JP2009050622A - Tumor diagnosis supporting system, tumor diagnosis supporting method, and tumor diagnosis supporting program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technique for reducing false positive rate of a tumor diagnosis supporting system. <P>SOLUTION: A tumor image is detected from a medical image, and a blood vascular image existing around the detected tumor image is extracted, and the extracted blood vascular image is separated at each space between blood vascular bifurcation and divided into a blood vascular segment, and the serpentine-curved degree of each blood vascular segment is computed and output visibly. Additionally, the bifurcation degree of each bifurcation existing in the blood vascular image is computed and output visibly. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a technique for supporting diagnosis of a malignant tumor.

  Malignant tumors are one of the leading causes of death in many countries. However, with the recent development of medical technology, if a tumor can be detected early, a high survival rate can be maintained by early treatment. Therefore, in order to increase the early detection rate, image diagnostic apparatuses such as an X-ray diagnostic apparatus, MRI apparatus, ultrasonic diagnostic apparatus, and nuclear medicine diagnostic apparatus have been developed, and the level is improved by improving the temporal and spatial resolution of the image diagnostic apparatus. Now it is possible to image the tumor of the subject. However, while the success rate of tumor imaging is increased, the slice thickness, which is the interval between imaging of medical images, is becoming increasingly thinner. As a result, the number of medical image data to be referred to for tumor detection has become enormous. There is a limit to the amount of data that can be diagnosed by a doctor. In addition, performing a huge amount of image diagnosis may cause considerable fatigue and may cause misinterpretation or oversight.

  Therefore, a tumor diagnosis support system that detects a shadow of a tumor from a medical image using a computer is provided in order to support a doctor's image diagnosis (see, for example, “Patent Document 1” and “Patent Document 2”). ). The tumor diagnosis support system is also called a CAD system (Computer Aided Detection).

  However, in the conventional tumor diagnosis support system, since there are many false positive rates in which a tumor is erroneously determined to be positive, it is used as a screening tool that screens in advance and narrows down candidates to be judged by a doctor as a malignant tumor. Limited.

  In order to pursue the efficiency and accuracy of image diagnosis by a doctor, it is necessary to reduce the false positive rate of this tumor diagnosis support system.

JP 2006-239005 A JP 2007-125169 A

  The present invention has been made in view of the above problems, and an object thereof is to provide a technique for reducing the false positive rate of a tumor diagnosis support system.

  In order to solve the above-described problem, a first aspect of the present invention includes a detection unit that detects a tumor image from a medical image, an extraction unit that extracts a blood vessel image present around the detected tumor image, and an extracted blood vessel Separating means for separating an image between blood vessel branch points and dividing it into blood vessel segments; calculation means for calculating the meandering degree of each blood vessel segment as a malignant tumor index; and output means for outputting the meandering degree in a visible manner; It is characterized by providing.

  In order to solve the above-described problem, the second aspect of the present invention includes a detection unit that detects a tumor image from a medical image, an extraction unit that extracts a blood vessel image from around the detected tumor image, and the blood vessel image. And a calculating means for calculating the branching degree of each branching point as a malignant tumor index, and an output means for outputting the branching degree in a visually recognizable manner.

  As the degree of meandering, the dispersion of the separation length of the blood vessel segment and the smooth curve obtained by smoothing the blood vessel extending shape of the blood vessel segment, the ratio of the length of the blood vessel segment and the distance from the start end to the end of the blood vessel segment, Obtained by converting a separation length curve indicating the separation length for each distance from the start end to the end of the blood vessel segment and the smooth curve into a spatial frequency domain. The peak frequency or centroid frequency of the frequency band, the ratio of the integrated value of the component size over the entire frequency band obtained by similar conversion and the integrated value of the component size of the predetermined frequency band, or similar conversion It can be set as the upper limit or lower limit frequency value of the frequency distribution band having a predetermined ratio of the integral values of the component sizes in the entire frequency band.

  The degree of branching may be the number of branches of the blood vessel image or the average branching angle of each branch point.

  Further, the separating means includes grouping means for grouping the blood vessel segments into the same distance range from the tumor image, and the calculating means is a dispersion of a separation length between the blood vessel segments and the smooth curve, The ratio between the length of the blood vessel segment and the distance from the beginning to the end of the blood vessel segment, the average or the sum or the maximum of the separation length of the blood vessel segment and the smooth curve, the distance from the start to the end of the blood vessel segment and the smooth curve Peak frequency or barycentric frequency of the frequency band obtained by converting the separation length curve indicating the separation length for each distance into the spatial frequency domain, the integrated value of the component size over the entire frequency band obtained by the same conversion, and a predetermined frequency It has a ratio with the integrated value of the component size of the band, or a predetermined ratio of the integrated value of the component size over the entire frequency band obtained by the same conversion. The upper or lower frequency value of the frequency distribution band, the ratio of the total length of each vascular segment and the total distance between the start and end, the average distance between the start and end of each vascular segment, or the The total or average of the length is calculated for each group as a group value, and the normalized value obtained by weighting the group value according to the representative distance between the group and the tumor image is calculated as the meandering degree. Also good.

  And a dividing unit that divides the space of the medical image into a plurality of regions according to a distance from the tumor image, wherein the calculating unit calculates the number of branches of the blood vessel image or an average branching angle of each branch point. A normalization value obtained by weighting according to a representative distance between the region and the tumor image may be calculated as the degree of branching.

  According to the present invention, a blood vessel image existing around a tumor image is extracted from a medical image, and the meandering degree of the blood vessel image is calculated. In addition, the branching degree of each branch point existing in the blood vessel image is calculated. Based on the knowledge that blood vessels around malignant tumors are complicated to bend and branch, the degree of meandering and branching of blood vessels can be used as a malignant index of tumor. Can be reduced.

  Also, by correcting the meandering degree and branching degree of blood vessels according to the distance from the tumor image, based on the knowledge that the closer to the tumor, the more likely the influence of the malignant tumor on the blood vessels is, The degree of branching can be used as a more accurate tumor malignancy index, and the false positive rate by the tumor diagnosis support system can be further reduced.

  Hereinafter, preferred embodiments of the tumor diagnosis support technology according to the present invention will be described in detail with reference to the drawings.

(Basic configuration)
FIG. 1 is a block diagram illustrating a configuration of a tumor diagnosis support system that performs tumor diagnosis support. The tumor diagnosis support system has a configuration in which an input device 2 and a display device 3 or a printing device 4 are connected to a workstation 1 as an input / output device.

  The workstation 1 is a so-called computer and includes an arithmetic control unit 5, a main storage unit 6, and an external storage unit 7 therein. The external storage unit 7 is an HDD (Hard Disk Drive), for example, and stores data and programs. The external storage unit 7 stores an operating system program and a tumor diagnosis support program. The arithmetic control unit 5 is a CPU or the like (Central Processing Unit or graphic chip), and appropriately decodes and executes a tumor diagnosis support program to control arithmetic and peripheral devices. The main storage unit 6 is a RAM (Random Access Memory), and calculation results, data, and a tumor diagnosis support program are developed as a work area of the calculation control unit 5.

  The input device 2 is a keyboard or a mouse having a wheel tracking function, and outputs an operation signal to the arithmetic control unit 5 in accordance with a user operation. The display device 3 is an LCD (Liquid Crystal Display) display or a CRT (Cathode Ray Tube) display, and outputs the drawing data created by the arithmetic control unit 5 so as to be visible on the screen. The printing device 4 is a printer that forms an image on a paper medium according to the drawing data created by the arithmetic control unit 5.

  In this tumor diagnosis support system, the arithmetic control unit 5 decodes and executes a tumor diagnosis support program stored in the external storage unit 7, so that the shape of the blood vessel image around the tumor image existing in the medical image is detected. The degree of meandering and branching is calculated as a malignant tumor index. The calculation of the meandering degree and the degree of branching and the visible output assist the doctor in determining whether the tumor image is malignant or benign, or whether the meandering degree or degree of branching is malignant Determine if it is.

  The degree of meandering is the degree of blood vessel bending. The degree of branching is the degree of branching of the blood vessel tree.

  FIG. 2 is a schematic diagram showing a benign tumor and a blood vessel image existing in the vicinity thereof, and FIG. 3 is a schematic diagram showing a malignant tumor and a blood vessel image existing in the vicinity thereof. As shown in FIGS. 2 and 3, the blood vessel image 200b existing around the malignant tumor 100b is more complex than the blood vessel image 200a existing around the benign tumor 100a, and the blood vessels meander. You can see that Since the malignant tumor 100b grows by gathering surrounding tissues, blood vessels gather in the surrounding region of the tumor. As a result, the blood vessel image 200b is bent and the density of the blood vessels is increased. When the density of blood vessels is high, the degree of branching of blood vessels around the malignant tumor appears high. The degree of bending and the degree of branching tends to be higher as the malignant tumor is closer.

  In the tumor diagnosis support system according to the present embodiment, the influence of the malignant tumor 100b on the blood vessel is grasped from the viewpoint of meandering and branching, and it is determined whether or not the tumor is malignant or benign.

  Hereinafter, the determination or determination support technique based on the meandering degree and the determination or determination support technique based on the branching degree will be described separately.

(Tumor diagnosis support system for meandering degree calculation)
FIG. 4 is a functional block diagram showing a tumor diagnosis support system for calculating the degree of meandering. The system composed of computers is shown mainly with the functions realized by starting the tumor diagnosis support program.

  This tumor diagnosis support system detects a tumor image from a medical image, extracts a blood vessel image existing around the tumor image, extracts the blood vessel image into segments, calculates a meandering degree for each segment, Malignant or benign is judged and the meandering degree and judgment result are output so as to be visible.

  The tumor image detection configuration includes a medical image storage unit 10, a tumor image detection unit 20, and a tumor image center point storage unit 25.

  The medical image storage unit 10 includes an external storage unit 7. The medical image storage unit 10 stores medical images. A medical image is an image in a subject obtained by imaging the subject with an X-ray diagnostic apparatus, an X-ray CT apparatus, an MRI apparatus, an ultrasonic diagnostic apparatus, or a nuclear medicine diagnostic apparatus. It is an image in which a tumor image and surrounding blood vessel images exist, and is a two-dimensional image or a three-dimensional image.

  The tumor image detection unit 20 includes a display device 3, an input device 2, and a calculation control unit 5. The tumor image center point storage unit 25 includes the main storage unit 6. The tumor image detection unit 20 detects a tumor image from a medical image and calculates the coordinates of the center point. When a medical image is displayed on the display device 3 and a point on the tumor image is designated using the input device 2, the shape of the tumor image is specified by the region expansion method or the like around that point. When the shape of the tumor image is specified, the average of the coordinate positions of the voxels constituting the tumor image is taken, and the center point of the tumor image is calculated. The detection of the tumor image is to obtain a base point from which a surrounding blood vessel image is extracted in order to determine whether the tumor image is malignant or benign, and the coordinates of the center point are the base point. The coordinates of the center point are stored in the tumor image center point storage unit 25.

  As a blood vessel image extraction configuration, a blood vessel image extraction unit 30 and a centerline storage unit 35 are provided.

  The blood vessel image extraction unit 30 includes a display device 3, an input device 2, and an arithmetic control unit 5. The centerline storage unit 35 includes the main storage unit 6. The blood vessel image extraction unit 30 extracts a blood vessel image and calculates a center line of the blood vessel image. When a medical image is displayed and a point on the blood vessel image is designated using the input device 2, the shape of the blood vessel image is specified by the region expansion method or the like around that point. The range of the blood vessel image that specifies the shape is a predetermined range that affects the blood vessel shape by the tumor image around the center point of the tumor image stored in the tumor image center point storage unit 25. The center line of the blood vessel image is calculated by a thinning method or the like. The calculated center line is stored in the center line storage unit 35. This center line indicates the extending shape of the blood vessel.

  As a configuration for dividing the blood vessel image into segments, a segment separation unit 40 and a segment storage unit 45 are provided.

  The segment separation unit 40 includes the calculation control unit 5. The segment storage unit 45 includes the main storage unit 6. The segment separation unit 40 divides the blood vessel image into blood vessel segments. A blood vessel segment is a center line without a branch existing between two adjacent branch points of a blood vessel image spreading in a branch shape. This is because the meandering degree of the blood vessel is calculated by eliminating the influence of the branching on the recognition of the bending of the blood vessel.

  The configuration for calculating the meandering degree includes a meandering degree calculating unit 50, a smooth curve storing unit 54, and a meandering degree storing unit 58.

  The meandering degree calculation unit 50 includes an arithmetic control unit 5. The meandering degree calculation unit 50 calculates the meandering degree from the bending shape of the blood vessel segment. As will be described later, the meandering degree is the dispersion of the separation length between the smooth curve obtained by smoothing the extending shape of the tortuous blood vessel segment and the center line of the blood vessel segment, the average of the separation length, the total of the separation length, or the maximum of the separation length. Value, or the ratio of the length of the blood vessel segment to the distance from the start to the end, or the peak frequency, barycentric frequency indicated by the spatial frequency spectrum of the blood vessel segment, the ratio of the total integral value to the integral value of the predetermined frequency region, or the total integral value Is a frequency corresponding to an integral value corresponding to a predetermined ratio. The meandering degree calculation unit 50 generates a smooth curve obtained by smoothing the extension shape of the blood vessel segment in advance. The generated smooth curve is stored in the smooth curve storage unit 54 and is called when the meandering degree is calculated. The smooth curve storage unit 54 includes the main storage unit 6. The calculated meandering degree is stored in the meandering degree storage unit 58. The meandering degree storage unit 58 includes the external storage unit 7.

  The configuration for determining malignancy or benign tumor from the meandering degree includes a determination unit 60 and a threshold storage unit 65, and an output unit 70 as an output configuration.

  The determination unit 60 includes the calculation control unit 5. The threshold storage unit 65 includes the external storage unit 7. The determination unit 60 compares the threshold stored in the threshold storage unit 65 in advance with the degree of meandering to determine whether the tumor is malignant or benign.

  The output unit 70 includes the display device 3 or the printing device 4. The output unit 70 outputs the calculated meandering degree and the determination result of malignancy or benignity so as to be visible.

  FIG. 5 is a schematic flowchart showing a tumor diagnosis support operation performed by the tumor diagnosis support system.

  First, the tumor diagnosis support system reads a medical image from the medical image storage unit 10 and displays it on the display device 3 (S01). A medical image and a cursor for designating a position on the display screen are displayed on the display screen. The movement destination of the cursor is redrawn according to the direction instruction operation using the input device 2.

  When a tumor image is selected (S02), the tumor image detection unit 20 identifies and detects the tumor image (S03).

  FIG. 6 is a schematic diagram showing tumor image detection and blood vessel image extraction. The selection of the tumor image is performed using the input device 2. The user uses the input device 2 to position the cursor on the tumor image and point to a point on the tumor image. When the cursor is positioned on the tumor image by the direction instruction operation using the input device 2 and a point on the tumor image is pointed out by pressing the enter button, the tumor image detection unit 20 expands the region around the pointed point. The shape of the tumor image is identified by the method.

  In the region expansion method, the pixel values of surrounding voxels are checked with the determined region as the center, and if the pixel value is equal to or greater than a predetermined value, it is incorporated into the determined region. The determination of this area and the search for the surrounding pixel values are repeated, and the determined area is gradually expanded. The predetermined value is a threshold value indicating that a tumor image tissue exists. If the medical image is grayscale, the predetermined value is a boundary value between the tissue and the non-tissue, or a pixel value of one point indicated. The voxel for checking the pixel value is a voxel located in the non-determined area among the 23 points around the voxel located at the boundary of the decided area. The surrounding pixel values are searched, and when any voxel becomes less than the predetermined value, the detection of the tumor image is terminated.

  When the tumor image is detected, the tumor image detection unit 20 calculates the center point of the detected tumor image (S04). When the center point is calculated, the blood vessel image extraction unit 30 determines a region of interest around the center point (S05).

  The center point of the tumor image is simply the average of the coordinates of each voxel constituting the tumor image. As shown in FIG. 6, the blood vessel image extraction unit 30 sets a predetermined range around the center point as a region of interest. This region of interest may be determined manually by selecting the boundary point of the region of interest with a cursor. The region of interest is a range where blood vessel images are extracted.

  When the region of interest is determined, the blood vessel image extraction unit 30 extracts a blood vessel image around the tumor image existing in the region of interest (S06). Extraction of the blood vessel image is performed by a region expansion method from one point on the blood vessel indicated by the input device 2.

  When the blood vessel image is extracted, the blood vessel image extraction unit 30 specifies the center line of the blood vessel image (S07).

  The center line of the blood vessel image is specified by thinning the blood vessel image, and the extended shape of the blood vessel appears. Thinning is performed by gradually removing voxels located at the contour of the blood vessel image from the blood vessel image. Until the continuity of the voxels constituting the blood vessel image is interrupted, the voxels located in the contour are gradually removed to narrow the contour.

  When the center line of the blood vessel image is specified, the segment separation unit 40 segments the blood vessel image and divides it into blood vessel segments (S08). Each blood vessel segment is stored in the segment storage unit 45.

  The segment separation unit 40 reads a center line as a blood vessel image, divides the center line at a branch point, and acquires a blood vessel segment. The branch point is determined for each determination point by scanning the determination point along the center line from the start end of the read center line. It is determined whether or not there are a plurality of voxels constituting the center line on a predetermined voxel line passing through the determination point and perpendicular to the center line. As a result of the determination, if there are a plurality of voxels constituting the center line, the determination point before that is set as a branch point.

  When the blood vessel image is divided into blood vessel segments, the meandering degree calculation unit 50 sequentially reads out each blood vessel segment from the segment storage unit 45, and performs processing for calculating the meandering degree described later for each blood vessel segment (S09).

  When the meandering degree is calculated, the determination unit 60 reads the threshold value from the threshold value storage unit 65, and compares the meandering degree with the threshold value (S10). As a result of the comparison, if there is a meandering degree that is equal to or greater than the threshold value (S10, Yes), the output unit 70 determines the malignancy of “suspected malignant” together with each meandering degree stored in the meandering degree storage unit 58. A character string indicating the result is output (S11). On the other hand, when there is no meandering degree that is equal to or greater than the threshold (S10, No), the output unit 70 indicates the judgment result “possibility of benign” together with each meandering degree stored in the meandering degree storage unit 58. A character string is output (S12).

(First embodiment of meandering degree calculation: dispersion of separation length)
Next, a first embodiment by the meandering degree calculation unit 50 will be described.

  FIG. 7 is a schematic diagram of the meandering degree calculation according to the first embodiment. As shown in FIG. 7, the meandering degree calculation unit 50 calculates the variance of the separation lengths 230 between the smooth curve 220 and the center line 210 of the blood vessel segment as the meandering degree.

  The smooth curve 220 is a curve obtained by smoothing the extending shape of the blood vessel segment. It is obtained by generating a B-spline curve using each point of the center line 210 as a control point. Alternatively, each voxel of the center line 210 may be sequentially set as a reference point, and a moving average coordinate of coordinate values of surrounding voxels including a machine order point may be set as a point on the smooth curve 220 with respect to the reference point.

  Each separation length 230 is a distance from each point on the center line 210 to the smooth curve. Each point on the center line 210 is sequentially set as a survey point, and a line with a plurality of angles is drawn toward the smooth curve 220. Among these line lengths, the minimum value is the separation length at the survey point.

  Variance is the statistical sample variance. Each separation length 230 is used as a sample, and indicates how much the sample deviates from the sample average. That is, the meandering degree calculation unit 50 obtains a sample average by taking the average of the separation lengths 230 and sets the meandering value as the mean value of the square of the difference between the sample average and each separation length 230.

  FIG. 8 is a flowchart showing the meandering degree calculating operation of the meandering degree calculating unit 50 according to the first embodiment.

  First, the meandering degree calculation unit 50 calculates a smooth curve obtained by smoothing the extending shape of the blood vessel segment by the B-spline method using each point of the center line 210 as a control point (S21). The smooth curve 220 is stored in the smooth curve storage unit 54.

  Next, a separation length 230 is sequentially calculated for each point on the center line 210 of the blood vessel segment (S22).

  When the separation length 230 at each point on the center line 210 is calculated, the variance of the separation length 230 is calculated as the meandering degree (S23). This meandering degree is calculated for each blood vessel segment. The calculated meandering degree is stored in the meandering degree storage unit 58.

(Second embodiment of meandering degree calculation: average of separation length)
Next, a second embodiment by the meandering degree calculation unit 50 will be described. The meandering degree calculation unit 50 according to the second embodiment calculates an average of the separation length 230 between the smooth curve 220 and the center line 210 as the meandering degree.

  FIG. 9 is a flowchart showing the meandering degree calculating operation of the meandering degree calculating unit 50 according to the second embodiment.

  First, the meandering degree calculation unit 50 calculates a smooth curve 220 obtained by smoothing the extending shape of the blood vessel segment by the B-spline method using each point of the center line 210 as a control point (S31). The smooth curve is stored in the smooth curve storage unit 54.

  Next, a separation length 230 is sequentially calculated for each point on the center line 210 of the blood vessel segment (S32).

  When the separation length at each point on the center line 210 is calculated, an average of these separation lengths 230 is calculated as the meandering degree (S33). This meandering degree is calculated for each blood vessel segment. The calculated meandering degree is stored in the meandering degree storage unit 58.

(Third embodiment of meandering degree calculation: total separation length)
Next, a third embodiment by the meandering degree calculation unit 50 will be described. The meandering degree calculation unit 50 according to the third embodiment calculates the total of the separation length 230 between the smooth curve 220 and the center line 210 as the meandering degree.

  FIG. 10 is a flowchart showing the meandering degree calculating operation of the meandering degree calculating unit 50 according to the third embodiment.

  First, the meandering degree calculation unit 50 calculates a smooth curve 220 obtained by smoothing the extending shape of the blood vessel segment by the B-spline method using each point of the center line 210 as a control point (S41). The smooth curve 220 is stored in the smooth curve storage unit 54.

  Next, the separation length 230 is sequentially calculated for each point on the center line 210 of the blood vessel segment (S42).

  When the separation length 230 at each point on the center line 210 is calculated, the total of the separation lengths 230 is calculated as the meandering degree (S43). This meandering degree is calculated for each blood vessel segment. The calculated meandering degree is stored in the meandering degree storage unit 58.

(Fourth embodiment of meandering degree calculation: maximum separation length)
Next, a fourth embodiment by the meandering degree calculation unit 50 will be described. The meandering degree calculation unit 50 according to the fourth embodiment extracts the maximum value of the separation length 230 between the smooth curve 220 and the center line 210 as the meandering degree.

  FIG. 11 is a flowchart showing the meandering degree calculating operation of the meandering degree calculating unit 50 according to the fourth embodiment.

  First, the meandering degree calculation unit 50 calculates a smooth curve 220 obtained by smoothing the extending shape of the blood vessel segment by the B-spline method using each point of the center line 210 as a control point (S51). The smooth curve 220 is stored in the smooth curve storage unit 54.

  Next, a separation length 230 is sequentially calculated for each point on the center line 210 of the blood vessel segment (S52).

  When the separation length 230 at each point on the center line 210 is calculated, the maximum value is extracted from the separation length 230 as the meandering degree (S53). This meandering degree is calculated for each blood vessel segment. The calculated meandering degree is stored in the meandering degree storage unit 58.

(Fifth embodiment of meandering degree calculation: ratio of distance to length)
Next, a fifth embodiment by the meandering degree calculation unit 50 will be described.

  FIG. 12 is a schematic diagram of the meandering degree calculation according to the fifth embodiment. As shown in FIG. 12, the meandering degree calculation unit 50 calculates the ratio between the distance 250 of the blood vessel segment and the total length 240 as the meandering degree. The blood vessel segment distance 250 is a distance obtained by linearly connecting the start and end of the center line. The total length 240 of the blood vessel segment is the length of the center line.

  FIG. 13 is a flowchart showing the meandering degree calculating operation of the meandering degree calculating unit 50 according to the fifth embodiment.

  First, the meandering degree calculation unit 50 reads the coordinates of the start and end of the center line 210, and calculates the distance 250 between the start and end of the blood vessel segment from the coordinates of the start and start (S61).

  Next, the distance between the points on the center line 210 of the blood vessel segment is accumulated, and the total length 240 of the blood vessel segment is calculated (S62).

  When the distance 250 and the total length 240 of the start and end of the blood vessel segment are calculated, the ratio of the distance 250 and the total length 240 is calculated as the meandering degree (S63). This meandering degree is calculated for each blood vessel segment. The calculated meandering degree is stored in the meandering degree storage unit 58.

(Sixth embodiment of meandering degree calculation: peak frequency)
Next, a sixth embodiment by the meandering degree calculation unit 50 will be described.

  The meandering degree calculation unit 50 according to the sixth embodiment extracts a frequency having a peak value from the spatial frequency spectrum of the blood vessel segment as the meandering degree. FIG. 14 is a flowchart showing the operation of the meandering degree calculation unit 50 according to the sixth embodiment.

  First, the meandering degree calculation unit 50 calculates a smooth curve 220 obtained by smoothing the extending shape of the blood vessel segment by the B-spline method using each point of the center line 210 as a control point (S71). The smooth curve 220 is stored in the smooth curve storage unit 54.

  Next, the separation length 230 is sequentially calculated for each point on the center line 210 of the blood vessel segment (S72).

  When the separation length 230 at each point on the center line 210 is calculated, the meandering degree calculation unit 50 generates a separation length curve 260 from the center line 210 indicating the extending shape of the blood vessel segment and the smooth curve 220 (S73).

  FIG. 15 is a diagram illustrating a separation length curve generated by the meandering degree calculation unit 50. The separation length curve 260 is a plot of the separation length 230 from each point on the center line 210 to the smooth curve 220 corresponding to each distance from the start to the end.

  When the separation length curve 260 is generated, the meandering degree calculation unit 50 obtains a spatial frequency spectrum of the blood vessel segment by converting the separation length curve 260 into spatial frequency domain data by Fourier transform or the like (S74).

  FIG. 16 is a spectrum diagram obtained by converting the separation length curve 260 into the spatial frequency domain. This spectrum is a plot of the magnitude of the frequency component corresponding to the frequency. The magnitude of the frequency component represents how much the frequency appears. When a spectrum appears in a high frequency band, it indicates that the blood vessel segment meanders in a complicated manner.

  When converted to the spatial frequency domain, the meandering degree calculation unit 50 searches for the peak frequency Pf having the peak value Pv of the magnitude of the frequency component from the spectrum shown in FIG. 16 as the meandering degree (S75). This meandering degree is calculated for each blood vessel segment. The calculated meandering degree is stored in the meandering degree storage unit 58.

(Seventh embodiment of meandering degree calculation: center-of-gravity frequency)
Next, a seventh embodiment by the meandering degree calculation unit 50 will be described.

  The meandering degree calculation unit 50 according to the seventh embodiment extracts the centroid frequency of the spatial frequency spectrum of the blood vessel segment as the meandering degree. The centroid frequency is a frequency corresponding to the centroid of the region surrounded by the envelope of the spatial frequency spectrum and the frequency axis. FIG. 17 is a flowchart showing the operation of the meandering degree calculation unit 50 according to the seventh embodiment.

  First, the meandering degree calculation unit 50 calculates a smooth curve 220 obtained by smoothing the extending shape of the blood vessel segment by the B-spline method using each point of the center line 210 as a control point (S81). The smooth curve 220 is stored in the smooth curve storage unit 54.

  Next, a separation length 230 is sequentially calculated for each point on the center line 210 of the blood vessel segment (S82).

  When the separation length 230 at each point on the center line 210 is calculated, the meandering degree calculation unit 50 calculates the smooth curve 220 and then separates the separation length curve from the center line 210 and the smooth curve 220 indicating the extending shape of the blood vessel segment. 260 is generated (S83).

  When the separation length curve 260 is generated, the meandering degree calculation unit 50 converts the separation length curve 260 into a spatial frequency domain by Fourier transform or the like (S84).

  When converted to the spatial frequency domain, the meandering degree calculation unit 50 calculates the center-of-gravity frequency from the spatial frequency spectrum as the meandering degree (S85). This meandering degree is calculated for each blood vessel segment. The calculated meandering degree is stored in the meandering degree storage unit 58.

(Eighth embodiment of meandering degree calculation: integral ratio of whole spectrum and high frequency band)
Next, an eighth embodiment by the meandering degree calculation unit 50 will be described.

  The meandering degree calculation unit 50 according to the eighth embodiment calculates, as the meandering degree, an integration ratio of the component sizes in the entire frequency band indicated by the spatial frequency spectrum of the blood vessel segment and the predetermined wavenumber band. FIG. 18 is a flowchart showing the operation of the meandering degree calculation unit 50 according to the eighth embodiment.

  First, the meandering degree calculation unit 50 calculates a smooth curve 220 obtained by smoothing the extending shape of the blood vessel segment by the B-spline method using each point of the center line 210 as a control point (S91). The smooth curve 220 is stored in the smooth curve storage unit 54.

  Next, the separation length 230 is sequentially calculated for each point on the center line 210 of the blood vessel segment (S92).

  When the separation length 230 at each point on the center line 210 is calculated, the meandering degree calculation unit 50 calculates the smooth curve 220 and then separates the separation length curve from the center line 210 and the smooth curve 220 indicating the extending shape of the blood vessel segment. 260 is generated (S93).

  When the separation length curve 260 is generated, the meandering degree calculation unit 50 converts the separation length curve 260 into a spatial frequency domain by Fourier transform or the like (S94).

  FIG. 19 is a schematic diagram of a spatial frequency spectrum showing the total integral value and the predetermined band integral value. When the frequency decomposition is performed, the meandering degree calculation unit 50 integrates the magnitudes of the frequency components over the entire frequency band of the spatial frequency spectrum to obtain an overall integration value IVa (S95). The total integral value IVa is an area surrounded by the envelope of the spatial frequency spectrum and the frequency axis. Further, the meandering degree calculation unit 50 integrates the magnitude of the frequency component over the predetermined frequency band fr of the spatial frequency spectrum to obtain the predetermined band integration value IVp (S96). The predetermined band integral value IVp is an area surrounded by the envelope portion and the frequency axis corresponding to the predetermined frequency band fr of the spatial frequency spectrum.

  The meandering degree calculation unit 50 stores the upper limit frequency Hf and the lower limit frequency Lf of the predetermined frequency band fr as predetermined values. In order to calculate the meandering degree, it is preferable to store an upper limit frequency Hf and a lower limit frequency Lf so that the predetermined frequency band fr is a high frequency band.

  When the total integral value IVa and the predetermined band integral value IVp are obtained, the meandering degree calculation unit 50 calculates the ratio of the total integral value IVa and the predetermined band integral value IVp as the meandering degree (S97). This meandering degree is calculated for each blood vessel segment. The calculated meandering degree is stored in the meandering degree storage unit 58.

(Ninth embodiment of meandering degree calculation: upper limit or lower limit of frequency band having integral value of predetermined ratio)
Next, a ninth embodiment by the meandering degree calculation unit 50 will be described.

  The meandering degree calculation unit 50 according to the ninth embodiment calculates an upper limit or a lower limit frequency value of a frequency band having an integral value of a predetermined ratio as the meandering degree. The predetermined ratio is a ratio with respect to the integral value of the magnitude of the frequency component over the entire frequency band. FIG. 20 is a flowchart showing the operation of the meandering degree calculation unit 50 according to the ninth embodiment.

  First, the meandering degree calculation unit 50 calculates a smooth curve 220 obtained by smoothing the extending shape of the blood vessel segment by the B-spline method using each point of the center line 210 as a control point (S101). The smooth curve 220 is stored in the smooth curve storage unit 54.

  Next, the separation length 230 is sequentially calculated for each point on the center line 210 of the blood vessel segment (S102).

  When the separation length 230 at each point on the center line 210 is calculated, the meandering degree calculation unit 50 calculates the smooth curve 220 and then separates the separation length curve from the center line 210 and the smooth curve 220 indicating the extending shape of the blood vessel segment. 260 is generated (S103).

  When the separation length curve 260 is generated, the meandering degree calculation unit 50 converts the separation length curve 260 into a spatial frequency domain by Fourier transform or the like (S104). FIG. 21 is a schematic diagram of a spatial frequency spectrum showing the total integral value IVa, the survey integration value IVr, and the survey frequency Rf.

  When converted to the spatial frequency domain, the meandering degree calculation unit 50 integrates the magnitudes of the frequency components over the entire frequency band of the spatial frequency spectrum to obtain an overall integral value IVa (S105). The total integral value IVa is an area surrounded by the envelope of the spatial frequency spectrum and the frequency axis.

  Further, the meandering degree calculation unit 50 sets a predetermined survey frequency Rf lower than the maximum frequency Maxf indicated by the spatial frequency spectrum (S106), and for the frequency band between the maximum frequency Maxf and the survey frequency Rf, the frequency band A survey integration value IVr is obtained by integrating the magnitudes of the frequency components over (S107). The initial value of the survey frequency Rf is stored in the external storage unit 7 in advance. The survey integration value IVr is an area surrounded by the envelope portion of the spatial frequency spectrum starting from the highest frequency Maxf and ending at the survey frequency Rf and the frequency axis.

  Next, a ratio between the total integral value IVa and the survey integral value IVr is calculated (S108), and it is determined whether or not the calculated ratio is a predetermined ratio (S109). A value corresponding to the predetermined ratio is stored in the external storage unit 7 in advance. For example, the predetermined ratio is 50%.

  If the calculated ratio is smaller than the predetermined ratio (S109, No), the survey frequency Rf is set again to a frequency lower by the predetermined frequency (S110), and S107 to S110 are repeated. That is, the survey frequency Rf is the lower limit frequency of the frequency band between the highest frequency Maxf and the survey frequency Rf.

  If the calculated ratio is the same value as the predetermined ratio (S109, Yes), the set survey frequency Rf is set as the meandering degree and stored in the meandering degree storage unit 58 (S111).

  Note that a frequency band between the lowest frequency of the spatial frequency spectrum and the survey frequency Rf may be used. In this case, the survey frequency Rf is an upper limit frequency of this frequency band.

  As described above, in this tumor diagnosis support system, the meandering degree is calculated in any one of the embodiments, and the judgment unit 60 compares the meandering degree with a threshold value to determine whether the tumor image is malignant or benign, and the output unit At 70, the determination result is output together with the meandering degree so as to be visible.

(Tumor diagnosis support system for branching degree calculation)
Next, a technique for calculating the degree of branching of a blood vessel image present around a tumor image to assist tumor diagnosis will be described.

  FIG. 22 is a functional block diagram showing a tumor diagnosis support system for calculating the degree of branching. The system composed of computers is shown mainly with the functions realized by starting the tumor diagnosis support program.

  This tumor diagnosis support system detects a tumor image from a medical image, extracts a blood vessel image existing around the tumor image, calculates a branching degree in the blood vessel image, determines malignancy or benignity from the branching degree, determines the branching degree and The determination result is output so as to be visible.

  The detection configuration of the tumor image, the extraction configuration of the blood vessel image, the configuration for determining the malignancy or benignity of the tumor from the degree of meandering, and the output configuration are the same as those of the tumor diagnosis support system for calculating the degree of meandering. , Tumor image detection unit 20, tumor image center point storage unit 25, blood vessel image extraction unit 30, center line storage unit 35, determination unit 60, threshold storage unit 65, and output unit 70.

  As a configuration for calculating the branching degree, a branching degree calculation unit 80 and a branching degree storage unit 85 are provided.

  The branching degree calculation unit 80 includes the display device 3 and the calculation control unit 5. The branching degree calculation unit 80 calculates the branching degree from the branching of the blood vessel image. As will be described later, the degree of branching is the number of branches or the average branching angle. The degree of branching is performed by scanning the blood vessel image from one end to the other end. One end of the blood vessel image is on the side close to the heart, a medical image is displayed, and is designated by the user using the input device 2. The calculated branching degree is stored in the branching degree storage unit 85. The branching degree storage unit 85 includes the external storage unit 7.

  FIG. 23 is a schematic flowchart showing a tumor diagnosis support operation performed by the tumor diagnosis support system.

  First, the tumor diagnosis support system reads a medical image from the medical image storage unit 10 and displays it on the display device 3 (S121). When a tumor image is selected (S122), the tumor image detection unit 20 identifies and detects the tumor image (S123).

  When the tumor image is detected, the tumor image detection unit 20 calculates the center point of the detected tumor image (S124). When the center point is calculated, the blood vessel image extraction unit 30 determines a region of interest around the center point (S125). When the region of interest is determined, the blood vessel image extraction unit 30 extracts a blood vessel image around the tumor image existing in the region of interest (S126).

  When the blood vessel image is extracted, the blood vessel image extraction unit 30 specifies the center line of the blood vessel image (S127).

  When the center line of the blood vessel image is specified, the branching degree calculation unit 80 reads the center line of the blood vessel image from the center line storage unit 35 and performs a process of calculating a branching degree described later from each branch point of the center line ( S128).

  When the branching degree is calculated, the determination unit 60 reads the threshold value from the threshold value storage unit 65 and compares the branching degree with the threshold value (S129). As a result of the comparison, if there is a branching degree that is equal to or greater than the threshold value (S129, Yes), the output unit 70 determines the malignancy of “suspected malignant” together with each branching degree stored in the branching degree storage unit 85. A character string indicating the result is output (S130). On the other hand, when there is no degree of branching that is equal to or greater than the threshold (S129, No), the output unit 70 indicates the determination result “possibility of benign” together with each degree of branching stored in the degree of branching storage unit 85. A character string is output (S131).

(First embodiment of branching degree calculation: number of branches)
Next, a first embodiment of branching degree calculation by the branching degree calculation unit 80 will be described.

  FIG. 24 is a schematic diagram of branching degree calculation according to the first embodiment. The blood vessel image is configured by branching blood vessels into branches. The branching degree calculation unit 80 detects branch points 270 from the blood vessel image and counts the number of branch points 270. The search point 280 is sequentially shifted from one end of the center line 210 to the downstream and the branch point 270 is searched. One end of the center line 210 is determined as an end 290 closer to the heart. When the branch point 270 is found, the number of branches is counted up.

  FIG. 25 is a flowchart showing the branching degree calculating operation of the branching degree calculating unit 80 according to the first embodiment.

  First, the branching degree calculation unit 80 secures a storage area for storing the branching point 270, the scanned points and the unscanned points in the main storage unit 6, initializes the number of branches to 0, and displays a medical image. (S141). The scanned point is a point on the center line ahead of the branch point and has already become the survey point 280. The unscanned point is a point on the center line ahead of the branch point and has not yet become the survey point 280.

  When an end 290 on the side close to the heart of the blood vessel image is designated from the medical image using the input device 2 (S142), one end on the designated side of the center line 210 is set as an upstream to the investigation point 280 (S143). . When the blood vessel image has already branched into a plurality of blood vessels from outside the region of interest, or when there are a plurality of blood vessel images in the region of interest, the user designates all ends of the blood vessel image close to the heart.

  When the survey point 280 is set, the downstream adjacent points of the center line 210 with respect to the survey point 280 are counted (S144), and the number obtained by subtracting 1 from the count obtained by the count is added to the number of branches (S145). When the count number is 1, since the count number is 0 instead of the branch point, there is no change in the branch number.

  When the update of the count and the number of branches is completed, the survey point 280 is shifted to one downstream side on the center line 210.

  At this time, if the count number is 2 or more (S146, Yes), the survey point 280 is transferred to the adjacent point having the smallest coordinate value excluding the scanned point, and the storage as the unscanned point is deleted (S147). . The coordinate value is, for example, a scalar quantity. Then, the newly set survey point 280 is stored in the storage area as the scanned point, and the branch destination point that has not yet been scanned is stored in the storage area (S148). When newly scanned points and unscanned points are stored, S144 to S148 are repeated, and branch points are counted.

  When the count number is not 2 or more (S146, No), that is, when the survey point 280 is not the branch point 270, it is determined whether the survey point 280 can be shifted to one downstream side on the center line 210 (S149).

  When a point on the center line 210 exists on the downstream side of the survey point 280 (S149, Yes), the survey point 280 is moved to an existing adjacent point (S150), and S144 to S148 are repeated.

  If there is no point on the centerline 210 downstream of the survey point 280 (S149, No), that is, if the survey point 280 reaches the other downstream end of the centerline 210, is there an unscanned point? Judgment is made (S151). When an unscanned point exists (S151, Yes), the survey point 280 is set to the lowest unscanned point on the most downstream side (S152), and S144 to S148 are repeated.

  On the other hand, when there is no unscanned point (S151, No), the counting of the number of branches ends. The counted branch number is stored in the branch degree storage unit 85 as a branch degree.

(Second embodiment of branching degree calculation: average branching angle)
Next, a second embodiment by the branching degree calculation unit 80 will be described.

  FIG. 26 is a schematic diagram of branching degree calculation according to the second embodiment. The blood vessel image is configured by branching blood vessels into branches. The branching degree calculation unit 80 detects a branch point 270 from the blood vessel image and calculates a branch angle 300 of the branch point 270. The branch angle 300 is calculated for each branch point 270, and the average branch angle is obtained as the branch degree.

  The branch angle 300 is an angle formed by two center lines 210 extending from the branch point 270. The coordinates of the branch point 310 separated by a predetermined voxel from the branch point 270 on the branch center line 210 are acquired, and the branch angle 300 is calculated from the coordinates of the branch point 270 and the coordinates of the two branch points 310.

  FIG. 27 is a flowchart showing the branching degree calculating operation of the branching degree calculating unit 80 according to the second embodiment.

  First, the branching degree calculation unit 80 runs the survey point 280 on the center line 210, and sequentially detects the branch point 270 (S161). The detected coordinates of the branch point 270 are temporarily stored in the main storage unit 6.

  When the branch point 270 is detected, the coordinates of the branch point 300 separated by a predetermined voxel from the branch point 270 on the center line 210 branching from the branch point 270 are acquired (S162). When the coordinates of the two branch points 310 are acquired, the branch angle 300 formed by the two branches of the center line 210 is sequentially calculated for each branch point 270 from the coordinates of the branch point 270 and the two branch points 310 (S163).

  When the branch angle 300 is calculated for all branch points 270 (S164, Yes), the average of all the calculated branch angles 300 is taken, and the average branch angle is acquired as the branch degree (S165). The calculated average branch angle is stored in the branch degree storage unit 85 as a branch degree.

  As described above, in this tumor diagnosis support system, the degree of branching is calculated in any of the embodiments, and the determination unit 60 compares the degree of branching with the threshold value to determine whether the tumor image is malignant or benign. At 70, the determination result is output together with the degree of branching so as to be visible.

(Modification example of meandering degree and branching degree calculation)
Next, modified examples of the above-described meandering degree and branching degree will be described.

  A malignant tumor grows by gathering surrounding tissues, and the influence on the surrounding tissues is proportional to the distance from the malignant tumor. That is, the closer to the malignant tumor, the higher the aggregation rate of surrounding tissues. Therefore, the degree of meandering and branching tends to increase as the malignant tumor is closer. Therefore, in the tumor diagnosis support system according to the modified example of the meandering degree and the degree of branching, the meandering degree and the degree of branching are weighted and corrected according to the distance between the tumor image and the meandering degree and the branching degree calculation point, The tendency to increase as the degree of branching is closer to a malignant tumor is added to the determination support or determination of malignancy or benignity.

  FIG. 28 is a functional block diagram of the tumor diagnosis support system for calculating the meandering degree according to this modification. In the tumor diagnosis support system according to this modification, a group division unit 90 is provided between the processing of the segment separation unit 40 and the meandering degree calculation unit 50, and blood vessel segments that are in the same region according to the distance from the tumor image And blood vessel image parts are divided into the same group. The group dividing unit 90 includes the calculation control unit 5. Also in the tumor diagnosis support system for calculating the degree of branching according to this modification, the group dividing unit 90 is provided between the processing of the blood vessel image extracting unit 30 and the processing of the degree of branching calculating unit 80.

  Then, the meandering degree calculation unit 50 calculates a value indicating the degree of meandering of the group by taking the average, total, or maximum of the group. Further, the calculated value is weighted according to the representative distance between the group and the tumor image to obtain the meandering degree. In addition, the branching degree calculation unit 80 calculates a value indicating the degree of branching for each region, and weights this value according to the representative distance between the region and the tumor image to obtain the branching degree.

  FIG. 29 is a schematic diagram showing a shell space division according to a modified example of the meandering degree and the branching degree. As shown in FIG. 29, in calculating the meandering degree and the branching degree, the medical image space is divided into a plurality of shell spaces 400a, 400b, 400c... Concentrically around the tumor image. The shell spaces 400a, 400b, and 400c are divided into a plurality of stages from the innermost shell space 400a to the outermost shell space 400c, and each has a predetermined thickness.

  Blood vessel segments that fit in the same shell are grouped as one group within the same distance from the tumor image. The blood vessel image is divided by a plurality of shells, and the portion of the blood vessel image that fits in the same shell is divided as a region at the same distance from the tumor image.

  The division is performed by storing, for each shell space, shell space information indicating an area occupied by the shell spaces 400a, 400b, 400c. The shell space information is constituted by the distance between the tumor image and the boundary of the shell space on the tumor image side, and the distance between the tumor image and the boundary of the shell space on the opposite side of the tumor image.

  FIG. 30 is a table for storing weight values. The meandering degree calculation unit 50 and the branching degree calculation unit 80 store a table for storing the weight values in advance. This table stores identification information for identifying one of the plurality of shell spaces 400a, 400b, 400c. The weighting value is a value proportional to the distance from the tumor image of the shell spaces 400a, 400b, 400c. Since the closer the distance from the tumor image is, the more affected the malignant tumor is, the group at a close distance is regarded as important. Among the predetermined thicknesses of the shell spaces 400a, 400b, 400c,..., For example, the center position is a distance from the tumor image in the shell space.

  The meandering degree calculation unit 50 reads out the shell space information, and for the blood vessel segments existing in the shell spaces 400a, 400b, 400c... Indicated by the shell space information, the dispersion of the separation length between the blood vessel segment and the smooth curve, The ratio of the length of the blood vessel segment to the distance from the start to the end of the blood vessel segment, the average or the total or maximum of the separation length of the blood vessel segment and the smooth curve, the distance for each distance from the start to the end of the blood vessel segment and the smooth curve The peak frequency or centroid frequency of the frequency band obtained by converting the separation length curve indicating the length into the spatial frequency, the integrated value of the component size over the entire frequency band obtained by the same conversion, and the integrated value of the predetermined frequency band Ratio, the upper or lower frequency value of the frequency band having a predetermined ratio of the integral value of the magnitude of the component of the entire frequency range obtained by the same conversion Ratio of the total distance between the total and the beginning and end of the length of each vessel segment, the average of the distance between the beginning and end of each vessel segment, or total or calculates the average of the length of each vessel segment.

  The ratio of the total length of each blood vessel segment and the total distance between the start end and the end, the average of the distance between the start end and the end of each blood vessel segment, or the total or average of the length of each blood vessel segment, This is performed by a method of calculating the length of the blood vessel segment and the distance between the start end and the end. That is, the ratio of the sum of the length of each blood vessel segment and the sum of the distance between the start end and the end is calculated by calculating the length of each blood vessel segment and calculating the distance between the start end and the end of each blood vessel segment. Sum and take the ratio of the total length and the total distance. The average distance between the start end and the end of each blood vessel segment is calculated by calculating the distance between the start end and the end of each blood vessel segment, and taking the average of the calculated distances. For the total or average of the lengths of the respective blood vessel segments, the lengths of the respective blood vessel segments are calculated and totaled, or the calculated lengths are totaled.

  Similarly, the branching degree calculation unit 80 reads out the shell space information, and for the portion of the blood vessel image existing in the shell spaces 400a, 400b, 400c... Indicated by the shell space information, The average branch angle of each branch point is acquired for each region.

  This calculation is performed on the blood vessel segments and blood vessel image portions existing in the shell spaces 400a, 400b, 400c,.

  And the table which memorize | stores a weighting value is read, and the value which calculated the weighting value paired with the identification information corresponding to either of shell space 400a, 400b, 400c ... which shell space information shows is multiplied. Normalization is performed, and the normalized value is defined as a meandering degree or a branching degree.

  FIG. 31 is a flowchart showing a calculation operation according to this modification example of the meandering degree and the branching degree. In this description, the technique for calculating the meandering degree based on the ratio of the total length of each blood vessel segment and the total distance between the start end and the end will be described. The calculation of the basic numerical value is the same as in each of the above-described embodiments.

  First, the group division unit 90 reads the center point of the tumor image from the tumor image center point storage unit 25, and converts the medical image space into a plurality of shell spaces 400a, 400b, 400c... According to the distance from the tumor image. Divide (S171). When divided into shell spaces 400a, 400b, 400c,..., Blood vessel segments that belong to the same shell space and are within the same distance range from the tumor image are associated with the same group (S172).

  When grouping is performed, the meandering degree calculation unit 50 calculates the length and the distance from the start end to the end for each blood vessel segment associated with the same group (S173). Next, the values of the lengths are summed, the values of the distances are summed, and the ratio of the sum of the lengths and the sum of the distances is calculated (S174).

  When the ratio is calculated, the meandering degree calculation unit 50 reads a table storing weight values (S175), and obtains the weight values corresponding to the group for which the ratio is calculated from the table (S176).

  When the weighting value is acquired, a value obtained by multiplying the calculated ratio by the weighting value as the meandering degree and normalizing the calculated ratio according to the distance from the tumor image is obtained (S177). The acquired meandering degree is stored in the meandering degree storage unit 58.

  When the meandering degree is calculated for all the groups (S178, Yes), the meandering degree calculation process ends, and if there is a group whose meandering degree has not been calculated yet (S178, No), the meandering degree has not been calculated yet. S173 to S177 are repeated for the group.

  As described above, the blood vessel image existing around the tumor image is extracted from the medical image, and the meandering degree of the blood vessel image is calculated. In addition, the branching degree of each branch point existing in the blood vessel image is calculated. Based on the knowledge that blood vessels around malignant tumors are complicated to bend and branch, the degree of meandering and branching of blood vessels can be used as a malignant index of tumor. Can be reduced.

  In addition, the meandering degree and branching degree are calculated by weighting the degree of blood vessel bending and branching according to the distance from the tumor image. As a result, based on the knowledge that the closer the tumor is to the influence of the malignant tumor on the blood vessel, the meandering degree and branching degree of the blood vessel can be used as a more accurate malignant index of the tumor. The false positive rate due to can be further reduced.

It is a block diagram which shows the structure of the tumor diagnosis assistance system which implements tumor diagnosis assistance. It is a schematic diagram which shows the benign tumor and the blood vessel image which exists in the periphery. It is a schematic diagram which shows the malignant tumor and the blood vessel image which exists in the periphery. It is a functional block diagram which shows the tumor diagnosis assistance system which calculates a meandering degree. It is a schematic flowchart which shows the tumor diagnosis support operation implemented by the tumor diagnosis support system which calculates a meandering degree. It is a schematic diagram which shows tumor image detection and blood vessel image extraction. It is a schematic diagram of the meandering degree calculation by the 1st Example. It is a flowchart which shows the meander degree calculation operation | movement by a 1st Example. It is a flowchart which shows the meander degree calculation operation | movement which concerns on a 2nd Example. It is a flowchart which shows the meander degree calculation operation | movement which concerns on a 3rd Example. It is a flowchart which shows the meander degree calculation operation | movement which concerns on a 4th Example. It is a schematic diagram of the meander degree calculation by the 5th Example. It is a flowchart which shows the meander degree calculation operation | movement which concerns on a 5th Example. It is a flowchart which shows the meander degree calculation operation | movement which concerns on a 6th Example. It is a figure which shows a separation length curve. It is the spectrum figure obtained by frequency-resolving a separation length curve. It is a flowchart which shows the meander degree calculation operation | movement which concerns on a 7th Example. It is a flowchart which shows the meander degree calculation operation | movement which concerns on an 8th Example. It is a schematic diagram of the spatial frequency spectrum which shows a total integral value and a predetermined zone | band integral value. It is a flowchart which shows the meander degree calculation operation | movement which concerns on a 9th Example. It is a schematic diagram of the spatial frequency spectrum which shows a total integral value, a survey integration value, and a survey frequency. It is a functional block diagram which shows the tumor diagnosis assistance system which calculates a branch degree. It is a schematic flowchart which shows the tumor diagnosis assistance operation implemented by the tumor diagnosis assistance system which concerns on branching degree calculation. It is a schematic diagram of the branching degree calculation by the 1st Example. It is a flowchart which shows the branch degree calculation operation | movement by a 1st Example. It is a schematic diagram of branching degree calculation by the 2nd Example. It is a flowchart which shows the branch degree calculation operation | movement by a 2nd Example. It is a functional block diagram of a tumor diagnosis support system according to a modified example of meandering degree and branching degree. It is a schematic diagram which shows the shell space division | segmentation which concerns on the calculation modification of a meandering degree and a branching degree. It is a table which memorize | stores a weighting value. It is a flowchart which shows the calculation operation | movement which concerns on the calculation modification of a meandering degree and a branching degree.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Workstation 2 Input device 3 Display device 4 Printing device 5 Arithmetic control unit 6 Main storage unit 7 External storage unit 10 Medical image storage unit 20 Tumor image detection unit 25 Tumor image center point storage unit 30 Blood vessel image extraction unit 35 Center line storage Unit 40 segment separation unit 45 segment storage unit 50 meandering degree calculation unit 54 smooth curve storage unit 58 meandering degree storage unit 60 determination unit 65 threshold value storage unit 70 output unit 80 branching degree calculation unit 85 branching degree storage unit 100a benign tumor 100b Tumor 200a, 200b Blood vessel image 210 Center line 220 Smooth curve 230 Separation length 240 Length 250 Distance 260 Separation length curve 270 Branch point 280 Investigation point 290 End point close to heart 300 Branch angle 310 Branch point Pv Peak value Pf Peak frequency Lf Upper limit frequency Hf Lower limit frequency Rf Investigation frequency Maxf Maximum frequency fr Predetermined Frequency band IVa Overall integration value IVp Predetermined band integration value IVr Survey integration value

Claims (30)

Detection means for detecting a tumor image from a medical image;
An extraction means for extracting a blood vessel image present around the detected tumor image;
Separating means for separating the extracted blood vessel image into blood vessel segments by separating each blood vessel branch point,
Calculating means for calculating the meandering degree of each blood vessel segment as a malignant tumor index;
Output means for outputting the meandering degree so as to be visible;
Providing
A tumor diagnosis support system characterized by the above. Detection means for detecting a tumor image from a medical image;
An extraction means for extracting a blood vessel image from around the detected tumor image;
A calculating means for calculating a branching degree of each branching point present in the blood vessel image as a malignant tumor index;
An output means for outputting the branching degree so as to be visible;
Providing
A tumor diagnosis support system characterized by the above. Storage means for storing a threshold;
A determination means for determining malignancy or benign of the tumor image based on the meandering degree and the threshold;
Further comprising
The output means outputs either or both of the meandering degree and the determination result so as to be visible,
The tumor diagnosis support system according to claim 1, wherein: Storage means for storing a threshold;
A determination means for determining malignancy or benign of the tumor image based on the branching degree and the threshold;
Further comprising
The output means outputs either or both of the degree of branching and the result of the determination so as to be visible;
The tumor diagnosis support system according to claim 2, wherein: The calculating means includes
A curve calculating means for calculating a smooth curve obtained by smoothing the blood vessel extending shape of the blood vessel segment;
As the meandering degree, a dispersion calculating means for calculating a dispersion of a separation length between the blood vessel segment and the smooth curve;
Including,
The tumor diagnosis support system according to claim 1. The calculating means includes
Segment length calculating means for calculating the length of the blood vessel segment;
Distance calculating means for calculating the distance between the start and end of the blood vessel segment;
A ratio calculating means for calculating a ratio between the length and the distance as the meandering degree;
Including,
The tumor diagnosis support system according to claim 1. The calculating means includes
A curve calculating means for calculating a smooth curve obtained by smoothing the blood vessel extending shape of the blood vessel segment;
As the meandering degree, a separation length calculation means for calculating an average or total or maximum of separation lengths of the blood vessel segment and the smooth curve;
Including,
The tumor diagnosis support system according to claim 1. The calculating means includes
A curve calculating means for calculating a smooth curve obtained by smoothing the blood vessel extending shape of the blood vessel segment;
A frequency conversion means for converting a separation length curve indicating a separation length for each distance from the start end to the end of the blood vessel segment and the smooth curve into a spatial frequency domain;
As the meandering degree, a frequency extraction means for extracting a peak frequency or a centroid frequency from the frequency distribution obtained by the conversion;
Including,
The tumor diagnosis support system according to claim 1. The calculating means includes
A curve calculating means for calculating a smooth curve obtained by smoothing the blood vessel extending shape of the blood vessel segment;
A frequency conversion means for converting a separation length curve indicating a separation length for each distance from the start end to the end of the blood vessel segment and the smooth curve into a spatial frequency domain;
As the meandering degree, from the frequency distribution obtained by the conversion, a frequency calculating means for calculating a ratio between an integrated value of the component size over the entire frequency band and an integrated value of the component size over the predetermined frequency band;
Including,
The tumor diagnosis support system according to claim 1. The calculating means includes
A curve calculating means for calculating a smooth curve obtained by smoothing the blood vessel extending shape of the blood vessel segment;
A frequency conversion means for converting a separation length curve indicating a separation length for each distance from the start end to the end of the blood vessel segment and the smooth curve into a spatial frequency domain;
As the meandering degree, from the frequency distribution obtained by the conversion, a frequency calculating means for calculating an upper limit or a lower limit frequency value of a frequency band having a predetermined ratio of an integral value of the component size over the entire frequency band;
Including,
The tumor diagnosis support system according to claim 1. The calculating means includes
A counting means for counting the number of branches of the blood vessel image as the degree of branching;
The tumor diagnosis support system according to claim 2. The calculating means includes
A branch angle calculating means for calculating an average branch angle of each branch point as the branch degree;
The tumor diagnosis support system according to claim 2. The separating means includes
Grouping means for grouping the blood vessel segments for each within the same distance range from the tumor image,
The calculating means includes
A curve calculating means for calculating a smooth curve obtained by smoothing the blood vessel extending shape of the blood vessel segment;
Dispersion calculating means for calculating a dispersion of a separation length between the blood vessel segment and the smooth curve;
A group value calculating means for calculating an average or sum of the variances or a maximum value for each group;
Normalization value calculating means for calculating a normalization value obtained by weighting the value according to a representative distance between the group and the tumor image as the meandering degree;
Including,
The tumor diagnosis support system according to claim 1. The separating means includes
Grouping means for grouping the blood vessel segments for each within the same distance range from the tumor image,
The calculating means includes
A curve calculating means for calculating a smooth curve obtained by smoothing the blood vessel extending shape of the blood vessel segment;
A group value calculating means for calculating an average or total or maximum value of the separation length between the blood vessel segment and the smooth curve for each group;
Normalization value calculating means for calculating a normalization value obtained by weighting the value according to a representative distance between the group and the tumor image as the meandering degree;
Including,
The tumor diagnosis support system according to claim 1. The separating means includes
Grouping means for grouping the blood vessel segments for each within the same distance range from the tumor image,
The calculating means includes
A curve calculating means for calculating a smooth curve obtained by smoothing the blood vessel extending shape of the blood vessel segment;
A frequency conversion means for converting a separation length curve indicating a separation length for each distance from the start end to the end of the blood vessel segment and the smooth curve into a spatial frequency domain;
A group value calculating means for calculating the average or maximum value of the peak frequency or centroid frequency for each group from the frequency distribution obtained by the conversion;
Normalization value calculating means for calculating a normalization value obtained by weighting the value according to a representative distance between the group and the tumor image as the meandering degree;
Including,
The tumor diagnosis support system according to claim 1. The separating means includes
Grouping means for grouping the blood vessel segments for each within the same distance range from the tumor image,
The calculating means includes
A curve calculating means for calculating a smooth curve obtained by smoothing the blood vessel extending shape of the blood vessel segment;
A frequency conversion means for converting a separation length curve indicating a separation length for each distance from the start end to the end of the blood vessel segment and the smooth curve into a spatial frequency domain;
A group for calculating the average or maximum value of the ratio of the integrated value of the component size over the entire frequency band and the integrated value of the component size over the predetermined frequency band for each group from the frequency distribution obtained by the conversion A value calculating means;
Normalization value calculating means for calculating a normalization value obtained by weighting the value according to a representative distance between the group and the tumor image as the meandering degree;
Including,
The tumor diagnosis support system according to claim 1. The separating means includes
Grouping means for grouping the blood vessel segments for each within the same distance range from the tumor image,
The calculating means includes
A curve calculating means for calculating a smooth curve obtained by smoothing the blood vessel extending shape of the blood vessel segment;
A frequency conversion means for converting a separation length curve indicating a separation length for each distance from the start end to the end of the blood vessel segment and the smooth curve into a spatial frequency domain;
Group value calculation that calculates the average or maximum value of the upper or lower frequency value of the frequency band having a predetermined ratio of the integral value of the component size over the entire frequency band from the frequency distribution obtained by the conversion for each group Means,
Normalization value calculating means for calculating a normalization value obtained by weighting the value according to a representative distance between the group and the tumor image as the meandering degree;
Including,
The tumor diagnosis support system according to claim 1. The separating means includes
Grouping means for grouping the blood vessel segments for each within the same distance range from the tumor image,
The calculating means includes
A total length calculating means for calculating the total length of the blood vessel segments for each group;
An overall distance calculating means for calculating a total distance between the start and end of the blood vessel segment for each group;
A group value calculating means for calculating a ratio of the total length and the total distance of the same group;
Normalization value calculating means for calculating a normalization value weighted by the ratio according to a representative distance between the group and the tumor image as the meandering degree;
Including,
The tumor diagnosis support system according to claim 1. The separating means includes
Grouping means for grouping the blood vessel segments for each within the same distance range from the tumor image,
The calculating means includes
A group value calculating means for calculating an average value of the distance between the start and end of the blood vessel segment for each group;
Normalization value calculating means for calculating a normalization value obtained by weighting the average value according to a representative distance between the group and the tumor image as the meandering degree;
Including,
The tumor diagnosis support system according to claim 1. The separating means includes
Grouping means for grouping the blood vessel segments for each within the same distance range from the tumor image,
The calculating means includes
A group value calculating means for calculating a total or average value of the lengths of the blood vessel segments for each group;
Normalization value calculating means for calculating a normalization value obtained by weighting the value according to a representative distance between the group and the tumor image as the meandering degree;
Including,
The tumor diagnosis support system according to claim 1. Further comprising a dividing means for dividing the space of the medical image into a plurality of regions according to the distance from the tumor image,
The calculating means includes
Counting means for counting the number of branches of the blood vessel image for each region;
Normalization value calculating means for calculating a normalization value obtained by weighting the number of branches according to a representative distance between the region and the tumor image as the degree of branching;
Including,
The tumor diagnosis support system according to claim 2. Further comprising a dividing means for dividing the space of the medical image into a plurality of regions according to the distance from the tumor image,
The calculating means includes
Branch angle calculation means for calculating an average value of branch angles of the branch points for each region. As the branch degree, a normalized value obtained by weighting the average value according to a representative distance between the region and the tumor image is calculated. Normalized value calculating means;
Including,
The tumor diagnosis support system according to claim 2. Detect tumor image from medical image,
Extract blood vessels around the detected tumor image,
The extracted blood vessel image is separated into blood vessel segments for each blood vessel branch point,
Calculating the meandering degree of the blood vessel segment as a malignant tumor index,
A tumor diagnosis support method characterized by the above. Detect tumor image from medical image,
Extract blood vessels around the detected tumor image,
Calculating the branching degree of each branch point present in the blood vessel image as a malignant tumor index,
A tumor diagnosis support method characterized by the above. Determining malignancy or benign tumor based on the meandering degree and a predetermined threshold;
The tumor diagnosis support method according to claim 23, wherein: Determining malignancy or benign tumor based on the degree of branching and a predetermined threshold;
The tumor diagnosis support method according to claim 24, wherein: Computer
Detection means for detecting a tumor image from a medical image;
An extraction means for extracting a blood vessel image present around the detected tumor image;
Separating means for separating the extracted blood vessel image into blood vessel segments by separating each blood vessel branch point,
Calculating means for calculating the meandering degree of each blood vessel segment as a malignant tumor index;
Output means for outputting the meandering degree so as to be visible;
Make it work,
A tumor diagnosis support program characterized by Computer
Detection means for detecting a tumor image from a medical image;
An extraction means for extracting a blood vessel image from around the detected tumor image;
A calculating means for calculating a branching degree of each branching point present in the blood vessel image as a malignant tumor index;
An output means for outputting the branching degree so as to be visible;
To function as a
A tumor diagnosis support program characterized by Storage means for storing a threshold;
Determination means for determining malignancy or benign tumor based on the meandering degree and the threshold;
To make it work,
The output means outputs either or both of the meandering degree and the determination result so as to be visible,
28. The tumor diagnosis support program according to claim 27, wherein: Storage means for storing a threshold;
A determination means for evaluating malignancy or benign tumor based on the degree of branching and the threshold;
To make it work,
The output means outputs either or both of the degree of branching and the result of the determination so as to be visible;
30. The tumor diagnosis support program according to claim 28.