. If the position is outside the three-dimensional image data space, the process proceeds to step Q11. .
In step Q8, data at adjacent points is created by interpolation of three-dimensional image data, etc., and diffusion tensor analysis is performed to obtain eigenvectors, FA values, and eigenvalues.
In step Q9, if the FA value is equal to or greater than the threshold, the fiber tracking has not reached the end of the brain white matter fiber, so the process proceeds to step Q10 to continue the fiber tracking. If the FA value is less than the threshold, the end of the brain white matter fiber has been reached. Proceed to step Q11 to end fiber tracking.

In step Q10, eigenvalues of diffusion tensors at adjacent points are set to λ1, λ2, and λ3, eigenvectors are set to e1, e2, and e3, a tracking direction vector is set to d _{i + 1,} and the previous adjacent point or tracking start point is set. Let d _i be the tracking direction vector of
d _{i + 1} = {λ1 (e ₁ · d _i ) e ₁ + λ ₂ (e ₂ · d _i ) e ₂ + λ 3 (e ₃ · d _i ) e ₃ } / | λ1 (e ₁ · d _i ) e ₁ + λ 2 (E ₂ · d _i ) e ₂ + λ3 (e ₃ · d _i ) e ₃ |
And
FIG. 13 is a conceptual diagram showing the tracking direction vector d _{i + 1} .
Then, the process returns to step Q7.

  In this way, steps Q7 to Q10 are repeated until there is no 3D image data or fiber tracking reaches the end of the brain white matter fiber. For example, as shown in FIG. 7, from tracking start point S1 to adjacent points N1, N2 , N3, and so on. At that time, connectivity is determined by using an inner product of vectors.

In step Q11, the region from the tracking start point to the last adjacent point is stored as one brain white matter fiber.
In step Q12, if there is still a tracking start point that has not been selected in step Q5, the process returns to step Q5, and if not, the process proceeds to step Q13 in FIG.

In step Q13 in FIG. 18, it is determined whether or not there is an intersection between the obtained fiber and the end-side region of interest R2 (or the end-side region of interest volume region), and only the fibers having the intersection are selected.
In step Q14 ′, for example, as shown in FIG. 20, an image is generated in which only the selected brain white matter fiber f is viewed from a desired observation direction.

In step Q15, the opacity at the tracking start point and X _o. When the opacity at a certain adjacent point is X _{n + 1} , the FA value at the previous adjacent point or the tracking start point is FA _n , and the opacity is X _n ,
X _{n + 1} = FA _n · X _n
And

In step Q16, when the eigenvalues of the diffusion tensor are λ1, λ2, and λ3, the display color (R, G, B) is changed to
R: G: B = 1: λ2 / λ1: λ3 / λ1
And

  In step Q17, an image of the selected fiber is displayed using the opacity X and the display colors (R, G, B).

In step Q18, when the first eigenvalue of the diffusion tensor for the selected fiber is λ1, the FA value is FA, and the total length of the fiber is L, it is the sum of all the selected fibers.
M_Value = Σλ1 · FA / L
Is calculated and displayed.

According to the MRI apparatus of the third embodiment, the following effects can be obtained in addition to the effects of the second embodiment.
(5) Since only the nerve fiber f passing through the two parts is depicted, the connection state of the nerve fiber f between the two parts can be visually recognized.
(6) Quantitative evaluation is possible by using M_Value as an index of the strength with which nerve fibers connect two parts.

The average M_Value may be displayed by dividing M_Value by the number of selected fibers.
Further, the display brightness and display color of the fibers may be changed according to M_Value.

1 is a block diagram showing an MRI apparatus according to a first embodiment. It is a flowchart which shows the fiber drawing process which concerns on 1st Embodiment. FIG. 3 is a flowchart subsequent to FIG. 2. It is an illustration figure of the screen which sets a region of interest. It is an illustration figure of the lattice point arranged regularly. It is an illustration figure of the tracking start point which shifted the position irregularly. It is a conceptual diagram which shows a fiber tracking state. It is an illustration figure of the image which looked at the calculated | required fiber from the desired observation direction. It is an illustration figure of the image which looked at the calculated | required fiber from another observation direction. It is a flowchart which shows the fiber drawing process which concerns on 2nd Embodiment. FIG. 11 is a flowchart subsequent to FIG. 10. FIG. 12 is a flowchart subsequent to FIG. 11. It is a conceptual diagram which shows a tracking direction vector. It is a conceptual diagram which shows that a tracking direction is not mistaken even when a fiber crosses. It is explanatory drawing which shows that a tracking direction is not mistaken also in a nerve fiber crossing part. It is a flowchart which shows the fiber drawing process which concerns on 3rd Embodiment. FIG. 17 is a flowchart subsequent to FIG. 16. FIG. 18 is a flowchart subsequent to FIG. 17. It is an illustration figure of the screen which sets a start side region of interest and an end side region of interest. It is an illustration figure of the screen which displays only the fiber which connects a start side region of interest and an end side region of interest. It is a flowchart which shows the conventional fiber drawing process. It is an illustration figure of the screen which sets a region of interest. It is an illustration figure of the tracking start point arranged regularly. It is a conceptual diagram which shows a fiber tracking state. It is an illustration figure of the image which looked at the calculated | required fiber from the desired observation direction. It is an illustration figure of the image which looked at the calculated | required fiber from another observation direction. It is a conceptual diagram which shows that a fiber cross | intersects in a nerve fiber crossing part. It is explanatory drawing which shows that a tracking direction is mistaken also in a nerve fiber crossing part.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 MRI apparatus 1 Magnet assembly 6 Display apparatus 7 Calculator 13 Console

Claims (8)

  The tracking start point is set in the 3D image data collected by the diffusion tensor method with the MRI apparatus, the diffusion tensor analysis is performed at each tracking start point in the 3D image data, the direction of the principal axis vector and the diffusion anisotropy value Each of the tracked fibers is tracked by selecting a neighboring point along the direction of the principal axis vector, performing a diffusion tensor analysis at that neighboring point, and repeatedly obtaining the direction of the principal axis vector and the diffusion anisotropy value. A fiber rendering method characterized by generating an image as seen from a desired observation direction and displaying the image with opacity reflecting the diffusion anisotropy value at each tracking start point and each adjacent point.   2. The fiber rendering method according to claim 1, wherein an FA value is used as the diffusion anisotropy value. In fiber rendering method of claim 2, and X _{n + 1} the opacity of the neighboring points in, and the FA value in the previous neighboring points or tracking start point and FA _n, opacity and X _n and when,
X _{n + 1} = FA _n · X _n
A fiber rendering method characterized by:   The tracking start point is set in the 3D image data collected by the diffusion tensor method with the MRI apparatus, and the diffusion tensor analysis is performed at each tracking start point in the 3D image data to determine the direction of the principal axis vector and the eigenvalue of the diffusion tensor. Find the adjacent point along the direction of the principal axis vector, perform the diffusion tensor analysis at that neighboring point, repeat the calculation of the direction of the principal axis vector and the eigenvalue of the diffusion tensor, and track the fibers, and each tracked fiber desired When the images as seen from the observation direction are generated and the eigenvalues of the diffusion tensors at the tracking start points and the adjacent points are λ1, λ2, and λ3, the display colors (R, G, B) are set to R: G : A fiber rendering method characterized by displaying in a display color of B = 1: λ2 / λ1: λ3 / λ1.   Means for setting the tracking start point in the 3D image data collected by the diffusion tensor method on the MRI apparatus, and the diffusion tensor analysis at each tracking start point in the 3D image data to determine the direction of the principal axis vector and the diffusion anisotropy Means for determining the sex value and means for tracking the fiber by selecting adjacent points along the direction of the main axis vector and performing diffusion tensor analysis at the adjacent points to determine the direction of the main axis vector and the diffusion anisotropy value And means for generating an image as if each tracked fiber was viewed from a desired observation direction and displaying the image with opacity reflecting the diffusion anisotropy value at each tracking start point and each adjacent point. Fiber rendering device.   6. The fiber rendering apparatus according to claim 5, wherein an FA value is used as the diffusion anisotropy value. In fiber rendering apparatus of claim 6, and X _{n + 1} the opacity of the neighboring points in, and the FA value in the previous neighboring points or tracking start point and FA _n, opacity and X _n and when,
X _{n + 1} = FA _n · X _n
A fiber rendering device characterized by that.   Means for setting a tracking start point in 3D image data collected by the diffusion tensor method with an MRI apparatus, and performing a diffusion tensor analysis at each tracking start point in the 3D image data to determine the direction of the principal axis vector and the diffusion tensor Means for determining eigenvalues, means for selecting a neighboring point along the direction of the principal axis vector, performing diffusion tensor analysis at the neighboring point, and repeatedly determining the direction of the principal axis vector and the eigenvalue of the diffusion tensor, and tracking the fiber; Display color (R, G, B) when an image is generated as if each tracked fiber is viewed from a desired observation direction, and eigenvalues of diffusion tensors at each tracking start point and each adjacent point are λ1, λ2, and λ3. A fiber rendering device comprising: means for displaying in a display color of R: G: B = 1: λ2 / λ1: λ3 / λ1.

Images