JP4087640B2 - Disease candidate information output system - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a disease candidate information output system, and more particularly to a disease candidate information output system that automatically outputs disease candidate information such as a disease candidate name based on subject image data and disease state information.
[0002]
[Prior art]
In medical facilities such as private hospitals and general hospitals, diagnosis of patients is performed by one or many doctors. Usually, each patient has a doctor in charge, and the doctor in charge asks about the progress to date, subjective symptoms, treatment status, etc., and checks the symptoms by direct examination.
[0003]
In addition, medical images such as blood tests, urine tests, X-ray images, CT images, and MR images are taken at the discretion of the doctor in charge, and the doctor in charge makes a final diagnosis with reference to these test results, Determine the name of the disease and the need for additional testing and treatment policy. The patient is treated by additional examination or medication based on the determined diagnosis result, and the doctor in charge continues to check the cure status or progress status of the disease.
[0004]
[Problems to be solved by the invention]
Conventionally, a method in which diagnosis and diagnosis are performed by one doctor in charge is common, but the diagnostic ability and technology of a doctor depend on experience, etc. There was a difference. Also, in cases other than unusual cases or fields of expertise, even experienced doctors may be lost in diagnosis.
[0005]
In response to these problems, a medical diagnosis support apparatus that searches for and outputs disease information based on input disease state information has already been proposed by the present applicant (Japanese Patent Laid-Open No. 2002-32476). This device inputs disease state information such as image interpretation results, diagnosis (diagnosis) results, inquiry results, pathological examination results, etc., so that the name of the disease corresponding to the entered disease state information, a representative image of the disease The purpose is to enable appropriate diagnosis without depending on doctor's experience and skills, etc. by utilizing various medical information. It is what.
[0006]
However, this medical diagnosis support apparatus is generally used to input information that can be entered in a medical chart as medical condition information, and is required to input findings by an interpreter when making a diagnosis accompanied by image interpretation. Therefore, the output information may be influenced by the ability of the image interpreter.
[0007]
In view of the above circumstances, the present invention outputs disease candidate information that can output diagnosis support information for determining an appropriate diagnosis result and treatment policy without being affected by the ability of a doctor or an image interpreter. The purpose is to provide a system.
[0008]
[Means for Solving the Problems]
The disease candidate information output system according to the present invention relates to an image analysis unit that obtains image analysis information by image analysis of a subject image based on image data representing a subject image accompanied by disease state information, and a diagnosis performed in the past. Based on image analysis information and medical condition information using case database storing a lot of case image data and / or diagnostic data and case image data and / or diagnostic data stored in case database A disease candidate information determination unit that determines disease candidate information for an image and an output unit that outputs the determined disease candidate information are provided.
[0009]
Here, the medical condition information means medical examination information, inquiry information, examination data information, medical history information, and the like. The examination information indicates a medical condition confirmed by a doctor such as swelling of the throat, body temperature, coughing, and so on. Interview information is information that has been heard from the patient, such as the progress of a medical condition such as a sore throat or cough from several days ago, the progress of treatment by medication, etc., the age, sex, and weight of the patient. is there. The test data information indicates the results of tests such as blood tests and urine tests, and the medical history information indicates the past medical history (disease name, time, degree, etc.) of the patient himself / her relative.
[0010]
The disease candidate information obtained by the disease candidate information determination unit means a highly likely disease candidate name and its possibility, a medicine to be administered, a treatment policy, a necessary additional examination, etc., and the disease candidate information determination unit It is preferable to obtain a plurality of disease candidate names in order from the most likely ones together with the treatment policy and additional examinations for the disease candidate names.
[0011]
Diagnostic data stored in the case database means past condition information and disease information (disease name, treatment policy, additional examination, etc. determined by the doctor), and related image data is stored in the case database. If it has been stored, it is stored with the associated case image data.
[0012]
The image analysis unit analyzes the subject image by various image analysis. For example, the image analysis unit detects an abnormal shadow in the image based on the image data of the subject, and detects the detection result (the type and position of the detected abnormal shadow). , Size, degree of malignancy, etc.) can be obtained as image analysis information, or case image data representing an image similar to a subject image from case image data stored in a case database It is also possible to search by analysis and obtain the searched case image data and / or diagnostic data accompanying the case image data as image analysis information.
[0013]
Here, the case image data representing an image similar to the subject image means image data having a portion whose image characteristics are similar to image data representing the whole and / or part of the subject image. It is. When searching, similarity may be determined by a predetermined similarity determination item.
[0014]
The similarity determination items include, for example, the shape, size, shading pattern characteristics, position, direction, number, feature amount, etc. of the shadow appearing in the image, shading pattern characteristics of the entire image, structure (lung, heart, etc.) Etc.) such as shape and size, and the like, which means that various items can be determined based on image data. The similarity may be determined, or the similarity may be determined for a combination of a plurality of items. The feature amount is an index value used to determine the malignancy degree of the detected abnormal shadow. In the case of a tumor shadow, the circularity, edge information of the outline, density histogram information of the region, etc. (See JP-A-9-167238).
[0015]
“Determination of disease candidate information based on image analysis information and disease state information using case image data and / or diagnosis data stored in the case database” means that the data stored in the case database is This means that disease candidate information is determined based on image analysis information obtained from the processing target image (subject image) and information indicating a medical condition accompanying the processing target image.
[0016]
The disease candidate information output system of the present invention further includes an input unit for inputting image data representing a subject image from a client terminal connected via a network, and the output unit transmits the disease candidate information via the network to the client terminal. Can also be output.
[0017]
Here, the network means an intranet in the hospital, the Internet, a dedicated line, or the like.
[0018]
【The invention's effect】
According to the disease candidate information output system of the present invention configured as described above, a case image related to a diagnosis made in the past based on image analysis information and disease state information obtained by image analysis of a subject image Since disease candidate information is determined using data and diagnostic data, appropriate disease candidate information (disease candidate name, treatment policy, etc.) is not affected by the diagnostic ability and image interpretation ability of doctors and image interpreters. Can be output.
[0019]
In particular, it is possible to reduce oversight when the subject image is a rare case or when the doctor has never seen a doctor.
[0020]
If the image analysis unit retrieves case image data similar to the subject image and obtains the retrieved case image data and diagnostic data associated with the case image data as image analysis information, the subject image Even when it is difficult to interpret and image diagnosis, appropriate disease candidate information can be obtained by referring to similar cases.
[0021]
In addition, if image data representing a subject image is input from a client terminal via a network and disease candidate information is output to the client terminal, a case database or a case database for storing a large number of case image data and diagnostic data Even in a small-scale medical facility that cannot provide a device (image analysis unit) that performs image analysis of a subject image in its own facility, it is possible to obtain disease candidate information using a large number of cases by simply providing a client terminal. Obtainable.
[0022]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of a disease candidate information output system according to the present invention will be described below with reference to the drawings. FIG. 1 is a system configuration diagram of a disease candidate information output system according to the present invention.
[0023]
The disease candidate information output system 1 according to the present embodiment includes a case database 20 storing a large number of case image data and diagnosis data related to a diagnosis performed in the past, and disease state information accompanied by image data P1 representing a subject image. The processing unit 10 that performs processing for determining disease candidate information for the image data P1 using the case image data and diagnosis data stored in the case database 20 and the disease candidate information obtained by the processing unit 10 And an output unit 30 that outputs. Further, a display device 2 is connected to the disease candidate information output system 1, and the disease candidate information output by the output unit 30 is displayed on a monitor screen provided in the display device 2.
[0024]
The case database 20 includes an image database 20a that stores a large number of case image data and a diagnostic database 20b that stores a large number of diagnostic data. The image database 20a stores a large number of case image data of various subjects previously diagnosed at the own facility or other facilities, and the diagnosis database 20b diagnoses when diagnosed in the past at the own facility or other facilities. Many data (disease state information and disease information) are stored. In addition, the case image data stored in the image database 20a is stored in a state associated with the associated diagnostic data stored in the diagnostic database 20b.
[0025]
The medical condition information attached to the image data P1 is medical examination information (medical condition confirmed by a doctor), inquiry information (information such as medical conditions heard from a patient), examination data information, and medical history information. It is recorded in a state associated with the image data P1 on a recording medium such as a hard disk or a magnetic disk.
[0026]
Based on the input image data P1, the processing unit 10 searches the image database 20a of the case database 20 for image data P2 (case image data) representing a similar case of the image data P1 by image analysis. A case database based on the image analysis unit 11 for obtaining the image data P2 obtained as image analysis information, the image data P2 (image analysis information) obtained by the image analysis unit 11 and the medical condition information attached to the image data P1 The diagnosis data stored in the 20 diagnosis databases 20b is used to configure a disease candidate information determination unit 12 that determines disease candidate information for the image data P1.
[0027]
The image analysis unit 11 detects an abnormal shadow in the image based on the image data P1, and further selects similar cases based on the type, shape, size, etc. of the detected abnormal shadow from the image database 20a. Search. The similarity determination items (items for determining similarity) when searching for similar cases can be set in various ways. Here, the “shape” and “size” of the abnormal shadow are adopted. And That is, the image analysis unit 11 searches the image database 20a for case image data having an abnormal shadow similar in shape and size to the abnormal shadow detected based on the image data P1. The image analysis unit 11 can search for two or more similar cases of the image data P1 in descending order of similarity, and here, the similarity is high based on a certain criterion (in the case where it can be expressed numerically). (Similarity of 80% or more, etc.) All cases are searched. As the similarity, for example, a ratio of sizes and a degree of overlapping of shapes (such as a ratio of the area of the overlapping portion to the entire area) can be used.
[0028]
FIG. 2 shows a chest image P to be processed and an ROI image P ′ showing the ROI in the chest image P. In the right lung portion of the chest image P, a circular shadow suspected of being a tumor shadow appears. The ROI image P ′ is detected by the iris filter processing based on the image data P1, and has a high overall brightness and includes an isolated circular shadow. Note that details of detection of a tumor shadow and extraction of an ROI image using iris filter processing are described in Japanese Patent Application Laid-Open No. 8-294479 and the like, and are omitted here.
[0029]
Next, the operation of the disease candidate information output system of the present embodiment configured as described above will be described.
[0030]
When image data P1 to be diagnosed and accompanying medical condition information are input to this system, first, an image analysis unit 11 of the processing unit 10 detects a tumor shadow in the image based on the image data P1, and extracts an ROI image P ′. Further, based on the similarity determination item set in advance, image data P2 having a tumor shadow similar in shape and size to the tumor shadow included in the ROI image P ′ from the image database 20a is obtained. Searched.
[0031]
The search result in the image analysis unit 11 is input from the image analysis unit 11 to the disease candidate information determination unit 12, and the medical condition information input to the system accompanying the image data P1 is input to the disease candidate information determination unit 12. The Here, diagnostic data D2 related to the image data P2 and diagnostic data D3 including pathological information similar to the pathological condition information are retrieved from the diagnostic database 20b and input to the disease candidate information determining unit 12, and the diagnostic data D2, D3 Is used to determine candidate disease information for the image data P1. That is, since there are diagnostic data stored in the diagnostic database 20b with or without image data, the diagnostic data D2 related to the image data P2 that is a similar case to the image data P1 and the same as the medical condition information The candidate disease information is determined with reference to diagnostic data D3 including pathological condition information (including those that do not have related image data).
[0032]
For example, if the diagnosis information included in the medical condition information includes “fever” and the test data information includes blood test data, the diagnosis data D3 including similar data is retrieved from the diagnosis database 20b and the disease candidate information determination unit 12, and referring to the diagnosis data D3 and the diagnosis data D2, the name of the disease candidate, its possibility, the treatment policy by medication, the necessity of additional examination, and the like are determined. At this time, a predetermined number of disease candidate names are determined in descending order of possibility.
[0033]
The disease candidate name determined by the disease candidate information determination unit 12 is output and displayed on the display device 2 by the output unit 30. FIG. 3 shows an example of the display screen. The possibility is displayed as a score together with the disease candidate name. The type of medicine to be administered and the type of additional examination are also displayed.
[0034]
According to the disease candidate information output system according to the present embodiment, it is possible to automatically detect abnormal shadows and search for similar cases simply by inputting image data to be diagnosed and its disease state information, and to highly likely disease candidates. Since the name is displayed, it is possible to prevent oversight when the case is rare or difficult to interpret. In addition, since a plurality of similar cases are searched from the case database and a plurality of disease candidate names are displayed in order from the most likely one, the selection range is widened when the doctor makes a final decision, which helps to prevent misdiagnosis. In other words, even a doctor who has a lot of experience is difficult to completely prevent oversight, so more reliable diagnosis can be expected by supporting diagnosis with several cases. In addition, since diagnosis candidate information is determined using diagnosis data that does not have related image data, a wider range of diagnosis support information can be obtained.
[0035]
In the above embodiment, the image analysis unit 11 detects abnormal shadows and searches for similar cases, and the disease candidate information determination unit 12 searches the diagnosis data D2 of the searched similar cases and the diagnosis data D3 based on the disease state information. However, the type of image analysis in the image analysis unit 11, the type of image analysis information to be obtained, and the input of image data and diagnostic data to be used in the disease candidate information determination unit 12 The route and type are not limited to the above embodiment.
[0036]
For example, the image analysis unit 11 only detects abnormal shadows, and the disease candidate information determination unit 12 determines similar cases (similar cases from the case database 20 based on the detection results (types of abnormal shadows, progress, etc.) and disease state information. It may be configured to search for a case having an abnormal shadow and including similar pathological information. Further, the image analysis unit 11 obtains the image data P2 and the diagnosis data D2 as image analysis information, and the disease candidate information determination unit 12 accompanies the obtained image analysis information (image data P2 and diagnosis data D2) and the image data P1. The disease candidate information may be determined without searching for or searching for the diagnostic data D3 based on the medical condition information to be performed, or only the diagnostic data D2 associated with the image data P2 is detected by the image analysis unit 11 as image analysis information. The disease candidate information determination unit 12 may determine the disease candidate information based only on the diagnosis data D2 and the disease state information. Further, the image analysis unit 11 may obtain the image data P2 by inputting the image data in which the position of the abnormal shadow is designated by an operator or the like without detecting the abnormal shadow, or depending on the presence or absence of the abnormal shadow. Instead, the image data P2 of similar cases may be obtained. That is, the image analysis unit 11 and the disease candidate information determination unit 12 can combine various input paths and types of image data and diagnosis data to be used. The type of image analysis in the image analysis unit 11 is not limited to detection of abnormal shadows or similar image search, and any image analysis based on the image data P1 may be used.
[0037]
Further, the items of the disease state information and the disease candidate information are not limited to the contents described in the above embodiment, and various information related to the patient obtained before diagnosis can be adopted as the disease state information. As candidate information, various information that can be candidates for diagnosis results can be adopted. For example, if you prepare a medication table that records the side effects and interactions of the drug, and enter the presence / absence of allergies, age, information on the drug you are currently using, etc. Can be determined more finely and output as disease candidate information, which also helps prevent medication errors. In addition, information such as the sex of the patient may be included in the medical condition information.
[0038]
In addition, the similarity determination item in the image analysis unit 11 is not only the above-mentioned abnormal shadow `` shape '' `` size '', but also the shading pattern characteristics, position, direction, number, feature amount, etc. of the shadow appearing in the image, Combination of one or more of various items whose similarity can be determined based on image data, such as the density pattern characteristics of the entire image and the shape and size of the structure (lung, heart, etc.) Things can be adopted. The image analysis unit 11 may further include a determination item selection unit that can change the similarity determination item.
[0039]
Note that a case database used for searching for similar cases in the image analysis unit 11 and a case database used in the disease candidate information determination unit 12 may be provided separately.
[0040]
Next, another embodiment of the disease candidate information output system of the present invention will be described. FIG. 4 is a system configuration diagram of a remote system including a disease candidate information output system according to the present invention. In addition, the same code | symbol is attached | subjected about the element similar to the said 1st Embodiment, and the description is abbreviate | omitted.
[0041]
The disease candidate information output system (server) 5 according to the present embodiment includes an input unit 41 that inputs image data P1 representing a subject image together with accompanying medical condition information from a client terminal 6 connected via a network 3, Case database 20 storing a lot of case image data and diagnosis data related to the diagnoses, and determining disease candidate information for image data P1 using case image data and diagnosis data stored in case database 20 The processing unit 10 that performs the processing to be performed, and the output unit 31 that outputs the disease candidate information obtained by the processing unit 10 to the client terminal 6 via the network 3.
[0042]
The client terminal 6 is provided at a location away from the installation location of the server 5. Here, it is assumed that the server 5 is installed in a large general hospital, and the client terminal 6 is installed in a plurality of small hospitals affiliated with the general hospital. Various installation forms of the client terminal 6 and the server 5 are possible, such as a form in which the server 5 is provided in the data center and a plurality of client terminals 6 are provided in the medical facility.
[0043]
Next, the operation of the disease candidate information output system of the present embodiment configured as described above will be described.
[0044]
When the server 5 is accessed from the client terminal 6, the disease candidate information output system processing screen is displayed on the screen of the client terminal 6, and from this screen, the operator can treat the image data P1 to be processed and the medical condition information related to the image data P1 ( (Examination information, inquiry information, examination data information, medical history information) is inputted to the processing unit 10 from the input unit 41 via the network 3.
[0045]
The processing unit 10 performs the same processing as in the first embodiment, and the obtained disease candidate information is output to the client terminal 6 via the network by the output unit 31, and the disease candidate information as shown in FIG. It is displayed on the monitor screen of the client terminal 6.
[0046]
According to the disease candidate information output system configured as described above, disease candidate information can be obtained by inputting image data to be processed and disease state information from a client terminal connected via a network. Even in a large hospital, disease candidate information can be obtained using a large number of case image data and diagnosis data.
[0047]
It should be noted that usage information recording means for recording the number of times of information output by each client terminal 6 and the data transmission / reception capacity may be provided in the server 5 so that the accounting information is calculated in the server 5. In this way, the client terminal 6 can be managed by the server 5.
[0048]
In addition, a case registration means is provided in the client terminal 6 connected via the network 3, and image data that has been diagnosed by a doctor, its medical condition information, and disease information (disease name and treatment policy determined by the doctor, additional examination, etc.) Can be transmitted from the client terminal 6 to the server 5, and image data and diagnostic data can be further stored in the case database 20. Further, a form in which this case registration means is directly provided in the disease candidate information output system 1 without using the network 3 is also possible. By providing a registration means for accumulating and storing data in the case database in this way, a larger number of cases can be referred to, and thus more accurate and wide range of disease candidate information can be obtained. The disease information stored in the case database 20 is preferably disease information determined by a doctor, but the disease candidate information determined by the disease candidate information determination unit 12 may also be stored.
[0049]
In each of the above embodiments, the disease candidate information is output to the display device by the output units 30 and 31, but the disease candidate information may be output to a printer and printed.
[Brief description of the drawings]
FIG. 1 is a system configuration diagram of a disease candidate information output system according to an embodiment of the present invention. FIG. 2 is a diagram illustrating an example of a processing target image. FIG. 3 is a diagram illustrating an example of a display screen on a display device. System configuration diagram of remote system of disease candidate information output system according to another embodiment of the present invention
1 Disease candidate information output system
10 Processing section
11 Image analysis unit
12 Disease candidate information decision section
20 case database
20a image database
20b Diagnostic database
30,31 Output section
41 Input section

Claims (2)

A diagnosis database including pathological information and disease information related to a diagnosis performed in the past, and a case database storing a large number of case image data associated with the diagnosis data; and
A case image representing an image similar to the subject image from case image data stored in the case database by image analysis of the subject image based on image data representing the subject image accompanied by medical condition information An image analyzer that retrieves data and obtains the diagnostic data corresponding to the retrieved case image data ;
A search unit for searching said diagnostic data corresponding to the condition information being associated with said image data from said case database,
A disease for determining disease candidate information for the subject image based on the disease data included in the obtained diagnosis data using the diagnosis data obtained by the search unit and the diagnosis data obtained by the image analysis unit A candidate information determination unit;
A disease candidate information output system comprising: an output unit that outputs the determined disease candidate information. An input unit that inputs image data representing the subject image from a client terminal connected via a network; and the output unit outputs the disease candidate information to the client terminal via the network. claim 1 Symbol placement disease candidate information output system and said.

Images