JP5661890B2 - Information processing apparatus, information processing method, and program - Google Patents

Description

The present invention relates to an information processing apparatus and an information processing method for processing Cases data, a program for executing the information processing method in a computer, as well as a computer-readable storage medium storing the program.

  Conventionally, for the purpose of reducing the burden on a doctor who interprets a medical image (such as an X-ray image, CT image, or MRI image), a medical site is automatically detected by digitizing the medical image and analyzing the image, Medical image processing apparatuses that perform computer-aided diagnosis have been developed. Hereinafter, the computer-aided diagnosis is referred to as CAD (Computer-Aided Diagnosis).

  In such CAD, abnormal shadow candidates are automatically detected as disease sites. In this abnormal shadow detection process, image data representing a radiographic image such as an X-ray image is processed by a computer to detect abnormal tumor shadows due to cancer or the like, high-density microcalcification shadows, and the like. Then, by presenting the detection result, it is possible to reduce the burden on the doctor's interpretation and improve the accuracy of the interpretation result.

  Further, among CAD techniques, a technique for automatically diagnosing benign / malignant and disease types from detected abnormal shadow candidates has been developed. As a technique for realizing this, the following Non-Patent Document 1 introduces a technique for estimating the malignancy of breast cancer in mammography using a Bayesian network. Here, the Bayesian network is a probability model in which a qualitative dependency relationship between uncertain events is represented by a graph and a quantitative relationship between individual events is represented by a conditional probability table. By inputting events that can be observed before diagnosis, such as analysis data of medical images and clinical data such as patient attributes and test results, this probability model is not observed before diagnosis such as the malignancy of breast cancer. The probability distribution of events that are

  By the way, when a doctor interprets a medical image and performs a medical practice such as diagnosis or treatment, there are cases where the affected part shown in the medical image being interpreted has an unfamiliar image feature or multiple diseases with similar image features In such cases, it may be difficult to determine the diagnosis name or the treatment. In such a case, the doctor may consult other veteran doctors or read a medical text or other literature to read a description of the image features regarding the suspected disease name. Alternatively, doctors should consult medical literature with photographs, find photographs similar to the affected area in the medical image being interpreted, and refer to the name of the disease indicated in the photograph for reference to medical practice. It may be.

  However, there are not always other doctors who can consult with you, and just because you looked at the literature, you may not always find a comment or description of the image features similar to the affected area in the medical image being interpreted. . Therefore, when a doctor interprets a medical image and performs a medical practice such as diagnosis or treatment, hesitation in the medical practice may not be solved.

  In order to solve such problems by electronic means, a similar case search apparatus for searching for similar cases has been proposed in recent years. The basic idea of this similar case search device is to support medical practice by searching for case data based on some criteria from case data accumulated in the past and displaying it to the doctor. is there.

  As a general similar image search method, there is known a method of searching image data of a similar medical image whose image feature quantity is similar to a medical image being interpreted from an image database accumulated in the past. Further, when a similar case is presented on the same screen as a medical image to be treated during a medical practice, it is effective in supporting the medical practice because reference information for medical treatment can be grasped at a time. However, since the amount of information that can be displayed on the same screen during a medical practice is limited, it is necessary to present information efficiently. On the other hand, in the following Patent Document 1, many similar cases are presented when the patient to be diagnosed is a first visit, and past images of the patient are presented when the patient to be diagnosed is other than the first visit. A method of presenting fewer cases is described.

JP 2008-102665 A

Burnside ES, Rubin DL, Fine JP, Shachter RD, Sisney GA, Leung WK, "Bayesian network to predict breast cancer risk of mammographic microcalcifications and reduce number of benign biopsy results: initial experience", Radiology. 2006 Sep; 240 (3) : 666-73.

  However, since the technique described in Patent Literature 1 only changes the number of similar cases presented by determining whether or not it is the first visit, the display method is not necessarily changed according to the necessity of similar case presentation. Absent. With the technique described in Patent Document 1, for example, in the case of easy-to-treat cases, many similar cases are presented when it is determined that the first visit is made even though the need for similar case presentation is low. . On the other hand, with the technique described in Patent Document 1, for example, in cases where it is difficult to perform medical treatment, only a few similar cases are presented if it is determined that the case is other than the first visit, even though the need for similar case presentation is high. become.

  This invention is made in view of such a problem, and it aims at providing the mechanism which can change the display method of the said similar case according to the necessity of presentation of a similar case in the case of a medical treatment act. .

The information processing apparatus of the present invention stores a plurality of case data, and each case data includes case data storage means that includes information on the degree of difficulty of medical practice determined for each case data. , diagnosis 療対 a similar case search means for searching for at least one obtains the similar case data is a case data including examination information similar examination information and class of elephant picture, the difficulty of the intervention in the similar case data Acquisition means for acquiring information on the degree of difficulty of the medical practice in the medical image .
According to another aspect of the information processing apparatus of the present invention, a plurality of case data are stored, and each case data includes information on the image feature amount and information on the difficulty level of the medical practice determined for each case data. When performing a search for acquiring at least one similar case data that is similar to the medical object image, from the case data storage means including the information on the image feature amount in the medical object image is acquired. , Similar case retrieval means for retrieving case data including image feature amount information similar to the image feature amount information in the acquired medical object image as the similar case data, and difficulty of the medical practice in the similar case data Acquisition means for acquiring information on the degree of difficulty of the medical practice in the medical image based on the information on the degree.
The present invention also includes an information processing method by the information processing apparatus described above, a program for causing a computer to execute the information processing method, and a computer-readable storage medium that stores the program.

  According to the present invention, it is possible to change the display method of the similar case according to the necessity of presenting the similar case during the medical practice. For this reason, it becomes possible to improve the efficiency of the medical practice by the doctor.

It is a schematic diagram which shows an example of schematic structure of the information processing system which concerns on the 1st Embodiment of this invention. It is a flowchart which shows an example of the process sequence of the information processing method by the information processing apparatus which concerns on the 1st Embodiment of this invention. It is a schematic diagram which shows the 1st Embodiment of this invention and shows an example of the content of the case data stored in the case database shown in FIG. FIG. 8 is a schematic diagram illustrating an example of a display method M _{1 according to} the first embodiment of the present invention when the initial diagnosis flag F is True (that is, the initial diagnosis) and the display recommendation degree P is P <T. Shows a first embodiment of the present invention, is a schematic view showing an example of a display method M ₂ at the time of and display recommendation degree P when initial diagnosis flag F is True (ie first visit) is P ≧ T. FIG. 9 is a schematic diagram illustrating an example of a display method M ₃ when the first diagnosis flag F is False (that is, other than the first diagnosis) and the recommended display degree P is P <T according to the first embodiment of this invention. It shows a first embodiment of the present invention, a schematic diagram and displays the recommendation degree P when initial diagnosis flag F is False (i.e., other than the first visit) shows an example of the display method M ₄ when the P ≧ T. FIG. 8 is a schematic diagram illustrating an example of a display method M _{5 according to} the first embodiment of the present invention when the initial diagnosis flag F is True (that is, initial diagnosis) and the display recommendation degree P is P ≧ T ′. It is a flowchart which shows an example of the detailed process sequence in the display recommendation degree calculation process of the similar case performed by step S104 of FIG. FIG. 10 is a schematic diagram of a correction function u = f (x) showing an example of a relationship between a doctor's skill level x and a correction coefficient u, which is used when calculating a display recommendation level correction coefficient in step S907 of FIG. The correction function v = g showing an example of the relationship between the correction coefficient v and the degree of difference y between the attribute of the diagnosis target image and the doctor's specialty, which is used when calculating the correction coefficient for the display recommendation level in step S907 in FIG. It is a schematic diagram of (y). It is a schematic diagram which shows an example of schematic structure of the information processing system which concerns on the 2nd Embodiment of this invention. It is a flowchart which shows an example of the process sequence of the information processing method by the information processing apparatus which concerns on the 2nd Embodiment of this invention. It is a flowchart which shows an example of the detailed process sequence in the display recommendation degree calculation process of the similar case performed by step S203 of FIG. It is a flowchart which shows an example of the process sequence of the information processing method by the information processing apparatus which concerns on the 3rd Embodiment of this invention. It is a schematic diagram which shows the 3rd Embodiment of this invention and shows an example of the content of the case data stored in the case database shown in FIG. The third embodiment of the present invention is shown, and the display when the initial diagnosis flag F is False (that is, other than the initial diagnosis) and the parameter α _P of the display recommendation degree _P is α _P ≧ T and α _P > β _P is a schematic view showing an example of a method M _8. The third embodiment of the present invention is shown, and the display when the initial diagnosis flag F is False (that is, other than the initial diagnosis) and the parameter β _P of the display recommendation degree _P is β _P ≧ T and β _P ≧ α _P is a schematic view showing an example of a method M _9. It is a schematic diagram which shows an example of the hardware constitutions of the information processing system which concerns on the 4th Embodiment of this invention.

  Hereinafter, embodiments (embodiments) for carrying out the present invention will be described with reference to the drawings.

(First embodiment)
First, a first embodiment of the present invention will be described.

  In the information processing system according to the present embodiment, examination information including a medical image of a subject (patient) is associated with information indicating the difficulty of a medical practice based on the examination information (diagnostic practice in this embodiment). It is assumed that a case database for storing case data is provided. Here, the difficulty of the diagnostic action used in the present embodiment is defined as “diagnosis difficulty”.

  In addition, in this embodiment, before displaying the diagnosis screen, a similar case is searched from the case database, and it is recommended to present a similar case based on the “diagnosis difficulty” associated with the searched similar case. The display method of a similar case is determined according to this value. This is because the “diagnosis difficulty” of a case related to a medical image similar to the medical image to be diagnosed corresponds to the “diagnosis difficulty” of the medical image to be diagnosed, and it is necessary to present a similar case according to this value It is based on the idea that gender changes. Thereby, in the diagnostic action, information presentation to an appropriate doctor according to the necessity of presenting similar cases due to the difficulty of diagnosis can be realized. Here, the degree of recommendation of similar case presentation is defined as “display recommendation degree”.

FIG. 1 is a schematic diagram illustrating an example of a schematic configuration of an information processing system 10-1 according to the first embodiment of the present invention.
As shown in FIG. 1, the information processing system 10-1 includes an information processing apparatus 100-1, a case database 210, a medical image database 220, and a medical information database 230.

Here, the information processing apparatus 100-1 processes case data including examination information of a subject (patient).
The case database 210 is case data storage means for storing case data in which examination information of a subject is associated with information on the degree of difficulty of medical practice based on the examination information. That is, the case database 210 stores case data of cases where medical treatment such as already diagnosed or treated has already been performed.
The medical image database 220 is a medical image storage unit that stores image information related to a medical image captured by a medical image capturing apparatus (not shown), and an existing PACS or the like can be used. Examples of types of medical images include simple X-ray images (X-ray images), X-ray CT images, MRI images, PET images, SPECT images, and ultrasonic images. The medical image image information stored in the medical image database 220 stores not only the medical image itself but also supplementary information of the medical image. For example, when the medical image is a chest X-ray CT image, imaging information such as imaging date / time, imaging modality (X-ray CT), imaging region (chest), and imaging conditions can be stored as supplementary information.
The medical information database 230 is medical information storage means in which medical information is stored, and examples of the medical information include basic information about diseases, information on contents such as medical information journals and academic journal papers. The medical information database 230 may use information on the Web that can be used for Internet search as a database.

  The information processing apparatus 100-1 includes a medical image acquisition unit 101, a feature amount extraction unit 102, a past image acquisition unit 103, a similar case search unit 104, a display recommendation level calculation unit 105, a medical reference information search unit 106, a storage unit 107, The display method determining unit 108 and the display unit 109 are included.

  The medical image acquisition unit 101 performs processing for acquiring, from the medical image database 220, image information related to a medical image of a subject that is a target of a medical practice such as diagnosis or treatment. Here, a medical image of a subject that is a subject of a medical practice is referred to as a “medical target image”. In the present embodiment, a diagnostic action is applied as a medical practice, and a medical image of a subject that is a target of the diagnostic action is referred to as a “diagnosis target image”.

  The feature amount extraction unit 102 performs a process of extracting an image feature amount from the medical treatment target image acquired by the medical image acquisition unit 101 (diagnosis target image in the present embodiment).

  The past image acquisition unit 103 performs processing for acquiring image information related to the patient's past image in the medical treatment target image (diagnosis target image in the present embodiment) from the medical image database 220.

  The similar case retrieval unit 104 performs a process of retrieving similar case data, which is case data including examination information whose characteristics are similar to the examination information of the medical treatment target, from the case database 210.

  The display recommendation level calculation unit 105 calculates the display recommendation level of the similar case data based on the difficulty level of the medical practice in the similar case data searched by the similar case search unit 104 (diagnosis difficulty level in this embodiment). Process.

The medical reference information search unit 106 performs processing for searching medical reference information such as a disease from the medical information database 230.
The storage unit 107 stores various types of information acquired by the medical image acquisition unit 101, the feature amount extraction unit 102, the past image acquisition unit 103, the similar case search unit 104, the display recommendation degree calculation unit 105, and the medical reference information search unit 106. Remember.

  The display method determination unit 108 performs processing for determining a display method of the similar case data based on the difficulty level of the medical practice in the similar case data searched by the similar case search unit 104 (diagnosis difficulty level in the present embodiment). Do. More specifically, the display method determination unit 108 uses the display recommendation level calculated based on the difficulty level of the medical practice in the display recommendation level calculation unit 105 (diagnosis difficulty level in this embodiment). Processing to determine the display method is performed.

  The display unit 109 performs a process of displaying similar case data and the like according to the display method determined by the display method determination unit 108.

Next, the operation of each component of the information processing apparatus 100-1 according to the first embodiment will be described using the flowchart of FIG.
FIG. 2 is a flowchart illustrating an example of a processing procedure of an information processing method by the information processing apparatus 100-1 according to the first embodiment of the present invention.

<Step S101>
First, in step S101 of FIG. 2, the medical image acquisition unit 101 acquires, from the medical image database 220, image information related to a medical image (diagnosis target image) that is a target of image diagnosis designated by a doctor.

  Then, the medical image acquisition unit 101 transmits the acquired image information of the diagnosis target image to the feature amount extraction unit 102 and the storage unit 107.

<Step S102>
Subsequently, in step S102, the feature amount extraction unit 102 first acquires image information of the diagnosis target image from the medical image acquisition unit 101. Then, the feature amount extraction unit 102 extracts an image feature amount from a medical image (diagnosis target image) included in the image information acquired from the medical image acquisition unit 101.

  Specifically, when extracting the image feature amount, when the acquired diagnostic object image is a chest X-ray CT image, the feature amount extraction unit 102 first performs image analysis on the chest X-ray CT image to detect an affected part. After performing the above, an area surrounding the affected area is set. This region is defined as “region of interest”. Then, the feature amount extraction unit 102 extracts the image feature amount of the affected part from the set “region of interest”. For example, when the affected area is an isolated nodule shadow, for example, the size of the affected area (long diameter / short diameter / average diameter diameter, area, etc.), the shape characteristics of the affected area (the ratio between the long and short diameters, the length of the contour line) The ratio of the average diameter, the fractal dimension of the contour line, etc.), the average density value of the affected area, the density distribution pattern of the affected area, etc. are extracted. In addition to this, various image feature amounts can be extracted.

  Then, the feature quantity extraction unit 102 transmits the image feature quantity extracted from the diagnosis target image to the storage unit 107.

<Step S103>
Subsequently, in step S103, the similar case search unit 104 first acquires the image feature amount of the diagnosis target image stored in the storage unit 107 in step S102. Next, the similar case search unit 104 performs similar case search processing for extracting case data having image feature amounts similar to the image feature amounts of the acquired diagnosis target image from the case database 210 (similar case search step).

FIG. 3 is a schematic diagram illustrating an example of contents of case data stored in the case database 210 illustrated in FIG. 1 according to the first embodiment of this invention.
As shown in FIG. 3, the case data stored in the case database 210 of the present embodiment includes identification information 310, clinical information 320, image information 330, diagnosis information 340, and diagnosis difficulty level information 350. It is. Specifically, the case data of this embodiment includes, for example, examination information of a subject including identification information 310, clinical information 320, image information 330, and diagnostic information 340, and diagnostic difficulty information 350 indicating the difficulty of medical practice. Are stored in association with each other.

Specifically, as the identification information 310, case ID (image data ID) information corresponding to the medical image to be diagnosed is set.
As the clinical information 320, information such as the ID, name, age, and sex of a patient (subject) that is a diagnosis target is set.
Further, as the image information 330, image data of medical images, photographing information, and information on image feature amounts extracted from the medical images are set. Here, as shooting information, information such as shooting date / time, shooting modality, shooting site, shooting conditions, and the like is set.
Further, as the diagnosis information 340, information on image findings and a definitive diagnosis name is set.
Further, as the diagnosis difficulty level information 350, a diagnosis difficulty level (1 to 10) indicating the difficulty level of the diagnostic action for the medical image is set. Here, in the example shown in FIG. 3, a value of 1 to 10 is used as the diagnostic difficulty. However, the value is not limited to this, and any numerical value representing the diagnostic difficulty such as 0.0 to 1.0 may be used. May be set.

Next, similar case search will be described.
For example, when the image feature quantity of the lung nodule shadow is obtained from the diagnosis target image, for example, the similar case search unit 104 first creates a feature quantity vector having the above-described image feature quantity as an element. Also, as shown in FIG. 3, since the image feature amount of each case data is included in the case database 210, for example, the similar case search unit 104 also creates a feature vector corresponding to each case data. Then, the similar case search unit 104 acquires, as a search result, case data having a high similarity to the diagnosis target image in the feature space having each element of the created feature quantity vector as an axis.

More specifically, the similar case retrieval unit 104, the feature vector F _i of the i-th case data stored in the feature vector F _t and case database 210 of the diagnostic target images, respectively, the following equations (1) and It is given by equation (2).

In the equations (1) and (2), m is the number of vector elements. At this time, the similar case search unit 104 gives the similarity S _i based on the Euclidean distance of the feature vector F _i of the i-th case data with respect to the feature quantity vector F _t of the diagnosis target image by the following equation (3). .

The similar case search unit 104 calculates the similarity S _i for all i satisfying 1 ≦ i ≦ N, where N is the total number of case data in the case database 210, and the value of the similarity S _i is Case data is acquired for k constants in descending order, and these are set as similar cases. In the present embodiment, for example, k = 3.

  Then, the similar case search unit 104 transmits k pieces of similar case data acquired by the similar case search process to the display recommendation degree calculation unit 105 and the storage unit 107.

<Step S104>
Subsequently, in step S <b> 104, the display recommendation level calculation unit 105 first acquires similar case data from the similar case search unit 104. Next, the display recommendation level calculation unit 105 calculates the display recommendation level P of the similar case based on the diagnosis difficulty level in the acquired similar case data.

  Then, the display recommendation level calculation unit 105 transmits the calculated display recommendation level P to the medical reference information search unit 106 and the storage unit 107.

<Step S105>
Subsequently, in step S <b> 105, the past image acquisition unit 103 acquires image information related to the patient's past image in the diagnosis target image from the medical image database 220. At this time, the past image acquisition unit 103 acquires the patient ID of the diagnosis target image input to the medical image acquisition unit 101, for example, and searches and acquires the image information of the past image using the patient ID as a key. Thereby, when image diagnosis is performed a plurality of times in the past for the same patient, the image information of the past image can be acquired. At this time, if the past image cannot be acquired by searching the medical image database 220, the past image acquisition unit 103 sets a flag F (which will be referred to as “first diagnosis flag”) F indicating “first diagnosis” to “True”. To "". Conversely, the past image acquisition unit 103 sets the initial diagnosis flag F to “False” when a past image can be acquired by searching the medical image database 220.

  Then, the past image acquisition unit 103 transmits the acquired image information of the past image and the value of the initial diagnosis flag F to the storage unit 107.

<Step S106>
Subsequently, in step S <b> 106, the medical reference information search unit 106 first acquires the display recommendation level P from the display recommendation level calculation unit 105 and the value of the initial diagnosis flag F from the storage unit 107. Then, the medical reference information search unit 106 determines whether or not the acquired value of the recommended display degree P is less than the first threshold T (P <T) and the initial diagnosis flag F is True (that is, initial diagnosis). Judging.

  Then, as a result of the determination in step S106, if the value of the display recommendation degree P is less than the first threshold T and the first visit is made, the process proceeds to step S107. On the other hand, as a result of the determination in step S106, if the value of the recommended display degree P is not less than the first threshold T or not the first visit, the process proceeds to step S108.

  Here, in the present embodiment, for example, the first threshold T = 0.5. The process of step S106 performs the medical reference information search process of step S107 only when the display recommendation degree P of the diagnosis target image is smaller than the first threshold T and the current image diagnosis is the first diagnosis. Means. This is because it is less necessary to present and refer to a similar case when diagnosis of an image to be diagnosed is easier than a certain criterion based on the first threshold T, and past images are This is because the display area having a margin is effectively used because it does not exist. In addition, even when diagnosis is easy, by presenting medical reference information (medical reference information) on the presumed disease, various knowledge such as the latest information on the disease can be used as future reference. It makes sense to present medical reference information.

<Step S107>
As described above, if the result of determination in step S106 is that the display recommendation level P is less than the first threshold T and this is the first visit, the process proceeds to step S107.
In step S107, the medical reference information search unit 106 acquires the image information of the diagnosis target image from the storage unit 107, and performs processing for searching the medical reference information of the disease from the medical information database 230.

  For example, when the feature amount extraction unit 102 extracts the “region of interest” of the lung nodule shadow and the image feature amount from the diagnosis target image, the medical reference information search unit 106 firstly extracts the “pulmonary nodule shadow” as the extracted information. Get information that it is a finding. Then, the medical reference information search unit 106 stores medical reference information related to a disease predicted from this finding (for example, primary lung cancer, metastatic lung tumor, benign tumor, etc.) in the medical information database 230. Search from within. Here, as described above, the medical information database 230 is a database in which medical information is stored. This medical information is, for example, basic information on the disease (onset site, cause of onset, symptom, diagnosis method, treatment method, precautions in diagnosis / treatment, etc.), medical information magazines, academic journal papers, and the like.

  The medical reference information search unit 106 transmits the acquired medical reference information to the storage unit 107.

<Step S108>
As described above, if the result of determination in step S106 is that the recommended display level P is not less than the first threshold value T or if it is not the first visit, the process proceeds to step S108. Moreover, also when the process of step S107 is complete | finished, it progresses to step S108.
In step S108, the display method determination unit 108 first stores, from the storage unit 107, image information of the diagnosis target image, image information of the past image, similar case data, medical reference information, the value of the initial diagnosis flag F, and the display recommendation level. Get the value of P. Next, the display method determination unit 108, based on the acquired value of the initial diagnosis flag F and the display recommendation level P, acquires the image information of the acquired diagnosis target image, the image information of the past image, the similar case data, and the medical reference information. A content display method M (M is any one of M ₁ to M ₆ described later) is determined. This step S108 constitutes a “display method determining step”.

  Then, the display method determination unit 108 displays the determined display method M information, the image information of the diagnosis target image acquired from the storage unit 107, the image information of the past image, the similar case data, and the medical reference information on the display unit 109. Send.

<Step S109>
Subsequently, in step S109, the display unit 109, first, the display method determining unit 108, (one of M ₆ from M ₁ to be described later) the value of the display method M, the image information of the diagnosis target image, past image Image information, similar case data, and medical reference information. Then, according to the acquired display method M, the display unit 109 displays, on the diagnosis screen of a monitor (not shown), diagnosis of image information of the acquired diagnosis target image, image information of past images, similar case data, medical reference information, and the like. Information is displayed (display step). Thus, the doctor diagnoses the patient who is the subject while referring to the diagnostic information displayed on the diagnostic screen of the monitor (not shown).

  When the process of step S109 ends, the process of the information processing method according to the first embodiment shown in FIG. 2 ends.

  Next, the display method M determined in step S108 will be described with reference to FIGS.

FIG. 4 shows the first embodiment of the present invention, and the display method M when the initial diagnosis flag F is True (that is, the initial diagnosis) and the recommended display level is less than the first threshold (P <T). ₁ is a schematic diagram illustrating an example of _1. FIG.

In the diagnosis screen 400 shown in FIG. 4, a diagnosis target image display area 410 is arranged on the left side, a medical reference information display area 420 is arranged on the right side, and a similar case display button 430 is arranged on the upper right side. The reason why the similar case is not presented on the diagnosis screen 400 is that, as described in step S106, when diagnosis of the diagnosis target image is easy, it is less necessary to present and refer to the similar case. It is. That is, in the display method M ₁ shown in FIG. 4, it is possible to prevent the feel troublesome to the physician to view the unnecessarily similar cases when diagnosis is easy. Furthermore, the reason why the medical reference information is presented on the right side of the diagnostic screen 400 is to effectively use the display area with sufficient margin as described in step S106. In the diagnosis screen 400, when the similar case display button 430 is pressed, the medical reference information display area 420 is replaced with a similar case, or a similar case display screen is newly displayed. it can. Thus, similar cases can be displayed only when the doctor feels necessary.

FIG. 5 shows the first embodiment of the present invention, and the display method M when the initial diagnosis flag F is True (that is, the initial diagnosis) and the recommended display level is equal to or higher than the first threshold (P ≧ T). ₂ is a schematic diagram showing an example of 2. FIG.

In the diagnosis screen 500 shown in FIG. 5, a diagnosis target image display area 510 is arranged on the left side, and a similar case display area 520 is arranged on the right side. Here, in the similar case display area 520 shown in FIG. 5, cases [1] to [3] are listed in descending order when the number of cases is k = 3. In the display method M ₂ shown in FIG. 5, the display recommendation degree P is a predetermined value or more, i.e. for diagnosis of the diagnosis target image is difficult, can be referred to the diagnostic act of physicians by presenting the similar cases. Furthermore, when diagnosis is difficult, a specific similar case may be more effective for diagnosis than medical information of a general disease. Therefore, in the display method M ₂ shown in FIG. The medical reference information display area 420 shown in FIG. 4 is replaced with a similar case display area 520 and displayed.

FIG. 6 shows the first embodiment of the present invention, and the display method when the initial diagnosis flag F is False (that is, other than the initial diagnosis) and the recommended display level is less than the first threshold (P <T). it is a schematic diagram showing an example of M _3.

In the diagnosis screen 600 shown in FIG. 6, a diagnosis target image display area 610 is arranged on the upper stage, a past image display area 620 is arranged on the lower stage, and a similar case display button 630 is arranged on the upper right side. In the display method M ₃ shown in FIG. 6, not displaying similar cases, not only preventing unnecessary presentation of similar cases and causing annoyance to the doctor, as in FIG. Since it is necessary to display the past image, it is possible to prevent the display area from being occupied unnecessarily. Further, in this diagnosis screen 600, when the similar case display button 630 is pressed, the number of diagnosis target images and past images is reduced and, for example, similar cases are displayed on the right side, or a similar case display screen is newly displayed. Can be displayed. Thereby, similar to the case of FIG. 4, similar cases can be displayed only when the doctor feels necessary.

FIG. 7 shows the first embodiment of the present invention, and the display method when the initial diagnosis flag F is False (that is, other than the initial diagnosis) and the recommended display level is equal to or higher than the first threshold (P ≧ T). it is a schematic diagram illustrating an example of M _4.

In the diagnosis screen 700 shown in FIG. 7, the diagnosis target image display area 710 is arranged in the upper part of the left area, the past image display area 720 is arranged in the lower part of the left area, and the similar case display area 730 is arranged on the right side. Has been. Here, in the similar case display area 730 shown in FIG. 7, as in the example shown in FIG. 5, the cases [1] to [3] are listed in descending order of similarity when the number of cases k = 3. It is displayed. In the display method M ₄ shown in FIG. 7, it is difficult diagnosis of the diagnosis target image, it can be referred to the doctor by presenting the similar cases. Furthermore, since there is a great need for past images even when diagnosis is difficult, both past images and similar cases are displayed and presented.

Further, in the above-described example, when displaying similar cases, an example in which medical images of similar cases are displayed as a list in one column as in the display method M ₂ shown in FIG. 5 and the display method M ₄ shown in FIG. 7 is shown. ing. However, the display method of similar cases is not limited to this. For example, a method of displaying similar cases in descending order of similarity for each disease classification can be applied. At this time, the display methods M ₂ and M ₄ are changed to the display methods for displaying each classification of the diseases, and the display methods are M ₅ and M ₆ , respectively. However, in this case, the similar case display area becomes larger than the display methods M ₂ and M ₄ , and the diagnosis target image display area and the past image display area become smaller correspondingly, so that the display recommendation degree P is particularly high. Take an example. Specifically, in the present embodiment, the display recommendation degree satisfies the second threshold or higher (P ≧ T ′) using the second threshold T ′ (T ′> T) larger than the first threshold T. In this case, these display methods M ₅ and M ₆ shall be applied. Here, for example, the second threshold T ′ = 0.8 is used.

FIG. 8 shows the first embodiment of the present invention, and the display method when the initial diagnosis flag F is True (that is, the initial diagnosis) and the recommended display level is equal to or greater than the second threshold (P ≧ T ′). it is a schematic diagram illustrating an example of M _5.

  In the diagnosis screen 800 shown in FIG. 8, a diagnosis target image display area 810 is arranged on the left side, and a similar case display area 820 is arranged on the right side. However, in the similar case display area 820 shown in FIG. 8, when the number of cases is k = 9, cases [1] to [9] are displayed in descending order of similarity for each disease classification.

Further, a description thereof will be omitted with regard display method M _6, since the display method similar case is the same. In this way, by displaying similar cases for each disease classification, it is possible to compare each set of similar cases belonging to the disease, and to present useful information to the doctor in determining the disease of the diagnosis target image. .

Next, the details of the similar case display recommendation degree calculation processing executed in step S104 of FIG. 2 will be described.
FIG. 9 is a flowchart illustrating an example of a detailed processing procedure in the similar case display recommendation degree calculation processing executed in step S <b> 104 of FIG. 2.

<Step S901>
When the process of step S104 of FIG. 2 is started, first, in step S901 of FIG. 9, the display recommendation degree calculation unit 105 sets the number of similar cases to k, and sets the current similar case to be processed to i. In this case, i = 1 is set. At this time, for example, it is assumed that the similar cases to be processed are arranged in descending order of similarity.

<Step S902>
Subsequently, in step S <b> 902, the display recommendation level calculation unit 105 acquires the value of the diagnostic difficulty in the case data related to the similar case i from the similar case data acquired from the similar case search unit 104. In this case, the diagnosis difficulty of acquired similar case i and D _i.

<Step S903>
Subsequently, in step S903, the display recommendation level calculation unit 105 calculates a reference difficulty level R _{i in} which the reference level of the similar case i with respect to the diagnosis target image is added to the diagnosis difficulty level D _{i in the} similar case i. By calculating the reference difficulty level R _i , the weight of the similar case when calculating the recommended recommendation degree P of the similar case can be changed according to how much the similar case is used as a reference. At this time, the reference difficulty R _i is a value obtained by weighting the reference difficulty of the similar case i with respect to the diagnosis target image with respect to the diagnosis difficulty D _i . Here, as the medical image of the similar case i is more similar to the diagnosis target image, the similar case i is considered to be a reference to the diagnosis target image. Therefore, the reference degree of the similar case i with respect to the diagnosis target image is defined as follows.
(Reference level of similar case i for diagnosis target image)
= (Similarity of similar case i to diagnosis target image)

Therefore, when the similarity of the similar case i to the diagnosis target image is S _i , the reference difficulty R _i is given by the following equation (4).

<Step S904>
Subsequently, in step S904, the display recommendation level calculation unit 105 sets a variable i indicating a similar case to be processed as i = i + 1. Thereby, the similar case i to be processed is changed and set.

<Step S905>
Subsequently, in step S905, the display recommendation degree calculation unit 105 compares the value of the similar case i to be processed with the value of the number k of similar cases, and determines whether i ≦ k.
As a result of this determination, if i ≦ k, the process returns to step S902, and the processes after step S902 are performed again. On the other hand, if i ≦ k is not satisfied (ie, i> k) as a result of the determination in step S905, the process proceeds to step S906.

<Step S906>
In step S906, the display recommendation level calculation unit 105 calculates the general display recommendation level _Pst of similar cases based on the reference difficulty R _i for each similar case i calculated in the processing up to step S905. Here, the “general display recommendation level” represents a degree of recommending similar case presentation for a diagnosis target image, which does not depend on a doctor who performs diagnosis. In the processing of step S906, the reference value of “diagnosis difficulty” of a similar case similar to the diagnosis target image corresponds to a general diagnosis difficulty for the diagnosis target image, and is referred to as “general display recommendation” of the similar case. It means to apply as "degree".

As a specific method of calculating the general display recommendation level P _st , for example, the reference difficulty level of the similar case having the largest reference difficulty level among the similar cases is applied as the general display recommendation level P _st of the similar case. . That is, the general display recommendation degree P _st can be given by the following equation (5).

This is because if there is a similar case with a high degree of reference difficulty among similar cases, the necessity of presenting it to the doctor as a similar case increases, that is, it is necessary to increase the general display recommendation level. Because it is. However, the method for calculating the general display recommendation degree P _st is not limited to the above-described method. For example, it is also possible to consider the reference degree of difficulty of the whole similar cases, for example, a general display recommendation degree P _st the average of the values of R _i. The general display recommendation degree P _{st at} this time can be given by the following equation (6).

<Step S907>
Subsequently, in step S907, the display recommendation level calculation unit 105 calculates a display recommendation level correction coefficient according to the skill level and expertise of the doctor who performs the diagnosis. This is because the difficulty evaluation of the same case changes depending on the skill level and expertise of the doctor, and the necessity of presenting the similar case to the doctor changes accordingly.

  First, in step S907, the display recommendation level calculation unit 105 acquires information on the skill level and expertise of the doctor who performs the diagnosis. Specifically, information on the skill level of image diagnosis of a doctor who performs diagnosis (for example, the number of diagnoses for each past diagnosis difficulty level) and information on expertise (for example, clinical departments / special modalities / specialized parts engaged) Get information about. Here, the number of diagnoses for each past diagnosis difficulty level is a value representing how many cases the doctor has made in the past for each diagnosis difficulty level associated with the case. The number of diagnoses for each past diagnosis difficulty level is, for example, extracted from the case data stored in the case database 210 only for cases diagnosed by the doctor, and appears for each diagnosis difficulty level associated with the case. It can be calculated by counting the frequency. Also, for example, when the information acquired is stored in advance as user information of the information processing apparatus 100-1 for each doctor on a magnetic disk or the like built in the information processing apparatus, and the doctor logs in to the information processing apparatus. It can also be obtained by loading the information.

  Next, an example in which the display recommendation level correction coefficient is calculated based on the acquired information on the skill level and expertise of the doctor will be described.

  FIG. 10 is a schematic diagram of the correction function u = f (x) showing an example of the relationship between the doctor's skill level x and the correction coefficient u, which is used when calculating the display recommendation level correction coefficient in step S907 of FIG. It is.

The skill level x is calculated by the following method, for example.
When the diagnosis difficulty level is expressed in 10 stages of 1 to 10, the number of diagnosis for each diagnosis difficulty level of the doctor is n _{1 to} n _{10 for} each stage. At this time, a value (referred to as “weighted diagnosis frequency”) N obtained by weighting the total diagnosis frequency with the diagnostic difficulty is obtained by the following equation (7).

Then, when the upper limit value of the number of times of weighted diagnosis is N _max , the skill level x is obtained by the following equation (8).

Here, for example, N _max = 10000. The proficiency level x is not limited to that based on the past number of diagnoses for each diagnosis difficulty level. For example, the skill level x may be calculated based on the years of experience of image diagnosis by a doctor. In this case, it can be calculated by replacing N and N _{max in} equation (8) with the doctor's years of experience and the upper limit of years of experience, respectively.

  For example, when the value of the skill level x thus obtained is equal to or less than the average value (here, x = 0.2), the value of the skill level x is assigned to the correction function f to cope with it. The value of the correction coefficient u to be calculated can be calculated (in this case, u = 1.3). This is because when the skill level x of the image diagnosis of the doctor is smaller than the average value, the same case tends to feel more difficult than the doctor having the average experience value due to lack of experience, and the correction coefficient u is 1. The reason why it becomes larger than 0.0 is to cope with such a situation.

  FIG. 11 is a correction showing an example of the relationship between the correction coefficient v and the degree of difference y between the attribute of the diagnosis target image and the doctor's expertise, which is used when calculating the correction coefficient for the display recommendation level in step S907 in FIG. It is a schematic diagram of the function v = g (y).

  The degree of difference y is set to 0 when all the items of specialty of the doctor (the engaged medical department / special modality / special site) match the attributes of the diagnosis target image, and vice versa. 0. In addition, when the attribute of the diagnosis target image and the specialty of the doctor are different for each item, an added value of the degree of difference is given for each item (however, the total of the added values is 1.0 at maximum).

Here, for example, it is assumed that the diagnosis target image is a chest X-ray CT image.
In this case, for example, if the medical department engaged by the doctor is a radiology department, the clinical departments to which the doctors belong coincide with each other, so the added value is set to 0. In addition, for example, if the medical department engaged in by the doctor is a medical department related to image diagnosis to some extent, such as internal medicine or mammary gland department, the addition value is set to 0.2, and the medical department engaged in by the doctor is not related to other image diagnosis. If it is a medical department, the added value is set to 0.4.
If the doctor's specialized modality is an X-ray CT image, the modalities match, so the addition value is set to zero. Further, for example, if the doctor's specialized modality is an X-ray image (mammography, etc.) or an image such as MRI that is slightly different from the X-ray CT image, the addition value is 0.15, and the doctor's specialized modality is other than that (image If the diagnosis is not performed, the added value is set to 0.3.
Further, if the doctor's specialized site is the chest, the site matches, so the added value is set to 0. Also, for example, if the doctor's specialized part is the abdomen, head, breast, etc., the focus of image diagnosis is different, so the added value is 0.15, and the doctor's specialized part is otherwise (no image diagnosis is performed) Etc.), the added value is set to 0.3.

  According to this example, when the diagnosis target image is a chest X-ray CT image, if the medical department engaged by the doctor is the radiology department, the specialized modality is the MRI, and the specialized part is the head, the sum of the added values is 0.3, that is, the degree of difference y = 0.3. By substituting the value of this difference y into the correction function g, the value of the corresponding correction coefficient v can be calculated (in this case, v = 1.2). This is because if there is a difference between the attributes of the image to be diagnosed and the doctor's expertise, the same case tends to feel more difficult than a doctor who has a perfect match due to lack of knowledge. The reason why v is larger than 1.0 is to cope with such a situation.

  In the present embodiment, it is assumed that the correction functions f and g are given in advance. These can be given, for example, by the following method.

First, discrete values obtained by dividing the horizontal axis (skill level x and dissimilarity y) in FIG. 10 and FIG. 11 at regular intervals are set as skill level a _j and dissimilarity b _j , respectively. However, 1 ≦ j ≦ n, where n is the number of divisions at regular intervals on the horizontal axis. At this time, with respect to the same case, an average value of the diagnostic difficulty evaluated by a plurality of doctors having attributes of skill level x = (average value) and difference level y = 0 is set as D _base, and a reference value of the diagnostic difficulty level And

Further, the average value of the diagnostic difficulty evaluated by a plurality of doctors having attributes of skill level x = a _j and difference level y = 0 is D1 _j , skill level x = (average value), and difference level y = b _j . Let D2 _{j be} the average value of the diagnostic difficulty evaluated by a plurality of doctors having attributes. Next, D1 _j and D2 _j are obtained for all values of j that satisfy 1 ≦ j ≦ n. Then, the discrete values (correction coefficients u _j and v _j ) on the vertical axis in FIGS. 10 and 11 are calculated by taking the ratio of each of the obtained D1 _j and D2 _j and D _base . The calculation formulas can be given by the following formulas (9) and (10), respectively.

The coordinate points (x _j , u _j ) and (y _j , v _j ) thus obtained are plotted on the graph for all values of j that satisfy 1 ≦ j ≦ n. Later, by performing regression analysis on the combination of these points, the correction functions f and g can be obtained. Here, the least square method is applied as a method of regression analysis.

  Here, the description returns to FIG. 9 again.

<Step S908>
When the process of step S907 ends, the process proceeds to step S908.
In step S908, the display recommendation calculation unit 105, a general display recommendation degree P _st calculated in step S906, and corrected using a correction coefficient u and v of the display recommendation degree calculated in step S907, similar case Display recommendation degree P is calculated. This means that a value obtained by correcting the general display recommendation level P _st of the similar case according to the attribute of the doctor, which does not depend on the attribute of the doctor, becomes the display recommendation level P of the similar case for the doctor. At this time, the diagnosis difficulty level of the case data stored in the case database 210, which is a base for calculating the general display recommendation level _Pst , is standardized in advance, and corresponds to the skill level and expertise of the doctor. There shall be no variation.

Specifically, by taking the product of the general display recommendation degree P _st of the similar case, which is a standardized value, and the correction coefficients u and v, the display of the specific similar case corresponding to the doctor who performs the diagnostic action The recommendation degree P is calculated. That is, the display recommendation degree P can be calculated by the following equation (11).

  However, P in the formula (11) is a value in the range of 1 ≦ P ≦ 10, and when the calculation result of the display recommendation degree P according to the formula (11) is above or below this range, It shall be changed to a value or a lower limit value.

  By performing the processes in steps S901 to S908 described above, the display recommendation degree calculation process for similar cases executed in step S104 in FIG. 2 is performed.

As described above, in the information processing apparatus 100-1 according to the first embodiment, the display recommendation degree P of the similar case is calculated based on the diagnosis difficulty level in the case data of the similar case, and the display recommendation degree The display method of similar cases at the time of diagnosis is determined according to P.
According to this configuration, it is possible to present appropriate diagnostic information according to the necessity of presenting similar cases to a doctor.
Moreover, the display recommendation degree P with high reliability can be set by calculating the display recommendation degree P of the similar case based on the diagnosis difficulty level of the similar case given appropriately in advance.
Furthermore, by calculating the display recommendation degree P in consideration of the skill level of the doctor and the proficiency level of the modality to be interpreted, it is possible to display according to the necessity of presenting similar cases, and to improve the diagnosis efficiency. be able to.

(Second Embodiment)
Next, a second embodiment of the present invention will be described.

  In the information processing apparatus 100-1 according to the first embodiment described above, the display recommendation degree P of the similar case is calculated based on the diagnosis difficulty level in the similar case. On the other hand, in the information processing apparatus 100-2 according to the second embodiment, the similarity is based on both the diagnosis difficulty level in a similar case and the disease determination difficulty level obtained by image analysis of the diagnosis target image. The case is different from the first embodiment in that the case recommendation degree P is calculated. Here, the “disease discrimination difficulty level” represents the degree of difficulty of disease discrimination performed by image analysis of the diagnosis target image. Thereby, even when the similarity of the similar case with respect to the diagnosis target image is low, that is, the reference degree of the similar case is low, the reliability of the display recommendation degree P is increased by increasing the weight of the disease determination difficulty level, which is another index. Deterioration can be prevented.

  FIG. 12 is a schematic diagram illustrating an example of a schematic configuration of an information processing system 10-2 according to the second embodiment of the present invention. In FIG. 12, the same components as those shown in FIG. 1 are denoted by the same reference numerals, and detailed description thereof will be omitted in the following description.

  As shown in FIG. 12, the information processing system 10-2 includes an information processing apparatus 100-2, a case database 210, a medical image database 220, and a medical information database 230.

  Here, the information processing apparatus 100-2 illustrated in FIG. 12 is different from the configuration of the information processing apparatus 100-1 illustrated in FIG. 1 in that a disease estimation unit 1201 is added. The configuration of the information processing apparatus 100-1 shown in FIG.

  The disease estimation unit 1201 performs a disease estimation process based on the image feature amount extracted by the feature amount extraction unit 102.

Next, the operation of each component of the information processing apparatus 100-2 according to the second embodiment will be described using the flowchart of FIG.
FIG. 13 is a flowchart showing an example of a processing procedure of an information processing method by the information processing apparatus 100-2 according to the second embodiment of the present invention. In FIG. 13, the same processing steps as those shown in FIG. 2 are denoted by the same step numbers, and detailed description thereof will be omitted in the following description.

  In the flowchart shown in FIG. 13, first, processing similar to that in step S101 shown in FIG. 2 (medical image (diagnosis target image) acquisition processing) is performed.

<Step S201>
Subsequently, in step S <b> 201, the feature amount extraction unit 102 first acquires image information of the diagnosis target image from the medical image acquisition unit 101. Then, the feature amount extraction unit 102 extracts an image feature amount from a medical image (diagnosis target image) included in the image information acquired from the medical image acquisition unit 101. The specific processing content is the same as that in step S102 shown in FIG.

  Then, the feature quantity extraction unit 102 transmits the image feature quantity extracted from the diagnosis target image to the storage unit 107 and the disease estimation unit 1201.

<Step S202>
Subsequently, in step S <b> 202, the disease estimation unit 1201 first acquires the image feature amount of the diagnosis target image from the feature amount extraction unit 102. Then, the disease estimation unit 1201 performs a disease estimation process in which the processing result is obtained as a continuous value based on the acquired image feature amount.

  Specifically, for example, by applying the Bayesian network technique described in Non-Patent Document 1, an estimated value (predicted value) of a disease can be calculated as a probability. For example, when the diagnosis target image is a chest X-ray CT image and an image feature amount of a lung nodule shadow is extracted, the estimated values of the diseases are set in descending order of probability, “primary lung cancer”: 0.5, “ It can be calculated as “metastatic lung tumor”: 0.4, “benign tumor”: 0.1.

  Then, the disease estimation unit 1201 transmits information on the calculated estimated value of the disease to the storage unit 107.

  Then, the same process (similar case search process) as step S103 shown in FIG. 2 is performed.

<Step S203>
Subsequently, in step S <b> 203, the display recommendation level calculation unit 105 first acquires similar case data from the similar case search unit 104, and acquires information on an estimated value of a disease from the storage unit 107. Next, the display recommendation level calculation unit 105 calculates the “disease determination difficulty” of the diagnosis target image from the estimated value (predicted value) of the disease acquired from the storage unit 107. Thereafter, the display recommendation level calculation unit 105 displays the similar case display recommendation level based on both the diagnosis difficulty level in the similar case data acquired from the similar case search unit 104 and the calculated disease determination difficulty level of the diagnosis target image. P is calculated.

  Then, the display recommendation level calculation unit 105 transmits the calculated display recommendation level P to the medical reference information search unit 106 and the storage unit 107.

  Thereafter, the processing similar to steps S105 to S109 shown in FIG. 2 is performed, and the processing of the information processing method according to the second embodiment shown in FIG. 13 ends.

Next, details of the display recommendation degree calculation process for similar cases executed in step S203 of FIG. 13 will be described.
FIG. 14 is a flowchart showing an example of a detailed processing procedure in the similar case display recommendation degree calculation processing executed in step S203 of FIG. In FIG. 14, process steps similar to those shown in FIG. 9 are given the same step numbers, and detailed description thereof will be omitted in the following description.

  In the flowchart shown in FIG. 14, first, processing similar to steps S <b> 901 to S <b> 906 shown in FIG. 9 is performed.

<Step S1401>
Subsequently, in step S <b> 1401, the display recommendation degree calculation unit 105 first acquires information on an estimated value (predicted value) of a disease from the storage unit 107. Then, the display recommendation level calculation unit 105 calculates the “disease determination difficulty” of the diagnosis target image based on the information on the estimated value of the disease acquired from the storage unit 107.

Specifically, in this embodiment, depending on the difference between the values of probabilities in the two diseases are most likely estimate of the disease (predicted value), and calculates the disease determination difficulty D _x of the diagnosis target image. This is because it is considered difficult to discriminate a disease if there is no difference in the estimated (predicted) disease probability. Here, when the probability of the two diseases having the highest probability of the estimated value of the disease is p and q (0 ≦ p ≦ 1, 0 ≦ q ≦ 1, p ≧ q), respectively, the disease of the diagnosis target image The discrimination difficulty level D _x is calculated from the following equation (12).

However, in the equation (12), when D _x <1, it is assumed that D _x = 1. Using equation (12), when the probabilities of two diseases are p = 0.5 and q = 0.4, respectively, D _x = 9, and when the difference between the probabilities of the two diseases is small, it is difficult to determine the disease. The degree D _x takes a high value. Conversely, when the probabilities of two diseases are p = 0.9 and q = 0.1, respectively, D _x = 2, and when the difference between the probabilities of the two diseases is large, the disease discrimination difficulty D _x is small. Takes a value.

<Step S1402>
Subsequently, in step S1402, the display recommendation calculation unit 105 changes based on the general display recommendation degree P _st similar case calculated in step S906, the disease determination difficulty D _x of the diagnosis target image calculated in step S1401 Perform the process.

Details of the processing in step S1402 will be described below.
In the first embodiment, when calculating the reference difficulty R _i of the similar case in step S903, the similarity of the similar case is used as the reference degree of the similar case with respect to the diagnosis target image. In the second embodiment, the display recommendation level calculation unit 105 calculates the similarity S _type as a reference level that represents the entire similar case. Further, the similarity of the similar case having the highest similarity among the similar cases is applied as the similarity S _type . This can be given by the following equation (13), for example.

Next, the display recommendation level calculation unit 105 weights between the general display recommendation level P _st calculated in step 906 and the disease determination difficulty level D _x calculated in step S1401 by the value of the similarity S _type. The general display recommendation degree P ′ _st after the change is obtained. This calculation can be performed using, for example, the following equation (14).

(14) In the equation, if a large value of similarity S _type, the proportion of the general display recommendation degree P _st occupying the calculation of the general display recommendation degree P _'st after the change is large. On the other hand, when the value of the similarity S _type is small, the ratio of the disease determination difficulty D _x in the calculation of the general display recommendation P ′ _st after the change is large.

When the processing of step S1402 is completed, processing similar to steps S907 and S908 shown in FIG. 9 is performed, and similar case display recommendation degree calculation processing executed in step S203 of FIG. 13 is performed.
At this time, for the general display recommendation degree P _st used in the process of step S907, in this embodiment, the changed general display recommendation degree P ′ _st calculated in step S1402 is applied.

  As described above, according to the second embodiment, even when the similarity of the similar case with respect to the diagnosis target image is low, that is, when the reference degree of the similar case is low, the weight of the disease determination difficulty level that is another index By increasing the value, it is possible to prevent the reliability of the display recommendation level from being lowered.

(Third embodiment)
Next, a third embodiment of the present invention will be described.

  In the information processing apparatus 100-1 according to the first embodiment described above, the display method of the similar case is determined based on the diagnostic difficulty in the case data of the similar case. On the other hand, in the information processing apparatus according to the third embodiment, the display method of the similar case is determined based on the “degree of difficulty in judgment at the time of treatment” that represents the difficulty of judgment at the time of treatment. Different from the first embodiment.

  At this time, there are difficulty in judging the treatment method and difficulty in judging the timing of treatment as difficulty in judgment at the time of treatment. When it is difficult to judge the treatment method, it is possible to refer to the treatment method of similar cases, and when it is difficult to judge the treatment timing, it is possible to refer to the treatment timing of similar cases. is there.

  Therefore, in the third embodiment, the “degree of difficulty in judgment at the time of treatment” is expressed by two parameters, that is, the difficulty of judgment of the treatment method and the difficulty of judgment of the timing of treatment. Determine the display method. In this way, by determining the display method of similar cases based on the difficulty of judgment at the time of treatment, at the conference etc. where examination of the treatment policy by the attending physician or the examination of the treatment policy is made, to the doctor who comes from difficulty of judgment at the time of treatment Appropriate information presentation according to the necessity of presentation of similar cases can be realized.

  The schematic configuration of the information processing system including the information processing apparatus according to the third embodiment is the same as that of the information processing system 10-1 including the information processing apparatus 100-1 according to the first embodiment shown in FIG. Therefore, the description thereof is omitted.

Next, the operation of each component of the information processing apparatus 100-1 according to the third embodiment will be described using the flowchart of FIG.
FIG. 15 is a flowchart illustrating an example of a processing procedure of an information processing method performed by the information processing apparatus 100-1 according to the third embodiment of the present invention. In FIG. 15, process steps similar to the process steps shown in FIG. 2 are given the same step numbers, and detailed description thereof will be omitted in the following description.

In the flowchart shown in FIG. 15, first, processing similar to that in steps S101 and S102 shown in FIG. 2 is performed.
At this time, in step S101, the medical image acquisition unit 101 acquires, from the medical image database 220, image information related to the medical image of the subject that is the target of the medical practice designated by the doctor. Here, a medical image of a subject that is the subject of a medical practice is referred to as a “medical target image”. In step 102, the feature amount extraction unit 102 extracts an image feature amount from the medical treatment target image, and transmits the extracted image feature amount of the medical treatment target image to the storage unit 107.

<Step S301>
Subsequently, in step S301, the similar case search unit 104 first acquires the image feature amount of the medical treatment target image stored in the storage unit 107 in step S102. Next, the similar case search unit 104 performs similar case search processing for extracting case data having an image feature amount similar to the image feature amount of the acquired medical treatment target image from the case database 210.

  FIG. 16 is a schematic diagram showing an example of the contents of case data stored in the case database 210 shown in FIG. 1 according to the third embodiment of the present invention. In FIG. 16, the same components as those shown in FIG. 3 are denoted by the same reference numerals, and detailed description thereof will be omitted in the following description.

  As shown in FIG. 16, the case data stored in the case database 210 of the present embodiment includes identification information 310, clinical information 320, image information 330, treatment information 1610, and treatment determination difficulty information 1620. It is included. That is, in the case data in the third embodiment shown in FIG. 16, the treatment information 1610 is used instead of the diagnosis information 340 with respect to the case data in the first embodiment shown in FIG. Instead of the information, the treatment difficulty information 1620 is used. Specifically, the case data of the present embodiment includes, for example, examination information of a subject including identification information 310, clinical information 320, image information 330, and treatment information 1610, and a determination level at the time of treatment indicating the difficulty level of a medical practice. Information 1620 is associated and stored.

Specifically, as the treatment information 1610, information on a treatment method and treatment timing (treatment time) is set.
In addition, the treatment determination difficulty level information 1620 includes a parameter α (value of 1 to 10) related to the determination difficulty level of the treatment method and a parameter β (value of 1 to 10) related to the determination difficulty level of the treatment timing. ) Are set. Note that the example shown in FIG. 16 shows an example in which both the parameter α related to the difficulty of determination of the treatment method and the parameter β related to the difficulty of determination of the treatment timing are applied as the treatment determination difficulty information 1620. However, either one can be applied.

  Details of the similar case search process in step 301 are the same as step S103 shown in FIG. 2 of the first embodiment, and a description thereof will be omitted.

  Then, the similar case search unit 104 transmits k pieces of similar case data acquired by the similar case search process to the display recommendation degree calculation unit 105 and the storage unit 107.

<Step S302>
Subsequently, in step S <b> 302, the display recommendation degree calculation unit 105 first acquires similar case data from the similar case search unit 104. Next, the display recommendation level calculation unit 105 calculates a display recommendation level P for similar cases based on the degree of difficulty in determination at the time of treatment in the acquired similar case data.

First, in the processing of step 302, when the difficulty level of the similar case is determined as D = {α, β}, the display recommendation degree of the similar case calculated based on this is set to P = {α _p, β _p }. And This means that the display recommendation degree P includes a parameter α _p representing the display recommendation degree related to the treatment method for similar cases and a parameter β _p representing the display recommendation degree related to the treatment timing for similar cases. This process is applied by replacing the location of the diagnosis difficulty level and the display recommendation level with these parameters in the flowchart (FIG. 9) relating to the details of the display recommendation level calculation process for similar cases in the first embodiment. Since it can, the description is omitted.

Then, the display recommendation level calculation unit 105 transmits the calculated display recommendation level P = {α _p, β _p } to the medical reference information search unit 106 and the storage unit 107.

  Subsequently, the same processing (past image acquisition processing) as step S105 shown in FIG. 2 is performed.

<Step S303>
Subsequently, in step S <b> 303, the medical reference information search unit 106 first acquires the display recommendation level P = {α _p, β _p } from the display recommendation level calculation unit and the value of the initial diagnosis flag F from the storage unit 107. Then, the medical reference information search unit 106 determines that the values of the parameters α _p and β _{p in} the acquired display recommendation degree P are less than the first threshold T (α _p <T and β _p <T), and the first visit flag It is determined whether or not F is True (that is, the first visit).

As a result of the determination in step S303, if both the values of the parameters α _p and β _p in the display recommendation degree P are less than the first threshold T and the first visit, the process proceeds to step S107. On the other hand, as a result of the determination in step S303, if at least one of the values of the parameters α _p and β _p in the display recommendation degree P is not less than the first threshold T or not the first visit, the process proceeds to step S304.

Here, in the present embodiment, for example, the first threshold value T = 0.5, as in the first embodiment. The processing in step S303 is performed only in the case where the values of the parameters α _p and β _p in the display recommendation degree P of the medical object image are smaller than the first threshold T and the current medical treatment is the first medical examination. This means that reference information search processing is performed.

<Step S304>
As described above, as a result of the determination in step S303, if at least one of the values of the parameters α _p and β _p in the display recommendation degree P is not less than the first threshold T or is not the first visit, the process proceeds to step S304. . Moreover, also when the process of step S107 is complete | finished, it progresses to step S304.
In step S304, the display method determination unit 108 first stores, from the storage unit 107, image information of the medical object image, image information of the past image, similar case data, medical reference information, the value of the initial diagnosis flag F, and the display recommendation level. The value of P = {α _p, β _p } is acquired. Next, the display method determination unit 108, based on the acquired value of the initial medical examination flag F and the value of the recommended display degree P, the acquired image information of the medical object image, the image information of the past image, the similar case data, and the medical reference information A content display method M (M is any one of M ₇ to M ₁₁ described later) is determined.

  Then, the display method determination unit 108 displays the determined display method M information, the image information of the medical treatment target image acquired from the storage unit 107, the image information of the past image, the similar case data, and the medical reference information on the display unit 109. Send.

<Step S305>
Subsequently, in step S305, the display unit 109, first, the display method determining unit 108, (one of M ₁₁ from M ₇ to be described later) the value of the display method M, the image information of the medical care subject image, past image Image information, similar case data, and medical reference information. Then, according to the acquired display method M, the display unit 109 displays medical information such as acquired image information of the medical object image, image information of the past image, similar case data, and medical reference information on a medical screen of a monitor (not shown). Display information. Thus, the doctor performs medical treatment such as treatment of the patient as the subject while referring to the medical information displayed on the medical screen of the monitor (not shown).

  When the process of step S305 ends, the process of the information processing method according to the third embodiment illustrated in FIG. 15 ends.

  Next, the display method M determined in step S304 will be described with reference to FIGS.

First, when the parameters α _p and β _p of the display recommendation degree P are α _p <T and β _p <T with respect to the first threshold T, both the treatment method and the treatment timing of similar cases are displayed. Therefore, the display method M ₇ that does not present similar cases is adopted. The display method M ₇ is in the first embodiment, a display method similar to the effect and a display method M ₁ which does not present a similar case, the description of the detailed drawings will be omitted.

FIG. 17 shows a third embodiment of the present invention, where the initial diagnosis flag F is False (that is, other than the initial diagnosis), and the parameter α _p of the display recommendation degree P is α _p ≧ T and α _p > β _p is a schematic view showing an example of a display method M ₈ when the. That is, the example shown in FIG. 17 is a display method M ₈ when the recommended display level (α _p ) related to the treatment method is greater than or equal to the first threshold T and larger than the recommended display level (β _p ) related to the treatment timing. It is an example.

  In the medical treatment screen 1700 shown in FIG. 17, the medical object image display area 1710 is arranged at the upper part of the left area, the past image display area 1720 is arranged at the lower part of the left area, and the similar case display area 1730 is arranged on the right side. Has been. In the similar case display area 1730 shown in FIG. 17, cases [1] to [9] are displayed in descending order of similarity for each type of treatment method. In this way, when the display recommendation degree P of the treatment method is particularly high, by displaying similar cases for each type of treatment method, it is possible to compare each set of similar cases belonging to the treatment method, and in the future of the image to be treated Information useful for determining the treatment method can be presented to the doctor.

FIG. 18 shows a third embodiment of the present invention, where the initial diagnosis flag F is False (that is, other than the initial diagnosis), and the parameter β _p of the display recommendation degree P is β _p ≧ T and β _p ≧ α _p is a schematic view showing an example of a display method M ₉ when the. That is, the example shown in FIG. 18 is a display when the recommended display level (β _p ) related to the treatment timing is equal to or higher than the first threshold T and higher than the recommended display level related to the treatment method (β _p ≧ α _p ). it is an example of how M _9.

  In the medical treatment screen 1800 shown in FIG. 18, the medical object image display area 1810 is arranged at the upper part of the left area, the past image display area 1820 is arranged at the lower part of the left area, and the similar case display area 1830 is arranged on the right side. Has been. In the similar case display area 1830 shown in FIG. 18, time series information of treatment progress regarding the patients of each similar case (case [1] to case [3]) is displayed. In this way, when the recommended display timing P of treatment timing is particularly high, by displaying time series information of treatment progress in each similar case, information useful for determining the future treatment timing of the image to be treated can be obtained. Can be presented to a doctor.

Furthermore, the display method when the initial diagnosis flag F is True (that is, the initial diagnosis) is only to display the medical object image without displaying the past image in the display area on the left side in the display methods M ₈ and M ₉ described above. Detailed description thereof will be omitted. These display methods are referred to as display methods M ₁₀ and M ₁₁ in order.

As described above, in the information processing apparatus 100-1 according to the third embodiment, the display recommendation degree P of the similar case is calculated based on the degree of difficulty in determination at the time of treatment in the case data of the similar case, and this display The display method of similar cases at the time of treatment is determined according to the recommendation degree P.
According to this configuration, it is possible to present appropriate medical information according to the necessity of presenting a similar case when a doctor examines a treatment policy.
In addition, by calculating the display recommendation level P of the similar case based on the difficulty level of the similar case given in advance appropriately, it is possible to set the display recommendation level P with high reliability.

(Fourth embodiment)
The present invention can take an embodiment as, for example, an information processing system, an information processing apparatus, an information processing method, a program, or a storage medium. Specifically, the present invention may be applied to an information processing system including a plurality of devices, or may be applied to an information processing device including a single device.

  FIG. 19 is a schematic diagram illustrating an example of a hardware configuration of an information processing system 10-4 according to the fourth embodiment of the present invention. In FIG. 19, the same components as those shown in FIG. 1 are denoted by the same reference numerals, and detailed description thereof will be omitted in the following description.

  As illustrated in FIG. 19, the information processing system 10-4 includes an information processing apparatus 100-4, a case database 210, a medical image database 220, a medical information database 230, and a network 240.

  The information processing apparatus 100-4 includes a central processing unit (CPU) 141, a main memory 142, a magnetic disk 143, a display memory 144, a monitor 145, a mouse 146, a keyboard 147, and a common bus 148.

  The CPU 141 mainly controls the operation of each component of the information processing apparatus 100-4 and controls the operation of the information processing apparatus 100-4 in an integrated manner. For example, the CPU 141 communicates with the case database 210, the medical image database 220, and the medical information database 230 via the network 240 by executing a program stored in the main memory 142. Further, for example, the CPU 141 executes various programs such as overall control of the information processing apparatus 100-4 by executing a program stored in the main memory 142.

  The main memory 142 stores a program executed by the CPU 141 and provides a work area when the CPU 141 executes the program.

  The magnetic disk 143 stores an operating system (OS), peripheral device driver, various application software, and work data generated or used by various application software.

  The display memory 144 temporarily stores display data for display on the monitor 145.

  The monitor 145 includes, for example, a CRT monitor, a liquid crystal monitor, and the like, and displays various images and various information based on display data in the display memory 144 under the control of the CPU 141. The execution status and execution result of the program executed by the CPU 141 are displayed on the monitor 145 as necessary.

  The mouse 146 and the keyboard 147 are used for a pointing input and a character input by the user, respectively. A user (doctor or the like) inputs various commands (instructions / commands) to the information processing apparatus 100-4 by operating the mouse 146 and the keyboard 147.

  The common bus 148 connects the constituent elements of the information processing apparatus 100-4 so that they can communicate with each other, and connects the information processing apparatus 100-4 and the network 240 so that they can communicate with each other.

  The network 240 connects the information processing apparatus 100-4 to the case database 210, the medical image database 220, and the medical information database 230 so that they can communicate with each other. The network 240 is configured by, for example, a local area network (LAN) using Ethernet (registered trademark), the Internet, or the like. Instead of the network 240, an interface such as USB or IEEE1394 may be applied.

  Here, the schematic configuration (functional configuration) of the information processing apparatus 100-1 shown in FIG. 1, the schematic configuration (functional configuration) of the information processing apparatus 100-2 shown in FIG. 12, and the information processing apparatus 100-4 shown in FIG. An example of a correspondence relationship with the hardware configuration will be described.

For example, from the CPU 141, the program stored in the main memory 142, the mouse 146, and the keyboard 147 in FIG. 19, the medical image acquisition unit 101, the past image acquisition unit 103, the similar case search unit 104, and the medical reference in FIGS. An information search unit 106 is configured.
Further, for example, the feature amount extraction unit 102, the display recommendation level calculation unit 105, and the display method determination unit 108 in FIG. 1 and FIG. 12 are configured from the programs stored in the CPU 141 and the main memory 142 in FIG. The disease estimation unit 1201 is configured.
Further, for example, the storage unit 107 includes the main memory 142 or the magnetic disk 143 in FIG.
Further, for example, the display unit 109 in FIG. 19 includes the CPU 141 in FIG. 19, the program stored in the main memory 142, the display memory 144, the monitor 145, the mouse 146, and the keyboard 147.

(Other embodiments)
The present invention can also be realized by executing the following processing.
That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, or the like) of the system or apparatus reads the program. It is a process to be executed.
This program and a computer-readable recording medium storing the program are included in the present invention.

  Note that each of the embodiments of the present invention described above is merely an example of implementation in carrying out the present invention, and the technical scope of the present invention should not be construed as being limited thereto. Is. That is, the present invention can be implemented in various forms without departing from the technical idea or the main features thereof.

10-1: Information processing system, 100-1: Information processing apparatus, 101: Medical image acquisition unit, 102: Feature quantity extraction unit, 103: Past image acquisition unit, 104: Similar case search unit, 105: Display recommendation calculation Part: 106: medical reference information search part, 107: storage part, 108: display method determination part, 109: display part, 210: case database, 220: medical image database, 230: medical information database

Claims (16)

Be one that stores a plurality of case data, inspection of the case data storage means associates the difficulty of interventions determined for each relevant case data is included in each case data, diagnosis 療対 elephant image a similar case search means for performing at least one obtains search for similar case data is case data which includes inspection information to similar information and s,
An information processing apparatus comprising: an acquisition unit configured to acquire information on a difficulty level of a medical practice in the medical image based on information on a difficulty level of the medical practice in the similar case data . The information processing apparatus according to claim 1, wherein the difficulty level of the medical practice includes at least one of a diagnostic difficulty level based on examination information of a subject and a determination level at the time of treatment.   The information processing apparatus according to claim 2, wherein the determination difficulty level at the time of treatment is one or both of a determination difficulty level of a treatment method and a determination difficulty level of a treatment timing. Based on the difficulty of the intervention in the retrieved the similar case data in the similar case search means, a display recommendation degree calculation means for calculating a display recommendation degree of the similar case data,
A display method determining means for determining a method of displaying the similar case data with pre Symbol Display recommendation degree
The information processing apparatus according to claim 1 , further comprising:   The display recommendation level calculation means calculates the display recommendation level of the similar case data using a value obtained by weighting the difficulty level of the medical practice in the similar case data based on the similarity level of the similar case data. The information processing apparatus according to claim 4, wherein the information processing apparatus calculates the information processing apparatus.   The display recommendation degree calculation means uses the value obtained by correcting the difficulty level of the medical practice in the similar case data according to at least one of the skill level and the expertise of a doctor who performs the medical practice. The information processing apparatus according to claim 4, wherein the display recommendation degree of data is calculated.   The display method determination means determines a display method that does not display the similar case data when the display recommendation level is smaller than a first threshold value, and the display recommendation level is equal to or higher than the first threshold value. 5. The information processing apparatus according to claim 4, wherein a display method for displaying the similar case data is determined. The difficulty level of the medical practice is a diagnostic difficulty level based on the examination information of the subject ,
The display method determining means determines the display method to display the similar case data for each disease classification when the recommended degree of display is equal to or greater than a second threshold value that is greater than the first threshold value. The information processing apparatus according to claim 7. The degree of difficulty of the medical practice is the degree of difficulty in judging at the time of treatment based on the examination information of the subject ,
The display recommendation level calculated by the display recommendation level calculation means is composed of a display recommendation level related to a treatment method and a display recommendation level related to a treatment timing.
The display method determining means is configured to display the similar case for each type of treatment method when the recommended display level related to the treatment method is greater than or equal to the first threshold value and greater than the recommended display level related to the treatment timing. The information processing apparatus according to claim 7, wherein a display method for displaying data is determined. The degree of difficulty of the medical practice is the degree of difficulty in judging at the time of treatment based on the examination information of the subject ,
The display recommendation level calculated by the display recommendation level calculation means is composed of a display recommendation level related to a treatment method and a display recommendation level related to a treatment timing.
The display method determination means, when the recommended display level related to the treatment timing is equal to or higher than the first threshold and equal to or higher than the recommended display level related to the therapeutic method, a time series of treatment progress in the similar case data The information processing apparatus according to claim 7, wherein a display method for displaying information is determined. The recommendability calculating means further the diagnosis 療対 be those analyzes inspection information elephant image to calculate the disease determination difficulty, the at retrieved the similar case data in the similar case search means 11. The information processing apparatus according to claim 4, wherein a display recommendation level of the similar case data is calculated using a difficulty level of medical practice and the disease determination difficulty level. Display method determining means for determining a display method of the similar case data based on the difficulty level of the medical practice in the similar case data;
Display means for displaying the similar case data according to the display method determined by the display method determination means;
The information processing apparatus according to claim 1, further comprising: Case data storage means for storing a plurality of case data, each case data includes information on the image feature amount and information on the difficulty level of the medical practice determined for each case data, When performing a search for acquiring at least one similar case data which is case data similar to the medical image, the information on the image characteristic amount in the medical image is acquired, and the image feature in the acquired medical image Similar case retrieval means for retrieving case data including image feature amount information similar to the information of the case as the similar case data,
Acquisition means for acquiring information on the degree of difficulty of the medical practice in the medical image based on information on the degree of difficulty of the medical practice in the similar case data;
An information processing apparatus comprising: Be one that stores a plurality of case data, inspection of the case data storage means associates the difficulty of interventions determined for each relevant case data is included in each case data, diagnosis 療対 elephant image a similar case retrieval step of performing at least one obtains search for similar case data is case data which includes inspection information to similar information and s,
An information processing method comprising: an acquisition step of acquiring information on a difficulty level of a medical practice in the medical image based on information on a difficulty level of the medical practice in the similar case data . Be one that stores a plurality of case data, inspection of the case data storage means associates the difficulty of interventions determined for each relevant case data is included in each case data, diagnosis 療対 elephant image a similar case retrieval step of performing at least one obtains search for similar case data is case data which includes inspection information to similar information and s,
The program for making a computer perform the acquisition step which acquires the information of the difficulty level of the medical practice in the said medical treatment object image based on the information of the difficulty level of the medical practice in the said similar case data . A computer-readable storage medium storing the program according to claim 15 .

Images