JP2015221141A - Medical image diagnosis support apparatus, operation method for medical image diagnosis support apparatus, and medical image diagnosis support program - Google Patents

Abstract

[Problem] A medical image diagnosis support device capable of outputting appropriate information for determining whether a lesion or lesion candidate included in a group of tomographic images is a malignant lesion during follow-up observation, an operating method thereof, and Provide programs. SOLUTION: An image acquisition unit 41 acquires a group of tomographic images 21N taken this time and a group 21P of tomographic images taken last time. The coordinate acquisition unit 42 acquires the coordinates of the three-dimensional center of the lesion 22 included in the tomographic image groups 21N and 21P. The pixel value distribution acquisition unit 43 acquires a plurality of pixel value distributions of target pixels 60A of the tomographic image groups 21N and 21P on a spherical surface with a radius R from the center coordinates, using the radius R as a variable. The pixel value distribution image generation unit 44 generates pixel value distribution images 63AN and 63AP in which pixel value distributions are arranged with respect to radius R. The input/output control unit 40 outputs pixel value distribution images 63AN and 63AP. [Selection diagram] Figure 5

Description

  The present invention relates to a medical image diagnosis support apparatus, a method for operating a medical image diagnosis support apparatus, and a medical image diagnosis support program.

  Conventionally, in the medical field, image diagnosis based on a tomographic image group composed of a plurality of tomographic images obtained by photographing a subject (patient) with a modality such as a CT (Computed tomography) apparatus or an MRI (Magnetic Resonance Imaging) apparatus has been performed. ing. In such image diagnosis, images are taken over a certain period of time, and a plurality of tomographic image groups obtained at different imaging times are comparatively interpreted, for example, a lesion included in the tomographic image group is a malignant lesion such as cancer. Follow-up observation is performed to determine whether or not. If it is determined that the lesion is suspected of being a malignant lesion, appropriate measures such as detailed examination and treatment are taken.

  Patent Document 1 describes a medical image diagnosis support apparatus that calculates the degree of growth of a lesion and displays the calculated degree of growth in order to assist in determining whether or not the lesion in follow-up observation is a malignant lesion. . This medical image diagnosis support apparatus first took a previous image with a plurality of tomographic image groups having different imaging timings, for example, the tomographic image group acquired this time, based on the position of the structure of the living organ such as the slice position information of the bronchial bifurcation. Three-dimensional alignment of tomographic image groups is performed. Next, a lesion is extracted from the current tomographic image group, and the extracted lesion and a lesion in the previous tomographic image group existing at the same position as the extracted lesion are analyzed, and a feature amount of each lesion is calculated. Then, the degree of growth of the lesion is calculated based on the calculated feature amount. Examples of feature values include the roughness, which is the ratio of the run length of the high pixel value region to the low pixel value region in the lesion, and the previous tomographic image among the pixels at the center of the lesion in the current tomographic image group. The number changed from the pixel value of the pixel of the lesion in the group, the major axis, minor axis, volume, etc. of the lesion are mentioned.

JP 2008-173213 A

  In Patent Document 1, after the three-dimensional alignment of the current and previous tomographic image groups, the three-dimensional alignment is extracted from the current tomographic image group on the assumption that the three-dimensional alignment is performed accurately. The lesion and the lesion in the previous tomographic image group existing at the same position as the extracted lesion are analyzed, and the feature amount of each lesion is calculated. For this reason, if the three-dimensional alignment is not accurately performed, the reliability of the growth amount calculated based on the feature amount, and hence the feature amount, is lowered. In addition, the reliability of the growth degree is lowered even when the lesion is not accurately extracted from the current tomographic image group.

  Here, lesions included in a plurality of groups of tomographic images with different imaging times are subject to changes in subject position, subject orientation, and subject breathing even if the lesions have no temporal change. The position of the lesion and the direction of the lesion differ depending on changes in the movement of surrounding organs such as the heartbeat of the subject. The position of the lesion in the subject is determined to a certain extent if, for example, a single reference point such as the center and the center of gravity of the lesion is determined and alignment is performed between a plurality of tomographic image groups having different imaging timings based on the reference point. Although it is possible to suppress the effects of a decrease in alignment accuracy, it is difficult to perfectly match the direction of the lesion in consideration of the resolution of the tomographic image (slice thickness, number of pixels, etc.) and some deflection of the lesion itself. . Here, the direction of the lesion means that in a two-dimensional tomographic image, in a two-dimensional tomographic image group, a two-dimensional rotational direction having an axis that passes through the center of the lesion and is perpendicular to the two-dimensional plane as a rotational axis. This is a three-dimensional rotation direction with the center of the lesion as the rotation center. In particular, a lesion whose shape and internal structure are not uniform has a greater effect on the reliability of the growth degree due to the failure to match the orientation. In addition, the lesion does not always have a clear boundary that can be distinguished from the adjacent normal region, and the difficulty in accurately extracting the lesion in this case also affects the accuracy of the orientation.

  As described above, the reliability of the degree of growth described in Patent Document 1 depends on the alignment accuracy and the lesion extraction accuracy. Therefore, the degree of growth described in Patent Document 1 is inappropriate as information for determining whether or not a lesion included in a tomographic image group is a malignant lesion in follow-up observation.

  Furthermore, there are various sizes of lesions, and some are as small as about 1 cm. Since these relatively small lesions may be an early stage of early cancer and are one of the important follow-up targets, they can correctly recognize subtle changes over time in the size and density of the lesions (pixel values). Thus, it is necessary to determine whether or not the lesion is a malignant lesion. However, since a relatively small lesion has a smaller amount of image information than a relatively large lesion, it is more difficult to accurately perform positioning and lesion extraction. For this reason, for relatively small lesions, the reliability of the degree of growth assuming accurate alignment is lower, and it is not possible to judge whether the relatively small lesion is a malignant lesion based on the degree of growth. There is a risk. In addition, a qualitative temporal change cannot be correctly recognized because a subtle change in the size and concentration of the lesion cannot be visually recognized only by a quantitative value such as the degree of growth.

  The present invention has been made in view of the above problems, and has little influence on reliability due to alignment accuracy and lesion extraction accuracy, and a lesion or lesion candidate included in a tomographic image group in a follow-up observation is a malignant lesion. It is an object of the present invention to provide a medical image diagnosis support apparatus, an operation method of a medical image diagnosis support apparatus, and a medical image diagnosis support program capable of outputting appropriate information for determining whether or not.

  In order to achieve the above object, the medical image diagnosis support apparatus of the present invention includes an image acquisition unit that acquires a tomographic image group composed of a plurality of tomographic images obtained by photographing a subject, and a tomographic image acquired by the image acquisition unit. A coordinate acquisition unit that acquires coordinates of a three-dimensional center or center of gravity of a lesion or lesion candidate included in the group, and a pixel value distribution of target pixels of a tomographic image group on a spherical surface having a radius R from the coordinates acquired by the coordinate acquisition unit Alternatively, the pixel value distribution acquisition unit that acquires a plurality of pixel value distributions of the target pixel of the tomographic image on the circumference of the radius R from the coordinates acquired by the coordinate acquisition unit and the pixel value distribution acquisition unit. A pixel value distribution image generation unit that generates a pixel value distribution image in which the pixel value distributions are arranged with respect to the radius R, and an output control unit that outputs the pixel value distribution image generated by the pixel value distribution image generation unit.

  The output control unit preferably outputs a plurality of pixel value distribution images generated by the pixel value distribution image generation unit based on a plurality of tomographic image groups having different imaging timings.

  A difference image generation unit that generates a difference image of a plurality of pixel value distribution images generated by the pixel value distribution image generation unit based on a plurality of tomographic image groups having different imaging times is provided, and the output control unit is generated by the difference image generation unit It is preferable to output the difference image.

  The pixel value distribution acquisition unit generates a pixel value distribution based on the histogram generated by the histogram generation unit that generates a histogram that represents the distribution amount of the pixel value in degrees and that represents the range of the pixel value by a class, and the histogram generated by the histogram generation unit. It is preferable to have a pixel value distribution generation unit.

  For example, the pixel value distribution generation unit takes a pixel value distribution amount in the length direction, expresses the size of the pixel value distribution amount as a length, and expresses a pixel value distribution that expresses the range of the pixel value as a difference in color. Generate based on the histogram.

  For example, the pixel value distribution generation unit generates a linear pixel value distribution whose length is proportional to the square of the radius R, and the pixel value distribution image generation unit arranges the linear pixel value distribution with respect to the radius R, A stepwise pixel value distribution image is generated. The pixel value distribution generation unit generates an arc-shaped pixel value distribution whose length is proportional to the square of the radius R, and the pixel value distribution image generation unit arranges the arc-shaped pixel value distribution with respect to the radius R, A fan-shaped pixel value distribution image may be generated. Alternatively, the pixel value distribution generation unit generates a linear pixel value distribution in which the lengths are normalized and aligned, and the pixel value distribution image generation unit arranges the linear pixel value distribution with respect to the radius R, and forms a rectangular shape. A pixel value distribution image may be generated.

  The pixel value distribution generation unit may generate a pixel value distribution in which the pixel value range is expressed in the length direction and the distribution amount of the pixel value is expressed by the difference in color based on the histogram.

  The distribution amount of pixel values is, for example, the number of target pixels for each range of pixel values, the number of target pixels for each range of pixel values, which is obtained by dividing the number of target pixels for each range of pixel values by the number of all target pixels. Ratio, the area of the target pixel for each pixel value range when the minute area obtained by dividing the spherical surface into a plurality of target pixels, or the range of pixel values when the minute area obtained by dividing the spherical surface into a plurality of target pixels Is the ratio of the area of the target pixel obtained by dividing the area of the target pixel by the area of the spherical surface.

  In addition, according to the method for operating the medical image diagnosis support apparatus of the present invention, an image acquisition step of acquiring a tomographic image group composed of a plurality of tomographic images obtained by photographing an object by the image acquisition unit, and an image by the coordinate acquisition unit A coordinate acquisition step for acquiring the coordinates of the three-dimensional center or center of gravity of the lesion or lesion candidate included in the tomographic image group acquired in the acquisition step, and a radius R from the coordinates acquired in the coordinate acquisition step by the pixel value distribution acquisition unit A plurality of pixel value distributions of the target pixels of the tomographic image group on the spherical surface or the pixel value distributions of the target pixels of the tomographic image on the circumference of the radius R from the coordinates acquired in the coordinate acquisition step are acquired using the radius R as a variable. The pixel value distribution acquisition step and the pixel value distribution image generation unit generate a pixel value distribution image in which the pixel value distributions acquired in the pixel value distribution acquisition step are arranged with respect to the radius R. A pixel value distribution image generating step, the output control unit, and an output control step of outputting the pixel value distribution image generated by the pixel value distribution image generating step.

  Furthermore, the medical image diagnosis support program of the present invention includes an image acquisition function for acquiring a tomographic image group composed of a plurality of tomographic images obtained by photographing a subject, and a lesion or a lesion included in the tomographic image group acquired by the image acquisition function. A coordinate acquisition function that acquires the coordinates of the three-dimensional center or center of gravity of a lesion candidate, and a pixel value distribution of a target pixel in a tomographic image group on a spherical surface having a radius R from the coordinates acquired by the coordinate acquisition function, or a coordinate acquisition function A pixel value distribution acquisition function that acquires a plurality of pixel value distributions of target pixels of a tomographic image on the circumference of radius R from the acquired coordinates using radius R as a variable, and a pixel value distribution acquired by the pixel value distribution acquisition function as a radius A computer executes a pixel value distribution image generation function for generating pixel value distribution images arranged with respect to R and an output control function for outputting a pixel value distribution image generated by the pixel value distribution image generation function. To.

  According to the present invention, the pixel value distribution of the target pixel of the tomographic image group on the spherical surface having the radius R from the coordinates of the three-dimensional center or the center of gravity of the lesion or lesion candidate included in the tomographic image group, or included in the tomographic image group. A pixel value distribution image in which pixel value distributions of target pixels of a tomographic image on the circumference of radius R are arranged with respect to radius R from the coordinates of the three-dimensional center or center of gravity of the lesion or lesion candidate to be generated, and the generated pixel value Since the distribution image is output, it is only necessary to align the three-dimensional center or the center of gravity of the lesion with one reference point. Further, the difference in the orientation of the lesion is canceled by acquiring the pixel value distribution of the target pixel of the tomographic image group on the spherical surface with the radius R or the pixel value distribution of the target pixel of the tomographic image on the circumference of the radius R. And there is no need to precisely align the orientation of the lesion. Furthermore, by expressing the pixel value distribution two-dimensionally in the pixel value distribution image, it is possible to correctly recognize a subtle change in the size and density of the lesion. Therefore, the accuracy of the alignment and the extraction accuracy of the lesion have little impact on the reliability, and appropriate information for determining whether the lesion or candidate lesion included in the tomographic image group is a malignant lesion in follow-up observation It is possible to provide a medical image diagnosis support apparatus capable of outputting, an operation method of the medical image diagnosis support apparatus, and a medical image diagnosis support program.

It is a figure which shows a medical information network system. It is explanatory drawing which shows a tomographic image and a tomographic image group. It is a figure which shows the content of the image file. It is a block diagram which shows the computer which comprises a assistance apparatus. It is a block diagram which shows the function of CPU of a assistance apparatus. It is explanatory drawing which shows the mode of the extraction of the lesion by a lesion extraction part, and the calculation of the coordinate of the center or gravity center of a lesion by a coordinate calculation part. It is explanatory drawing which shows a mode that the pixel value distribution of the object pixel on spherical surface SP (K) of radius R (K) is acquired from the center of a lesion. It is explanatory drawing which shows a mode that the pixel value distribution of the object pixel on spherical surface SP (K) of radius R (K) is acquired from the center of a lesion. It is explanatory drawing which shows the object pixel on spherical surface SP (K) of radius R (K) from the center of a lesion. It is a top view of FIG. 9A. It is explanatory drawing which shows the object pixel on spherical surface SP (K) of radius R (K) from the center of a lesion. FIG. 10B is a plan view of FIG. 10A. It is a figure which shows a histogram. It is explanatory drawing which shows a mode that pixel value distribution is produced | generated based on a histogram. It is a figure which shows a pixel value distribution image. It is a figure which shows the comparative interpretation window which displays the tomographic image image | photographed this time and the last time. It is a figure which shows the comparative interpretation window which displays a tomographic image and a pixel value distribution image. It is a figure which shows the comparative interpretation window which expanded and displayed the pixel value distribution image. It is explanatory drawing which shows exchange of the various information between a support apparatus, a medical department terminal, and a medical image DB server. It is a flowchart which shows the operation | movement procedure of a assistance apparatus. It is a block diagram which shows the function of CPU of the assistance apparatus which provided the difference image generation part. It is a figure which shows a difference image. It is explanatory drawing which shows a mode that an arc-shaped pixel value distribution is produced | generated from a linear pixel value distribution, and a fan-shaped pixel value distribution image is produced | generated from an arc-shaped pixel value distribution. It is a figure which shows the rectangular pixel value distribution image produced | generated from the linear pixel value distribution which normalized and arranged the length. It is explanatory drawing which shows a mode that the pixel value distribution which took the range of the pixel value in the length direction and expressed the magnitude of the distribution amount of the pixel value by the difference in color is generated based on the histogram. It is a figure which shows the rectangular pixel value distribution image produced | generated from the pixel value distribution of FIG. It is explanatory drawing which shows a mode that spherical surface SP (K) is divided | segmented into a some micro area | region, and a micro area | region is made into the object pixel on a spherical surface.

[First embodiment]
In FIG. 1, a medical information network system 10 is constructed in a medical facility such as a hospital, and includes a CT apparatus 11, a medical image database (hereinafter abbreviated as DB) server 12, a medical department terminal 13, a medical image diagnosis support apparatus (hereinafter referred to as a medical image diagnosis support apparatus). 14) and a network 15 that connects these devices so that they can communicate with each other. The network 15 is, for example, a LAN (Local Area Network) or a mobile communication network laid in a medical facility.

  In addition to the above, the medical information network system 10 includes a hospital information system (HIS) and a radiation information system (RIS). HIS accepts various information related to medical practices at medical facilities, such as electronic medical records, accounting information, examination reservation information, prescription information, etc., from medical departments such as clinical departments and radiology departments, and stores and manages various information To do. The examination reservation information is issued on the electronic medical record of the medical department terminal 13 by a doctor (diagnostic doctor) in the medical department. The examination reservation information includes an imaging order which is an examination reservation for tomographic imaging by the CT apparatus 11 in addition to various examination reservations such as blood tests and endoscopic examinations. The imaging order is for instructing tomographic imaging to the user of the CT apparatus 11 such as a doctor or a radiographer.

  Subject information is stored in the HIS. Subject information is registered in the electronic medical record. The subject information includes items such as subject ID (IDentification) for identifying individual subjects, subject name, sex, date of birth, age, height, weight, and the like (see FIG. 3).

  The RIS receives the imaging order from the medical department terminal 13, and stores and manages the imaging order. The imaging order includes, for example, an order ID for identifying each order, a doctor ID of a diagnostician who issued the imaging order, a subject ID of a tomographic image to be imaged by the imaging order, and tomographic imaging by the imaging order. It has items such as a user ID in charge, an imaging region and / or orientation (see FIG. 3). There are various parts of the human body such as the head, chest, and abdomen in the imaging region. In addition, there are supine, prone, lying down and so on. In addition to these items, the imaging order includes items such as the date and time when the imaging order was received by RIS, the purpose of tomographic imaging such as follow-up observation, and items to be delivered from the diagnostician to the user.

  The RIS transmits the imaging order to the CT apparatus 11. The user confirms the imaging order on the display of the console of the CT apparatus 11, sets the imaging conditions corresponding to the confirmed imaging order in the CT apparatus 11, and performs tomographic imaging. The imaging conditions include items such as a tube voltage applied to an X-ray source that irradiates X-rays toward a subject, a tube current, an X-ray irradiation time, a slice thickness that is an imaging interval of each tomographic image, and the like (see FIG. 3).

  The medical department terminal 13 is operated by a diagnostician. The medical department terminal 13 includes a display 16 and an input device 17. The display 16 displays various operation screens according to operations of the input device 17 such as a mouse and a keyboard. A GUI (Graphical User Interface) is arranged on the operation screen, and the medical department terminal 13 receives an operation instruction input from the input device 17 through the GUI.

  The CT apparatus 11 performs tomographic imaging according to the set imaging conditions, and obtains a tomographic image group 21 including a plurality of tomographic images 20 as shown in FIG. In FIG. 2, the chest is designated as an imaging region, and a plurality of sheets obtained as a result of performing tomographic imaging by sequentially scanning in the Z direction perpendicular to the XY plane represented by the tomographic image 20 with the set slice thickness. A tomographic image 20 is shown. FIG. 2 shows a state in which a lesion 22 exists in the lower left lung field, and the lesion 22 is reflected in three consecutive tomographic images 20. The CT apparatus 11 creates an image file 23 of the obtained tomographic image 20 and transmits the created image file 23 to the medical image DB server 12. The medical image DB server 12 is a so-called PACS (Picture Archiving and Communication System) server, receives the image file 23 transmitted from the CT apparatus 11, and stores and manages the image file 23.

  In FIG. 3, an image file 23 has a file format conforming to, for example, the DICOM (Digital Imaging and Communication in Medicine) standard, and includes a tag storage area 25 and an image storage area 26. The tomographic image 20 is stored in the image storage area 26. In the tag storage area 25, subject information such as subject ID, subject name, order ID, imaging order such as imaging region and / or orientation, tube voltage (unit kV), tube current (unit mA), X-ray irradiation Each item of imaging conditions such as time (unit: msec) and slice thickness (unit: mm) is provided.

  The tag storage area 25 is provided with items of file ID, photographing date and time, and lesion ID. The file ID is a number for identifying each image file 23 and is automatically assigned by the CT apparatus 11 when the image file 23 is created. As described with reference to FIG. 2, in tomographic imaging, a tomographic image group 21 composed of a plurality of tomographic images 20 is obtained for one imaging order, so the file ID is “F216” or the like. The code assigned to each imaging order and the code representing the number of the tomographic images 20 such as “1” are connected with a hyphen. The tomographic images 20 of the image file 23 having the same code assigned for each photographing order constitute a tomographic image group 21 obtained by one photographing order.

  The lesion ID is a number for identifying each lesion 22 and is given when the lesion 22 is extracted from the tomographic image 20 stored in the image storage area 26.

  The medical image DB server 12 searches for the tomographic image group 21 corresponding to the search key in response to an acquisition request for the tomographic image group 21 from an external device using each item of the incidental information as a search key, and searches the tomographic image group searched for. 21 is transmitted to the external device of the acquisition request source. The acquisition request includes, in addition to the search key, identification information (IP (Internet Protocol) address) of the external device that is the acquisition request source. The medical image DB server 12 specifies the transmission destination of the tomographic image group 21 based on this identification information. The medical image DB server 12 can search and transmit a specific tomographic image 20 in response to an acquisition request for the specific tomographic image 20 as well as the tomographic image group 21. Although the tomographic image 20 and the tomographic image group 21 are expressed as being transmitted to an external device, the image file 23 is actually transmitted to the external device.

  The medical image DB server 12 obtains a plurality of tomographic image groups with different imaging timings obtained by performing imaging with a certain period of time on the same subject, for example, the tomographic image group 21N captured this time and the tomographic image captured last time. The group 21P acquisition request 56 (see FIG. 5) is received from the support apparatus 14, the tomographic image groups 21N and 21P are searched, and the searched tomographic image groups 21N and 21P are transmitted to the support apparatus 14. In the following description, the subscript “N” is added to the image related to the current shooting, and the subscript “P” is added to the image related to the previous shooting.

  As a search key included in the acquisition request, any item stored in the tag storage area 25 of the image file 23 such as a subject ID, shooting date / time, order ID, and user ID can be designated. It is also possible to specify a search range, for example, by specifying all tomographic image groups 21 whose shooting date is six months ago, and the tomographic image group 21 having a subject ID of P0500 and a user ID of R0001. It is also possible to perform a so-called AND search or OR search using a plurality of items.

  The medical department terminal 13 transmits an acquisition request for the tomographic image group 21 to the medical image DB server 12 and receives the tomographic image group 21 transmitted from the medical image DB server 12 in response to the acquisition request. Further, the medical department terminal 13 transmits an analysis request 55 (see FIG. 5) of the tomographic image group 21 to the support device 14, and the pixel value distribution image 63A (FIG. 5) transmitted from the support device 14 in response to the analysis request 55. 13). The analysis request 55 for the tomographic image group 21 includes a search key for designating the tomographic image group 21 from which the pixel value distribution image 63 </ b> A is generated, and identification information (IP of the medical department terminal 13 that issued the analysis request 55. (Internet Protocol) address).

  The display 16 displays the tomographic image 20 from the medical image DB server 12, the pixel value distribution image 63A from the support apparatus 14, and the like (see FIGS. 14, 15, and 16). The medical department terminal 13 receives an input of a diagnosis result of the diagnostician regarding the tomographic image 20 displayed on the display 16 via the input device 17. The diagnosis result includes a result of follow-up observation in which the tomographic image group 21N and the tomographic image group 21P are comparatively interpreted and it is determined whether or not the lesion 22 is a malignant lesion such as cancer. If the lesion is determined to be a suspicious malignant lesion, appropriate measures such as detailed examination and treatment are taken.

  The support apparatus 14 uses the tomographic image groups 21N and 21P as appropriate information for determining whether or not the lesion 22 is a malignant lesion such as cancer when the diagnostician performs follow-up on the medical department terminal 13. The generated pixel value distribution images 63AN and 63AP are output to the clinical department terminal 13.

  In FIG. 4, the computer configuring the support apparatus 14 includes a storage device 30, a memory 31, a CPU (Central Processing Unit) 32, and a communication unit 33. These are interconnected via a data bus 34.

  The storage device 30 is a hard disk drive that is built in a computer constituting the support apparatus 14 or connected through a cable or a network. The storage device 30 stores various application programs including a control program such as an operating system and a medical image diagnosis support program (hereinafter simply referred to as a support program) 35. The support program 35 is a program for causing a computer to function as the support device 14.

  The memory 31 is a work memory for the CPU 32 to execute processing. The CPU 32 loads the program stored in the storage device 30 into the memory 31 and executes processing according to the program, thereby comprehensively controlling each part of the computer.

  The communication unit 33 is a network interface that transmits various types of information to and from an external device via the network 15. The communication unit 33 receives the tomographic image groups 21N and 21P analysis request 55 from the medical department terminal 13, the tomographic image groups 21N and 21P according to the acquisition request 56 from the medical image DB server 12, and the medical image DB server. The acquisition request 56 is transmitted to 12 and the pixel value distribution images 63AN and 63AP are transmitted to the medical department terminal 13, respectively.

  In FIG. 5, when the support program 35 is activated, the CPU 32 cooperates with the memory 31 to input / output control unit (corresponding to an output control unit) 40, an image acquisition unit 41, a coordinate acquisition unit 42, and a pixel value distribution acquisition unit. 43 and a pixel value distribution image generation unit 44. The coordinate acquisition unit 42 includes a lesion extraction unit 50 and a coordinate calculation unit 51. The pixel value distribution acquisition unit 43 includes a histogram generation unit 52 and a pixel value distribution generation unit 53.

  The input / output control unit 40 has an input control function for receiving various information received by the communication unit 33 and an output control function for outputting various information to the communication unit 33. Specifically, the input / output control unit 40 is transmitted from the medical image DB server 12 and received by the communication unit 33 in response to the analysis request 55 and the acquisition request 56 of the tomographic image groups 21N and 21P from the medical department terminal 13. The tomographic image groups 21N and 21P are received. Further, the input / output control unit 40 outputs an acquisition request 56 corresponding to the received analysis request 55 to the communication unit 33, and transmits the acquisition request 56 to the medical image DB server 12 via the communication unit 33. Further, the input / output control unit 40 outputs the pixel value distribution images 63AN and 63AP to the communication unit 33 and transmits the pixel value distribution images 63AN and 63AP to the diagnosis request source medical department terminal 13 via the communication unit 33.

  The image acquisition unit 41 has an image acquisition function of acquiring the tomographic image groups 21N and 21P from the medical image DB server 12 received by the input / output control unit 40. The image acquisition unit 41 outputs the acquired tomographic image groups 21N and 21P to the coordinate acquisition unit 42 and the pixel value distribution acquisition unit 43.

  The coordinate acquisition unit 42 has a coordinate acquisition function of acquiring the coordinates of the three-dimensional center of the lesion 22 included in the tomographic image groups 21N and 21P acquired by the image acquisition unit 41. The lesion extraction unit 50 of the coordinate acquisition unit 42 extracts the lesion 22 for all of the plurality of tomographic images 20N and 20P constituting the tomographic image groups 21N and 21P. For example, the lesion extraction unit 50 compares the pixel values (CT values) of the pixels of the tomographic images 20N and 20P with a preset threshold value for identifying the lesion 22 and other regions, and compares the level with a threshold value greater than or equal to the threshold value. A pixel region having a pixel value is extracted as a lesion 22. The lesion extraction unit 50 outputs the extraction result of the lesion 22 to the coordinate calculation unit 51.

  The coordinate calculation unit 51 calculates the three-dimensional center coordinates of the lesion 22 based on the extraction result of the lesion 22 from the lesion extraction unit 50. The coordinate calculation unit 51 recognizes a lesion 22 having the same or similar coordinates calculated based on each of the tomographic images 20N and 20P as the same lesion 22, and determines the lesion ID of each image file 23 of the tomographic images 20N and 20P. Give the item the same number.

  The center of the lesion 22 is literally the geometric center of the pixel region extracted as the lesion 22 in the tomographic images 20N and 20P. For example, the pixel region extracted as the lesion 22 in the tomographic images 20N and 20P is circular. Is the center of the circle.

  When the same lesion 22 is reflected in a plurality of consecutive tomographic images 20, the coordinate calculation unit 51 has the largest size among the same lesions 22 extracted from the plurality of consecutive tomographic images 20. Calculate the coordinates of the center of the thing. The coordinate calculation unit 51 attaches a lesion ID to the image file 23 of the tomographic image 20 in which the coordinates of the center of the lesion 22 are the largest among the plurality of continuous tomographic images 20. The coordinate acquisition unit 42 outputs the coordinates calculated by the coordinate calculation unit 51 to the pixel value distribution acquisition unit 43.

  FIG. 6 schematically shows how the lesion extraction unit 50 extracts the lesion 22 and the coordinate calculation unit 51 calculates the center coordinates of the lesion 22. The level of the pixel value of the pixel 60 of the tomographic image 20 is indicated by hatching, and the lower the hatching density, the higher the pixel value. A region 57 surrounded by a lower solid line is a region of the pixel 60 having a pixel value equal to or greater than the threshold value, and is a region extracted as the lesion 22 by the lesion extraction unit 50. The region 57 has a circular or square planar shape, and a portion with a high pixel value in the region 57 is unevenly distributed on the upper right side. Reference symbol C <b> 1 is the center of the lesion 22. The coordinates of the center C1 are represented by three-dimensional coordinates (X, Y, Z) indicating the position of the pixel 60 on the center C1 in the tomographic image group 21.

  The pixel value distribution acquisition unit 43 is a pixel value of a pixel 60 (hereinafter, referred to as a target pixel 60A) of the tomographic image group 21 on the spherical surface SP having a radius R from the coordinates of the center C1 of the lesion 22 acquired by the coordinate acquisition unit 42. It has a pixel value distribution acquisition function for acquiring a plurality of distributions 62A (see FIG. 12) using the radius R as a variable.

  7 and 8 are diagrams showing a spherical surface SP (K) having a radius R (K) from the coordinates of the center C1, FIG. 7 is a two-dimensional XY, and FIG. 8 is a three-dimensional XYZ including the Z direction of the scanning direction. It is expressed in dimensions. K is a natural number of 0 or more indicating the number of the radius R and the spherical surface SP (K = 0, 1, 2, 3,..., N−2, N−1, N). The spherical surface SP (0) having the radius R (0) is the center C1, and the target pixel 60A in this case is one pixel 60 on the center C1. A difference ΔR between the radius R (K) and the radius R (K + 1) is, for example, a value equal to or smaller than the size of the pixel 60. The radius R (N) is the maximum value of the radius R, and is a size including the lesion 22 and the vicinity of the periphery thereof. The radius R (N) is, for example, the size of several tens to several hundreds of pixels 60. The radius R (N) is set to about 1.5 cm to 2.0 cm in terms of a meter when the lesion 22 having a size that is likely to be an early state of early cancer is about 1 cm.

  As shown by hatching in FIGS. 9A and 10A, in this embodiment, among the pixels 60 of the tomographic image group 21, the center of the pixel 60 is included in the spherical surface SP (K), and the spherical surface of the spherical surface SP (K). The pixel 60 of the tomographic image group 21 at the upper position is set as a target pixel 60A. 9A shows the target pixel 60A of the tomographic image 20A having the center C1, and FIG. 10A shows the target pixel 60A of the tomographic images 20B and 20C that are before and after the tomographic image 20A and intersect the spherical surface SP (K). . 9B and 10B are plan views of FIGS. 9A and 10A, respectively.

  In FIG. 9A, there are a total of 24 target pixels 60A in the tomographic image 20A. In FIG. 10A, there are 16 target pixels 60A in the tomographic images 20B and 20C. That is, there are a total of 56 target pixels 60A in the case of these three tomographic images 20A, 20B, and 20C. The number of target pixels 60 </ b> A is the smallest when the radius R (0) is 1, and increases as the radius R (K) increases.

  The histogram generation unit 52 of the pixel value distribution acquisition unit 43 has pixel values of -1000 to -901, -900 to -801, ..., -100 to -1, 0 to 99, 100 to the target pixel 60A. A pixel value range such as 199 is determined, and the number of target pixels 60A for each pixel value range is counted. In the present embodiment, the number of target pixels 60A is used as the distribution amount of pixel values.

  As shown in FIG. 11, the histogram generation unit 52 generates a histogram 61 that represents the number of target pixels 60A counted for each pixel value range in degrees on the vertical axis and the pixel value range in terms of horizontal axis on the radius. Generated for each R (K). The closer the lesion 22 is to the center C1, the higher the pixel value tends to be. Therefore, the smaller the radius R (K), the higher the histogram 61 has a higher frequency on the right side of the horizontal axis. On the contrary, as the radius R (K) is larger, the histogram 61 has a higher frequency on the left side of the horizontal axis where the pixel value range is lower. The histogram generation unit 52 outputs the generated histogram 61 to the pixel value distribution generation unit 53.

  As shown in FIG. 12, the pixel value distribution generation unit 53 takes the number of target pixels 60A in the length direction based on the histogram 61 generated by the histogram generation unit 52, and expresses the size of the number of target pixels 60A in length. Then, a pixel value distribution 62A that expresses the level of the pixel value range by the difference in color is generated. More specifically, the pixel value distribution generation unit 53 stacks vertical bars indicating the number of target pixels 60A in the pixel value range of the histogram 61 in order from the right side of the horizontal axis where the pixel value range is high. A straight band indicating the number of target pixels 60A in the pixel value range is formed, and a color corresponding to the height of the pixel value range is added to this band to form a pixel value distribution 62A.

  In FIG. 12, the level of the pixel value range is indicated by shades of hatching, and the lower the hatching depth is, the higher the range of pixel values is. For example, the pixel value range is expressed as white, gray in the middle, black in the middle, and white, gray, and black gradations in the pixel value range. Represent. Alternatively, instead of monochrome of white, gray, and black, it may be expressed in multiple colors such as a portion with a high pixel value range expressed in red, a middle portion in green, and a low portion in blue.

  In the pixel value distribution 62A, the smaller the radius R (K), the smaller the number of target pixels 60A, so the length is shorter. The smaller the radius R (K), the closer to the center C1 of the lesion 22, the pixel value range. The proportion of the high part is large. Further, since the area of the spherical surface SP (K) is proportional to the square of the radius R (K) and the number of target pixels 60A is also proportional to the square of the radius R (K), the length of the pixel value distribution 62A is also the radius. It is proportional to the square of R (K). The pixel value distribution acquisition unit 43 outputs the pixel value distribution 62 </ b> A generated by the pixel value distribution generation unit 53 to the pixel value distribution image generation unit 44.

  As illustrated in FIG. 13, the pixel value distribution image generation unit 44 represents the pixel value distribution 62A in two dimensions by arranging the pixel value distributions 62A acquired by the pixel value distribution acquisition unit 43 with respect to the radius R (K). It has a pixel value distribution image generation function for generating the pixel value distribution image 63A. More specifically, the pixel value distribution image generation unit 44 arranges the pixel value distribution 62A having the radius R (1) on the right side of the pixel value distribution 62A having the radius R (0) and aligning the lower ends thereof. By repeating in order up to (N) pixel value distribution 62A, a stepwise pixel value distribution image 63A that rises to the right as shown in the figure is generated.

  In the pixel value distribution image 63A, the smaller the radius R (K), the greater the proportion of the portion with the higher pixel value range, and the greater the radius R (K), the greater the proportion of the portion with the lower pixel value range. Become more. The boundary where the color in the vicinity of the radius R (N) of the pixel value distribution image 63A changes indicates the boundary between the lesion 22 and the normal region, and thus indicates the size of the lesion 22. The pixel value distribution image generation unit 44 outputs the generated pixel value distribution image 63A to the input / output control unit 40.

  The pixel value distribution acquisition unit 43 generates pixel value distributions 62AN and 62AP based on the coordinates of the center C1 of the lesion 22 calculated based on the tomographic images 20N and 20P, respectively. The pixel value distribution image generation unit 44 generates pixel value distribution images 63AN and 63AP corresponding to the tomographic images 20N and 20P based on the pixel value distributions 62AN and 62AP generated by the pixel value distribution acquisition unit 43. The pixel value distribution image generation unit 44 attaches the lesion ID given to the tomographic images 20N and 20P to the pixel value distribution images 63AN and 63AP. Thereby, the tomographic images 20N and 20P and the corresponding pixel value distribution images 63AN and 63AP are associated with each other by the lesion ID.

  When a plurality of lesions 22 are extracted by the lesion extraction unit 50, coordinates are calculated, a pixel value distribution 62A is generated, and a pixel value distribution image 63A is generated for each of the extracted lesions 22. Therefore, when a plurality of lesions 22 are extracted, pixel value distribution images 63A are generated for the number of extracted lesions 22. Whether the pixel value distribution image 63A is generated based on which of the plurality of lesions 22 is identified by the lesion ID.

  In the follow-up observation, the clinical department terminal 13 causes the display 16 to display a comparative interpretation window 70A shown in FIG. The comparative interpretation window 70A is a screen for comparative interpretation of the tomographic images 20N and 20P, and a display frame 71A and a display frame 71A for displaying the tomographic images 20N and 20P transmitted from the medical image DB server 12 side by side in response to the acquisition request. 71B and a display frame 72 for displaying supplementary information common to the tomographic images 20N and 20P such as subject information. Above the display frames 71A and 71B, a display frame 73 for displaying the file ID and a display frame 74 for displaying the shooting date and time are provided.

  An instruction button 75 for instructing display of the pixel value distribution image 63A is arranged in the comparative interpretation window 70A. When the instruction button 75 is selected, the medical department terminal 13 transmits an analysis request 55 for the tomographic image groups 21N and 21P to which the tomographic images 20N and 20P displayed in the display frames 71A and 71B belong to the support apparatus 14. .

  When the instruction button 75 is selected, the display of the comparative interpretation window 70A transitions to the comparative interpretation window 70B shown in FIG. Display frames 76A and 76B for displaying the pixel value distribution images 63AN and 63AP are added to the lower part of the display frames 71A and 71B in the comparative interpretation window 70B. The tomographic images 20N and 20P displayed in the display frames 71A and 71B are associated with the pixel value distribution images 63AN and 63AP displayed in the display frames 76A and 76B by the lesion ID.

  As shown in FIG. 16, the display frames 76A and 76B are enlarged and displayed in the comparative interpretation window 70C independently of the comparative interpretation window 70B so that the pixel value distribution images 63AN and 63AP can be comparatively interpreted in more detail. It is possible to do. By enlarging and displaying the pixel value distribution images 63AN and 63AP, a slight change in density at the boundary between the lesion 22 and the normal region can be recognized without overlooking.

  Hereinafter, the operation of the above configuration will be described with reference to FIGS. 17 and 18. First, an imaging order is issued from the medical department terminal 13 to the RIS by the diagnostician. The imaging order is transmitted from the RIS to the CT apparatus 11 and displayed on the display of the console of the CT apparatus 11. The user confirms the contents of the imaging order on the display, sets the imaging conditions corresponding to the imaging order in the CT apparatus 11, and performs tomographic imaging. A tomographic image group 21 composed of a plurality of tomographic images 20 obtained in this way is transmitted to the medical image DB server 12 as an image file 23 to which a file ID with the same code is allocated, and the medical image DB server 12 Stored and managed.

  The diagnostician operates the clinical department terminal 13 and performs follow-up observation. The diagnostician transmits an acquisition request for the tomographic image groups 21N and 21P using the subject ID and the photographing date and time as search keys to the medical image DB server 12 through the medical department terminal 13. The medical image DB server 12 searches for the tomographic image groups 21N and 21P according to the acquisition request from the medical department terminal 13, and transmits the searched tomographic image groups 21N and 21P to the medical department terminal 13.

  The medical department terminal 13 receives the tomographic image groups 21N and 21P from the medical image DB server 12, and displays the comparative interpretation window 70A on the display 16 based on the received group. The diagnostician comparatively interprets the tomographic images 20N and 20P displayed in the comparative interpretation window 70A, and performs follow-up observation.

  During comparative interpretation, the diagnostician selects the instruction button 75 to display the pixel value distribution images 63AN and 63AP. As shown in FIG. 17A, when the instruction button 75 is selected, an analysis request 55 for the tomographic image groups 21N and 21P to which the tomographic images 20N and 20P displayed in the display frames 71A and 71B belong at that time is displayed. It is transmitted from the medical department terminal 13 to the support device 14.

  In the support device 14, the analysis request 55 from the medical department terminal 13 is received by the communication unit 33. As shown in step S <b> 10 of FIG. 18, the analysis request 55 received by the communication unit 33 is accepted by the input / output control unit 40. As shown in step S11 of FIG. 17B and FIG. 18, an acquisition request 56 corresponding to the analysis request 55 received by the input / output control unit 40 is transmitted to the medical image DB server 12 via the communication unit 33. The The medical image DB server 12 searches for the tomographic image groups 21N and 21P in response to the acquisition request 56 from the support apparatus 14, and as shown in FIG. Send to.

  In the support device 14, the tomographic image groups 21 </ b> N and 21 </ b> P from the medical image DB server 12 are received by the communication unit 33. The tomographic image groups 21N and 21P received by the communication unit 33 are received by the input / output control unit 40. The tomographic image groups 21N and 21P received by the input / output control unit 40 are output to the image acquisition unit 41. Thereby, the tomographic image groups 21N and 21P are acquired by the image acquisition unit 41 (step S12 in FIG. 18). The tomographic image groups 21N and 21P are output to the coordinate acquisition unit 42 and the pixel value distribution acquisition unit 43.

  In the coordinate acquisition unit 42, the lesion 22 included in the tomographic image groups 21N and 21P is extracted by the lesion extraction unit 50 (step S13 in FIG. 18), and the extraction result is output to the coordinate calculation unit 51. The coordinate calculation unit 51 calculates the coordinates of the center C1 of the lesion 22 (step S14 in FIG. 18). Thereby, the coordinates of the center C1 of the lesion 22 are acquired (step S15 in FIG. 18). The coordinates of the center C1 of the lesion 22 are output to the pixel value distribution acquisition unit 43.

  The lesion 22 does not always have a clear boundary that can be distinguished from the adjacent normal region. Therefore, if the lesion 22 is to be extracted more accurately, a complicated process is required and the process takes time. However, the extraction of the lesion 22 by the lesion extraction unit 50 is a pre-process for calculating the coordinates of the center C1 of the lesion 22 by the subsequent coordinate calculation unit 51, and therefore it is not necessary to accurately extract the lesion 22, and the lesion is extracted. Even if a part of the normal region is included in the region 57 extracted as 22, it does not matter. Even if the lesion 22 is extracted more accurately, compared to the case where the lesion 22 is extracted by a method with low extraction accuracy, a large error does not occur in the coordinates of the center C1 of the lesion 22. The extraction of 22 does not require a high degree of accuracy. Therefore, the extraction of the lesion 22 can be completed in a short time with a simple process.

  In the pixel value distribution acquisition unit 43, the histogram generation unit 52 displays a histogram 61 representing the number of target pixels 60 </ b> A counted for each pixel value range in degrees and representing the pixel value range as a class for each radius R (K). (Step S16 in FIG. 18). Based on the histogram 61, the pixel value distribution generation unit 53 generates pixel value distributions 62AN and 62AP (step S17 in FIG. 18). Thereby, pixel value distributions 62AN and 62AP are acquired (step S18 in FIG. 18). The pixel value distributions 62AN and 62AP are output to the pixel value distribution image generation unit 44.

  The pixel value distribution image generation unit 44 generates pixel value distribution images 63AN and 63AP in which the pixel value distributions 62AN and 62AP are arranged with respect to the radius R (K) (step S19 in FIG. 18). The pixel value distribution images 63AN and 63AP are output to the input / output control unit 40. The pixel value distribution images 63AN and 63AP output to the input / output control unit 40 are output to the communication unit 33, and as shown in step S20 of FIG. 17D and FIG. 13 is transmitted.

  In the clinical department terminal 13, the comparative interpretation window 70B is displayed on the display 16, and the pixel value distribution images 63AN and 63AP can be browsed in addition to the tomographic images 20N and 20P. The diagnostician displays the comparative interpretation window 70C as necessary, and enlarges and observes the pixel value distribution images 63AN and 63AP in detail.

  The lesions 22 included in the tomographic image groups 21N and 21P are caused by changes in the position of the subject and the orientation of the subject at the current and previous imaging, changes in the movement of surrounding organs such as the subject's breathing, and the heartbeat of the subject. Although the position of the lesion 22 and the orientation of the lesion 22 are different, the difference in the position of the lesion 22 is only the alignment by one reference point of the coordinates of the three-dimensional center C1 of the lesion 22 of the tomographic images 20N and 20P. Can be solved. The difference in orientation of the lesion 22 can be offset by acquiring the distribution state of the pixel values of the target pixel 60A on the spherical surface SP (K) having a radius R (K) that has no anisotropy with respect to the lesion 22. The orientation of the lesion 22 does not need to be accurately matched. Therefore, it is not necessary to consider the resolution (slice thickness, number of pixels) of the tomographic image 20 and the bending of the lesion 22 itself. Even when the lesion 22 itself is deflected, the difference between the pixel value distribution image 63A as a whole is a slight difference compared to the case where there is no deflection, and therefore the lesion 22 for determining whether or not the lesion 22 is a malignant lesion. The influence of its own deflection can be offset.

  For a lesion 22 having a size of about 1 cm which may be an early state of early cancer, it is more difficult to accurately perform alignment and lesion extraction. It is more effective because accurate alignment and accurate lesion extraction are not necessary. For this reason, it can contribute to the early detection and rapid treatment of cancer.

  The pixel value distribution image 63 </ b> A itself allows the diagnostician to visually recognize the size and density of the lesion 22 that is not known with quantitative numerical values such as the degree of growth of Patent Document 1. Therefore, by outputting the pixel value distribution images 63AN and 63AP to the clinical department terminal 13 so as to be viewable, whether or not the lesion 22 is a malignant lesion, such as an increasing tendency of the size of the lesion 22 or an increasing tendency of the density. Therefore, it is possible to provide a diagnostic doctor with important judgment materials when judging whether or not.

  Conventionally, when it is determined whether or not a lesion 22 is a malignant lesion, it is necessary to comparatively interpret a plurality of tomographic images 21N and 21P, which takes time. However, in the present embodiment, it is possible to determine whether or not the lesion 22 is a malignant lesion simply by comparing and interpreting one pixel value distribution image 63AN and 63AP with respect to one lesion 22. Work can be made more efficient.

  Since the pixel value distribution image 63A shows not the pixel values of the tomographic image 20, but the distribution thereof, the surrounding organs such as the change in the position of the subject at the time of photographing, the orientation of the subject, the breathing of the subject, the heartbeat of the subject, etc. Even if there is a change in the pixel values of the normal region in the vicinity of the lesion 22 and its periphery between the tomographic image group 20N and the tomographic image 20P due to a change in the movement of the tomographic image 20N and the like, Has little impact on the perception of

  Note that the coordinates of the center of gravity of the lesion 22 may be acquired instead of the coordinates of the center of the lesion 22. The center of gravity of the lesion 22 is determined in consideration of the level of the pixel value of the region 57 extracted as the lesion 22. For this reason, the center of gravity of the lesion 22 does not necessarily coincide with the geometric center C1 of the region 57, as indicated by reference numeral C2 in FIG. For example, when a portion with a high pixel value is biased to one side of the region 57, the geometric center is shifted toward a portion with a high pixel value of the lesion 22 to obtain the center of gravity of the lesion 22. The center of gravity C2 shown in FIG. 6 is shifted from the center C1 to the upper right side according to a portion with a high pixel value unevenly distributed on the upper right side.

  How much the geometric center C1 is shifted depends on a difference in pixel values between a portion having a high pixel value in the region 57 and another portion, a ratio occupied by a portion having a high pixel value in the region 57, and the like. To decide.

  Further, instead of setting the pixel 60 of the tomographic image group 21 on the spherical surface SP (K) having the radius R (K) from the coordinates acquired by the coordinate acquiring unit 42 as the target pixel 60A, the radius from the coordinates acquired by the coordinate acquiring unit 42 is used. The pixel 60 of the tomographic image 20 on the circumference of the circle of R (K) may be the target pixel 60A. In this case, the target pixel 60A is, for example, only 24 target pixels 60A of the tomographic image 20A shown in FIG. 9A, and the 16 target pixels 60A of the tomographic images 20B and 20C shown in FIG. 10A are excluded.

[Second Embodiment]
In the first embodiment, the support device 14 generates the pixel value distribution images 63AN and 63AP based on the tomographic image groups 21N and 21P, and outputs the generated pixel value distribution images 63AN and 63AP to the medical department terminal 13. However, a difference image between the pixel value distribution images 63AN and 63AP may be generated, and the generated difference image and the pixel value distribution images 63AN and 63AP may be output to the medical department terminal 13.

  In FIG. 19, the CPU 80 of the support apparatus according to the present embodiment functions as a difference image generation unit 81 in addition to the function units 40 to 44 of the first embodiment. The difference image generation unit 81 receives the pixel value distribution images 63AN and 63AP from the pixel value distribution image generation unit 44. Then, a difference image 82 between the pixel value distribution images 63AN and 63AP shown in FIG. 20 is generated.

  Since the pixel value distribution images 63AN and 63AP express the level of the pixel value range by the difference in color, for example, the difference image generation unit 81 has the same pixel value range in the pixel value distribution images 63AN and 63AP. The portion is black (indicated by dark hatching in FIG. 20), and a difference image 82 colored to change from gray to white (indicated by light hatching in FIG. 20) depending on the number of pixel value ranges is generated. For example, the pixel value distribution image 63AN is 100 to 199, the pixel value distribution image 63AP is 0 to 99, and a portion where the pixel value range is different by one step is gray near black, and the pixel value distribution image 63AN is 100 to 199, In the pixel value distribution image 63AP, the portion where the pixel value range is different in six stages from -500 to -401 is set to gray close to white. The difference image generation unit 81 outputs the generated difference image 82 to the input / output control unit 40.

  The input / output control unit 40 outputs the difference image 82 to the communication unit 33 in addition to the pixel value distribution images 63AN and 63AP, and transmits the difference image 82 to the medical department terminal 13 via the communication unit 33. The medical department terminal 13 displays the difference image 82 in the comparative interpretation windows 70B and 70C in addition to the pixel value distribution images 63AN and 63AP.

  When there is no change in the entire difference image 82, it can be seen that there is no difference between the pixel value distribution images 63AN and 63AP, that is, there is no change over time in the lesion 22. On the contrary, when there is a change in the difference image 82, the lesion 22 It can be seen that the lesion 22 has changed over time. Therefore, by displaying the difference image 82 in addition to the display of the pixel value distribution images 63AN and 63AP, it is possible to correctly recognize the enlargement tendency of the size of the lesion 22 and the increase tendency of the density.

[Third Embodiment]
In each of the above embodiments, a linear pixel value distribution 62A whose length is proportional to the square of the radius R (K) is simply arranged with respect to the radius R (K) to generate a stepwise pixel value distribution image 63A. However, the shape of the pixel value distribution 62 and the pixel value distribution image 63 is not limited to a linear shape and a staircase shape. For example, as shown in FIG. 21, the pixel value distribution generation unit 53 generates an arcuate pixel value distribution 62B from a linear pixel value distribution 62A generated based on the histogram 61, and this arcuate pixel value distribution. 62B may be arranged with respect to the radius R (K) by the pixel value distribution image generation unit 44 to generate a fan-shaped pixel value distribution image 63B. The arc-shaped pixel value distribution 62B is proportional to the square of the radius R (K), like the pixel value distribution 62A of the first embodiment.

  The stepped pixel value distribution image 63A has a significantly different shape from the rectangular tomographic image 20, and the larger the radius R (K), the longer the pixel value distribution 62A and the radius R (K ) Is conspicuous as compared with the pixel value distribution 62A having a short length, so that some habituation is required for observation. On the other hand, the arc-shaped pixel value distribution 62B is similar to the shape of the tomographic image 20 compared to the staircase-shaped pixel value distribution image 63A, and has a long pixel value distribution 62B when the radius R (K) is large. Since it is less conspicuous than the short pixel value distribution 62B, it can be observed with a feeling close to that of the tomographic image 20.

[Fourth Embodiment]
In each of the above embodiments, the pixel value distributions 62A and 62B having different lengths are arranged with respect to the radius R (K) to generate the stepwise pixel value distribution image 63A or the fan-shaped pixel value distribution 62B. Is not limited to this. It is also possible to normalize and align the lengths of the pixel value distributions 62, arrange the pixel value distributions 62 having the same lengths with respect to the radius R (K), and generate a rectangular pixel value distribution image 63.

  In this case, the histogram generation unit 52 uses the number of target pixels 60A for each pixel value range as the distribution amount of pixel values, instead of the number of target pixels 60A for each pixel value range in each of the above embodiments. A ratio of the number of target pixels 60A for each range of pixel values divided by the number of 60A is calculated. The number of all target pixels 60A is the number of target pixels 60A for a certain spherical surface SP (K). In the example of FIGS. 9 and 10, the number of all target pixels 60A is 56.

  For example, when the number of target pixels 60A in the range of pixel values from −1000 to −901 is four and the number of all target pixels 60A is 100, the pixel values in the range of pixel values from −1000 to −901 Is 4/100 = 0.04. When the ratio of the number of target pixels 60A in the range of each pixel value is added, it becomes 1. The histogram generation unit 52 generates a histogram that represents the ratio of the number of target pixels 60A for each pixel value range in terms of frequency and represents the pixel value range in terms of class.

  Similarly to the first embodiment, the pixel value distribution generation unit 53 stacks vertical bars indicating the ratio of the number of target pixels 60A for each pixel value range of the histogram in order from the pixel value range having the highest range, A linear band indicating the magnitude of the ratio of the number of target pixels 60A is created, and a pixel value distribution 62C (see FIG. 22) in which a color corresponding to the height of the pixel value range is added to the band. The pixel value distributions 62A and 62B of the first and third embodiments have a shorter length because the number of target pixels 60A is smaller as the radius R (K) is smaller, but the pixel value distribution 62C of the present embodiment is Since the ratio of the number of target pixels 60A that are 1 when added is an image, the length is the same regardless of the radius R (K). Therefore, the pixel value distribution 62C is obtained by normalizing and aligning the length of the pixel value distribution 62A.

  As shown in FIG. 22, the pixel value distribution image generation unit 44 arranges linear pixel value distributions 62C whose lengths are normalized and aligned with respect to the radius R (K), and generates a rectangular pixel value distribution image 63C. To do. The rectangular pixel value distribution image 63C is not a characteristic shape such as the staircase-like pixel value distribution image 63A and the arc-shaped pixel value distribution 62B, but has the same shape as the tomographic image 20, and thus a tomographic image. It can be observed with a sense close to 20.

[Fifth Embodiment]
In each of the above embodiments, the distribution amount of the pixel values such as the number and ratio of the target pixels 60A is taken in the length direction, the size of the distribution amount of the pixel values is expressed by the length, and the level of the pixel value range is expressed by the difference in color. The pixel value distributions 62A to 62C are generated, but the present invention is not limited to this. A pixel value distribution may be generated in which the range of pixel values is taken in the length direction and the distribution amount of the pixel values is expressed by the difference in color.

  In FIG. 23, a pixel value distribution generation unit 53 takes a pixel value range in the length direction based on the histogram 61 generated by the histogram generation unit 52, and expresses the magnitude of the distribution amount of pixel values by the difference in color. A value distribution 62D is generated. More specifically, the pixel value distribution generation unit 53 creates a plurality of sections divided at equal intervals in the linear band according to the range of each pixel value of the histogram 61, and the distribution amount of the pixel values in this section The pixel value distribution 62D is obtained by adding colors according to the size of.

  In FIG. 23, the magnitude of the distribution amount of the pixel value is shown by hatching, and the distribution amount of the pixel value is larger as the hatching density is lighter. The expression of the distribution amount of the pixel value is, for example, a portion where the distribution amount of the pixel value is large as white, a middle portion as gray, and a small portion (including the case where the distribution amount of pixel value is 0) as black. The magnitude of the value distribution is represented by white, gray, and black gradations. As in the case of the pixel value distribution 62A, it may be expressed in multiple colors.

  As shown in FIG. 24, the pixel value distribution image generation unit 44 arranges the pixel value distributions 62D with respect to the radius R (K), and generates a rectangular pixel value distribution image 63D. Since this rectangular pixel value distribution image 63D has the same shape as the tomographic image 20, like the pixel value distribution image 63C of the fourth embodiment, it can be observed with a sense closer to the tomographic image 20.

  In each of the above embodiments, the tomographic image group 21 is located in the spherical surface SP (K) having the radius R (K) from the coordinates of the center C1 of the lesion 22 and located on the spherical surface of the spherical surface SP (K). Although the pixel 60 is the target pixel 60A, the pixel 60 of the tomographic image group 21 at the position on the spherical surface of the spherical surface SP (K) is the target regardless of whether the center is included in the spherical surface SP (K). The pixel 60A may be used.

  Further, as shown in FIG. 25, the spherical surface SP (K) may be divided into a plurality of minute regions, and these minute regions may be used as the target pixel 60A. Since the target pixel 60A in this case is a virtual pixel created on the spherical surface SP (K), it does not have a pixel value as it is. Therefore, the pixel value of the target pixel 60A is calculated by interpolation from the pixel value of the pixel 60 of the tomographic image 20 in the vicinity of the target pixel 60A. For example, when there are four pixels 60 of the tomographic image 20 near a certain target pixel 60A and the pixel values of the four pixels 60 are 15, 25, 38, and 46, respectively, the pixel value of the target pixel 60A is When the average value of the pixel values of the four pixels 60 is an interpolation value, (15 + 25 + 38 + 46) / 4 = 31.

  In this case, the histogram generation unit 52 integrates the area of the target pixel 60A (area of the minute region) for each pixel value range as the distribution amount of the pixel value, expresses the area of the target pixel 60A in degrees, and the pixel value A histogram representing the range of is generated for each radius R (K). The pixel value distribution generation unit 53 takes the area of the target pixel 60A for each range of pixel values in the length direction, expresses the size of the area of the target pixel 60A by length, and expresses the level of the pixel value range by color difference. The generated pixel value distribution 62 is generated. The pixel value distribution 62 may be linear like the pixel value distribution 62A of the first embodiment, or may be arcuate like the pixel value distribution 62B of the third embodiment. Alternatively, as in the pixel value distribution 62C of the fifth embodiment, the pixel value range may be taken in the length direction, and the size of the area of the target pixel 60A for each pixel value range may be expressed by the difference in color.

  As the distribution amount of the pixel values, instead of the area of the target pixel 60A for each pixel value range, the area of the target pixel 60A obtained by dividing the area of the target pixel 60A for each pixel value range by the area of the spherical surface SP (K). You may employ | adopt the ratio of an area. If the distribution amount of the pixel values is the ratio of the area of the target pixel 60A, the pixel value distribution 62 can be obtained by normalizing and aligning the lengths as in the pixel value distribution 62D of the fourth embodiment.

  In the first embodiment, the lesion 22 having the same or similar coordinates calculated based on the tomographic images 20N and 20P is recognized as the same lesion 22, but the present invention is not limited to this. For example, in the comparative interpretation window 70A, the diagnostic doctor designates the region of the lesion 22 that appears to be identical on the tomographic images 20N and 20P, and the lesion extraction, coordinate calculation, pixel value distribution 62 and The pixel value distribution image 63 may be generated. In this case, the file ID of the image file 23 of the tomographic images 20N and 20P in which the region considered to be the lesion 22 is designated and the coordinate information of the designated region are attached to the analysis request 55. The support apparatus 14 extracts a lesion only in a designated area. Therefore, it is more efficient than extracting the lesion 22 or the like for all of the plurality of tomographic images 20N and 20P constituting the tomographic image groups 21N and 21P.

  The coordinates of the center C1 or the center of gravity C2 of the lesion 22 may be manually specified. For example, in the comparative interpretation window 70A, the diagnostician designates the pixel 60 that seems to be the center C1 or the center of gravity C2 of the lesion 22 on the tomographic images 20N and 20P. In this case, the coordinates of the designated pixel 60 are added to the analysis request 55. The coordinate acquisition unit 42 receives the analysis request 55 from the input / output control unit 40 and acquires the coordinates attached to the analysis request 55. In this case, it is not necessary to extract a lesion and calculate coordinates, so that the processing can be simplified. However, since the center C1 or the center of gravity C2 of the lesion 22 is left to the sense of the diagnostician, it is preferable to extract the lesion and calculate the coordinates as in the above embodiments when more accuracy is important.

  Based on the position of the structure of the living organ such as the slice position information of the bronchial bifurcation, the tomographic image groups 21N and 21P may be three-dimensionally aligned, and the lesion may be extracted. In this case, the lesion extraction and the coordinate calculation may be performed only on the tomographic image group 21N, and the lesion extraction result and the coordinate calculation result on the tomographic image group 21N may be applied to the tomographic image group 21P.

  When the analysis request 55 is transmitted from the medical department terminal 13 to the support device 14, since the medical department terminal 13 has received the tomographic image groups 21 N and 21 P from the medical image DB server 12, the tomographic image is added to the analysis request 55. Groups 21N and 21P may be attached. In this case, it is not necessary to transmit the acquisition request 56 corresponding to the analysis request 55 from the support apparatus 14 to the medical image DB server 12.

  In each of the above embodiments, the support device 14 is provided separately from the medical image DB server 12 and the medical department terminal 13, but the medical image DB server 12 or the medical department terminal 13 may have the function of the support device 14. . Further, the functions of the medical department terminal 13 and the support device 14 may be integrated, and the comparative interpretation windows 70A to 70C may be displayed on the display of the support device 14. In this case, the output control unit has an output control function of the pixel value distribution image 63 to the display.

  The transmission destination of the pixel value distribution image 63 is not limited to the medical department terminal 13 of each of the above embodiments, and may be an interpretation terminal operated by an interpretation doctor who performs interpretation of the tomographic image group 21. Further, instead of acquiring the tomographic image group 21 from the medical image DB server 12, the tomographic image group 21 may be acquired from a storage medium such as a compact disk in which the tomographic image group 21 is stored. Furthermore, the CT apparatus 11 and the medical department terminal 13 may be one or plural. In addition to or instead of the CT apparatus 11, another modality such as an MRI apparatus may be provided.

  In each of the above-described embodiments, an example in which a pixel value distribution image is generated based on two current tomographic image groups and the previous tomographic image group has been described. A pixel value distribution image may be generated based on a group of tomographic images.

  The lesion 22 includes, in addition to a narrowly defined lesion such as a malignant lesion that has been finally diagnosed as having cancer, for example, by the diagnostician (corresponding to the “lesion” in claims 1, 11 and 12). In the initial stage of diagnosis, malignancy is suspected and a candidate for a malignant lesion, but a lesion candidate that is finally diagnosed as a benign lesion (corresponding to a “lesion candidate” in claims 1, 11 and 12) included.

14 Medical image diagnosis support device (support device)
20 tomographic images 21, 21N and 21P tomographic image groups 22 lesions 32, 80 CPU
35 Medical diagnostic imaging support program (support program)
40 Input / output control unit (output control unit)
41 image acquisition unit 42 coordinate acquisition unit 43 pixel value distribution acquisition unit 44 pixel value distribution image generation unit 52 histogram generation unit 53 pixel value distribution generation unit 60A target pixel 61 histogram 62A, 62B, 62C, 62D pixel value distribution 63A, 63AN and 63AP, 63B, 63C, 63D Pixel value distribution image 81 Difference image generation unit 82 Difference image

Claims (12)

An image acquisition unit for acquiring a tomographic image group composed of a plurality of tomographic images obtained by photographing a subject;
A coordinate acquisition unit that acquires coordinates of a three-dimensional center or center of gravity of a lesion or lesion candidate included in the tomographic image group acquired by the image acquisition unit;
The pixel value distribution of the target pixel of the tomographic image group on the spherical surface with the radius R from the coordinates acquired by the coordinate acquisition unit, or the tomographic image on the circumference with the radius R from the coordinates acquired by the coordinate acquisition unit. A pixel value distribution acquisition unit that acquires a plurality of pixel value distributions of the target pixel using the radius R as a variable;
A pixel value distribution image generation unit that generates a pixel value distribution image in which the pixel value distributions acquired by the pixel value distribution acquisition unit are arranged with respect to the radius R;
A medical image diagnosis support apparatus comprising: an output control unit that outputs the pixel value distribution image generated by the pixel value distribution image generation unit.   The medical image diagnosis support apparatus according to claim 1, wherein the output control unit outputs the plurality of pixel value distribution images generated by the pixel value distribution image generation unit based on the plurality of tomographic image groups having different imaging timings. A difference image generation unit that generates a difference image of the plurality of pixel value distribution images generated by the pixel value distribution image generation unit based on the plurality of tomographic image groups having different imaging timings;
The medical image diagnosis support apparatus according to claim 1, wherein the output control unit outputs the difference image generated by the difference image generation unit. The pixel value distribution acquisition unit generates a histogram that represents a distribution amount of the pixel value in degrees and represents a range of the pixel value in a class;
The medical image diagnosis support apparatus according to claim 1, further comprising: a pixel value distribution generation unit that generates the pixel value distribution based on the histogram generated by the histogram generation unit.   The pixel value distribution generation unit takes the distribution amount of the pixel value in the length direction, expresses the size of the distribution amount of the pixel value by length, and expresses the height of the range of the pixel value by a difference in color. The medical image diagnosis support apparatus according to claim 4, wherein a value distribution is generated based on the histogram. The pixel value distribution generation unit generates the linear pixel value distribution whose length is proportional to the square of the radius R,
The medical image diagnosis support apparatus according to claim 5, wherein the pixel value distribution image generation unit generates the stepwise pixel value distribution image by arranging the linear pixel value distributions with respect to the radius R. The pixel value distribution generation unit generates the arc-shaped pixel value distribution whose length is proportional to the square of the radius R,
The medical image diagnosis support apparatus according to claim 5, wherein the pixel value distribution image generation unit generates the fan-shaped pixel value distribution image by arranging the arc-shaped pixel value distributions with respect to the radius R. The pixel value distribution generation unit generates the linear pixel value distribution in which the lengths are normalized and aligned,
The medical image diagnosis support apparatus according to claim 5, wherein the pixel value distribution image generation unit generates the rectangular pixel value distribution image by arranging the linear pixel value distributions with respect to the radius R.   5. The pixel value distribution generation unit generates the pixel value distribution in which a range of the pixel value is taken in a length direction and the distribution amount of the pixel value is expressed by a difference in color based on the histogram. The medical image diagnosis support apparatus according to 1. The distribution amount of the pixel value is the number of the target pixels for each range of the pixel value,
A ratio of the number of target pixels for each range of pixel values, obtained by dividing the number of target pixels for each range of pixel values by the number of all target pixels;
The area of the target pixel for each range of the pixel value when the target area is a minute area obtained by dividing the spherical surface into a plurality of areas,
Or the ratio of the area of the target pixel obtained by dividing the area of the target pixel for each range of the pixel value by the area of the spherical surface when the target area is a minute area obtained by dividing the spherical surface into a plurality of areas. The medical image diagnosis support apparatus according to any one of claims 4 to 9. An image acquisition step of acquiring a tomographic image group composed of a plurality of tomographic images obtained by photographing an object by the image acquisition unit;
A coordinate acquisition step of acquiring coordinates of a three-dimensional center or center of gravity of a lesion or lesion candidate included in the tomographic image group acquired in the image acquisition step by a coordinate acquisition unit;
The pixel value distribution acquisition unit obtains the pixel value distribution of the target pixels of the tomographic image group on the spherical surface having the radius R from the coordinates acquired in the coordinate acquisition step, or the circle having the radius R from the coordinates acquired in the coordinate acquisition step. A pixel value distribution acquisition step of acquiring a plurality of pixel value distributions of target pixels of the tomographic image on the circumference using the radius R as a variable;
A pixel value distribution image generation unit that generates a pixel value distribution image in which the pixel value distribution acquired in the pixel value distribution acquisition step is arranged with respect to the radius R by a pixel value distribution image generation unit;
An output control step of outputting the pixel value distribution image generated in the pixel value distribution image generation step by an output control unit, and a method for operating a medical image diagnosis support apparatus, comprising: An image acquisition function for acquiring a tomographic image group composed of a plurality of tomographic images obtained by photographing a subject;
A coordinate acquisition function for acquiring coordinates of a three-dimensional center or center of gravity of a lesion or lesion candidate included in the tomographic image group acquired by the image acquisition function;
The pixel value distribution of the target pixel of the tomographic image group on the spherical surface having the radius R from the coordinates acquired by the coordinate acquiring function, or the tomographic image on the circumference of the radius R from the coordinates acquired by the coordinate acquiring function. A pixel value distribution acquisition function for acquiring a plurality of pixel value distributions of the target pixel using the radius R as a variable;
A pixel value distribution image generation function for generating a pixel value distribution image in which the pixel value distributions acquired by the pixel value distribution acquisition function are arranged with respect to the radius R;
A medical image diagnosis support program for causing a computer to execute an output control function for outputting the pixel value distribution image generated by the pixel value distribution image generation function.

Images