Prognosis determination device, prognosis determination program, and prognosis determination method

Abstract

A prognosis determination device (100) is provided with: an acquisition unit (155) for acquiring a biological parameter and a medical image about a disease from which a target person has suffered; and an output unit (156) that inputs, to a discriminator, information about the medical image, information about the biological parameter, and information about a treatment scheduled to be performed, and that outputs, by the discriminator, information about a prognosis when the treatment has been performed. The discriminator is generated by a machine learning process based on actual treatment data and interpolation treatment data. The actual treatment data is about the treatment having been actually performed, and includes information about a medical image and a biological parameter of a patient before treatment, information about a treatment which has been performed on the patient, and information about a prognosis of the patient after the treatment. The interpolation treatment data is information for performing interpolation for prognosis of the patient after the treatment, and includes information that is generated from the actual treatment data and a typical model when the treatment has been performed.

Description

Technical Field

[0001] The present invention relates to a prognosis determination device, a prognosis determination program, and a prognosis determination method.

Background Art

[0002] In medical diagnosis, medical images are important materials for making judgments in the diagnosis of various diseases. For example, an optical coherence tomography (OCT) device enables three-dimensional imaging, three-dimensional structural analysis, and functional analysis of the eye to be examined, and is widely used in ophthalmic diagnosis of the retina and the like. In the cardiovascular field, a catheter-type OCT device is also used in the diagnosis of coronary arteries. diagnosis.

[0003] The tomographic image of the fundus obtained by using an OCT device in ophthalmic diagnosis of the retina includes important information for diagnosis, such as information on the diseases of the subject's eyes and information on visual acuity. However, since reading of the tomographic image of the fundus requires interpretation by a specialist, there is a problem that accurate diagnosis is difficult when there is no specialist with sufficient experience.

[0004] Therefore, in recent years, the importance of image analysis of medical images has been increasing. For example, in the above example, an image processing method for detecting abnormalities in the fundus is being developed. More specifically, Patent Document 1 discloses a method of acquiring choroid information from a fundus image of an eye to be examined and comparing the choroid information with a standard database of the choroid to determine whether there is an abnormality in the fundus.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] In addition to diagnosing the subject's disease from medical images, specialists are also required to determine an appropriate treatment method according to the subject's disease condition and the prognosis when such treatment is performed. Therefore, it is desirable to realize a mechanism capable of determining the prognosis when a predetermined treatment is performed from medical images.

[0007] The present invention has been made in view of such circumstances, and an object thereof is to provide a prognosis determination device, a prognosis determination program, and a prognosis determination method capable of determining the prognosis when a predetermined treatment is performed based on medical images.

Means for Solving the Problems

[0008] A prognosis determination device according to an embodiment of the present invention includes an acquisition unit that acquires a medical image and biological parameters related to a disease suffered by a subject, inputs information related to the medical image, information related to the biological parameters, and information related to a treatment planned to be performed on the subject to a discriminator, and an output unit that outputs information related to the prognosis of the subject when the treatment is performed by the discriminator. The discriminator is generated by machine learning processing based on actual treatment data and interpolated treatment data. The actual treatment data is data related to the actually performed treatment, and includes information related to the medical image and biological parameters of the patient before the treatment, information related to the treatment performed on the patient, and information related to the prognosis of the patient after the treatment. The interpolated treatment data is information for interpolating the prognosis of the patient after the treatment, and includes information generated from the actual treatment data and a typical model when the treatment is performed.

[0009] According to this configuration, it becomes possible to determine the prognosis when a predetermined treatment is performed based on a medical image without relying on a specialist. Further, when there are multiple treatment method options, since it is possible to determine the prognosis for each treatment, it becomes possible to present reference information for a doctor or the subject when choosing a treatment method. Furthermore, since the discriminator is obtained by using actual treatment data and interpolated treatment data as learning data, even when there is little actual treatment data, it becomes possible to prepare sufficient learning data, and the determination accuracy of the information regarding the prognosis to be output is further improved.

[0010] Further, the output unit of the prognosis determination device may be configured to further output information regarding the prognosis in the case where treatment is not performed by the discriminator.

[0011] Thereby, as comparative information in the case of performing treatment, it is also possible to provide information regarding the prognosis in the case where treatment is not performed. Also, by comparing the prognosis in the case of performing treatment and the case of not performing treatment, it becomes possible to evaluate the effect of the treatment.

[0012] Furthermore, the prognosis determination device may be configured such that at least any one of information regarding the treatment received by the subject in the past, the age of the subject, or the gender of the subject is further input to the discriminator.

[0013] Thereby, the discriminator can also consider the history of past treatments, the age and gender of the subject, and the determination accuracy of the information regarding the prognosis output by the prognosis determination device is improved.

[0014] Further, the prognosis determination device may further include a display unit that performs display control on the information regarding the prognosis for each type of treatment, for each cost of treatment, or for each prognosis.

[0015] Thereby, it is possible to compare and display the cost of treatment and the prognosis in the case of performing that treatment according to the type of treatment.

[0016] Furthermore, the typical model in the case of performing treatment may be created from information related to the treatment and information related to the prognosis after the treatment, and may not include the medical image before the treatment.

[0017] Such a typical model can be created based on publicly available information such as medical papers and clinical trial data that describe treatment methods and their prognoses.

[0018] Also, the medical image may be a fundus image, and the prognosis may be a visual acuity prognosis. Further, the fundus image may be a horizontal tomogram and a vertical tomogram.

[0019] Thereby, the prognosis determination device of the present embodiment functions as a device for determining the prognosis of visual acuity. Also, by using a horizontal tomogram and a vertical tomogram as the fundus image, the prognosis determination system tends to be further improved. In addition, without being limited to the above, the prognosis determination device of the present embodiment functions as a device for determining the prognosis of an arbitrary disease by setting the type of medical image and the disease to be targeted, etc. respectively.

[0020] Furthermore, the prognosis determination program according to the embodiment of the present invention causes the prognosis determination device to execute steps of acquiring a medical image and a biological parameter related to a disease suffered by a subject, inputting information related to the medical image, information related to the biological parameter, and information related to a treatment to be performed on the subject into a discriminator, and outputting information related to the prognosis of the subject when the treatment is performed by the discriminator. The discriminator is generated by machine learning processing based on actual treatment data and interpolated treatment data. The actual treatment data is data related to actual treatment, and includes information related to a medical image and a biological parameter of a patient before treatment, information related to the treatment performed on the patient, and information related to the prognosis of the patient after the treatment. The interpolated treatment data is information for interpolating the prognosis of the patient after treatment, and includes information generated from the actual treatment data and a typical model in the case of performing the treatment.

[0021] In addition, in the prognosis determination method according to the embodiment of the present invention, a prognosis determination device acquires a medical image and a biological parameter related to a disease suffered by a subject, inputs information related to the medical image, information related to the biological parameter, and information related to a treatment planned to be performed on the subject into a discriminator, and the discriminator outputs information related to the prognosis of the subject when the treatment is performed. The discriminator is generated by machine learning processing based on actual treatment data and interpolated treatment data. The actual treatment data is data related to actual treatment, including information on the medical image and biological parameters of the patient before treatment, information on the treatment performed on the patient, and information on the prognosis of the patient after the treatment. The interpolated treatment data is information for interpolating the prognosis of the patient after treatment, including information generated from the actual treatment data and a typical model when the treatment is performed.

Advantages of the Invention

[0022] According to the present invention, it is possible to provide a prognosis determination device, a prognosis determination program, and a prognosis determination method capable of determining a prognosis when a predetermined treatment is performed based on a medical image.

Brief Description of the Drawings

[0023]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Mode for Carrying Out the Invention

[0024] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. However, the present invention is not limited thereto, and various modifications are possible without departing from the gist thereof.

[0025] <Hardware Configuration> In one embodiment of the present invention, for example, a prognosis determination system 1 for determining the prognosis when a predetermined treatment is performed based on a medical image is constructed by an information processing device 100 that is a prognosis determination device.

[0026] FIG. 1 shows a block diagram showing an information processing device 100 included in a prognosis determination system according to an embodiment of the present invention. The information processing device 100 typically includes one or more processors 110, a communication interface 120 that controls wired or wireless communication, an input / output interface 130, a memory 140, a storage 150, and one or more communication buses 160 for interconnecting these components. Through their cooperation, the processes, functions, or methods described in the present disclosure are realized.

[0027] The processor 110 executes the processes, functions, or methods realized by the code or instructions included in the program stored in the memory 140. The processor 110 includes, by way of example and not limitation, one or more central processing units (CPUs) and GPUs (Graphics Processing Units).

[0028] The communication interface 120 transmits and receives various data to and from other information processing devices via a network. Such communication may be performed either by wire or wirelessly, and any communication protocol may be used as long as mutual communication can be executed. For example, the communication interface 120 is implemented as hardware such as a network adapter, various communication softwares, or a combination thereof.

[0029] The input / output interface 130 includes an input device that inputs various operations to the information processing device 100, and an output device that outputs the processing results processed by the information processing device 100. For example, the input / output interface 130 includes information input devices such as a keyboard, a mouse, and a touch panel, image input devices such as an OCT device and a fundus camera, and an image output device such as a display. Note that the information processing device 100 may receive a predetermined input by connecting an external input / output interface 130. For example, an image input device such as an OCT device or a fundus camera may be externally attached to the information processing device 100.

[0030] Note that the image input device can be appropriately selected according to the medical image to be acquired. The medical image in the present embodiment is not particularly limited, and examples thereof include images obtained by X-ray, CT, MRI, nuclear medicine (PET, SPECT), ultrasonic, and endoscope. The medical image may be a still image or a dynamic image (video). Further, the medical image may be image information in a general format such as JPEG, or may be image information in a medical format conforming to an international standard such as DICOM (Digital Imaging and COmmunication in Medicine). Such medical images conforming to a medical format may be accompanied by tag information such as examination conditions and examination dates of CT, etc., and the age and gender of the patient determined by other standards. The image information in the medical format may be accompanied by, for example, tag information such as examination conditions and examination dates of CT, etc., and the age and gender of the patient determined by other standards.

[0031] Memory 140 temporarily stores the program loaded from storage 150 and provides a working area for processor 110. Various data generated while processor 110 is executing the program are also temporarily stored in memory 140. Memory 140 may be, by way of non-limiting example, a high-speed random access memory such as DRAM, SRAM, DDR RAM, or other random access solid-state storage devices, and these may be combined.

[0032] Storage 150 stores programs, various functional units, and various data. Storage 150 may be, by way of non-limiting example, a non-volatile memory such as a magnetic disk storage device, an optical disk storage device, a flash memory device, or other non-volatile solid-state storage devices, and these may be combined. Other examples of storage 150 may include one or more storage devices installed remotely from processor 110.

[0033] In one embodiment of the present invention, storage 150 stores programs, functional units, and data structures, or subsets thereof. Information processing apparatus 100 is configured to function as acquisition unit 155, output unit 156, display unit 157, and learning unit 158 as shown in FIG. 1 by processor 110 executing instructions included in the programs stored in storage 150.

[0034] Operating system 151 includes, for example, procedures for processing various basic system services and executing tasks using hardware.

[0035] Network communication unit 152 is used, for example, to connect information processing apparatus 100 to other computers via communication interface 120 and one or more communication networks such as the Internet, other wide area networks, local area networks, metropolitan area networks, etc.

[0036] The target person data 153 records the personal information of the person whose prognosis is to be determined, and may also record medical images and biological parameters input to the discriminator. For example, as shown in FIG. 2, the target person data 153 may include information such as the gender, age, disease name, treatment history, medical images, vision, and blood test information of the target person. Further, the information on medical images, vision, and blood tests may include not only current information but also past information. The target person data 153 may be an electronic medical record stored in the storage 150, or may be an electronic medical record stored in a remote server.

[0037] In addition, in the present embodiment, the "biological parameter" is information of the target person other than medical images. For example, it includes information such as gender, age, height, weight, vision, hearing, blood pressure, blood test, urine test, electrocardiogram test results, diagnosis at the time of input, disease stage, past history, treatment history, etc. Further, the biological parameters may include values that serve as indicators when judging the prognosis, such as vision in eye diseases. In the following, eye diseases may be used as an example for explanation, but unless otherwise specified, vision is used as an example of a value that serves as an indicator when judging the prognosis.

[0038] In addition, in the present embodiment, the "target person" refers to a person whose prognosis is determined by the prognosis determination device of the present embodiment. On the other hand, the "patient" refers to a person who is the source of actual treatment data and the like stored in the learning data described later.

[0039] The learning data 154 is a data set used to generate a discriminator. The learning data 154 includes actual treatment data obtained in actual treatment, and further includes interpolated treatment data generated from the actual treatment data.

[0040] For example, it includes information on medical images before treatment, information on treatment, and information on the prognosis after treatment. The learning data 154 may be stored in the storage 150, or may be stored in a remote server.

[0041] The learning data 154 includes actual treatment data obtained in actual treatment, and further includes interpolated treatment data generated from the actual treatment data.

[0042] The actual treatment data included in the learning data 154 is data regarding actual treatment, and includes information on medical images and biological parameters before actual treatment, information on the treatment performed on the patient, and information on the prognosis of the patient after treatment. Note that the information on the prognosis of the patient after treatment may include information on the medical images and biological parameters of the patient. Such actual treatment data can be collected from doctors, nurses, other medical personnel, or medical institutions. The actual treatment data collected from such medical institutions, as shown in FIG. 3, is not a regularly and evenly recorded prognosis of treatment, but is mostly obtained at intervals depending on the patient's visit timing.

[0043] If an attempt is made to create a prognosis prediction model based only on such actual treatment data, a model like the dashed line in FIG. 3 will be obtained. However, since the model obtained in this way has a low acquisition frequency of the post-treatment progress data, there is no confirmation that it accurately indicates the prognosis. Also, even if it were possible to collect a significant amount of such actual treatment data from a plurality of patients, since each individual actual treatment data is insufficient to create an accurate model, it is difficult to create a model that can accurately discriminate the prognosis.

[0044] Furthermore, in reality, even for the same disease, the types of treatments administered to patients vary widely. Therefore, it is itself difficult to collect a significant amount of actual treatment data of patients who have received the same treatment for the same disease. Also, in order to further improve the prediction accuracy of the prognosis, it may be considered to collect actual treatment data for patient groups with similar ages, genders, etc. However, it is even more difficult to collect a significant amount of such actual treatment data.

[0045] In contrast, in the present embodiment, interpolation treatment data for supplementing actual treatment data is created and used as part of the learning data. The interpolation treatment data is information for interpolating the prognosis of a patient after treatment, and includes information generated from the actual treatment data and a typical model in the case of performing treatment. This interpolation treatment data supplements information on the prognosis predicted from the typical model when, for example, it is assumed that a certain treatment method is implemented for an individual having information on a certain medical image.

[0046] The typical model may include, for example, information on treatment and information on the prognosis after the treatment. The typical model may be created from information that does not include the pre-treatment medical image, or may be created from information that includes the pre-treatment medical image. Similarly, the typical model may or may not consider the pre-treatment medical image as a model showing the prognosis of a certain treatment.

[0047] An image of the typical model is shown in FIG. 4. FIG. 4 shows a typical model composed of the progress when the visual prognosis of a patient when a certain treatment is performed is recorded every other day. Such a typical model can be created based on publicly available information such as medical papers and clinical trial data that describe the treatment method and its prognosis.

[0048] Such information on medical papers or clinical trials regarding the treatment method that is the basis of the typical model includes information on the prognosis when the treatment method is implemented, particularly information on the medical effect of the treatment method, and due to its nature of showing the medical effect, the prognosis is recorded at shorter intervals than the actual treatment data routinely obtained at medical institutions. Therefore, compared with the prognosis prediction model composed only of the actual treatment data (see the dashed line in FIG. 3), the typical model (see the solid line in FIG. 4) shows the prognosis accurately.

[0049] In addition, in actual treatment data, for example, the next treatment may be performed two months after the first treatment. Therefore, there may be no data regarding the prognosis three months after the first treatment without the next treatment, or the prognosis four months after the first treatment without the next treatment (see the dashed line in Figure 3). However, in the case of a typical model, the prognosis over a longer period can be represented (see the solid line in Figure 4).

[0050] Then, the above typical model is applied to the actual treatment data to interpolate the prognosis between the actual treatment data and create interpolated treatment data. Figure 5 shows an image of the interpolated treatment data. In Figure 5, between the actual treatment data A0 immediately before treatment (0 months) and the actual treatment data A2 two months after treatment, between the actual treatment data A2 two months after treatment and the actual treatment data A3 three months after treatment, and after the actual treatment data A3 three months after treatment, the data is interpolated. The method for creating this interpolated treatment data is not particularly limited. For example, information regarding the prognosis after treatment may be interpolated or extrapolated from the actual treatment data.

[0051] By applying the above typical model to the actual treatment data to generate interpolated treatment data in this way, even if there are circumstances such as the actual treatment data not regularly recording the prognosis of the treatment, the deficiency of the actual treatment data can be compensated for by the interpolated treatment data. Therefore, the discriminator generated by the machine learning process based on the actual treatment data and the interpolated treatment data can show the prognosis more accurately.

[0052] The interpolated treatment data does not have to interpolate all of the actual treatment data as long as it interpolates the values that are indicators when judging the prognosis. For example, in Figure 5, the actual treatment data includes, in addition to the visual acuity that is an indicator when judging the prognosis, information regarding the medical images of the patient before treatment and biochemical parameters other than visual acuity. However, as long as the interpolated treatment data includes information regarding the number of days elapsed since treatment and visual acuity, it does not have to include information regarding medical images or biochemical parameters other than visual acuity.

[0053] Also, in the actual treatment data, for example, even if all of the actual treatment data A0, A2, and A3 in FIG. 5 do not include biochemical parameters other than medical images and vision, for example, only the actual treatment data A0 may include biochemical parameters other than medical images and vision, and the actual treatment data A2 and A3 may include only information regarding vision that serves as an index when determining the prognosis.

[0054] Among these, it is preferable that the actual treatment data includes at least the medical image (actual treatment data A0) in the test results immediately before treatment. The learning data configured in this case includes, as the actual treatment data, the medical image in the test results immediately before treatment, information regarding the treatment performed, and information regarding the prognosis of the patient after treatment, and includes, as the interpolated treatment data, information regarding the interpolated prognosis. That is, this learning data becomes data indicating how the prognosis will change if a certain treatment is performed on an individual in a predetermined state of a medical image.

[0055] A discriminator generated by performing machine learning processing with a sufficient number of such learning data outputs information regarding the prognosis when treatment is performed based on information regarding the medical image, information regarding the biochemical parameter, and information regarding the treatment to be performed.

[0056] The acquisition unit 155 executes a process of acquiring information regarding the medical image, information regarding the biological parameter, and information regarding the treatment to be performed on the subject from the communication interface 120 or the input / output interface 130. Each piece of information acquired by the acquisition unit 155 can be stored in the subject data 153. Hereinafter, the acquisition unit 155 will be described by taking an eye disease as an example, but the present embodiment is not limited to eye diseases.

[0057] In the case of eye diseases, the medical images acquired by the acquisition unit 155 include fundus images. Further, the information regarding the fundus images is not particularly limited, and may include, for example, tomographic images of the fundus obtained by an OCT device and fundus photographs obtained by a fundus camera. Furthermore, the fundus images may be tomographic images in the horizontal direction and tomographic images in the vertical direction.

[0058] In the case of eye diseases, the information regarding the biological parameters acquired by the acquisition unit 155 may include information regarding visual acuity, which is an index when judging the prognosis. Examples of such information regarding visual acuity include the values of visual acuity obtained by a visual acuity test using a Landolt ring, an E chart, a Snellen chart, or the like, or a visual acuity testing device. Further, the information regarding visual acuity may include known test results regarding how things look, such as an Amsler chart.

[0059] Furthermore, the information regarding the treatment planned to be performed by the acquisition unit 155 is information regarding the treatment to be performed on the disease of the subject's eye. For example, for a subject suffering from age-related macular degeneration, treatment methods such as anti-vascular endothelial growth factor therapy (anti-VEGF therapy), photodynamic therapy, and laser photocoagulation are known. Also, for anti-VEGF therapy, as types of intravitreal injection of anti-VEGF drugs, Lucentis (ranibizumab), Macugen (pegaptanib), Eylea (aflibercept), etc. are known. Although the frequency and number of injections vary depending on the state of the disease, as an example, in anti-VEGF therapy, the anti-VEGF drug is directly injected into the eye every 4 weeks or every 6 weeks. Also, there may be cases where anti-VEGF therapy and photodynamic therapy are used in combination.

[0060] The information regarding the treatment planned to be performed is information regarding such treatment methods. For example, when performing anti-VEGF therapy once every 4 weeks, when performing anti-VEGF therapy once every 6 weeks, when using a combination of anti-VEGF therapy performed once every 4 weeks and photodynamic therapy, etc., it may include a plurality of treatment patterns that can be performed on the disease of the subject's eye.

[0061] In addition, the information on the treatment to be performed may include the name of the disease, instead of or in addition to the specific treatment method as described above. For example, when the information on the treatment to be performed includes the disease name "age-related macular degeneration", it can be regarded as specifying a known treatment method for age-related macular degeneration. Thereby, information on the prognosis after treatment can be output for each of a plurality of treatment methods for age-related macular degeneration.

[0062] Furthermore, the information on the treatment to be performed may include the doctor's findings necessary for diagnosing the disease name, instead of or in addition to the disease name. For example, when the information on the treatment to be performed includes the findings "only the central part of the visual field appears distorted (metamorphopsia)" or "the central part appears dark (central scotoma)", it can be regarded as specifying a treatment method for the disease related to the findings. Thereby, information on the prognosis after treatment can be output for each of a plurality of treatment methods effective for the findings.

[0063] Note that in the above, age-related macular degeneration is taken as an example, but the target disease is not limited to the above, and it can be targeted at epiretinal membrane, macular hole, macular edema, glaucoma, cataract, and other eye-related diseases in general.

[0064] Furthermore, the information on the treatment to be performed may be associated with information on the cost related to the treatment. Examples of the information on the cost related to the treatment include, for example, the medical points for each treatment and the medical expenses in cases such as 10% co-payment, 30% co-payment, or full payment.

[0065] Note that the storage 150 may have data including the treatment and information on the cost associated with the treatment. In this case, even when the information on the treatment to be performed does not include information on the cost of the treatment, when the acquisition unit 155 acquires the information on the treatment to be performed, it can refer to the above database and associate information on the cost with the information on the treatment to be performed.

[0066] In addition, the acquisition unit 155 may acquire at least any one of information on past treatments performed on the subject, the age of the subject, or the gender of the subject. These pieces of personal information can be used as elements to improve the accuracy of the information on the prognosis after treatment output by a discriminator described later.

[0067] Each piece of information acquired by the acquisition unit 155 can be input to the information processing apparatus 100 by a doctor, a nurse, other medical personnel, or a medical institution via the communication interface 120. Also, each piece of information acquired by the acquisition unit 155 may be received by the acquisition unit 155 from another terminal or server via the input / output interface 130.

[0068] The output unit 156 executes a process of inputting information on a medical image, information on a biological parameter, and information on a treatment to be performed into the discriminator and outputting information on the prognosis when the treatment is performed. The discriminator is not particularly limited, and may be, for example, a learned model generated by machine learning processing based on the above learning data.

[0069] In addition to the above information, the output unit 156 may further input at least any one of information on past treatments performed on the subject, the age of the subject, or the gender of the subject into the discriminator. Since the prognosis can also be affected by the treatment history, age, or gender, adding these pieces of information to the information input to the discriminator can further improve the accuracy of the information on the prognosis when the treatment is performed.

[0070] Examples of the information on the prognosis when the treatment is performed include information on future changes in values that are indicators when determining the prognosis and information on the future progression of the disease. Also, when there are a plurality of the above treatment patterns, the information on the prognosis may include information on the prognosis associated with each treatment pattern. Thereby, it is possible to compare and contrast how the prognosis changes according to each treatment pattern. Also, it is possible to compare and contrast the costs associated with each treatment pattern.

[0071] Furthermore, the output unit 156 may be configured to further output information regarding the prognosis in the case where treatment is not performed, by the discriminator. This enables acquisition of information regarding the prognosis in the case where treatment is not performed. Therefore, it is possible to compare and contrast how the prognosis changes between the case where treatment is not performed and the case where treatment is performed.

[0072] By having the configuration as described above, prognosis determination can be performed based on a medical image without relying on a specialist, and it becomes possible to determine the prognosis in the case where a predetermined treatment is performed.

[0073] Also, the information processing apparatus 100 may have a display unit 157 that executes a process of display control of the information regarding the prognosis output by the discriminator on a display device. For example, the display unit 157 can perform display control of the information regarding the prognosis for each type of treatment, for each cost of treatment, or for each prognosis.

[0074] Taking an eye disease as an example, the display unit 157 can control an image output device such as a display to plot and display on the vertical axis the visual acuity and on the horizontal axis the time axis, the visual acuity one week later, two weeks later, three weeks later, etc. predicted when treatment methods A, B, C,... are performed, for each treatment method. Also, the display unit 157 can control to list-display the treatment methods on the display or the like in descending or ascending order of cost, or list-display the treatment methods in descending or ascending order of the prognosis one month later.

[0075] This enables comparison and contrast of how the prognosis changes for each type of treatment, or comparison and contrast of the treatment costs for each type of treatment. Such comparison and contrast can be used as a reference material when a doctor explains a medical treatment policy to a subject, or as a reference material when the subject selects a treatment policy.

[0076] Further, the information processing apparatus 100 may include a learning unit 158 that executes a process of generating a discriminator by machine learning processing based on learning data. Examples of the machine learning processing include supervised learning and semi-supervised learning.

[0077] As described above, the learning data 154 includes, for example, information on medical images before treatment, information on treatment, and information on prognosis after treatment. Further, the learning data includes actual treatment data acquired in actual treatment and interpolated treatment data generated from the actual treatment data.

[0078] By further including such interpolated treatment data in this way, it is possible to assume cases where treatment method A, treatment method B, treatment method C, etc. are respectively performed for the information on a medical image before one actual treatment, and it is possible to obtain information on prognosis after a plurality of treatments. Therefore, the learning data can be efficiently collected, and the accuracy of the discriminator obtained using the learning data is further improved.

[0079] Further, the learning unit 158 may further execute a process of creating interpolated treatment data generated from the actual treatment data. In creating the interpolated treatment data, for example, the learning unit 158 creates a typical model that predicts the prognosis when the treatment method is implemented based on information on the medical effect of the treatment method included in a medical paper or the like. Then, the learning unit 158 applies the typical model to the actual treatment data to interpolate the prognosis between the actual treatment data and create interpolated treatment data.

[0080] Also, when creating interpolated treatment data by interpolating or extrapolating information on prognosis after treatment from the actual treatment data, interpolation or extrapolation can be performed using the typical model that predicts the prognosis.

[0081] By creating interpolated treatment data in this way, the number of cases included in the learning data can be increased. By performing machine learning processing using the learning data including the actual treatment data and such interpolated treatment data, the information on prognosis output by the generated discriminator is further improved.

[0082] <Operation processing> Next, the operation of the information processing apparatus 100 according to the embodiment of the present invention configured as described above will be described.

[0083] (Creation of discriminator) FIG. 6 is a flowchart illustrating a machine learning process in which the learning unit 158 of the information processing apparatus 100 creates a discriminator using learning data.

[0084] In step S201, the learning unit 158 collects actual treatment data as learning data. At this time, the learning unit 158 can collect the actual treatment data from the subject data 153 of the information processing apparatus 100 or the subject data stored in another information processing apparatus or server connected via a network.

[0085] In step S202, the learning unit 158 acquires a typical model for predicting the prognosis when a predetermined treatment method is implemented. The learning unit 158 may acquire a typical model created by a doctor, other medical personnel, or a medical institution, or may acquire a typical model stored in another information processing apparatus or server connected via a network. Further, the learning unit 158 may create a typical model by machine learning processing based on information on the medical effects of treatment methods included in medical papers or the like and acquire it.

[0086] Note that, since the typical model is for the purpose of pre-learning the overall trend in the model, even if there is no case report under exactly the same conditions, the prognosis may be extracted from the closest literature or report and used. Also, if there are reports of clinical studies or post-marketing surveys, the average thereof may be used, and if there is no data of the same race as the learning target, data of other races may be used. Furthermore, if there is no clinical study for rare diseases, case reports can also be used.

[0087] Also, for example, regarding visual prognosis, since it is examined using logMAR vision, other vision indices such as ETDRS or decimal vision may be converted and used. Since the prognoses usually reported have a wide interval such as once a month, while based on linear interpolation, if it is presumed from other literature or expert knowledge that the course is not linear, such as improving once and then deteriorating, it may be followed accordingly. For the extrapolation of prognosis data for periods without reports, although basically extended linearly, for example, in cases where there are only reports that the next administration is carried out before the effect of the drug wears off, the prognosis without additional administration may be such that the data of the natural prognosis is applied thereafter. Furthermore, if there are related diseases, the prognosis data thereof can also be used.

[0088] In step S203, the learning unit 158 creates interpolated treatment data based on the actual treatment data and the typical model. For example, the learning unit 158 inputs information on the medical image before the actual treatment of the actual treatment data into the typical model, and obtains information on the assumed treatment and information on the prognosis after the assumed treatment. Thereby, the learning unit 158 can obtain interpolated treatment data including information on the medical image before the actual treatment of the actual treatment data, information on the assumed treatment, and information on the prognosis after the assumed treatment.

[0089] In step S204, the learning unit 158 creates a discriminator from the learning data including the actual treatment data and the interpolated treatment data. For example, the learning unit 158 performs machine learning using the learning data with information on the prognosis after treatment as the correct label.

[0090] (Estimation of prognosis) FIG. 7 is a flowchart illustrating the process in which the output unit 156 of the information processing apparatus 100 outputs information regarding the prognosis when treatment is performed by the discriminator. FIG. 7 is a flowchart illustrating the process in which the output unit 156 of the information processing apparatus 100 outputs information regarding the prognosis when treatment is performed by the discriminator.

[0091] In step S301, the acquisition unit 155 acquires information regarding a medical image and information regarding a biological parameter. At this time, the acquisition unit 155 may acquire information regarding a medical image and information regarding a biological parameter from another information processing device or measuring device via the communication interface 120, or may acquire information regarding a medical image and information regarding a biological parameter from an information input device such as a keyboard or a touch panel via the input / output interface 130.

[0092] In step S302, the acquisition unit 155 may acquire information regarding a treatment to be performed. For example, by the doctor operating an information input device such as a keyboard or a touch panel, the acquisition unit 155 can acquire information regarding a treatment to be performed. When the doctor's opinion on the subject's disease is age-related macular degeneration, the doctor can input that information, or can also input an opinion such as "only the central part of the visual field appears distorted" instead of the disease name. Alternatively, a specific treatment method such as "perform anti-VEGF therapy once every six weeks" can also be input. The acquisition unit 155 can acquire information regarding a treatment to be performed according to these inputs.

[0093] Further, instead of the input operation from the doctor, the acquisition unit 155 may acquire information regarding a treatment to be performed, such as the subject's disease name or symptoms, from the electronic medical record information stored in the storage of the information processing device or a server accessible by the information processing device.

[0094] In step S303, the output unit 156 inputs information regarding a medical image, information regarding a biological parameter, and information regarding a treatment to be performed to the discriminator, and the discriminator outputs information regarding the prognosis when the treatment is performed. At this time, when the information regarding the treatment to be performed includes a plurality of treatment methods, the output unit 156 can output information regarding the prognosis for each treatment method.

[0095] In addition, the output unit 156 can attach information regarding the cost related to the treatment for each treatment method and output information regarding the prognosis. Thereby, for each treatment method, the cost when performing that treatment can be compared.

[0096] In step S304, the display unit 157 performs display control of the output information regarding the prognosis on an image output device such as a display. The display control method is not particularly limited. For example, a graph plotting the progress of the prognosis with the vertical axis being visual acuity and the horizontal axis being the time axis, a graph in which such graphs are superimposed for each treatment method, or a list arranged in ascending or descending order of the prognosis visual acuity or treatment cost, etc. may be mentioned.

[0097] In the prognosis determination system provided by the information processing apparatus 100 according to the embodiment of the present invention described above, for example, without relying on a doctor with higher expertise for a specific disease among ophthalmologists such as macular specialists, it becomes possible to determine the prognosis when performing a predetermined treatment based on a medical image. Also, when there are multiple treatment method options, since it is possible to determine the prognosis for each treatment, it becomes possible to present an index for a doctor or the subject when choosing a treatment method.

[0098] As described above, the present invention is not limited to the above-described embodiments and examples, and various modifications are possible without departing from the gist thereof. That is, the above embodiments are merely illustrative in every respect and are not to be construed in a limiting sense.

[0099] For example, the acquisition unit 155 may acquire information regarding the prognosis after treatment from the electronic medical record information and pass it to the learning unit 158. Thereby, the learning unit 158 can add information regarding the prognosis after treatment to the learning data and perform additional learning.

[0100] Also, the above prognosis determination system may be configured by a server 200, which is a prognosis determination device, and a user terminal 300 that are communicably connected via a communication network such as the Internet.

[0101] Figure 8 shows the processing sequence of a prognosis determination system composed of a server 200, which is a prognosis determination device, and a user terminal 300. The server 200 is communicably connected to the user terminal 300 via a communication network such as the Internet. The server 200 provides a prognosis determination system that inputs information related to a medical image, information related to vision, and information related to a planned treatment received from the user terminal 300 into a discriminator, outputs information related to the prognosis when the treatment is performed, and transmits the output information to the user terminal 300.

[0102] Here, the server 200 is an example of an information processing device that implements all or part of the prognosis determination system of the present invention, and can have the hardware configuration and the configuration of the functional units of the above-described information processing device 100. Also, the user terminal 300 may be a normal computer equipped with a display capable of transmitting and receiving information and displaying information.

[0103] In step S401, the user terminal 300 transmits information related to a medical image, information related to vision, and information related to a planned treatment to the server 200. Then, in step S402, the acquisition unit of the server 200 acquires information related to a medical image, information related to vision, and information related to a planned treatment.

[0104] In step S403, the output unit of the server 200 inputs information related to a medical image, information related to vision, and information related to a planned treatment received from the user terminal 300 into a discriminator, and outputs information related to the prognosis when the treatment is performed.

[0105] In step S404, the server 200 transmits the output information related to the prognosis to the user terminal 300. Then, in step S405, the user terminal 300 can perform display control of the information related to the prognosis on a display device.

[0106] FIG. 9 shows an example of a graph in which the progress of the prognosis is superimposed and displayed for each treatment method on the display device of the user terminal 300. The vertical axis of the graph indicates visual acuity, and the horizontal axis indicates the number of weeks. In FIG. 9, the current visual acuity of the subject is 0.5, and the progress of the prognosis when each treatment method is implemented is shown. According to such a graph, it can be seen that if treatment is not performed (no treatment), the visual acuity of the subject will decrease, and if treatment is performed, recovery of visual acuity can be expected. Also, taking aflibercept as an example, it is possible to determine the treatment policy at the next medical examination, such as that visual acuity can be maintained by performing treatment again 10 to 16 weeks later.

[0107] Thus, any user terminal 300 that can access the server 200 can use the prognosis determination system, and the regional gap in medical levels can be eliminated.

Description of Reference Numerals

[0108] 1... prognosis determination system, 100... information processing device, 110... processor, 120... communication interface, 130... input / output interface, 140... memory, 150... storage, 151... operating system, 152... network communication unit, 153... subject data, 154... learning data, 155... acquisition unit, 156... output unit, 157... display unit, 158... learning unit, 160... communication bus, 200... server, 300... user terminal

Claims

1. An acquisition unit that acquires medical images and biological parameters related to a disease suffered by a subject; Information regarding the medical image, information regarding the biological parameter, and information regarding a treatment planned to be performed on the subject are input to a discriminator, and an output unit that outputs information regarding the prognosis of the subject when the treatment is performed by the discriminator; and The discriminator is generated by machine learning processing based on actual treatment data and interpolated treatment data; The actual treatment data is data regarding the treatment actually performed, and includes information regarding a medical image and a biological parameter of a patient before treatment, information regarding the treatment performed on the patient, and information regarding the prognosis of the patient after the treatment; The interpolated treatment data is information for interpolating the prognosis of the patient after treatment, and includes information generated from the actual treatment data and a typical model when the treatment is performed. The typical model is created based on publicly available information including medical papers and / or clinical trial data that disclose information recording the prognosis at intervals shorter than the actual treatment data for the actual treatment and its prognosis; A prognosis determination device.

2. The output unit further outputs, by the discriminator, information regarding the prognosis when the treatment is not performed; The prognosis determination device according to Claim 1.

3. At least any one of information regarding a treatment received by the subject in the past, the age of the subject, or the gender of the subject is further input to the discriminator; The prognosis determination device according to Claim 1 or 2.

4. The device further includes a display unit that controls display of the information regarding the prognosis for each type of treatment, for each cost of the treatment, or for each prognosis; The prognosis determination device according to any one of Claims 1 to 3.

5. The typical model when the treatment is performed includes information regarding the treatment and information regarding the prognosis after the treatment, and is created from information that does not include a medical image before the treatment; The prognosis determination device according to any one of Claims 1 to 4.

6. The medical image is a fundus image, and the prognosis is a visual prognosis; The prognosis determination device according to any one of Claims 1 to 5.

7. The fundus image is a horizontal tomographic image and a vertical tomographic image; The prognosis determination device according to Claim 6.

8. As the information regarding the prognosis, a display unit that controls display to further include a graph in which graphs plotting the progress of the prognosis with the vertical axis being visual acuity and the horizontal axis being the time axis are superimposed for each treatment method is provided. The prognosis determination device according to any one of claims 1 to 3.

9. In the prognosis determination device, a step of acquiring a medical image and biological parameters regarding the disease suffered by the subject; information regarding the medical image, information regarding the biological parameters, and information regarding the treatment planned for the subject are input to a discriminator, and the discriminator outputs information regarding the prognosis of the subject when the treatment is performed; and the steps are executed. The discriminator is generated by machine learning processing based on actual treatment data and interpolated treatment data. The actual treatment data is data regarding actual treatment, and includes information regarding the medical image and biological parameters of the patient before treatment, information regarding the treatment performed on the patient, and information regarding the prognosis of the patient after the treatment. The interpolated treatment data is information for interpolating the prognosis of the patient after treatment, and includes information generated from the actual treatment data and a typical model when the treatment is performed. The typical model is created based on publicly available information including medical papers and / or clinical trial data that disclose information recording the prognosis at intervals shorter than the actual treatment data for the actual treatment and its prognosis. A prognosis determination program.

10. When the prognosis determination device acquires a medical image and biological parameters regarding the disease suffered by the subject; information regarding the medical image, information regarding the biological parameters, and information regarding the treatment planned for the subject are input to a discriminator, and the discriminator outputs information regarding the prognosis of the subject when the treatment is performed; and the steps are executed. The discriminator is generated by machine learning processing based on actual treatment data and interpolated treatment data. The actual treatment data is data regarding actual treatment, and includes information regarding the medical image and biological parameters of the patient before treatment, information regarding the treatment performed on the patient, and information regarding the prognosis of the patient after the treatment. The interpolation treatment data is information for interpolating the prognosis of the patient after treatment, and includes information generated from the actual treatment data and a typical model in the case of performing the treatment. The typical model is created based on publicly available information including medical papers and / or clinical trial data that disclose information recording prognoses at intervals shorter than the actual treatment data for the actual treatment and its prognosis. Prognosis determination method.

Images