JP2018128933A - Diagnosis support device, diagnosis support method, and diagnosis support program - Google Patents

Abstract

CONSTITUTION: An image diagnosis support device 10 is a device that supports diagnosis based on a request comment and a plurality of medical images created for a common patient, and a CPU 12pr executes the following processing; first, at least one of the plurality of medical images is displayed on a sub monitor 12m2; an opinion of a radiologist on the request comment is acquired through a keyboard/mouse 12km and displayed on an X-ray interpretation report screen on a main monitor 12m1; subsequently, a candidate disease presentation request attached with the plurality of medical images and a candidate disease presentation request attached with the request comment and opinion are issued toward a machine learning system 40; a result of analysis transmitted back from the special machine learning system 40 is displayed multiply on the X-ray interpretation report screen.EFFECT: Diagnosis support performance is improved.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a diagnosis support apparatus, a diagnosis support method, and a diagnosis support program, and particularly supports diagnosis of a patient in cooperation with an analysis system that analyzes a patient's disease based on the medical image of the patient and the opinion of the diagnostician. The present invention relates to a diagnosis support apparatus, a diagnosis support method, and a diagnosis support program.

  In a medical facility, an interpreting doctor analyzes a plurality of medical images output from a modality for photographing a patient and creates an interpretation report. In the interpretation report, a medical image representing a lesion is attached, and the name of a disease diagnosed from the attached medical image, a treatment method, and the like are described.

  In order to improve the diagnostic accuracy of the interpretation doctor, that is, the reliability of the interpretation report, it is necessary to accurately perform both “existence diagnosis” for diagnosing the presence of a lesion and “qualitative diagnosis” for diagnosing the nature of the lesion. However, the number of examinations and the number of images are increasing due to the recent improvement in the performance of CT apparatuses and MRI apparatuses, which increases the burden on the interpreting physician. Such an increase in burden causes oversight of lesions, input errors, forgetting to write, etc., so it is noticed how to support the diagnosis of an interpreting doctor.

  In this connection, Patent Document 1 discloses disease candidate information that performs detection of an abnormal shadow and search for similar cases based on image data to be diagnosed and disease state information, and displays a disease candidate name having a high possibility. An output system is disclosed.

  In addition, Patent Document 2 calculates a point for each disease name using a point map when a plurality of pieces of disease state information are input, and representative points are associated with the disease names in order of disease names having the highest points. A medical diagnosis support apparatus that displays an image is disclosed.

Japanese Patent No. 4087640 JP 2009-59381 A

  In both Patent Documents 1 and 2, image diagnosis is performed by cooperation of an analysis system that analyzes a patient's disease based on image data and / or disease state information to be diagnosed. Here, in order to increase the dependency of the analysis system in the image diagnosis, an improvement in diagnosis accuracy by the analysis system is required. However, the design of an analysis system for improving diagnosis accuracy may induce misdiagnosis and thus reduce diagnosis support performance.

  Therefore, a main object of the present invention is to provide a diagnosis support apparatus capable of improving support performance.

  A diagnosis support apparatus according to the present invention (10: reference numerals corresponding to the embodiments; the same applies hereinafter) is a diagnosis support apparatus that supports diagnosis based on a plurality of medical images and request sentences created for a common patient. Therefore, medical image output means (S9, S25 to S27) for outputting at least one of a plurality of medical images to the diagnostician, and obtaining the diagnostician's opinion regarding the request text in relation to the processing of the medical image output means Obtaining means (S11 to S17), first analysis requesting means (S41, S45) for requesting the analysis apparatus (40) a first analysis process for analyzing a patient's disease based on image information including a plurality of medical images, Second analysis requesting means for requesting the analyzer to perform second analysis processing for analyzing a patient's disease based on the character information including the request sentence and the opinion acquired by the acquiring means in relation to the processing of the first analysis requesting means. (S39, S43) and the first analysis obtained by the first analysis process Fruit and second analysis result and outputs the analysis result obtained by the second analysis process comprising output means (S47, S49).

  Preferably, the first analysis process is executed based only on the former of the image information and the character information, and the second analysis process is executed based only on the latter of the image information and the character information.

  Preferably, the analysis result output means simultaneously outputs the first analysis result and the second analysis result toward the diagnostician.

  Preferably, the medical image output means includes update means (S25, S27) for updating the output object in response to the update operation, and the analysis result output means is the update means and the acquisition unit after outputting the first analysis result and the second analysis result. Restart at least one of the means.

  Preferably, an interpretation report display means (S7) for displaying an interpretation report having a request text field (REQ) and an opinion entry field (RMK, DGN) on the first monitor (12m1) is further provided, and the obtaining means views the opinion. Includes opinion display means (S13, S17) to be displayed in the entry field.

  In one aspect, the medical image output means includes medical image display means (S9, S27) for displaying at least one of a plurality of medical images on a second monitor (12m2) associated with the first monitor.

  In another aspect, the analysis result output means includes additional display means (S49) for additionally displaying the first analysis result and the second analysis result on the first monitor.

  Preferably, each of the first analysis request unit and the second analysis request unit issues an analysis request in a different manner depending on the patient's attribute.

  Preferably, each of the first analysis request unit and the second analysis request unit issues an analysis request to the analysis device when a request issuance operation is accepted.

  Preferably, each of the first analysis result and the second analysis result includes a plurality of candidate disease names and a corresponding probability of each disease candidate.

  More preferably, it further comprises an emphasizing means (S57) for emphasizing candidate disease names satisfying a predetermined condition among a plurality of candidate disease names, and the leakage condition that the predetermined condition is missing from the opinion acquired by the acquisition means, And a probability condition that the corresponding probability exceeds the standard.

  A diagnosis support method according to the present invention is a diagnosis support method executed by a diagnosis support apparatus (10) for supporting diagnosis based on a plurality of medical images and request sentences created for a common patient. A medical image output step (S9, S25 to S27) for outputting at least one of the medical images to the diagnostician, and an acquisition step for acquiring the diagnostician's view of the request text in relation to the processing of the medical image output step ( S11 to S17), a first analysis requesting step (S41, S45) for requesting the analysis device (40) to perform a first analysis process for analyzing a patient's disease based on image information including a plurality of medical images, and obtaining a request sentence A second analysis requesting step (S39, S43) for requesting a second analysis process for analyzing the patient's disease based on the character information including the opinion acquired by the step in relation to the process of the first analysis requesting step to the analyzing apparatus. ) And the first analysis process It comprises outputting a second analysis result obtained by the first analysis result and the second analysis obtained I analysis result output step (S47, S49).

  The diagnosis support program according to the present invention includes a plurality of medical images stored in a processor (12pr) of a diagnosis support apparatus (10) that supports diagnosis based on a plurality of medical images and request sentences created for a common patient. A medical image output step (S9, S25 to S27) for outputting at least one to the diagnostician, and an acquisition step (S11 to S17) for acquiring the diagnostician's view on the request text in relation to the processing of the medical image output step The first analysis requesting step (S41, S45) for requesting the analysis device (40) to perform a first analysis process for analyzing a patient's disease based on image information including a plurality of medical images, obtained by a request sentence and an acquisition step. A second analysis request step (S39, S43) for requesting the analysis apparatus to perform a second analysis process for analyzing a patient's disease based on the character information including the received opinion in relation to the process of the first analysis request step; Obtained by one analysis process 1 the analysis results and for executing the analysis result output step (S47, S49) for outputting a second analysis result obtained by the second analysis processing, a diagnosis support program.

  By using a plurality of medical images as the target of the first analysis process, it is possible to reduce a concern that a medical image in which a lesion appears is overlooked and diagnosis is performed. In addition, by making the request text and the opinion of the diagnostician with respect to this request the subject of the second analysis process, diagnosis can be performed without paying attention to some diseases (that is, without noticing some of the disease names). Concern is reduced. This improves the diagnosis support performance.

  The above object, other objects, features and advantages of the present invention will become more apparent from the following detailed description of embodiments with reference to the drawings.

It is a block diagram which shows the structure of the medical-diagnosis assistance system applied to this Example. It is an illustration figure which shows an example of a structure of the table referred by a machine learning system. It is an illustration figure which shows an example of the structure of the DICOM file produced by modality. It is an illustration figure which shows an example of the diagnostic assistance screen displayed on the main monitor of an image diagnosis assistance apparatus. It is an illustration figure which shows another example of the diagnostic assistance screen displayed on the main monitor of an image diagnosis assistance apparatus. It is an illustration figure which shows an example of the diagnostic assistance screen displayed on the sub monitor of an image diagnosis assistance apparatus. It is an illustration figure which shows another example of the diagnostic assistance screen displayed on the main monitor of an image diagnosis assistance apparatus. It is an illustration figure which shows another example of the diagnosis assistance screen displayed on the main monitor of an image diagnosis assistance apparatus. It is an illustration figure which shows another example of the diagnostic assistance screen displayed on the main monitor of an image diagnosis assistance apparatus. It is an illustration figure which shows another example of the diagnostic assistance screen displayed on the main monitor of an image diagnosis assistance apparatus. It is an illustration figure which shows another example of the diagnosis assistance screen displayed on the main monitor of an image diagnosis assistance apparatus. It is a flowchart which shows a part of operation | movement of CPU provided in the image diagnosis assistance apparatus. It is a flowchart which shows a part of other operation | movement of CPU provided in the image diagnosis assistance apparatus. It is a flowchart which shows a part of other operation | movement of CPU provided in the image diagnosis assistance apparatus. It is a flowchart which shows further another part of operation | movement of CPU provided in the image diagnosis assistance apparatus. It is a flowchart which shows a part of other operation | movement of CPU provided in the image diagnosis assistance apparatus.

  Referring to FIG. 1, the medical diagnosis support system according to this embodiment includes an image diagnosis support device 10, a modality 20, a PACS 30, and a machine learning system (analysis device) 40 connected to each other via a hospital LAN 50.

  In the diagnostic imaging support apparatus 10, a communication I / F 12cm, a CPU 12pr, a keyboard / mouse 12km, a DRAM 12mm, an HDD 12hd, a main monitor (first monitor) 12m1, and a sub-monitor (second monitor) 12m2 are connected to the bus BS1.

  The sub monitor 12m2 is attached to the main monitor 12m1. An interpreting physician (diagnostic) faces the main monitor 12m1 and the sub monitor 12m2, and operates the keyboard / mouse 12km to perform image diagnosis.

  The machine learning system 40 includes a text specialized machine learning system 40at for adults, an image specialized machine learning system 40ai for adults, a text specialized machine learning system 40ct for children, and an image specialized machine learning system 40ci for children. . The disease candidate table shown in FIG. 2 is provided in the machine learning system 40 so as to be accessible from both the text specialized machine learning systems 40at and 40ct and the image specialized machine learning systems 40ai and 40ci.

  According to FIG. 2, “lung cancer”, “metastatic lung tumor”, “mediastinal tumor”, “pulmonary artery thrombosis”, “pulmonary edema”, “pleuritis”, “pyothorax”, “pneumothorax”, “chronic obstructive” Numerous character strings such as “pulmonary disease”, “pulmonary tuberculosis”, “interstitial pneumonia”, “sarcoidosis”, “pneumoconiosis”, “lymphangiomyomatosis”, “acute respiratory failure”, “bronchiectasis” Registered in the candidate table.

  The modality 20 is, for example, an MRI apparatus that outputs a DICOM file containing a plurality of medical images representing a patient's body axis section. The laboratory engineer who operates the modality 20 operates the modality 20 based on the request comment from the attending physician. As a result, a tomographic image at a desired position is created as a medical image, and the created medical image is stored in a DICOM file. The DICOM file created in this way has the data structure shown in FIG. 3 and is stored in the PACS 30 via the hospital LAN 50.

  Referring to FIG. 3, the DICOM file stores attached information in addition to a plurality of medical images. Here, the attached information includes a plurality of items (examination date, patient name, age at examination of the patient, accession number, etc.) each indicating examination attributes.

  The CPU 12pr provided in the image diagnosis support apparatus 10 executes the diagnosis support process shown in FIGS. 12 to 16 in order to support the image diagnosis by the interpretation doctor in cooperation with the machine learning system 40. A program corresponding to this flowchart is stored in the HDD 12hd as a diagnosis support program.

  Referring to FIG. 12, in step S1, the top screen is developed on DRAM 12mm. The developed top screen is read by the main monitor 12m1 and displayed as shown in FIG. According to FIG. 4, the top screen has a heading column IDX in which a plurality of headings each representing a plurality of examinations are listed. On the top screen shown in FIG. 4, the interpretation report corresponding to the headline is also “held”.

  In step S3, it is determined whether or not a headline selection operation (= operation for selecting any one of a plurality of headlines listed in the headline column IDX) has been performed with the keyboard / mouse 12km. In step S5, it is determined whether or not another operation has been performed with the keyboard / mouse 12km. If both the determination result of step S3 and the determination result of step S5 are NO, the process returns to step S3. If the determination result of step S3 is YES, the process proceeds to step S7. Proceed to the process.

  In step S7, an interpretation report screen corresponding to the headline selected by the headline selection operation is developed on the DRAM 12mm. The top screen on the DRAM 12 mm is updated by the interpretation report screen.

  The updated interpretation report screen is read by the main monitor 12m1 and displayed in the manner shown in FIG. According to FIG. 5, the interpretation report screen is provided with a request comment field (request text field) REQ, a finding field RMK, a diagnosis field DGN, and an attached image field IMG. Among these, the character string of “chest pain” in the request comment field REQ is described by the attending physician. Further, in this embodiment, the finding column RMK and the diagnosis column DGN are collectively referred to as an “opinion entry column”, and the findings described in the finding column RMK and the diagnosis comments described in the diagnosis column DGN are collectively referred to as “opinions”.

  In step S9, the PACS 30 is accessed through the communication I / F 12cm, and a DICOM file corresponding to the accession number assigned to the selected heading is acquired from the PACS 30. In step S9, any one of a plurality of medical images stored in the acquired DICOM file is developed on the DRAM 12mm. The developed medical image is read by the sub monitor 12m2 and displayed in the manner shown in FIG.

  In step S11, it is determined whether or not a finding input operation (an operation for inputting a desired character string by specifying the finding column RMK) has been performed with the keyboard / mouse 12km. If a determination result is NO, it will progress to step S15 as it is.

  On the other hand, if a determination result is YES, it will progress to step S13 and will display the character string acquired by finding input operation in the finding column RMK in the way shown in FIG. According to FIG. 7, in the finding column RMK, a character string “Recognizing calcification at the coronary artery origin” is displayed. When the display of the character string is completed, the process proceeds to step S15.

  In step S15, it is determined whether or not a diagnostic comment input operation (operation for specifying a diagnostic column DGN and inputting a desired character string) has been performed with the keyboard / mouse 12km. If a determination result is NO, it will progress to step S19 as it is.

  On the other hand, if a determination result is YES, it will progress to step S17 and will display the character string acquired by diagnostic comment input operation in the diagnostic column DGN in the way shown in FIG. According to FIG. 8, a character string “1. Suspected angina” is displayed in the diagnosis column DGN. When the display of the character string is completed, the process proceeds to step S19.

  In step S19, it is determined whether or not an image attaching operation (operation for moving a medical image displayed on the sub monitor 12m2 to the attached image column IMG by a drag and drop operation) has been performed with the keyboard / mouse 12km. If the determination result is NO, the process proceeds directly to step S25, while if the determination result is YES, the following processing is executed in step S21 and step S23, and then the process proceeds to step S31.

  That is, in step S21, a thumbnail image is created based on the medical image displayed on the sub monitor 12m2. In step S23, the created thumbnail image is displayed in the attached image column IMG (see FIGS. 7 and 8).

  In step S25, it is determined whether or not an image update operation (an operation for updating a medical image on the sub monitor 12m2) has been performed with the keyboard / mouse 12km. If a determination result is NO, it will progress to step S29 as it is.

  On the other hand, if a determination result is YES, it will progress to step S27 and will expand | deploy another medical image in the DICOM file acquired by step S9 to DRAM12mm. As a result, the medical image displayed on the sub monitor 12m2 is updated to another medical image. When the process of step S27 is completed, the process proceeds to step S29.

  In step S29, it is determined whether or not the “Confirm” button (see FIGS. 7 and 8) displayed in the vicinity of the attached image field IMG has been clicked by the keyboard / mouse 12km (whether or not a request issuing operation has been performed). To do. If the determination result in step S29 is NO, it is determined in step S31 whether or not the “confirm” button displayed on the upper right of the interpretation report screen has been clicked with the keyboard / mouse 12km, and other operations are performed with the keyboard / mouse 12km. It is determined in step S35 whether or not.

  If both the determination result in step S31 and the determination result in step S35 are NO, the process returns to step S11. If the determination result in step S35 is YES, the process proceeds to another process. If the determination result in step S31 is YES, the process returns to step S11. After the confirmation process is performed in step S33, the process returns to step S1.

  As a result, the display on the main monitor 12m1 returns from the interpretation report screen to the top screen. However, as indicated by bold circles in FIG. 9, the text corresponding to the interpretation report that has been subjected to the confirmation process is given a character string of “confirmed” instead of “hold”.

  If the determination result of step S29 is YES, it will be discriminate | determined by step S37 whether the age at the time of the examination described in the interpretation report screen is 15 years old or more. If the determination result is YES, the process proceeds to step S39 while if the determination result is NO, the process proceeds to step S43.

  In step S39, a candidate disease presentation request (analysis request) is transmitted to the text specialized machine learning system 40at, and in subsequent step S41, a candidate disease presentation request is transmitted to the image specialized machine learning system 40ai. In step S43, the candidate disease presentation request is transmitted to the text specialized machine learning system 40ct. In the subsequent step S45, the candidate disease presentation request is transmitted to the image specialized machine learning system 40ci. The candidate disease presentation request is issued through the communication I / F 12 cm and in different modes depending on the patient attributes.

  The candidate disease presentation request transmitted in step S39 or S43 is accompanied by the same character string as the character string displayed in the request comment field REQ as character information representing the request comment, and the character string displayed in the observation field RMK. The same character string is attached as character information representing a finding, and the same character string as the character string displayed in the diagnosis column DGN is attached as character information representing a diagnostic comment.

  In addition, all the medical images in the DICOM file acquired in step S9 are attached to the candidate disease presentation request transmitted in step S41 or S45 as image information.

  Each of the text specialized machine learning systems 40at and 40ct has artificial intelligence similar to that of a doctor who has abundant experience in diagnosing a patient's disease based on requested comments, findings, and diagnostic comments. Each of the image specialized machine learning systems 40ai and 40ci has artificial intelligence similar to that of a doctor who has abundant experience in diagnosing a patient's disease based on a medical image.

  Each of the text specialized machine learning systems 40at and 40ct analyzes the patient's disease based on the character information accompanying the candidate disease presentation request. On the other hand, each of the image specialized machine learning systems 40ai and 40ci analyzes the patient's disease based on the image information accompanying the candidate disease presentation request. That is, each of the text specialized machine learning systems 40at and 40ct executes an analysis process (first analysis process) based only on the former of character information and image information, and each of the image specialized machine learning systems 40ai and 40ci An analysis process (second analysis process) is executed based on only the latter of the character information and the image information.

  Each of the text specialized machine learning systems 40at and 40ct extracts a plurality of candidate diseases from the candidate disease table shown in FIG. 2 by an analysis process based on character information, and extracts the plurality of disease candidates and the probability corresponding to each disease candidate, Returns an answer containing Similarly, each of the image specialized machine learning systems 40ai and 40ci also extracts a plurality of candidate diseases from the candidate disease table shown in FIG. 2 by analysis processing based on image information, and corresponds to the extracted plurality of disease candidates and each disease candidate. The response including the probability of being returned is returned to the request source.

  When the processing of step S41 or S45 is completed, it is determined whether or not there is a response from both the text specialized machine learning system 40at and the image specialized machine learning system 40ai or from both the text specialized machine learning system 40ct and the image specialized machine learning system 40ci. And repeatedly determine.

  When the determination result is updated from NO to YES, the process proceeds to step S49, and the analysis result screen is developed on the DRAM 12mm. The developed analysis result screen is read by the main monitor 12m1, and is displayed in multiple (additional display) on the interpretation report screen in the manner shown in FIG.

  Referring to FIG. 10, the candidate disease name and the corresponding probability representing the analysis result based on the image information are displayed on the left side of the analysis result screen, and the candidate disease name and the corresponding probability representing the analysis result based on the character information are displayed on the analysis result screen. Displayed on the right side. The analysis result based on the image information and the analysis result based on the character information are displayed simultaneously.

  Here, the analysis results based on the image information are arranged in the order of “pleurisy 80%”, “anginal 50%”, “acute myocarditis 15%”, and “lung cancer 10%”. The analysis results based on the character information are arranged in the order of “angina 80%”, “myocardial infarction 40%”, “aortic valve disease 30%”, “acute myocarditis 10%”.

  In step S51, a candidate disease name leaking from the diagnosis column DGN is searched from the analysis result screen. In step S53, it is determined whether or not the search result indicates “detection”. If the determination result is NO, the leakage condition is regarded as not being satisfied, and the process returns to step S11. On the other hand, if the determination result is YES, it is considered that the leakage condition is satisfied, and the process proceeds to step S55.

  In step S55, it is determined whether the corresponding probability of the detected candidate disease name exceeds a reference value (= 70%). If the determination result is NO, the probability condition is regarded as not being satisfied, and the process returns to step S11. On the other hand, if a determination result is YES, it will be considered that the probability condition was satisfied and it will progress to step S57. In step S57, the detected disease candidate name and the corresponding probability are highlighted with a thick frame for alert. When the enhancement process is completed, the process returns to step S11.

  According to FIG. 10, “pleurisy” is leaking from the diagnostic column DGN, while “angina” is listed in the diagnostic column DGN. Therefore, only “pleurisy 80%” among “pleurisy 80%” and “angina 80%” is highlighted in a bold frame.

  By making all the medical images stored in the DICOM file subject to analysis processing by the image specialized machine learning system 40ai or 40ci, the concern of overlooking the medical image in which a lesion appears and making a diagnosis can be reduced. Also, request comments, findings, and diagnosis comments are subject to analysis processing by the text specialized machine learning system 40at or 40ct, so that some of the diseases are not taken care of (that is, some of the disease names are not noticed). ) Reduces concerns about making a diagnosis.

  By returning to step S11 after the display of the analysis result screen is completed, the processing from step S11 to step S27 is restarted, and the character string displayed in each of the finding column RMK and the diagnostic column DGN and the attached image column IMG are displayed. The displayed thumbnail image can be edited (see FIG. 11).

  According to the description in the thick frame shown in FIG. 11, the character string “recognizing inflammation in the left pleura” is added to the observation column RMK, and the description of the diagnosis column DGN is “1. pleurisy” and “2. It is changed to “suspected psychosis”. When the “Confirm” button is clicked after editing, the processing from step S37 to step S57 is restarted, and the analysis result screen is updated.

  As can be seen from the above description, the image diagnosis support apparatus 10 is an apparatus that supports diagnosis based on a plurality of medical images and request comments created for a common patient, and the CPU 12pr constituting the apparatus supports the following. Processing is executed.

  First, at least one of the plurality of medical images is displayed on the sub-monitor 12m2 facing the interpretation doctor (S9, S25 to S27). Further, the interpretation doctor's opinion (findings and / or diagnosis comments) on the requested comment is acquired through the keyboard / mouse 12 km and displayed on the interpretation report screen on the main monitor 12 m 1 (S 11 to S 17).

  Thereafter, a candidate disease presentation request accompanied by a plurality of medical images is issued to the image specialized machine learning system 40ai or 40ci (S41, S45), and a candidate disease presentation request accompanied by request comments and opinions is issued by a text specialized machine. Issued to the learning system 40at or 40ct (S39, S43). The former candidate disease presentation request requests analysis of a patient's disease based on image information, and the latter candidate disease presentation request requests analysis of the patient's disease based on character information.

  The analysis result by the image specialized machine learning system 40ai or 40ci and the analysis result by the text specialized machine learning system 40at or 40ct are displayed in multiple on the interpretation report screen on the main monitor 12m1 (S47, S49).

  By using a plurality of medical images as a target of analysis processing by the image specialized machine learning system 40ai or 40ci, the concern that a medical image in which a lesion appears and a diagnosis is performed is reduced. In addition, by making the requested comment and the interpretation doctor's opinion on this subject as an object of the analysis processing by the text specialized machine learning system 40at or 40ct, it is possible to pay attention to some diseases (that is, to a part of the disease name). Reduces concerns about diagnosis (without noticing). As a result, even an interpreting doctor with little experience can accurately perform image diagnosis.

  In this embodiment, patient attributes are classified into two categories, “adult” and “child”. However, patient attributes may be classified into two categories: “male” and “female”, or “adult male”, “adult female”, “pediatric male”, and “pediatric female”. Further, the classification may be made so as to have more diversity.

  In this embodiment, when the “Confirm” button is clicked, the diagnostic imaging support apparatus 10 transmits a candidate disease presentation request to the specialized image machine learning system 40ai or 40ci (step S41 in FIG. 15). (See S45). Further, the image specialized machine learning system 40ai or 40ci analyzes the patient's disease based on the image information attached to the transmitted candidate disease presentation request, and returns the analysis result to the image diagnosis support apparatus 10.

  However, instead of this, when the DICOM file is saved in the PACS 30, the PACS 30 sends an analysis request to the image specialized machine learning system 40ai or 40ci (with all the medical images in the DICOM file and the DICOM file number attached). ), And when the “Confirm” button is clicked, the diagnostic imaging support apparatus 10 transmits a candidate disease presentation request (with a DICOM file number) to the specialized image machine learning system 40ai or 40ci. May be.

  In this case, the image specialized machine learning system 40ai or 40ci analyzes the patient's disease based on the image information accompanying the analysis request, and returns the analysis result to the image diagnosis support apparatus 10 in response to the candidate disease presentation request.

  Accordingly, the image specialized machine learning system 40ai or 40ci can analyze the patient's disease before the “confirm” button is clicked. As a result, the response characteristic with respect to the click of the “confirm” button is improved.

  When an analysis request is transmitted from the PACS 30 to the specialized image machine learning system 40ai or 40ci, the apparatus configured by the image diagnosis support apparatus 10 and the PACS 30 corresponds to the “diagnosis support apparatus” of the present invention.

10 ... diagnostic imaging support device 12pr ... CPU
12km ... keyboard / mouse 12hd ... HDD
12m1 ... main monitor 12m2 ... sub-monitor 40 ... machine learning system 40at, 40ct ... text specialized machine learning system 40ai, 40ci ... image specialized machine learning system

Claims (13)

A diagnosis support device that supports diagnosis based on a plurality of medical images and request sentences created for a common patient,
Medical image output means for outputting at least one of the plurality of medical images to a diagnostician;
An acquisition means for acquiring a diagnosis person's opinion on the request sentence in relation to the processing of the medical image output means;
First analysis requesting means for requesting the analysis device to perform first analysis processing for analyzing the disease of the patient based on image information including the plurality of medical images;
A second analysis process for analyzing the disease of the patient based on the character information including the request sentence and the opinion acquired by the acquisition unit is requested to the analysis apparatus in relation to the process of the first analysis request unit. A diagnosis support apparatus comprising: a second analysis request unit; and an analysis result output unit that outputs a first analysis result obtained by the first analysis process and a second analysis result obtained by the second analysis process. The first analysis process is executed based only on the former of the image information and the character information,
The diagnosis support apparatus according to claim 1, wherein the second analysis process is executed based on only the latter of the image information and the character information.   The diagnosis support apparatus according to claim 1 or 2, wherein the analysis result output means outputs the first analysis result and the second analysis result simultaneously toward the diagnostician. The medical image output means includes an update means for updating an output object in response to an update operation,
4. The diagnosis support apparatus according to claim 1, wherein the analysis result output unit restarts at least one of the update unit and the acquisition unit after outputting the first analysis result and the second analysis result. An interpretation report display means for displaying an interpretation report having a request text field and an opinion entry field on the first monitor;
The diagnosis support apparatus according to claim 1, wherein the acquisition unit includes an opinion display unit that displays the opinion in the opinion entry column.   6. The diagnosis support apparatus according to claim 5, wherein the medical image output means includes medical image display means for displaying at least one of the plurality of medical images on a second monitor associated with the first monitor.   The diagnosis support apparatus according to claim 5 or 6, wherein the analysis result output means includes additional display means for additionally displaying the first analysis result and the second analysis result on the first monitor.   8. The diagnosis support apparatus according to claim 1, wherein each of the first analysis request unit and the second analysis request unit issues an analysis request in a different manner depending on the attribute of the patient.   9. The diagnosis support apparatus according to claim 1, wherein each of the first analysis request unit and the second analysis request unit issues an analysis request to the analysis apparatus when a request issue operation is received.   The diagnosis support apparatus according to any one of claims 1 to 9, wherein each of the first analysis result and the second analysis result includes a plurality of candidate disease names and corresponding probabilities of the respective disease candidates. Emphasis means for highlighting candidate disease names satisfying a predetermined condition among the plurality of candidate disease names,
The diagnosis support apparatus according to claim 10, wherein the predetermined condition includes a leakage condition that the opinion is leaked from the opinion acquired by the acquisition unit and a probability condition that the corresponding probability exceeds a reference. A diagnosis support method executed by a diagnosis support apparatus that supports diagnosis based on a plurality of medical images and request sentences created for a common patient,
A medical image output step of outputting at least one of the plurality of medical images to a diagnostician;
An acquisition step of acquiring a diagnosis person's view on the request sentence in association with the processing of the medical image output step,
A first analysis requesting step for requesting an analysis device to perform a first analysis process for analyzing a disease of the patient based on image information including the plurality of medical images;
A second analysis process for analyzing the patient's disease based on the character information including the request sentence and the opinion acquired by the acquisition step is requested to the analysis apparatus in relation to the process of the first analysis request step. A diagnosis support method comprising: a second analysis request step; and an analysis result output step of outputting a first analysis result obtained by the first analysis process and a second analysis result obtained by the second analysis process. In the processor of the diagnosis support apparatus that supports diagnosis based on a plurality of medical images and request sentences created for a common patient,
A medical image output step of outputting at least one of the plurality of medical images to a diagnostician;
An acquisition step of acquiring a diagnosis person's view on the request sentence in association with the processing of the medical image output step,
A first analysis requesting step for requesting an analysis device to perform a first analysis process for analyzing a disease of the patient based on image information including the plurality of medical images;
A second analysis process for analyzing the patient's disease based on the character information including the request sentence and the opinion acquired by the acquisition step is requested to the analysis apparatus in relation to the process of the first analysis request step. Diagnosis support for executing a second analysis request step and an analysis result output step for outputting the first analysis result obtained by the first analysis process and the second analysis result obtained by the second analysis process program.

Images