Medical information processing apparatus, medical information processing method, and program

Abstract

To improve the accuracy of determining a lesion.SOLUTION: A medical information processing apparatus according to an embodiment includes a first analysis unit, a second analysis unit, a determination unit, and a third analysis unit. The first analysis unit executes a first analysis based on a medical image of a subject. The second analysis unit executes a second analysis based on a specimen examination result related to the subject. The determination unit determines the degree of matching between the result of the first analysis executed by the first analysis unit and the result of the second analysis executed by the second analysis unit. The third analysis unit executes a third analysis based on the medical image and the specimen examination result according to analysis conditions based on a result of the determination of the matching degree executed by the determination unit, and determines diagnosis support information based on the medical image and the specimen examination result.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The embodiments disclosed in this specification and the drawings relate to a medical information processing device, a medical information processing method, and a program. [Background technology]

[0002] Conventionally, imaging diagnosis is known for diagnosing cancer in a subject based on image information of the subject. However, since the image that should be judged as a malignant cancer may differ depending on the cancer subtype, it may be difficult to diagnose cancer based on image information alone.

[0003] Liquid biopsy, which uses bodily fluids such as blood to test for cancer in a subject, has been known as a test that places less of a burden on the subject. In recent years, diagnostic techniques have been developed that add test information from liquid biopsy to image information obtained by image diagnosis. However, image information is morphological information, whereas test information from liquid biopsy is property information, so there are cases where the two do not correspond. Therefore, it has been difficult to improve the accuracy of lesion determination in the past. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Special Publication No. 2023-530453 [Patent Document 2] Special Publication No. 2023-99261 [Patent Document 3] Special Publication No. 2022-23821 Summary of the Invention [Problem to be solved by the invention]

[0005] One of the problems to be solved by the embodiments disclosed in this specification and the drawings is to improve the accuracy of lesion determination. However, the problems to be solved by the embodiments disclosed in this specification and 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]

[0006] The medical information processing apparatus according to the embodiment includes a first analysis unit, a second analysis unit, a determination unit, and a third analysis unit. The first analysis unit executes a first analysis based on a medical image of a subject. The second analysis unit executes a second analysis based on a specimen test result of the subject. The determination unit executes a determination of the degree of consistency between a result of the first analysis by the first analysis unit and a result of the second analysis by the second analysis unit. The third analysis unit executes a third analysis based on the medical image and the specimen test result under analysis conditions based on the determination result of the degree of consistency by the determination unit, and determines diagnostic support information based on the medical image and the specimen test result. [Brief description of the drawings]

[0007] [Figure 1] FIG. 1 is a block diagram showing an example of the configuration of a medical information processing system according to an embodiment. [Diagram 2] FIG. 2 is a block diagram showing an example of the arrangement of a medical image processing apparatus according to an embodiment. [Diagram 3] FIG. 3 is a flowchart showing an example of the operation of the medical image processing apparatus according to the embodiment. [Figure 4] FIG. 4 is a diagram showing an example of calculation of a consistency degree in an example of the operation of the medical image processing apparatus according to the embodiment. [Diagram 5] FIG. 5 is a flowchart showing a process of setting analysis conditions for a third analysis in an example of the operation of the medical image processing apparatus according to the embodiment. [Figure 6] FIG. 6 is a diagram showing an example of setting analysis conditions for a third analysis in an example of the operation of the medical image processing apparatus according to the embodiment. [Figure 7]FIG. 7 is a diagram showing the contribution of a biomarker attribute to a second analysis in an example of the operation of the medical information processing apparatus according to the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Hereinafter, an embodiment of a medical image processing device will be described with reference to the drawings. In the following description, components having substantially the same functions and configurations are denoted by the same reference numerals, and duplicated descriptions will be given only when necessary.

[0009] Fig. 1 is a block diagram showing an example of the configuration of a medical information processing system 100 according to an embodiment. Fig. 2 is a block diagram showing an example of the configuration of a medical information processing apparatus 3 according to the first embodiment.

[0010] As shown in FIG. 1, the medical information processing system 100 according to the first embodiment includes a medical image diagnostic apparatus 1, an inspection apparatus 2, and a medical information processing apparatus 3. The medical image diagnostic apparatus 1, the inspection apparatus 2, and the medical information processing apparatus 3 are communicably connected via a network 200. That is, the medical information processing apparatus 3 is provided on the medical information processing system 100 connected via the network 200. Note that the medical information processing apparatus 3 may be an offline apparatus. In that case, the medical information processing apparatus 3 may acquire information from the medical image diagnostic apparatus 1 and the inspection apparatus 2 via a portable recording medium.

[0011] The devices included in the medical information processing system 100 may be in a state in which they can communicate with each other directly or indirectly, for example, via an in-hospital LAN (Local Area Network) installed in a hospital. In addition, the medical information processing system 100 shown in FIG. 1 may be communicably connected to devices other than those shown in the figure (for example, a server that stores medical information, etc.).

[0012] For example, the medical information processing system 100 may include various systems such as an Electronic Health Records (EHR) system, a Picture Archiving and Communication System (PACS), a Hospital Information System (HIS), a Radiology Information System (RIS), a diagnostic report system, and a Laboratory Information System (LIS).

[0013] The medical image diagnostic apparatus 1 captures an image of a subject to collect medical images. Then, the medical image diagnostic apparatus 1 transmits the collected medical images to the medical information processing apparatus 3. For example, the medical image diagnostic apparatus 1 may be an X-ray diagnostic apparatus such as a mammography apparatus, an X-ray CT (Computed Tomography) apparatus, an MRI (Magnetic Resonance Imaging) apparatus, an ultrasound diagnostic apparatus, a SPECT (Single Photon Emission Computed Tomography) apparatus, a PET (Positron Emission Computed Tomography) apparatus, or the like.

[0014] The testing device 2 is operated by a clinical laboratory technician or the like to perform specimen testing to analyze specimens such as blood and urine obtained from patients. The specimen testing includes, for example, pathological testing, blood and biochemical testing, liquid biopsy testing, general testing of urine and stool, etc., immunoserological testing, genetic testing, microbiological testing, blood transfusion and organ transplant-related testing, etc. The testing device 2 then transmits the test results of the specimen testing (measurement data, etc.) to the medical information processing device 3.

[0015] For example, the testing device 2 may be an automatic biochemical analyzer, an automatic immunoanalyzer, a flow cytometer, a genetic analyzer, a protein analyzer, an extracellular vesicle analyzer, a blood circulating tumor cell detection device, or the like. The flow cytometer is an instrument that performs flow cytometry. Flow cytometry is an analytical method in which a suspension of a measurement object is treated as a high-speed fluid, and the fluid is irradiated with laser light, mercury light, or the like to generate scattered light and fluorescence, thereby measuring the size, amount, and the like of the measurement object.

[0016] A gene analyzer is a device that analyzes gene sequences. For example, a gene analyzer is a device that amplifies nucleic acid molecules extracted from a biological sample using a PCR (Polymerase Chain Reaction) method or the like, and determines the presence or absence of a specific gene sequence or quantifies the amount of the sequence, or a device that analyzes sequence information of a nucleic acid molecule. A protein analyzer is, for example, an automatic immunoassay analyzer or a highly sensitive protein detector (e.g., Quanterix's Single Molecule Assay device SIMOA (registered trademark)).

[0017] The extracellular vesicle analysis device is, for example, a single extracellular vesicle analysis device (e.g., EXOVIEW (registered trademark) from NanoView Biosciences, Inc., etc.), and the blood circulating tumor cell detection device is, for example, a blood circulating tumor cell detection device using a nanomagnetic bead method (e.g., CELLSEARCH (registered trademark) system from Veridex, Inc., etc.), etc.

[0018] The medical information processing device 3 acquires various information from the medical image diagnostic device 1 and the inspection device 2, and performs various information processing using the acquired information. For example, the medical information processing device 3 is realized by computer equipment such as a server, a workstation, a personal computer, or a tablet terminal.

[0019] 2, the medical information processing device 3 includes an input interface 31, an output interface 32, a communication interface 33, a storage circuit 34, and a processing circuit 35. The input interface 31, the output interface 32, the communication interface 33, and the storage circuit 34 are connected to the processing circuit 35.

[0020] The input interface 31 accepts various instructions and input operations of information from an operator. Specifically, the input interface 31 converts the input operations accepted from the operator into electrical signals and outputs them to the processing circuit 35. For example, the input interface 31 is realized by a trackball, a switch button, a mouse, a keyboard, a touch pad that performs input operations by touching the operation surface, a touch screen in which a display screen and a touch pad are integrated, a non-contact input circuit using an optical sensor, and a voice input circuit. Note that the input interface 31 is not limited to only those that have physical operation parts such as a mouse and a keyboard. For example, an example of the input interface 31 also includes an electrical signal processing circuit that receives an electrical signal corresponding to an input operation from an external input device provided separately from the device and outputs the electrical signal to a control circuit.

[0021] The output interface 32 outputs various information. For example, the output interface 32 includes a display. The display outputs medical images generated by the processing circuitry 35 and a GUI (Graphical User Interface) for receiving various operations from an operator. For example, as the display, a liquid crystal display (LCD), a cathode ray tube (CRT) display, an organic electro luminescence display (OELD), a plasma display, or any other arbitrary display can be used as appropriate. The output interface 32 may include a speaker.

[0022] The communication interface 33 controls communication between each device in the medical information processing system 100. Specifically, the communication interface 33 receives various information from each device and outputs the received information to the processing circuit 35. For example, the communication interface 33 is realized by a network card, a network adapter, a NIC (Network Interface Controller), or the like.

[0023] The memory circuitry 34 is a non-transient storage device that stores various information, and is, for example, a hard disk drive (HDD), an optical disk, a solid state drive (SSD), an integrated circuit storage device, etc. The memory circuitry 34 stores, for example, a control program that controls the medical information processing device 3 and various data used in executing the control program. The memory circuitry 34 may be a drive device that reads and writes various information between a portable storage medium such as a compact disc (CD), a digital versatile disc (DVD), and a flash memory, or a semiconductor memory element such as a random access memory (RAM), in addition to an HDD or an SSD.

[0024] The processing circuit 35 is a circuit that controls the operation of the entire medical information processing device 3 in response to an electrical signal of an input operation input from the input interface 31. For example, the processing circuit 35 includes a first analysis function 351, a second analysis function 352, a determination function 353, a third analysis function 354, and an analysis condition setting function 355. The first analysis function 351 is an example of a first analysis unit. The second analysis function 352 is an example of a second analysis unit. The determination function 353 is an example of a determination unit. The third analysis function 354 is an example of a third analysis unit. The analysis condition setting function is an example of an analysis condition setting unit.

[0025] Here, for example, each processing function executed by a first analysis function 351, a second analysis function 352, a determination function 353, a third analysis function 354, and an analysis condition setting function 355, which are components of the processing circuit 35 shown in Fig. 2, is recorded in the storage circuit 34 in the form of a program executable by a computer. The processing circuit 35 is, for example, a processor. The processor constituting the processing circuit 35 realizes a function corresponding to each read program by reading each program from the storage circuit 34 and executing it. In other words, the processing circuit 35 in a state in which each program has been read has each function shown in the processing circuit 35 of Fig. 2.

[0026] 2 shows a case where the first analysis function 351, the second analysis function 352, the determination function 353, the third analysis function 354, and the analysis condition setting function 355 are realized by a single processing circuit 35, but the embodiment is not limited to this. For example, the processing circuit 35 may be configured by combining multiple independent processors, and each processor may execute each program to realize each processing function. Furthermore, each processing function of the processing circuit 35 may be realized by being appropriately distributed or integrated in a single or multiple processing circuits.

[0027] The first analysis function 351 executes a first analysis based on a medical image related to a subject. In the first analysis, the first analysis function 351 may determine the presence or absence of a mass in the medical image. In addition, when it is determined in the first analysis that a mass is present, the first analysis function 351 may determine whether the mass is benign or malignant. The first analysis may be performed by setting a region of interest (ROI) on the medical image where the presence of a mass is suspected. The first analysis may be performed using machine learning ML. The first analysis may be performed using a trained model that has learned about the mass on the medical image. The machine learning ML may be deep learning DL using a neural network. When information indicating the region of interest set in the medical image diagnostic device 1 is acquired by the medical information processing device 3, the first analysis function 351 may perform the first analysis based on the information indicating the region of interest. The first analysis function 351 may calculate the result of the first analysis together with its likelihood.

[0028] The first analysis function 351 may perform the first analysis further based on past medical information of the subject. For example, the first analysis function 351 may obtain the past medical information of the subject from an electronic medical chart or a personal health record (PHR) and use the obtained medical information in the first analysis.

[0029] The second analysis function 352 executes a second analysis based on the specimen test results of the subject. The specimen test results may include values ​​of biomarkers contained in a specimen (e.g., body fluids such as blood and urine) collected from the subject. The second analysis function 352 may execute the second analysis based on the values ​​of the biomarkers. That is, the second analysis function 352 may execute the second analysis based on the results of a liquid biopsy of the subject. In the liquid biopsy, analysis results of genes (DNA, RNA, mRNA, cfDNA, etc.), proteins, circulating cancer cells (CTCs), etc. are obtained. The second analysis function 352 may determine the presence or absence of a lesion based on whether or not a value of a biomarker indicating a lesion is detected in the second analysis. The lesion is, for example, cancer such as breast cancer. The second analysis function 352 may execute the second analysis using machine learning ML. The second analysis function 352 may execute the second analysis using deep learning DL as the machine learning ML. The second analysis function 352 may calculate the result of the second analysis together with its likelihood. In addition, in the second analysis using machine learning, the contribution of each biomarker value to the analysis result may be calculated together with the analysis result.

[0030] The second analysis function 352 may perform the second analysis further based on past medical information of the subject. For example, the second analysis function 352 may obtain the past medical information of the subject from an electronic medical chart or a personal health record (PHR) and use the obtained information in the second analysis.

[0031] The determination function 353 executes a determination of the degree of consistency between the result of the first analysis by the first analysis function 351 and the result of the second analysis by the second analysis function 352. The determination function 353 may determine the degree of consistency by taking into consideration the likelihood of the result of the first analysis and the likelihood of the result of the second analysis.

[0032] The third analysis function 354 executes a third analysis based on the medical image and the specimen test results under analysis conditions based on the result of the consistency determination by the determination function 353. The third analysis function 354 executes the third analysis to determine diagnostic support information based on the medical image and the specimen test results. The diagnostic support information may be information indicating whether the tumor is benign or malignant. The third analysis function 354 executes the third analysis using machine learning, i.e., a learning algorithm.

[0033] The analysis condition setting function 355 sets analysis conditions for the third analysis by setting at least one of machine learning parameters (i.e., algorithm parameters), mode, and network used in the third analysis. The third analysis function 354 executes the third analysis according to the analysis conditions set by the analysis condition setting function 355. The machine learning parameters used in the third analysis may include weights of attributes of biomarkers contributing to the third analysis using machine learning.

[0034] The analysis condition setting function 355 may determine the purpose of the third analysis, i.e., the algorithm used for the third analysis, based on the result of the consistency determination by the determination function 353, the result of the first analysis by the first analysis function 351, and the result of the second analysis by the second analysis function 352. That is, the analysis condition setting function 355 may determine the purpose of the third analysis based on the result of the consistency determination by the determination function 353 and the result of the comprehensive determination of the lesion based on the result of the first analysis and the result of the second analysis. Furthermore, the analysis condition setting function 355 may determine the attribute of the biomarker to be noted based on the determined purpose of the third analysis, i.e., the algorithm. Then, the analysis condition setting function 355 may set the weight of the attribute of the biomarker contributing to the third analysis by weighting the determined attribute according to the contribution of the determined attribute to the second analysis.

[0035] The biomarker attribute of interest may be the use of the biomarker, for example, (1) a general biomarker for cancer diagnosis and treatment evaluation, (2) subtype determination, (3) differentiation from normal cell types such as luminal, basal, and myoepithelial, (4) differentiation of proliferation or disease progression, (5) a biomarker associated with cancer immunohistochemistry prognosis, etc.

[0036] Furthermore, when applied to breast cancer, noteworthy biomarker attributes (i.e., uses) may be (1) general tumor markers (i.e., tumor markers common to cancers in general), (2) determination of triple-negative breast cancer, (3) determination of subtypes other than triple-negative breast cancer, (4) detection of metastasis, (5) early detection, (6) determination of fibroadenomas such as cysts, (7) determination of cancer recurrence, etc.

[0037] The analysis condition setting function 355 may determine the attributes of the biomarkers to be noted based on the correspondence between the types of biomarkers (biomarkers used in the second analysis) and the attributes of the biomarkers stored in advance in the memory circuit 34. Furthermore, the analysis condition setting function 355 may acquire the correspondence between the types of biomarkers used in the second analysis and their attributes based on information held in an external database, information on academic papers and conference reports on an external network, and research information stored in the operator's own computer, and may determine the attributes of the biomarkers to be noted based on the acquired correspondence.

[0038] The analysis condition setting function 355 may weight the attributes of the determined notable biomarker more heavily when the contribution of the determined notable biomarker to the second analysis is low than when the contribution is high.

[0039] The analysis condition setting function 355 may obtain the contribution of the attribute of the noteworthy biomarker to the second analysis from the result of the machine learning used in the second analysis. Furthermore, the analysis condition setting function 355 may calculate a score according to the obtained contribution. The score when the contribution is high may be a larger value than the score when the contribution is low. Then, the analysis condition setting function 355 may weight the determined attribute of the noteworthy biomarker based on the calculated score.

[0040] The analysis condition setting function 355 may set analysis conditions (i.e., parameters, modes, networks, etc.) to reduce false negative determinations when the consistency as to whether the tumor is benign or malignant is greater than a first threshold and the results of the first and second analyses are benign.

[0041] In addition, the analysis condition setting function 355 may set analysis conditions for reducing false positive determinations or for detecting metastasis and recurrence of a lesion when the consistency is greater than a first threshold and the results of the first and second analyses are malignant.

[0042] In addition, the analysis condition setting function 355 may set the analysis conditions for reducing false negative determinations or for detecting metastasis and recurrence of a lesion when the consistency is equal to or less than a first threshold and equal to or greater than a second threshold that is smaller than the first threshold.

[0043] Furthermore, the analysis condition setting function 355 may set analysis conditions for early detection of a lesion when the degree of consistency is less than the second threshold and the result of the first analysis is benign or absent.

[0044] In addition, when the consistency is less than a second threshold and the result of the first analysis is benign, the analysis condition setting function 355 may set analysis conditions for reducing false positive determinations based on images or for detecting metastasis and recurrence of a lesion.

[0045] In addition, the analysis condition setting function 355 may set analysis conditions for reducing false negative judgments based on images or for detecting metastasis and recurrence of a lesion when the consistency is less than a second threshold and the result of the first analysis is malignant.

[0046] Next, an operation example of the medical information processing device 3 according to the first embodiment configured as above will be described. Fig. 3 is a flowchart showing an operation example of the medical information processing device 3 according to the first embodiment. Note that the following operation example is premised on the specimen test being liquid biopsy (LB).

[0047] First, as shown in FIG. 3, the first analysis function 351 executes a first analysis based on a medical image (step S1).

[0048] Furthermore, the second analysis function 352 executes a second analysis based on the sample test results relating to the subject (step S2).

[0049] Next, the determination function 353 executes a determination of the degree of consistency between the result of the first analysis by the first analysis function 351 and the result of the second analysis by the second analysis function 352 (step S3). In addition, the determination function 353 performs a comprehensive determination based on the result of the first analysis by the first analysis function 351 and the result of the second analysis by the second analysis function 352.

[0050] FIG. 4 is a diagram showing an example of calculation of the consistency in an example of the operation of the medical information processing device 3 according to the embodiment. For example, when the subject is "specimen 1" in FIG. 4, the determination function 353 determines that the consistency is 76% because the result of the second analysis (LB result) is benign (likelihood of benign: 83%, likelihood of malignancy: 14%) and the result of the first analysis (image findings) is benign (likelihood of benign: 76%, likelihood of malignancy: 24%). This consistency corresponds to the likelihood of benignity in the result of the first analysis which has a smaller likelihood of benignity. In this case, the determination function 353 determines that the lesion of "specimen 1" is benign as a comprehensive determination based on the result of the first analysis and the result of the second analysis, as shown in FIG. 4.

[0051] Furthermore, when the specimen is "specimen 2" in FIG. 4, the judgment function 353 judges that the consistency is 68% because the result of the second analysis is malignant (likelihood of benign: 32%, likelihood of malignancy: 68%) and the result of the first analysis is malignant (likelihood of benign: 47%, likelihood of malignancy: 73%). This consistency corresponds to the likelihood of malignancy in the result of the second analysis which has a smaller likelihood of malignancy. In this case, the judgment function 353 judges that the lesion in "specimen 2" is malignant as a comprehensive judgment based on the results of the first analysis and the results of the second analysis, as shown in FIG. 4.

[0052] In addition, when the subject is "specimen 3" in FIG. 4, the determination function 353 determines that the consistency is 47% because the result of the second analysis is benign (likelihood of benign: 51%, likelihood of malignancy: 48%) and the result of the first analysis is malignant (likelihood of benign: 47%, likelihood of malignancy: 53%). This consistency corresponds to the likelihood of benignity in the result of the first analysis, which has a smaller likelihood of benignity. In the example shown in FIG. 4, when the result of the second analysis is different from the result of the first analysis in terms of benignity and malignancy, such as "specimen 3", the determination function 353 determines the smallest likelihood of the four likelihoods, the likelihood of benignity and malignancy of the result of the second analysis and the likelihood of benignity and malignancy of the result of the first analysis, as the consistency. This is also true for "specimen 4" to "specimen 6" described later. In this case, as shown in FIG. 4, the result of the overall determination by the determination function 353 for "specimen 3" is that re-determination is required.

[0053] Furthermore, when the specimen is "Sample 4" in FIG. 4, the judgment function 353 judges that the consistency is 34% because the result of the second analysis is malignant (likelihood of benign: 34%, likelihood of malignancy: 66%) and the result of the first analysis is benign (likelihood of benign: 52%, likelihood of malignancy: 48%). This consistency corresponds to the likelihood of benignity in the result of the second analysis which has a smaller likelihood of benignity. In this case, as shown in FIG. 4, the result of the overall judgment by the judgment function 353 for "Sample 4" is that re-judgment is required.

[0054] Furthermore, when the specimen is "specimen 5" in FIG. 4, the judgment function 353 judges that the consistency is 38% because the result of the second analysis is malignant (likelihood of benign: 38%, likelihood of malignancy: 64%) and the result of the first analysis is benign (likelihood of benign: 69%, likelihood of malignancy: 39%). This consistency corresponds to the likelihood of benignity in the result of the second analysis which has a smaller likelihood of benignity. In this case, as shown in FIG. 4, the result of the overall judgment by the judgment function 353 for "specimen 5" is that re-judgment is required.

[0055] Furthermore, when the specimen is "specimen 6" in FIG. 4, the judgment function 353 judges that the consistency is 32% because the result of the second analysis is benign (likelihood of benign: 51%, likelihood of malignancy: 48%) and the result of the first analysis is malignant (likelihood of benign: 32%, likelihood of malignancy: 68%). This consistency corresponds to the likelihood of benignity in the result of the first analysis which has a smaller likelihood of benignity. In this case, as shown in FIG. 4, the result of the overall judgment by the judgment function 353 for "specimen 6" is that re-judgment is required.

[0056] After the consistency determination and the overall determination are performed, the analysis condition setting function 355 sets analysis conditions for the third analysis (step S4), as shown in Fig. 3. Fig. 5 is a flowchart showing a process of setting analysis conditions for the third analysis in an operation example of the medical information processing device 3 according to the embodiment.

[0057] Specifically, as shown in FIG. 5, first, the analysis condition setting function 355 determines the purpose of the third analysis, i.e., the algorithm to be used for the third analysis, based on the consistency judgment result and the overall judgment result (step S41).

[0058] Fig. 6 is a diagram showing an example of setting analysis conditions for the third analysis in an example of the operation of the medical information processing device according to the embodiment. In the example shown in Fig. 4 and Fig. 6, when the subject is "Sample 1", the analysis condition setting function 355 determines that the purpose, i.e., the algorithm, of the third analysis is to reduce false negatives ("ML / DL-Type F") based on the degree of consistency being 76%, which is greater than 60% (i.e., the first threshold value), and the result of the overall judgment being benign.

[0059] In addition, when the specimen is "Sample 2", the analysis condition setting function 355 determines that the purpose of the third analysis, i.e., the algorithm, is to reduce false positives or detect metastasis and recurrence ("ML / DL-Type E") based on the consistency being 68%, which is greater than 60%, and the overall judgment result being malignant.

[0060] Furthermore, when the subject is "Sample 3," the analysis condition setting function 355 determines that the purpose of the third analysis, i.e., the algorithm, is to reduce false negatives on the image side or to detect metastasis and recurrence ("ML / DL-Type D") based on the fact that the consistency is 47%, which is greater than or equal to 40% (i.e., the second threshold) and less than or equal to 60%, and the result of the overall judgment is that re-evaluation is required.

[0061] In addition, when the specimen is "Sample 4", the analysis condition setting function 355 determines that the purpose of the third analysis, i.e., the algorithm, is to achieve early detection and reduce false positives on the image side ("ML / DL-Type C") based on the fact that the consistency is 34%, which is less than 40%, and the result of the overall judgment is that re-evaluation is required.

[0062] In addition, when the specimen is "Sample 5", the analysis condition setting function 355 determines that the purpose of the third analysis, i.e., the algorithm, is to reduce false positives on the image side or to detect metastasis and recurrence ("ML / DL-Type B") based on the fact that the consistency is 38%, which is less than 40%, and the result of the overall judgment is that re-evaluation is required.

[0063] In addition, when the specimen is "Sample 6", the analysis condition setting function 355 determines that the purpose of the third analysis, i.e., the algorithm, is to reduce false negatives on the image side or to detect metastasis and recurrence ("ML / DL-Type A") based on the fact that the consistency is 32%, which is less than 40%, and the result of the overall judgment is that re-evaluation is required.

[0064] After determining the purpose of the third analysis, as shown in FIG. 5, the analysis condition setting function 355 determines the attributes of the biomarkers to be noted based on the determined purpose of the third analysis (step S42). In the example shown in FIG. 4 and FIG. 6, when the specimen is "specimen 1", the analysis condition setting function 355 determines that the attributes of the biomarkers to be noted are subtype determination, metastasis determination, and distinction from normal cell type. When the specimen is "specimen 2", the analysis condition setting function 355 determines that the attributes of the biomarkers to be noted are subtype determination, progression discrimination, and prognosis estimation. When the specimen is "specimen 3", the analysis condition setting function 355 determines that the attributes of the biomarkers to be noted are subtype determination and distinction from normal cell type. When the specimen is "specimen 4", the analysis condition setting function 355 determines that the attributes of the biomarkers to be noted are general tumor marker, distinction from normal cell type, and subtype determination. Furthermore, when the specimen is "specimen 5", the analysis condition setting function 355 determines that the attributes of the biomarkers to be noted are a general tumor marker, distinction from normal cell types, and subtype determination. Furthermore, when the specimen is "specimen 6", the analysis condition setting function 355 determines that the attributes of the biomarkers to be noted are a general tumor marker, distinction from normal cell types, subtype determination, and progression determination.

[0065] After determining the attributes of the biomarkers to be noted, as shown in FIG. 5, the analysis condition setting function 355 calculates a score according to the degree of contribution of the determined attributes of the biomarkers to be noted to the second analysis (step S43).

[0066] FIG. 7 is a diagram showing the contribution of the attributes of the biomarkers to the second analysis in an example of the operation of the medical information processing device 3 according to the embodiment. The vertical axis in FIG. 7 indicates the contribution of each attribute of the biomarkers A to T to the second analysis. The contribution of each attribute of the biomarkers A to T varies as shown in FIG. 7. The analysis condition setting function 355 calculates a score according to the total value of the contributions in the top 60% tile for each attribute (each of the biomarkers A to T), for example. For example, the analysis condition setting function 355 may calculate a score of "5" when the contribution (i.e., the total value) is 80-100%. Also, the analysis condition setting function 355 may calculate a score of "4" when the contribution is 79-60%. Also, the analysis condition setting function 355 may calculate a score of "3" when the contribution is 59-40%. Furthermore, the analysis condition setting function 355 may calculate a score of "2" when the contribution rate is 39-20%. Furthermore, the analysis condition setting function 355 may calculate a score of "1" when the contribution rate is 19-0%.

[0067] In the examples shown in FIG. 4 and FIG. 6, when the subject is “Sample 1”, the analysis condition setting function 355 calculates a score “4” based on the contribution of the determined attribute of the notable biomarker to the second analysis being 77%. When the subject is “Sample 2”, the analysis condition setting function 355 calculates a score “4” based on the contribution of the determined attribute of the notable biomarker to the second analysis being 68%. When the subject is “Sample 3”, the analysis condition setting function 355 calculates a score “3” based on the contribution of the determined attribute of the notable biomarker to the second analysis being 54%. When the subject is “Sample 4”, the analysis condition setting function 355 calculates a score “2” based on the contribution of the determined attribute of the notable biomarker to the second analysis being 39%. Furthermore, when the specimen is "Sample 5", the analysis condition setting function 355 calculates a score of "5" based on the contribution of the determined attribute of the notable biomarker to the second analysis being 81%. Furthermore, when the specimen is "Sample 6", the analysis condition setting function 355 calculates a score of "3" based on the contribution of the determined attribute of the notable biomarker to the second analysis being 43%.

[0068] After calculating the scores, as shown in FIG. 5, the analysis condition setting function 355 performs weighting based on the calculated scores to set weights that the determined attributes of the noteworthy biomarkers will contribute to the third analysis (step S44).

[0069] 6, when the specimen is "Sample 1," the analysis condition setting function 355 sets the weighting strength for the attributes "subtype determination, metastasis determination, and distinction from normal cell type" to "medium" for the third analysis according to the algorithm "ML / DL-Type F" for reducing false negatives. Such weighting is based on the fact that the score of "Sample 1" is the second highest value of "4."

[0070] Furthermore, when the subject is "specimen 2", the analysis condition setting function 355 sets the weighting strength for the attribute "subtype determination, progression discrimination, prognosis estimation" to "medium" for the third analysis according to the algorithm "ML / DL-Type E" for reducing false positives or detecting metastasis and recurrence. Such weighting is based on the fact that the score of "specimen 2" is the second highest value of "4". Note that, for "specimen 2", the analysis condition setting function 355 may weight "specimen 2" by increasing the weight of biomarkers for determining the subtype of breast cancer (subtyping) and the weight of tumor biomarkers of sites other than the breast (e.g., prostate, etc.).

[0071] Furthermore, when the subject is "specimen 3", the analysis condition setting function 355 sets the weighting strength for the attribute "determination of subtype, distinction from normal cell type" to "strong" for the third analysis according to the algorithm "ML / DL-Type D" for reducing false negatives on the image side or for detecting metastasis and recurrence. Such weighting is based on the score of "specimen 3" being a medium value "3".

[0072] Furthermore, when the subject is "specimen 4," the analysis condition setting function 355 sets the strength of weighting for the attributes "general tumor marker, distinction from normal cell type, subtype determination" to "strong" for the third analysis according to the algorithm "ML / DL-Type C" for early detection and reducing false positives on the image side. Such weighting is based on the fact that the score of "specimen 4" is "2," which is the second lowest.

[0073] For example, triple-negative breast cancer generally has a clear border and appears benign in images. For this reason, the analysis condition setting function 355 can increase the weight (contribution) of biomarkers that are common in triple-negative breast cancer for "sample 4" whose analysis result of the first analysis is benign, thereby reducing the oversight of triple-negative breast cancer and leading to early detection. Biomarkers that are common in triple-negative breast cancer include, for example, epidermal growth factor receptor (EGFR), estrogen receptor (ER), progesterone receptor (PR), and HER2 / neu.

[0074] In addition, even if the consistency is less than 40% and there is no analysis result of the first analysis (i.e., the cancer has not grown enough to be determined from the image), the analysis condition setting function 355 may perform weighting for the third analysis according to the algorithm "ML / DL-Type C" for early detection. In this case, the analysis condition setting function 355 may perform weighting to increase the weight of a biomarker having a characteristic of early cancer detection. Examples of biomarkers having a characteristic of early cancer detection include DNA methylation analysis and Tumor Associated Auto-antibody analysis.

[0075] Furthermore, when the subject is "specimen 5", the analysis condition setting function 355 sets the weighting strength for the attribute "biomarker attribute is general tumor marker, distinction from normal cell type, subtype determination" to "medium" for the third analysis according to the algorithm "ML / DL-Type B" for reducing false positives on the image side or detecting metastasis and recurrence. Such weighting is based on the fact that the score of "specimen 5" is the highest value "5". For example, the analysis condition setting function 355 may weight "specimen 5" by increasing the weights of fibroblast and tumor biomarkers.

[0076] Furthermore, when the subject is "specimen 6", the analysis condition setting function 355 sets the weighting strength for the attributes "general tumor marker, distinction from normal cell type, subtype determination, and progression determination" to "strong" for the third analysis according to the algorithm "ML / DL-Type A" for reducing false negatives on the image side or detecting metastasis and recurrence. Such weighting is based on the score of "specimen 6" being a medium value of "3". For example, the analysis condition setting function 355 may weight "specimen 6" by increasing the weight of a biomarker for breast cancer subtyping or a tumor biomarker for a site other than the breast (e.g., prostate, etc.).

[0077] When increasing the weight of a biomarker for breast cancer subtyping, the analysis condition setting function 355 may increase the weight of a biomarker associated with a subtype corresponding to the image feature acquired by the first analysis. For example, when the image feature is that the image does not include calcification and has a clear border, the analysis condition setting function 355 may increase the weight of epidermal growth factor receptor (positive), estrogen receptor (negative), progesterone receptor (negative), and HER2 / neu (negative) associated with triple-negative breast cancer. When the image feature is that the image includes calcification, the analysis condition setting function 355 may increase the weight of HER2 (positive) associated with HER2-positive breast cancer. When the image feature is that the border is unclear or includes spicules, the analysis condition setting function 355 may increase the weight of estrogen receptor (positive), progesterone receptor (positive), and HER2 / neu (positive / negative) associated with luminal breast cancer.

[0078] After the analysis conditions for the third analysis are set as described above, as shown in FIG. 3, the third analysis function 354 executes the third analysis in accordance with the set analysis conditions (step S5).

[0079] For example, when the specimen is “Sample 1” in FIG. 4 and FIG. 6, i.e., when the results of the first and second analyses are benign (step S51: benign), the third analysis function 354 performs the third analysis according to analysis conditions (ML / DL-Type F) for reducing false negatives (step S53).

[0080] Furthermore, when the specimen is "Sample 2" in Figures 4 and 6, i.e., when the results of the first and second analyses are malignant (step S51: malignant), the third analysis function 354 performs a third analysis according to analysis conditions (ML / DL-Type E) for reducing false positives or detecting metastasis and recurrence (step S54).

[0081] Also, when the subject is “Sample 3” in FIG. 4 and FIG. 6, the third analysis function 354 executes the third analysis according to the analysis conditions (ML / DL-Type D) for reducing false negatives on the image side or for detecting metastasis and recurrence (step S55).

[0082] Furthermore, if the subject is "Specimen 4" in Figures 4 and 6, i.e., if the result of the first analysis is benign or absent (step S52: benign / absent), the third analysis function 354 executes the third analysis according to analysis conditions (ML / DL-Type C) for early detection and reducing false positives on the image side (step S56).

[0083] Furthermore, when the subject is “Specimen 5” in FIG. 4 and FIG. 6, that is, when the result of the first analysis is benign (step S52: benign), the third analysis function 354 performs the third analysis according to analysis conditions (ML / DL-Type B) for reducing false positives on the image side or for detecting metastasis and recurrence (step S57).

[0084] Furthermore, when the subject is "Specimen 6" in Figures 4 and 6, i.e., when the result of the first analysis is malignant (step S52: malignant), the third analysis function 354 performs a third analysis according to analysis conditions (ML / DL-Type A) for reducing false negatives on the image side or for detecting metastasis and recurrence (step S58).

[0085] As described above, in the embodiment, the first analysis function 351 executes a first analysis based on a medical image related to a subject. The second analysis function 352 executes a second analysis based on a specimen test result related to the subject. The determination function 353 executes a determination of the degree of consistency between the result of the first analysis by the first analysis function 351 and the result of the second analysis by the second analysis function 352. The third analysis function 354 executes a third analysis based on the medical image and the specimen test result under analysis conditions based on the determination result of the degree of consistency by the determination function 353, and determines diagnostic support information based on the medical image and the specimen test result.

[0086] As a result, the third analysis is performed under analysis conditions based on the result of the determination of the degree of consistency between the result of the first analysis based on the medical image of the subject and the result of the second analysis based on the specimen test result of the subject, so that appropriate diagnostic support information can be determined. Therefore, according to the first embodiment, the accuracy of lesion determination can be improved.

[0087] In the embodiment, the sample test result includes values ​​of biomarkers contained in the sample collected from the subject. The second analysis function 352 executes the second analysis based on the values ​​of the biomarkers. The third analysis function 354 executes the third analysis using machine learning. The analysis condition setting function 355 sets analysis conditions by setting at least one of machine learning parameters, mode, and network. The third analysis function 354 executes the third analysis according to the analysis conditions set by the analysis condition setting function 355.

[0088] This allows the third analysis to be performed with high accuracy, thereby appropriately improving the accuracy of lesion determination.

[0089] Also, in an embodiment, the parameters are weights of attributes of biomarkers that contribute to a third analysis using machine learning.

[0090] This allows the third analysis to be performed appropriately by setting the weights of the attributes of the biomarkers in consideration of the result of the consistency determination.

[0091] Furthermore, in the embodiment, the analysis condition setting function 355 determines the purpose of the third analysis based on the result of the consistency determination by the determination function 353, the result of the first analysis by the first analysis function 351, and the result of the second analysis by the second analysis function 352. Furthermore, the analysis condition setting function 355 determines the attribute of the biomarker to be noted based on the determined purpose of the third analysis. Furthermore, the analysis condition setting function 355 sets the weight of the attribute of the biomarker contributing to the third analysis by weighting the determined attribute according to the contribution of the determined attribute to the second analysis.

[0092] As a result, even if the attribute of a biomarker to be noted is in the minority (i.e., the attribute has a low contribution to the second analysis), it is possible to set analysis conditions for the third analysis taking into consideration the minority attribute. Therefore, according to the embodiment, even if a minority attribute is an attribute to be noted, the attribute can be appropriately reflected in the results of the third analysis.

[0093] In addition, in the embodiment, when the contribution of the determined attribute to the second analysis is low, the analysis condition setting function 355 weights the determined attribute more heavily than when the contribution is high.

[0094] This makes it possible to more appropriately set analysis conditions for the third analysis that takes into account minority attributes.

[0095] In the embodiment, the second analysis function 352 executes the second analysis using machine learning. The analysis condition setting function 355 obtains a contribution degree from the result of the machine learning used in the second analysis, calculates a score according to the obtained contribution degree, and weights the determined attribute based on the calculated score.

[0096] This allows appropriate weighting to be done based on the scores.

[0097] Moreover, in the embodiment, the first analysis function 351 performs the first analysis further based on past medical information of the subject.

[0098] This can further improve the accuracy of lesion determination.

[0099] Also, in an embodiment, the second analysis function 352 performs a second analysis further based on past medical information of the subject.

[0100] This can further improve the accuracy of lesion determination.

[0101] In the embodiment, the determination function 353 determines the degree of consistency by taking into account the likelihood of the result of the first analysis and the likelihood of the result of the second analysis.

[0102] This allows minority attributes to be reflected in the consistency level, further improving the accuracy of lesion determination.

[0103] Furthermore, in the embodiment, when the consistency is greater than the first threshold and the results of the first and second analyses are benign, the analysis condition setting function 355 sets analysis conditions for reducing false negative determinations.

[0104] This makes it possible to appropriately perform a determination based on the degree of consistency and the result of the second analysis.

[0105] In addition, in an embodiment, when the consistency is greater than a first threshold and the results of the first and second analyses are malignant, the analysis condition setting function 355 sets analysis conditions for reducing false positive determinations or for detecting metastasis and recurrence of the lesion.

[0106] This makes it possible to appropriately perform a determination based on the degree of consistency and the result of the second analysis.

[0107] In addition, in an embodiment, when the consistency is equal to or less than a first threshold and equal to or greater than a second threshold that is smaller than the first threshold, the analysis condition setting function 355 sets analysis conditions for reducing false negative determinations or for detecting metastasis and recurrence of a lesion.

[0108] This allows appropriate determination to be made according to the degree of consistency.

[0109] Furthermore, in the embodiment, when the second threshold is less than the second threshold and the result of the first analysis is benign or absent, the analysis condition setting function 355 sets analysis conditions for early detection of a lesion.

[0110] This makes it possible to appropriately perform a determination according to the degree of consistency and the result of the first analysis.

[0111] In addition, in an embodiment, when the result of the first analysis is benign because the second threshold is less than the second threshold, the analysis condition setting function 355 sets analysis conditions for reducing false positive determinations based on images or for detecting metastasis and recurrence of the lesion.

[0112] This makes it possible to appropriately perform a determination based on the degree of consistency and the result of the first analysis.

[0113] In addition, in an embodiment, when the consistency is less than the second threshold and the result of the first analysis is malignant, the analysis condition setting function 355 sets analysis conditions for reducing false negative judgments based on images or for detecting metastasis and recurrence of the lesion.

[0114] This makes it possible to appropriately perform a determination according to the degree of consistency and the result of the first analysis.

[0115] In the above embodiment, the input data is the medical image and the result of the specimen test, but the embodiment is not limited to this, and other input data can be used in combination. For example, the input data can be electronic medical record data. In this case, in step S3, the consistency between the medical image, the result of the specimen test, and the data of the electronic medical record can be determined. This makes it possible to provide diagnostic support information by adding information that cannot be considered only from the medical image and the result of the specimen test.

[0116] The term "processor" used in the above description means a circuit such as a central processing unit (CPU), a graphics processing unit (GPU), an application specific integrated circuit (ASIC), a programmable logic device (e.g., a simple programmable logic device (SPLD), a complex programmable logic device (CPLD), and a field programmable gate array (FPGA)). The processor realizes its function by reading and executing a program stored in a memory circuit. Instead of storing a program in a memory circuit, the processor may be configured to directly incorporate the program in the circuit. In this case, the processor realizes its function by reading and executing the program incorporated in the circuit. The processor is not limited to being configured as a single processor circuit, and may be configured as a single processor by combining multiple independent circuits to realize its function. Furthermore, the multiple components in FIG. 2 may be integrated into a single processor to realize its function.

[0117] According to at least one of the embodiments described above, it is possible to improve the accuracy of lesion determination.

[0118] Although several embodiments have been described above, these embodiments are presented only as examples and are not intended to limit the scope of the invention. The novel device and method described in this specification can be implemented in various other forms. In addition, various omissions, substitutions, and modifications can be made to the forms of the device and method described in this specification without departing from the spirit of the invention. The appended claims and their equivalents are intended to include such forms and modifications that fall within the scope and spirit of the invention. [Explanation of symbols]

[0119] 3 Medical information processing equipment 351 1st analysis function 352 Second analysis function 353 Judgment Function 354 Third analysis function 355 Analysis condition setting function

Claims

1. a first analysis unit that executes a first analysis based on a medical image related to a subject; A second analysis unit that performs a second analysis based on a sample test result related to the sample; a determination unit that executes a determination of a degree of consistency between a result of the first analysis by the first analysis unit and a result of the second analysis by the second analysis unit; a third analysis unit that executes a third analysis based on the medical image and the specimen test result under analysis conditions based on the result of the determination of the degree of consistency by the determination unit, and determines diagnostic support information based on the medical image and the specimen test result; A medical information processing device comprising:

2. the specimen test result includes values ​​of biomarkers contained in a specimen collected from the subject; The second analysis unit performs the second analysis based on the value of the biomarker; The third analysis unit performs the third analysis using machine learning; The medical image processing device further includes an analysis condition setting unit that sets the analysis conditions by setting at least one of a parameter, a mode, and a network of the machine learning, The medical image processing apparatus according to claim 1 , wherein the third analysis unit executes the third analysis in accordance with the analysis conditions set by the analysis condition setting unit.

3. The medical image processing device according to claim 2 , wherein the parameter is a weight of an attribute of the biomarker that contributes to the third analysis using the machine learning.

4. 4. The medical information processing device according to claim 3, wherein the analysis condition setting unit determines a purpose of the third analysis based on a result of the consistency determination by the determination unit, a result of the first analysis by the first analysis unit, and a result of the second analysis by the second analysis unit, determines attributes of biomarkers to be noted based on the determined purpose of the third analysis, and sets weights of attributes of the biomarkers contributing to the third analysis by weighting the determined attributes in accordance with a contribution degree of the determined attributes to the second analysis.

5. The medical image processing device according to claim 4 , wherein the analysis condition setting unit weights the determined attribute more heavily when the degree of contribution of the determined attribute to the second analysis is low than when the degree of contribution is high.

6. The second analysis unit performs the second analysis using machine learning; 5. The medical information processing device according to claim 4, wherein the analysis condition setting unit obtains the degree of contribution from a result of machine learning used in the second analysis, calculates a score according to the obtained degree of contribution, and weights the determined attribute based on the calculated score.

7. The medical image processing apparatus according to claim 1 , wherein the first analysis unit executes the first analysis further based on past medical information of the subject.

8. The medical image processing apparatus according to claim 1 , wherein the second analysis unit executes the second analysis further based on past medical information of the subject.

9. The medical image processing apparatus according to claim 1 , wherein the determining unit determines the degree of consistency by taking into consideration a likelihood of the result of the first analysis and a likelihood of the result of the second analysis.

10. The medical information processing device according to claim 2 , wherein the analysis condition setting unit sets the analysis conditions for reducing false negative determinations when the consistency is greater than a first threshold and the results of the first and second analyses are benign.

11. 11. The medical information processing device according to claim 10, wherein the analysis condition setting unit sets the analysis conditions for reducing false positive determinations or detecting metastasis and recurrence of a lesion when the consistency is greater than the first threshold and the results of the first and second analyses are malignant.

12. 11. The medical information processing device according to claim 10, wherein the analysis condition setting unit sets the analysis conditions for reducing false negative determinations or detecting metastasis and recurrence of a lesion when the consistency is equal to or less than the first threshold and equal to or greater than a second threshold that is smaller than the first threshold.

13. The medical information processing device according to claim 10 , wherein the analysis condition setting unit sets the analysis conditions for early detection of a lesion when the degree of consistency is less than a second threshold value that is smaller than the first threshold value and a result of the first analysis is benign or absent.

14. 11. The medical information processing device according to claim 10, wherein the analysis condition setting unit sets the analysis conditions for reducing false positive determinations based on images or detecting metastasis and recurrence of a lesion when the consistency is less than a second threshold value that is smaller than the first threshold value and a result of the first analysis is benign.

15. 11. The medical information processing device according to claim 10, wherein the analysis condition setting unit sets the analysis conditions for reducing false negative determinations based on images or for detecting metastasis and recurrence of a lesion when the consistency is less than a second threshold value that is smaller than the first threshold value and a result of the first analysis is malignant.

16. The medical image processing apparatus according to claim 4 , wherein the attribute of the biomarker is an application of the biomarker.

17. The sample test result includes values ​​of a plurality of biomarkers, The medical image processing apparatus according to claim 2 , wherein the second analysis unit executes the second analysis based on values ​​of the plurality of types of biomarkers.

18. The medical information processing device according to claim 16, wherein the attributes of the biomarkers include at least one of a general tumor marker, a determination of triple-negative breast cancer, a determination of a subtype other than triple-negative breast cancer, detection of metastasis, early detection, a determination of fibroadenoma such as a cyst, and a determination of cancer recurrence.

19. performing a first analysis based on a medical image of a subject; performing a second analysis based on the specimen test results relating to the specimen; determining a degree of consistency between a result of the first analysis and a result of the second analysis; A medical information processing method, comprising: performing a third analysis based on the medical image and the specimen test results using analysis conditions based on the result of the consistency determination; and determining diagnostic support information based on the medical image and the specimen test results.

20. On the computer, performing a first analysis based on a medical image of a subject; performing a second analysis based on the sample test results relating to the sample; A step of performing a determination of the degree of consistency between the result of the first analysis and the result of the second analysis; and a step of performing a third analysis based on the medical image and the specimen test result under analysis conditions based on the result of the determination of the degree of consistency, and determining diagnostic support information based on the medical image and the specimen test result; A program for executing the above.

Images