Medical determination support device and non-transitory recording medium

Abstract

A medical determination support device and a non-transitory recording medium, the subject being to calculate an analysis result relating to a site to be examined of a subject on the basis of a plurality of types of medical information by using a plurality of arithmetic units corresponding to each of the plurality of types of medical information, the analysis result being calculated in consideration of the importance of each arithmetic unit. The analysis unit outputs an analysis result using a group of arithmetic units that correspond to each of the plurality of types of medical information acquired by the medical information processing unit and perform analysis on the basis of the corresponding type of medical information. The memory stores importance information indicating the importance of each type of medical information when an analysis result regarding an analysis item is output. The analysis unit determines the importance of each piece of medical information (in other words, each arithmetic unit) by referring to the importance information before analysis, and calculates an analysis result on the basis of the partial analysis information output by each arithmetic unit when each piece of medical information is input to the corresponding arithmetic unit and the determined importance of each arithmetic unit.

Description

Technical Field

[0001] This specification discloses improvements of a medical judgment support device and a medical judgment support program. Background Art

[0002] In the past, doctors made judgments (e.g., diagnoses) on the determination target site of the subject based on various medical information about the subject obtained using one or more medical devices. In addition, in this specification, judgments related to the determination target site refer to judgments on the progression state or prognosis (prevention prediction, progression prediction, risk of onset (including recurrence), etc.) of a certain lesion in the determination target site. Here, doctors sometimes analyze according to one or more analysis items based on various medical information and make judgments on the determination target site based on the analysis results of the one or more analysis items.

[0003] For example, when the determination target site is the heart and the determination content is a determination related to ischemic heart disease, the doctor analyzes the determination target site, i.e., the heart, according to one or more analysis items related to the heart (e.g., ultrasonic images, CT (Computed Tomography) images, magnetic resonance images, etc.), and makes a determination related to ischemic heart disease based on the analysis results.

[0004] In the past, there has been proposed a device that supports the diagnosis of a subject using multiple medical information (multiple image data). For example, Patent Document 1 discloses an image diagnosis support system that processes multiple image data of a subject's test site acquired using different medical equipment (e.g., a mammography or ultrasonic diagnostic device) to generate benign or malignant determination data of the test site, wherein probability values ​​of malignancy of the detected lesions are generated respectively based on the characteristic data of the detected lesions, and a comprehensive probability value is calculated by weighting or referring to a LUT (Look Up Table) corresponding to a combination of the contents of the signal indicating the presence or absence of the lesion. In addition, Patent Document 2 discloses an image diagnosis support technology that supports disease discrimination using multiple measurement values ​​acquired by a medical image acquisition device.

[0005] Patent Document 1: Japanese Patent No. 4633298

[0006] Patent Document 2: Japanese Patent Application Publication No. 2020-192068

[0007] Consider using arithmetic units corresponding to a plurality of types of medical information to perform analysis based on the plurality of types of medical information for obtaining analysis results related to the inspection object part of the subject. The arithmetic unit is, for example, a learning model or a LUT (LookUp Table), etc., but is not limited thereto. Here, the importance of each of the plurality of types of medical information when obtaining the analysis result is sometimes different from each other. That is, the importance of the arithmetic units corresponding to the plurality of types of medical information is sometimes different from each other. For example, when obtaining an analysis result related to the cardiac function of the heart as the inspection object part, when an ultrasonic image of the heart at rest as medical information is more important than an ultrasonic image of the heart under load as medical information, the importance of the arithmetic unit that processes the ultrasonic image of the heart at rest is greater than that of the arithmetic unit that processes the ultrasonic image of the heart under load. Summary of the invention

[0008] The medical judgment support device disclosed in this specification aims to calculate the analysis result considering the importance of each operator when calculating the analysis result about the inspection target part of the subject based on the various medical information using a plurality of operators corresponding to the various medical information.

[0009] The medical judgment support device disclosed in the present specification comprises: an analysis unit, which outputs an analysis result related to an inspection object part of a subject based on multiple medical information related to the subject, and uses a group of operators that respectively correspond to the types of the medical information and output partial analysis information for calculating the analysis result based on the corresponding types of the medical information to output the analysis result. The medical judgment support device is characterized in that the analysis unit performs the following processing: determines the importance of each operator included in the operator group based on importance information indicating the importance of each type of the medical information when outputting the analysis result; and calculates the analysis result based on the partial analysis information output by each operator when the corresponding medical information is input and the determined importance of each operator.

[0010] Preferably, in the importance information, the importance of each type of the medical information is associated with the diagnostic condition of the subject, and the analysis unit determines the importance of each of the operators included in the operator group based on the importance information and the current diagnostic condition of the subject.

[0011] Preferably, the analyzing unit does not use a part of the computing units in the computing unit group to calculate the analysis result according to the determined importance of each computing unit.

[0012] Preferably, the analysis unit has a plurality of operator groups corresponding to a plurality of analysis items related to the inspection object part, and the analysis unit performs the following processing: for each of the analysis items, determining the importance of each operator included in the plurality of operator groups based on the importance information showing the importance of each type of medical information; and for each of the analysis items, calculating the analysis result based on the partial analysis information output by each operator when the corresponding medical information is input and the determined importance of each operator, thereby outputting a plurality of analysis results regarding the plurality of analysis items.

[0013] Preferably, the importance of a computing unit for one of the analysis items is different from the importance of another computing unit for another of the analysis items and corresponding to the medical information of the same type as that of the computing unit.

[0014] Preferably, there is also a comprehensive judgment unit, which judges the state of the inspection object part based on the output of the comprehensive judgment learning model when the multiple analysis results output by the analysis unit are input to the comprehensive judgment learning model, and the comprehensive judgment learning model is trained to predict and output the state of the inspection object part based on multiple analysis results of the multiple analysis items.

[0015] Preferably, the inspection object part is the heart, the analysis unit outputs the disease status related to each of the multiple disease categories related to the heart as the multiple analysis results, and the comprehensive judgment unit comprehensively judges the state of the heart of the subject based on the disease status related to each of the multiple disease categories.

[0016] Preferably, the medical determination support device further includes a display control unit configured to display the medical information serving as a basis for the analysis by the analysis unit and the determination result of the comprehensive determination unit on a display unit.

[0017] Preferably, the display control unit displays, on the display unit, a screen for selecting any one of the plurality of medical information serving as a basis for analysis by the analysis unit or the determination result of the comprehensive determination unit.

[0018] Preferably, the display control unit performs display in a manner that allows the importance of the medical information to be determined based on the importance information.

[0019] Furthermore, the medical judgment support program disclosed in the present specification is characterized in that the medical judgment support program enables a computer to function as an analysis unit, and the analysis unit outputs an analysis result related to the inspection object part of the subject based on a plurality of medical information related to the subject, and outputs the analysis result using a group of operators that respectively correspond to the types of the medical information and are trained in a manner of outputting partial analysis information for calculating the analysis result based on the corresponding types of the medical information, and the analysis unit performs the following processing: determining the importance of each operator included in the operator group based on importance information indicating the importance of each type of the medical information when outputting the analysis result; and calculating the analysis result based on the partial analysis information output by each operator when the corresponding medical information is input and the determined importance of each operator.

[0020] Effects of the Invention

[0021] According to the medical judgment support device disclosed in this specification, when multiple computing units corresponding to multiple types of medical information are used to compute analysis results related to the inspection target part of the subject based on the multiple types of medical information, the analysis results can be computed taking into account the importance of each computing unit. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic diagram of the structure of the medical judgment support system involved in this embodiment.

[0023] Figure 2 It is a schematic diagram of the structure of the medical judgment support device involved in this embodiment.

[0024] Figure 3 This is a functional block diagram of the medical information processing unit.

[0025] Figure 4 This is a functional block diagram of the analysis unit.

[0026] Figure 5 This is a conceptual diagram showing a group of operators for each combination of a type of medical information and an analysis item.

[0027] Figure 6 This is a conceptual diagram showing the content of importance information.

[0028] Figure 7 This is a conceptual diagram showing an example of a computing unit used in a process for obtaining an analysis result of each analysis item.

[0029] Figure 8 This is a conceptual diagram showing how a computing unit is formed by combining a structural model and learning information.

[0030] Fig. 9This is a diagram showing a first display example of the processing result of the medical judgment supporting device.

[0031] Fig.10 This is a diagram showing a second display example of the processing result of the medical judgment supporting device.

[0032] Explanation of symbols

[0033] 10-medical judgment support system, 12-medical equipment, 14-user terminal, 16-medical judgment support device, 18-communication line, 20-communication interface, 22-memory, 24-medical information DB, 26-calculator, 28-importance information, 30-learning model for comprehensive judgment, 32-prognosis information DB, 34-processor, 36-medical information processing unit, 38-analysis unit, 40-comprehensive judgment unit, 42-display control unit. DETAILED DESCRIPTION

[0034] Figure 1 1 is a schematic diagram of the structure of the medical judgment support system 10 involved in the present embodiment. The medical judgment support system 10 is configured to include one or more medical devices 12, a user terminal 14, and a medical judgment support apparatus 16. In the present embodiment, the medical judgment support system 10 has a plurality of medical devices 12, but the medical device 12 may be one. The plurality of medical devices 12, the user terminal 14, and the medical judgment support apparatus 16 are connected in a communicative manner via a communication line 18 including a WAN (Wide Area Network) or a LAN (Local Area Network).

[0035] The medical device 12 is used to obtain medical information about the subject (especially its judgment target part). The multiple medical devices 12 possessed by the medical judgment support system 10 are medical devices 12 of different types. For example, the multiple medical devices 12 are ultrasonic diagnostic devices, X-ray CT devices, MRI (Magnetic Resonance Imaging) devices, nuclear medicine examination devices, blood test devices, gene test devices, electrocardiographs, or cardiac catheter examination devices. The different types of medical devices 12 obtain different medical information about the judgment target part. For example, the ultrasonic diagnostic device obtains ultrasonic image data of the judgment target part, the X-ray CT device obtains CT image data of the judgment target part, and the MRI device obtains a nuclear magnetic resonance image of the judgment target part. The medical information obtained by each medical device 12 is sent to the medical judgment support device 16 via the communication line 18.

[0036] The user terminal 14 is a terminal used by a user of the medical judgment support system 10. The user terminal 14 may be, for example, a personal computer or a tablet terminal, etc. The user terminal 14 is configured to include a processor, a memory, a communication interface, an input interface, and a display.

[0037] Figure 2 : is a schematic diagram of the structure of the medical judgment support device 16. In this embodiment, the medical judgment support device 16 is a server, and the medical judgment support device 16 can be any device as long as it can perform the functions described below. In addition, as the medical judgment support device 16, each function described below can also be achieved by the cooperation of multiple physically separated devices.

[0038] The communication interface 20 is composed of, for example, a network adapter, etc. The communication interface 20 performs the function of communicating with other devices (medical equipment 12 and user terminal 14) via the communication line 18. In particular, in the present embodiment, the communication interface 20 receives medical information from a plurality of medical equipment 12. In addition, the communication interface 20 receives instructions from the user terminal 14 and sends information indicating processing results to the user terminal 14.

[0039] The memory 22 is configured to include a HDD (Hard Disk Drive), an SSD (Solid State Drive), an eMMC (embedded Multi Media Card), a ROM (Read Only Memory), or a RAM (Random Access Memory). The memory 22 stores a medical judgment support program for operating each part of the medical judgment support device 16. In addition, the medical judgment support program can also be stored in a computer-readable non-temporary storage medium such as a USB (Universal Serial Bus) memory or a CD-ROM. The medical judgment support device 16 can read and execute the medical judgment support program from such a storage medium.

[0040] And, if Figure 2As shown, the memory 22 stores a medical information DB (DataBase) 24, a computing unit 26, importance information 28, a comprehensive judgment learning model 30, and a prognosis information DB 32. The medical information DB 24 is a database that accumulates and stores medical information received from each medical device 12. In addition, the medical information sent from each medical device 12 is attached with attribute information indicating the attributes of the medical information, such as the device ID that uniquely identifies the medical device 12 that has acquired the medical information, the subject ID that uniquely identifies the subject, the determination target part information indicating the determination target part, and the action condition information indicating the action condition (for example, action mode, etc.) of the medical device 12 when the medical information is acquired. The medical information is stored in the medical information DB 24 in association with the attribute information. In addition, the medical information stored in the medical information DB 24 also includes information obtained through the doctor's interview of the subject (for example, electronic medical records, etc.).

[0041] The calculation unit 26, the importance information 28, the comprehensive determination learning model 30, and the prognosis information DB 32 will be described later.

[0042] The processor 34 is configured to include at least one of a general-purpose processor (e.g., a CPU (Central Processing Unit)) and a dedicated processor (e.g., a GPU (Graphics Processing Unit), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a programmable logic device). The processor 34 may not be composed of a single processing device, but may be composed of a plurality of processing devices that exist in physically separate locations in cooperation with each other.

[0043] like Figure 2 As shown, the processor 34 functions as a medical information processing unit 36, an analysis unit 38, a comprehensive determination unit 40, and a display control unit 42 according to the medical determination support program stored in the memory 22. In detail, as described later, the processor 34 determines the state of the determination target part of the subject by performing the above functions.

[0044] First, the summary of the determination process of the state of the determination target part of the subject as the whole of the processor 34 is described. The processor 34 receives determination instruction information including the subject as the determination target, the determination target part of the subject, and the determination content from the user terminal 14. In the present embodiment, the determination target part is the heart, and the determination content is to determine the state of the heart of the subject, especially the state related to ischemic heart disease (specifically, the lesion state and prognosis (prevention prediction, progression prediction, recurrence risk, etc.) of ischemic heart disease). Of course, the determination target part and the determination content are not limited to this. When the processor 34 receives the determination instruction information, it analyzes a plurality of different analysis items based on the medical information about the subject stored in the medical information DB 24 to determine the determination content (in the present embodiment, the determination of the state related to ischemic heart disease) to obtain a plurality of analysis results about the plurality of analysis items. In the present embodiment, the plurality of analysis items used for determining the state related to ischemic heart disease are coronary artery stenosis, myocardial ischemia, myocardial viability, cardiac function, and ischemic heart disease risk. Then, the processor 34 obtains a determination result on the determination content by comprehensively determining the plurality of analysis results on the plurality of analysis items. The following describes in detail the functions performed by the processor 34 .

[0045] Figure 3 2 is a functional block diagram of the medical information processing unit 36. The medical information processing unit 36 ​​performs processing for providing information necessary for analysis to the analyzing unit 38 described later, regarding each analysis item for determining the determination content indicated by the determination instruction information.

[0046] First, the medical information processing unit 36 ​​determines a plurality of different analysis items for determining the determination content indicated by the received determination instruction information. For example, by pre-storing analysis item information that associates the determination content with the plurality of analysis items in the memory 22, the medical information processing unit 36 ​​can determine the plurality of analysis items corresponding to the determination content by referring to the analysis item information.

[0047] like Figure 3 As shown, the medical information processing unit 36 ​​has functions as a detection unit 36a and a measurement unit 36b.

[0048] The detection unit 36a extracts information required for the analysis of the determined multiple analysis items related to the subject indicated by the received judgment instruction information from the medical information DB 24. The information required for the analysis of each analysis item can also be associated with the above-mentioned analysis item information. The detection unit 36a can determine the information required for the analysis of each analysis item by referring to the analysis item information and the attribute information associated with the medical information. Figure 3In the figure, for convenience, each detection unit 36a (coronary artery stenosis information detection unit 36a-1, myocardial ischemia information detection unit 36a-2, cardiac function information detection unit 36a-3 and ischemic heart disease risk information detection unit 36a-4) for extracting required information is shown separately according to the analysis item.

[0049] The coronary artery stenosis information detection unit 36a-1 extracts information necessary for analyzing the analysis item "coronary artery stenosis" from the medical information DB24. For example, the coronary artery stenosis information detection unit 36a-1 extracts information necessary for analyzing the analysis item "coronary artery stenosis" from the medical information DB24.

[0050] Resting cardiac ultrasound images, stress cardiac ultrasound images, or intravascular ultrasound images (IVUS (IntraVascular UltraSound)) acquired by ultrasound diagnostic equipment

[0051] Coronary CT angiography (CCTA (Coronary CT Angiography)) or FFR (Fractional Flow Reserve)-CT images obtained by X-ray CT equipment,

[0052] Coronary artery MRA (Magnetic Resonance Angiography) images or stress myocardial perfusion MRI images acquired by an MRI device, or

[0053] Cardiac nuclear medicine examination results obtained by nuclear medicine examination equipment

[0054] At least 1 of .

[0055] The myocardial ischemia information detection unit 36a-2 extracts information necessary for analyzing the analysis items "myocardial ischemia" and "myocardial vitality" from the medical information DB24.

[0056] Resting cardiac ultrasound images, stress cardiac ultrasound images, myocardial contrast ultrasound images or myocardial strain ultrasound images acquired by an ultrasonic diagnostic device,

[0057] Dual-energy CT images, photon counting CT images, delayed angiography CT images, coronary CT angiography images (CCTA (Coronary CT Angiography)), FFR (Fractional Flow Reserve)-CT images or stress perfusion CT images acquired by X-ray CT devices,

[0058] Dynamic (Cine) MRI images, delayed contrast MRI images, or stress myocardial perfusion MRI images acquired by an MRI device, or

[0059] Cardiac nuclear medicine examination results obtained by nuclear medicine examination equipment

[0060] At least 1 of .

[0061] The cardiac function information detection unit 36a-3 extracts information necessary for analyzing the analysis item "cardiac function" from the medical information DB24. For example, the cardiac function information detection unit 36a-3 extracts information necessary for analyzing the analysis item "cardiac function" from the medical information DB24.

[0062] Ultrasonic cardiac images acquired by ultrasonic diagnostic equipment,

[0063] Cardiac CT images or FFR (Fractional Flow Reserve)-CT images acquired by X-ray CT devices,

[0064] A cardiac MRI image or delayed contrast MRI image acquired by an MRI device, or

[0065] Cardiac nuclear medicine examination results obtained by nuclear medicine examination equipment

[0066] At least 1 of .

[0067] The ischemic heart disease risk information detection unit 36a-4 extracts information required for analyzing the analysis item "ischemic heart disease risk" from the medical information DB 24. For example, the ischemic heart disease risk information detection unit 36a-4 extracts biochemical information and gene factor information obtained from a blood tester, a gene tester, etc., and information obtained by a doctor's interview of a subject (electronic medical records, etc.) from the medical information DB 24.

[0068] As described above, the detection unit 36 ​​a can acquire various types of medical information acquired from various types of medical devices 12 as information required for analysis of each analysis item.

[0069] The measuring unit 36b measures the parameters required for the analysis of the determined multiple analysis items based on the information detected by the detecting unit 36a. The parameters required for the analysis of each analysis item may also be associated with the above-mentioned analysis item information, and the measuring unit 36b can determine the parameters required for the analysis of each analysis item by referring to the analysis item information. Figure 3 For convenience, the measurement units 36b (coronary artery stenosis measurement unit 36b-1, myocardial ischemia measurement unit 36b-2, myocardial vitality measurement unit 36b-3, and cardiac function measurement unit 36b-4) that measure the required parameters are shown separately according to the analysis items.

[0070] The coronary artery stenosis measuring unit 36b-1 measures parameters required for the analysis of the analysis item "coronary artery stenosis" based on the information extracted by the coronary artery stenosis information detecting unit 36a-1. For example, the coronary artery stenosis measuring unit 36b-1 measures at least one of left ventricular output (LVEF (Left Ventricle Ejection Fraction: Left ventricular ejection fraction)), left ventricular end-systolic volume, left ventricular long-axis systolic function index (GLS (Global Longitudinal Strain: Global longitudinal strain), S', MAPSE (Mitral Annular Plane Systolic Excursion: Mitral valve annular plane systolic displacement)), coronary artery calcification score (CACS (Coronary Artery Calcium Score)), napkin ring sign, lesion blood vessel diameter increase value, coronary artery blood flow reserve capacity (CFR (Coronary Flow Reserve)) or coronary blood flow reserve ratio (FFR).

[0071] The myocardial ischemia measuring unit 36b-2 measures parameters required for the analysis of the analysis item "myocardial ischemia" based on the information extracted by the myocardial ischemia information detecting unit 36a-2 or the ischemic heart disease risk information detecting unit 36a-4. For example, the myocardial ischemia measuring unit 36b-2 measures at least one of a parameter related to the medical history collection or physical sign of the subject, a parameter related to the exercise stress electrocardiogram, a parameter related to the dynamic (Holter) electrocardiogram, or a parameter related to the cardiac catheter.

[0072] The myocardial vitality measuring unit 36b-3 measures parameters required for the analysis of the analysis item "myocardial vitality" based on the information extracted by the myocardial ischemia information detecting unit 36a-2. For example, the myocardial vitality measuring unit 36b-3 measures at least one of the parameters related to myocardial blood flow, wall motion, myocardial fat metabolism, contractile reserve capacity, myocardial fibrosis, myocardial extracellular fluid volume quantification (ECV (Extracellular Volume Fraction)), wall thickness, coronary blood flow reserve ratio (FFR), left ventricular output (LVEF), end systolic volume (ESV (End Systolic Volume)), end diastolic volume (EDV (End Diastolic Volume)), myocardial weight, fatty degeneration, calcification, cardiac catheter or electrocardiogram.

[0073] The cardiac function measuring unit 36b-4 measures parameters required for the analysis of the analysis item "cardiac function" based on the information extracted by the cardiac function information detecting unit 36a-3. For example, the cardiac function measuring unit 36b-4 measures at least one of the parameters related to left ventricular output (LVEF), primary output or cardiac power output, left ventricular wall motion score index (WMSI (Wall Motion Score Index)), left ventricular long-axis systolic function index (GLS, S', MAPSE), left ventricular pressure increase rate, left ventricular inflow blood flow velocity waveform (maximum velocity ratio of early diastolic wave to atrial systolic wave or deceleration time of early diastolic wave), pulmonary venous blood flow velocity waveform, systolic positive wave, early diastolic positive wave, atrial systolic negative wave, early diastolic mitral annular maximum movement velocity, tricuspid regurgitation maximum blood flow velocity, systolic pulmonary artery pressure, right ventricular systolic function, right ventricular diastolic function, myocardial strain analysis, and identification of ischemic / non-ischemic myocardial damage.

[0074] In addition, the above parameters are known parameters, and the existing methods can be used to measure them, so the detailed contents of the measurement methods of the parameters are omitted here. In this specification, the term "medical information" is a concept that also includes information required for the analysis of multiple analysis items detected by the detection unit 36a and parameters about multiple analysis items measured by the measurement unit 36b. In the medical information, the information required for the analysis of multiple analysis items detected by the detection unit 36a and the parameters about multiple analysis items measured by the measurement unit 36b are specifically recorded as "medical information obtained by the medical information processing unit 36".

[0075] Figure 4 : is a functional block diagram of the analysis unit 38. The analysis unit 38 analyzes a plurality of analysis items based on a plurality of medical information related to the subject (particularly its determination target part) acquired by the medical information processing unit 36, and outputs a plurality of analysis results related to the determination target part of the subject for the plurality of analysis items. In particular, in the present embodiment, the analysis unit 38 outputs a disease state related to each of a plurality of disease categories related to the determination target part, i.e., the heart, as a plurality of analysis results. Figure 4 In the figure, for convenience, each analysis unit 38 (coronary artery stenosis analysis unit 38-1, myocardial ischemia analysis unit 38-2, myocardial vitality analysis unit 38-3, cardiac function analysis unit 38-4 and ischemic heart disease risk analysis unit 38-5) for extracting required information is shown separately according to the analysis item.

[0076] In this embodiment, each analysis unit 38 outputs analysis results using a group of computing units 26 that correspond to the plurality of types of medical information acquired by the medical information processing unit 36 ​​and analyze the corresponding types of medical information. In this specification, the information output by each computing unit 26 is described as partial analysis information.

[0077] refer to Figure 5 Provide detailed explanation. Figure 5 1 is a conceptual diagram showing a group of computing units 26 for each combination of a type of medical information and an analysis item. Figure 5 The illustrated set of operators 26 is stored in the memory 22. Figure 5 In the figure, multiple analysis items (coronary artery stenosis, myocardial ischemia, myocardial vitality and cardiac function) are arranged horizontally, and multiple medical information (resting cardiac ultrasound, stress cardiac ultrasound, CCTA, FFR-CT, coronary artery MRA, stress myocardial perfusion MRI, cardiac nuclear medicine examination and IVUS) are arranged vertically. In this embodiment, Figure 5 The various medical information shown includes medical information acquired using a plurality of different medical devices 12 (eg, ultrasonic diagnostic apparatus, X-ray CT apparatus, MRI apparatus, etc.). In particular, in the present embodiment, Figure 5 The illustrated plurality of medical information includes a plurality of medical images acquired using a plurality of different medical devices 12 .

[0078] Figure 5 The respective computing units 26 shown correspond to the respective analysis items and the respective types of medical information, and are different computing units 26. For example, the computing unit 26-1 that outputs partial analysis information related to coronary artery stenosis based on resting cardiac ultrasound, the computing unit 26-2 that outputs partial analysis information related to myocardial ischemia based on resting cardiac ultrasound, and the computing unit 26-3 that outputs partial analysis information related to coronary artery stenosis based on stress cardiac ultrasound are different computing units 26.

[0079] Here, in the present embodiment, each operator 26 is a learning model for analysis. That is, each operator 26 as a learning model for analysis is pre-trained in a manner that predicts and outputs partial analysis information about the corresponding analysis item with high accuracy when the corresponding type of medical information is input. The learning model for analysis may be, for example, a neural network, and as a learning model for analysis, any model may be used as long as it performs the functions described below. Furthermore, each operator 26 does not necessarily have to be a learning model. Each operator 26 may be any operator as long as it can output partial analysis information about the corresponding analysis item based on the corresponding type of medical information. For example, each operator may be a LUT.

[0080] The coronary artery stenosis analyzing unit 38-1 performs the operation of the plurality of computing units 26 (in Figure 5In the embodiment, a plurality of operators 26) corresponding to the analysis item "coronary artery stenosis" and arranged vertically respectively input corresponding types of medical information to obtain a plurality of partial analysis information. Then, the coronary artery stenosis analysis unit 38-1 calculates the analysis result on the analysis item "coronary artery stenosis" based on the plurality of partial analysis information output by the plurality of operators 26.

[0081] The myocardial ischemia analysis unit 38-2 performs a plurality of operations 26 (in Figure 5 In the example, a plurality of operators 26) corresponding to the analysis item "myocardial ischemia" and arranged vertically respectively input corresponding types of medical information to obtain a plurality of partial analysis information. Then, the myocardial ischemia analysis unit 38-2 calculates the analysis result of the analysis item "myocardial ischemia" based on the plurality of partial analysis information output by the plurality of operators 26.

[0082] The myocardial vitality analyzing unit 38-3 performs a plurality of operations 26 (in Figure 5 In the example, the plurality of operators 26) corresponding to the analysis item "myocardial vitality" and arranged vertically respectively input the corresponding types of medical information to obtain a plurality of partial analysis information. In addition, the myocardial vitality analysis unit 38-3 calculates the analysis result of the analysis item "myocardial vitality" based on the plurality of partial analysis information output by the plurality of operators 26.

[0083] The cardiac function analysis unit 38-4 performs the analysis of the cardiac function and the plurality of computing units 26 (in Figure 5 In the example, a plurality of operators 26) corresponding to the analysis item "cardiac function" and arranged vertically respectively input corresponding types of medical information to obtain a plurality of partial analysis information. In addition, the cardiac function analysis unit 38-4 calculates the analysis result of the analysis item "cardiac function" based on the plurality of partial analysis information output by the plurality of operators 26.

[0084] In addition, Figure 5 (and the following Figure 6 and Figure 7 ) does not describe the computing unit 26 used by the ischemic heart disease risk analysis unit 38-5 that outputs the analysis result related to the analysis item "ischemic heart disease risk". However, the ischemic heart disease risk analysis unit 38-5 also uses the computing unit (learning model for analysis or LUT) 26 to output the analysis result, similarly to other analysis units 38. Specifically, the ischemic heart disease risk analysis unit 38-5 performs analysis related to the analysis item "ischemic heart disease risk" by inputting the information extracted by the ischemic heart disease risk information detection unit 36a-4 into the computing unit 26, and outputs the analysis result.

[0085] Figure 6 2 is a conceptual diagram showing the contents of the importance information 28 stored in the memory 22. As described above, each analysis unit 38 calculates the analysis results for each analysis item based on a plurality of medical information, but there is a case where the importance of each type of medical information in each analysis item is different. For example, as described below, when analyzing the analysis item "coronary artery stenosis", the medical information "FFR-CT" is more important than other types of medical information (that is, it has a greater impact on the analysis result than other types of medical information), and the medical information "IVUS" is less important than other types of medical information (that is, it has a smaller impact on the analysis result than other types of medical information).

[0086] The importance information 28 is information indicating the importance of each type of medical information when outputting the analysis results for each analysis item. In the present embodiment, the importance information 28 indicates the importance of each combination of the analysis item and the type of medical information in three stages: "large", "medium", and "small". Figure 6 In the figure, the combination indicated by hatching (ie, the operator 26a) indicates "large" importance, the combination indicated by white (ie, the operator 26b) indicates "medium" importance, and the combination indicated by black (ie, the operator 26c) indicates "small" importance.

[0087] For example, from the analysis item "coronary artery stenosis", the importance of the medical information "resting echocardiography", "stress echocardiography", "CCTA", "coronary artery MRA", "stress myocardial perfusion MRI", and "cardiac nuclear medicine examination" is "medium", the importance of the medical information "FFR-CT" is "high", and the importance of the medical information "IVUS" is "low". On the other hand, from the analysis item "myocardial vitality", the importance of the medical information "stress echocardiography" and "cardiac nuclear medicine examination" is "medium", the importance of the medical information "resting echocardiography" is "high", and the importance of the medical information "CCTA", "FFR-CT", "coronary artery MRA", "stress myocardial perfusion MRI", and "IVUS" is "low". In this way, the importance of a certain operator 26 (hereinafter referred to as the operator of interest) can be different from the importance of other operators 26 corresponding to the same type of medical information as the operator of interest for other analysis items. That is, as described below, from the medical information "resting cardiac ultrasound", the importance of the analysis item "coronary artery stenosis" is "medium", but the importance of the analysis item "myocardial vitality" is "large".

[0088] Before performing analysis, the analysis unit 38 determines the importance of each medical information (in other words, each operator 26) by referring to the importance information 28. In the present embodiment, a plurality of analysis units 38 are provided corresponding to the analysis items, so the analysis unit 38 determines the importance of the plurality of operators 26 for each analysis item based on the importance information 28. Then, the analysis unit 38 calculates the analysis result for each analysis item based on the partial analysis information output by each operator 26 when each medical information is input to the corresponding operator 26 and the determined importance of each operator 26.

[0089] For example, when the content of importance information 28 is Figure 6 In the case of the contents shown, when calculating the analysis result on coronary artery stenosis, the importance of the medical information “FFR-CT” is “large”, and therefore the coronary artery stenosis analyzing unit 38-1 assigns a larger weight (at least larger than the importance of the medical information with “medium” importance) to the partial analysis information output by the computing unit 26 corresponding to the medical information, and then calculates the analysis result on the analysis item “coronary artery stenosis” based on the partial analysis information output by the plurality of computing units 26 corresponding to the plurality of medical information. Furthermore, when calculating the analysis result on coronary artery stenosis, the importance of the medical information “IVUS” is “small”, and therefore the coronary artery stenosis analyzing unit 38-1 assigns a smaller weight (at least smaller than the medical information with “medium” importance) to the partial analysis information output by the computing unit 26 corresponding to the medical information, and then calculates the analysis result on the analysis item “coronary artery stenosis” based on the partial analysis information output by the plurality of computing units 26 corresponding to the plurality of medical information. The other analysis units 38 also calculate analysis results in the same manner.

[0090] According to the determined importance of each medical information (i.e., each computing unit 26), each analyzing unit 38 may not use a part of the computing units 26 in the computing unit 26 group to calculate the analysis result. Figure 7 As shown, each analysis unit 38 may not use the medical information with a "small" importance to calculate the analysis result. For example, the myocardial vitality analysis unit 38-3 may calculate the analysis result of the analysis item "myocardial vitality" using only the medical information "resting cardiac ultrasound", "stress cardiac ultrasound" and "cardiac nuclear medicine examination". In this case, the myocardial vitality analysis unit 38-3 does not use the calculation unit 26 corresponding to the medical information "CCTA", "FFR-CT", "coronary artery MRA", "stress myocardial perfusion MRI" and "IVUS" to calculate the analysis result.

[0091] Here, the importance of each type of medical information about each analysis item may change depending on the diagnosis status (in other words, state) of the subject (especially the determination target part). Therefore, in the importance information 28, it is preferable to associate the importance of each type of medical information with the diagnosis status of the subject. In this case, each analysis unit 38 determines the importance of each medical information (in other words, each operator 26) based on the importance information 28 and the diagnosis status of the subject (especially the determination target part) before performing analysis.

[0092] In this embodiment, the analysis unit 38 uses a plurality of computing units 26 corresponding to a plurality of analysis items and a plurality of medical information to calculate the analysis results. Figure 5 As shown in FIG. 1 , it is necessary to prepare a plurality of computing units 26. As the number of analysis items and types of medical information increases, the number of computing units 26 also becomes large. In particular, when the computing unit 26 is a learning model for analysis, if a plurality of computing units 26 are stored in the memory 22 in advance, the storage capacity of the memory 22 is compressed.

[0093] To address this problem, Figure 8 As shown, in this embodiment, the operator 26 as the analytical learning model is divided into a structural model 50 and learning information 52 and stored in the memory 22. The structural model 50 is data that specifies the structure of the analytical learning model (for example, the number of layers of the neural network, the number of neurons in one layer, etc.). In other words, the structural model 50 is data specified by the hyperparameters of the analytical learning model (parameters that are not adjusted by the learning process). Multiple structural models 50 can be stored in the memory 22. The learning information 52 is a parameter that is adjusted by the learning process (for example, the weight of each node of the neural network or the bias of each neuron, etc.). The learning information 52 is information obtained by the learning process.

[0094] The analysis unit 38 forms the operator 26 as the analysis learning model by combining the structural model 50 and the learning information 52. For example, when it is desired to obtain partial analysis information based on the medical information "resting echocardiography" regarding the analysis item "coronary artery stenosis", the coronary artery stenosis analysis unit 38-1 forms the operator 26 by combining the structural model 50 for combining the analysis item "coronary artery stenosis" with the medical information "resting echocardiography" and the learning information 52 for combining. Furthermore, when it is desired to obtain partial analysis information based on the medical information "resting echocardiography" regarding the analysis item "myocardial ischemia", the myocardial ischemia analysis unit 38-2 forms the operator 26 by combining the structural model 50 for combining the analysis item "myocardial ischemia" with the medical information "resting echocardiography" and the learning information 52 for combining.

[0095] The learning information 52 is different for each combination of the analysis item and the type of medical information, and therefore needs to be prepared separately, but the structural model 50 can be used in common for a plurality of combinations of the analysis item and the type of medical information. Therefore, the structural model 50 and the learning information 52 are separated and stored separately in the memory 22, and when the analysis unit 38 performs calculations, the structural model 50 and the learning information 52 are combined to form the required operator 26, thereby reducing the storage capacity of the memory 22 at least compared to the case where all the operators 26 are stored in the memory 22.

[0096] In addition, in the invention in which the analysis unit 38 determines the importance of each medical information based on the importance information 28 and calculates the analysis result based on the partial analysis information output by each calculation unit 26 while considering the determined importance, it is not necessary for the analysis unit 38 to perform analysis on a plurality of analysis items. In other words, the invention can also be applied when the analysis unit 38 analyzes a single analysis item.

[0097] Return to Figure 2 The comprehensive determination unit 40 determines the state of the determination target site based on the analysis results of the plurality of analysis items related to the determination target site obtained by the analysis unit 38. Specifically, the comprehensive determination unit 40 determines the state of the determination target site based on the output of the comprehensive determination learning model 30 when the plurality of analysis results output by the analysis unit 38 are input to the comprehensive determination learning model 30 stored in the memory 22.

[0098] The comprehensive determination learning model 30 may be, for example, a neural network. As the comprehensive determination learning model 30, any model may be used as long as the function described below is fulfilled. The comprehensive determination learning model 30 is a learning model that is pre-trained in a manner to predict and output the state of the determination target site based on the analysis results of multiple analysis items related to the determination target site. The comprehensive determination learning model 30 is trained in a learning device as the medical determination support device 16 or other device. For example, the learning device inputs the analysis results of multiple analysis items related to the determination target site into the comprehensive determination learning model 30 as learning data. The comprehensive determination learning model 30 outputs a prediction result (the state of the determination target site) for the learning data. The learning device calculates the difference between the output of the comprehensive determination learning model 30 and the state of the determination target site as the training data, and adjusts the parameters of the comprehensive determination learning model 30 in a manner that reduces the difference. By repeatedly performing this process, the learning process is improved. The trained comprehensive determination learning model 30 can predict and output the state of the determination target site with high accuracy based on the analysis results of multiple analysis items related to the determination target site.

[0099] As described above, in this embodiment, the determination object part is the heart, and the analysis unit 38 outputs the disease state related to each of the multiple disease categories related to the heart as multiple analysis results. Therefore, the comprehensive determination learning model 30 comprehensively determines the state of the heart as the determination object part based on the disease state related to each of the multiple disease categories.

[0100] Assuming that the medical judgment support device 16 does not have an analysis unit 38, all medical information acquired by the medical information processing unit 36 ​​is input into the comprehensive judgment learning model 30. Even in this case, functionally, the comprehensive judgment learning model 30 can predict and output the state of the judgment object part based on the medical information acquired by the medical information processing unit 36. However, if this is done, the number of inputs of the comprehensive judgment learning model 30 will become huge, and the amount of calculation for predicting the state of the judgment object part will also become huge, so the scale of the comprehensive judgment learning model 30 will also become huge. However, in this embodiment, the processing based on the analysis unit 38 is performed before the processing of the comprehensive judgment learning model 30 is performed, so the scale of the comprehensive judgment learning model 30 can be reduced.

[0101] The comprehensive determination unit 40 preferably determines the prognosis of the determination target site based on the prognosis information DB32 stored in the memory 22 and the state of the determined determination target site. The prognosis information DB32 is a database that accumulates and stores state change information such as time changes or treatment processes of the state of the determination target site (heart in this embodiment) of multiple subjects in the past. The comprehensive determination unit 40 determines the state change information indicating the state of the determined determination target site from the state change information stored in the prognosis information DB32, and determines the prognosis of the determination target site based on the state of the determination target site after the determined state indicated by the determined state change information. The prognosis of the determination target site includes, for example, the prognosis when the status quo is maintained without treatment and the prognosis when appropriate treatment (treatment included in the state change information) is performed.

[0102] The display control unit 42 performs processing to display the processing result of the medical judgment support device 16 on the display unit. For example, the display control unit 42 displays the processing result of the medical judgment support device 16 on the display of the user terminal 14. In particular, the display control unit 42 displays the medical information that is the basis for the analysis by the analysis unit 38 and the determination result of the comprehensive determination unit 40 on the display of the user terminal 14.

[0103] Fig. 9 FIG. 1 is a diagram showing a first display example of the processing result of the medical judgment supporting device 16. Fig. 9In the example of , an ultrasonic image USI and a coronary CT angiography image CCTA are displayed as medical information that is the basis for analysis by the analysis unit 38, a report R as the result of the determination by the comprehensive determination unit 40, and a menu area MA. The menu area MA is a GUI for inputting a user's selection or enlarged display instruction from the medical information that is the basis for analysis by the analysis unit 38 or the result of the determination by the comprehensive determination unit 40, and thumbnails or titles are displayed. In the report R, as the state of the subject's heart, it is shown that there is coronary artery calcification, no vulnerable plaque, the stenosis of the left anterior descending artery (LAD) is 0.78, and there is abnormality in stress myocardial perfusion. Furthermore, in the report R, as the risk of severity of the prognosis predicted by the comprehensive determination unit 40, it is shown that the current incidence of infarction in 5 years is 34 to 45% and the incidence of infarction after appropriate treatment is 8 to 25%.

[0104] Fig.10 FIG. 2 is a diagram showing a second display example of the processing result of the medical judgment supporting device 16. Fig.10 In the example of , there is an example in which the inspection process and diagnostic evidence (category, level, grade) of the medical information obtained as the basis for the analysis by the analysis unit 38 are displayed, and a report R' is displayed as the judgment result of the comprehensive judgment unit 40. In particular, the inspection process or diagnostic evidence of the medical information involved in the importance information 28 can be displayed in a manner that can distinguish the importance of the medical information by highlighting, etc. In addition, the diagnostic evidence can be sorted from top to bottom in the order of importance. In the report R', blood pressure, blood sugar level, regularity of smoking, presence or absence of lipid abnormalities, etc. are shown as risk factors for the onset of the disease. Moreover, in the report R', as the onset prediction predicted by the comprehensive judgment unit 40, it is shown that the current incidence of infarction in 10 years is 30 to 65% and the incidence of infarction when appropriate prevention is implemented is 13 to 30%.

[0105] As mentioned above, although embodiment which concerns on this invention was demonstrated, this invention is not limited to the said embodiment, Unless it deviates from the summary of this invention, various changes can be made.

Claims

1. A medical judgment support device comprising: an analysis unit that outputs an analysis result related to an inspection target part of the subject based on a plurality of types of medical information related to the subject, and outputs the analysis result using a group of operators that respectively correspond to the types of the medical information and output partial analysis information for operating the analysis result based on the corresponding types of the medical information, The medical judgment support device is characterized in that the analysis unit performs the following processing: determining the importance of each of the computing units included in the computing unit group based on importance information indicating the importance of each type of the medical information when outputting the analysis result; and The analysis result is calculated based on the partial analysis information output by each of the calculation units when the corresponding medical information is input and the determined importance of each of the calculation units.

2. The medical judgment support device according to claim 1, characterized in that: In the importance information, the importance of each type of the medical information is associated with the diagnostic condition of the subject. The analyzing unit determines the importance of each of the computing units included in the computing unit group based on the importance information and the current diagnostic condition of the subject.

3. The medical judgment support device according to claim 1, wherein: The analysis unit calculates the analysis result without using some of the operators in the operator group, based on the determined importance of each of the operators.

4. The medical judgment support device according to claim 1, wherein: The analysis unit includes a plurality of the computing unit groups corresponding to a plurality of analysis items related to the inspection object part. The analysis unit performs the following processing: For each of the analysis items, determining the importance of each of the computing units included in the plurality of computing unit groups based on the importance information indicating the importance of each type of medical information; and For each of the analysis items, the analysis result is calculated based on the partial analysis information output by each of the calculation units when the corresponding medical information is input and the importance of each of the calculation units determined, thereby outputting a plurality of analysis results for the plurality of analysis items.

5. The medical judgment support device according to claim 4, characterized in that: The importance of a computing unit for one of the analysis items is different from the importance of another computing unit corresponding to the same type of medical information as the computing unit for another of the analysis items.

6. The medical judgment support device according to claim 4, characterized in that: The medical judgment support device also includes a comprehensive judgment unit, which judges the state of the inspection object part based on the output of the comprehensive judgment learning model when the comprehensive judgment learning model is input with the multiple analysis results output by the analysis unit, and the comprehensive judgment learning model is trained to predict and output the state of the inspection object part based on multiple analysis results of the multiple analysis items.

7. The medical judgment support device according to claim 6, characterized in that: The inspection object part is the heart, the analyzing section outputs a disease state associated with each of a plurality of disease categories related to the heart as the plurality of analysis results, The comprehensive determination unit comprehensively determines the state of the heart of the subject based on the disease state associated with each of a plurality of disease types.

8. The medical judgment support device according to claim 6, wherein: The medical determination support device further includes a display control unit configured to display the medical information serving as a basis for the analysis by the analysis unit and the determination result of the comprehensive determination unit on a display unit.

9. The medical judgment support device according to claim 8, characterized in that: The display control unit displays, on the display unit, a screen for selecting any one of the plurality of medical information serving as a basis for analysis by the analysis unit or the determination result of the comprehensive determination unit.

10. The medical judgment support device according to claim 8, wherein: The display control unit displays the medical information in a manner that allows the importance of the medical information to be determined based on the importance information.

11. A computer-readable non-transitory recording medium having a medical judgment support program recorded thereon, characterized in that: The medical judgment support program causes the computer to function as an analysis unit. The analysis unit outputs an analysis result related to an inspection target part of the subject based on a plurality of types of medical information related to the subject, and outputs the analysis result using a group of operators that respectively correspond to the types of the medical information and are trained to output partial analysis information for calculating the analysis result based on the corresponding types of the medical information, The analysis unit performs the following processing: determining the importance of each of the computing units included in the computing unit group based on importance information indicating the importance of each type of the medical information when outputting the analysis result; and The analysis result is calculated based on the partial analysis information output by each of the calculation units when the corresponding medical information is input and the determined importance of each of the calculation units.

Images