JP7480601B2 - Medical diagnosis support device, method for controlling medical diagnosis support device, and program - Google Patents

Description

The present invention relates to a medical diagnostic support device, a control method for a medical diagnostic support device, and a program.

As a device for determining the urgency of injured or sick people, a triage supplementary device has been proposed that uses AI (Artificial Intelligence) to determine the urgency of injured or sick people from the characteristics of images captured by a drone (see, for example, Patent Document 1). In addition, a device has been disclosed that relates to a rescue plan presentation system that calculates the degree of danger, which indicates the physical danger of each rescue target among multiple rescue targets who require rescue, and determines the rescue priority (see, for example, Patent Document 2).

JP 2019-192029 A JP 2018-129075 A

However, in Patent Document 1, the degree of illness is determined from captured images (photographs), making it difficult to determine the urgency of the sick person. In addition, in Patent Document 2, rescue priorities are determined according to the degree of danger to the person to be rescued, making it impossible to recognize the urgency of whether the person to be rescued needed to be rescued early.

The present invention aims to provide a medical diagnostic support device, a control method for the medical diagnostic support device, and a program that use AI to calculate the probability of each disease from multiple vital signs, thereby improving the diagnostic flow and providing diagnostic support by assigning priorities to patients and diseases.

That is, the above-mentioned object of the present invention is achieved by the following configuration.
(1) an acquisition unit for acquiring a plurality of pieces of biological information of each patient;
an estimation unit that estimates a disease and a possibility of the disease from the plurality of biological information of each patient acquired by the acquisition unit using a trained identification device;
a control unit that outputs to an output unit the disease and the possibility of the disease of each patient estimated by the estimation unit;
A calculation unit that calculates a risk level of each of the diseases based on the possibility of each of the diseases estimated by the estimation unit,
The calculation unit is
calculating a list of modalities recommended for use with each of the patients from the calculated risk level of each of the diseases of each of the patients;
Medical diagnostic support equipment.

( 2 ) The calculation unit is
Calculating an overall risk level for each of the patients from the risk level of the disease for each of the patients;
The medical diagnosis support device according to ( 1 ).

( 3 ) The calculation unit
performing triage for the plurality of patients based on the overall risk levels of the plurality of patients;
A medical diagnosis support device according to ( 2 ).

( 4 ) acquiring a plurality of pieces of biometric information for each patient;
A step of estimating a disease and a possibility of the disease from a plurality of biometric information of each patient using a trained identification device;
outputting the disease and the possibility of disease of each of the patients to an output unit;
A calculation step of calculating a risk level of each of the diseases based on the estimated possibility of each of the diseases;
Further, in the calculation step,
calculating a list of modalities recommended for use with each of the patients from the calculated risk level of each of the diseases of each of the patients;
A method for controlling a medical diagnosis support device.

( 5 ) To the computer,
acquiring a plurality of biometric data for each patient;
estimating a disease and its possibility from a plurality of biometric information of each patient using a trained identification device;
outputting the disease and the possibility of disease of each of the patients to an output unit;
A calculation step of calculating a risk level of each of the diseases based on the estimated possibility of suffering from each of the diseases,
Further, in the calculation step,
calculating a list of modalities recommended for use with each of the patients from the calculated risk level of each of the diseases of each of the patients;
A program characterized by causing a user to execute the following:

According to the present invention, by understanding the disease of a subject and the possibility of contracting the disease, it is possible to prevent diseases from being overlooked at an early stage, and to extract subjects who should be given priority for diagnosis in situations where there are too many subjects to be tested at emergency sites, etc.

This is expected to improve the prognosis for each individual patient, reduce medical costs, and improve patient quality of life (QOL).

1 is a diagram illustrating an example of the configuration of a medical diagnosis support apparatus according to a first embodiment. 2 is a functional block diagram showing a configuration of a CPU of the medical diagnosis support device according to the first embodiment. FIG. 4 is a flowchart showing an operation of the medical diagnosis support device according to the first embodiment. This is an explanatory diagram showing the calculation of the risk level for each disease based on a patient's disease and the possibility of the disease. FIG. 1 is an explanatory diagram showing the results of triage performed on multiple patients based on their overall risk levels. 13 is a flowchart showing a process in which the medical diagnosis support apparatus according to the second embodiment calculates a list of modalities recommended for use with each patient. 13 is a flowchart showing an operation of the medical diagnosis support device according to the third embodiment. 11 is a flowchart showing the processing of a learning method for a trained model of a medical diagnosis support device. FIG. 11 is an explanatory diagram showing an example of biometric information prepared in advance.

The following describes in detail the form for carrying out the present invention. Note that the embodiment described below is one example for realizing the present invention, and should be appropriately modified or changed depending on the configuration of the device to which the present invention is applied and various conditions, and the present invention is not limited to the following embodiment. In addition, a configuration may be made by appropriately combining parts of each embodiment described below. Note that the same reference numerals are used for the same members, and the description will be omitted as appropriate.

First Embodiment
[Overall configuration of medical diagnosis support device]
1A is a diagram for explaining an example of the configuration of a medical diagnosis support device 100 according to the first embodiment. The medical diagnosis support device 100 according to the first embodiment is configured to include a CPU 101, an operation reception unit 102, a trained model 103, an output unit 104, a display unit 105, a ROM (Read Only Memory) 106, a RAM (Random Access Memory) 107, and an external storage device 108. The external storage device 108 is configured to include a disease estimation program 109.

As shown in FIG. 1A, the CPU 101 is an arithmetic processing device that controls the entire medical diagnosis support device 100. The CPU 101 performs various controls by executing programs stored in the ROM 106 and the external storage device 108. For example, the CPU 101 executes a disease estimation program 109 stored in the external storage device 108 to realize an acquisition unit 111, an estimation unit 112, a control unit 113, and a calculation unit 114 shown in FIG. 1B.

Figure 1B is a functional block diagram showing the configuration of the CPU 101 of the medical diagnosis support device 100 according to the first embodiment.

The acquisition unit 111 has a function of acquiring multiple pieces of biometric information (vital data) for each patient. Vital data refers to information such as age, sex, blood pressure, respiratory rate, pulse rate, electrocardiogram, body temperature, oxygen saturation, medical history, smoking history, purpose of visit, and method of visit. The acquisition unit 111, for example, acquires the measurement results automatically or manually based on input from the user.

The estimation unit 112 has a function of estimating a disease and its possibility of occurrence from multiple pieces of biometric information of each patient acquired by the acquisition unit 111, using the trained model (trained identification device) 103. For example, the estimation unit 112 estimates the disease and its possibility of occurrence for each patient as follows: cerebral aneurysm: 60%, diabetes: 55%, asthma: 40%, lung cancer: 30%, fracture: 25%, pneumothorax: 10%, stones: 3%, etc.

The control unit 113 has the function of outputting the disease and the possibility of disease of each patient estimated by the estimation unit 112 to the output unit 104.

The calculation unit 114 has a function of calculating the risk level of each disease based on the possibility of contracting each disease. The calculation unit 114 can also calculate the overall risk level according to the risk level (score) of each disease. Note that the risk level indicates the degree to which immediate treatment is required, including both the degree of danger and the degree of urgency.

In this embodiment, the CPU 101 controls the entire medical diagnosis support device 100. However, instead of controlling the entire medical diagnosis support device 100, for example, multiple hardware devices may share the processing to control the entire medical diagnosis support device 100.

The operation reception unit 102 receives input of patient information and patient biometric information. The operation reception unit 102 is composed of operation members such as a keyboard, a mouse, various switches, buttons, and a touch panel that receive various operations from the user.

The trained model 103 includes a pre-trained classifier and stores various parameters. Details of the trained model 103 will be described later.

The output unit 104 outputs the disease and the possibility of suffering from the disease to the display unit 105. The output unit 104 is configured, for example, by a medical examination terminal used by a doctor or the like.

The display unit 105 displays the disease and the possibility of suffering from the disease. The display unit 105 is composed of a display device such as a liquid crystal display, an organic electroluminescence display, or a printer, or software for displaying the disease (viewer).

The ROM 106 stores, for example, a control program for controlling the medical diagnosis support apparatus 100 .
The RAM 107 is configured, for example, by a dynamic random access memory (DRAM), and functions as a working memory for temporarily storing data and image data required for the CPU 101 to execute the control program and the disease prediction program 109 .

The external storage device 108 is, for example, a hard disk drive, and stores the disease risk of each patient calculated by the calculation unit 114.

[Operation of medical diagnosis support device]
Next, the operation of the medical diagnosis support device 100 configured as described above will be described using a flowchart with reference to FIGS. 1A and 1B.

Figure 2 is a flowchart showing the operation of the medical diagnosis support device 100 according to the first embodiment.

First, the CPU 101 of the medical diagnosis support device 100 acquires multiple pieces of biometric information for each patient in the acquisition unit 111 (step S001). The acquisition unit 111 accepts, for example, input of patient information and patient biometric information (vital data) for each patient. Note that when acquiring multiple pieces of biometric information for each patient, the CPU 101 may automatically acquire the measurement results for each patient, or may accept input from the operation acceptance unit 102.

The CPU 101 of the medical diagnosis support device 100 uses the learned model 103 in the estimation unit 112 to estimate one or more diseases and the possibility of suffering from each disease for each patient from the multiple pieces of biometric information of each patient acquired by the acquisition unit 111 (step S003).

The CPU 101 of the medical diagnosis support device 100 calculates the risk level of each disease in the calculation unit 114 based on the possibility of contracting each disease (step S005). In this case, the CPU 101 calculates the risk level for each patient in the calculation unit 114 from the priority score and the possibility of contracting each disease (also simply called the possibility).

Figure 3 is an explanatory diagram showing the calculation of the risk level for each disease from the disease and its possibility for a certain patient (e.g., Mr. C). As shown in Figure 3, the horizontal axis has columns for "priority score," "possibility," and "risk level," while the vertical axis has columns for "stroke," "cerebral aneurysm," "fracture," "lung cancer," "diabetes," etc.

The priority score is external data that is determined as a default value. For example, if there is a national standard, this value can be adopted as this priority score, or it can be edited arbitrarily in accordance with the doctor's authority and the hospital's regulations. In addition, this priority score can be made variable depending on the priority order of each hospital, the spread of infectious diseases, etc.

The probability indicates the probability of contracting each disease estimated by the estimation unit 112. The risk level is a value calculated by calculating the priority score and the probability of the disease. In this embodiment, this value is explained using the value obtained by multiplying the priority score and the probability of the disease, but this embodiment is not limited to this. In addition to the priority score and the possibility of the disease, correction can also be made using biological information such as age, sex, and blood pressure. Various methods can be used as long as the value can be calculated using the priority score and the possibility of the disease.

In the case of Mr. C shown in FIG. 3, the priority score of "100" and the possibility of "0.8" are multiplied together to calculate the risk level of "80" for "stroke." Similarly, the priority score of "90" and the possibility of "0.6" are multiplied together to calculate the risk level of "54" for "cerebral aneurysm." The priority score of "60" and the possibility of "0.2" are multiplied together to calculate the risk level of "12" for "fracture." The priority score of "90" and the possibility of "0.2" are multiplied together to calculate the risk level of "18" for "lung cancer." The priority score of "20" and the possibility of "0.4" are multiplied together to calculate the risk level of "8" for "diabetes." In this embodiment, the calculation unit 114 may not only calculate the "risk level," but also display the reason and details for the output of a disease with a high degree of severity when the disease is output.

Returning to the flowchart of FIG. 2, the CPU 101 of the medical diagnosis support device 100 calculates the overall risk level for each patient from the disease risk level of each patient in the calculation unit 114 (step S007).

In this case, the calculation unit 114 calculates the sum of the risk levels for each disease for each patient, and calculates an overall risk level for each patient. For example, in the case of Mr. C in FIG. 3, the calculation unit 114 adds up the risk level of stroke of "80", the risk level of cerebral aneurysm of "54", the risk level of fracture of "12", the risk level of lung cancer of "18", and the risk level of diabetes of "8", and calculates an overall risk level of "172".

The CPU 101 of the medical diagnosis support device 100 outputs the total risk level for each patient calculated by the calculation unit 114 to the output unit 104. The output unit 104 displays the total risk level for each patient on the display unit 105 (step S009). In this case, the calculation unit 114 of the CPU 101 performs triage for multiple patients based on the total risk levels of the multiple patients, displays the triage results on the display unit 105, and ends the processing of the medical diagnosis support device 100.

Triage refers to determining and selecting treatment priorities based on the severity of a patient's condition.

Figure 4 is an explanatory diagram showing the results of triage performed on multiple patients based on their overall risk levels. The CPU 101 can display patients in descending order of overall risk level by performing triage in the calculation unit 114.

As shown in Figure 4, for example, columns for the overall risk level and modality used are provided for each patient name, and the name column displays four patients in descending order of overall risk level. Specifically, in descending order of overall risk level, Mr. C has an overall risk level of "172," Mr. B has an overall risk level of "135," Mr. D has an overall risk level of "96," and Mr. A has an overall risk level of "47."

This allows medical staff to determine the priority of examinations and treatments, starting with Mr. C, who has the highest overall risk level. In addition, when a medical staff member selects a patient's name, the medical diagnosis support device 100 may display the selected patient's disease and the possibility of the disease, and may display an explanatory diagram (for example, see FIG. 3) that calculates the risk level for each disease.

The modality to be used by each patient is displayed in the "Modality to be used" field. The modalities to be used will be explained in the second embodiment.

As described above, the CPU 101 of the medical diagnosis support device 100 according to the first embodiment is configured to include an acquisition unit 111, an estimation unit 112, a control unit 113, and a calculation unit 114. In the estimation unit 112, the CPU 101 estimates a disease and its possibility of being affected from multiple pieces of biometric information of each patient acquired by the acquisition unit 111, using the trained model 103. In addition, in the calculation unit 114, the CPU 101 calculates the risk level of each disease based on the possibility of being affected by each disease, and calculates an overall risk level for each patient from the risk level of the disease of each patient.

As a result, the medical diagnosis support device 100 according to the first embodiment can prevent illnesses from being overlooked early by identifying the subject's disease and the possibility of contracting it, and can also extract subjects who should be given priority for diagnosis in situations where there are too many subjects to be examined at emergency sites, etc.

Furthermore, the medical diagnosis support device 100 according to the first embodiment is expected to improve the prognosis of each individual patient, thereby reducing medical costs and improving patient QOL.

Second Embodiment
The CPU 101 of the medical diagnosis support device 100 according to the second embodiment calculates a list of modalities (hereinafter, also referred to as "used modalities") recommended for use with each patient based on the risk level of each disease of each patient in the calculation unit 114. Note that the used modalities also include the meaning of the modalities that the patient plans to use in the first embodiment.

The CPU 101 of the medical diagnosis support device 100 according to the second embodiment can display a list of modalities to be used for each patient, allowing medical personnel such as doctors and nurses to determine the priorities of examinations and treatments for patients.

[Operation of medical diagnosis support device]
Next, the operation of the medical diagnosis support device 100 according to the second embodiment will be described with reference to the flowchart shown in Fig. 5. Note that, in the operation of the medical diagnosis support device 100 according to the second embodiment, the same operations as those of the medical diagnosis support device 100 according to the first embodiment shown in Fig. 2 are denoted by the same reference numerals, and the description thereof will be omitted as appropriate.

Figure 5 is a flowchart showing the process by which the medical diagnosis support device 100 according to the second embodiment calculates a list of modalities recommended for use with each patient.

First, in steps S001 and S003 shown in FIG. 5, the same processing as in steps S001 and S003 shown in FIG. 2 is performed.

Next, the CPU 101 of the medical diagnosis support device 100 calculates the risk level of each disease based on the possibility of contracting each disease in the calculation unit 114, and calculates a list of modalities to be used for each patient from the risk level of each disease for each patient (step S101).

For example, referring to FIG. 3, "MRI (Magnetic Resonance Imaging)" is determined as the modality recommended for use based on the possibility of stroke or cerebral aneurysm, and "CT (Computed Tomography)" is determined as the modality recommended for use based on the possibility of fracture. Also, "X-ray" is determined as the modality recommended for use based on the possibility of lung cancer, and "US (Ultrasonography)" is determined as the modality recommended for use based on the possibility of diabetes.

As a result, in step S009, the modality to be used for each patient can be displayed as shown in FIG. 4. Also, in FIG. 4, an indicator is displayed according to the ratio at which the modality to be used is recommended based on the likelihood of disease, and for example, it shows that MRI is likely to be selected (used) as the modality to be used.

As described above, the CPU 101 of the medical diagnosis support device 100 according to the second embodiment calculates, in the calculation unit 114, a list of modalities recommended for use for each patient based on the risk level of each disease for each patient.

As a result, the CPU 101 of the medical diagnosis support device 100 according to the second embodiment can display a list of modalities to be used for each patient, allowing medical professionals such as doctors and nurses to determine the priorities of examinations and treatments for patients.

Third Embodiment
The CPU 101 of the medical diagnosis support device 100 according to the third embodiment executes a combination of the operations of the medical diagnosis support device 100 according to the first embodiment and the operations of the medical diagnosis support device 100 according to the second embodiment.

[Operation of medical diagnosis support device]
Next, the operation of the medical diagnosis support device 100 according to the third embodiment will be described with reference to the flowchart shown in Fig. 6. Note that the operation of the medical diagnosis support device 100 according to the third embodiment executes the same processing as that of the flowchart shown in Fig. 2, with the processing of step S101 of the flowchart shown in Fig. 5 added. Note that the same processes are denoted by the same reference numerals, and the description will be omitted as appropriate.

Figure 6 is a flowchart showing the operation of the medical diagnosis support device 100 according to the third embodiment.

The medical diagnosis support device 100 according to the third embodiment first acquires multiple pieces of biometric information of each patient in the acquisition unit 111 (step S001). The acquisition unit 111 accepts input of, for example, patient information and patient biometric information (vital data) of each patient.

The CPU 101 of the medical diagnosis support device 100 uses the learned model 103 in the estimation unit 112 to estimate one or more diseases and the possibility of suffering from each disease for each patient from the multiple pieces of biometric information of each patient acquired by the acquisition unit 111 (step S003).

The CPU 101 of the medical diagnosis support device 100 calculates the risk level of each disease based on the possibility of contracting each disease in the calculation unit 114 (step S005). In the calculation unit 114, the CPU 101 calculates the risk level for each patient from the priority score and the possibility of contracting each disease.

The CPU 101 of the medical diagnosis support device 100 calculates the overall risk level for each patient from the risk level of each patient's disease in the calculation unit 114 (step S007). The calculation unit 114 calculates the sum of the risk levels for each disease for each patient, and calculates the overall risk level for each patient.

The CPU 101 of the medical diagnosis support device 100 also calculates (determines) a list of modalities to be used for each patient from the overall risk level of each patient in the calculation unit 114 (step S101). Note that if the modalities to be used can be calculated (determined) from the risk level of each disease of each patient, the calculation unit 114 may calculate (determine) the modalities to be used from the risk level of each disease of each patient.

The CPU 101 of the medical diagnosis support device 100 displays the results of the triage shown in FIG. 4 on the display unit 105 according to the overall risk level for each patient calculated by the calculation unit 114 (step S009), and displays a list of the modalities used, and ends the processing of the medical diagnosis support device 100.

As described above, the CPU 101 of the medical diagnosis support device 100 according to the third embodiment is configured to include an acquisition unit 111, an estimation unit 112, a control unit 113, and a calculation unit 114. The CPU 101 uses the learned model 103 in the calculation unit 114 to estimate a disease and its possibility of being affected from multiple pieces of biometric information of each patient acquired by the acquisition unit 111. The CPU 101 calculates the risk level of each disease based on the possibility of being affected by each disease, and calculates an overall risk level and a list of modalities to be used for each patient from the risk level of the disease of each patient.

As a result, the medical diagnosis support device 100 according to the first embodiment can identify the disease of the subject and the possibility of the subject being afflicted, thereby enabling the selection of high-priority patients for treatment, and can also easily arrange for the modalities to be used, allowing many patients to be rescued and treated efficiently.

Next, we will explain how to generate a trained model for the medical diagnosis support device 100.

[Generating a trained model]
The trained model 103 uses prepared biological information (such as vital data) as input, and a large number (i sets (i is, for example, several thousand to several hundred thousand)) of data sets with disease possibility as output as training sample data, and performs machine learning using this. As a learning device (not shown), a standalone high-performance computer using a CPU and GPU processor, or a cloud computer can be used.

In this embodiment, the learning method using a neural network formed by combining perceptrons as a learning device (not shown) is described using a flowchart, but this embodiment is not limited to this, and various methods can be used as long as they are supervised learning.

Specifically, the learning device (not shown) can apply random forests, support vector machines (SVMs), boosting, Bayesian, network linear discrimination methods, nonlinear discrimination methods, etc.

[How to train a trained model]
FIG. 7 is a flowchart showing the process of the learning method of the trained model 103 of the medical diagnosis support device 100.

The learning device (not shown) reads the learning sample data, which is the teacher data. For example, if it is the first time, the first set of learning sample data (biometric information, disease possibility) is read, and if it is the i-th time, the i-th set of learning sample data is read (step S201).

Figure 8 is an explanatory diagram showing biometric information (vital data) prepared in advance for an arbitrary patient X. As shown in Figure 8, the biometric information (vital data) is composed of input data such as age, sex, blood pressure, respiratory rate, pulse rate, electrocardiogram, body temperature, oxygen saturation, medical history, smoking history, purpose of visit, method of visit, and others, and teacher data such as cerebral aneurysm, fracture, lung cancer, diabetes, and others. The trained model 103 then reads this biometric information (vital data) in order as training sample data.

The learning device (not shown) inputs the input data from the loaded learning sample data into the neural network (step S203).

The learning device (not shown) compares the neural network's estimation results, i.e., the estimated disease, with the input training data (step S205).

The learning device (not shown) adjusts the parameters based on the comparison results (step S207). For example, the learning device (not shown) adjusts and updates the parameters by performing a process called back-propagation (backpropagation method) so as to reduce the error in the comparison results.

When the learning device (not shown) has completed all processing of all data from the first to the i-th set (YES in step S209), it proceeds to step S211. On the other hand, when the learning device has not completed all processing of all data (NO in step S209), it returns to step 201, reads the next learning sample data, and repeats the processing from step S201 onwards.

The learning device (not shown) stores the trained model 103 constructed in the processing up to step S207 (step S211) and ends the processing.

As a result, the CPU 101 of the medical diagnosis support device 100 according to the fourth embodiment can use the learned model 103 in the estimation unit 112 to estimate a disease and its possibility of being affected from multiple pieces of biometric information of each patient acquired by the acquisition unit 111.

100 Medical diagnosis support device 101 CPU
102 Operation reception unit 103 Trained model 104 Output unit 105 Display unit 106 ROM
107 RAM
108 External storage device 109 Disease estimation program 111 Acquisition unit 112 Estimation unit 113 Control unit 114 Calculation unit

Claims (5)

An acquisition unit that acquires a plurality of pieces of biological information of each patient;
an estimation unit that estimates a disease and a possibility of the disease from the plurality of biological information of each patient acquired by the acquisition unit using a trained identification device;
a control unit that outputs to an output unit the disease and the possibility of the disease of each patient estimated by the estimation unit;
A calculation unit that calculates a risk level of each of the diseases based on the possibility of each of the diseases estimated by the estimation unit,
The calculation unit is
calculating a list of modalities recommended for use with each of the patients from the calculated risk level of each of the diseases of each of the patients;
Medical diagnostic support equipment. The calculation unit is
Calculating an overall risk level for each of the patients from the risk level of the disease for each of the patients;
The medical diagnosis support device according to claim 1 . The calculation unit is
performing triage for the plurality of patients based on the overall risk levels of the plurality of patients;
The medical diagnosis support device according to claim 2 . acquiring a plurality of biometric information for each patient;
A step of estimating a disease and a possibility of the disease from a plurality of biometric information of each patient using a trained identification device;
outputting the disease and the possibility of disease of each of the patients to an output unit;
A calculation step of calculating a risk level of each of the diseases based on the estimated possibility of each of the diseases;
Further, in the calculation step,
calculating a list of modalities recommended for use with each of the patients from the calculated risk level of each of the diseases of each of the patients;
A method for controlling a medical diagnosis support device. On the computer,
acquiring a plurality of biometric data for each patient;
estimating a disease and its possibility from a plurality of biometric information of each patient using a trained identification device;
outputting the disease and the possibility of disease of each of the patients to an output unit;
A calculation step of calculating a risk level of each of the diseases based on the estimated possibility of suffering from each of the diseases,
Further, in the calculation step,
calculating a list of modalities recommended for use with each of the patients from the calculated risk level of each of the diseases of each of the patients;
A program characterized by causing a user to execute the following:

Images