JP3788846B2 - X-ray computed tomography system - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention is an X-ray computed tomography apparatus that irradiates a subject with cone-beam X-rays, measures intensity data of the X-rays transmitted through the subject, and performs reconstruction calculation processing based on the measurement data ( Hereinafter, it is abbreviated as “CT”).
[0002]
[Prior art]
(1) Fan beam CT and cone beam CT
The fan beam CT irradiates a fan beam (fan) X-ray, and as shown in FIG. 13A, a single slice detector in which the detection elements 100 are arranged in a one-dimensional direction (channel direction). 101. In the figure, F indicates the X-ray focal point, and α indicates the fan angle of the fan beam.
[0003]
The cone beam CT irradiates cone beam (quadrangular pyramid) X-rays, and as shown in FIG. 13B, a surface on which the detection elements 200 are arranged in a two-dimensional direction of a column direction and a channel direction. A detector 201 is provided. In the figure, β represents the cone angle. The column direction of the detection elements 200 is parallel to the Z axis (the body axis direction and the slice direction of the subject).
(2) Fan beam reconstruction Image reconstruction in the fan beam CT is performed as follows, for example. That is, X-rays are irradiated while rotating the X-ray tube around the subject, and projection data is obtained based on the transmitted X-ray intensity collected by a detector disposed opposite to the X-ray tube, After performing the correction process, an image is reconstructed by a convolution backprojection (CBP) method.
[0004]
FIG. 14 is a diagram showing the flow of reconstruction processing by the CBP method. In FIG. 2A, a thick arrow schematically shows the subject. In the CBP method, projection data at all angles is collected and corrected (FIG. (B)), and a convolution operation with a reconstruction function and a back projection operation at all angles are performed (FIG. (C)). ). FIG. 4D shows back projection at a certain angle.
(3) Half reconstructed image reconstruction using a fan beam When using a parallel beam, 180 ° data is necessary, but when using a fan beam, 180 ° + fan angle data is necessary. In this case, as shown in FIG. 15, it is only necessary to collect data on the angle of the focal position from Fs to Fe, and the method of reconstructing an image using data of 180 ° + fan angle in this way is half reconstruction. It is called. Half-reconstruction is a well-known image reconstruction method. In the conventional half-reconstruction, when data obtained at individual focal angles from the focal position Fs to Fe is back-projected, This is performed for all pixels (pixels).
(4) Half Reconstruction in SPECT (Nuclear Medicine Diagnostic Equipment) SPECT (Single Photon Emission Computed Tomography), which is a kind of nuclear medicine diagnostic equipment, also has a half reconstruction method in the fan beam similar to that of the CT described above. In particular, a half reconstruction method (hereinafter referred to as “nothing”) for obtaining an optimum focal angle (referred to as a back projection angle) Fs (pixel (r, θ)) → Fe (pixel (r, θ)) for each pixel in the image. (Referred to as "circle method"). For the circleless method, see Reference 1 (“High spatial resolution reconstruction technique for SPECT using a fan-beam collimator”, IEEE TRANSACTIONS ON NUCLEAR SCIENCE, Vol. 40, No. 4, August 1993, pp. 1149-1153, T. Ichihara, K. Nambu, N. Motomura) or Reference 2 (“SPECT's problems and advantages”, Journal of Japanese Society of Radiological Technology, Vol. 50, No. 9, P.1639-1650 (P.318) -329), Takashi Ichihara, September 1994), “Fan-beam high-resolution reconstruction method” (P.327 or P.1648).
[0005]
In the non-circular method, smoothing processing may be performed in a boundary region in which the back projection angle is slightly wider than the optimum angle in order to reduce artifacts.
(5) There are various methods for cone beam reconstruction (Feldkamp reconstruction method) and reconfigurable region cone beam reconstruction. The so-called Feldkamp reconstruction, which is an extension of the CBP method in fan beam reconstruction, The approximate reconstruction method called is representative. The Feldkamp reconstruction method is described in Reference 3 (“Practical cone-beam algorithm”, J. Opt. Soc. Am. A, 1, pp. 612-619 (1984) Feldkamp LA, Davis LCand Kress JW). ing.
[0006]
A region that can be reconstructed by this method is considered three-dimensionally as shown in FIG. That is, only a cylindrical region having a height of H determined by the following equation (1) in the slice direction can be reconstructed.
[0007]
H = 2 × (FCD-w) tan β (1)
FIG. 17 is a diagram showing an FOV (Field of View) in cone beam reconstruction. FIG. 17A shows the FOV viewed from the side, and FIG. 17B shows the FOV viewed from the top. It is.
(6) Half reconstruction and reconfigurable area in cone beam reconstruction Here, the combination of cone beam reconstruction by the CBP method described in (5) above and half reconstruction described in (3) above think of. In this case, considering the voxel in the reconstruction area, not the pixels in the reconstruction image, only the data of the same focal angle is used for the entire reconstruction area. The height H of the cylindrical region representing the reconfigurable region is unchanged.
[0008]
[Problems to be solved by the invention]
(1) Expansion of reconfigurable area Since the reconfigurable area (height H) shown in FIG. 16 is small and there is a lot of useless exposure, it is desired to increase the reconfigurable area. .
(2) Degradation of image quality due to the effect of cone angle As described above, since the Feldkamp reconstruction method is an approximate reconstruction method, the image quality degradation becomes significant as the cone angle (β in FIG. 16 ) increases. There is.
(3) Noise countermeasure for half reconstruction (S / N improvement)
When half reconstruction is performed, the number of data backprojected per voxel or pixel is approximately halved, so that there is a problem that S / N deteriorates.
[0009]
Accordingly, it is an object of the present invention to provide an X-ray computed tomography apparatus that is capable of enlarging a reconfigurable region, reducing image quality deterioration due to the effect of cone angle, and reducing noise caused by half reconstruction.
[0010]
[Means for Solving the Problems]
In order to solve the above problems and achieve the object, the present invention is configured as follows. That is, the X-ray computed tomography apparatus of the present invention has a surface detector, and in the X-ray computed tomography apparatus that collects cone beam data and performs three-dimensional reconstruction, the optimum focal angle for each voxel is obtained. In order to three-dimensionally project data, backprojection area determination means for determining a three-dimensional backprojection area corresponding to a focal angle and voxel coordinates is provided.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
(First embodiment)
FIG. 1 is a block diagram showing a configuration of a cone beam CT according to the first embodiment of the present invention. In FIG. 1, a cone beam CT10 includes an X-ray tube 21 for irradiating cone-beam shaped X-rays, and a surface detector 23 disposed opposite to the X-ray tube 21 with a subject not shown in the figure. It is attached to the gantry 25 so that it can rotate integrally. The X-ray tube 21 is supplied with a high voltage generated in the high voltage generator 19 under the control of the X-ray controller 17. The X-ray control device 17 is controlled by the system control unit 11.
[0012]
The gantry / bed control unit 13 drives the X-ray tube 21 and the surface detector 23 by controlling a gantry rotating unit (not shown) of the gantry 25, while the subject is placed by controlling the bed moving unit 15. The couch 15a is moved in the Z-axis direction.
[0013]
A data collection unit 27 is connected to the surface detector 23. The data collection unit 27 obtains projection data from multiple directions of the subject based on the detector output of the surface detector 23 under the control of the system control unit 11. The projection data obtained by the data collection unit 27 is sent to the image reconstruction unit 31 and is subjected to image reconstruction processing described below. The image of the subject obtained by the reconstruction process in the image reconstruction unit 31 is displayed on the display unit 33.
[0014]
FIG. 2 is a block diagram showing a detailed configuration of the image reconstruction unit 31.
The image reconstruction unit 31 includes a control device 50, a memory device 51, and a back projection calculation device 54. The memory device 51 includes a back projection area table 52 and a data storage unit 53. The control device 50 reads out necessary data from the data storage unit 53 of the memory device 51, performs predetermined processing such as convolution, and then outputs the data to the backprojection calculation device 54. In addition, the control device 50 reads the back projection region table of the corresponding view from the back projection region table 52 of the memory device 51 and outputs it to the back projection calculation device 54. The back projection computing device 54 performs back projection computation on a predetermined area according to the data output from the control device 50 and the back projection area table. The calculation result is stored in the data storage unit 53 of the memory device 51.
[0015]
Here, the non-circular Feldkamp reconstruction method according to the feature of the present invention will be described.
First, as shown in FIG. 3, when considering the “quality” of the beam irradiated from a certain focal position F, the beam passing through the voxel A has a low cone quality because the cone angle is large, and the beam passing through the voxel B has a cone angle. Is small so the quality is good. Although the cause is different from the change in data quality due to collimator accuracy in the SPECT apparatus, a similar result is obtained in the cone beam CT.
[0016]
Therefore, as shown in FIG. 4A, back projection is performed from an focal point from Fs (C) to Fe (C) on an arbitrary voxel C. FIG. 4A shows an optimum projection angle for the voxel C. FIG.
[0017]
When this is considered from an arbitrary focal position, as shown in FIG. 4B, a circle having a diameter of a line segment connecting the focal point F and the rotation center O (referred to as “no circle”), Considering a region (referred to as “Higan”), data from the focal position is backprojected to “Higan”.
[0018]
Considering again from an arbitrary voxel C, as shown in FIG. 5, it can be seen that the voxel C is included in the “cluster” region from the focal point Fs to Fe.
When this reconstruction area “Higan” is drawn three-dimensionally, it is as shown in FIG.
[0019]
As a result, as shown in FIG. 7, it is possible to obtain H ′ that is approximately double the reconfigurable region H (see FIG. 16). At this time, H ′ is expressed by the following equation (2).
[0020]
H ′ = 2 × FCD · tan β
= FCD / (FCD-w) × H (2)
Further, in this embodiment, the area to be three-dimensional backprojected is slightly expanded, and smoothing processing by weighting calculation is performed on the data in the boundary area . As a result, it is possible to reduce artifacts caused by a clear boundary between the region to be back-projected (“Higan”) and the region not to be back-projected (in the “no circle”).
[0021]
Further, the backprojection boundary area called “no circle” is not a perfect circle, but an ellipse as shown in FIG. 8B or 8C, and backprojection is performed in an area where the data quality does not change much. You may comprise as follows.
[0022]
Here, consider a case where volume scanning is performed by repeating one rotation → subject movement → one rotation. In the case of the conventional three-dimensional reconstruction based on 360 ° data, the movement distance for one time becomes H as shown in FIG.
[0023]
On the other hand, in the present embodiment, the focal angle data that is the optimum data for each voxel is backprojected to reconstruct three-dimensionally. For this reason, the region to be three-dimensionally backprojected is represented by the focal angle and voxel coordinates (here In FIG. 9 (b), the moving distance can be set to H 'so that a wide range can be imaged at high speed. It becomes possible.
[0024]
Also, even if the focal angle for backprojecting is the same over all voxels, so-called “dorming” occurs in the reconstructed image under the following conditions.
(1) The number of back projections (number of views) depends on the voxel
(2) Multiplying the convolution data by the inverse ratio of the square of the focal point-voxel distance, that is, 1 / (distance) ² to reduce the image quality degradation due to such doming during back projection. For example, a uniform phantom like a cylinder is scanned and reconstructed in advance, the correction coefficient is calculated from the degree of doming, and the reconstructed image is multiplied to improve the reliability of the voxel value of the reconstructed image it can.
[0025]
As described above, according to the non-circular Feldkamp reconstruction in the cone beam CT according to the present embodiment, the data of the focal angle that is the optimum data for each voxel is back-projected to perform the three-dimensional reconstruction. For this reason, the area to be three-dimensionally backprojected is determined according to the focal angle and the voxel coordinates (two-dimensional coordinates in this case), so that the movement distance can be expanded once and a wide range can be photographed at high speed. It becomes possible to do.
[0026]
In addition, the 3D backprojection area is slightly expanded, and the data in the boundary area is subjected to a smoothing process by weighting, so that the backprojection area ("Higan") and the non-backprojection area (in "no circle") Artifacts caused by a clear boundary can be reduced.
[0027]
As a modification of the present embodiment, a focal angle for back projection may be set for each region instead of for each voxel. In the present embodiment, the description has been made on the assumption that the volume scan is performed. However, a helical scan may be performed. In this case, for example, the above H ′ is set to a helical pitch.
[0028]
(Second Embodiment)
Next, a second embodiment of the present invention will be described. The second embodiment relates to an apparatus in which a non-circular shape is deformed along the slicing direction, and a 360 ° region and a non-circular region are separated. When the maximum cone angle β of the cone beam CT system is not so large, the image quality is sufficiently good even if the area of the height H near the center is reconstructed with data of normal 360 °. Further, the S / N of the image is improved when the focal angle for backprojection is wide, that is, when the number of backprojection data is large.
[0029]
As is apparent from FIG. 16, among the reconfigurable regions shown in FIG. 7, the region of height H near the center can be three-dimensionally backprojected using 360 ° data. Therefore, as shown in FIG. 10A, the reconfigurable region H ′ is divided into two regions (regions A and B) along the slice direction, and for the region A, normal data using 360 ° data is used. As described in the first embodiment, the region B is subjected to the three-dimensional reconstruction based on the 180 + fan angle data of the focus angle optimized for each voxel, as described in the first embodiment. For this reason, in this embodiment, the back projection area table 52 of the image reconstruction unit 31 is dependent on the Z-axis coordinates.
[0030]
Therefore, since the region to be three-dimensionally backprojected is determined corresponding to the focal angle and the voxel coordinates (three-dimensional coordinates including the Z-axis coordinate here), the reconstruction region can be enlarged and S / N can be increased. Can be improved.
[0031]
As described above, the reconfigurable area is not strictly divided into two areas A and B, and the noncircular shape (boundary) may be smoothly changed according to the Z coordinate as shown in FIG. 11 or FIG.
[0032]
Further, as shown in FIG. 10B, the non-circular shape may be changed in the area B according to the Z coordinate, and the area A may be back-projected using 360 ° data.
Also in the second embodiment, as described in the first embodiment, the boundary of the region to be three-dimensional backprojected may be slightly widened, and smoothing may be applied to data in the vicinity of the boundary.
[0033]
(Third embodiment)
Next, a third embodiment of the present invention will be described. The third embodiment relates to an apparatus in which non-circular regions obtained by two scans are overlapped and weighted and added. As described in the second embodiment, when back projection using 360 ° data is performed in the region A, a difference in S / N occurs at the boundary between the regions A and B.
[0034]
Therefore, as shown in FIG. 9C, the moving distance of one time is dared to be larger than H and smaller than H ′, for example, H ″. At this time, the reconstruction areas are F (k) and F (k + 1). ) Obtained by scanning with regions A (k) and A (k + 1) and F (k) and F (k + 1) reconstructed with 360 ° data obtained by scanning according to the second embodiment. Regions B (k) and B (k + 1) reconstructed by the method are classified into regions C (k, k + 1) that overlap in two scans.
[0035]
Here, in the region C (k, k + 1), the data obtained by the back projection of the data obtained by the scan of F (k) using the non-circle method of the second embodiment and the data obtained by the scan of F (k + 1). Similarly, the backprojected data is weighted and added with a weight depending on the Z coordinate and reconstructed. The weighted addition may be addition of data (that is, voxel data) obtained by performing a back projection operation for a necessary focal angle, addition of back projection operation data for one focal angle (one view), or cone. It is also possible to ignore the difference in corners and add convolution data before back projection calculation.
[0036]
As a modification of the present embodiment, the weight at the time of weighted addition can be changed as appropriate, such as making it dependent on the focal angle instead of a constant value or Z coordinate, or making it variable for each voxel.
[0037]
According to the third embodiment, the reconfigurable regions are overlapped, that is, the data for two rotations of the focal angle optimized for each voxel or each region is back-projected. S / N degradation due to the configuration can be suppressed.
[0038]
The present invention is not limited to the above-described embodiment, and can be implemented with various modifications. For example, in determining the three-dimensional backprojection region, the actual backprojection calculation may be combined with any of the following methods or other methods after using the method of the present invention.
[1] “Back projection using centering surface”
[2] "Direct back projection"
[3] “Back-projection method using corn-para transformation”
Alternatively, various 3D reconstruction methods using backprojection operations that apply Feldkamp reconstruction, such as convolution of 2D projection data of one view in the slice direction and then 3D backprojection, are known. The present invention is applicable to at least various three-dimensional image reconstruction methods using backprojection operation, and is not limited to the Feldkamp reconstruction method.
Further, the present invention is not limited to CT, and can be applied to various modalities such as SPECT and PET.
[0039]
【The invention's effect】
As described above, according to the present invention, it is possible to provide an X-ray computed tomography apparatus capable of expanding a reconfigurable region, reducing image quality deterioration due to the effect of cone angle, and reducing noise caused by half reconstruction. .
[Brief description of the drawings]
FIG. 1 is a block diagram showing a schematic configuration of a cone beam CT according to a first embodiment of the present invention.
FIG. 2 is a block diagram showing a detailed configuration of an image reconstruction unit according to the first embodiment.
FIG. 3 is a diagram showing a relationship between individual pixels in the same plane as the X-ray beam path according to the first embodiment.
FIG. 4 is a diagram showing an optimal backprojection angle for the voxel C according to the first embodiment.
FIG. 5 is a view showing “no circle” and “higashi” according to the first embodiment.
FIG. 6 is a diagram three-dimensionally showing “Higan” according to the first embodiment.
FIG. 7 is a view showing a reconfigurable area according to the first embodiment.
FIG. 8 is a view showing a modification of the “no circle” shape according to the first embodiment.
FIG. 9 is a view showing a reconfigurable area in the volume scan according to the first embodiment.
FIG. 10 is a diagram showing a reconfigurable area according to the second embodiment of the present invention.
FIG. 11 is a view showing a modification of the “no circle” shape according to the second embodiment.
FIG. 12 is a view showing a modification of the “no circle” shape according to the second embodiment.
FIG. 13 is a diagram showing a configuration of a fan beam CT detector according to a conventional example of the present invention.
FIG. 14 is a diagram showing a flow of reconstruction processing by a CBP method according to a conventional example of the present invention.
FIG. 15 is a diagram showing a focal angle in half reconstruction according to a conventional example of the present invention.
FIG. 16 is a diagram three-dimensionally showing a reconfigurable area in the Feldkamp reconstruction method according to the conventional example of the present invention.
17A and 17B are a top view and a side view showing an FOV in cone beam reconstruction according to a conventional example of the present invention.
[Explanation of symbols]
10 ... Cone beam CT
DESCRIPTION OF SYMBOLS 11 ... System control part 13 ... Base and bed control part 15 ... Bed movement part 17 ... X-ray control device 19 ... High voltage generator 21 ... X-ray tube 23 ... Surface detector 25 ... Base 27 ... Data collection part 31 ... Image Reconfiguration unit 33 ... display unit

Claims (3)

In an X-ray computed tomography apparatus having a surface detector and collecting cone beam data and performing three-dimensional reconstruction,
Storage means for storing the collected cone beam data;
Of the cone beam data stored in the storage means, an image is reconstructed by back projecting data having a smaller cone angle for each voxel, and a circular region that does not perform back projection and the circular region And a reconstruction processing means configured to perform back projection processing by determining a region for back projection outside the region, and to perform smoothing by weighting calculation at the circular region boundary. X-ray computed tomography apparatus. X-ray computer tomography apparatus according to claim 1, further comprising a means for adding weighted backprojection data for each obtained by two scans before and after moving the focal point along the Z axis. The X-ray computed tomography apparatus according to claim 1, wherein the shape of the circular region is changed based on a Z-axis coordinate .

Images