Medical image processing device, method and program

Abstract

Medical image processing device comprising: a determining device (22) that determines whether a first medical image (V1) of a subject contains at least one section of an upper end vertebra and at least one section of a lower end vertebra; an image acquisition device (21) that, if the determination is negative, acquires a second medical image (V2) which allows the identification of a vertebra marking of the subject; and a marking device (23) that aligns the first medical image (V1) with the second medical image (V2) and marks the vertebra contained in the first medical image (V1).

Description

Background of the invention Field of invention The present invention relates to a medical image processing device, a medical image processing method and a medical image processing program for detecting a vortex contained in a medical image. Description of the state of the art The spinal cord is a crucial structure that plays a vital role in transmitting messages between the brain and the body. For this reason, the spinal cord is protected by several vertebrae (the vertebral column). Furthermore, it is standard practice to confirm the presence or absence of damage or injury to a vertebra by interpreting a tomographic image obtained by scanning a patient. This requires identifying each vertebra, for example, to create a report on a damaged or injured vertebra. Therefore, various image processing algorithms exist for separating and identifying multiple vertebrae based on a tomographic image obtained by scanning a subject (patient), and for marking each vertebra. For example, patent specification 1 proposes a method in which, using tomographic images of a calculated tomography image (CT), a magnetic resonance imaging (MRI) image, or the like as a target, tomographic images of planes intersecting each vertebra and running parallel to the central axis of each vertebra are generated, a characteristic value representing the sharpness of a cross-sectional shape in each tomographic image is calculated, and a characteristic value representing a regularity of an arrangement of vertebrae is calculated, a region of each vertebra is identified by identifying a location of an intervertebral disc between the respective vertebrae based on these characteristic values, and furthermore, the identified region of each vertebra is marked. Non-patent 1 proposes a method for extrapolating vertebrae and a sacrum using the Deformable Part Model and Dynamic Programming. The method described in non-patent 1 enables the extrapolation of a sacrum and the marking of vertebrae with reference to the sacrum. Patent specification 2 describes user interfaces and workflows for display and navigation based on pre-identified anatomical landmarks, as well as the linking of multiple images and label transfer within a common frame of reference. Patent 3 discloses an automated system for the detection and numbering of intervertebral discs / vertebrae in sagittal MR / CT images with heuristic search and chaining algorithms, optional user interaction, and the use of additional sequences / data sets (e.g., IDEAL) and image combinations generated in the same examination setting. [List of printed publications] [Patent specification] PTL 1: Japanese unexamined patent publication JP 2011 - 131 040 APTL2: Published US patent application US 2012 / 0172700 A1 PTL 3: Published US patent application US 2013 / 0287276 A1 [Non-patent specification] NPL 1: “Vertebrae Detection and Labeling in Lumbar MR Images”, MICCAI Workshop: Computational Methods and Clinical Applications for Spine Imaging, 2013. REVELATION OF THE INVENTION In the case that the target image only includes part of a spine, for example only the thoracic spine, the positions of the vertebrae cannot be recognized using the method described in patent specification 1 or in non-patent specification 1, so that as a result the vertebrae cannot be marked. The present invention was developed in view of the circumstances described above, and it is an objective of the present invention to enable the marking of the vertebra contained in the image, even in the case of an image which contains only a part of a vertebra, within the framework of a medical image processing device, a method and a program. A medical image processing device according to the present invention comprises: a determining device that determines whether a first medical image of a subject contains at least one section of an upper end vertebra and at least one section of a lower end vertebra; an image acquisition device that, if the determination is negative, acquires a second medical image which allows the detection of a marker of the subject's vertebra; and a marking device that aligns the first medical image with the second medical image and marks the vertebra contained in the first medical image. The term "at least a section of an upper end vertebra" refers to a section of several vertebrae as part of a spinal column, which can be identified as the upper end vertebra. It can therefore refer to the entire upper end vertebra or a section thereof. The term "at least a section of a lower end vertebra" refers to a part of several vertebrae that form a spine, which can be identified as the lower end vertebra. Consequently, it can refer to the entire lower end vertebra or a section thereof. The "marking" can be any information as long as it allows for the identification of the vertebra's anatomical location. For example, the marking can be an anatomical symbol, a number, or a combination of a symbol and a number specific to that particular vertebra. Furthermore, the marking can be a symbol, a number, a combination of a symbol and a number, and so on, referring to a specific vertebra. In the medical image processing device according to the invention, the marking device can be a device which, if the determination is positive, marks the vertebra contained in the first medical image based on information about either at least one section of the upper end vertebra and at least one section of the lower end vertebra contained in the first medical image. Furthermore, in the medical image processing device according to the invention, the image acquisition device can be a device that acquires a medical image as the second medical image, which contains at least either at least one section of the upper end vertebra or at least one section of the lower end vertebra. Furthermore, in the medical image processing device according to the invention, the image acquisition device can be a device which acquires an image as the second medical image in which a marking of at least one vertebra of the subject is known. Furthermore, in the medical image processing device according to the invention, the upper end vertebra can be at least one of a first and a second cervical vertebra. Furthermore, in the medical image processing device according to the invention, the lower end vertebra can be at least one of a fifth lumbar vertebra and a sacrum. Furthermore, in the medical image processing device according to the invention, the image acquisition device can be a device which, as the second medical image, acquires a medical image which was acquired using the same imaging method as the first medical image when several medical images are available that allow for the recognition of a marking of the subject's vertebrae. Furthermore, in the medical image processing device according to the invention, the image acquisition device can be a device which, as the second medical image, acquires a medical image which, with regard to the imaging time, is close to the first medical image when several medical images are available that enable the recognition of a marking of the subject's vertebrae. Regarding the “medical image which is closest in terms of imaging time”, preferably a medical image which is closest in imaging time is used, however this is not restrictive, one can also use a medical image which is within a predetermined number of days of the image acquisition time of the first medical image. A medical image processing method according to the invention comprises the following steps: determining whether a first medical image of a subject contains at least one segment of an upper end vertebra and at least one segment of a lower end vertebra; if the determination is negative, obtaining a second medical image that allows for the identification of a vertebra of the subject; and aligning the first medical image with the second medical image and marking the vertebra contained in the first medical image. Note that the medical processing method according to the invention can be designed as a program that is executed by a computer. To mark vertebrae, it is necessary to use at least one upper and one lower end vertebra as a reference. According to the invention, a determination is made as to whether at least one segment of an upper end vertebra and at least one segment of a lower end vertebra are contained in a first medical image of a person. If the determination is negative, a second medical image is obtained, which allows the vertebra to be identified. The first and second medical images are aligned, and the vertebra contained in the first medical image is marked using the second medical image. Thus, even if the first medical image contains only a segment of a vertebra, the vertebra contained in the first medical image can be marked with the help of the second medical image. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a hardware configuration diagram of a diagnostic support system in which a medical image processing device according to an embodiment of the invention is used, showing an overall view of the system. Fig. 2 shows a medical image processing device realized by installing a medical image processing program on a computer, its schematic setup being shown. Fig. 3 schematically shows a sagittal view to illustrate an arrangement of vortices. Fig. 4 shows an example of a first three-dimensional image. Fig. 5 shows an example of a second three-dimensional image. Fig. 6 shows another example of a second three-dimensional image. Fig. 7 is a drawing to explain an alignment. Fig. 8 is a flowchart illustrating the processing performed in the present embodiment. DETAILED DESCRIPTION OF PREFERRED EXECUTION FORMS The following describes embodiments of the invention with reference to the accompanying drawings. Fig. 1 shows a hardware configuration of a diagnostic support system in which a medical image processing device according to one embodiment of the invention is used, and an overall arrangement of the system is shown. As shown in Fig. 1, the system includes a medical image processing device 1 according to the present embodiment, a three-dimensional image acquisition device 2, and an image storage server 3, which is connected via a network 4 for communication. The three-dimensional image acquisition device 2 is a device that acquires a diagnostic target zone of a patient and generates a three-dimensional image representing that zone. In particular, the device 2 is a CT scanner, an MRI scanner, a positron emission tomography (PET) scanner, or the like. The two-dimensional image generated by the three-dimensional image acquisition device 2 is sent to the image storage server 3 and stored there. In the present embodiment, it is assumed that the patient's diagnostic target zone is a vertebra, that the three-dimensional image acquisition device 2 is an MRI scanner, and that the three-dimensional image is an MRI image. The image storage server 3 is a computer that stores and manages various types of data and includes a high-capacity external storage device and database management software. The image storage server 3 communicates with other devices via a wired or wireless network 4 to send and receive image data and similar information. Specifically, the image storage server 3 receives image data of a three-dimensional image generated by the three-dimensional image acquisition device 2 and similar data via the network, and it manages the data by storing it on a storage medium, such as a high-capacity external storage device. The image data storage format and the communication between each device are based on a protocol such as Digital Imaging and Communications in Medicine (DICOM). Furthermore, a DICOM standard tag is attached to the three-dimensional image. This tag contains various types of information, including a patient's name, information about the imaging device, the date and time of image acquisition, the imaged area, and the like. The medical image processing device 1 is a computer in which a medical image processing program according to the invention is installed. The computer can be a workstation or a personal computer operated directly by the physician performing the diagnosis, or it can be a server computer connected via a network. The medical image processing program is distributed as a recording on a recording medium, for example, DVD, CD-ROM, or the like, and is installed from the recording medium onto the computer. Alternatively, the program is stored in a storage device of the server computer connected to the network, or is stored in a network storage device in an externally accessible state, and is downloaded and installed on the computer used by the physician upon request. Fig. 2 shows a medical image processing device implemented by installing the medical image processing program on a computer, the schematic structure of which is depicted. As can be seen from Fig. 2, the medical image processing device 1 includes a CPU 11, a memory 12, and a storage 13 in the configuration of a standard workstation. Furthermore, a display 14 and an input device 15, for example in the form of a mouse, are connected to the medical image processing device 1. Memory 13 contains various types of information, including a three-dimensional image obtained from image storage server 3 via network 4, and an image generated by the processing performed in medical image processing device 1. Memory 12 stores the medical image processing program. The medical image processing program defines, in the form of processes executed by CPU 11, a first image acquisition process to obtain a first three-dimensional image V1 containing a diagnostic target vertebra of a person; a determination process to ascertain whether the first three-dimensional image V1 contains at least one section of an upper end vertebra and at least one section of a lower end vertebra; a second image acquisition process to obtain, if the determination is negative, a second three-dimensional image V2, which allows the identification of a vertebra of the person; and a marking process to align the first three-dimensional image V1 with the second three-dimensional image V2 and to mark the vertebra contained in the first three-dimensional image V1. The execution of these processes by CPU 11 according to the program causes the computer to function as an image acquisition unit 21, a detection unit 22, and a marking unit 23. The medical image processing device 1 can contain multiple CPUs for executing the first and second image acquisition processes, the detection process, and the marking process, respectively. The image acquisition unit 21 acquires the first and second three-dimensional images V1, V2 from the image storage server 3. In the event that the three-dimensional images V1, V2 are already stored in memory 13, the image acquisition unit 21 can retrieve the images from memory 13. The determination unit 22 determines whether the first three-dimensional image V1 contains at least one segment of an upper end vertebra and at least one segment of a lower end vertebra, and it outputs the result of this determination to the image acquisition unit 21 and the labeling unit 23. Fig. 3 schematically illustrates a sagittal image depicting the arrangement of vertebrae. As shown in Fig. 3, each vertebra is anatomically labeled. Here, the vertebral column consists of four sections: the cervical spine, the thoracic spine, the lumbar spine, and the sacrum. The cervical spine consists of seven vertebrae, anatomically labeled C1 to C7. The thoracic spine consists of twelve vertebrae and is anatomically labeled Th1 to Th12. The lumbar spine consists of five vertebrae and is anatomically labeled L1 to L5.The sacrum consists of only one bone with the anatomical designation S1. The present embodiment makes use of these anatomical markings as markers to be applied to the vertebrae. As shown in Fig. 3, vertebrae have a cylindrical shape; however, the first and second cervical vertebrae at the upper end and the fifth lumbar vertebra and the sacrum at the lower end have characteristic shapes that differ from those of the other vertebrae. For this reason, the determining unit 22 of the present embodiment uses at least one of the first and second cervical vertebrae as the upper end vertebra and at least one of the fifth lumbar vertebra and the sacrum as the lower end vertebra as the objective for reaching a decision. Note that the determination unit 22 performs at least one of the following: a first determination, which is a determination as to whether the first three-dimensional image V1 contains at least a section of the upper end vortex, that is, contains at least an area in which the upper end vortex is recognizable; and a second determination, which is a determination as to whether the first three-dimensional image V1 contains at least a section of the lower end vortex, that is, contains at least an area in which the lower end vortex can be recognized. In the present embodiment, only the first or the second determination is performed, or both the first and the second determination are performed.Furthermore, the first determination can be performed first, and if this first determination is negative, the second determination can be performed, whereas if the first determination is positive, the second determination can be omitted. In contrast, the second determination can be performed first, and if this second determination is negative, the first determination can be performed, whereas if the second determination is positive, the first determination can be omitted. If, in this case, both the first and second determinations are negative, then the determination as to whether the first three-dimensional image V1 contains at least one section of the upper end-vertebral and at least one section of the lower end-vertebral is negative; otherwise, it is positive. The identification unit is equipped with a template containing a pattern that represents a three-dimensional shape of at least one of the first and second cervical vertebrae for the first identification. Furthermore, identification unit 22 is equipped with a template containing a pattern that represents a three-dimensional shape of at least one of the fifth lumbar vertebrae and the sacrum for the second identification. Identification unit 22 compares the first three-dimensional image with the template to locate an area that has a shape identical to that of the template pattern. If the area is then found, both the first and second identifications are positive. If the first three-dimensional image V1 contains only a section of the thoracic spine, as shown in Fig. 4, then the finding made by the determining unit 22 is negative. On the other hand, if the first three-dimensional image V1 contains at least one of the first and second cervical vertebrae, as shown in Fig. 5, or if the first three-dimensional image V1 contains at least one of the fifth lumbar vertebra and the sacrum, as shown in Fig. 6, then the finding made by the determining unit 22 is positive. It should be noted that in each of Figs. 6 to 7, a sagittal image passing through the center of the vertebrae, generated from a three-dimensional image, is shown for illustrative purposes. The marking unit 23 performs a first marking process if the determination unit 22's finding is negative; if it is positive, it performs a second marking process. The first marking process will be explained first. If a finding by the determination unit 22 is negative, the image acquisition unit 21 acquires a second three-dimensional image V2 from the image storage server 3, which enables the detection of a marking of a vertebra of the person corresponding to the first three-dimensional image V1. A DICOM standard tag is attached to a three-dimensional image stored on image storage server 3. This tag contains a patient name, information about the imaging device, the date and time of image acquisition, the imaged area, and similar data. Image acquisition unit 21 acquires a second three-dimensional image, V2, whose tag contains the same patient name as the first three-dimensional image, V1. This acquired image allows for the identification of a vertebral marker. Regarding the three-dimensional image that enables the identification of a vertebral marker, an image can be used in which vertebrae have already been marked and whose tag contains the relevant information.Alternatively, an image can be used with a tag containing information indicating that at least part of the upper end vertebra or at least part of the lower end vertebra is included in the image. If multiple three-dimensional images of the same patient, enabling the identification of a vertebral marker, are stored in image storage server 3, a second three-dimensional image (V2) is obtained using the same imaging technique as the first three-dimensional image (V1). For example, if the first three-dimensional image (V1) is an MRI image, then an MRI image will be obtained as the second three-dimensional image (V2). Furthermore, the second three-dimensional image (V2) can be chosen to have the closest imaging time to the first three-dimensional image (V1).Furthermore, a determination can be made as to whether a three-dimensional image obtained using the same imaging method as the first three-dimensional image V1 is stored in the image storage server 3 or not, and if the determination is negative, a second three-dimensional image V2 can be obtained which is closest to the first three-dimensional image V1 with regard to the image acquisition time. If the second three-dimensional image V2 is an image whose tag contains marking information about vertebrae contained in the image, then the marking unit 23 marks the vertebra contained in the first three-dimensional image V1 using the three-dimensional image V2. If, on the other hand, the second three-dimensional image V2 is an image whose tag does not contain marking information about vertebrae contained in the image, then the image contains at least one section of the upper end vertebra and at least one section of the lower end vertebra, and the marking unit 23 detects the position of each vertebra in the second three-dimensional image V2, and it marks each vertebra. The processing for detecting the position of each vertebra and marking each vertebra is identical to a second marking process, which is explained below, so its detailed description is omitted here. The marking unit 23 then performs an alignment between the first three-dimensional image V1 and the second three-dimensional image V2. As an alignment method, a comparison between the first three-dimensional image V1 and the second three-dimensional image V2 can be performed to locate an area containing a vortex with the same shape as a vortex contained in the first three-dimensional image V1. If the imaging method for the first three-dimensional image V1 differs from that of the second three-dimensional image V2, for example, if the first three-dimensional image V1 is an MRI image and the second three-dimensional image V2 is a CT image, the alignment can be performed using the method described in "Multi-modal volume alignment by maximization of mutual information," Medical Image Analysis (1996), Vol. 1, No. 1, pp. 35-51 (Reference Literature 1). This leads to an alignment of the first three-dimensional image V1 with the second three-dimensional image V2, as shown in Fig. 7. In the second three-dimensional image V2, the markings of the vortices are already known. If the markings in the second three-dimensional image V2 are Th2, Th3, and Th4, corresponding to the vortices contained in the area corresponding to the first three-dimensional image V1, then the vortices contained in the first three-dimensional image V1 can be marked Th2, Th3, and Th4, respectively. Next, a second marking process will be described. This second marking process is a procedure for marking vertebrae contained in the first three-dimensional image V1 based on at least one piece of information from at least one section of the upper end vertebra and at least one section of the lower end vertebra contained in the first three-dimensional image V1. The marking unit 23 performs the second marking process using, for example, the method described in patent specification 1. That is, the marking unit 23 identifies a central axis of each vertebra from the first three-dimensional image V1 and generates tomographic images of cross-sectional planes parallel to the identified central axis of each vertebra.The marking unit 23 then identifies the position of each vertebra based on a characteristic value representing the sharpness of a cross-sectional shape in each tomographic image and a characteristic value representing the regularity of the vertebrae's arrangement. If the first three-dimensional image V1 contains at least a section of the upper end vertebra, the marking unit 23 marks the identified vertebrae in sequence starting from the upper end. If the first three-dimensional image V1 contains at least a section of the lower end vertebra, the marking unit 23 marks the identified vertebrae in sequence starting from the lower end. Note that if the first three-dimensional image V1 contains a sacrum, the marking unit 23 can detect the sacrum using the method described in non-patent 1 and mark the vertebrae with reference to the sacrum. The procedure for marking vertebrae is not limited to the procedures described above; any suitable procedure can be used. The following describes processing according to the present embodiment. Fig. 8 is a flowchart illustrating the processing performed in the present embodiment. First, the image acquisition unit 21 acquires a first three-dimensional image V1 from the image storage server 3, serving as a diagnostic target (step ST1). The determination unit 22 then performs a determination to ascertain whether the first three-dimensional image V1 contains at least a section of the upper end vortex, that is, whether it contains at least an area in which the upper end vortex can be detected (determination process, step ST2). If step ST2 is negative, the marking unit 23 performs the first marking process (step ST3). If, on the other hand, step ST2 is positive, the marking unit 23 performs the second marking process (step ST4).The marking unit 23 then saves the first three-dimensional image V1 in which the vortices are marked (step ST5), and the processing is completed. If, in the present embodiment, a determination by the determining unit 22 is negative in this way, a second three-dimensional image V2 is obtained. Subsequently, the first three-dimensional image V1 and the second three-dimensional image V2 are aligned, and a vortex contained in the first three-dimensional image V1 is marked with the aid of the second three-dimensional image V2. Thus, even if the first three-dimensional image V1 contains only part of a vortex, the vortex contained in the first three-dimensional image V1 can be marked. If several three-dimensional images are available that could be used for the second three-dimensional image V2, then one of these images, acquired using the same imaging method as the first three-dimensional image V1, is selected as the second three-dimensional image V2. This allows for the precise marking of the vortex to be marked, which is contained in the first three-dimensional image V1. Furthermore, if several three-dimensional images are available that are suitable as the second three-dimensional image V2, the image that most closely matches the first three-dimensional image V1 in terms of its acquisition time is selected as the second three-dimensional image V2. This allows for the precise marking of the vortex to be marked, which is contained in the first three-dimensional image V1. In the above embodiment, during the execution of the first marking process, it may happen that no image is stored in the image storage server 3 that would allow the detection of a vortex marker. In this case, the marking unit 23 can display a message on the display 14 indicating that the vortex cannot be marked, that the vortex contained in the first three-dimensional image is not markable. Instead of or in addition to this message display, a voice message can be output. Furthermore, in the above embodiment, a vertebra is marked using three-dimensional images as the first and second medical images according to the invention. However, the marking of a vertebra can also be performed with a two-dimensional X-ray image obtained by X-ray imaging, or with a tomographic image of a sagittal section through the center of the vertebra as the target. In this case, different image types can be used for the first and second medical images, with, for example, the first medical image being a three-dimensional image and the second medical image being an X-ray image. Furthermore, in this case, a pseudo-two-dimensional X-ray image can be generated by projecting the three-dimensional image of the first medical image and then marking a vertebra using the pseudo-X-ray image as the target.

Claims

Medical image processing device comprising: a determining device (22) that determines whether a first medical image (V1) of a subject contains at least one section of an upper end vertebra and at least one section of a lower end vertebra; an image acquisition device (21) that, if the determination is negative, acquires a second medical image (V2) which allows the identification of a vertebra marking of the subject; and a marking device (23) that aligns the first medical image (V1) with the second medical image (V2) and marks the vertebra contained in the first medical image (V1). Medical image processing device according to claim 1, characterized in that the marking device (23) is a device which, if the determination is positive, marks the vertebra contained in the first medical image (V1), based on information about either at least one section of the upper end vertebra or at least one section of the lower end vertebra contained in the first medical image. Medical image processing device according to claim 1 or 2, characterized in that the image acquisition device (21) is a device that acquires a medical image which contains at least either at least one section of the upper end vertebra or at least one section of the lower end vertebra of the subject as a second medical image (V2). Medical image processing device according to claim 1 or 2, characterized in that the image acquisition device (21) is a device that acquires an image in which a marking of at least one vertebra of the subject is known as the second medical image (V2). Medical image processing device according to one of claims 1 to 4, characterized in that the upper end vertebra is at least one of a first cervical vertebra and a second cervical vertebra. Medical image processing device according to one of claims 1 to 5, characterized in that the lower end vertebra is at least one of a fifth lumbar vertebra and a sacrum. Medical image processing device according to one of claims 1 to 6, characterized in that the image acquisition device (21) is a device which, when several medical images are available that enable the detection of a marking of a vertebra of the subject, acquires a medical image as the second medical image (V2) which is acquired using the same imaging method as the first medical image (V1). Medical image processing device according to one of claims 1 to 6, characterized in that the image acquisition device (21) is a device which, when several medical images are available that enable the detection of a marking of the subject's vertebrae, acquires a medical image (V2) which is close to the first medical image (V1) with respect to the image acquisition time. A medical image processing procedure comprising the following steps: Determining whether at least one of at least one segment of an upper end vertebra and at least one segment of a lower end vertebra is included in a first medical image (V1) of a subject; if the determination is negative, obtaining a second medical image (V2) that allows recognition of a marker of the subject's vertebra; and aligning the first medical image (V1) with the second medical image (V2) and marking the vertebra included in the first medical image (V1). Medical image processing program for causing a computer to perform the following steps: Determine whether a first medical image (V1) of a subject contains at least one of at least one segment of an upper end vertebra and at least one segment of a lower end vertebra; if the determination is negative, obtain a second medical image (V2) that allows recognition of a marker of the subject's vertebra; and align the first medical image (V1) with the second medical image (V2) and mark the vertebra contained in the first medical image (V1).

Images