Medical report generation method, medical report generation system, and program
The medical report generation system automates the addition of new clinical indicators and image processing, addressing inefficiencies by allowing physicians to input new conditions and update algorithms, thereby reducing the workload in medical report preparation.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- CANON KK
- Filing Date
- 2022-06-30
- Publication Date
- 2026-07-29
AI Technical Summary
Existing medical report generation systems require physicians to manually launch additional tools and perform measurements for newly discovered conditions, leading to increased workload and inefficiency in report preparation.
A medical report generation system that allows physicians to input new clinical indicators, presents matching search labels, performs automatic second image processing, and updates first image processing based on user input, reducing the need for manual data entry and tool usage.
The system automates the addition of new clinical indicators and image processing, reducing the burden on physicians by integrating new conditions into medical reports efficiently and updating image processing algorithms for future use.
Smart Images
Figure 0007897057000001 
Figure 0007897057000002 
Figure 0007897057000003
Abstract
Description
Technical Field
[0001] The embodiments disclosed in this specification and the drawings relate to a medical report generation method, a medical report generation system, and a program.
Background Art
[0002] When a doctor (e.g., a radiologist) reads a medical image, observes abnormalities in the medical image, diagnoses a disease, and creates a medical diagnosis report (hereinafter referred to as a medical report). A medical report summarizes and describes important findings in a medical image (e.g., an X-ray image, a computed tomography (CT) image, a magnetic resonance imaging (MRI) image, an ultrasonic image, etc.) generated by a medical image generation device. Therefore, a medical report is generally considered to be a very important component of a patient's medical profile.
[0003] In order to improve the efficiency of generating medical reports, AI algorithms for assisting in reading medical images and creating medical reports are being increasingly developed. As a typical example of a medical report generation method based on an AI algorithm, the following steps are performed.
[0004] First, a step of acquiring a medical image generated by a medical image generation device and transmitting it to a Picture Archiving and Communication System (PACS) to store the medical image in the PACS is performed.
[0005] Next, a step of transmitting the medical image to an AI workstation, automatically analyzing the medical image by the AI workstation to obtain information on the condition (clinical index) in the medical image, and performing image processing corresponding to the condition information to obtain a detection result, which is part of the data necessary for diagnosis corresponding to the image processing, is performed.
[0006] Next, the detection results are sent to the automated report generation system. The automated generation system then enters (pre-fills) the aforementioned detection results into the medical report and generates the pre-filled medical report.
[0007] Next, the radiologist completes the medical report by editing the pre-filled medical report.
[0008] Here, the image processing described above is usually stored in the memory unit in association with the disease state (clinical indicator) information in the medical image, and may also be stored in association with the attributes of the medical image. For example, if the medical image is a CT image and the AI workstation automatically analyzes that the clinical indicator is "lung cancer," then the image processing corresponding to "lung cancer" from the image processing stored in the memory unit will be executed on that medical image. Then, as an example of executing the image processing corresponding to "lung cancer," the lung region of the medical image is divided (segmented) and analyzed to obtain detection results, which are data necessary for diagnosing lung cancer. Subsequently, the detection results are automatically entered into the medical report.
[0009] However, real-world disease diagnosis is more complex, and the image processing described above, being pre-configured during software development, typically cannot meet the complex clinical demands of reality. For example, after a preliminary medical report has been created, a physician may discover a new condition while interpreting medical images and wish to supplement the report with information about that new condition.
[0010] In contrast, Patent Document 1 discloses that if, after creating a medical report (image interpretation report), the physician discovers a new disease in the medical image or if there are findings missing from the image interpretation report, the physician can input additional keywords via voice or other means, and these additional keywords will be added to the image interpretation report.
[0011] Patent Document 2 discloses that after a medical report (image interpretation report) has been created, if a physician issues a correction instruction to either the diseased area in the medical image or the description of the diseased area in the image interpretation report, the other description of the diseased area in the medical image or the description of the diseased area in the image interpretation report will also be corrected in accordance with the correction instruction.
[0012] However, in the case of the two patent documents mentioned above, physicians need to separately launch other tools such as CAD (Computer-Aided Diagnosis) software to measure the organs and body parts related to the new condition, obtain the data necessary for diagnosing the new condition, and manually supplement this data in the medical report. Furthermore, as mentioned above, the image processing is pre-configured during the software development stage and cannot usually be edited and saved by the user. Therefore, even if the same condition is newly discovered in the same type of medical image (e.g., a CT image) the next time, the physician will need to launch other tools such as CAD software to perform measurements and manually enter the results into the medical report again. [Prior art documents] [Patent Documents]
[0013] [Patent Document 1] Japanese Patent Publication No. 2019-153250 [Patent Document 2] Japanese Patent Publication No. 2019-149130 [Overview of the Initiative]
[0014] One of the problems that the embodiments disclosed herein and in the drawings aim to solve is to reduce the burden of preparing medical reports. However, the problems solved by the embodiments disclosed herein and in the drawings are not limited to the above problem. Problems corresponding to the effects of each configuration shown in the embodiments described later can also be positioned as other problems. [Means for solving the problem]
[0015] The medical report generation method according to this embodiment includes an acquisition step, a first image processing step, a preliminary medical report entry step, an image interpretation information input step, a search label presentation step, a second image processing step, and a medical report completion step. The acquisition step acquires a medical image that includes the body part the user wishes to diagnose and to which clinical indicator information has been added. The first image processing step performs a first image processing, which is a predetermined image processing stored in advance in association with the body part and the clinical indicator information relating to the medical image, on the medical image. The preliminary medical report entry step generates and displays a preliminary medical report according to the detection results obtained by performing the first image processing. The image interpretation information input step generates image interpretation information keywords based on new clinical indicators entered by the user for the medical report. The search label presentation step presents to the user search labels that are candidates for new clinical indicators that match the image interpretation information keywords, and that are sorted according to the degree of matching between the label description corresponding to the search label and the image interpretation information keywords. The second image processing step performs a second image processing on the medical image corresponding to the search label selected by the user from among the presented search labels. The medical report completion step completes the medical report according to the detection results obtained from the second image processing. [Brief explanation of the drawing]
[0016] [Figure 1] Figure 1 is a diagram showing the configuration of the medical report generation system according to this embodiment. [Figure 2] Figure 2 is a flowchart showing the processing (medical report generation method) by the medical report generation system according to this embodiment. [Figure 3] Figure 3 shows how a user discovers a new clinical indicator and enters the image interpretation information related to that new clinical indicator into a pre-filled medical report. [Figure 4A]FIG. 4A is a diagram showing a state in which a search label as a candidate for a new clinical index is presented to a user. [Figure 4B] FIG. 4B is a diagram showing a state in which, after a user selects a search label, image processing corresponding to the search label is automatically executed. [Figure 4C] FIG. 4C is a diagram showing a state in which, after image processing, the result of a corresponding detection item is introduced into a medical report. [Figure 5] FIG. 5 is a diagram showing an example of second image processing stored in a second storage table. [Figure 6A] FIG. 6A is a diagram showing an example of first image processing stored in a first storage table. [Figure 6B] FIG. 6B is a diagram showing an example of the first image processing corrected by executing step S108. [Figure 7] FIG. 7 is a diagram showing a state in which a new search label added to a second storage table is presented to a user.
MODE FOR CARRYING OUT THE INVENTION
[0017] Hereinafter, embodiments of a medical report generation method, a medical report generation system, and a program will be described with reference to the drawings.
[0018] FIG. 1 is a configuration diagram showing a medical report generation system 10 according to the present embodiment. The medical report generation system 10 according to the present embodiment includes a medical image acquisition unit 1, a processing unit 2, an input unit 3, a display unit 4, and a storage unit 5. The processing unit 2 and the storage unit 5 are provided in a medical report generation apparatus 20.
[0019] For example, the medical image acquisition unit 1 is a communication interface connected to the medical report generation apparatus 20. The medical image acquisition unit 1 acquires a medical image including a site where a user desires a diagnosis and to which clinical index information is added.
[0020] For example, the medical report generation device 20 contains a processing unit 2 and a storage unit 5. Specifically, the medical report generation device 20 is a computer and includes processing circuits that execute each processing function and storage circuits that store each of these processing functions in the form of a program that can be executed by the computer. The processing unit 2 corresponds to each processing function that executes steps S101 to S108 described later.
[0021] For example, input unit 3 is an input interface connected to the medical report generation device 20. Input unit 3 allows the user to discover new clinical indicators and input new image interpretation information into a pre-filled medical report.
[0022] For example, the display unit 4 is a display device connected to the medical report generation device 20. The display unit 4 displays the medical report generated by the processing unit 2.
[0023] For example, the memory unit 5 has a first memory table 51 and a second memory table 52, which will be described later.
[0024] The method for generating a medical report using the medical report generation system 10 according to this embodiment will be explained below with reference to Figures 2 to 6. Figure 2 is a flowchart showing the processing (medical report generation method) by the medical report generation system 10 according to this embodiment. Here, we will explain using as an example the case in which a user (doctor) newly discovers that the patient has "liver metastases" when diagnosing whether or not the patient has lung cancer.
[0025] First, the processing unit 2 of the medical report generation system 10 executes step S101 in Figure 2.
[0026] In step S101, the processing unit 2 uses the medical image acquisition unit 1 to acquire a medical image generated by the medical image generation device, or to read and acquire the medical image from a medical image database that stores medical images. The medical image database is, for example, located within a PACS. The medical image includes the body part that the user wishes to diagnose (here, "chest" as an example), and clinical indicator information is attached to the medical image.
[0027] Here, the medical image may be a 2D or 3D image, and may be generated from any of the following imaging modes: for example, functional MRI (e.g., f(Functional)MRI, DCE (Dynamic Contrast Enhanced)-MRI, and diffusion MRI), cone-beam CT (CBCT), spiral CT, positron emission tomography (PET), single-photon emission computed tomography (SPECT), X-ray imaging, optical tomography, fluorescence imaging, ultrasound imaging, and radiotherapy portal imaging, or a combination thereof. In this embodiment, the medical image is assumed to be a CT image. Furthermore, the clinical indicator information may be disease condition information automatically determined by analyzing the medical image using an AI workstation that has performed machine learning, or it may be a disease name such as "lung cancer" included in a description such as "I want to be tested for the presence or absence of lung cancer" that is entered incidentally when user information is registered.
[0028] Next, the processing unit 2 of the medical report generation system 10 executes step S102 in Figure 2.
[0029] In step S102, the processing unit 2 performs a first image processing on the medical image based on the human body part (here, "chest") related to the medical image and the clinical indicator information, "lung cancer". The first image processing is a predetermined image processing that has been stored in the first storage table 51 in association with the human body part and clinical indicator information in advance, and is for detecting predetermined detection items using a predetermined algorithm.
[0030] Figure 6A shows an example of the first image processing stored in the first memory table 51. As shown in Figure 6A, when the body part in question is the "chest" and the clinical indicator information is "lung cancer", the first image processing performs an image segmentation process called the "lung nodule analysis pipeline" on the medical image as a predetermined image processing, and then derives the data necessary for lung detection according to a predetermined algorithm corresponding to the said processing (the detection items are not shown in Figure 6A). In other words, when the processing unit 2 of the medical report generation system 10 acquires a medical image in step S101 that includes the body part that the user wishes to diagnose and to which clinical indicator information is attached, it automatically executes the first image processing related to the body part and clinical indicator information related to the medical image and obtains the detection results of the detection items corresponding to the first image processing.
[0031] Here, although not shown in Figure 6A, the attributes of the medical image and the detection items corresponding to the first image processing are pre-stored in the first storage table 51.
[0032] Examples of attributes of medical images include patient ID, examination ID, device ID, and image series ID. Specifically, a patient's medical image is an image to which a DICOM (Digital Imaging and Communications in Medicine) tag is attached as an attribute (ancillary information) of the medical image. These attributes include, for example, patient ID, examination ID, device ID, and image series ID, and are standardized according to the DICOM standard.
[0033] Examples of detection items corresponding to the first image processing include count, shape, size, and location. For example, count represents the number of tumors, while shape, size, and location represent the shape, size, and location of the tumors, respectively.
[0034] After the first image processing is performed, the processing unit 2 of the medical report generation system 10 executes step S103 in Figure 2.
[0035] In step S103, the processing unit 2 generates detection results for the corresponding detection items obtained by the first image processing, that is, some data on the lungs necessary for diagnosing lung cancer (for example, whether the airways and thoracic blood vessels are normal, the presence or absence of nodules, and the size of the nodules). The processing unit 2 then enters the above detection results into a medical report (preliminary entry), generates a medical report with the preliminary entry, and displays it on the display unit 4.
[0036] Next, the processing unit 2 of the medical report generation system 10 executes step S104 in Figure 2.
[0037] Here, we assume that when a user interprets a medical image, they discover a new clinical indicator, "liver metastasis," in the patient, and attempt to input this newly discovered clinical indicator into the medical report using the input unit 3. Figure 3 shows the user discovering a new clinical indicator and inputting the interpretation information related to that new clinical indicator into the pre-filled medical report.
[0038] As shown in Figure 3, suppose the user, by operating the input unit 3, enters the description "metastasis confirmed" (corresponding to "image interpretation information for medical images") into the pre-filled medical report regarding the newly discovered clinical indicator, "liver metastasis." In this case, in step S104, the processing unit 2 generates image interpretation information keywords based on the image interpretation information "metastasis confirmed" entered by the user. Specifically, the processing unit 2 generates image interpretation information keywords based on the image interpretation information "metastasis confirmed" entered by the user and the contents of the pre-filled medical report.
[0039] As an example, processing unit 2 searches for a parent item (upper abdomen in Figure 3) that is an editable item related to the body part "chest" based on the context of the pre-filled medical report. Processing unit 2 also extracts the keyword "metastasis" from the image interpretation information "metastasis confirmation" entered by the user. Then, processing unit 2 combines the searched parent item "upper abdomen" with the keyword "metastasis" extracted from the image interpretation information "metastasis confirmation" to generate the image interpretation information keyword "upper abdomen metastasis," which includes the aforementioned parent item "upper abdomen" and the aforementioned keyword "metastasis." Here, the user may input the image interpretation information in various input formats such as voice or text, and when extracting the keyword "metastasis" from the image interpretation information, known keyword extraction techniques may be used to extract the keyword from the original voice or text input by the user using natural language processing.
[0040] Next, the processing unit 2 of the medical report generation system 10 executes step S105 in Figure 2.
[0041] Figure 4A shows how the system presents the user with search labels as candidates for new clinical indicators. In step S105, for example, as shown in Figure 4A, the processing unit 2 presents the user with search labels that are candidates for new clinical indicators that match the image interpretation information keyword "upper abdominal metastasis," such as "1. Liver metastasis 2. Adrenal metastasis 3....". The presented search labels are sorted according to the degree of match (score) between the label description corresponding to the search label and the image interpretation information keyword, with higher match scores ranking higher.
[0042] Here, if the degree of matching between the label description corresponding to the search label and the image interpretation information keyword is the same, processing unit 2 further sorts according to the past frequency of use of the search label. Specifically, processing unit 2 sorts by the probability of clinical occurrence, based on the frequency of user usage in similar past cases, and sorts labels that have been selected more often by users in similar past cases to a higher position. This probability of clinical occurrence can be statistically calculated and updated in real time by the developer.
[0043] Furthermore, the search labels are classified by body part, and the label descriptions corresponding to the search labels are pre-stored in the second memory table 52, associated with the search labels.
[0044] Figure 5 shows an example of the second image processing stored in the second memory table 52. As shown in Figure 5, the second image processing is stored in the second memory table 52 in association with the search label, label description, detection items, and the second image processing. The search labels are classified by body part, such as "liver metastasis," "adrenal metastasis," "bone metastasis," and "lymph node metastasis," and the label description, detection items, and second image processing corresponding to the search label are stored in the second memory table 52 in association with the search label. For example, if the search label is "liver metastasis," the label description corresponding to the search label "liver metastasis" is "liver metastasis analysis" and "tumor coefficient, shape, size, and location." The detection items corresponding to the search label "liver metastasis" are "count," "shape," "size," and "location." Furthermore, the second image processing corresponding to the search label "liver metastasis" is performed as an algorithm pipeline, starting with liver segmentation, then liver segmentation and tumor extraction, and finally the statistical engine.
[0045] Because the search labels are stored in the second memory table 52, for example, after the user enters "metastasis confirmed" as image interpretation information, the display unit 4 can display search labels such as "1. Liver metastasis 2. Adrenal metastasis 3...." as shown in Figure 4A. In other words, the search labels as candidates for new clinical indicators are automatically highlighted in a pop-up display or similar, allowing the user to select a search label.
[0046] Then, the processing unit 2 of the medical report generation system 10 executes step S106 in Figure 2.
[0047] In step S106, suppose the user selects the search label "1. Liver metastasis" from the presented search labels, as shown in Figure 4A. In this case, the processing unit 2 performs a second image processing (liver segmentation, liver segmentation, tumor extraction, statistical engine) on the medical image, corresponding to the search label "1. Liver metastasis" selected by the user, in the second memory table 52. At this time, the second image processing corresponding to the search label "Liver metastasis" sequentially acquires several data necessary for diagnosis by detecting predetermined detection items (count, shape, size, position) using a predetermined algorithm (algorithm pipeline).
[0048] Figure 4B shows how the processing unit 2 automatically performs a second image processing operation corresponding to the search label "liver metastasis" after the user selects the search label "1. liver metastasis". More specifically, during the execution of this second image processing operation, for example, as shown by the dashed box in Figure 4B, the medical report displayed on the display unit 4 shows the detection items (count, shape, size, location) corresponding to "liver metastasis" in the second memory table 52. However, since it takes a certain amount of time to derive the data for each detection item, it is displayed as "loading" in the figure.
[0049] After the second image processing is performed, the processing unit 2 of the medical report generation system 10 executes step S107 in Figure 2.
[0050] In step S107, the processing unit 2 performs a second image processing on the medical image and, according to the detection results obtained, supplements the medical report and displays it on the display unit 4. As shown by the dashed frame in Figure 4C, the processing unit 2 derives the detection results as data for each detection item (count, shape, size, position) and supplements them into the medical report. For example, as shown in Figure 4C, as a detection result for the search label "liver metastasis", the tumor count representing the number of tumors is "2", and for the first tumor "#1", the tumor size, shape, and position are derived to be "3.8 × 3.5 mm", "elliptical", and "VII", respectively, and are supplemented into the medical report. Similarly, for the second tumor "#2", the tumor size, shape, and position are derived and supplemented into the medical report.
[0051] Here, it is preferable to cache the specific parameters of each derived detection item for a predetermined time. This allows, for example, when multiple systems are connected via a LAN (Local Area Network) within a hospital, if the specific parameters of each detection item for a patient have already been obtained within the predetermined time, they can be directly incorporated into the medical report as shown in Figure 4C, and the loading process shown in Figure 4B can be omitted. When deriving each detection item, known acceleration techniques such as parallel computing, distributed computing, cluster computing, and cloud computing may be employed.
[0052] After step S107, the processing unit 2 of the medical report generation system 10 executes step S108 in Figure 2.
[0053] In step S108, the processing unit 2 updates the first image processing in the first memory table 51 to add the second image processing, and stores the updated first image processing associated with the human body part and the first keyword. In the example described above, after the processing unit 2 supplements the medical report with the detection results related to the search label "liver metastasis" in step S107, it adds the second image processing, "liver metastasis pipeline," which corresponds to the search label "liver metastasis," to the first image processing described above, as shown in Figure 6B, to "(lung nodule analysis pipeline) and (liver metastasis pipeline)," and stores it in the first memory table 51. As a result, when the medical report generation system 10 next acquires a medical image that includes the chest and has the clinical indicator information "lung cancer" added to it, it can automatically operate the updated first image processing, "(lung nodule analysis pipeline) and (liver metastasis pipeline)," as shown in Figure 6B.
[0054] In step S108, more preferably, the processing unit 2 displays a screen on the display unit 4 containing the message "Do you want to save this image processing method?" and the operation button "Save," thereby confirming with the user whether or not to perform the update of the first image processing described above. If the user determines that the same algorithm as the one executed in the medical report should be run for all newly acquired medical images of the same genre (e.g., CT images), in other words, if the user wants to directly obtain a medical report like the one in Figure 4C for all newly acquired medical images of the same genre, they click the operation button "Save" on the screen. In this case, the first image processing in the first memory table is updated to include the second image processing, as shown in Figure 6B, and the next time the medical report generation system 10 acquires a medical image containing the same human body part and with the same clinical indicator information added, it will execute a new first image processing that includes the second image processing. If the user clicks "Do not save," the first image processing is not updated. In other words, the next time the medical report generation system 10 acquires a medical image containing the same body part and with the same clinical indicator information, it will execute the original first image processing as is (Figure 6A).
[0055] Furthermore, more preferably, the function that allows the user to confirm whether or not to save the image processing may be subject to editing permissions. For example, permissions for this function may be granted to only certain doctors within the hospital, or permissions for this function may be granted to certain doctors for only certain organs.
[0056] As described above, in the medical report generation system 10 according to this embodiment, when a physician discovers a new disease, "liver metastasis," while interpreting medical images and edits the medical report, the system complements the medical report based on the context of the medical report. As a result, when a user discovers a new disease, "liver metastasis," in the medical report generation system 10 according to this embodiment, they only need to input the relevant image interpretation information, "metastasis confirmation," to execute the corresponding image processing and obtain the detection results of the detection items (count, shape, size, location) necessary for diagnosing the disease, without having to separately launch other tools such as CAD software. In this way, the medical report generation system 10 according to this embodiment can reduce the burden of creating medical reports.
[0057] Specifically, if a user discovers another medical condition while interpreting a medical image and wishes to analyze or measure it, the medical report generation system 10 according to this embodiment eliminates the need to separately launch other tools such as CAD software, perform measurements or analysis, and then manually enter the results into the medical report. Taking the above embodiment as an example, the user simply enters "metastasis confirmed" into the pre-generated medical report, and the search labels "1. Liver metastasis 2. Adrenal metastasis 3…" shown in Figure 4A are automatically highlighted in a pop-up display or similar. When the user selects the search label "Liver metastasis," the medical report generation system 10 automatically performs a second image processing corresponding to the search label "Liver metastasis," and reflects the detection results obtained from this second image processing into the medical report, as shown in Figure 4C. Furthermore, in the medical report generation system 10 according to this embodiment, the system confirms with the user whether or not to perform an update to the first image processing. If the user determines that the same algorithm used in the medical report should be applied to all newly acquired medical images of the same genre, the system updates the first image processing to add the second image processing by clicking the "Save" operation button on the screen. In this case, when generating a medical report for newly acquired medical images of the same genre the next time, the medical report generation system 10 according to this embodiment automatically adds and executes the image processing corresponding to "liver metastasis" without the user having to input the description "metastasis confirmation," thereby generating the medical report shown in Figure 4C. Therefore, the medical report generation system 10 according to this embodiment can reduce the workload associated with creating medical reports.
[0058] Furthermore, in the medical report generation system 10 according to this embodiment, new disease conditions and corresponding image processing can be added by the software developer during updates and maintenance of the medical report generation system 10. Taking the above embodiment as an example, if the search labels "bone metastasis" and "lymph node metastasis" shown in Figure 5 were not previously stored but were newly added by the software developer, then after the user inputs "metastasis confirmed" as image interpretation information, the medical report generation system 10 according to this embodiment will automatically highlight the search labels "1. liver metastasis 2. adrenal metastasis 3. bone metastasis 4. lymph node metastasis" in a pop-up display on the display unit 4, as shown in Figure 7, allowing the user to select a search label.
[0059] Thus, in the medical report generation system 10 according to this embodiment, when the system matches newly entered image interpretation information, it highlights it, making it easy for the user to be notified that a new medical condition has been added to the system. In the embodiment described above, even if the user ultimately selects "liver metastasis," the newly appearing search labels "3. Bone metastasis" and "4. Lymph node metastasis" options are visible on the display unit 4, as shown in Figure 7. This also has the advantage that if the user diagnoses "bone metastasis" or "lymph node metastasis" next time, the system will activate the corresponding image processing by directly entering the relevant description of "bone metastasis" or "lymph node metastasis" without having to launch other tools such as CAD software.
[0060] Furthermore, in the medical report generation system 10 according to this embodiment, if the degree of matching between the label description corresponding to the search label and the image interpretation information keyword is the same, the search label is sorted according to the frequency of use by the user in similar cases in the past. Therefore, even if the user has not directly diagnosed "bone metastasis" or "lymph node metastasis," they can see that it is displayed higher up, and thus know that it has been selected many times by other users. As a result, the medical report generation system 10 according to this embodiment focuses on the possibility that it is a clinically common disease and further interprets the medical image to determine whether the patient has a disease such as "bone metastasis" or "lymph node metastasis," thereby preventing the oversight of the disease.
[0061] In the embodiment described above, the first image processing is stored in the first memory table 51 in association with human body parts and clinical indicator information. However, the first image processing may also be stored in the first memory table 51 in association with the image format of the medical image, the number of medical images, etc. Examples of image formats here include CT images and MRI images.
[0062] Furthermore, if, after the user has entered the image interpretation information "metastasis confirmed" for the medical image described in the medical report, only one search label is presented as a candidate for a new clinical indicator in step S105, then in step S106, the processing unit 2 may skip the step of having the user select the search label and directly and automatically execute the second image processing corresponding to that single search label.
[0063] Furthermore, in the embodiments described above, the method by which the user inputs the image interpretation information is not limited; it may be input using a mouse, keyboard, etc., or it may be input by voice using a microphone.
[0064] Furthermore, this embodiment can also be implemented as a medical report generation program, a computer-readable recording medium on which the program is recorded, and the like.
[0065] According to at least one embodiment described above, the burden of preparing medical reports can be reduced.
[0066] While several embodiments have been described, these embodiments are presented as examples only and are not intended to limit the scope of the invention. These embodiments can be implemented in a variety of other forms, and various omissions, substitutions, modifications, and combinations of embodiments are possible without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents. [Explanation of Symbols]
[0067] 1. Medical Image Acquisition Unit 2 Processing Units 3. Input section 4 Display section
Claims
1. The acquisition step involves obtaining medical images that include the area the user wishes to diagnose and to which clinical indicator information is added. A first image processing step in which a first image processing is performed on the medical image, which is a predetermined image processing that has been stored in advance in association with the human body parts and clinical indicator information relating to the medical image, based on the human body parts and clinical indicator information relating to the medical image; A medical report preliminary entry step is performed to generate and display a medical report with preliminary entries in accordance with the detection results obtained by the first image processing described above, When the user inputs new clinical indicators into the medical report, the process includes an input step for image interpretation information, which generates image interpretation information keywords based on the image interpretation information. A search label presentation step, which presents to the user search labels that are candidates for new clinical indicators that match the aforementioned image interpretation information keywords, and which are sorted according to the degree of matching between the label description corresponding to the search label and the aforementioned image interpretation information keywords; A second image processing step involves performing a second image processing on the medical image corresponding to the search label selected by the user from among the presented search labels. A medical report completion step is performed to complete the medical report according to the detection results obtained by the second image processing described above, A method for generating medical reports, which is performed by a medical report generation system.
2. After the aforementioned medical report supplementation step, A first image processing update step involves updating the first image processing to apply the second image processing described above, and storing the updated first image processing in association with the human body part and the clinical indicator information. A method for generating a medical report according to claim 1, further comprising:
3. The authority to execute the first image processing update step is granted to a designated user. The method for generating a medical report according to claim 2.
4. In the aforementioned image interpretation information input step, the image interpretation information keyword is generated based on the image interpretation information entered by the user and the contents of the pre-filled medical report. A method for generating a medical report according to claim 1 or 2.
5. In the image interpretation information input step, based on the context of the pre-filled medical report, a parent item related to the body part is identified, and the parent item and keywords extracted from the image interpretation information entered by the user are combined to generate the image interpretation information keyword that includes the parent item and the keywords. The method for generating a medical report according to claim 4.
6. In the step of presenting the search label, the search label is highlighted and presented to the user. A method for generating a medical report according to claim 1 or 2.
7. Detection items are stored in advance in association with the first image processing and the second image processing, respectively, and after the execution of the first image processing or the second image processing, the detection results of the corresponding detection items are entered into the medical report. A method for generating a medical report according to claim 1 or 2.
8. The detection results are cached for a predetermined time. The method for generating a medical report according to claim 6.
9. In the search label presentation step, if the degree of matching between the label description corresponding to the search label and the image interpretation information keyword is the same, the search labels are further sorted according to their past frequency of use. A method for generating a medical report according to claim 1 or 2.
10. The first image processing is further stored in association with the attributes of the medical image. A method for generating a medical report according to claim 1 or 2.
11. In the search label presentation step, if only one search label is presented, In the second image processing step described above, the second image processing corresponding to the single search label is performed directly and automatically. A method for generating a medical report according to claim 1 or 2.
12. A medical image acquisition unit that acquires medical images including the area the user wishes to diagnose, with clinical indicator information added; Input section, Display unit and Memory unit and, A processing unit that performs a first image processing, which is a predetermined image processing stored in the storage unit in association with the human body parts and clinical indicator information related to the medical image, on the medical image based on the human body parts and clinical indicator information, generates a pre-filled medical report according to the detection result obtained from the first image processing and displays it on the display unit, and when the user inputs new clinical indicator reading information for the medical report using the input unit, generates reading information keywords based on the reading information, presents the user with search labels that are candidates for new clinical indicators that match the reading information keywords, and sorts the search labels according to the degree of matching between the label description corresponding to the search label and the reading information keyword, performs a second image processing on the medical image corresponding to the search label selected by the user from among the presented search labels, and complements the medical report according to the detection result obtained from the second image processing and displays it on the display unit, A medical report generation system equipped with [features / equipment].
13. The system acquires medical images that include the area the user wishes to diagnose, with clinical indicator information added. Based on the human body parts and clinical indicator information relating to the medical image, a first image processing, which is a predetermined image processing stored in advance in association with the human body parts and clinical indicator information, is performed on the medical image. Based on the detection results obtained from the first image processing described above, a pre-filled medical report is generated and displayed. When the user inputs new clinical indicators for the medical report, the system generates interpretation information keywords based on that interpretation information. The user is presented with search labels that are candidates for new clinical indicators that match the aforementioned image interpretation information keywords, and these search labels are sorted according to the degree of matching between the label description corresponding to the search label and the aforementioned image interpretation information keywords. A second image processing step is performed on the medical image, corresponding to the search label selected by the user from among the presented search labels. In accordance with the detection results obtained by the second image processing described above, the medical report is supplemented as follows: A program that instructs a computer to perform a process.