JP2018187384A - Medical information processing system and medical information processing apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a medical information processing system and a medical information processing apparatus capable of accurately predicting a result of pathological examination based on a medical image.SOLUTION: The medical information processing system according to embodiments includes a first acquisition part, a second acquisition part, and an associating part. The first acquisition part identifies the position of a tissue from first medical image data being an image of a target site acquired before the tissue in the target site is collected and obtains an image feature value of the tissue. The second acquisition part obtains a result of a pathological examination performed on the tissue. The associating part associates the image feature value of the tissue with the result of the pathological examination.SELECTED DRAWING: Figure 1

Description

  Embodiments described herein relate generally to a medical information processing system and a medical information processing apparatus.

  In recent years, the mechanism of diseases and drug efficacy has been elucidated molecularly, and accumulated knowledge has been applied to clinical tests. In particular, in the field of pathology, many tests are carried out to detect the expression of specific proteins and genes (hereinafter also referred to as molecular markers) that are specific to specific diseases and pathological conditions and can predict prognosis and drug efficacy. It is like that. Such examination is performed using a part of a diseased tissue collected by surgery or biopsy as a sample. However, the molecular marker is not always expressed homogeneously in the diseased tissue. For this reason, there is a risk that a false positive or false negative test result may be obtained.

  Therefore, in order to solve such a problem, the examination results based on medical images collected by a CT (Computed Tomography) apparatus, an MRI (Magnetic Resonance Imaging) apparatus, a nuclear medicine apparatus, etc. capable of rendering the entire diseased tissue, and disease Attempts have been made to link the results of physical molecular markers and use them for diagnosis. Such a field is called, for example, Radiomics.

Japanese translation of PCT publication No. 2002-517836 Japanese Patent No. 5094770

  The problem to be solved by the present invention is to provide a medical information processing system and a medical information processing apparatus capable of accurately predicting a test result of a pathological examination based on a medical image.

  The medical information processing system according to the embodiment includes a first acquisition unit, a second acquisition unit, and an association unit. The first acquisition unit specifies the position of the tissue from first medical image data that is an image of the target region before the tissue in the target region is collected, and acquires an image feature amount of the tissue. The second acquisition unit acquires a test result of a pathological test on the tissue. The association unit associates the image feature amount of the tissue with the examination result of the pathological examination.

FIG. 1 is a diagram illustrating an example of a configuration of a medical information processing system including a medical information processing apparatus according to the first embodiment. FIG. 2 is a diagram for explaining calculation of an expression intensity score of a molecular marker by immunostaining according to the first embodiment. FIG. 3 is a diagram illustrating an example of image information according to the first embodiment. FIG. 4 is a diagram illustrating an example of related information according to the first embodiment. FIG. 5 is a diagram for explaining an example of a pixel specifying process corresponding to a tissue by the first acquisition function according to the first embodiment. FIG. 6 is a diagram for explaining an example of calculation of a feature amount by the first acquisition function according to the first embodiment. FIG. 7 is a diagram for explaining an example of the prediction process by the prediction function according to the first embodiment. FIG. 8 is a flowchart illustrating a processing procedure performed by the medical information processing apparatus according to the first embodiment. FIG. 9 is a flowchart illustrating a processing procedure performed by the medical information processing apparatus according to the first embodiment. FIG. 10 is a diagram for explaining an example of a pixel specifying process corresponding to a tissue by the first acquisition function according to the second embodiment.

  Exemplary embodiments of a medical information processing system and a medical information processing apparatus according to the present application will be described below in detail with reference to the accompanying drawings. In the following, a medical information processing system including a medical information processing apparatus having the function of the medical information processing apparatus according to the present application will be described as an embodiment. However, the medical information processing system and the medical information processing apparatus according to the present application are described below. It is not limited by the embodiment shown.

(First embodiment)
First, the configuration of a medical information processing system including the medical information processing apparatus according to the first embodiment will be described. FIG. 1 is a diagram illustrating an example of a configuration of a medical information processing system including a medical information processing apparatus according to the first embodiment. For example, as illustrated in FIG. 1, the medical information processing system according to the first embodiment includes a medical image diagnostic apparatus 100, a server apparatus 200, an inspection apparatus 300, and a medical information processing apparatus 400. Each device is connected via a network. The medical information processing system may be further connected to another information processing apparatus such as a user terminal via a network.

  The medical image diagnostic apparatus 100 includes an X-ray diagnostic apparatus, an X-ray CT (Computed Tomography) apparatus, an MRI (Magnetic Resonance Imaging) apparatus, an ultrasonic diagnostic apparatus, a SPECT (Single Photon Emission Computed Tomography) apparatus, and a PET (Positron Emission Tomography). An SPECT-CT apparatus in which a SPECT apparatus and an X-ray CT apparatus are integrated, a PET-CT apparatus in which a PET apparatus and an X-ray CT apparatus are integrated, or a group of these apparatuses. The medical image diagnostic apparatus 100 collects medical image data of a subject (a target part in the subject), and generates a medical image for display from the collected medical image data. Then, the medical image diagnostic apparatus 100 displays the generated medical image on the display of the own apparatus. The medical image diagnostic apparatus 100 transmits the collected medical image data and the generated medical image to each apparatus connected to the network.

  For example, the medical image diagnostic apparatus 100 collects medical image data before and after biopsy of the same subject. That is, the medical image diagnostic apparatus 100 collects medical image data of a target part before the tissue is collected and medical image data of the target part after the tissue is collected in the same subject. Here, in the medical image diagnostic apparatus 100, medical image data may be collected by different apparatuses for the same target portion of the same subject. For example, an X-ray CT apparatus, an MRI apparatus, and a PET apparatus collect medical image data before and after biopsy of the same target region of the same subject. In the following, the types of images for each apparatus in the medical image diagnostic apparatus 100 (for example, CT images collected by an X-ray CT apparatus, MR images collected by an MRI apparatus, etc.) and images collected by the same apparatus Are also referred to as image types.

  Here, in the medical image diagnostic apparatus 100, various types of information can be acquired by an imaging method. For example, in an MRI system, the relaxation time of in vivo hydrogen atoms can be measured by various imaging sequences, but the existence of biopolymers such as proteins and the presence of pathological conditions such as inflammation and edema can be estimated from the length of the relaxation time. Can do. The MRI apparatus can also obtain information on the abundance ratio and pH of a specific metabolite in the imaging region. In the X-ray CT apparatus, the electron density information of the tissue can be obtained using a plurality of X-ray energies. Further, in the PET apparatus, metabolic information of a plurality of macromolecules can be obtained by changing the tracer administered to the subject. Image types in the present embodiment include image types for obtaining various information as described above.

  The server apparatus 200 stores medical image data collected by various medical image diagnostic apparatuses, and stores related information generated by the medical information processing apparatus 400. For example, the server apparatus 200 acquires medical image data and medical images from the medical image diagnostic apparatus 100 via a network, and stores the acquired medical image data and medical images in a storage circuit provided inside or outside the apparatus. . The server apparatus 200 acquires related information from the medical information processing apparatus 400 via a network, and stores the acquired related information in a storage circuit provided inside or outside the apparatus. Details of the related information will be described later.

  The examination apparatus 300 acquires examination information (biopsy information) for pathological examination, stores the obtained biopsy information, or transmits the biopsy information to the medical information processing apparatus 400. Specifically, the examination apparatus 300 stores biopsy information for a tissue collected from a target region of the subject, or transmits biopsy information to the medical information processing apparatus 400. For example, the inspection apparatus 300 stores biopsy information for a tissue sampled from a target region from which medical image data before and after biopsy is collected by the medical image diagnostic apparatus 100, or stores the biopsy information in the medical information processing apparatus 400. Send inspection information.

  Here, examples of the pathological examination for the tissue collected from the target site include cytological examination, histological examination, molecular pathological examination, and the like. For example, when a cytological examination or a histological examination is performed, a collected tissue or cell specimen is observed by a pathologist. For example, a pathologist observes a specimen under a microscope to make a morphological diagnosis such as tumor benign / malignant discrimination, diagnosis name, lesion nature, lesion spread, therapeutic effect, and prognosis. The inspection apparatus 300 receives these morphological diagnosis results and identification information (a patient ID, an inspection ID of a pathological examination, etc.) of a tissue (cell) on which a pathological examination has been performed, and stores them in association with each other. Furthermore, the inspection apparatus 300 transmits the received diagnosis result to the medical information processing apparatus 400 together with the identification information.

  In addition, for example, when molecular pathological examination is performed, immunostaining using collected tissues and cells, in situ hybridization, RT (Reverse Transcription) -PCR (Polymerase Chain Reaction) based techniques, etc. The expression of molecular markers (such as specific proteins and genes) is examined. Here, the molecular marker is a specific protein or gene that is specific to a specific disease or pathological condition and can predict prognosis or drug efficacy. For example, in immunostaining, the expression of a molecular marker is examined using an antigen-antibody reaction using an antibody specific for a molecular marker (antigen protein) in a tissue. Here, in immunostaining, an antigen-antibody reaction is visualized by color development by a fluorescent antibody method or an enzyme antibody method. For example, the pathologist determines the expression intensity of the molecular marker based on the number of colored cells (stained) and the degree of coloring (staining).

  In addition, for example, in situ hybridization uses a nucleic acid molecule complementary to a molecular marker (specific DNA or mRNA) in a tissue (cell), and uses the formation of a complex between the molecular marker and the nucleic acid molecule. Marker expression is examined. Here, in in situ hybridization, the formation of a complex between a molecular marker and a nucleic acid molecule is detected by a fluorescent substance or detected by immunohistochemistry. For example, the pathologist determines whether the expression of the molecular marker is positive / negative based on the presence or absence of complex formation.

  Further, for example, in a technique based on RT-PCR, RNA is extracted from a tissue, and cDNA is synthesized from the RNA by reverse transcriptase. Then, the expression of the molecular marker is examined by real-time PCR using cDNA as a template and a primer having a sequence complementary to the molecular marker (specific gene). Here, in the technique based on RT-PCR, the rise of the amplification curve in real-time PCR reflects the amount of cDNA in the sample. Since the amount of cDNA in the sample is proportional to the expression level of the molecular marker, the speed of the rise of the amplification curve in real-time PCR indicates the expression level of the molecular marker. For example, the pathologist judges the expression intensity of the molecular marker based on the expression level of the molecular marker with respect to the expression level of the control gene (gene expressed in a certain amount regardless of the disease).

  The inspection apparatus 300 accepts the examination result of the molecular pathological examination as described above and the identification information of the tissue (cell) on which the examination is performed, and stores them in association with each other. Furthermore, the inspection apparatus 300 transmits the received inspection result together with the identification information to the medical information processing apparatus 400. Here, the inspection apparatus 300 can score the inspection result of the molecular pathological inspection described above. For example, in the case of immunostaining, the inspection apparatus 300 sets “5 points” as the most strongly stained cells, counts the number of cells at each of the 6 stages of 0 to 5 and counts “(staining intensity × cell number)”. Is the expression intensity score of the molecular marker in the collected tissue.

  Hereinafter, calculation of the expression intensity score of the molecular marker by immunostaining will be described with reference to FIG. FIG. 2 is a diagram for explaining calculation of an expression intensity score of a molecular marker by immunostaining according to the first embodiment. Here, in FIG. 2, the horizontal axis shows the staining intensity by immunostaining, and the vertical axis shows the histogram. Further, FIG. 2 shows an example in which the staining intensity is not divided into the above six steps but is further divided. For example, the pathologist counts the number of cells for each staining intensity while observing the immunostained tissue with a microscope. The inspection apparatus 300 receives a count by a pathologist and generates a histogram indicating the number of cells for each staining intensity as shown in FIG. And the inspection apparatus 300 calculates the expression intensity score of a molecular marker based on a histogram after completion | finish of the count by a pathologist. Furthermore, the inspection apparatus 300 transmits the calculated expression intensity score of the molecular marker to the medical information processing apparatus 400.

  When calculating the expression intensity score of the molecular marker in in situ hybridization, for example, the test apparatus 300 calculates negative expression of the molecular marker as “score: 0” and positive as “score: 1”. Further, when calculating the expression intensity score of a molecular marker in a technique based on RT-PCR, the test apparatus 300 calculates, for example, a score according to the ratio of the expression level of the molecular marker to the expression level of the control gene. . Alternatively, for example, the test apparatus 300 calculates negative of the expression of the molecular marker as “score: 0” and positive as “score: 1”. Then, the testing apparatus 300 transmits the calculated expression intensity score of the molecular marker to the medical information processing apparatus 400.

  Returning to FIG. 1, the medical information processing apparatus 400 includes an I / F (interface) circuit 410, an input circuit 420, a display 430, a storage circuit 440, and a processing circuit 450. For example, the medical information processing apparatus 400 is realized by a computer device such as a workstation.

  The I / F circuit 410 is connected to the processing circuit 450, and controls transmission and communication of various data performed between the medical image diagnostic apparatus 100, the server apparatus 200, and the examination apparatus 300 connected via a network. For example, the I / F circuit 410 is realized by a network card, a network adapter, a NIC (Network Interface Controller), or the like.

  The input circuit 420 is connected to the processing circuit 450, converts an input operation received from the operator into an electrical signal, and outputs the electrical signal to the processing circuit 450. For example, the input circuit 420 is realized by a switch button, a mouse, a keyboard, a touch panel, and the like.

  The display 430 is connected to the processing circuit 450 and displays various information and various medical images output from the processing circuit 450. For example, the display 430 is realized by a liquid crystal monitor, a CRT (Cathode Ray Tube) monitor, a touch panel, or the like.

  The storage circuit 440 is connected to the processing circuit 450 and stores various data. For example, the storage circuit 440 is realized by a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, a hard disk, an optical disk, or the like. In the present embodiment, the storage circuit 440 stores medical image data received from the medical image diagnostic apparatus 100 or the server apparatus 200, biopsy information received from the inspection apparatus 300, processing results by the processing circuit 450, and the like.

  The processing circuit 450 controls each component included in the medical information processing apparatus 400 in accordance with an input operation received from the operator via the input circuit 420. For example, the processing circuit 450 is realized by a processor. In the present embodiment, the processing circuit 450 causes the storage circuit 440 to store medical image data, biopsy information, processing results, and the like output from the I / F circuit 410. The processing circuit 450 reads various information, medical image data, and the like from the storage circuit 440, executes various processes, and displays the processing results on the display 430.

  The configuration of the medical information processing system including the medical information processing apparatus 400 has been described above. Based on such a configuration, the medical information processing apparatus 400 according to the present embodiment makes it possible to accurately predict a pathological examination result based on a medical image. Specifically, the medical information processing apparatus 400 generates and accumulates related information that associates the image feature amount of the tissue collected in the pathological examination with the examination result of the pathological examination of the tissue. Thereby, by referring to such related information, the examination result of the pathological examination can be accurately predicted from the image feature amount of the medical image.

  As illustrated in FIG. 1, the storage circuit 440 according to the present embodiment stores, for example, image information 441, biopsy information 442, and related information 443. In addition, the processing circuit 450 according to the present embodiment executes a control function 451, a first acquisition function 452, a second acquisition function 453, a calculation function 454, and a prediction function 455. Here, the control function 451 is an example of an association unit in the claims. Moreover, the 1st acquisition function 452 is an example of the 1st acquisition part in a claim. The second acquisition function 453 is an example of a second acquisition unit in the claims. The calculation function 454 is an example of an acquisition unit in the claims. The prediction function 455 is an example of a prediction unit in the claims.

  The image information 441 is medical image data or medical images collected by examination, and is acquired from the medical image diagnostic apparatus 100 or the server apparatus 200 by the processing circuit 450. The medical image data acquired by the processing circuit 450 may be three-dimensional medical image data (volume data) collected by the medical image diagnostic apparatus 100. Here, the image information 441 includes medical image data and medical images collected before and after biopsy. FIG. 3 is a diagram illustrating an example of the image information 441 according to the first embodiment. For example, the image information 441 includes a medical image of a lesion before biopsy and an image of a lesion after biopsy as shown in FIG. Here, in the medical image of the lesion before biopsy and the image of the lesion after biopsy, identification information for uniquely identifying the correspondence with the biopsy information (for example, patient ID or pathological examination ID) Are associated. That is, the image information 441 is information in which an image is associated with identification information.

  The image information 441 includes medical image data and medical images of a plurality of image types related to the same target part of the same subject. For example, when the medical images before and after the biopsy shown in FIG. 3 are CT images, the image information 441 may include MR images of the same lesion, ultrasonic images, PET images, and the like. That is, the subject collects CT images, MR images, ultrasonic images, PET images, and the like related to the same lesion before biopsy (before tissue collection), and CT images related to the same lesion after biopsy (after tissue collection). MR images, ultrasound images, PET images, etc. are collected. Further, when the medical images before and after the biopsy shown in FIG. 3 are CT images, it may include a CT image collected with X-ray energy different from the CT image shown in FIG.

  As described above, the storage circuit 440 stores medical image data before and after biopsy and image information 441 that is a medical image for each target region including a tissue on which a pathological examination is performed. In addition, the storage circuit 440 stores medical image data and medical images of a part where a pathological examination is not performed as the image information 441. For example, the storage circuit 440 stores medical image data and medical images of a part different from the lesion shown in FIG. 3 in the subject from which the medical images shown in FIG. 3 are collected. As an example, the storage circuit 440 stores a medical image including a portion suspected of having a tumor although a pathological examination is not performed.

  The biopsy information 442 is examination information of a pathological examination acquired by the inspection apparatus 300, and is acquired from the inspection apparatus 300 by the processing circuit 450. For example, the biopsy information 442 is information in which a morphological diagnosis result by cytodiagnosis examination or tissue examination is associated with identification information of a tissue subjected to pathological examination. The biopsy information 442 is information in which a test result of a molecular pathological examination (for example, an expression intensity score of a molecular marker) is associated with identification information of a tissue subjected to the examination.

  Here, the biopsy information 442 is associated with the examination results of all the pathological examinations performed for each piece of tissue identification information. For example, a tissue collected from a target site is often divided into a plurality of tissue pieces, and different pathological examinations are often performed on each tissue piece. For example, the expression score of a molecular marker may be calculated by dividing a tissue sampled from a target site into a plurality of tissue pieces and using methods based on immunostaining, in situ hybridization, and RT-PCR, respectively. is there. Therefore, in the biopsy information 442, the examination result of the pathological examination performed is associated with each collected tissue.

  The related information 443 is information in which the image feature amount of the tissue subjected to the pathological examination and the biopsy information are associated with each other, is generated by the processing of the processing circuit 450, and is stored in the storage circuit 440. FIG. 4 is a diagram illustrating an example of the related information 443 according to the first embodiment. For example, as shown in FIG. 4, the related information 443 is information in which “case”, “score (marker A)”, and “feature” are associated with each other. Here, the “case” indicates a disease symptom, a diagnosis result, or the like. The “score (marker A)” indicates the expression intensity score of the molecular marker calculated by the inspection apparatus 300. The “feature amount” indicates an image feature amount acquired from the image information by the processing circuit 450. Here, “feature amount A” to “feature amount F” in FIG. 4 are image feature amounts respectively acquired from images of different image types.

  For example, the storage circuit 440 reads “Case: 1, Score (Marker A): 4, Feature A: 1, Feature B: 1, Feature C: 5, Feature D: 2, Feature E: 0, The related information 443, which is “feature amount F: 1”, is stored. Here, the numerical value in each feature amount indicates a predetermined index value. Similarly, the storage circuit 440 stores related information 443 in which the expression intensity score of the molecular marker is associated with the image feature amount for each case. Note that the related information 443 illustrated in FIG. 4 is merely an example, and the associated image feature amount is not limited to the example of FIG. That is, not only a predetermined index value but also various statistics may be associated. FIG. 4 shows a case where one “score” is associated, but a case where a plurality of “scores” are associated may also be used.

  Returning to FIG. 1, the control function 451 in the processing circuit 450 performs overall control of the medical information processing apparatus 400. For example, the control function 451 causes the storage circuit 440 to store medical image data, medical images, biopsy information, and the like output from the I / F circuit 410. The control function 451 executes various processes in accordance with an input operation received from an operator via the input circuit 420. In addition, the control function 451 stores the related information 443 in the storage circuit 440 and displays various processing results by the processing circuit 450 on the display 430. Further, the control function 451 can also store the processing result by the processing circuit 450 in an external device such as the server device 200. For example, the control function 451 can also store the related information 443 in the server device 200.

  The first acquisition function 452 specifies the position of the tissue from the first medical image data that is an image of the target part before the tissue in the target part is collected, and acquires the image feature amount of the tissue. Specifically, the first acquisition function 452 compares the first medical image data with the second medical image data that is the image of the target region after the tissue is collected, and the first medical image data And an image feature amount of the tissue based on the change between the first medical image data and the second medical image data. For example, the first acquisition function 452 includes first medical image data that is an image of a target part before the tissue is collected in the target part, and a second image that is an image of the target part after the tissue is collected. An image feature amount of a tissue based on difference information with medical image data is acquired. That is, the first acquisition function 452 identifies the pixel corresponding to the tissue collected by the pathological examination on the image by subtracting the first medical image data and the second medical image data. Get the feature value of the pixel.

  Here, when a tissue is collected by pathological examination, distortion occurs in the target region. Therefore, the first acquisition function 452 performs alignment for correcting tissue strain that occurs after biopsy. Specifically, the first acquisition function 452 performs non-linear alignment between the first medical image data and the second medical image data, and then performs the first medical image data and the second medical image data. By calculating difference information from the medical image data, an image feature amount of the tissue is acquired. Note that the non-linear alignment performed by the first acquisition function 452 may be performed using any known technique.

  And the 1st acquisition function 452 specifies the pixel corresponding to the structure | tissue extract | collected by the biopsy by subtracting the 1st medical image data and 2nd medical image data after alignment. For example, the first acquisition function 452 performs alignment between the first medical image generated from the first medical image data and the second medical image generated from the second medical image data, After the alignment, the pixel value corresponding to the tissue collected by the biopsy is specified by subtracting the pixel value between each pixel in the first medical image and each corresponding pixel in the second medical image.

  Here, the first acquisition function 452 compares the difference value for each pixel corresponding to the first medical image and the second medical image with a threshold value, so that the pixel is applied to the tissue collected by biopsy. It is determined whether or not the corresponding pixel. That is, the pixel value corresponding to the tissue collected by biopsy has a large change in pixel value compared to other pixels. Therefore, the first acquisition function 452 determines a position where the difference information exceeds a predetermined threshold as a tissue position, and acquires an image feature amount of the tissue position. The threshold value may be arbitrarily set by the user, or may be automatically set based on pixel value change information before and after biopsy. Further, the threshold value may be set for each image type and each target region.

  Here, the pixel corresponding to the tissue collected by the biopsy may be specified after the alignment of the entire image as described above. However, the accuracy of the alignment is improved and the processing is performed. In order to shorten the time, the alignment may be performed only for a predetermined region including the tissue. In such a case, the first acquisition function 452 is between the first region including the tissue position in the first medical image data and the second region including the tissue position in the second medical image data. Then, non-linear alignment is performed, and difference information between the first region and the second region is calculated, thereby obtaining an image feature amount of the tissue.

  FIG. 5 is a diagram for explaining an example of the pixel specifying process corresponding to the tissue by the first acquisition function 452 according to the first embodiment. For example, as shown in FIG. 5, when the images before and after biopsy of image type A are aligned, first, the control function 451 displays the post-biopsy image collected after the tissue is collected by biopsy. 430 is displayed. The input circuit 420 receives a designation operation for the region R1 including the tissue from the user.

  The first acquisition function 452 aligns the region included in the region R1 received by the input circuit 420 with the pre-biopsy image, so that the region R2 is a region corresponding to the region R1 in the pre-biopsy image. To extract. Thereafter, the first acquisition function 452 collects the values of the pixels included in the region R1 from the pixel values of the corresponding pixels in the region R2, and compares the value after the difference with a threshold value, thereby collecting by biopsy. A pixel group corresponding to the specified tissue is specified.

  As described above, when the pixel group corresponding to the tissue collected by biopsy is specified, the first acquisition function 452 calculates the feature amount (image feature amount) of the pixel group corresponding to the tissue collected by biopsy. get. Here, the first acquisition function 452 can acquire an arbitrary feature amount as the image feature amount. For example, the first acquisition function 452 uses various statistical quantities such as an average pixel value and pixel value variation of the specified pixel group as an image feature amount, and an index represented by the sum of products of the pixel value and the appearance frequency. Calculate the value. In addition, the first acquisition function 452 can also calculate a feature amount indicating a change in contrast in the pixel group as an image feature amount. The first acquisition function 452 calculates the image feature amount described above for each image type. Note that the first acquisition function 452 can also calculate a plurality of feature amounts from one image type. Further, the number of feature amounts may be different depending on the image type.

  FIG. 6 is a diagram for explaining an example of feature amount calculation by the first acquisition function 452 according to the first embodiment. Here, in FIG. 6, differences in histograms of pixel values before and after biopsy (collected tissues) for four image types (MR T1-weighted image, MR T2-weighted image, CT image, and FDG-PET image). The case where the image feature amount is calculated from the pixel value profile) of FIG. Moreover, the upper stage in FIG. 6 shows the target site before and after biopsy in which images of four image types are collected. Each histogram shows, from the top, the MR T1-weighted image histogram, the MR T2-weighted image histogram, the CT image histogram, and the FDG-PET image histogram, and the left side shows the histogram before biopsy. The right side shows a histogram after biopsy.

  For example, the first acquisition function 452 generates the histogram shown in FIG. 6 for each image type, and extracts the pixel value of the pixel corresponding to the tissue collected by the biopsy from the change in the histogram before and after the biopsy. Thus, the image feature amount is calculated. For example, the first acquisition function 452 creates a histogram of pixel values of a lesion before biopsy. Further, the first acquisition function 452 generates a histogram of pixel values of other pixels excluding pixels corresponding to the tissue in the lesion after biopsy. Then, the first acquisition function 452 calculates the image feature amount from the change in the histogram before and after the biopsy. For example, in the case of a T1-weighted image (in the case of the feature amount A), as shown in FIG. 6, the pixel value of the tissue shows a relatively low value in the entire image. The first acquisition function 452 extracts the pixel value profile of such a tissue as an image feature amount for each image type.

  As described above, the first acquisition function 452 calculates an image feature amount for each image type. Here, the first acquisition function 452 can also correct the difference in imaging conditions for each image type. For example, when the first acquisition function 452 calculates an image feature amount from a predetermined CT image (CT images before and after biopsy), a CT image in which various parameters included in the imaging condition are acquired under the predetermined condition. After performing the correction process on the predetermined CT image so as to approximate, the image feature amount is calculated from the predetermined CT image. In other words, the first acquisition function 452 can also perform correction so as to align the imaging conditions for each image type when calculating the image feature amount. Thereby, in the related information 443 stored in the storage circuit 440, it is possible to store the image feature amount in which the difference due to the difference in the imaging conditions is eliminated in association with it. In addition, the predetermined conditions in various parameters can be set arbitrarily.

  Returning to FIG. 1, the second acquisition function 453 acquires the examination result of the pathological examination for the tissue. Specifically, the second acquisition function 453 acquires the examination information (biopsy information) of the pathological examination stored by the examination apparatus 300 and stores it in the storage circuit 440. For example, the second acquisition function 453 acquires a test result related to the expression of the molecular marker in the tissue.

  As described above, when the first acquisition function 452 acquires an image feature amount and the second acquisition function 453 acquires biopsy information, the control function 451 generates related information in which each piece of information is associated. The generated related information is stored in the storage circuit 440. For example, the control function 451 generates related information in which image feature amounts and biopsy information are associated with each other based on identification information such as a patient ID and a pathological examination ID, and stores the generated related information in the storage circuit 440. Store. For example, the control function 451 generates the related information 443 shown in FIG. 4 and stores it in the storage circuit 440.

  Here, the control function 451 sequentially generates and updates the related information every time the image feature amount and the biopsy information are acquired, and accumulates the related information. The related information may be appropriately stored not only in the storage circuit 440 but also in the server device 200 or the like.

  The calculation function 454 calculates an image feature amount in medical image data that is a target of pathological examination prediction. For example, the calculation function 454 calculates the image feature amount of the target part from the medical image data of the subject before the tissue is collected. Here, the calculation function 454 can calculate the same image feature amount as the first acquisition function 452 described above. That is, the calculation function 454 can calculate an image feature amount for each image type. For example, like the first acquisition function 452 described above, the calculation function 454 can calculate the image feature amount after performing correction that matches the imaging conditions for the medical image data to be predicted. . In addition, when performing the correction for aligning the imaging conditions, the correction is performed with the same correction content as that for calculating the image feature amount associated with the related information 443. In addition, medical image data that is a target of pathological examination is, for example, a medical image of a subject that is newly suspected to be a lesion in interpretation of a medical image, or a tissue that has undergone a pathological examination. Is a medical image or the like of a part (for example, metastasis) different from the target part from which the sample is collected.

  The prediction function 455 predicts a test result of a pathological examination in a part different from the target part based on the related information. Specifically, the prediction function 455 predicts a result when a pathological examination is performed on the part for which the calculation function 454 has calculated the image feature amount based on the related information stored by the storage circuit 440. Note that the prediction of prediction of the inspection result by the prediction function 455 is accurate when the correction for aligning the imaging conditions is performed when calculating the image feature amount at the time of generating the related information and the prediction of the inspection result. Will be improved. FIG. 7 is a diagram for explaining an example of the prediction process by the prediction function 455 according to the first embodiment. Here, in FIG. 7, it shows about the case where the expression prediction of a molecular marker is performed.

  For example, as illustrated in FIG. 7, when the user sets two ROIs for a captured image, the calculation function 454 calculates an image feature amount of each ROI. Then, the prediction function 455 performs molecular marker expression prediction of each ROI based on the image feature amount of each ROI calculated by the calculation function 454. For example, as illustrated in FIG. 7, the prediction function 455 calculates “marker expression prediction score: 4.2” for one ROI. Further, the prediction function 455 calculates a marker expression prediction score: 0.8 ”for the other ROI.

  Here, the prediction function 455 can perform prediction based on various analysis methods. That is, an analysis method according to the amount and quality of the related information 443 accumulated in the storage circuit 440 can be used as appropriate. In other words, the calculation function 454 calculates an image feature amount (for example, a predetermined image feature amount or a predetermined combination of image feature amounts) according to the accumulated related information 443, and the prediction function 455 calculates The result is predicted using the image feature amount.

  In the above-described example, the case where the user sets the ROI for the captured image has been described. However, the embodiment is not limited to this, and an area specified by CAD (Computer Aided Diagnosis) may be set as the ROI.

  The processing functions of the processing circuit 450 have been described above. Here, for example, each processing function described above is stored in the storage circuit 440 in the form of a program executable by a computer. The processing circuit 450 reads each program from the storage circuit 440 and executes each read program, thereby realizing a processing function corresponding to each program. In other words, the processing circuit 450 in a state where each program is read has the processing functions shown in FIG.

  In addition, although the example in case each processing function is implement | achieved by the single processing circuit 450 was demonstrated in FIG. 1, embodiment is not restricted to this. For example, the processing circuit 450 may be configured by combining a plurality of independent processors, and each processor may implement each processing function by executing each program. In addition, each processing function of the processing circuit 450 may be realized by being appropriately distributed or integrated into a single or a plurality of processing circuits.

  Next, a processing procedure performed by the medical information processing apparatus 400 according to the first embodiment will be described with reference to FIGS. 8 and 9. 8 and 9 are flowcharts showing a processing procedure performed by the medical information processing apparatus 400 according to the first embodiment. Note that FIG. 8 shows processing in a case where a region is specified in an image and a pixel corresponding to a tissue is specified by subtracting the specified region. Further, FIG. 8 shows processing after acquisition of medical image data, pathological examination, and diagnosis by a doctor are completed. Further, FIG. 9 shows a process in the case where an inspection result is predicted for an acquired image using the related information 443 accumulated in the storage circuit 440. Further, FIG. 9 shows a case where an area for predicting the inspection result is designated by the user.

  First, FIG. 8 will be described. Steps S101, S102, and S107 in FIG. 8 are realized, for example, when the processing circuit 450 calls and executes a program corresponding to the control function 451 and the second acquisition function 453 from the storage circuit 440. Further, steps S103 to S106 are realized, for example, when the processing circuit 450 calls and executes a program corresponding to the first acquisition function 452 from the storage circuit 440. Further, step S108 is realized, for example, when the processing circuit 450 calls and executes a program corresponding to the calculation function 454 and the prediction function 455 from the storage circuit 440.

  For example, as shown in FIG. 8, in the medical information processing apparatus 400 according to the present embodiment, first, the processing circuit 450 acquires images before and after biopsy and biopsy information (step S101). Then, the processing circuit 450 displays a medical image on the display 430, and accepts designation of a region including a part collected by biopsy via the input circuit 420 (step S102).

  Then, the processing circuit 450 aligns the images before and after the biopsy (step S103) and makes a difference between the regions (step S104). Further, the processing circuit 450 executes threshold processing (step S105), and specifies a pixel group corresponding to a tissue collected by biopsy. Thereafter, the processing circuit 450 calculates an image feature amount (step S106).

  Then, the processing circuit 450 generates related information in which the image feature amount and the biopsy information are associated with each other, and stores the related information in the storage circuit 440 (step S107). Thereafter, the processing circuit 450 predicts a test result of a biopsy of a part where biopsy is not performed based on the related information (step S108).

  Next, FIG. 9 will be described. Steps S201, S202, and S205 in FIG. 9 are realized, for example, when the processing circuit 450 calls and executes a program corresponding to the control function 451 from the storage circuit 440. Further, step S203 is realized, for example, when the processing circuit 450 calls and executes a program corresponding to the calculation function 454 from the storage circuit 440. S204 is realized, for example, when the processing circuit 450 calls and executes a program corresponding to the prediction function 455 from the storage circuit 440.

  For example, as illustrated in FIG. 9, in the medical information processing apparatus 400 according to the present embodiment, first, the processing circuit 450 acquires an image including a target portion for predicting a test result of a pathological examination (step S201). Here, the image including the target part for predicting the examination result of the pathological examination may be an image used for generating the related information 443, or a new image not used for generating the related information 443. It may be an image. In addition, the image including the target portion for predicting the examination result of the pathological examination may be an image collected from an arbitrary subject.

  Then, the processing circuit 450 displays the acquired image on the display 430, and accepts designation of an area via the input circuit 420 (step S202). Further, the processing circuit 450 calculates the image feature amount of the designated area (step S203).

  Then, the processing circuit 450 predicts the inspection result based on the image feature amount and the related information (step S204). Thereafter, the processing circuit 450 displays the predicted inspection result (prediction result) on the display 430 (step S108).

  In FIG. 9, the case where the medical information processing apparatus 400 calculates the image feature amount has been described. However, the embodiment is not limited thereto, and for example, the medical information processing apparatus 400 is calculated by another apparatus. The image feature amount may be acquired. In such a case, for example, the processing circuit 450 acquires an image feature amount calculated by another device and an image for which the image feature amount has been calculated. Then, the processing circuit 450 predicts the inspection result based on the acquired image feature amount and related information, and displays the prediction result together with the acquired image.

  As described above, according to the first embodiment, the first acquisition function 452 collects the first medical image data that is an image of the target part before the tissue in the target part is acquired, and the tissue. An image feature amount of the tissue based on the difference information with the second medical image data that is the image of the target region after the acquisition is acquired. The second acquisition function 453 acquires the examination result of the pathological examination for the tissue. The control function 451 stores related information in which the image feature amount of the tissue and the examination result of the pathological examination are associated with each other in the storage circuit 440. Therefore, the medical information processing apparatus 400 according to the first embodiment can accumulate the related information in which the image feature amount of the tissue subjected to the pathological examination is associated with the result of the pathological examination of the tissue, By using this related information, it is possible to accurately predict the examination result of the pathological examination based on the medical image.

  For example, in recent years, molecular targeted therapy using therapeutic drugs that act on specific proteins and genes (molecular markers) as targets has been performed as a treatment for cancer and the like, and it is expected to improve the therapeutic effect and reduce side effects Has been. In such molecular targeted therapy, therapeutic drugs targeting molecules (proteins, genes) that are specifically expressed in cancer cells are used. Here, when performing molecular target therapy, prior to the treatment, for example, it is confirmed whether or not a molecular marker serving as a target molecule of the therapeutic drug is expressed in cancer cells, and a therapeutic method (therapeutic drug) is determined according to the expression state. ) Is selected. Here, if a biopsy cannot be performed due to the physical reason of the subject or the location of the disease, it cannot be confirmed whether or not the molecular marker that is the target molecule of the therapeutic agent is expressed. , Effective molecular targeted therapy may be difficult to perform. Even in such a case, the medical information processing apparatus 400 according to the present application can predict the expression state of the molecular marker from the medical image of the subject, and as a result, enables effective molecular target therapy. .

  Further, according to the first embodiment, the first acquisition function 452 performs the first medical image data after performing non-linear alignment between the first medical image data and the second medical image data. By calculating difference information between the image data and the second medical image data, the image feature amount of the tissue is acquired. Therefore, the medical information processing apparatus 400 according to the first embodiment can specify the pixel corresponding to the tissue after correcting the distortion of the region due to the tissue sampling, and more accurately determine the position of the collected tissue. Make it possible to identify.

  In addition, according to the first embodiment, the first acquisition function 452 includes the first region including the tissue position in the first medical image data and the first region including the tissue position in the second medical image data. The non-linear alignment is performed between the two regions and the difference information between the first region and the second region is calculated, thereby acquiring the image feature amount of the tissue. Therefore, the medical information processing apparatus 400 according to the first embodiment can perform alignment within a narrower range, improve alignment accuracy, and shorten processing time.

  Further, according to the first embodiment, the first acquisition function 452 determines a position where the difference information exceeds a predetermined threshold as a tissue position, and acquires an image feature amount of the tissue position. Therefore, the medical information processing apparatus 400 according to the first embodiment can more accurately specify the position of the collected tissue.

  Further, according to the first embodiment, the first acquisition function 452 acquires image feature amounts of tissues based on difference information in different image types. The control function 451 stores, in the storage circuit 440, related information in which the examination result of the pathological examination is further associated with the image feature amount of the tissue in different image types. Therefore, the medical information processing apparatus 400 according to the first embodiment can generate related information including a large amount of information, and can perform highly accurate prediction.

  Further, according to the first embodiment, the second acquisition function 453 acquires a test result related to the expression of the molecular marker in the tissue. The control function 451 stores related information in which the image feature amount of the tissue is associated with the test result regarding the expression of the molecular marker in the tissue in the storage circuit 440. Therefore, the medical information processing apparatus 400 according to the first embodiment can accumulate information on the expression of the molecular marker, and can predict the expression of the molecular marker.

  Further, according to the first embodiment, the prediction function 455 predicts a test result of a pathological examination in a part different from the target part based on the related information stored in the storage circuit 440. Therefore, the medical information processing apparatus 400 according to the first embodiment makes it possible to predict the result of the pathological examination using the related information.

(Second Embodiment)
In the above-described embodiment, the case where images before and after biopsy are collected for each image type has been described. In the second embodiment, a case will be described in which images before and after biopsy are collected with one image type, and only images before biopsy are collected with other image types. Note that the medical information processing apparatus 400 according to the second embodiment differs from the medical information processing apparatus 400 according to the first embodiment in the processing content by the first acquisition function 452. Hereinafter, this will be mainly described.

  The first acquisition function 452 according to the second embodiment is an image of a target region before a tissue of an image type different from the first medical image data and the second medical image data is collected. By aligning the medical image data and the first medical image data, the position of the tissue in the third medical image data is specified, and the image feature quantity of the tissue position in the third medical image data is acquired. To do. FIG. 10 is a diagram for explaining an example of the pixel specifying process corresponding to the tissue by the first acquisition function 452 according to the second embodiment. Here, FIG. 10 shows processing after medical image data before and after biopsy is collected with the image type A, and a pixel group corresponding to a biopsy collection site is specified by alignment between regions.

  For example, the first acquisition function 452 corresponds to the biopsy collection site on each image by aligning the pre-biopsy image of the image type A and the pre-biopsy images of the image types B to F, respectively. A pixel group to be specified is specified. Here, also in the above-described alignment between images before biopsy, alignment including distortion correction is performed in order to eliminate the influence of respiration and peristalsis. For example, the first acquisition function 452 aligns the image types B to F with the region R2 of the image type A through nonlinear alignment processing. Thereby, the 1st acquisition function 452 can each extract the pixel group corresponding to the pixel group of the biopsy collection site | part of the image type A in each image of image types BF.

  Since the image types to be aligned are different, the meanings of the pixel values are different. Therefore, the first acquisition function 452 performs alignment processing in consideration of the relationship between pixel values of different types of images. For example, the first acquisition function 452 can perform the alignment process using a method for calculating the mutual image information amount. In addition, when the image type is different, the pixel size may be different. Therefore, the first acquisition function 452 performs, for example, an interpolation process for matching a pixel having a large pixel size with a pixel having a small pixel size.

  As described above, according to the second embodiment, the first acquisition function 452 uses the target region before the tissue of the image type different from the first medical image data and the second medical image data is collected. The position of the tissue in the third medical image data is specified by aligning the third medical image data and the first medical image data, which are the images of the first medical image data, and the tissue in the third medical image data. The image feature amount of the position is acquired. The control function 451 stores, in the storage circuit 440, related information in which the examination result of the pathological examination is further associated with the image feature quantity of the tissue position in the third medical image data. Therefore, the medical information processing apparatus 400 according to the second embodiment can reduce the number of images collected after biopsy, and can reduce the burden on the subject.

(Third embodiment)
The first embodiment has been described so far, but may be implemented in various different forms other than the first and second embodiments described above.

  In the above-described embodiment, a case has been described in which a pixel group corresponding to a tissue collected by biopsy is specified by subtracting pixels of a medical image. However, the embodiment is not limited to this. For example, when three-dimensional medical image data (volume data) is collected, a voxel group corresponding to the tissue is obtained by subtracting voxel values of the volume data. May be specified. In this case, the tissue collected by biopsy can be specified in three dimensions.

  In the above-described embodiment, the case where the expression of the molecular marker is predicted has been described. However, the embodiment is not limited to this. For example, the metabolic information of a polymer that can be acquired using a PET apparatus is acquired as an image feature amount and stored in association with the diagnosis result of a pathological examination. May be. In such a case, the first medical image data and the second medical image data are PET images, and the first acquisition function 452 uses the difference information between the first medical image data and the second medical image data. Based on this, an image feature amount based on the dynamics of the tracer in the tissue is acquired.

  For example, the first acquisition function 452 acquires, as an image feature amount, the tracer dynamics in a pixel group from which a tissue is collected in a PET image collected using a protein specifically metabolized in a predetermined cancer as a tracer. To do. Then, the control function 451 stores in the storage circuit 440 related information that associates the collected tissue diagnosis and the tracer dynamics acquired as the image feature amount. The calculation function 454 obtains the dynamics of the tracer at a site where no tissue is collected. The prediction function 455 refers to the related information to determine whether or not a site from which tissue is not collected from the tracer dynamics is a predetermined cancer.

  Further, in the above-described embodiment, the case where the medical information processing apparatus 400 performs acquisition of image feature amounts, storage of related information, and prediction processing has been described as an example. However, the embodiment is not limited to this. For example, the above-described processing may be executed in a distributed manner by a plurality of devices included in the medical information processing system. For example, the image feature amount is calculated in each device included in the medical image diagnostic apparatus 100, and the medical information processing apparatus 400 acquires biopsy information from the inspection apparatus and acquires the image feature amount from the medical image diagnostic apparatus 100. Thus, the related information may be generated.

  The term “processor” used in the description of each embodiment described above is, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or an application specific integrated circuit (ASIC). Circuits such as programmable logic devices (for example, Simple Programmable Logic Device (SPLD), Complex Programmable Logic Device (CPLD), and Field Programmable Gate Array (FPGA)) Means. Here, instead of storing the program in the storage circuit, the program may be directly incorporated in the circuit of the processor. In this case, the processor realizes the function by reading and executing the program incorporated in the circuit. In addition, each processor of the present embodiment is not limited to being configured as a single circuit for each processor, but may be configured as a single processor by combining a plurality of independent circuits to realize its function. Good.

  Here, the program executed by the processor is provided by being incorporated in advance in a ROM (Read Only Memory), a storage unit, or the like. This program is a file in a format installable or executable in these apparatuses, such as a CD (Compact Disk) -ROM, an FD (Flexible Disk), a CD-R (Recordable), a DVD (Digital Versatile Disk), or the like. It may be provided by being recorded on a computer-readable storage medium. The program may be provided or distributed by being stored on a computer connected to a network such as the Internet and downloaded via the network. For example, this program is composed of modules including each functional unit. As actual hardware, the CPU reads a program from a storage medium such as a ROM and executes it, whereby each module is loaded on the main storage device and generated on the main storage device.

  According to at least one embodiment described above, it is possible to accurately predict a pathological examination result based on a medical image.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and the equivalents thereof.

400 Medical Information Processing Device 450 Processing Circuit 451 Control Function 452 First Acquisition Function 453 Second Acquisition Function 454 Calculation Function 455 Prediction Function

Claims (13)

A first acquisition unit that specifies a position of the tissue from first medical image data that is an image of the target region before the tissue in the target region is collected, and acquires an image feature amount of the tissue;
A second acquisition unit that acquires a test result of a pathological test on the tissue;
An associating unit associating the image feature amount of the tissue with the examination result of the pathological examination;
A medical information processing system comprising:   The first acquisition unit compares the first medical image data with the second medical image data that is an image of the target region after the tissue is collected. The medical information processing system according to claim 1, wherein an image feature amount of the tissue based on a change from the second medical image data is acquired.   The first acquisition unit is configured based on difference information between the first medical image data and second medical image data that is an image of the target region after the tissue is collected. The medical information processing system according to claim 2, wherein:   The first acquisition unit performs non-linear alignment between the first medical image data and the second medical image data, and then performs the first medical image data and the second medical image. The medical information processing system according to claim 3, wherein an image feature amount of the tissue is acquired by calculating difference information with image data.   The first acquisition unit includes a first region including the position of the tissue in the first medical image data and a second region including the position of the tissue in the second medical image data. The medical information processing system according to claim 3, wherein non-linear alignment is performed, and difference information between the first region and the second region is calculated to acquire an image feature amount of the tissue.   The said 1st acquisition part determines the position where the said difference information exceeded the predetermined threshold value as the position of the said structure | tissue, and acquires the image feature-value of the position of the said structure | tissue. Medical information processing system described in 1. The first acquisition unit acquires image feature amounts of the tissue based on difference information between the first medical image data and the second medical image data in different image types,
The medical information processing system according to any one of claims 3 to 6, wherein the association unit further associates an image feature amount of the tissue in the different image types with an examination result of the pathological examination. The first acquisition unit is a third medical image data which is an image of the target region before the tissue of an image type different from the first medical image data and the second medical image data is collected. And the position of the tissue in the third medical image data are specified by aligning the first medical image data with the first medical image data, and the image feature amount of the position of the tissue in the third medical image data is determined. Acquired,
The medical information processing according to any one of claims 3 to 6, wherein the associating unit further associates an image feature amount of the tissue position in the third medical image data with an examination result of the pathological examination. system. The second acquisition unit acquires a test result relating to expression of a molecular marker in the tissue,
The medical information processing system according to any one of claims 3 to 8, wherein the association unit associates an image feature amount of the tissue with a test result regarding expression of a molecular marker in the tissue. The first medical image data and the second medical image data are PET (Positron Emission Tomography) images,
The first acquisition unit acquires an image feature amount based on a dynamic of a tracer in the tissue based on difference information between the first medical image data and the second medical image data,
The medical information processing system according to any one of claims 3 to 6, wherein the association unit associates an image feature amount based on a tracer dynamics in the tissue with an examination result of the pathological examination.   The image processing apparatus according to claim 1, further comprising: a prediction unit that predicts the examination result of the pathological examination in a part different from the target part based on related information in which the image feature amount of the tissue and the examination result of the pathological examination are associated. The medical information processing system according to any one of 10 to 10. A first acquisition unit that specifies a position of the tissue from first medical image data that is an image of the target region before the tissue in the target region is collected, and acquires an image feature amount of the tissue;
A second acquisition unit that acquires a test result of a pathological test on the tissue;
An associating unit associating the image feature amount of the tissue with the examination result of the pathological examination;
A medical information processing apparatus comprising: An acquisition unit that acquires an image feature amount of a target tissue included in the medical image data;
A prediction unit that predicts a pathological examination test result for the target tissue from an image feature quantity of the target tissue based on related information in which an image feature quantity of the tissue and a pathological examination test result for the tissue are associated;
An output control unit that controls to output a prediction result for the target tissue;
A medical information processing system comprising:

Images