JP2008200373A - Similar case retrieval apparatus and its method and program and similar case database registration device and its method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable to retrieve with lower processing load and higher precision in a similar case retrieval about tomographic images acquired by three-dimensional medical images. <P>SOLUTION: The similar case retrieval apparatus associates a similar case database 9 with the slice thickness, the characteristic amount showing the content characteristics of tomographic images, and a support information of interpretation/diagnosis of radiogram based on the tomographic images about a plurality of tomographic images having different slice thickness. During the retrieval, the similar case retrieval part 25 refers to the similar case database 9 and obtains the similar case information associated with the slice thickness and the characteristic amount in which the coincidence of the slice thickness T<SB>1</SB>of the retrieval subject image I<SB>1</SB>and the characteristic amount C<SB>1</SB>showing the content characteristics of the retrieval subject image I<SB>1</SB>obtained by the characteristic amount obtaining part 24 meets the predetermined standard. During the registration, the image reconfiguration part 31 reconfigures an image I<SB>n</SB>with different slice thickness from that of the image I<SB>1</SB>, and a similar case information registration part 36 associates the slice thickness, the characteristic amount and the interpretation/diagnosis support information about the image I<SB>1</SB>and the reconfigured image I<SB>n</SB>and registers. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an apparatus and method for searching for similar cases and registering a database for a tomographic image reconstructed from a three-dimensional medical image, and a program for causing a computer to execute the method.

  In order to support the interpretation of medical images by interpreting doctors and diagnosis based on medical images by clinicians, medical images and information related to the images, such as interpretation results and diagnosis results, are stored in a database, and the interpretation / diagnosis target A similar case retrieval system has been proposed that retrieves from a database information such as an image similar to a medical image and information such as interpretation / diagnosis results associated with the image.

  For example, search similar image data having a portion whose image characteristics are similar to the input search image data from an image database storing a large number of image data representing subject images, and output the search results. A system is known (for example, Patent Document 1).

The feature amount (shape information and texture information) of the subject image and diagnostic information related to the image are stored in a database, and the feature amount calculated from the input search image data and the feature stored in the database are stored. A system that searches for diagnostic information related to an image similar to a search image based on the amount and outputs a search result is known (for example, Patent Document 2).
JP 2004-005364 A Japanese Patent Laying-Open No. 2005-065728

  When a similar case search for a tomographic image representing a cross-section obtained by slicing a region to be examined obtained by reconstruction of a three-dimensional medical image obtained by CT, MRI or the like is to be realized with a conventional similar case search system, a three-dimensional medical Since a large number of tomographic images can be obtained from the image, the database capacity of the system described in Patent Document 1 in which the image data is stored in the image database as it is becomes very large. In addition, in order to determine the similarity of images, it is necessary to obtain a feature value that represents the characteristics of the image. In this system, it is necessary to obtain the similarity of the feature values that represent the characteristics of the image in the database each time the search is performed. As a result, if the number of images stored in the database increases, the processing load on the system also becomes very high.

  In addition, in a tomographic image obtained by reconstructing a three-dimensional medical image, when the slice thickness is changed, the characteristics of the image also change, and the feature amount representing the characteristics of the image also changes. Since the slice thickness is not taken into consideration in the system, if the slice thickness of the tomographic image to be searched is different from the slice thickness of the tomographic image registered in the database, images with similar image characteristics cannot be searched correctly. There is a possibility that the search accuracy will be lowered.

  The present invention has been made in view of the above circumstances, and an apparatus, a method, and a program that enable a higher-accuracy search with a lower processing load in a similar case search for a tomographic image obtained from a three-dimensional medical image. Is intended to provide.

  The present invention relates to a slice thickness of a tomographic image representing a cross section obtained by slicing a specific examination target site of one or more subjects with a predetermined slice thickness, and a feature amount representing a content feature of at least a part of the tomographic image. Slice thickness in a similar case database in which similar case information in which tomographic images having similar content features and / or interpretation / diagnosis support information supporting diagnosis based on the similar images is associated is registered. Similar case information about a plurality of tomographic images having different values is registered, and similar cases are searched using this database.

  That is, the similar case retrieval apparatus of the present invention includes a slice thickness of a tomographic image representing a cross section obtained by slicing a specific examination target region of one or more subjects with a predetermined slice thickness, and at least a partial region of the tomographic image. Similar case information is registered in which feature quantities representing content features and interpretation / diagnosis support information that supports interpretation of tomographic images with similar content features and / or diagnosis based on similar images are registered A similar case database, in which similar case information for a plurality of tomographic images having different slice thicknesses is registered, and at least one of tomographic images to be searched representing a cross section obtained by slicing the specific examination target part of the subject The feature amount acquisition means for acquiring the feature amount representing the content feature of the part area, the similar case database, and the slice thickness and the search target image Characterized in that the degree of match between the feature amount is provided a similar case search means for obtaining the similar case information associated with slice thickness and the feature meets a predetermined criterion.

  The similar case search method of the present invention includes a step of acquiring a feature amount representing a content feature of at least a partial region of a tomographic image to be searched that represents a cross section obtained by slicing a specific examination target region of a subject, and one or more steps A slice thickness of a tomographic image representing a cross section obtained by slicing the specific examination target portion of the subject with a predetermined slice thickness, a feature amount representing the content feature of at least a partial region of the tomographic image, and A similar case database in which similar case information associated with interpretation / diagnosis support information associated with interpretation / diagnosis support information for assisting diagnosis based on interpretation of tomographic images with similar content features and / or similar images is registered. By referring to the registered similar case information for multiple different tomographic images, the degree of matching with the slice thickness and feature amount of the image to be searched satisfies the predetermined criteria. Characterized by a step of acquiring similar case information associated with slice thickness and the feature.

  The similar case search program of the present invention causes a computer to execute each step of the above similar case search method.

  The similar case database registration apparatus according to another aspect of the present invention includes a slice thickness of a tomographic image representing a cross section obtained by slicing a specific examination target region of one or more subjects with a predetermined slice thickness, and at least one of the tomographic images. A feature amount representing the content feature of a part area and an interpretation / diagnosis support information that supports interpretation of a tomographic image with similar content feature and / or diagnosis based on the similar image are registered in the database in association with each other. The similar case database registration apparatus having registration means includes a region of interest in the three-dimensional medical image with a plurality of different slice thicknesses from the three-dimensional medical image representing the examination target part of one subject. Image reconstruction means for reconstructing a plurality of tomographic images representing sliced cross sections, and the content of at least the region of interest of each reconstructed tomographic image A feature amount acquisition unit that acquires a feature amount representing a symptom, and the registration unit is reconstructed by the image reconstruction unit with respect to the interpretation / diagnosis support information about the examination target region of the one subject. In addition, the slice thickness and the feature amount of each tomographic image are registered in association with each other.

  The similar case database registration method of the present invention slices from a three-dimensional medical image representing a specific examination target region of one subject so as to include a region of interest in the three-dimensional medical image with a plurality of different slice thicknesses. Reconstructing a plurality of tomographic images representing a cross section; obtaining a feature quantity representing at least the content feature of the region of interest of each reconstructed tomographic image; and Each of the reconstructed information for the interpretation / diagnosis support information for supporting the interpretation of the tomographic image representing the cross-section of the examination target region and the tomographic image having the similar content feature and / or the diagnosis based on the similar image. And a step of associating the slice thickness of the tomographic image with the feature quantity and registering them in a database.

  The similar case database registration program of this invention makes a computer perform each step of said similar case database registration method.

  Next, details of the present invention will be described.

  The “subject” is an individual who has taken an image representing a region to be examined, and may be other than a human being.

  A specific example of the “specific examination site” includes a human chest.

  The “tomographic image” is reconstructed from a three-dimensional medical image acquired by a modality such as CT or MRI. Specifically, a cross-section perpendicular to a predetermined line-of-sight direction in the three-dimensional medical image This is obtained by a process of projecting a pixel in a three-dimensional region specified by a slice thickness representing a thickness in the line-of-sight direction onto a projection plane perpendicular to the line-of-sight direction.

  Specific examples of the “partial region of the tomographic image” include a region of interest in the tomographic image. The region of interest is a region that is noticed when interpretation or diagnosis is performed, and has a large influence on the diagnosis by the characteristics of the region, and is usually a region called ROI (Region of Interest). This region of interest may be specified by the observer of the tomographic image by using a pointing device or the like while viewing the screen on which the image is displayed, or may be detected by a known image analysis method. For example, a contour extracted by the latter method may be extracted from the region specified by the former method, or the former may be extracted from the region detected by the latter method. It may be selected by the method.

  The “content feature” is a feature related to the content of the image, and specific examples include features such as brightness, edge, shape, size, and isolated region. Further, a combination of a plurality of types of features may be used.

  The “feature amount” is obtained by quantifying the content features described above, and can be obtained by a known image analysis method.

  “Representing the characteristic features of at least a part of the tomographic image” may represent the characteristic features of a part of the tomographic image, such as the region of interest, or the entire tomographic image. May represent a characteristic of the image, may represent both of them, or may represent a characteristic of a part of a region in a tomographic image, and a content of a region other than the part It may be a combination with a feature.

  Specific examples of the “interpretation / diagnosis support information” include an interpretation report of the tomographic image by a radiologist and the like, final definitive diagnosis information on the examination target part of the subject, and the like of the subject. It may be a tomographic image itself of the examination target site. Details of each specific example will be described below.

  The interpretation report and the final confirmed diagnosis information are obtained by associating the interpretation information indicating the interpretation result of the medical image of the subject and / or the diagnostic information indicating the diagnostic content of the subject with the medical image of the subject. It may be generated based on interpretation information and / or diagnosis information acquired from the storage means for storing.

  Specific examples of this storage means include Radiology Information Management System (RIS), Medical Image Storage and Communication System (PACS), Hospital Information System HIS (Hospital Information System: HIS), etc. A database for storing and storing interpretation reports and electronic medical records provided in a system for managing medical information.

  In addition, “stores (interpretation information and / or diagnosis information) in a state in which the medical image is associated with the medical image” is based on the interpretation information and / or the diagnosis information and a part of information of the medical image. It means storing directly or indirectly in a state where it can access the other. As a specific example of “directly accessible state”, the interpretation information includes a report ID for identifying itself and an image ID for identifying the interpreted medical image, while the medical image identifies itself. An image ID to be identified and a report ID to identify an interpretation report of the image are attached. From the interpretation information side, a medical image is directly accessed based on the image ID, or conversely, from the medical image side. A case where the information is stored so that the interpretation information can be directly accessed based on the report ID. As a specific example of “indirectly accessible state”, the diagnostic information includes a patient ID for identifying itself and an examination ID for identifying an examination that the patient has received, while medical images include An image ID for identifying itself and an examination ID for identifying an examination including imaging of the image are attached, and further, an examination ID for identifying the examination, a patient ID for identifying the patient who has undergone the examination, and an examination were taken. There is examination information including an image ID for identifying an image. From the diagnosis information side, examination information is specified based on the examination ID, and an image photographed in the examination based on the image ID included in the examination information is displayed. By specifying, the medical image is indirectly accessed based on the diagnostic information, or from the medical image side, the examination information is identified based on the examination ID, and the patient is identified based on the patient ID included in the examination information. By identifying the diagnostic information, it includes conditions that can be or indirect access to diagnostic information based on the medical image.

  Note that the “storage unit” does not need to store the medical image itself associated with the interpretation information and / or the diagnostic information.

  Further, the “interpretation information” may be an interpretation report in which the interpretation results by a plurality of radiologists are described, and the “diagnosis information” includes the final confirmed diagnosis information and the findings based on the medical image. The present invention is not limited to this, and may include all information generated in the course of diagnosis for the subject, such as the results of other examinations and interviews, and the contents of treatment.

  “Generate interpretation / diagnosis support information based on interpretation information and / or diagnosis information” means that predetermined information is extracted from the interpretation information and / or diagnosis information to generate interpretation / diagnosis support information. The interpretation information and / or diagnosis information may be used as interpretation / diagnosis support information as they are.

  When registering an interpretation report and final confirmed diagnosis information in the similar case database, first, the three-dimensional medical image having a plurality of different slice thicknesses from a three-dimensional medical image representing a specific examination target region of one subject. And reconstructing a plurality of tomographic images representing sliced slices so as to include the region of interest, and at least one of the three-dimensional medical image and the tomographic image reconstructed from the three-dimensional medical image from the storage unit. Interpretation information and / or diagnosis information associated with one image is generated, interpretation / diagnosis support information is generated based on the acquired interpretation information and / or diagnosis information, and for the generated interpretation / diagnosis support information, The slice thickness and the feature amount of the reconstructed tomographic image may be registered in association with each other. First, the interpretation information and / or diagnostic information is stored from the storage unit. And generating interpretation / diagnosis support information based on the obtained interpretation information and / or diagnosis information, and from the three-dimensional medical image associated with the obtained interpretation information and / or diagnosis information, at a plurality of different slice thicknesses , Reconstructing a plurality of tomographic images representing cross sections sliced so as to include a region of interest in the three-dimensional medical image, and reconstructing slices of the reconstructed tomographic image with respect to the previously generated interpretation / diagnosis support information The thickness and the feature quantity may be registered in association with each other.

  Further, the “interpretation / diagnosis support information” includes two or more of the plurality of tomographic images of the examination target region of the subject, and each of the two or more tomographic images and the slice thickness of each image. When searching for similar cases, the degree of coincidence with the slice thickness of the image to be searched among two or more tomographic images included in the interpretation / diagnosis support information satisfies a predetermined criterion. It is preferable to acquire a slice thickness tomographic image.

  The “interpretation / diagnosis support information” may be link information indicating a place where the substantial content is stored. Specific examples include an interpretation report of the tomographic image, an electronic medical record of the subject, a name of a database in which the tomographic image itself is stored, search key information for acquiring the information, and the like.

  The “similar case information” is a plurality of different slice thicknesses representing a cross section of the examination target portion of one subject with respect to the interpretation / diagnosis support information about the examination target portion of one subject. The slice thickness of each tomographic image and the feature amount may be associated with each other. That is, “similar case information” can be obtained by associating a plurality of slice thicknesses with feature amounts for each slice thickness with respect to interpretation / diagnosis support information of one subject. The similar case database registration apparatus / method / program according to the present invention is for registering such similar case information.

  The “predetermined criterion (for the degree of matching of slice thickness and feature amount)” in the search is the slice thickness of the tomographic image to be searched and the slice thickness associated with the similar case information, especially for the slice thickness. Although it is preferable, it may be a criterion such that the difference between the thicknesses of the two slices is the minimum or not more than a predetermined threshold. As for the feature amount, the difference between the feature amount of the tomographic image to be searched and the feature amount associated with the similar case information is the minimum, or the difference is equal to or less than a predetermined threshold. A criterion such as being within a predetermined number from a small one can be considered. Therefore, a plurality of similar case information may be acquired by the search.

  As a processing flow at the time of search, first, similar case information associated with a slice thickness whose degree of match with the slice thickness of the tomographic image to be searched satisfies a predetermined criterion is acquired, and then the acquired similar case It is conceivable to acquire information associated with a feature amount whose degree of matching with a feature amount of a tomographic image to be searched satisfies a predetermined criterion. Further, after determining that there is similar case information associated with a slice thickness that satisfies the predetermined criterion, the feature amount of the tomographic image to be searched may be acquired. On the other hand, if similar case information with a slice thickness that satisfies the predetermined criteria for the degree of matching with the slice thickness of the search target image is not registered in the similar case database, the tomographic image of the search target is reconstructed. The sliced image is sliced from the 3D medical image so that the degree of matching with the slice thickness of the similar case information registered in the similar case database satisfies the predetermined criteria and includes the slice position of the search target image. A tomographic image representing a cross section may be reconstructed, and the reconstructed image may be searched again as a new search target image.

It is also conceivable to add means for outputting similar case information acquired by searching to the similar case searching apparatus of the present invention. Specifically, output by display on a display screen, output by printing, output to a storage medium of an apparatus as electronic data, and the like can be considered. Further, the output target may be all or a part of the similar case information, and if similar case information satisfying a predetermined criterion is not acquired by the search, this is indicated. It may be information.

  Note that the similar case retrieval apparatus of the present invention may be further provided with the similar case database registration apparatus of the present invention.

  In the similar case search of the present invention, the slice thickness and the feature amount of the tomographic image to be searched are referred to by referring to the similar case database in which the slice thickness, feature amount, and similar case information of a plurality of tomographic images having different slice thicknesses are associated. Similar case information associated with slice thicknesses and feature quantities whose degree of coincidence satisfies a predetermined criterion is acquired, so that similarity of content features of an image based on feature quantities is determined while considering slice thickness of the image And the accuracy of the search for similar cases is improved. In addition, the feature amount of the tomographic image is registered in the similar case database without registering the tomographic image itself in order to determine the similarity with the image to be searched, so the data capacity to be registered in the database is reduced. At the same time, it is not necessary to perform a process of calculating the feature amount from the tomographic image in the database every time the search is performed, and the processing load on the system is reduced.

  In addition, after acquiring similar case information associated with a slice thickness whose degree of match with the slice thickness of the tomographic image to be searched satisfies a predetermined criterion, the tomographic image to be searched is acquired from the acquired similar case information If the feature amount is related to a feature amount satisfying a predetermined criterion, the feature amount is determined for similar case information whose slice thickness does not satisfy the predetermined criterion. Since the slice thickness is different, similar case information associated with images having different image characteristics is not acquired and the search accuracy is improved. Furthermore, if it is determined that there is similar case information associated with a slice thickness that satisfies the predetermined criterion, the slice thickness satisfies the predetermined criterion if the feature amount of the tomographic image to be searched is acquired. When there is no similar case information, the feature amount is not acquired from the tomographic image to be searched, so that the processing load on the system is reduced.

  In addition, when registering in the similar case database, a plurality of tomographic images having different slice thicknesses are reconstructed from the three-dimensional medical image of one subject, and the interpretation / diagnosis support information of the subject is reconstructed. If the slice thickness and the feature quantity of each configured tomographic image are registered in association with each other, one interpretation / diagnosis support information can be efficiently registered, which contributes to effective use of similar case information. Further, when searching, the possibility of hitting the similar case information is increased, which contributes to improvement of search accuracy.

  Furthermore, if the interpretation / diagnosis support information includes the tomographic image itself associated with the information, the tomogram obtained as the interpretation / diagnosis support information at the time of interpretation of the search target image or diagnosis based on the search target image. It becomes possible to refer to images, which contributes to comparative interpretation of both images. Here, the interpretation / diagnosis support information includes a plurality of tomographic images associated with the information, and associates the plurality of tomographic images with the slice thickness of each image, and at the time of retrieval, the slice thickness of the retrieval target image. If a tomographic image whose degree of coincidence with a predetermined criterion is acquired as interpretation / diagnosis support information, a difference in slice thickness between the search target image and the tomographic image acquired as interpretation / diagnosis support information Therefore, comparative interpretation with higher accuracy is possible, which contributes to improvement of interpretation / diagnosis accuracy.

  Further, if the feature amount in the similar case information is set as the feature amount of the region of interest in the image and the search is performed, the degree of matching with the feature amount of the region of interest in the search target image is determined. By determining the similarity of the content features in the important part in the image, it is possible to search for a more appropriate similar case, which contributes to improvement in interpretation / diagnosis accuracy. At that time, if the region of interest in the search image is automatically detected, it contributes to a reduction in the work burden of the interpreter / diagnostic person and a reduction in oversight of a lesion in the image.

  As a result of the search, if similar case information having a slice thickness whose degree of matching with the slice thickness of the image to be searched satisfies a predetermined criterion is not registered in the similar case database, A slice obtained by slicing a 3D medical image to include a slice position of an image to be searched with a slice thickness that satisfies a predetermined standard with a matching degree with a slice thickness of similar case information registered in the similar case database If the tomographic image representing the image is reconstructed and the search is performed again based on the reconstructed image, the case where there is no corresponding similar case is reduced, the search accuracy is improved, and the interpretation / diagnosis accuracy is improved. Also contribute to.

  In addition, when registering in the similar case database, interpretation / diagnosis information stored in other medical information systems such as RIS, PACS, and HIS is acquired to generate interpretation / diagnosis support information. This makes it possible to save the trouble of manually generating the interpretation / diagnosis support information from scratch, and enhances the cooperation with other systems in the medical institution, thereby effectively utilizing the information resources. Furthermore, when the interpretation / diagnosis support information is link information for interpretation / diagnosis information or image data, it is possible to save waste of storage capacity by storing information of the same contents in a plurality of locations, and Data content inconsistency that occurs when only one of the information stored in the location is updated can be avoided, so the reliability of the information is improved, and data for avoiding such inconsistency Since this synchronization processing is also unnecessary, the processing load on the system is reduced. In particular, since information on electronic medical records, which is an example of diagnostic information, has the property of being updated as needed, the effect of avoiding this data content inconsistency problem is high.

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

  FIG. 1 shows a schematic configuration of a medical information system in which the similar case retrieval apparatus according to the first embodiment of the present invention is introduced. As shown in the figure, this system includes a medical image photographing device (modality) 1, an image quality check workstation (QA-WS) 2, an image diagnostic medical workstation 3 (3 a, 3 b), and an image information management server 4. The image information database 5, the interpretation report server 6, the interpretation report database 7, the similar case search server 8, and the similar case database 9 are connected to each other via a network 19 so as to communicate with each other. Each device except for various databases is controlled by a program installed from a recording medium such as a CD-ROM. The program may be installed after being downloaded from a server connected via a network such as the Internet.

  The modality 1 captures a region to be examined of a subject, generates image data of an image representing the region three-dimensionally, and adds incidental information defined by the DICOM standard to the image data. A device for outputting as image information is included. Specific examples include CT (Computed Tomography), MR (Magnetic Resonance), PET (Positron Emission Tomography), and an ultrasonic imaging apparatus. Hereinafter, a set of image data representing a subject and accompanying information of the image data is referred to as “image information”. That is, the “image information” includes text information related to the image.

  The QA-WS 2 includes a general-purpose processing device (computer), one or two high-definition displays, and input devices such as a keyboard and a mouse. The processing apparatus incorporates software for supporting the work of the laboratory technician. The QA-WS 2 receives image information compliant with DICOM from the modality 1 and displays the image data included in the received image information and the contents of the incidental information on the screen by a function realized by executing the software program. Urge the laboratory technician to confirm. Then, the image information confirmed by the inspection engineer is transferred to the image information management server 4 via the network 19 and the registration of the image information in the image information database 5 is requested.

  The image diagnostic medical workstation 3 is an apparatus used by an image diagnostician for interpretation of an image and generation of an interpretation report, and includes a processing device, one or two high-definition displays, and input devices such as a keyboard and a mouse. The In this apparatus, an image browsing request to the image information management server 4, display of an image received from the image information management server 4, automatic detection / highlighting of a part that appears to be a lesion in the image, support for creating an interpretation report, interpretation report Interpretation report registration request and browsing request to server 6, display of interpretation report received from interpretation report server 6, similar case information browsing request to similar case search server 8, and similar case information received from similar case search server 8 And a request for registration of similar case information to the similar case search server.

  The image information management server 4 is a general-purpose computer having a relatively high processing capability and incorporating a software program that provides a function of a database management system (DBMS). The image information management server 4 includes a large-capacity storage in which the image information database 5 is configured. This storage may be a large-capacity hard disk device connected to the image information management server 4 via a data bus, or connected to a NAS (Network Attached Storage) or SAN (Storage Area Network) connected to the network 19. It may be a disc device.

  In the image information database 5, image data representing a subject image and incidental information are registered. The accompanying information includes, for example, an image ID for identifying individual images, a patient ID for identifying a subject, an examination ID for identifying examinations, a unique ID (UID) assigned to each piece of image information, Examination date, examination time when image information was generated, type of modality used in examination to obtain the image information, patient information such as patient name, age, gender, examination site (imaging site), imaging conditions ( (Contrast agent used / used dye, radionuclide, radiation dose, etc.) may include information such as series number or collection number when multiple images are acquired in one examination. Image information can be managed as, for example, XML or SGML data.

  When the image information management server 4 receives a registration request for image information from the QA-WS 2, the image information management server 4 arranges the image information in a database format and registers it in the image information database 5.

  Further, when the image management server 4 receives a browsing request from the diagnostic imaging medical workstation 3 via the network 19, the image management server 4 searches the image information registered in the image information database 5 and uses the extracted image information as a request source. To the image diagnostic medical workstation 3.

  When an operation requesting browsing of an image to be interpreted is performed by a user such as an image diagnostician, the diagnostic imaging medical workstation 3 transmits a browsing request to the image management server 8 and acquires image information necessary for interpretation. To do. Then, the image information is displayed on the monitor screen, and an automatic lesion detection process is executed in response to a request from the image diagnostician.

  Furthermore, the image diagnostic medical workstation 3 displays a report creation screen for supporting the creation of an interpretation report on the monitor, and text indicating the contents of the diagnosis (findings, etc.) based on the interpretation is input from the radiologist. In some cases, an interpretation report is generated in which the input information and an image to be interpreted (hereinafter referred to as an interpretation target image) are recorded. When there are a plurality of interpretation target images, a representative image (hereinafter referred to as a representative image) having the greatest influence on the diagnosis is recorded in the interpretation report. The diagnostic imaging medical workstation 3 transfers the generated interpretation report to the interpretation report server 6 via the network 19 and requests registration of the interpretation report in the interpretation report database 7.

  The interpretation report server 6 incorporates a software program that provides a function of a database management system (DataBase Management System: DBMS) in a general-purpose computer having a relatively high processing capacity. When the registration request is received, the interpretation report is prepared in a database format and registered in the interpretation report database 7.

  The interpretation report database 7 includes, for example, an image ID for identifying an image to be interpreted or a representative image, an interpreter ID for identifying an image diagnostician who has performed interpretation, position information of a region of interest, findings, and certainty of findings. Such information is registered. In addition, the examination number and patient number acquired by referring to the supplementary information of the image information at the time of image interpretation, and the image data itself of the image to be interpreted or the representative image may also be included. The image data may be a copy of the image data registered in the image information database 5 as it is, or may be reduced image data having a smaller number of pixels (thinned out) than the image data registered in the image information database 5. Good. The image data may be link information indicating a storage location (folder or the like) of the image data in the image information database 5 and a file name. Further, the position information of the region of interest may be registered in the image information database 5 as supplementary information of the image data instead of the interpretation report database 7. The interpretation report can be managed as, for example, XML or SGML data.

  When the image interpretation report server 6 receives a browsing request from the image diagnostic medical workstation 3 via the network 19, the image interpretation report server 6 searches the image interpretation report registered in the image interpretation report database 7, and uses the extracted image interpretation report as the image diagnosis of the request source. Transmit to the medical workstation 3.

  The similar case search server 8 incorporates a software program that provides a function of a database management system (DBMS) in a general-purpose computer having a relatively high processing capability. In the similar case database 9, images read in the past and image interpretation results of the images are registered as similar case information. The similar case retrieval apparatus according to the embodiment of the present invention is implemented as a client-server type system in which a similar case retrieval server 8 having an image diagnostic medical workstation 3 and a similar case database 9 is connected by a network 19. Details will be described later.

  The network 19 is a local area network that connects various apparatuses in the hospital. However, when the diagnostic imaging medical workstation 3 is also installed in another hospital or clinic, the network 19 may be configured by connecting local area networks of each hospital via the Internet or a dedicated line. In any case, it is desirable that the network 9 can realize high-speed transfer of image information such as an optical network.

Hereinafter, taking as an example a case where a tomographic image obtained by reconstructing a three-dimensional medical image of a human chest acquired by CT with a predetermined slice thickness is used as an interpretation target, the first embodiment of the present invention and Details of the similar case retrieval apparatus will be described. FIG. 2 is a block diagram schematically showing the configuration of the similar case retrieval apparatus. As shown in the figure, this apparatus uses interpretation support information that supports interpretation of an image to be interpreted, that is, an image that has been interpreted in the past and an interpretation result of the image, the slice thickness of the image and the interest in the image. A similar case database 9 registered as similar case information in association with the feature quantity of the region, and a slice thickness for acquiring the slice thickness T ₁ of the image to be interpreted I ₁ acquired from the image information database 5 by the image acquisition unit 21 an acquiring unit 22, an ROI setting unit 23 that sets a region of interest R ₁ in the image to be interpreted I _1, and characteristic amount obtaining section 24 for obtaining the feature amount C ₁ in the set region of interest R _1, slice thickness based on the T ₁ and the feature amount C _1, interpretation support information a ₁ to support the interpretation of the interpretation target images I _1, i.e., representative of the image interpretation and diagnosis results of the image and the image similar to the image to be interpreted I ₁ It is comprised from the similar case search part 25 which acquires information from the similar case database 9. FIG.

Similar case retrieval unit 25, whether the similar case information the degree of agreement with the slice thickness T ₁ of the image to be interpreted I ₁ is associated with a slice thickness satisfies a predetermined criterion (and T ₁ ') is present Similarity that the slice thickness search unit 25a to search, the slice thickness is T ₁ ′, and the feature amount matches the feature amount C ₁ of the region of interest R ₁ of the image to be interpreted I ₁ satisfies a predetermined criterion composed of the feature retrieval unit 25b, an interpretation support information acquisition unit 25c for acquiring interpretation support information a ₁ on the basis of the similar case information P ₁ to find whether the case information (and P ₁₎ is present .

  FIG. 3 schematically shows the configuration of the similar case database 9. As shown in the figure, the similar case database 9 includes a slice thickness table 26 in which types of slice thicknesses registered in the database are registered, file names of images that have been interpreted in the past, and interpretation / interpretation of the images. A slice thickness feature quantity table 27 (27a, 27b, 27c) registered in association with the slice thickness of the image and the feature quantity of the region of interest in the image in the past. Diagnostic information in which an image file group 28 storing image files of images that have been interpreted, and diagnostic information files that include interpretation and diagnosis results such as findings of the images and final confirmed diagnosis information based on the images It consists of a file group 29. In this embodiment, as specific examples of the slice thickness, feature amounts, image file names, and diagnostic information file names relating to images having a slice thickness of 1 mm, 2 mm, and 5 mm are registered in the tables 27a, 27b, and 27c, respectively. To do. Therefore, three records of 1 mm, 2 mm, and 5 mm are registered in the slice thickness table 26. The image file names and diagnostic information file names in the slice thickness feature quantity tables 27a, 27b, and 27c function as pointers indicating the storage locations of the files in the image file group 28 and the diagnostic information file group 29, respectively. Specific examples of the feature amount include the position and size (diameter) of an abnormal shadow in the region of interest, a luminance histogram, and AAM (TF Cootes, GJ. Edwards, CJ Taylor, “Active Appearance Models”, In Proc. 5th European Conference on Computer Vision, Springer, Germany Springer), 1998, vol.2, pp484-498), and information on the shape and texture of the lesion in the region of interest. Note that the feature amount may be a combination of a plurality of types.

  Each processing unit has various functions such as the performance of the diagnostic imaging medical workstation 3, the similar case retrieval server 8 and the network 19, the number of installed diagnostic imaging medical workstations 3, the access frequency to the similar case retrieval server 8, and the response time. Depending on the performance requirements, it will be distributed as appropriate. For example, the image acquisition unit 21, the slice thickness acquisition unit 22, the region of interest setting unit 23, and the feature amount acquisition unit 24 are mounted on the diagnostic imaging medical workstation 3, and only the similar case search unit 25 is mounted on the similar image search server. The image acquisition unit 21, the slice thickness acquisition unit 22, and the region of interest setting unit 23 are mounted on the diagnostic imaging medical workstation 3, and the feature amount acquisition unit 24 and the similar case search unit 25 are mounted on the similar case search server 8. Can be considered.

  Next, the details of the similar case search process and the flow of the process will be described using the flowchart of FIG.

First, the image diagnostician performs user authentication by means of the image diagnostic medical workstation 3 using biometric information such as a user ID / password and fingerprint for accessing the medical information system. When the user authentication is successful, an examination (interpretation) target image list based on the image diagnosis order issued by the ordering system is displayed on the display. The image diagnostician selects an image I ₁ to be interpreted from the examination image list using an input device such as a mouse. The image acquisition unit 21 makes a browsing request to the image information management server 4 using the image ID of the selected image I ₁ as a search key, and the image information management server 4 that has received this request searches the image information database 5. The image file of the image to be interpreted I ₁ is acquired, and the image file is transmitted to the diagnostic imaging workstation 3 that transmitted the request. The diagnostic imaging medical workstation 3 that has received the image file displays the image to be interpreted I ₁ on the display based on the image data of the image file (# 1).

  As shown in the screen example of FIG. 5, when the image diagnostician performs a right click of the mouse, a menu listing various functions for supporting image interpretation is displayed. Here, when the image diagnostician selects “similar case search” using a mouse or the like and performs a left click, the image diagnostic medical workstation 3 detects a left click event for the “similar case search” item in the menu. (# 2; YES), the subsequent similar case search process is performed. On the other hand, when any other item is selected, the selected process is performed (# 2; NO).

When “similar case search” is selected (# 2; YES), the slice thickness acquisition unit 22 reads the DICOM information of the displayed interpretation target image I ₁ and uses the interpretation target image I ₁ as the original CT image data. reconstructed from then obtains the slice thickness T ₁ of the time generated by (# 3). Here, as a specific example, slice thickness T ₁ of the image to be interpreted I ₁ is assumed to be 2 mm.

Slice thickness searching unit 25a, the image to be interpreted I ₁ of the slice thickness T ₁ of the search key, with reference to the slice thickness table 26 similar case database 9, the degree of match between the slice thickness T ₁ is satisfies a predetermined criterion It is searched whether there is similar case information associated with the slice thickness T ₁ ′ (# 4). Here, when similar case information that matches the slice thickness of the image to be interpreted is registered, it is determined that a predetermined criterion is satisfied (that is, T ₁ = T ₁ ′). In the above specific example, the slice thickness T ₁ of the image to be interpreted I ₁ is 2 mm, and a record of the slice thickness “2 mm” exists in the slice thickness table 26, so it is determined that the predetermined criterion is satisfied (# 5 ; YES). If the slice thickness T ₁ does not satisfy the predetermined standard, that is, if there is no slice thickness record that matches the slice thickness T _{1 in the} slice thickness table 26 (# 5; NO), “ A message indicating “No similar case” is displayed on the display of the diagnostic imaging workstation 3 (# 11).

When the slice thickness T ₁ satisfies a predetermined criterion (# 5; YES), the image diagnostician responds to a message prompting the user to select a region of interest displayed on the display of the image diagnostic workstation 3. The region of interest R ₁ is selected using the above. Thereby, the region-of-interest setting unit 23 specifies the position information of the region of interest R ₁ (# 6).

The feature amount acquisition unit 24 calculates the feature amount C ₁ of the region of interest R ₁ by a predetermined method (# 7). Here, the predetermined method means the same method as the feature amount calculation method registered in the slice thickness feature amount table 27c.

The feature quantity search unit 25b refers to the slice thickness feature quantity table 27 having a slice thickness of T ₁ ′, and the degree of coincidence with the feature quantity C ₁ of the region of interest R ₁ of the interpretation target image I ₁ is predetermined. It is searched whether or not similar case information P ₁ which is a feature quantity satisfying the criterion exists (# 8). Here, when the difference between the feature amount registered in the slice thickness feature amount table 27 and the feature amount C ₁ obtained from the image to be interpreted I ₁ is minimum and the difference is equal to or less than a predetermined threshold. It shall be determined that a predetermined standard is satisfied. In the above specific example, using the calculated feature quantity C ₁ as a search key, the feature quantity table 27b with slice thickness = 2 mm is referred to, and a similar case information record P ₁ satisfying this criterion is acquired. If there is no record satisfying this criterion in the slice thickness feature amount table 27 (# 9; NO), a message indicating “no similar case” is displayed on the display of the diagnostic imaging workstation 3 (# 11).

On the other hand, when the record P ₁ satisfying this criterion exists in the slice thickness feature quantity table 27 (# 9; YES), the image interpretation support information acquisition unit 25c sets the image file name and diagnosis information file name of the record P _1. reading the file, each image file group 28, obtains from the diagnostic information file group 29, and transfers these files, namely the interpretation support information a ₁ to the image diagnosis medical workstation 3. In the diagnostic imaging medical workstation 3, the received image file and diagnostic information file are read, and the image and diagnostic information are displayed on the display of the diagnostic imaging medical workstation 3 (# 11).

Thus, in the similar case retrieval apparatus according to the first embodiment of the present invention, similar case search unit 25 refers to the similar case database 9, interpretation target images I ₁ slice thickness T ₁ and ROI R ₁ Since the interpretation support information A ₁ is acquired based on the similar case information associated with the slice thickness and the feature quantity, the degree of coincidence with the feature quantity C ₁ satisfies the predetermined criterion, the slice thickness of the image to be interpreted is considered. However, it becomes possible to determine the similarity of the content features of the image based on the feature amount, and the accuracy of searching for similar cases is improved. In addition, the feature amount of the tomographic image is registered in the similar case database 9 without registering the tomographic image itself in order to determine the similarity to the image to be interpreted, so the data capacity to be registered in the database is reduced. In addition to reducing the amount of processing, it is not necessary to perform processing for calculating feature amounts from tomographic images in the database each time retrieval is performed, and the processing load on the system is reduced.

Also, slice thickness searching portion 25a of the similar case search unit 25 refers to the slice thickness table 26 similar case database 9, the degree of match between the slice thickness T ₁ of the image to be interpreted I ₁ satisfies a predetermined reference slice the similar case information thick determine whether registered in the similar case database 9, only if there is a registration. Thus processing after setting the ROI R ₁ by the ROI setting unit 23 is carried out, Similar case information associated with an image that approximates the feature quantity of the region of interest in the image is not erroneously acquired even though the image characteristics are different due to different slice thicknesses. , Search accuracy is improved. Further, only when there is the above registration, processing such as setting of the region of interest R ₁ by the region of interest setting unit 23 and calculation of the feature amount C ₁ by the feature amount acquisition unit 24 is performed. If there is no appropriate similar case information for the image to be interpreted I ₁ because it does not satisfy the condition, the setting of the region of interest R ₁ and the calculation of the feature amount C ₁ , which result in unnecessary processing, are not performed, and the image diagnostic doctor Work load and system processing load are reduced. Furthermore, since the slice thickness table 26 in which the types of slice thickness registered in the similar case database 9 are indexed is provided, the slice thickness search unit 25a needs to access all the records registered in the similar case database 9. The system processing time is shortened and the processing load is reduced.

Further, in the similar case database 9, as interpretation support information A ₁ because it contains the tomographic image itself, during the interpretation of the image to be interpreted, to be capable of referring to the tomographic image obtained as interpretation support information A ₁ It contributes to comparative interpretation of both images.

  Furthermore, since the feature quantity handled in this similar case search is the feature quantity of the region of interest in the image, it is possible to search for a more appropriate similar case by determining the similarity of the content features in the important part of the image. Can contribute to the improvement of interpretation accuracy.

In the above embodiments, the imaging diagnostician is but a region of interest R ₁ in interpretation target images I ₁ had to be specified manually, automatically detects a lesion candidate region in the image I ₁ as a region of interest R ₁ You may do it. In this case, when the image to be interpreted I ₁ is displayed on the display in step # 1 in FIG. 4, the detected region of interest R ₁ may be displayed in a form that can be identified by marking or the like. For example, FIG. Add the item “Automatic detection result display” to the menu that is displayed by right-clicking the mouse as shown in the above, and if the image diagnostician left-click and select this item, the automatic detection process and the display of the detection result are displayed. You may make it perform. Specific examples of the automatic detection method may be known methods described in JP-A-2005-198887 and JP-A-2005-246032. Further, when there are a plurality of lesion candidate regions as a result of automatic detection, the image diagnostician may select _{one of them} as the region of interest R ₁ or correct the lesion candidate region by automatic detection. A user interface may be provided. Thus, if to detect a region of interest R ₁ in the image to be interpreted I ₁ automatically, contribute reduce or doctor work load, to reduce the oversight of a lesion in the image. Note that it is conceivable to combine the automatic detection processing even when the region of interest R ₁ is manually specified as in the above embodiment. For example, it is conceivable that the image diagnostician designates a rectangular region including a lesion candidate and automatically detects the lesion candidate region as the region of interest R _{1 from the} region.

In the above embodiment, the slice thickness search unit 25 a is based on the fact that the slice thickness of the image to be interpreted I ₁ matches the slice thickness registered in the slice thickness table 26. Among the slice thicknesses registered in ( ₁₎ , the _one having the smallest difference from the slice thickness of the image to be interpreted I ₁ may satisfy a predetermined criterion.

Similarly, in the feature quantity search unit 25b, the feature quantity in the similar case information and the interpretation target among the similar case information associated with the slice thickness searched by the slice thickness search unit 25a in the slice thickness-specific feature quantity table 27. A predetermined number of similar case information may satisfy a predetermined criterion from those having a small difference from the feature amount C ₁ of the region of interest R ₁ of the image I ₁ , and all of the above differences are equal to or smaller than a predetermined threshold value May satisfy a predetermined standard. In addition, when the feature quantity is composed of multiple types of scales, a comprehensive feature quantity is obtained by weighting the value of each scale, and the judgment based on a predetermined standard is made for the total feature quantity. Alternatively, it may be possible to use a value based on each scale as it is to obtain a multidimensional feature value, and to determine similarity / dissimilarity according to the distance in the multidimensional space.

Furthermore, the slice thickness can be handled as a measure of the feature amount. For example, it is conceivable that the total feature amount is obtained by making the weight for the slice thickness larger than other scales. In this case, the similar case search unit 25 does not include the slice thickness search unit 25a. In the similar case database 9, instead of the slice thickness table 26 and the slice thickness feature amount table 27, the comprehensive feature amount, the image file name, and the diagnosis The feature value search unit 25b uses a table in which information file names are associated with each other, and the feature value search unit 25b compares the total feature value of the image to be interpreted I ₁ with the total feature value registered in this table, and the similarity satisfying a predetermined criterion. Case information may be acquired.

In the above-described embodiment, if similar case information associated with a slice thickness that satisfies a predetermined criterion is not registered as a result of searching the similar case database 9, a message indicating “no similar case” is displayed for diagnostic imaging medical use. The slice thickness that was displayed on the display of the workstation 3 (FIG. 4: # 9; NO, # 11), but satisfies the predetermined criteria from the three-dimensional medical image that is the reconstruction source image of the image to be interpreted These images may be reconstructed and the similar case database 9 may be searched. FIG. 6 is a block diagram schematically showing the configuration of such a similar case retrieval apparatus. This device is acquired by the similar case search device (FIG. 2) of the first embodiment, and the three-dimensional image acquisition unit 30 that acquires the three-dimensional medical image I ₀ that is the reconstruction source of the image to be interpreted I _1. The image reconstruction unit 31 for reconstructing a slice image having the slice thickness T ₂ registered in the slice thickness table 26 from the three-dimensional medical image I ₀ is further added.

FIG. 7 is a flowchart showing the flow of processing by the similar case search apparatus of FIG. As shown in the figure, in this similar case search apparatus, after selecting / displaying the image I ₁ to be interpreted (# 1) in the same manner as in the above-described embodiment, the image diagnostician performs “similar case search”. When "is selected (# 2), acquired slice thickness acquisition unit 22 the interpretation target images I ₁ (# 3), slice thickness retrieval section 25a is, the search key slice thickness T ₁ of the image to be interpreted I ₁ The slice thickness table 26 of the similar case database 9 is referred to (# 4).

Here, when a similar case information the degree of match between the slice thickness T ₁ is associated with a slice thickness T ₁ 'satisfying a predetermined criterion does not exist (# 5; NO), 3-dimensional image acquisition unit 30 Obtains the slice thickness (T ₂ ) closest to the slice thickness T ₁ of the interpretation target image I ₁ out of the slice thicknesses registered in the slice thickness table 26, and further obtains this from the DICOM information of the interpretation target image I _1. Information for identifying the three-dimensional medical image (CT image) data I ₀ as the reconstruction source of the image is acquired, the image information database 5 is searched based on this information, and the three-dimensional medical image data I ₀ as the reconstruction source is obtained. Get (# 12). Next, the image reconstruction unit 31 reconstructs the image I ₂ having the slice thickness T ₂ from the three-dimensional medical image data I ₀ (# 13).

Subsequently, for the image I ₂ , as in the above embodiment, the region-of-interest setting unit 23 receives the selection of the region of interest R ₂ and specifies its position information (# 6), and the feature amount acquisition unit 24 interests. calculating a feature amount C ₂ in the region R ₂ (# 7), the feature retrieval unit 25b, slice thickness of the slice Atsubetsu feature quantity table 27 with reference to those of T _2, is registered in the table 27 of similar case information are to obtain what degree of match between the feature quantity C2 of features and the reconstructed image I ₂ ROI R ₂ in the similar case information satisfies a predetermined criterion (P ₂₎ (# 8, # 9), similar case information acquisition unit 25c acquires and displays the diagnostic information file and the image file associated with the similar case information P ₂ (# 10).

As described above, when the similar case information associated with the slice thickness satisfying the predetermined criterion with respect to the slice thickness of the image to be interpreted I ₁ is not registered in the similar case database 9, the three-dimensional image acquisition unit 30 The three-dimensional medical image I ₀ that is the reconstruction source of the interpretation target image I ₁ is acquired, and the image reconstruction unit 31 uses the similar case information registered in the similar case database 9 from the acquired three-dimensional medical image I _0. reconstructs the slice image I ₂ of the associated slice thickness T ₂ and, if the similar case information P ₂ of the image I _2, which is re-configured to acquire, cases that there is no similar case with respect to the image to be interpreted The search accuracy is improved, and the interpretation / diagnosis accuracy is improved.

  Next, a similar case database registration apparatus according to the second embodiment of the present invention will be described. This apparatus is for registering similar case information in the similar case database 9 accessed during the search by the similar case search apparatus.

FIG. 8 is a block diagram schematically showing the configuration of the similar case database registration apparatus. As shown, the apparatus includes a slice image acquisition unit 33 for acquiring a plurality of slice images S _n obtained by CT, interpretation result of the reconstructed image from the slice images S _n or slice images S _n the interpretation report acquiring unit 32 that acquires an interpretation report M ₀ representing from the image interpretation report database 7, from a plurality of slice images S _n, a key slice selection section 34 for accepting a selection of a desired slice images (key slices) S _k , a region of interest setting unit 23 that sets a region of interest R _k in key slice image S _k, reconstructs the key slice image S _k, a plurality of tomographic images I _n having different slice thickness and a vicinity of the slice image an image reconstruction unit 31, in the tomographic image I _n, a region of interest extraction unit 35 for extracting a region of interest R _n corresponding to the region of interest R _k, region of interest R _k, wherein the amount of R _n C _k, a feature amount acquisition unit 24 that acquires C _n, key slice image S _k, a similar case information registering unit 36 for registering the image interpretation report M ₀ similar case database 9 in association for each of the tomographic images I _n Consists of

  Each processing unit, for example, the slice image acquisition unit 33 is the same as the above-described image browsing processing performed by the image diagnostic medical workstation 3, the image information management server 4, and the image information database 5 connected to the network 19. The interpretation report acquisition unit 32 can be realized in the same manner as the above-described interpretation report browsing process by the diagnostic imaging workstation 3 and the interpretation report server 6 connected to the network 19 and the interpretation report database 7. The key slice image selection unit 34, the region of interest setting unit 23, the image reconstruction unit 31, the region of interest extraction unit 35, and the feature amount acquisition unit 24 are mounted on the image diagnostic medical workstation 3, and the similar case information registration unit 36 Although it may be implemented in the similar case retrieval server 8, the performance of the diagnostic imaging medical workstation 3, the similar case retrieval server 8 and the network 19, the number of installed diagnostic imaging medical workstations 3, and access to the similar case retrieval server 8 Depending on various performance requirements such as frequency and response time, it can be appropriately distributed and implemented.

  Next, details of the similar case information registration process and the flow of the process will be described with reference to the flowchart of FIG.

First, the imaging diagnostician carries out viewing request to the image information management server 4, image diagnosis medical workstation 3 operations and search key image ID of the slice images S _n of the registration target, receiving the requested image information management server 4 performs a search against the image information database 5, and acquires the image file of the slice images S _n, the image diagnostic medical workstation 3 that has sent the request to send the image file. The diagnostic imaging medical workstation 3 that has received the image file displays the slice image _Sn on the display based on the image data of the image file (# 21). As a display method, various methods such as a method of sequentially displaying each slice image one by one, a method of simultaneously displaying a plurality of slice images in one screen, a method of combining these, and the like can be considered.

Further, the imaging diagnostician, browse to the interpretation report server 6 to the image ID of the image diagnosis medical workstation 3 by operating the registration target slice images S _n or slice image S _n reconstructed image from the retrieval key Make a request. Here, the setting of the search key is set and a user interface image diagnostician enters a search key (image ID), as automatically search key image ID of the slice images S _n displayed on the display user It is conceivable to use an interface. Interpretation report server 6 that has received the request by performing a search for image interpretation report database 7, acquires an interpretation report M ₀ of the reconstructed image from the slice images S _n or slice images S _n, image diagnosis that has transmitted the request The interpretation report M ₀ is transmitted to the medical workstation 3. Imaging medical workstation 3 receives the image interpretation report M ₀ displays the image interpretation report M ₀ on the display (# 22). Note that the order of this step and step # 21 may be reversed.

Then, the imaging diagnostician, based on the slice image S _n and interpretation report M _0, which is displayed on the display, to select the most appropriate image as the similar case information as a key slice image S _k (# 23). This operation is realized, for example, by selecting “Select as Key Slice” from the menu displayed by right-clicking the image to be selected and left-clicking.

Furthermore, and selects the region of interest for the key slice image S _k of the imaging diagnostician is selected, the ROI setting unit 23 specifies the position information of the selected region of interest R _k (# 24).

At the same time, the image reconstruction unit 31 reconstructs a key slice image S _k, a plurality of tomographic images I _n having different slice thickness and a vicinity of the slice image (# 25). Here, the slice image to be reconstructed original tomographic image I _n is always contains key slice image S _k. For example, if the slice thickness of each slice image S _n was 1 mm, the image reconstruction unit 31, a key slice image S _k and the image of the slice thickness 3mm by reconstituting one by one before and after the slice images that Are further reconstructed by reconstructing the key slice image _Sk and two slice images before and after the key slice image Sk. When the configuration of the similar case database 9 is as shown in FIG. 3, it is preferable that the slice thickness at the time of reconstruction is already registered in the slice thickness table 26. A known method can be used as a method for reconstructing an image having each slice thickness.

ROI extraction unit 35 based on the position information of the region of interest R _k, each of the tomographic images I _n the reconstructed, extracts a region corresponding to the region of interest R _k as a region of interest R _n (# 26).

The feature quantity acquisition unit 24 calculates the feature quantities C _k and C _n of the regions of interest R _k and R _n (# 27).

Finally similar case information registration unit 36, as well as associating the key slice image S _k and interpretation report M _0, in each of the tomographic image I _n in association with the interpretation report M ₀ is registered in the similar case database 9 (# 28 ). Specifically, if the similar case database 9 having the configuration shown in FIG. 3, and save the image file of the key slice image S _k or the tomographic image I _n to be registered in the image file group 28, diagnostic radiology report M ₀ The feature amount C of the region of interest R _k or R _n of the image to be registered is stored in the information file group 29 and the slice thickness feature amount table 27 in which information of the same slice thickness as the registration target slice thickness is registered. and _k or C _n, to register and file name of the image, the file name of the image interpretation report M _0. It should be noted that the file names of the images and the interpretation report M ₀ are automatically determined by the similar case information registration unit 36 according to a predetermined naming rule in order to avoid duplication of file names in the image file group 28 and the diagnostic information file group 29, respectively. It is preferable to make a determination.

Thus, in a similar case database registration apparatus according to the second embodiment of the present invention, the image reconstruction unit 31, a key slice image S _k which is selected from a plurality of slice images S _n by the key slice selection section 34 a plurality of tomographic images I _n different from the slice images slice thickness in the vicinity reconstructed from a plurality of slice images S _n, is similar case information registration unit 36, directly or indirectly, in association with the slice image S _n against interpretation report M ₀ which has been registered in the image interpretation report database 7 Te, key slice image S _k and slice thickness and a region of interest R _k of the reconstructed each tomographic image I _n, the feature amount C _k, R _n , C _n are associated with each other and registered in the similar case database 9, so that one interpretation report M ₀ can be efficiently registered as similar case information. Contributes to effective use. Further, since the interpretation report M ₀ is associated with the feature amount at each slice thickness at the time of retrieval, there is a high possibility that the similar case will be hit accurately, which contributes to improvement of retrieval accuracy.

Further, in the similar case database 9, since different images slice thickness that key slice image S _k and the tomographic image I _n is included as interpretation / diagnosis support information, the time of search, the search target image The difference in slice thickness between images obtained from the similar case database 9 is reduced, and comparative interpretation with higher accuracy is possible, which contributes to improvement of interpretation / diagnosis accuracy.

In the above embodiment, as for each slice thickness that is registered in the similar case database 9, and registers the image of the same slice thickness and its slice thickness (key slice image S _k or the tomographic image I _n) However, the feature amount of the region of interest in the image of the slice thickness may be associated with each slice thickness, and the same image, for example, the key slice image _Sk may be associated and registered. This eliminates the need to register a plurality of images of the same subject, thereby reducing the database capacity.

A system that realizes a workflow that integrates the first and second embodiments is also conceivable. FIG. 10 is a block diagram schematically showing the configuration of such a similar case search / registration system. As shown in the figure, this system uses interpretation support information that supports interpretation of an image to be interpreted, that is, an image that has been interpreted in the past and an interpretation result of the image, the slice thickness of the image and the interest in the image. A similar case database 9 registered as similar case information in association with the feature quantity of the region, and a slice thickness for acquiring the slice thickness T ₁ of the image to be interpreted I ₁ acquired from the image information database 5 by the image acquisition unit 21 an acquiring unit 22, an ROI setting unit 23 that sets a region of interest R ₁ in the image to be interpreted I _1, and characteristic amount obtaining section 24 for obtaining the feature amount C _1, C _n region of interest R _1, R _n , based on the slice thickness T ₁ and the feature amount C _1, interpretation support information a ₁ to support the interpretation of the interpretation target images I _1, i.e., image interpretation and diagnosis results of the image and the image similar to the image to be interpreted I ₁ A similar case search unit 25 that acquires information representing the same from the similar case database 9, a three-dimensional image acquisition unit 30 that acquires a three-dimensional medical image I ₀ that is a reconstruction source of the interpretation target image I ₁ , and the acquired three-dimensional extracting an image reconstruction unit 31 slice thickness from the medical image I ₀ reconstructs a plurality of tomographic images I _n which is different from the T _1, in the tomographic image I _n, the region of interest R _n corresponding to the region of interest R ₁ The region-of-interest extraction unit 35, an interpretation report generation unit 37 that supports the creation of the interpretation report M ₀ by the image diagnostician and registers the created interpretation report M _{0 in} the interpretation report database 7, and the interpretation target image I ₁ , tomographic image I _n , slice thickness T ₁ , T _n of each image, feature amount C ₁ , C _{n of} region of interest R ₁ , R _n of each image, and interpretation report M ₀ are associated and registered in similar case database 9 Similar case information registration It consists of 36 Metropolitan.

  About each component, the same code | symbol is attached | subjected to the component similar to the above-mentioned embodiment, and it has the same function as the above-mentioned. The interpretation report generation unit 37 corresponds to the interpretation report creation support function and the interpretation report server 6 function in the image diagnostic medical workstation 3 described above.

  Next, a workflow in which similar case retrieval and registration realized by this system are integrated will be described using the flowchart of FIG.

First, the imaging diagnostician carries out the search of the similar case with respect to the image to be interpreted I _1, to obtain the interpretation support information A ₁ (# 31). The details of the processing performed here are the same as those in # 1 to # 11 in the flowchart of FIG.

Next, the image diagnostician reads the interpretation target image I ₁ while referring to the interpretation support information A ₁ displayed on the display of the diagnostic imaging workstation 3 and creates a report M _{0 of the} interpretation result. The created interpretation report M ₀ is registered in the interpretation report database 7 (# 32).

Here, the image diagnostician determines that the image to be interpreted I ₁ is useful as similar case information when interpreting other images, and operates the mouse on the screen displayed on the display as “similar case information”. When the “Register” menu is selected (# 33; YES), the following processing is automatically executed.

In this case, the three-dimensional image acquisition unit 30 acquires information for identifying the three-dimensional medical image (CT image) data I ₀ that is the reconstruction source of this image from the DICOM information of the image to be interpreted I ₁ , and based on this information Then, the image information database 5 is searched to obtain the reconstruction source three-dimensional medical image data I ₀ (# 34).

Then, the image reconstruction unit 31, slice thickness from acquired three-dimensional medical image I ₀ reconstructs a plurality of tomographic images I _n which is different from the T ₁ (# 35). Here, tomographic image I _n is preferably one that is reconstructed based on the area of three-dimensional medical image I in ₀ corresponding to the region of interest R ₁ in interpretation target image I _1.

Furthermore, the region of interest extraction unit 35, in the tomographic image I _n, extracts a region of interest R _n corresponding to the region of interest R ₁ (# 36), the feature amount acquisition unit 24, the feature amount of the region of interest R _n C _n is calculated (# 37).

Finally, the similar case information registration unit 36 interprets the interpretation target image I ₁ , the tomographic image I _n , the slice thicknesses T ₁ and T _n of each image, and the feature amounts C ₁ and C _n of the regions of interest R ₁ and R _n of each image. _n , The interpretation report M ₀ is associated and registered in the similar case database 9 (# 37).

  As described above, by appropriately combining the similar case retrieval apparatus of the first embodiment and the similar case registration apparatus of the second embodiment, when the image is interpreted, the acquisition of the interpretation target image is performed. An integrated workflow from search, interpretation, report creation, and registration of similar case information is realized, contributing to improvement in interpretation and diagnosis efficiency.

  Next, a similar case search / registration device according to a third embodiment of the present invention will be described. This apparatus is for enabling similar case retrieval in cooperation with data stored in other medical information systems such as RIS, PACS, and HIS.

  FIG. 12 shows a schematic configuration of a medical information system in which the apparatus of this embodiment is introduced. As shown in the figure, a medical image photographing device (modality) 1, an image quality check workstation (QA-WS) 2, an image diagnostic medical workstation 3 (3a, 3b), an image information management server 4, an image information database. 5, an interpretation report server 6, an interpretation report database 7, a similar case search server 8, a similar case database 9, a diagnostic medical workstation 10 (10a, 10b), an electronic medical record management server 11, and an electronic medical record database 12 are connected via a network 19. Connected to each other in a communicable state. Compared with the system configuration (FIG. 1) in the first and second embodiments, the present embodiment is such that the diagnostic medical workstation 10 (10a, 10b), the electronic medical record management server 11, and the electronic medical record database 12 are newly connected. It becomes the composition.

  The medical doctor workstation 10 is an apparatus used by the medical doctor to refer to the interpretation report and the image that is the subject of the interpretation, and includes input of a processing device, one or two high-definition displays, and a keyboard / mouse. Consists of equipment. This processing device includes a medical record information creation function for recording the history of examinations, examinations, and treatments (medications, etc.) of each patient as electronic data, and an order system for radiology by creating examination order information. The inspection order function to transmit to (not shown) is incorporated.

  The medical doctor workstation 10 creates new medical record information recording patient information, the purpose of medical examination, etc. at the first visit of the patient, transfers it to the electronic medical record management server 11 via the network 19, and an electronic medical record database of the medical record information. Request registration to 12. Upon receiving a medical record information registration request, the electronic medical record management server 11 arranges the medical record information in a database format and registers it in the electronic medical record database 12.

  When the examination is ordered by the examination order function during the examination, the examination medical workstation 10 adds the examination number and the examination date notified from the diagnostic imaging system to the medical chart information. Further, when the examination is completed and an interpretation report is created by the diagnostic imaging doctor, and when the diagnostic report is referred to by the diagnostic doctor, a part of the information recorded in the interpretation report (for example, findings information of the diagnostic imaging doctor) is obtained. , Add to the chart information.

  On the other hand, information entered by the examiner himself / herself is added to the chart information as needed. For example, in the case of a test that does not involve imaging, such as a blood test, finding information input by a doctor referring to the test result is added to the chart information. In addition, for the record of medication and other treatments, information input by the doctor is added to the chart information. The medical examination workstation 10 transfers the updated chart information or the difference information indicating only the changed part to the electronic chart management server 11 and requests the update of the corresponding chart information.

  When the electronic medical record management server 11 receives a request for updating the medical chart information, the electronic medical chart database 12 is searched, the corresponding medical chart information is called on the memory, the contents are updated on the memory, and the updated medical chart information is then stored. Write to storage. Thereby, the contents of the electronic medical record database 12 are updated.

  Examples of medical record information registered in the electronic medical record database 12 are managed as, for example, XML or SGML data. As described above, the patient information registered at the first visit, the reception history of the patient, the examination history, medication, etc. The medical calendar, and other information necessary for diagnosis and treatment, such as past illness, chronic illness, and allergies, are recorded. Further, reduced image data similar to the interpretation report and link information to the image information may be recorded as one item of the inspection history.

  FIG. 13 schematically shows the relationship and configuration of each database in the present embodiment. As shown in the figure, the similar case database 9 associates an image ID for identifying an image of a similar case, a patient ID for identifying a patient of the image, and a slice thickness and a feature amount used for determination of image similarity. Registered. The details of the registration process in the similar case database 9 will be described later. The image interpretation report database 7 includes an image ID for identifying an image that has been interpreted, an image interpreter ID for identifying an image interpreter (image diagnostician) of the image, position information of a region of interest (ROI) in the image, and findings of the image (Interpretation results) are registered in association with each other. The image information database 5 includes an image ID, an examination ID for identifying the examination for which the image was acquired, a slice position of the image, a slice thickness, an original image ID for identifying an image that is a reconstruction source of the image, and an image. Data etc. are related. The electronic medical record database 12 includes a patient information table 12a, a patient examination history table 12b, a patient examination history table 12c, an examination information table 12d, and a examination information table 12e. In the patient information table 12a, attributes of the patient are newly registered at the first visit of the patient. Specifically, the patient ID for identifying the patient, the date of birth of the patient, sex, and the like are associated with each other. ing. The patient examination history table 12b is registered at the time of examination reservation for a patient or when an examination is performed, and the history of examinations received by a patient by associating the patient ID with an examination ID for identifying the examination that the patient has undergone. Is managed. The patient examination history table 12c is registered when a patient is scheduled for examination or when the examination is performed, and a history of examinations received by the patient by associating the patient ID with the examination ID for identifying the examination received by the patient. Is managed. The examination information table 12d registers the contents of each examination. Specifically, the examination ID, examination date, examination type (CT imaging, blood examination, etc.), examination site (chest, abdomen, etc.), imaging. A representative image ID for identifying a representative one of the images is associated. In this table, for example, new information is registered at the time of examination reservation, but the representative image ID is set at the time of performing image diagnosis or at the time of definitive diagnosis by examination. The medical examination information table 12e registers the contents of each medical examination. Specifically, the medical examination ID, the date and time of the medical examination, the finalized diagnostic information indicating the diagnostic contents determined by the medical examination, and the like are associated with each other. In this table, for example, new information is registered at the time of a medical appointment, and definitive diagnostic information or the like is set at the time of medical examination. The registration / update timing of each database and table can be changed depending on the workflow in each medical institution.

  In addition, the image ID, examination ID, patient ID, and examination ID in each database are based on the same system, and these databases and tables are associated with each other by, for example, the patient ID in the patient information table 12a. Based on the patient examination history table 12b, the examination ID of the examination that the patient has received in the past is acquired, and based on the examination ID, chest CT imaging (examination that has been performed on the patient in the past from the examination information table 12d) The image ID of the representative image obtained by (type / part) can be acquired. Furthermore, based on the representative image ID, the image data of the representative image can be acquired from the image information database 5, and the findings for the image can be acquired from the interpretation report database 7.

In a similar case search processing in this embodiment, like the first embodiment described above, after obtaining the feature amount C ₁ in the image to be interpreted I ₁ of the slice thickness T ₁ and ROI R _1, similar case retrieval unit 25 acquires from the similar case database 9 the image ID and the patient ID associated with the slice thickness and the feature amount whose degree of coincidence with the slice thickness T ₁ and the feature amount C ₁ satisfies a predetermined criterion, and further, the image The image data of the image is acquired from the image information database 5 based on the ID, and the ROI information and findings of the image are acquired from the interpretation report database 7 based on the image ID. Then, based on the acquired image data and ROI information, an image marked on the ROI portion is generated and displayed on the display of the diagnostic imaging medical workstation 10 as an image of a similar case together with the findings upon interpretation of the image. Furthermore, based on the acquired patient ID, the date of birth and sex of the patient are acquired from the patient information table 12a, the age is calculated from the date of birth, and the age and sex are displayed on the display as similar case information. Based on the ID, further refer to the patient diagnosis history table 12c to obtain a diagnosis ID for identifying the diagnosis received by the patient, refer to the diagnosis information table 12e based on the diagnosis ID, and obtain the confirmed diagnosis information of the patient. It can also be acquired and displayed on a display.

On the other hand, in this embodiment, when performing registration in the similar case database 9 as in the second embodiment described above, the similar case information registration unit 36 includes the key slice image _Sk and each of the reconstructed images. tomographic image I _n the slice thickness and the region of interest R _k, the feature amount C _k of R _n, and C _n, the image ID of each image, the registration associating the patient ID identifying the patient of the image performed. At the same time, information of the key slice image S _k and reconstructed each tomographic image I _n is newly registered in the image information database 5.

Next, as another method for performing registration in the similar case database 9 in the present embodiment, a method for searching for an image satisfying a desired condition and registering an image extracted by the search as a similar case will be described. FIG. 14 is a block diagram schematically showing the configuration of a similar case database registration apparatus that implements this method. As shown in the figure, the setting of desired search conditions by an operator such as an image diagnostician or a doctor is accepted, and the image information database 5, the interpretation report database 7, and the electronic medical record database 12 are received based on the search conditions. A search is performed to obtain an image I ₁ satisfying the search condition, its original image I ₀ , position information R _{1 of the} region of interest in the image I ₁ , feature amount C _{1 of the} region of interest, and patient's electronic medical record information K ₁ of the image. the registered image retrieval unit 38, an image I _1, an image reconstruction unit 31 to reconstruct the original image I ₀ Metropolitan plurality of tomographic images I _n having different slice thickness from, in the tomographic image I _n, interest a region of interest extraction unit 35 for extracting a region of interest R _n corresponding to the region R _1, the characteristic amount obtaining section 24 for obtaining the feature amount C _n of the region of interest R _n, the image I _1, each of the tomographic images I _n Slice thickness, feature quantity C ₁ , C _n , electron It consists of a similar case information registration unit 36 that registers medical record information K ₁ in the similar case database 9 in association with each other.

Here, the registration target image search unit 38 is activated, for example, by selecting “search for similar case registration target images” from a menu displayed on the display of the workstation 3 or 10, and the operator sets the search conditions. Provides a user interface for Specific examples of the search conditions include attribute information such as the date of birth and sex of the patient, examination information such as definitive diagnosis information, examination information such as examination type and examination site, interpretation result information such as an interpreter ID and interpretation findings, Image information such as slice position, slice thickness, etc. can be mentioned. For example, the operator has a search condition such as “a chest CT image of a woman in her 40s who has been diagnosed with primary lung cancer and has findings at the time of interpretation”. Can be set. In this search condition example, the registration target image search unit 38 extracts, for example, a diagnosis ID registered as “primary lung cancer” in the definitive diagnosis information of the diagnosis information table 12e, and associates it with the extracted diagnosis ID. The patient ID is acquired from the patient examination history table 12c, and the patient ID of the “40's female” patient is extracted based on the date of birth and gender with reference to the patient information table 12a based on the acquired patient ID. Further, the examination ID associated with the extracted patient ID is acquired from the patient examination history table 12b, and the examination information table 12d is referred to based on the obtained examination ID, and the “chest CT imaging” is performed based on the examination type and the examination site. ”Is extracted. Then, based on the extracted representative image ID, the image interpretation report database 7 is referenced to extract the image ID of the image “findings” based on the findings. The image corresponding to the image ID finally extracted by the above search corresponds to the image I ₁ described above.

The details of the processing of the image reconstruction unit 31, the region of interest extraction unit 35, and the feature amount calculation unit 24 are the same as those in the second embodiment. The slice thickness of the image I ₁ can be acquired from the image information database 5, and the region of interest R ₁ in the image I ₁ can be acquired from the ROI information in the interpretation report database 7. Processing of the region-of-interest setting unit 23 and the feature amount acquisition unit 24 for the image I _{1 in} the embodiment is not necessary.

Similar case information registration unit 36, the image I _1, the image ID of each of the tomographic images I _n, slice thickness, the feature amount C _1, C _n, in association with the patient ID of a patient of the image (electronic medical record information K ₁₎ Register in the similar case database 9.

  As described above, in the similar case search / registration apparatus according to the third embodiment of the present invention, the diagnostic information stored in the electronic medical record database 12 is acquired and interpretation / diagnosis support is performed when registering in the similar case database. Since the information is generated, it is possible to save the trouble of manually generating the interpretation / diagnosis support information from scratch, and the cooperation with other systems such as an electronic medical record system is strengthened, and the information resources are effectively used. Is planned. Furthermore, the interpretation / diagnosis support information is limited to the image ID and patient ID, and the interpretation / diagnosis information can be acquired from the image information database 5, the interpretation report database 7, and the electronic medical record database 12 by links from these IDs. In addition, it is possible to eliminate the waste of storage capacity by storing information of the same content in a plurality of locations, and avoid inconsistencies in data content caused by updating only one of the information stored in the plurality of locations. Therefore, the reliability of the information is improved, and the data synchronization processing for avoiding such inconsistency is not required, so that the processing load of the system is reduced. In particular, since information on electronic medical records, which is an example of diagnostic information, has the property of being updated as needed, the effect of avoiding this data content inconsistency problem is high.

  In the above embodiment, the search / registration to the similar case database in which the tomographic image obtained by reconstructing the three-dimensional medical image of the human chest acquired by CT has been described has been described. The modality for acquiring the image may be MR or the like, the subject may be other than a human, and the examination site may be an abdomen or a head. In addition, when registering similar case information associated with images by a plurality of modalities in the similar case database, information for identifying the type of modality is also associated with the similar case information, and this modality is used when searching. The information to be identified is preferably a search key in order to maintain and improve search efficiency and search accuracy. Similarly, when a plurality of types of subjects and examination sites are registered in the similar case database, it is preferable that information for identifying each is associated with similar case information and used as a search key.

  In addition, the technical scope of the present invention includes various modifications made to the system configuration, processing flow, database configuration, user interface, and the like in the above embodiment without departing from the spirit of the present invention. . Moreover, said embodiment is an illustration to the last and all the above description should not be utilized in order to interpret the technical scope of this invention restrictively.

1 is a schematic configuration diagram of a medical information system in which a similar case retrieval apparatus according to a first embodiment of the present invention is introduced. The block diagram which showed typically the structure of the similar case search device in the 1st Embodiment of this invention. The figure which represented typically the structure of the similar case database in 1st, 2nd embodiment of this invention. The flowchart which shows the flow of the similar case search process in the 1st Embodiment of this invention. The figure which shows an example of the menu screen for performing similar case search The block diagram which showed typically the structure of the similar case search device used as the modification with respect to the 1st Embodiment of this invention. The flowchart which shows the flow of the similar case search process used as the modification with respect to the 1st Embodiment of this invention. The block diagram which showed typically the structure of the similar case registration apparatus in the 2nd Embodiment of this invention. The flowchart which shows the flow of the similar case registration process in the 2nd Embodiment of this invention. The block diagram which showed typically the structure of the similar case search and registration apparatus used as the modification with respect to the 1st, 2nd embodiment of this invention. The flowchart which shows the flow of the similar case search and registration process used as the modification with respect to the 1st, 2nd embodiment of this invention. Schematic configuration diagram of a medical information system in which a similar case search / registration device according to a third embodiment of the present invention is introduced The figure which represented typically the relation and the structure of the database in the medical information system in the 3rd Embodiment of this invention. The block diagram which showed typically the structure of the similar case search and registration apparatus used as the 3rd Embodiment of this invention.

Explanation of symbols

1 3D image capture device (modality)
2 Workstation for checking image quality (QA-WS)
3, 3a, 3b Image diagnostic medical workstation 4 Image information management server 5 Image information database 6 Interpretation report server 7 Interpretation report database 8 Similar case search server 9 Similar case database 10, 10a, 10b Diagnostic medical workstation 11 Electronic medical record management server 12 electronic medical record database 12a patient information table 12b patient examination history table 12c patient examination history table 12d examination information table 12e examination information table 19 network 21 image acquisition unit 22 slice thickness acquisition unit 23 region of interest setting unit 24 feature amount acquisition unit 25 similar cases Search unit 25a Slice thickness search unit 25b Feature quantity search unit 25c Similar case information acquisition unit 26 Slice thickness table 27, 27a, 27b, 27c Slice thickness feature quantity table 28 Image file Group 29 Diagnostic information file group 30 3D image acquisition unit 31 Image reconstruction unit 32 Interpretation report acquisition unit 33 Slice image acquisition unit 34 Key slice image selection unit 35 Region of interest extraction unit 36 Similar case information registration unit 37 Interpretation report generation unit 38 Registration target image search part

Claims (20)

Slice thickness of a tomographic image representing a cross-section obtained by slicing a specific examination target region of one or more subjects with a predetermined slice thickness, a feature amount representing a content feature of at least a part of the tomographic image, and the content A similar case database in which similar case information associated with interpretation / diagnosis support information associated with interpretation / diagnosis support information for assisting diagnosis based on the interpretation of the tomographic image having similar features is used. The registered similar case information for a plurality of different tomographic images, and
A feature amount acquisition means for acquiring a feature amount representing the content feature of at least a partial area of a tomographic image to be searched representing a cross section obtained by slicing the examination target portion of the subject;
A similar case that refers to the similar case database and obtains the similar case information associated with the slice thickness and the feature amount in which the degree of matching with the slice thickness and the feature amount of the image to be searched satisfies a predetermined criterion A similar case retrieval apparatus comprising a retrieval means.   The similar case search means acquires similar case information associated with a slice thickness whose degree of matching with the slice thickness of the tomographic image to be searched satisfies a predetermined criterion, and then the acquired similar case information The similar case according to claim 1, wherein a similar case is acquired from which the degree of coincidence with the feature amount of the tomographic image to be searched is associated with a feature amount satisfying a predetermined criterion. Search device.   When the feature amount acquisition unit determines that there is similar case information associated with a slice thickness whose degree of matching with the slice thickness of the tomographic image to be searched satisfies a predetermined criterion by the similar case search unit The similar case retrieval apparatus according to claim 2, further comprising a process of acquiring the feature amount.   The similar case information includes a plurality of slices having different slice thicknesses representing cross sections of the examination target part of the one subject with respect to the interpretation / diagnosis support information for the examination target part of one subject. The similar case retrieval apparatus according to any one of claims 1 to 3, wherein an image in which each slice thickness of the image is associated with the feature amount is included.   5. The similar case search apparatus according to claim 1, wherein the interpretation / diagnosis support information includes one or more of the plurality of tomographic images. 6. The interpretation / diagnosis support information includes two or more of the plurality of tomographic images, and the two or more tomographic images are associated with each slice thickness of the tomographic image. ,
The similar case retrieval unit selects a tomographic image having a slice thickness that satisfies a predetermined criterion in a degree of coincidence with a slice thickness of the image to be retrieved among the two or more tomographic images included in the interpretation / diagnosis support information. The similar case search device according to claim 5, wherein the similar case search device is acquired. The partial region of the tomographic image is a region of interest in the image,
The similar case search apparatus according to claim 1, wherein the feature amount acquisition unit further includes a unit for designating a region of interest in the search target image. The partial region of the tomographic image is a region of interest in the image,
The similar case search apparatus according to claim 1, wherein the feature amount acquisition unit further includes a unit that detects a region of interest in the search target image. The search target image is reconstructed based on a three-dimensional medical image representing the examination target part of the subject;
As a result of the processing by the similar case search means, when the similar case information of the slice thickness satisfying a predetermined standard with the degree of matching with the slice thickness of the image to be searched is not registered in the similar case database, Slice from a three-dimensional medical image so that the degree of matching with the slice thickness of the similar case information registered in the similar case database satisfies a predetermined criterion and includes the slice position of the search target image Further comprising image reconstruction means for reconstructing a tomographic image representing the cross section obtained,
9. The method according to claim 1, wherein the feature amount acquisition unit and the similar case search unit re-process the image reconstructed by the image reconstruction unit as a new search target image. The similar case search device according to claim 1.   The similar case search apparatus according to claim 1, further comprising search result output means for outputting the similar case information obtained by the similar case search means. Slice thickness of a tomographic image representing a cross-section obtained by slicing a specific examination target region of one or more subjects with a predetermined slice thickness, a feature amount representing a content feature of at least a part of the tomographic image, and the content A similar case database registration device comprising registration means for registering in the database in association with interpretation / diagnosis support information that supports interpretation of the tomographic image and / or diagnosis based on the similar image having similar characteristics,
A plurality of tomographic images representing cross sections sliced so as to include a region of interest in the three-dimensional medical image are reconstructed from a three-dimensional medical image representing the examination target part of one subject with different slice thicknesses. Image reconstruction means;
A feature quantity acquisition means for acquiring a feature quantity representing at least the content feature of the region of interest of each reconstructed tomographic image;
The registration unit registers the slice thickness of each reconstructed tomographic image and the feature amount in association with the interpretation / diagnosis support information for the examination target region of the one subject. A similar case database registration device characterized in that:   12. The similar case database registration apparatus according to claim 11, wherein the interpretation / diagnosis support information includes one or more of the reconstructed tomographic images. The interpretation / diagnosis support information includes two or more of the reconstructed tomographic images;
13. The similar case database registration apparatus according to claim 12, wherein the registration unit registers the two or more tomographic images in association with slice thicknesses of the tomographic images. From the storage means for storing the interpretation information representing the interpretation result of the medical image of the subject and / or the diagnostic information representing the diagnostic content of the subject in a state associated with the medical image of the subject, the three-dimensional The interpretation information and / or the diagnosis information associated with a medical image and at least one of the plurality of tomographic images is acquired, and the interpretation / diagnosis is performed based on the acquired interpretation information and / or the diagnosis information. It further comprises interpretation / diagnosis support information generating means for generating support information,
14. The registration unit according to any one of claims 11 to 13, wherein the registration unit registers the slice thickness and the feature amount in association with the generated interpretation / diagnosis support information. The similar case database registration device described. Interpretation information indicating the interpretation result of the medical image of the subject and / or diagnostic information indicating the diagnostic content of the subject in a state of being associated with the medical image of the subject and the interpretation information and And / or further comprising interpretation / diagnosis support information generating means for acquiring the diagnostic information and generating the interpretation / diagnosis support information based on the acquired interpretation information and / or the diagnostic information.
The image reconstruction means reconstructs the plurality of tomographic images from a three-dimensional medical image associated with the interpretation information and / or the diagnostic information;
The registration unit is configured to register the slice thickness and the feature amount in association with the interpretation / diagnosis support information generated by the interpretation / diagnosis support information generation unit. The similar case database registration device according to any one of 11 to 13.   The similar case retrieval apparatus according to any one of claims 1 to 10, further comprising the similar case database registration apparatus according to any one of claims 11 to 15. Obtaining a feature amount representing a characteristic feature of at least a partial region of a tomographic image to be searched representing a cross section obtained by slicing a specific examination target region of a subject; and
A slice thickness of a tomographic image representing a cross-section obtained by slicing the examination site of one or more subjects with a predetermined slice thickness, a feature amount representing the content feature of at least a partial region of the tomographic image, and the content A similar case database in which similar case information associated with interpretation / diagnosis support information associated with interpretation / diagnosis support information for assisting diagnosis based on the interpretation of the tomographic image having similar features is used. The slice thickness and the feature quantity satisfying a predetermined criterion with respect to the slice thickness of the image to be searched and the degree of coincidence with the feature quantity with reference to the registered similar case information for a plurality of different tomographic images And acquiring the similar case information associated with a similar case retrieval method. Reconstructing a plurality of tomographic images representing a cross-section sliced so as to include a region of interest in the three-dimensional medical image with a plurality of different slice thicknesses from a three-dimensional medical image representing a specific examination target part of one subject And steps to
Obtaining a feature amount representing a content feature of at least the region of interest of each reconstructed tomographic image;
For interpretation / diagnosis support information for supporting interpretation of a tomographic image having a similar content feature to a tomographic image representing a cross section of the examination target region of the one subject and / or diagnosis based on the similar image And a similar case database registration method comprising the step of registering the slice thickness of each reconstructed tomographic image and the feature amount in a database in association with each other. On the computer,
Obtaining a feature amount representing a characteristic feature of at least a partial region of a tomographic image to be searched representing a cross section obtained by slicing a specific examination target region of a subject; and
A slice thickness of a tomographic image representing a cross-section obtained by slicing the examination site of one or more subjects with a predetermined slice thickness, a feature amount representing the content feature of at least a partial region of the tomographic image, and the content A similar case database in which similar case information associated with interpretation / diagnosis support information associated with interpretation / diagnosis support information for assisting diagnosis based on the interpretation of the tomographic image having similar features is used. The slice thickness and the feature quantity satisfying a predetermined criterion with respect to the slice thickness of the image to be searched and the degree of coincidence with the feature quantity with reference to the registered similar case information for a plurality of different tomographic images And a step of acquiring the similar case information associated with a similar case search program. On the computer,
Reconstructing a plurality of tomographic images representing a cross-section sliced so as to include a region of interest in the three-dimensional medical image with a plurality of different slice thicknesses from a three-dimensional medical image representing a specific examination target part of one subject And steps to
Obtaining a feature amount representing a content feature of at least the region of interest of each reconstructed tomographic image;
For interpretation / diagnosis support information for supporting interpretation of a tomographic image having a similar content feature to a tomographic image representing a cross section of the examination target region of the one subject and / or diagnosis based on the similar image A similar case database registration program for executing a step of registering the slice thickness of each reconstructed tomographic image and the feature amount in association with each other in a database.

Images