JP5363962B2 - Diagnosis support system, diagnosis support program, and diagnosis support method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a diagnostic image of the best image quality when a series with a different image direction is designated in the comparative reading of series images. <P>SOLUTION: At least one point representing anatomical characteristics of a subject, included in both volume data V1<SB>org</SB>and V2<SB>org</SB>with different image directions, is identified, and set as a landmark L1 or L2. A unified coordinate system 9 between the volume data is set based on the set landmark. Information on the image direction is compared between the volume data, and if inconsistency in the image direction is detected, the image direction (sagittal) of the volume data (V2<SB>org</SB>) with the maximum slice interval is determined as the image direction of a reconstructed image. Reconstructed image series RSI<SB>1</SB>-RSI<SB>N</SB>with the same image direction and the same cross section are created from the other volume data (V1<SB>org</SB>), and are output together with slice images SI<SB>1</SB>-SI<SB>N</SB>constituting the volume data (V2<SB>org</SB>) with the maximum slice interval to a display screen. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to a system, a computer program, and a method for supporting image diagnosis using data acquired by a tomography apparatus.

  Examples of tomographic apparatuses used in hospitals include X-ray CT (Computed Tomography) apparatus, MRI (Magnetic Resonance Imaging) apparatus, ultrasonic (US: Ultrasonic) diagnostic apparatus, PET (Positron). Emission Tomography (positron tomography) apparatus and the like are known. Both devices have improved spatial resolution due to recent improvements, and have achieved great results in the detection and evaluation of lesion sites. In particular, CT apparatuses and MRI apparatuses are greatly reduced in imaging time due to the increase in the number of detectors and the introduction of a high-speed imaging method.

  In recent years, the number of photographing apparatuses conforming to the DICOM standard has increased, and the construction of an information processing system for storing acquired data and utilizing the stored data is progressing. Many diagnostic support functions using stored data have been proposed. Among them, an image acquired in the most recent examination (hereinafter referred to as the current image) and an image of the same part acquired in the past examination (hereinafter referred to as the past image). ) Are displayed side by side and are widely used for follow-up observation. In recent years, in order to improve diagnosis accuracy, there have been an increasing number of cases where imaging is performed with a plurality of types of imaging apparatuses, and therefore, a comparative interpretation function that displays images of the same part acquired by different apparatuses side by side is also provided.

  Here, since the data acquired by the tomography apparatus is composed of a series of slice images each representing a plurality of cross sections, when comparing two series, it is necessary to associate slice images representing the same cross section with each other. . On the other hand, Patent Documents 1 and 2 propose a method of automatically associating slice images representing cross sections at the same position based on position coordinates in the body axis direction.

Japanese Patent No. 3639030 JP 2008-43524 A

  As data sharing progresses in the hospital due to the construction of an information processing system, doctors in one medical department can use the results of tests performed in the past in other medical departments. As a result, it is expected that two data acquired without assuming the comparison will be used more frequently for the comparison. For example, there is a possibility that lumbar CT data acquired by an orthopedic examination and abdominal CT data acquired by an internal examination may be used for comparative interpretation.

  However, even if the imaging region is the same, the imaging conditions are usually different for different inspection purposes. In particular, in MRI, the image orientation of a slice image can be set arbitrarily, so even if there is MRI data of the same part of the same examinee, the image direction is not always the same. Even with data having different image directions, comparative interpretation can be performed if one data is reconstructed to generate a tomographic image having the same image direction as the other data. However, since the reconstructed image does not always provide an image quality suitable for diagnosis, some device is required to ensure the image quality.

  However, in conventional hospitals, data with different image directions are rarely subjected to comparative interpretation. If data is not shared between clinical departments, doctors can only diagnose using data collected by themselves, and in that case, they do not dare to specify different image orientations for examinations for comparison purposes. . From such a situation, Patent Documents 1 and 2 only propose a correspondence method when the image directions are the same. In the past, no particular consideration has been given to the image quality of reconstructed images generated for comparative interpretation.

  In view of the above circumstances, the present invention provides a comparative interpretation image having the best image quality within a range that can be realized using supplied data when data having different image directions is selected as a comparison target. Objective.

  The diagnosis support system according to the present invention includes, as means for achieving the above object, volume data storage means, volume data selection means, landmark setting means, coordinate system setting means, mismatch detection means, direction determination means, described below, Image reconstruction means and display control means are provided. Further, the diagnosis support program of the present invention includes one or a plurality of computers, volume data storage means, volume data selection means, landmark setting means, coordinate system setting means, mismatch detection means, and direction determination means described below. A program for causing an image reconstruction unit and a display control unit to function. The diagnosis support program is usually composed of a plurality of program modules, and the function of each means is realized by one or a plurality of program modules. These program module groups are recorded on a recording medium such as a CD-ROM or DVD, or recorded in a downloadable state in a storage attached to a server computer or a network storage, and provided to the user. The diagnosis support method of the present invention includes volume data storage means, volume data selection means, landmark setting means, coordinate system setting means, mismatch detection means, direction determination means, image reconstruction means, and display control means described below. This is a method for supporting diagnosis by executing the process.

  The volume data storage means stores a series of slice images acquired by tomography of the subject as volume data in a predetermined storage medium. At this time, the volume data is associated with the image direction, the slice interval, and the subject information of the slice images constituting the series.

  The volume data selection means selects a plurality of volume data representing the same subject from the volume data stored in the storage medium based on the subject information associated with each volume data. For example, data acquired by using the same modality on different shooting dates is selected as a plurality of volume data. Alternatively, data acquired using different modalities on the same shooting date is selected as a plurality of volume data. Alternatively, all volume data representing the same subject may be selected regardless of the imaging date and modality. The selection criterion is preferably switched based on a user instruction.

  The landmark setting means identifies at least one point that represents an anatomical feature of the subject and is commonly included in a plurality of selected volume data, and that point is an anatomical recognition point. That is, it is set as a landmark. As a point representing an anatomical feature, it is preferable to identify a point representing an invariant feature that the subject always has. For example, a specific bone protrusion or center point, a specific blood vessel branch point, or the like is identified.

  Points representing anatomical features are identified, for example, by automatic analysis of volume data. Alternatively, an operation for designating a position in the volume data may be received, and a point designated by the operation may be identified (recognized) as a point representing an anatomical feature. It is preferable that the landmark setting unit selectively executes identification processing based on automatic analysis and identification processing based on a user operation based on a user instruction. In addition, the user may be able to correct the identification result obtained by the automatic analysis.

  The coordinate system setting means sets a unified (common) coordinate system for a plurality of selected volume data based on the set landmark. This makes it possible to specify the same position (cross section, point, etc.) of the subject with the same coordinate value in the selected plurality of volume data.

  The mismatch detection means detects the mismatch of the image direction between the volume data by comparing the information of the image direction associated with the volume data among the selected plurality of volume data. For example, in the case where one of the two selected volume data is volume data composed of a series of axial images and the other is volume data composed of a series of sagittal images, the discrepancy detection means performs image direction Output data indicating inconsistencies.

  The direction determining means selects one volume data having the maximum slice interval from the plurality of volume data when a mismatch in image direction is detected for the selected volume data. Then, the image direction of the one volume data is determined as the image direction of the reconstructed image. For example, of the two selected volume data, one is volume data composed of a series of axial images with a slice interval of 1 mm, and the other is volume data composed of a series of sagittal images with a slice interval of 10 mm. In this case, the direction of the sagittal section is determined as the image direction of the reconstructed image.

  The image reconstructing means uses slices of slice images constituting the one volume data used for determining the image direction based on the unified coordinate system in the other volume data (one or more). Each slice position corresponding to the position is specified. Then, a reconstructed image representing a cross section in the determined image direction is generated at each identified slice position, and a series of reconstructed images is generated.

  The display control means outputs the series of slice images constituting the one volume data used for determining the image direction and the series of reconstructed images generated from the other volume data to the display screen. . The layout of the display screen is a layout in which the slice image at the corresponding slice position can be compared with the reconstructed image.

  According to the present invention, when performing comparative interpretation, when it is necessary to reconstruct an image with an image direction different from the image direction at the time of shooting for some volume data, it is best within the range that can be realized from the supplied volume data. Can be generated and displayed.

Diagram showing schematic configuration of diagnosis support system Figure illustrating the volume data acquisition process Figure illustrating the volume data acquisition process The figure which shows the outline | summary of the process of WS for diagnosis The figure for demonstrating the template T used for the landmark automatic setting process Flow chart showing the outline of landmark automatic setting processing Figure illustrating a landmark manual setting screen Figure illustrating a landmark manual setting screen Flowchart showing the outline of manual landmark setting process Diagram showing how to detect mismatches (when no mismatches are detected) Diagram showing how to detect mismatches (when mismatches are detected) Flow chart showing the outline of the direction determination process

  FIG. 1 shows a schematic configuration of a diagnosis support system according to an embodiment of the present invention. A diagnosis support system 1 according to the present embodiment includes an examination room system 3, a data server 4, and a diagnosis workstation (WS) 6 connected to each other via a local area network (LAN) 2.

  The examination room system 3 includes a modality group 32 for imaging a subject and an examination workstation (WS) 31 for confirming and adjusting an image output from each modality.

  When the modality 32 outputs two-dimensional slice data (for example, a CT (Computed Tomography) device, an MR (Magnetic Resonance) device, etc.), the inspection WS 31 is configured by reconstructing a slice data group. Dimensional volume data is generated, and the generated volume data is transferred to the data server 4 together with the accompanying information. If the modality 32 directly outputs volume data (for example, an MS (Multi Slice) CT device, a cone beam CT device, etc.), the inspection WS 31 receives the volume data together with incidental information in the data server 4. Forward to.

  The data server 4 is obtained by mounting a software program that provides a function of a database management server (DBMS) on a computer having a high performance processor and a large capacity memory and having a relatively high processing capability. The program is stored in the storage, loaded into the memory at startup, and executed by the processor. Thereby, the data server 4 functions as the volume data storage means 41 and the volume data selection means 42 on the server (S) side.

  The volume data storage unit 41 stores the volume data and the auxiliary information transferred from the inspection WS 31 as a file 10 in the large-capacity storage 5 connected to the data server 4. The file 10 is a file conforming to the DICOM (Digital Imaging and Communication in Medicine) standard, and includes a header area and an area for storing volume data.

  In the header area, incidental information transferred from the inspection WS 31 and incidental information for data search added in the data server 4 are recorded as DICOM tag information. For example, as information for specifying the subject, information on the identification number, name, age, sex, and imaging region (head, chest, abdomen, etc.) of the examinee (patient) is recorded. Further, as information for specifying the shooting date, information on the date when the inspection was performed and the time when the shooting was performed is recorded. In addition, information such as the modality used for imaging, imaging conditions (contrast agent used / used dye, used radionuclide, radiation dose, etc.) is also recorded. Further, when the volume data is data acquired by a tomographic apparatus such as CT or MRI, information on slice intervals and image directions at the time of imaging is also recorded in the header area.

  Note that the volume data stored as a file in the large-capacity storage 5 is subjected to some processing in the inspection WS 31, such as deleting volume data acquired by shooting, and deleting unnecessary information from the volume data acquired by shooting. The applied volume data may be stored. For example, volume data from which bone information has been removed or volume data from which only bone information has been extracted may be stored.

  In response to a search request from the diagnostic WS 6, the volume data selection means 42 selects a file that meets the search condition from the plurality of files 10 stored in the large-capacity storage 5, and sends it to the diagnostic WS 6. To do.

  The diagnostic WS 6 is a general-purpose workstation equipped with a standard processor, memory, and storage, and a plurality of programs for providing various diagnostic support functions. These programs are stored in the storage, loaded into the memory at startup, and executed by the processor. As a result, the diagnostic WS 6 includes the volume data selection means 61, the landmark setting means 62, the coordinate system setting means 63, the mismatch detection means 64, the direction determination means 65, the image reconstruction means 66, and the display control means shown in FIG. 67 function as various processing means. The diagnosis WS 6 is connected to a display 7 and an input device 8 such as a mouse or a keyboard.

Hereinafter, processing of the diagnostic WS 6 will be described. Here, in order to facilitate understanding of the invention, a case where comparative interpretation is performed using abdominal CT data acquired by an internal medicine examination and abdominal (lumbar) MRI data obtained by an orthopedic examination is illustrated. While explaining. The abdominal CT data V1 _org is obtained for examining the large intestine by an internal examination, and the abdominal MRI data V2 _org is obtained for examining an abnormality of the lumbar spine by an orthopedic examination. In addition, it is assumed that abdominal CT data and abdominal MRI data are acquired by the process described below.

FIG. 2A shows how the abdominal CT data V1 _org is acquired. Abdominal CT data V1 _org sets the direction perpendicular to the image orientation in the body axis of the examinee, i.e. a slice plane of axial plane AP ₁ ~AP _N, and those obtained by performing CT imaging. Volume data composed of series AI _{1 to} AI _N of slice images acquired by imaging is abdominal CT data V 1 _org . Here, the interval (pixel interval) of voxel data when configuring the volume data is set to the same value as the slice interval. The abdominal CT data V1 _org is stored in the mass storage 5 as the DICOM file 10a together with the header information.

In the internal medicine examination, volume data V1 _tgt is generated by extracting only the large intestine to be diagnosed from the abdominal CT data V1 _org . The volume data V1 _tgt is used for large intestine analysis processing and generation of a diagnostic image to be displayed on the display 7. The volume data V1 _tgt is stored in the mass storage 5 as a DICOM file 10b together with header information.

FIG. 2B shows how the abdominal MRI data V2 _org was acquired. The abdominal MRI data V2 _org is acquired by performing MRI imaging with the direction parallel to the body axis of the examinee set as the image direction. The figure exemplifies a case where photographing is performed using the sagittal planes SP _{1 to} SP _N as slice planes. Volume data consisting of series _SI 1 _{~SI N} of the acquired slice images by photographing a abdominal MRI data _{V2 org.} Again, it is assumed that the interval of voxel data when configuring the volume data is set to the same value as the slice interval. The abdominal MRI data V2 _org is stored in the large-capacity storage 5 as the DICOM file 10c together with the header information.

In the orthopedic examination, volume data V2 _tgt is generated by extracting only the lumbar spine to be diagnosed from the abdominal MRI data V2 _org . The volume data V2 _tgt is used for lumbar spine analysis processing and generation of diagnostic images to be displayed on the display 7. The volume data V2 _tgt is stored in the mass storage 5 as a DICOM file 10d together with header information.

  In the present embodiment, when a comparative interpretation function for a predetermined biological tissue is selected in the function selection menu of the diagnostic WS 6, a data selection process by the volume data selection means 61 is executed. The volume data selection means 61 detects an operation for designating a subject (for example, an operation for designating a patient and an imaging region), and selects volume data representing the designated subject. Specifically, information necessary for specifying the subject is transmitted to the volume data selection means 42 on the data server 4 side, and a search for a file stored in the large-capacity storage 5 is requested.

  The volume data selection means 42 selects a file related to the designated subject from the stored file group and returns it to the volume data selection means 61. The volume data selection means 61 stores the acquired file in the memory of the diagnostic WS 6.

  When the target of comparative interpretation is limited to data acquired by the same modality due to setting or user operation, the volume data selection means 61 accepts an operation for specifying a modality, and includes information for specifying a subject. Information specifying the modality is transmitted to the data server 4. In this case, the volume data selection unit 42 selects a file related to the specified subject and the specified modality from the stored file group, and returns the file to the volume data selection unit 61.

  On the other hand, when the target of comparative interpretation is limited to data acquired on the same shooting date due to setting or user operation, the volume data selection means 61 accepts an operation for specifying the shooting date, Information specifying the shooting date is transmitted to the data server 4 together with the specifying information. In this case, the volume data selection unit 42 selects a file related to the specified subject and the specified imaging date from the stored file group, and returns the file to the volume data selection unit 61.

FIG. 3 is a diagram showing an outline of processing of the diagnostic WS 6. As described above, here, a case where a file including abdominal CT data V1 _org and a file including abdominal MRI data V2 _org are selected will be described as an example. As described with reference to FIGS. 2A and 2B, the abdominal CT data V1 _org and the abdominal MRI data V2 _org are data before the observation target extraction process is performed. Since these data include information on all living tissues in the abdomen, any and the same observation target can be extracted as necessary. The observation target extraction process may be executed immediately after the volume data is selected, but is not particularly limited because it can be executed at another stage.

Abdominal CT data V1 _org was taken with the slice interval set to 0.25 mm and the image direction set to axial, and abdominal MRI data V2 _org was taken with the slice interval set to 1 mm and the image direction set to sagittal. It shall be assumed. As described above, the slice interval and image direction information are recorded as header information (DICOM tag) of a file including volume data.

  The landmark setting means 62 identifies a point representing an anatomical feature (for example, a specific bone or organ protrusion or a specific blood vessel branch point) in each of the volume data included in the selected file, The identified point is set as a landmark. The landmark is set by storing the position coordinates of the identified point in the memory.

The feature identified by the landmark setting means 62 must be a feature that is commonly included in the selected plurality of volume data. In the illustrated case, the abdominal CT data V1 _org and the abdominal MRI data V2 _org have different acquisition purposes, so the imaging ranges are not necessarily the same. However, since the large intestine and lumbar vertebra are diagnostic targets, the lumbar vertebra should be included in common. In this case, one or more points on the lumbar spine can be set as landmarks.

  Among the lumbar vertebrae, the fifth lumbar vertebra can be distinguished from other lumbar vertebrae in that it is adjacent to the sacrum. Further, it has a characteristic shape, and for example, the tips of the left and right protrusions are relatively easy to identify even by analysis processing by a computer. Therefore, hereinafter, a case where a point on the fifth lumbar vertebra is set as a landmark will be described as an example.

  In the present embodiment, the landmark setting means 62 executes two types of processing: automatic setting processing for setting landmarks without user intervention, and manual setting processing for setting landmarks based on user operations. Can do. Which process is executed is selected based on the setting made by the user.

  Hereinafter, the automatic setting process of the landmark setting unit 62 will be described. FIG. 4 is a diagram for explaining a template T used for automatic setting, and FIG. 5 is a flowchart showing an outline of automatic setting processing.

  A template used for landmark automatic setting processing is stored for each imaging region in the memory of the diagnostic WS 6. The template is data representing a three-dimensional shape pattern representing an anatomical structure having features that are distinctive in shape compared to surrounding structures and one or more feature points in the three-dimensional shape pattern. . FIG. 4 illustrates a template T on which a pattern representing the shape of the fifth lumbar vertebra and the positions of points Pa, Pb, and Pc in the pattern are recorded. Each template has its own coordinate system defined, and the position of the feature point is stored as a coordinate value of the coordinate system.

When the automatic setting function is selected, the landmark setting means 62 identifies the feature points by automatic analysis of volume data. First, the landmark setting unit 62 refers to one of the files selected by the volume data selection unit 61 and determines the imaging region based on the header information of the file. Alternatively, the imaging region is determined by acquiring information on the imaging region designated at the time of volume data selection (S101). In the illustrated case, since the abdominal CT data V1 _org and the abdominal MRI data V2 _org are both abdominal data, the imaging region is determined to be “abdomen”.

  Next, the landmark setting unit 62 reads a template stored for the imaging region from the memory (S102). In the illustrated case, the template T illustrated in FIG. 4, that is, data representing the shape pattern of the fifth lumbar vertebra and the feature points Pa, Pb, and Pc are read.

  Thereafter, the landmark setting means 62 reads the volume data from the files selected by the volume data selection means 61, respectively. Then, for each of the volume data, matching with a template is performed for each region, and a region (corresponding region) including a structure having the same shape as the template pattern is searched (S103). The landmark setting means 62 further searches the extracted corresponding area for a point corresponding to the feature point indicated by the template T (S104). Then, the XYZ coordinate values of the points identified by the search are stored in the memory as landmark coordinates (S105). Subsequently, the processing of steps S103 to S105 is repeated for the next volume data read from the file (S106).

In the illustrated case, a region corresponding to the fifth lumbar vertebra in the abdominal CT data V1 _org is extracted by matching each region of the abdominal CT data V1 _org with the template T in step S103. Subsequently, in step S104, the feature point _P a, points corresponding to _{P b} and _{P c L1} a, _{L1 b} and L1 _c are identified. Thereby, as shown in the upper part of FIG. 4, points L1 _a , L1 _b, and L1 _{c serving as} landmarks are set in the abdominal CT data V1 _org . Subsequently, the same processing as target abdominal MRI data _{V2 org} is performed, as shown in the lower part of FIG. 4, during abdominal MRI data _{V2 org,} point a landmark _L2 a, L2 _b and L2 _c is set Is done.

  The above processing is processing when two volume data are selected by the volume data selecting means 61. However, when three or more volume data are selected, the processing of S103 to S105 is repeated to similarly perform land processing. A mark can be set.

  In the present embodiment, “Matching Local Self-Similarities across Images and Videos” (Shechtman E, Irani M, Computer Vision and Pattern Recognition 2007, IEEE Conference on 17-22 June 2007, Apply the method described in pages 1-8). In this method, matching is performed for descriptors obtained by performing log-polar mapping of the surrounding area of individual pixels that make up a still or moving image frame. is there. Since this method determines the match / mismatch of images based on descriptors representing geometric features, if the geometric features are the same, even if the images have completely different colors and textures, Can be recognized.

  The method proposed by this document as a moving image matching method is applied to matching processing of volume data output by a tomography apparatus by regarding one frame of a moving image as a slice image and considering a frame group of the moving image as volume data. can do. In the case of data acquired by different shooting modalities, the value (color or density) of the voxel data and the sharpness of the edge are different even if the shooting target is the same. However, according to the above method, depending on the different shooting modalities Corresponding points can be accurately identified even between acquired data.

  However, the search method for the corresponding region and the corresponding point is not limited to the above method, and other known methods may be adopted. Also, a plurality of types of search programs may be installed in the diagnostic WS 6 and the search method may be switched depending on whether the modalities that acquired the plurality of volume data are the same or different.

  The landmark automatic setting process has been described above. Subsequently, the manual setting process of the landmark setting unit 62 will be described with reference to FIGS. 6A, 6B and FIG. In the present embodiment, the user can set a landmark by performing an operation of designating a desired point on a volume rendering (VR) image displayed on the screen of the display 7. 6A and 6B are diagrams showing display examples of the setting screen, and FIG. 7 is a flowchart showing an outline of the manual setting process.

When the manual setting function is selected, the landmark setting unit 62 generates a VR image from the volume data included in the file selected by the volume data selection unit 61. Then, the generated image is displayed on the screen of the display 7 (S201). In the illustrated case, VR images representing the abdomen are generated and displayed from the abdominal CT data V1 _org and the abdominal MRI data V2 _org . FIG. 6A shows a VR image generated from the abdominal CT data V1 _{org in the} image window W1 and a VR image generated from the abdominal MRI data V2 _{org in the} image window W2 on the setting screen where the image windows W1 and W2 are arranged. , Each displayed state.

  Next, the landmark setting unit 62 accepts an area designating operation for the VR image being displayed (S202). For example, an area designation operation by moving the cursor C is accepted in each of the image windows W1 and W2 on the screen shown in FIG. 6A. FIG. 6A illustrates a case where regions R1 and R2 including the fifth lumbar vertebra are designated.

  When an area is designated on the VR image, the landmark setting unit 62 extracts a corresponding area from each volume data based on the coordinate value indicating the area designated on the screen (hereinafter, the extracted area). Is referred to as “partial volume data”). Subsequently, an enlarged region VR image representing only the designated region is generated from the extracted partial volume data and displayed (S203). The region enlarged VR image is a VR image representing the structure included in the designated region with a higher resolution than the VR image generated and displayed in step S201.

For example, in the example shown in the figure, when the regions R1 and R2 including the fifth lumbar vertebra are designated on the screen shown in FIG. 6A, the regions corresponding to the regions R1 and R2 from the abdominal CT data V1 _org and the abdominal MRI data V2 _org , respectively. Is extracted, and an enlarged region VR image representing the periphery of the fifth lumbar vertebra is generated from the extracted partial volume data and displayed. For example, when an area designation operation is performed on the screen shown in FIG. 6A, image windows W3 and W4 appear below the image windows W1 and W2 as shown in FIG. 6B, and the abdominal CT data V1 _org is displayed in the image window W3. The enlarged region VR image generated from the volume data is displayed in the image window W4, and the enlarged region VR image generated from the partial volume data of the abdominal MRI data V2 _org is displayed.

  Next, the landmark setting unit 62 accepts a point designation operation on the region enlarged VR image (S204). Then, the XYZ coordinate values of the designated point are stored in the memory as landmark coordinates (S205). For example, a point designation operation by moving the cursor C is accepted in each of the image windows W3 and W4 on the screen shown in FIG. 6B. FIG. 6B illustrates a case where the tip of the protrusion of the fifth lumbar vertebra is designated on the image window W3.

  When three or more volume data are selected by the volume data selection means 61, the landmark setting means 62 arranges the number of image windows corresponding to the number of selected volume data on the screen, and each window In addition, a VR image and an enlarged VR image generated from individual volume data are displayed.

  Next, referring to FIG. 3 again, the processing of the coordinate system setting unit 63 will be described. The coordinate system setting unit 63 sets a coordinate system unified among the selected volume data based on the landmark set by the landmark setting unit 62. In other words, by arranging a plurality of selected volume data in a common coordinate space, the same point in terms of anatomy can be specified by the same coordinate value in any volume data.

Specifically, any one volume data coordinate system is used as a reference coordinate system, and other volume data is enlarged, reduced, rotated, or translated by affine transformation or the like, so that the volume data coordinate system becomes the reference coordinate system. Match with the system. The figure illustrates a case where the coordinate system of the abdominal MRI data V2 _org is the reference coordinate system 9, and the abdominal CT data V1 _org is moved to the coordinate space represented by the reference coordinate system 9 by translation. By this processing, the landmark L1 and the landmark L2 are represented by the same coordinate value under the reference coordinate system 9. Further, not only the landmark but also the points and cross sections in the abdominal CT data V1 _org and the abdominal MRI data V2 _org can be specified by the same coordinate value if the anatomical position is the same.

  Note that if the difference in the coordinate system of the two volume data is only a deviation in the body axis direction, one landmark is sufficient, but in order to be able to cope with any difference, at least three landmarks are set. It is necessary to do. If three landmarks are set so that all three points are not included in the same plane, any deviation in the XYZ directions can be dealt with in the rotation direction and the difference in photographing magnification.

  Next, the processing of the mismatch detection unit 64 will be described with reference to FIGS. 3, 8A and 8B. The mismatch detection unit 64 sequentially refers to the header area of the file selected by the volume data selection unit 61, and compares the image direction information recorded in the header area. 8A and 8B illustrate a case where three files are selected.

As shown in FIG. 8A, in all the files 10a, 10b, and 10c, when the image directions recorded in the header area are the same, that is, the image directions at the time of shooting of the volume data V1 _org , V2 _org , V3 _org are If they match, the mismatch detection means 64 outputs an output value (for example, 0) indicating the match. On the other hand, as shown in FIG. 8B, in some files 10d, when the image direction recorded in the header area is different from the image direction recorded in the header area of other files 10a and 10c, there is a mismatch. The detection unit 64 outputs an output value (for example, 1) indicating the mismatch of the image directions.

  Next, the processing of the direction determining unit 65 will be described with reference to FIGS. 3 and 9. The direction determining unit 65 determines whether or not the image direction mismatch is detected based on the output value of the mismatch detection unit 64 (S301). If no mismatch is detected, that is, if the image directions recorded in the header area in all files match as in the example of FIG. 8A, the image direction of the reconstructed image is set to the matching direction. (Axial in the example of FIG. 8A) is determined (S307).

When the image direction mismatch is detected, the direction determining unit 65 searches for data having the maximum slice interval while sequentially referring to the slice interval values recorded in the header area of each file (S302). . In the example of FIG. 3, the value of the slice interval recorded in the header area of each file is 0.25 mm for a file including abdominal CT data and 1 mm for a file including abdominal MRI data V2 _org. A file containing V2 _org is selected.

  If there is only one file containing data with the maximum slice interval (S303), the direction determining means 65 determines the image direction of the reconstructed image as the image direction of the data with the maximum slice interval (S306).

  On the other hand, when there are a plurality of files including data with the maximum slice interval (S303), the direction determining means 65 displays the file name and header information (including the image direction and slice interval values) of each file on the display 7. Is displayed in a predetermined format on the screen. A file selection operation by the user is accepted (S304), and the image direction of the reconstructed image is determined as the image direction of the data included in the file selected by the user.

  2A and 2B, when the slice interval and the voxel data interval are the same, the pixel interval (Pixel) is used instead of the tag information indicating the slice interval (Spacing Between Slices). Tag information indicating Spacing may be referred to.

Next, the image reconstruction unit 66 will be described. As shown in FIG. 1, the image reconstruction unit 66 is supplied with information on the reference coordinate system 9 set by the coordinate system setting unit 63 and information on the image direction determined by the direction determination unit 65. As shown in FIG. 3, the image reconstruction means 66 arranges the volume data (abdominal MRI data V2 _org ) used for determining the image direction in the coordinate space represented by the reference coordinate system 9, and the volume data is the position coordinates of each slice images composing (sagittal image _SI 1 _{~SI N),} to identify respectively. In the illustrated case, Y-axis of the reference coordinate system 9, because it is perpendicular to the sagittal image _SI 1 _{~SI N,} the position of the sagittal image _SI 1 _{~SI N} is the Y-direction coordinate values _y 1 ~y _N Shall be specified.

Subsequently, the image reconstruction unit 66 arranges other volume data (abdominal CT data V1 _org ) in the coordinate space represented by the reference coordinate system 9. Then, the position corresponding to the slice position coordinates y _{1 to} y _N specified in the volume data (abdominal MRI data V2 _org ) used for determining the image direction is specified in the volume data. In the illustrated case, a sagittal section whose slice position coordinates in the Y direction are y _{1 to} y _N is defined in the abdominal CT data V1 _org by this processing. Then, the image reconstruction unit 66 uses the voxel data constituting the abdominal CT data _{V1 org,} to reconstruct the series _RSI 1 _{~RSI N} images that represent those sagittal section.

Through the above processing, a series of slice images (sagittal images SI _{1 to} SI _N ) constituting volume data (abdominal MRI data V2 _org ) used for determining the image direction, and other volume data (abdominal CT data V1). _org ), a series of reconstructed slice images (reconstructed sagittal images RSI _{1 to} RSI _N ) are obtained.

Under the reference coordinate system 9, a sagittal plane in the abdominal CT data _{V1 org} position _{y i (1 ≦ i ≦ N} ) , the sagittal plane in the abdominal MRI data _{V2 org} position _{y i (1 ≦ i ≦ N} ) is Anatomically, the cross-sections are the same, and thus information useful for diagnosis can be obtained by comparing images at corresponding positions. The display control unit 67 displays the images at the corresponding positions on the screen in a layout that allows easy comparison.

For example, as illustrated in FIG. 3, a series _RSI 1 _{~RSI N} of reconstructed sagittal image are arranged side by side in the horizontal direction on a screen line of the display 7, similarly horizontal series _SI 1 _{~SI N} sagittal image on the screen lower part The images are arranged side by side in the direction, and the display is controlled so that the images at corresponding positions are arranged vertically. Furthermore, when a scroll operation that switches the display range of series images is performed, the display on the upper part of the screen and the display on the lower part of the screen are linked to control the display so that the images at the corresponding positions are always aligned vertically. .

  In the description with reference to FIG. 3, two volume data having different image directions are illustrated, but the processing executed by each means is the same when three or more volume data in an arbitrary image direction are selected. is there. In the above description, the comparison between CT data and MRI data has been exemplified. However, the processing executed by each unit is the same for data acquired by other modalities such as CT data and PET data.

In the above description, the case where the abdominal CT data V1 _org of FIG. 2A and the abdominal MRI data V2 _org shown in FIG. 2B are used for interpretation is illustrated. However, for example, comparative interpretation is performed for diagnosis of the lumbar spine. For example, abdominal MRI data V2 _tgt may be used instead of abdominal MRI data V2 _org . That is, an image for comparison may be generated using data obtained by performing a new bone extraction process on abdominal CT data V1 _org and abdominal MRI data V2 _tgt which has already been subjected to the bone extraction process.

  As described above, in the present embodiment, the image direction at the time of shooting does not match among a plurality of series designated as comparison targets, and at least a part of the volume data has a required image direction. When a slice image must be generated by reconstruction processing, the reconstruction image is generated from the coarsest volume data by setting the image direction of the reconstruction image to the image direction of the series having the largest slice interval. There is no such thing. For this reason, it is possible to obtain a comparative interpretation image having the best image quality within a range that can be realized using the designated series.

  If the system of this embodiment is used, the data acquired for the different test | inspection objectives can be effectively utilized as data for comparative interpretation like the case illustrated. If the number of examinations (imaging) can be reduced by using data obtained in other departments or data obtained by different examinations in the same department, the physical and economic burden on the examinee will be reduced. Is done.

  In the above-described embodiment, either the landmark automatic setting process or the manual setting process is selectively executed. The landmark is temporarily set by the above-described automatic setting process, and then illustrated in FIG. 6B. The temporarily set landmark may be displayed on the screen, and the position of the landmark may be adjusted by the manual setting process described above.

  Moreover, in the said embodiment, although the coordinate system setting means has set XYZ coordinate system as a unified coordinate system, the kind of coordinate system is not specifically limited, For example, a polar coordinate system may be sufficient.

  The above embodiment is a client / server type system, but a single computer has a volume data storage unit, a volume data selection unit, a landmark setting unit, a coordinate system setting unit, a mismatch detection unit, and a direction determination unit. Further, the image reconstruction unit and the display control unit may be provided.

  Thus, the present invention is not limited to the above embodiment, and various modifications can be made without departing from the spirit of the present invention.

1 diagnosis support system, 32 modality groups, 5 large capacity storage,
9 coordinate system, 10, 10a-10d file AP1-APN axial plane,
AI1-AIN A series of axial images,
SP1-SPN sagittal surface,
SI ₁ -SI _N sagittal image series,
V1 _org , V2 _org volume data,
V1 _tgt , V2 _tgt volume data,
L1, L2 landmark (point),
RSI ₁ -RSI _N reconstructed sagittal image,
T template, Pa ~ Pc feature point,
W1-W4 Image window, C cursor, R1, R2 area

Claims (7)

Volume data storage means for storing a series of slice images acquired by tomography of a subject in a predetermined storage medium as volume data associated with the image direction of the slice image, the slice interval, and the subject information; ,
Volume data selection means for selecting a plurality of volume data representing the same subject from the volume data stored in the storage medium based on the information of the subject;
A landmark setting means for identifying at least one point that is an anatomical feature of a subject and is commonly included in a plurality of selected volume data, and setting the point as a landmark;
Coordinate system setting means for setting a unified coordinate system for the plurality of volume data based on the set landmark,
A mismatch detection means for detecting a mismatch in the image direction between the volume data by comparing the information in the image direction between the plurality of volume data;
When the discrepancy between the image directions is detected, one volume data having the maximum slice interval is selected from the plurality of volume data, and the image direction of the one volume data is determined based on the image of the reconstructed image. Direction determining means for determining the direction;
In each of the other volume data, based on the unified coordinate system, the slice position corresponding to the slice position of each slice image constituting the one volume data is specified, and at each specified slice position, Image reconstruction means for generating a series of reconstructed images by generating a reconstructed image representing a cross-section in the determined image direction;
A series of slice images constituting the one volume data and a series of reconstructed images generated from the other volume data can be compared with slice images and reconstructed images at corresponding slice positions. And a display control means for outputting to the display screen.   The diagnosis support system according to claim 1, wherein the landmark setting unit identifies a point representing the anatomical feature by automatic analysis of the volume data. The landmark setting unit, wherein the receiving an operation for specifying a position in the volume data, a point specified by the manipulation, characterized in that said identified as points representing anatomical features of claim 1 Symbol placement Diagnosis support system.   4. The diagnosis support system according to claim 1, wherein the volume data selection means selects a plurality of volume data acquired by the same modality on different photographing dates. The volume data selecting means, the diagnosis support device according to any one of claims 1 to 3, characterized in that selects a plurality of volume data obtained by different modalities in the same shooting date. One or more computers,
Volume data storage means for storing a series of slice images acquired by tomography of a subject in a predetermined storage medium as volume data associated with the image direction of the slice image, the slice interval, and the subject information;
Volume data selection means for selecting a plurality of volume data representing the same subject from the volume data stored in the storage medium based on the information on the subject,
Landmark setting means for identifying at least one point that is an anatomical feature of the subject and is commonly included in the selected plurality of volume data, and setting the point as a landmark;
A coordinate system setting means for setting a unified coordinate system for the plurality of volume data based on the set landmark;
A mismatch detection means for detecting a mismatch in the image direction between the volume data by comparing the information in the image direction between the plurality of volume data;
When the discrepancy between the image directions is detected, one volume data having the maximum slice interval is selected from the plurality of volume data, and the image direction of the one volume data is determined based on the image of the reconstructed image. Direction determining means for determining the direction,
In each of the other volume data, based on the unified coordinate system, the slice position corresponding to the slice position of each slice image constituting the one volume data is specified, and at each specified slice position, An image reconstruction means for generating a series of reconstructed images by generating a reconstructed image representing a cross section in the determined image direction, a series of slice images constituting the one volume data, and the other The series of reconstructed images generated from each volume data of the above is functioned as a display control means for outputting to a display screen in a layout in which a slice image at a corresponding slice position and a reconstructed image can be compared. Diagnosis support program. Volume data storage processing for storing a series of slice images acquired by tomography of a subject in a predetermined storage medium as volume data associated with the image direction of the slice image, the slice interval, and the subject information; ,
Volume data selection processing for selecting a plurality of volume data representing the same subject based on the information on the subject from the volume data stored in the storage medium;
A landmark setting process that identifies at least one point that is an anatomical feature of the subject and is commonly included in a plurality of selected volume data, and sets the point as a landmark;
A coordinate system setting process for setting a unified coordinate system for the plurality of volume data based on the set landmark;
A mismatch detection process for detecting a mismatch in image direction between volume data by comparing the information in the image direction between the plurality of volume data;
When the discrepancy between the image directions is detected, one volume data having the maximum slice interval is selected from the plurality of volume data, and the image direction of the one volume data is determined based on the image of the reconstructed image. Direction determination processing to determine the direction;
In each of the other volume data, based on the unified coordinate system, the slice position corresponding to the slice position of each slice image constituting the one volume data is specified, and at each specified slice position, An image reconstruction process for generating a series of reconstructed images by generating a reconstructed image representing a cross-section in the determined image direction;
A series of slice images constituting the one volume data and a series of reconstructed images generated from the other volume data can be compared with slice images and reconstructed images at corresponding slice positions. And a display control process for outputting to a display screen.

Images