JP2008142482A - Apparatus and program for carrying out segmentation of domain to be excised by complete mediastinal lymphadenectomy to two or more zones - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus and a program which carry out the segmentation of a domain to be excised by complete mediastinal lymphadenectomy to two or more zones. <P>SOLUTION: The apparatus has a means which extracts a mediastinal image showing a domain to be excised by the complete mediastinal lymphadenectomy from a CT image, a means which extracts a blood vessel image from the CT image, a means which sets the artery as the first interface based on the blood vessel image, a means which sets at least one of interfaces in the axial direction attributed to anatomical features as the second interface based on the mediastinal image, a means which sets at least one interface along an intramediastinal structure attributed to anatomical features as the third interface based on the mediastinal image and a means which carries out the segmentation of the domain to be excised by the complete mediastinal lympnadenectomy to two or more zones using at least one of the first, second and third interfaces. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an apparatus and a program for segmenting a region resected by mediastinal lymph node dissection into a plurality of sections.

  As a standard therapy for non-small cell carcinoma (stage I and II), a method of performing surgical resection is known. However, in Japan, the standard method is pulmonary segmentectomy and mediastinal lymphadenectomy. Since the postoperative prognosis in Japan is better than overseas, Japanese surgical methods are expected to spread worldwide through international academic societies such as the International Thoracic Surgery Society.

Conventionally, techniques for estimating the position of the mediastinal lymph node in a three-dimensional chest X-ray CT image are known (see Non-Patent Document 1, Non-Patent Document 2, and Non-Patent Document 3). However, there is no technique for segmenting a region to be excised by mediastinal lymphadenectomy into a plurality of sections or a technique for distinguishing a plurality of sections obtained as a result of the segmentation and displaying them three-dimensionally.
Tomoaki Sugawara et al., “Estimation of the presence of mediastinal lymph nodes in 3D chest X-ray CT images”, IEICE, IEICE Technical, TECHNICICAL REPORT OF IEICE, DSP2002-5, IE2002-5, MI2002-5 ( 2002-04), 25-30 Takayuki Kitasaka et al., “Extraction of mediastinal arterial region from 3D chest X-ray CT image using model”, IEICE, IEICE Technical, TECHNICAL REPORT OF IEICE, MI2000-97 (2001-01), 115 ~ 120 Takayuki Kitasaka et al. , "Automated Extraction of Aorta and Pulmonary Artery in Media from 3D Chest X-ray CT Images without Contrast Medium 200, Medical Imaging. Michael Fitzpatrick, Editors, Proceedings of SPIE Vol. 4684 (2002), 1496-1507.

  An object of this invention is to provide the apparatus and program which segment the area | region resected by mediastinal lymph node dissection into a some area.

  An apparatus of the present invention is an apparatus for segmenting a region excised by mediastinal lymph node dissection into a plurality of sections, and a mediastinal image representing a region excised by mediastinal lymph node dissection is obtained from a CT image. A means for extracting, a means for extracting a blood vessel image from a CT image, a means for setting the aorta as a first boundary surface based on the blood vessel image, and an anatomical feature based on the mediastinum image Means for setting at least one of the boundary surfaces in the axial direction as a second boundary surface, and based on the mediastinum image, of the boundary surfaces along the mediastinal structure based on anatomical features Means for setting at least one as a third boundary surface, and at least one of the first boundary surface, the second boundary surface, and the third boundary surface, and the mediastinal lymph node Segment the area to be excised by dissection into multiple areas And a Deployment to means, thereby the objective described above being achieved.

  In one embodiment of the present invention, the axial interface based on the anatomical features includes a left subclavian vein lower edge (pulmonary apex), a height at which the left brachiocephalic artery intersects the tracheal midline, and an aortic arch. The upper edge, the lower edge of the aortic arch (the height at which the ascending aorta and the descending aorta appear separated in an axial slice), the upper edge of the aortic vein, the upper edge of the left main pulmonary artery, the upper edge of the right main pulmonary artery, and the tracheal bifurcation Including the right main pulmonary artery lower margin.

  In one embodiment of the present invention, the boundary surface along the mediastinal structure based on the anatomical features includes an anterior vena cava wall, an ascending aorta anterior wall, an anterior tracheal wall, a posterior tracheal wall, It includes the left side of the aorta, the posterior wall of the ascending aorta, the posterior wall of the descending aorta, the anterior wall of the main bronchus, the posterior wall of the main bronchus, the inner surface of the left bronchus, and the inner surface of the right bronchus.

  It is preferable that the apparatus of the present invention further includes means for three-dimensionally displaying the plurality of areas in different modes.

  The program of the present invention is a program for causing a computer to execute a mediastinal lymph node segmentation process for segmenting a region to be excised by mediastinal lymph node dissection into a plurality of sections, and the mediastinal lymph node segmentation process includes: Extracting a mediastinum image representing a region excised by mediastinal lymph node dissection from a CT image; extracting a blood vessel image from the CT image; and based on the blood vessel image, the aorta is a first boundary surface A step of setting an axial direction boundary surface based on the anatomical feature as a second boundary surface based on the mediastinum image, and an anatomical feature based on the mediastinum image Setting a boundary surface along the mediastinal structure based thereon as a third boundary surface, the first boundary surface, the second boundary surface, and the third boundary surface Chino using at least one, includes a step of segmenting region to be excised by said longitudinal septum lymphadenectomy multiple zones, thereby the objective described above being achieved.

  ADVANTAGE OF THE INVENTION According to this invention, the apparatus and program which segment the area | region resected by mediastinal lymph node dissection into a some area can be provided. An apparatus and a program according to the present invention include a surgical support tool for displaying an area to be excised by surgery, an educational tool for dissemination of mediastinal lymph node dissection, and a tool for determining a therapeutic effect from volume evaluation of each lymph node area. Can be used as etc.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

1. Configuration of Mediastinal Lymph Node Segmentation System FIG. 1 shows an example of a configuration of a mediastinal lymph node segmentation system 1 according to an embodiment of the present invention.

  The mediastinal lymph node segmentation system 1 includes a CT device 10, a computer 20 connected to the CT device 10, and a display device 30 connected to the computer 20.

  The CT apparatus 10 stacks slice images of a three-dimensional object (for example, an area excised by mediastinal lymph node dissection), and divides the stacked slice images into a plurality of voxels to arrange them in three dimensions. Generate a plurality of voxels. The plurality of voxels generated by the CT apparatus 10 are output to the computer 20.

In FIG. 1, reference numeral 12 schematically indicates a set (CT image) of a plurality of voxels V _{(i, j, k)} generated by the CT apparatus 10 and arranged three-dimensionally. Here, i = 1, 2,... L; j = 1, 2,... M; k = 1, 2,... N, and L, M, and N are each 1 or more. It is an arbitrary integer. Each V _{(i, j, k)} has a CT value. The CT value is a value measured in units of voxels by the CT apparatus 10, and is represented by an integer, for example. As the CT apparatus 10, any type of CT apparatus capable of outputting the set 12 of voxels shown in FIG. 1 can be used.

  Based on the set 12 of voxels, the computer 20 executes a mediastinal lymph node segmentation process that segments a region to be excised by mediastinal lymph node dissection into a plurality of sections. As the computer 20, any type of computer including a CPU that executes various programs and a memory that stores various data can be used.

  The display device 30 displays the result of mediastinal lymph node segmentation processing executed by the computer 20. The display device 30 is preferably a display device capable of displaying an image generated as a result of mediastinal lymph node segmentation processing in three-dimensional color. The display device 30 may be, for example, a liquid crystal display device or a CRT display device.

  FIG. 2 shows an exemplary configuration of the computer 20.

  The computer 20 includes a CPU 21, a memory 22, an input interface unit 23, an output interface unit 24, a user interface unit 25, and a bus 26.

  The CPU 21 executes a program. The program is, for example, a program that causes the computer 20 to execute mediastinal lymph node segmentation processing for segmenting a region excised by mediastinal lymph node dissection into a plurality of sections. The program and data necessary for executing the program are stored in the memory 22, for example. It does not matter how the program is stored in the memory 22. For example, when the memory 22 is a rewritable memory, the program may be stored in the memory 22 by installing the program from the outside of the computer 20. Alternatively, when the memory 22 is a non-rewritable memory (read-only memory), the program may be stored in the memory 22 in a form fixed (burned) in the memory 22.

  The input interface unit 23 functions as an input interface for receiving a CT image from the CT apparatus 10.

  The output interface unit 24 functions as an output interface for outputting the result of mediastinal lymph node segmentation processing to the display device 30.

  The user interface unit 25 functions as an interface with the user. For example, input devices such as a mouse 25a and a keyboard 25b are connected to the user interface unit 25.

  The bus 26 is used to connect the components 21 to 25 in the computer 20 to each other.

2. Overview of Mediastinal Lymph Node Segmentation Processing FIG. 3 shows an overview of mediastinal lymph node segmentation processing. Such mediastinal lymph node segmentation processing can be realized, for example, in the form of a program. Such a program is executed by the CPU 21, for example.

  Step 301: A CT image is acquired. Each of the plurality of voxels included in the CT image has a CT value. Such acquisition of CT images is achieved, for example, when the CPU 21 receives a CT image output from the CT apparatus 10 via the input interface unit 23. However, the aspect of acquiring the CT image is not limited to this. The computer 20 can acquire a CT image in any manner. For example, the computer 20 may acquire a CT image by reading out a CT image recorded on a recording medium such as a magnetic disk, or a part of the CT image received from the CT apparatus 10 (for example, a CT apparatus). 10 selected from the CT images received from the user 10).

  Step 302: A mediastinal image representing a region to be removed by mediastinal lymph node dissection is extracted from the CT image. Such mediastinal image extraction is performed, for example, by performing threshold processing using a difference in CT values for a region (for example, a rectangular parallelepiped region) designated by an operator among regions in a chest CT image. This is accomplished by removing blood vessels, trachea, esophagus, thyroid gland, bone, lung, etc. from the area specified by the operator. Here, the region excised by mediastinal lymph node dissection includes lymph nodes and soft tissue / adipose tissue around the lymph nodes.

  Step 303: An image (blood vessel image) representing the aorta and vena cava is extracted from the CT image. Such blood vessel image extraction is performed by, for example, threshold processing and connection using a difference in CT values for a region (for example, a rectangular parallelepiped region) designated by an operator among regions in a chest CT image. This is achieved by performing component processing.

  Step 304: Based on the blood vessel image, an aorta (eg, aortic arch) is set as the first boundary surface. Details of step 304 will be described later in “3. Setting of Boundary Surface”.

  Step 305: Based on the mediastinum image, at least one of the boundary surfaces in the axial direction (direction perpendicular to the body axis) based on the anatomical features is set as the second boundary surface. Details of step 305 will be described later in “3. Setting of Boundary Surface”.

  Step 306: Based on the mediastinum image, at least one of the boundary surfaces along the mediastinal structure based on anatomical features is set as the third boundary surface. Details of step 306 will be described later in “3. Setting of boundary surface”.

  Step 307: Using at least one of the first boundary surface, the second boundary surface, and the third boundary surface, the region to be excised with mediastinal lymphadenectomy is segmented into a plurality of sections. Details of step 307 will be described later in “4.

3. Setting of boundary surface Three types of boundary surfaces are used as a boundary surface for segmenting a region to be excised by mediastinal lymph node dissection into a plurality of sections. Hereinafter, a method for setting each boundary surface will be described.

3.1 Aorta The aorta (for example, aortic arch) in the blood vessel image extracted in step 303 of FIG. 3 is set as the first boundary surface. The setting of the first boundary surface is automatically performed simultaneously with the extraction of the blood vessel image.

3.2 Axial direction based on anatomical features (direction perpendicular to the body axis) The boundary surfaces based on anatomical features in the axial direction (direction perpendicular to the body axis) are the following nine types of boundaries: Including faces.
(1) Left subclavian vein lower edge (pulmonary apex)
(2) Height at which the left brachiocephalic artery intersects the tracheal midline (3) Upper edge of the aortic arch (4) Lower edge of the aortic arch (the height at which the ascending aorta and the descending aorta appear separated by an axial slice)
(5) Upper edge of the aortic vein (6) Upper edge of the left main pulmonary artery (7) Upper edge of the right main pulmonary artery (8) Tracheal branch (9) Lower edge of the right main pulmonary artery (1), (2), (5) to ( The boundary surface of 8) corresponds to the definition shown in the “Lung Cancer Handling Rules (6th edition)”. The interfaces (3), (4), and (9) are additional interfaces required in the present invention. These additional interfaces are in line with the text of the “Lung Cancer Handling Rules (6th edition)”.

  Each of the nine types of boundary surfaces is manually set by the operator by the operator specifying a slice image in which each anatomical feature is seen. However, the boundary surface (i.e., tracheal bifurcation) of (8) above is automatically applied to the front wall of the trachea and the rear wall of the trachea as described later in “3.2.3 Setting of the front wall of the trachea and the rear wall of the trachea”. It may be automatically set when setting. Further, the boundary surface (ie, the upper edge of the aortic arch) of (3) is used when the ascending aortic anterior wall surface is automatically set as described later in “3.2.2 Setting of the ascending aortic anterior wall surface”. It may be set automatically.

  FIG. 4 shows an example of a boundary surface set as a boundary surface in the axial direction (direction perpendicular to the body axis) based on anatomical features.

3.3 Boundary Surface along Mediastinal Structure Based on Anatomical Features The boundary surface along the mediastinal structure based on anatomical features includes the following 11 types of boundary surfaces.
(1) The anterior wall of the vena cava (2) The anterior wall of the ascending aorta (3) The anterior wall of the trachea (4) The posterior wall of the trachea (5) The left side of the aorta (6) The posterior wall of the ascending aorta (7) The posterior wall of the descending aorta Anterior bronchial wall (9) Main bronchial posterior wall (10) Left bronchial inner surface (11) Right bronchial inner surface Each of the above 11 types of boundary surfaces designates slice images in which the operator can see each anatomical feature Therefore, it is manually set by the operator. However, the boundary surface (i.e., the front wall of the vena cava) of (1) can be automatically set by pressing a predetermined button. The boundary surface (2) (that is, the front wall of the ascending aorta) can be automatically set by pressing a predetermined button. The boundary surfaces (3) and (4) (that is, the wall surface before the trachea and the wall surface after the trachea) can be automatically set by designating one point in the trachea.

  Note that an intersection line formed on each axial slice image when these boundary surfaces intersect with each axial slice image is referred to as a boundary line.

  FIG. 5 shows an example of a boundary surface set as a boundary surface along a mediastinal structure based on anatomical features.

  FIG. 6 shows a coordinate system for each axial slice image. In FIG. 6, “A” indicates “anteria”, “P” indicates “posterio”, “R” indicates “right”, and “L” indicates “left”.

3.3.1 Setting the anterior wall of the vena cava Automatically setting the anterior wall of the vena cava is performed by, for example, using a vena cava image (a blood vessel image representing the vena cava) on each slice image This is accomplished by creating a straight line through the most ventral point. However, the vein shape in the vena cava image extracted from the CT image tends to be unstable.

  FIG. 7 shows an example of a vena cava image extracted from a CT image. In FIG. 7, a shows an example of an ideal vena cava image, b shows an example of a vena cava image when only a portion having a high contrast effect of the vena cava is extracted as a vena cava, c, d shows an example of the vena cava image when the lower part of the vena cava is not extracted. As shown in b, c, and d of FIG. 7, when the shape of the vein in the vena cava image is not stable, it is preferable to manually set the anterior vena cava wall surface.

  The manual setting of the anterior wall of the vena cava includes, for example, (1) setting the contact point (point A: start point) of the anterior vena cava in the start slice image of the boundary surface, and the vena cava in the end slice image of the boundary surface Setting the anterior wall contact point (point A: end point) (see a and b in FIG. 8) and (2) passing through the anterior wall contact point (point A; starting point) in the starting slice image By drawing a straight line parallel to the line, this straight line is used as a boundary line. In the end slice image, this straight line is bounded by drawing a straight line passing through the contact point (point A; end point) of the anterior wall of the aorta and parallel to the x axis. This is achieved by forming a line (see b in FIG. 8) and (3) calculating a contact point by linear interpolation for each slice image between the start slice image and the end slice image. In this manner, the front wall of the vena cava is created by stacking the boundary lines on each slice image (see c in FIG. 8).

3.3.2 Setting the Ascending Aortic Anterior Wall Wall The automatic setting of the ascending aortic anterior wall surface means that, for example, when creating an aortic image (a blood vessel image representing the aorta), three blood vessels branch from the aortic arch ( Cutting off the brachiocephalic artery, left common carotid artery, left subclavian artery in advance (see a and b in FIG. 9), and setting the smallest z-coordinate point of the aortic image as the aortic arch upper edge slice (FIG. 9). And a straight line passing through the most ventral point of the aorta image (see a, b, c, and d in FIG. 10) on each slice image. . Here, even when the right main pulmonary artery is on the ventral side of the aorta, the anterior wall of the aorta is traced by searching for the minimum y coordinate within 5 pixels left and right of the contact point of the previous slice (determined by experience). (See c in FIG. 10).

  Manually setting the anterior wall of the ascending aorta is, for example, using the middle point together with the start and end points of the anterior wall of the aorta set when manually setting the anterior wall of the vena cava. For each slice image between the images, a plane ascending anterior artery wall is created by creating a contact point by linear interpolation (see a in FIG. 11), and created for each slice image by linear interpolation. This is accomplished by moving the contact point up and down parallel to the y-axis to create a curved ascending anterior artery wall (see e and f in FIG. 11). Here, as the midpoint, a point D1 on the start slice image described later in “3.2.4 Setting of left side of aorta and rear wall of ascending aorta” is used. The shape of the blood vessel on the slice image may be circular or elliptical. Using the aorta image, a point obtained by moving the point D1 parallel to the negative direction of the y-axis to the front end of the ascending aorta is defined as a point F1, and a point obtained by moving the point D1 in the negative direction of the y-axis parallel to the left line of the aorta And Of the points F1 and F2, the point with the smaller y coordinate is set as the midpoint of the aortic anterior wall line (see b to d in FIG. 11).

3.3.3 Pre-tracheal wall and post-tracheal wall settings Setting the pre-tracheal wall and the post-tracheal wall automatically sets, for example, (1) a seed point in the trachea on the axial slice image. (See a in FIG. 12), (2) extracting the region of the CT value of the trachea by region growing (see b in FIG. 12), and (3) for each slice image from the first set point This is achieved by automatically setting a new seed and extracting the tracheal region by region growing as well. The seeds to be set automatically are 1/3 each of the left and right edges of the region from the minimum and maximum y coordinates of the region extracted in the previous slice so that it can correspond to the extraction of the left and right bronchi branched in the tracheal branch slice Points are calculated and used (see c in FIG. 12). Simultaneously with the processes (1) to (3), the tracheal region on each slice image is labeled, and the extraction is completed with the slice image having the label number of 2 (see d in FIG. 12). As a result, a tracheal branch slice and a tracheal region up to the tracheal branch are extracted (see e in FIG. 12). For the tracheal region of each extracted axial slice image, the minimum value of the y coordinate and the maximum value of the y coordinate are set as the y coordinate of the tracheal anterior wall line and the y coordinate of the tracheal rear wall line, respectively (a to d in FIG. 12). See yellow line).

  The manual setting of the anterior tracheal wall and the posterior tracheal wall is performed, for example, in the same manner as when the anterior vena cava wall is manually set, for each of the start slice image and the end slice image, This is achieved by setting the rear wall contact point and calculating the contact point by linear interpolation for each slice image between the start slice image and the end slice image. In this way, the front wall of the trachea and the rear wall of the trachea are created by accumulating the boundary lines on each slice image.

3.3.4 Setting the left side of the aorta and the posterior wall of the ascending aorta Manually setting the left side of the aorta and the posterior wall of the ascending aorta is the same as when manually setting the anterior wall of the aorta, for example. For each of the slice image and the end slice image,
Each ascending aorta left contact (point A) and descending aorta left contact (point B) are set, and a straight line passing through points A and B is defined as the aorta left line, and each slice between the start slice image and the end slice image This is accomplished by creating a contact point for the image by linear interpolation and creating a planar left side of the aorta. For each slice image, using the aorta image, point A and point B are translated to the left and right along the x-axis so as to be the left end contacts of the ascending aorta and descending aorta, respectively, and are corrected to the correct contact points. Is created (see a in FIG. 13). For each slice image, a point C obtained by translating the corrected point A left and right in the negative direction of the x-axis to the right end of the ascending aorta using the aorta image is defined as a point C. For the upper slice image of the aortic arch, a point obtained by moving the point C in the positive direction of the x-axis by three-fourths of the movement distance is defined as a point D2, and for the lower slice image, the point C is one-half of the movement distance. Only the point translated in the positive direction of the x-axis is defined as a point D1 (see b1 and b2 in FIG. 13). Points D1 and D2 are moved in the positive direction of the y-axis in parallel with the left side of the aorta as points E1 and E2, respectively (see c in FIG. 13). For each slice image, a line that passes through the points E1 and E2 and is perpendicular to the left line of the aorta is created, and that line is taken as the aortic posterior wall line. However, the aortic posterior wall line is not created for a region where the y coordinate is smaller than the intersection of the aortic left line and the aortic posterior wall line (see d in FIG. 13).

3.3.5 Setting of the main bronchial anterior wall and the posterior main bronchial wall The bronchus runs diagonally (see Fig. 14a) or the left and right bronchus have different thicknesses (see Fig. 14b) ) In many cases.

  Setting the anterior wall of the main bronchus and the posterior wall of the main bronchus manually means, for example, setting the anterior wall contact and the posterior wall contact and setting the anterior wall of the ascending aorta manually. This is achieved by creating a wall surface and a curved posterior main bronchial wall surface (see a and b in FIG. 14).

4). Execution of segmentation The area to be removed by mediastinal lymphadenectomy is divided into nine sections (section # 1, section # 2, section # 3a, section # 3, section # 3p, section # 4, section # 5, section # 6, segmented into area # 7).

4.1 Boundary surfaces necessary for extracting each area It is not necessary to set all the boundary surfaces described above in “3. Setting of boundary surfaces ”. When the operator specifies the area to be extracted (for example, when the operator inputs the number of the area to be extracted), only the boundary surface necessary for extracting the specified area should be set. Good. For example, when the operator specifies the area to be extracted (for example, when the operator inputs the number of the area to be extracted), the name of the axial slice to be set is displayed, and when an axial slice is set, it is set on that slice. When there is an anatomical feature point to be displayed, a feature point name to be specified may be displayed. In this way, only necessary boundary surfaces may be set by setting necessary items according to the screen display (guidance). This makes it possible to extract a target area using the minimum necessary boundary surface.

  Table 1 shows the boundary surfaces necessary for extracting each area classified into a boundary surface perpendicular to the z-axis, a boundary surface perpendicular to the y-axis, and a boundary surface perpendicular to the x-axis. . In Table 1, the underlined boundary surface is not defined in the “Lung Cancer Handling Rules (6th edition)”, but the area to be excised by mediastinal lymph node dissection (mediastinal area) The invention of the present application is indispensable for segmenting into areas (area # 1, area # 2, area # 3a, area # 3, area # 3p, area # 4, area # 5, area # 6, area # 7) Indicates the boundary surface defined by the person. In Table 1, the shaded boundary surface indicates a boundary surface that is affected by individual differences.

Examples of the boundary surface perpendicular to the z-axis include the following boundary surfaces.

Z1: Lower margin of the left subclavian vein (pulmonary apex)
Z2: Height at which the left brachiocephalic artery intersects the tracheal midline Z3: Upper edge of the aortic arch Z4: Lower edge of the aortic arch Z5: Upper edge of the odd vein Z6: Upper edge of the left main pulmonary artery Z7: Upper edge of the right main pulmonary artery Z8: Tracheal branch Z9: Right main pulmonary artery lower edge Examples of the boundary surface perpendicular to the y-axis include the following boundary surfaces.

Y1: anterior wall of the vena cava Y2: anterior wall of the aorta Y3: anterior wall of the trachea Y4: anterior wall of the trachea Y5: anterior wall of the main bronchus Y6: anterior wall of the main bronchus Y7: anterior wall of the ascending aorta Y8: anterior wall of the descending aorta Examples of such boundary surfaces include the following boundary surfaces.

X1: Aortic arch X2: Left side of aorta X3: Left bronchial inner surface X4: Right bronchial inner surface In the following description, for simplicity of explanation, Z1 to Z9 indicate a boundary surface perpendicular to the z-axis, A boundary surface perpendicular to the y-axis is indicated by Y8, and a boundary surface perpendicular to the x-axis is indicated by X1 to X4.

4.2 Extraction of Area # 1 For example, the extraction of area # 1 specifies a pixel that satisfies the expression (# 1-1) among a plurality of pixels included in the mediastinal image extracted in step 302 of FIG. Then, a specific value (for example, a color value indicating yellowgreen) is assigned to the pixel value of the specified pixel. Such processing is executed by the computer 20.

Z1 ≦ z <Z2 (formula (# 1-1))

4.3 The boundary surface perpendicular to the x-axis of the extraction area # 2 of the area # 2 is above the slice image of the lower aortic arch (Z4) where the left side surface (X2) of the aorta begins to exist, and the aortic arch (X1 ) Is on the right side of the aortic arch (X1). The boundary surface perpendicular to the x-axis of the area # 2 is on the right side of the aortic left side surface (X2) because the aortic left side surface (X2) exists below the slice image of the lower edge of the aortic arch (Z4). . Therefore, in the present invention, in consideration of individual differences, when the slice image of the upper edge (Z5) of the aberrant vein that is the end slice image of the area # 2 is above the lower edge of the aortic arch (Z4), the aorta When the arch (X1) is used as the boundary surface, and the slice image of the aortic vein upper edge (Z5) is below the lower edge of the aortic arch (Z4), the aortic left side surface (X2) is used as the boundary surface. I decided to use it.

  The extraction of the area # 2 is performed by, for example, the expression (# 2-1) or the expression (# 2-2) depending on the condition among a plurality of pixels included in the mediastinal image extracted in step 302 of FIG. Alternatively, a pixel that satisfies any one of the formula (# 2-3) and the formula (# 2-4) is specified, and a specific value (for example, a color value representing a pink) is specified for the pixel value of the specified pixel. ) Is done. Such processing is executed by the computer 20.

  FIG. 15 shows an example in which the area # 2 is displayed by coloring the pink.

1. When Z2 ≦ z <Z5 and z ≦ Z3 (see (a) of FIG. 15)
Y3 ≦ y ≦ Y4 (formula (# 2-1))
2. If Z2 ≦ z <Z5 and z> Z3 If Z5 <Z4 (always the aortic arch as a boundary, see FIG. 15B)
Y3 ≦ y ≦ Y4 and x <X1 Expression (# 2-2)
else Z4 ≦ Z5 (the aortic arch and the left side of the aorta are the boundary surface)
(1) z <Z4 (see (b) of FIG. 15)
Y3 ≦ y ≦ Y4 and x <X1 Expression (# 2-3)
(2) Z4 ≦ z <Z5 (see (c) of FIG. 15)
Y3 ≦ y ≦ Y4 and x <X2 (formula (# 2-4))

4.4 Zone # 3a Extraction slice image perpendicular to the z-axis of extraction zone # 3a is not defined in the “Lung Cancer Handling Rules (6th edition)”. Therefore, in the present invention, the slice image of the right main pulmonary artery lower edge (Z9) is used as the end slice image perpendicular to the z-axis of the section # 3a.

  The extraction of the area # 3a is performed by, for example, the expression (# 3a-1) or the expression (# 3a-2) according to the condition among a plurality of pixels included in the mediastinal image extracted in step 302 of FIG. Alternatively, a pixel that satisfies any one of the expression (# 3a-3) or the expression (# 3a-4) is specified, and a specific value (for example, a color value representing orange) is specified as the pixel value of the specified pixel. ) Is done. Such processing is executed by the computer 20.

  FIG. 16 shows an example in which the area # 3a is displayed in an orange color.

1. In the case of Z2 ≦ z <Z9 and z ≦ Z3 (see (a) of FIG. 16)
y <Y1 Formula (# 3a-1)
2. If Z2 ≦ z <Z9 and z> Z3 if Y1 <Y2 (see FIG. 16B)
y <Y1 Formula (# 3a-2)
else Y2 ≦ Y1 (see (c) of FIG. 16)
(1) y <Y2
There is no condition for x ... Formula (# 3a-3)
(2) Y2 <y <Y1
x <X1 Formula (# 3a-4)

4.5 Extraction Zone of Zone # 3 The boundary surface perpendicular to the x-axis of the extraction zone # 3 is above the slice image of the lower aortic arch (Z4) where the left side of the aorta (X2) begins to exist. ) Is on the right side of the aortic arch (X1). The boundary plane perpendicular to the x-axis of the area # 3 is on the right side of the aortic left side surface (X2) because the aortic left side surface (X2) exists below the slice image of the lower edge of the aortic arch (Z4). . The boundary surface perpendicular to the y-axis of the area # 3 is the front wall of the trachea (Y3) above the slice image of the tracheal branch (Z8), and the front of the main bronchial wall (Y5) after the tracheal branch. Therefore, in the present invention, in consideration of individual differences, the slice image of the tracheal branch (Z8) is higher than the slice image of the upper edge of the right main pulmonary artery (Z7), which is a boundary surface perpendicular to the z-axis of the section # 3. In the case of the above, between the slice image of the tracheal anterior wall line (Z3) and tracheal branch (Z8) and the slice image of the right main pulmonary artery upper edge (Z7), the front side from the main anterior bronchial wall (Y5) It was decided.

  For example, the extraction of the area # 3 satisfies the expression (# 3-1) among a plurality of pixels included in the mediastinal image extracted in step 302 of FIG. # 3-2) or the expression (# 3-3) or the expression (# 3-4) or the expression (# 3-5)) is identified, and the pixel of the identified pixel This is done by assigning a specific value (for example, a color value representing blue) to the pixel value. Such processing is executed by the computer 20.

  FIG. 17 shows an example in which the area # 3 is displayed in blue.

Z2 ≦ z <Z7 (formula (# 3-1))
1. Z2 ≦ z <Z4 (while the left side of the aorta does not exist)
if z ≦ Z3 (see FIG. 17A)
Y1 <y <Y3
else z> Z3 (see FIG. 17B)
Y1 <y <Y3 and x <X1 Formula (# 3-2)
2. Z4 ≦ z <Z7 (while the left side of the aorta is present)
if Z7 <Z8 (When the tracheal wall is always used as the boundary surface, see FIG. 17C)
Y1 <y <Y3 and y <Y7 and x <X2 Formula (# 3-3)
else Z8 ≤ Z7 (when the anterior tracheal wall and main anterior bronchial wall are used as boundary surfaces)
(1) Z4 ≦ y <Z8 (see (c) of FIG. 17)
Y1 y <Y3 and y <Y7 and x <X2 (formula (# 3-4))
(2) Z8 <y <Z7 (see (d) of FIG. 17)
Y1 y <Y5 and y <Y7 and x <X2 Formula (# 3-5)

4.6 Extraction of Area # 3p The extraction of area # 3p satisfies, for example, the expression (# 3p-1) among a plurality of pixels included in the mediastinum image extracted in step 302 of FIG. (Depending on the condition, identify a pixel that satisfies formula (# 3p-2) or formula (# 3p-3) or formula (# 3p-4)), and This is done by assigning a specific value (for example, a color value representing red) to the pixel value. Such processing is executed by the computer 20.

  FIG. 18 shows an example in which the area # 3p is displayed by coloring red.

Z2 ≦ z <Z8 Formula (# 3p-1)
if z ≦ Z3 (see (a) of FIG. 18)
Y4 <y formula (# 3p-2)
else z> Z3 (see FIG. 18B)
Y4 <y and x <X1 Formula (# 3p-3)
However, since the area indicated by the following equation (# 3p-4) is not the area # 3p, the area is not extracted as the area # 3p (the white area in FIG. 18C).

Z4 ≦ z <Z8 and the y coordinate of the contact point between the left side of the aorta and the descending aorta <y and the descending aorta <x <X2 (formula (# 3p-4))

4.7 The boundary surface perpendicular to the x-axis of the extraction area # 4 of the area # 4 is above the slice image of the lower edge of the aortic arch (Z4) where the left side surface (X2) of the aorta begins to exist. ) Is on the right side of the aortic arch (X1). The boundary surface perpendicular to the x-axis of the area # 4 is below the slice image of the lower edge of the aortic arch (Z4) and to the right of the left side of the aorta (X2) because of the left side of the aorta (X2). . Therefore, in the present invention, in consideration of individual differences, when the slice image of the upper edge (Z5) of the aberrant vein that is the start slice image of the area # 4 is above the lower edge of the aortic arch (Z4), the aorta When the arch (X1) is used as the boundary surface, and the slice image of the aortic vein upper edge (Z5) is below the lower edge of the aortic arch (Z4), the aortic left side surface (X2) is used as the boundary surface. I decided to use it.

  The extraction of the area # 4 satisfies, for example, the expression (# 4-1) among a plurality of pixels included in the mediastinum image extracted in step 302 of FIG. # 4-2) or a pixel satisfying any one of the equations (# 4-3) or (# 4-4)) is specified, and a pixel value of the specified pixel (for example, a specific value (for example, , A color value representing yellow). Such processing is executed by the computer 20.

  FIG. 19 shows an example in which the area # 4 is displayed in a yellow color.

Z5 <z <Z8 (formula (# 4-1))
If Z4 <Z5 (when the left side of the aorta is always used as the boundary surface)
Y3 <y <Y4 and x <X2 Formula (# 4-2)
else Z5 ≤ Z4 (when both the aortic arch and the left side of the aorta are used as the boundary surface)
(1) Z5 ≦ z <Z4
Y3 <y <Y4 and x <X1 Formula (# 4-3)
(2) Z4 ≦ z <Z8
Y3 <y <Y4 and x <X2 Formula (# 4-4)

4.8 Extraction of Area # 5 In the extraction of area # 5, for example, a pixel satisfying the formula (# 5-1) is specified among a plurality of pixels included in the mediastinum image extracted in step 302 of FIG. Then, a specific value (for example, a color value representing green) is assigned to the pixel value of the specified pixel. Such processing is executed by the computer 20.

  FIG. 20 shows an example in which the area # 5 is displayed by coloring green.

Z4 ≦ z <Z6 and Y7 <y <Y8 and X2 <x Expression (# 5-1)

4.9 End Slice Image perpendicular to the z-axis of Extraction Area # 6 of Area # 6 is not defined in the “Lung Cancer Handling Rules (6th edition)”. Therefore, in the present invention, the slice image of the upper edge of the left main pulmonary artery (Z6) is used as the end slice image perpendicular to the z-axis of the section # 6. Also, according to the definition of “Lung Cancer Handling Rules (6th edition)”, the starting slice image perpendicular to the z-axis of the area # 6 is the slice of the upper aortic arch (Z3) where the aortic arch (X1) starts to exist. Although it is an image, the slice image of the upper edge of the aortic arch (Z3) may be above the slice image at the height (Z2) where the left brachiocephalic artery intersects the tracheal midline due to individual differences. Therefore, in the present invention, in such a case, priority is given to the definition that the upper side of the height (Z2) at which the left brachiocephalic artery intersects the tracheal midline is zone # 1, and is perpendicular to the z axis of zone # 6. As a starting slice image, a slice image at a height (Z2) at which the left brachiocephalic artery intersects the tracheal midline is used.

  The extraction of the area # 6 is performed by, for example, the expression (# 6-1) or the expression (# 6-2) depending on the condition among a plurality of pixels included in the mediastinal image extracted in step 302 of FIG. This is done by identifying a pixel that satisfies any one of the above and assigning a specific value (for example, a color value representing cyan) to the pixel value of the identified pixel. Such processing is executed by the computer 20.

  FIG. 21 shows an example in which the area # 6 is displayed in a colored shade.

if Z2 ≦ Z3
start_slice = Z3
else Z3 <Z2
start_slice = Z2
1. start_slice ≦ z <Z4 (see (a) of FIG. 21)
Y2 ≦ y ≦ Y8 and X1 <x Expression (# 6-1)
2. Z4 ≦ z <Z6 (see FIG. 21B)
Y2 ≦ y ≦ Y7 and X1 <x Expression (# 6-2)

4.10 The end slice image perpendicular to the z-axis of the extracted area # 7 of the area # 7 is not defined in the “Lung Cancer Handling Rules (6th edition)”. Therefore, in the present invention, the slice image of the lower edge of the right main pulmonary artery (Z9) is used as the end slice image perpendicular to the z-axis of the area # 7.

  In the extraction of the area # 7, for example, among the plurality of pixels included in the mediastinum image extracted in step 302 of FIG. 3, a pixel satisfying the expression (# 7-1) is specified, and This is performed by assigning a specific value (for example, a color value representing a purple) to the pixel value. Such processing is executed by the computer 20.

  FIG. 22 shows an example in which the area # 7 is displayed in a purple color.

Z8 ≦ z <Z9 and Y5 <y <Y6 and X4 <x <X3 (formula (# 7-1))

4.11 Usefulness of segmentation considering individual differences For the left boundary surface perpendicular to the x-axis of area # 2 and area # 4, the boundary surface is determined in consideration of the vertical relationship between Z4 and Z5 as described above. Even if it is not, it seems to be sufficient if the area is not the area # 5 or # 6. However, there are individual differences in the vertical relationship between the slice image of the left main pulmonary artery upper edge (Z6) and the slice image of the upper edge of the azygous vein (Z5), which are end slice images perpendicular to the z-axis of the areas # 5 and # 6. Although there is a ratio of several percent, there is a case where Z6 comes above Z5. In such a case, the area # 5 and the area # 6 do not exist on the left side of the area # 2 and the area # 4. Therefore, the segmentation in consideration of the individual differences described above is necessary for accurately segmenting the region (mediastinal region) excised by mediastinal lymphadenectomy into nine sections.

  FIG. 23 shows an example of the positional relationship between the section # 4 and the sections # 5 and # 6. 23A shows an example of a slice image in which both the area # 4 and the areas # 5 and # 6 exist. FIG. 23B shows the slice image in which the area # 4 exists but the areas # 5 and # 6 do not exist. An example of a slice image is shown.

  Compared to when the lymph node area was defined in the "Lung Cancer Handling Regulations (6th edition)", the accuracy of the imaging device was improved and images with a thinner slice thickness were obtained (formerly about 5 mm) But now it is about 1 to 0.625 mm). The average distance between slice images that cause individual differences is 6 to 10 mm. In a conventional imaging device, such a distance between thin slice images cannot be considered, and it is not necessary to consider it. However, segmentation according to the present invention is indispensable in order to create a diagnosis support image and a surgery support image using a CT image having a thinner slice thickness.

4.12 Usefulness of manually setting the boundary surface As for the slice image of the boundary surface perpendicular to the z-axis, there has conventionally been a slice image that cannot be automatically determined. In the previous research, it is almost fully automatic processing, but in the present invention, many manual operations are left. For example, the setting of the main bronchial anterior wall and the main bronchial posterior wall can be made more accurately by manual operation than by fully automatic).

  FIG. 24 is a diagram for explaining setting of the main bronchial front wall and the main bronchial rear wall. In FIG. 24, (a) shows a bronchus extraction area (green area) (extracted to the end bronchus). In FIG. 24, (b) shows that even if the bronchus at the end is removed, the boundary line becomes like a red line simply by connecting the maximum coordinates and the minimum coordinates of the left and right bronchi. (The correct answer line is a blue line). This could happen in 15 of 34 cases.

In the slice image, the bronchiole is often branched from the main bronchus after branching the trachea. In such a case, the maximum y coordinate of the trachea image is simply set as the tracheal posterior wall line, and the tracheal image This is because if the minimum coordinate of the y coordinate is the tracheal front wall line, an error is considered to occur. In order to create an accurate boundary surface, it is considered effective to draw a line by semi-auto like the present invention.

5. Extraction Example of Zone FIG. 25 shows a display example when only zone # 2 is extracted. Such a display is preferably a three-dimensional display. In order to extract the area # 2, as shown in Table 1, “the height at which the left brachiocephalic artery intersects the tracheal midline (Z2)” and “upper edge of the odd vein” (as shown in Table 1) Z5) ”is set,“ anterior tracheal wall (Y3) ”and“ rear tracheal wall (Y4) ”are set as boundaries perpendicular to the y-axis, and“ aortic arch (X1) is set as the interface perpendicular to the z-axis ” ”Or“ left side of the aorta (X2) ”, the computer 20 may execute the mediastinal lymph node segmentation process described in“ 4.3 Extraction of Section # 2 ”.

  FIG. 26 shows a display example when nine areas # 1 to # 7 are extracted. Such a display is preferably a three-dimensional display. In order to extract nine areas, all nine types of boundary surfaces described in “3.2 Boundary surfaces in the axial direction (direction perpendicular to the body axis) based on anatomical features” are set. It is necessary to set all eleven types of boundary surfaces described in “3.3 Boundary Surfaces along Mediastinal Structure Based on Anatomical Features”. These boundary surfaces are preferably set by a minimum number of setting operations.

6). Application example of the present invention By segmenting a region to be excised by mediastinal lymph node dissection into a plurality of sections, and displaying the plurality of sections in a different manner (for example, in different colors) in a three-dimensional manner, The device of the present invention can serve as a surgical support tool that displays the area to be surgically removed or an educational tool for the spread of mediastinal lymph node dissection. In addition, by counting the number of pixels corresponding to each extracted area, the volume of each area is calculated, and by evaluating the calculated volume for each area, the apparatus of the present invention can be used for radiotherapy and the like. It can serve as a tool to determine the effect of treatment. It is preferable that information such as the boundary line used to extract the area and the volume of the area is stored in a memory, and the stored information can be read from the memory when the same area is extracted next time. . In addition, by superimposing and displaying a semi-transparent segmented image on the CT image, the device of the present invention determines which lymph node region the enlarged lymph node belongs to in a case of lymph node enlargement. It can serve as a diagnostic support image tool.

  As mentioned above, although this invention has been illustrated using preferable embodiment of this invention, this invention should not be limited and limited to this embodiment. It is understood that the scope of the present invention should be construed only by the claims. It is understood that those skilled in the art can implement an equivalent range based on the description of the present invention and the common general technical knowledge from the description of specific preferred embodiments of the present invention.

  INDUSTRIAL APPLICABILITY The present invention is useful for providing an apparatus, a program, and the like that segment a region to be excised by mediastinal lymph node dissection into a plurality of sections. The apparatus and program of the present invention include a surgical support tool for displaying a region to be excised by surgery, an educational tool for dissemination of mediastinal lymph node dissection, a tool for determining a therapeutic effect from volume evaluation of each lymph node region, and the like. Can be used.

The figure which shows an example of a structure of the mediastinal lymph node segmentation system 1 of embodiment of this invention The figure which shows an example of a structure of the computer 20 Diagram showing mediastinal lymph node segmentation process overview The figure which shows an example of the boundary surface set as a boundary surface of the axial direction (direction perpendicular to a body axis) based on anatomical features The figure which shows an example of the boundary surface set as a boundary surface along the structure in the mediastinum based on the anatomical feature Diagram showing the coordinate system for each axial slice image The figure which shows the example of the vena cava image extracted from CT image Diagram for explaining how to set the vena cava anterior wall Diagram for explaining how to create a blood vessel image The figure for explaining the setting method of the ascending aortic anterior wall The figure for explaining the setting method of the ascending aortic anterior wall The figure for explaining the setting method of the tracheal front wall and the tracheal wall Diagram for explaining how to set the left side of the aorta and the posterior wall of the ascending aorta Diagram to explain how to set main anterior bronchial wall and main bronchial posterior wall The figure which shows the example which colored and displayed area # 2 to pink The figure which shows the example which colored and displayed area # 3a in orange The figure which shows the example which colored and displayed area # 3 in blue The figure which shows the example which colored and displayed area # 3p to red The figure which shows the example which colored and displayed area # 4 in yellow The figure which shows the example which colored and displayed area # 5 to green The figure which shows the example which colored and displayed area # 6 to cyan The figure which shows the example which colored and displayed area # 7 in purple The figure which shows the example of the positional relationship of area # 4 and area # 5, # 6 Diagram for explaining the settings of the main bronchial front wall and the main bronchial rear wall The figure which shows the example of a display at the time of extracting only area # 2 The figure which shows the example of a display at the time of extracting nine areas # 1- # 7

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Mediastinal lymph node segmentation system 10 CT apparatus 12 Collection of voxels 20 Computer 30 Display apparatus

Claims (5)

A device for segmenting a region resected by mediastinal lymphadenectomy into a plurality of sections,
Means for extracting from the CT image a mediastinal image representing an area excised by the mediastinal lymph node dissection;
Means for extracting a blood vessel image from a CT image;
Means for setting the aorta as a first boundary surface based on the blood vessel image;
Means for setting at least one of axial boundary surfaces based on anatomical features as a second boundary surface based on the mediastinum image;
Means for setting, based on the mediastinum image, at least one of boundary surfaces along the mediastinal structure based on anatomical features as a third boundary surface;
Means for segmenting a region resected by the mediastinal lymph node dissection into a plurality of sections using at least one of the first boundary surface, the second boundary surface, and the third boundary surface A device with and. The axial interface based on the anatomical features is:
The left subclavian vein lower edge (pulmonary apex),
The height at which the left brachiocephalic artery intersects the tracheal midline,
The upper edge of the aortic arch,
The lower edge of the aortic arch (the height at which the ascending and descending aorta appear separated by an axial slice);
The upper edge of the azygous vein,
The left main pulmonary artery upper edge,
The right main pulmonary artery upper edge,
Tracheal bifurcation,
The device according to claim 1, comprising: a right main pulmonary artery lower edge. The interface along the mediastinal structure based on the anatomical features is
The anterior vena cava wall;
An anterior wall of the ascending aorta,
A wall in front of the trachea,
The wall behind the trachea,
The left side of the aorta,
The posterior wall of the ascending aorta,
The posterior wall of the descending aorta, the anterior wall of the main bronchus,
The main posterior bronchial wall,
Left bronchial inner surface,
The device of claim 1, comprising: a right bronchial inner surface.   The apparatus according to claim 1, further comprising means for three-dimensionally displaying the plurality of areas in different manners. A program for causing a computer to execute a mediastinal lymph node segmentation process for segmenting a region to be removed by mediastinal lymph node dissection into a plurality of sections,
The mediastinal lymph node segmentation process is:
Extracting a mediastinum image representing a region resected by the mediastinal lymph node dissection from a CT image;
Extracting a blood vessel image from the CT image;
Setting the aorta as a first boundary surface based on the blood vessel image;
Setting an axial boundary surface based on anatomical features as a second boundary surface based on the mediastinum image;
Setting a boundary surface along the mediastinal structure based on anatomical features as a third boundary surface based on the mediastinum image;
Segmenting a region to be excised with the mediastinal lymphadenectomy into a plurality of sections using at least one of the first boundary surface, the second boundary surface, and the third boundary surface A program that includes and.