JP2852866B2 - Computer-aided image diagnosis learning support method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer-aided image diagnosis learning method for enabling a learner to self-learn image diagnosis using a database-based image for diagnosis and diagnostic information relating to the image. It is about.

[0002]

2. Description of the Related Art A conventional technique related to computer-aided image diagnosis learning is disclosed in Japanese Patent Application No. 197164/1993 (title: "Computer-aided Image Diagnosis Supporting Apparatus and Method") filed by the present applicant. There is technology. The outline of this technology is as follows. The image captured by the image capturing apparatus and displayed on the image display means, and the diagnostic information of the image including the disease name created by the diagnostic information creating means according to the pre-recorded standardized finding group item table, common to the The image and the diagnostic information are constructed by a database construction method comprising the steps of: providing an information symbol; and recording the image in image information storage means and recording the diagnostic information in diagnostic information storage means. Retrieving and displaying, from the image information storage means and / or the diagnostic information storage means, a relevant image of the image having the same finding symbol as the finding symbol in the information symbol of the image and / or related diagnostic information including a disease name, respectively. Computer-aided image diagnosis learning support method using a database capable of A step of learner selects the learning menu and diagnostic information; information symbol formation, to impart on the basis of the learning menus and diagnostic information are selected,
A step of searching and displaying from the database an image and / or diagnostic information having the same finding symbol as the finding symbol in the information symbol, and a step of selecting a description of an abnormal finding of the displayed image and / or diagnostic information. And further comprising: selecting image display conditions; and performing image processing according to the selected conditions.
This is a computer-aided image diagnosis learning support method.

According to the above prior art invention, a large number of images obtained in clinical image diagnosis are stored in a database together with diagnostic information created by an image diagnostician based on his or her experience, and the image and the image desired by the learner are stored. When the diagnostic information is selected, the image and its related images and the diagnostic information related thereto, and the explanation of the abnormal finding are further searched and displayed by the learner's further selection, so that the learner needs direct guidance from a specialist. Instead, I was able to learn diagnostic imaging techniques.

However, in this prior art,
It is possible to search and display images and their related images and diagnostic information related to them, and to learn by selecting explanations about abnormal findings related to them, but to present learners with questions and let them answer, There was no active means of image diagnosis learning support for evaluation and scoring, and so to say, only static and passive learning methods.

[0005]

SUMMARY OF THE INVENTION In order to solve the above-mentioned drawbacks of the prior art, the present invention utilizes a database constructed by the above-mentioned prior art to enable more effective learning support. That is, when you search for a learning image, a question about the image is automatically created, and when the learner answers the question, the answer is evaluated and graded, an image corresponding to the incorrect answer is displayed, and a response to the missing answer is displayed. The computer provides automatic and active learning support by computer at various points, such as displaying correct answers and images, so that the conventional technology can be changed from static and passive learning to dynamic and active learning. It is intended to make a qualitative change.

[0006]

The novel means employed by the present invention to solve the above problems is as follows.
Note that the image diagnosis support method according to the present invention uses a common information symbol for the image captured by the image capturing apparatus and the diagnostic information of the present image including the disease name created according to the standardized finding group item table. Using the database of the prior art, constructed by a database construction method comprising the steps of providing an image and recording diagnostic information of the image in the diagnostic information storage means while recording the image in the image information storage means. .

[0007] According to this database construction method, images newly photographed and displayed on the image display means are similarly processed.
According to the findings group item table, the diagnostic information of the image is stored with a common information symbol attached thereto, and the database is expanded by repeating this process, so that the image and all its related images and their diagnostic information are stored in the database. Search and display can be performed via the same finding symbol as the finding symbol in those information symbols.

[0008] A first diagnostic imaging support method according to the present invention comprises:
The image captured by the image capturing apparatus and displayed on the image display means, and the diagnostic information of the image including the disease name created by the diagnostic information creating means according to the pre-recorded standardized finding group item table, common to the The image and the diagnostic information are constructed by a database construction method comprising the steps of: providing an information symbol; and recording the image in image information storage means and recording the diagnostic information in diagnostic information storage means. Retrieving and displaying, from the image information storage means and / or the diagnostic information storage means, a relevant image of the image having the same finding symbol as the finding symbol in the information symbol of the image and / or related diagnostic information including a disease name, respectively. Computer-aided image diagnosis learning support method using a database capable of When the learner selects a desired learning image, the learning image is retrieved and displayed from the image information storage unit, and a question related to the learning image is created from the finding group item table and the diagnosis information. The learning image is displayed together with the learning image, and when the learner answers the question, the correctness of the answer is determined and displayed based on the diagnosis information.

[0009] A second image diagnosis support method according to the present invention comprises:
The first image diagnosis support method is further characterized in that the question includes a plurality of answer options.

A third diagnostic imaging support method according to the present invention comprises:
The first or second image diagnosis support method is further characterized in that the question comprises a plurality of questions, and the correct answer rate of the learner's answer can be calculated and displayed.

A fourth diagnostic imaging support method according to the present invention comprises:
The first, second, or third image diagnosis support method is further characterized in that an image corresponding to an incorrect answer can be searched and displayed.

[0012] A fifth diagnostic imaging support method according to the present invention comprises:
The learning support method according to any one of the first to fourth aspects, wherein a correct answer and an image corresponding to the missing answer can be searched and displayed.

[0013] A sixth image diagnosis support method according to the present invention comprises:
The learning support method according to any one of the first to fifth aspects, wherein a search for a learning image can be performed based on a disease name, a clinical symptom, a site, a patient age, and / or a patient gender.

[0014] A seventh image diagnosis support method according to the present invention comprises:
The learning support method according to any one of the first to sixth aspects, wherein a specific part can be designated and displayed on an image.

An eighth image diagnosis support method according to the present invention comprises:
The learning support method according to any one of the first to seventh aspects, wherein an image display condition can be changed.

A ninth diagnostic imaging support method according to the present invention comprises:
The learning support method according to any one of the second to eighth aspects, wherein an answer can be selected by clicking on a display screen.

According to a first aspect of the present invention, there is provided the learning support method according to any one of the first to ninth aspects, wherein the question is a disease name, a normal finding, an abnormal finding, or a clinical symptom. , Part and / or part detail designation.

[0018] A first image diagnosis support method according to the present invention is the learning support method according to any one of the first to first ○, wherein it is possible to display guidance items for a learner. And

[0019]

The doctor automatically creates a question for a large number of images created by a doctor at the site of clinical image diagnosis and various findings about the images, and the images and the questions are displayed on the computer display. Responds, the accuracy of the answer, the correct answer rate,
Since the correct answer to the wrong answer or the missing answer and the related guidance are displayed, the learner can learn the standard image reading technique for various images.

In addition, since each answer is evaluated and the evaluation of the whole answer is displayed as a correct answer rate, the labor required for the instructing doctor to score is omitted, and both the learner and the instructing doctor can immediately confirm the learning result. become. further,
Incorrect or missing answers from individual learners are displayed with images corresponding to individual learner answers,
Not only will you gain a broader understanding and knowledge of the causes, but also the correction of incorrect answers, and you will be able to learn at your own pace according to your level of understanding. The learning efficiency of the learner is much higher than in the case of taking the course.

[0021] The search of the learning image also includes the disease name, clinical symptoms, site,
Because the subject can be selected from a wide range, such as patient age and / or gender, it is possible to select the subject of interpretation that suits the individual learner, and it is also possible to specify and display a specific part on the screen. Learning in a situation closer to the site can be performed while viewing the image. Furthermore, since image display conditions can be changed and answers can be selected by clicking on the display screen, more detailed research on the case can be performed, and quick screen operation can be performed. In addition, you can ask questions up to the designation of disease name, normal findings, abnormal findings, clinical symptoms, site, site details,
The learner can efficiently absorb a wide range of empirical knowledge acquired by specialists in the field of clinical image diagnosis while viewing images.

Further, since the guidance items for the learner's answer can be displayed, the user can immediately proceed to the examination of the learned matter without waiting for the advice of a specialist after the answer. Therefore, according to the plan of the learner, it is possible to smoothly shift to the subsequent learning.

[0023]

An embodiment of the present invention will be described with reference to the drawings.

The present invention makes use of the prior art database described above (hereinafter referred to as the "database of the present application"). The database is simply X-ray,
Image diagnostic images such as CT, ultrasound, and MRI are stored in the image information storage means, and various diagnostic information including disease names obtained by a specialist from a theoretical and clinical experience are stored in the diagnostic information storage means. Is stored in

The diagnostic information is not arbitrarily created by a specialist, but is created in accordance with a standardized finding group item table similarly stored in the present database. In the finding group item table, parts, symptoms, and the like are hierarchically classified under a large classification according to the type of image, such as X-ray, CT, ultrasound, and MRI. Diagnostic information is created. Since each item of each category is provided with a finding symbol, the diagnostic information is expressed as a combination of such finding symbols, that is, an information symbol, and the information symbol is stored in the diagnostic information storage means and the corresponding image is displayed. The same information symbol is also added to the image information and stored in the image information storage means. In this way, by storing the same information symbol composed of a collection of finding symbols in accordance with the finding group item table to both the image and its diagnostic information, any image and its diagnostic information, and the same The related image having the finding symbol and the diagnostic information thereof can be easily and quickly searched and displayed.

FIG. 1 shows an example of a hierarchical classification in a finding group item table in the present database, and FIG. 2 shows an example of an input screen for the sub-classification and sub-classification. An example of the specific method of classification and information symbol assignment is as follows

As is clear from FIG. 1, the large classification is based on various types of images such as "X-ray images", "CT images", "ultrasonic images", and "MRI images". The middle classification is classified into, for example, “brain” and “head and neck” based on the imaging site of various images. The small classification is based on the interpretation subject for each imaging region, and is classified as, for example, “cerebroventricular” or “subarachnoid space”. Furthermore, the sub-classification is classified based on the finding items for each interpretation subject, for example, as "local expansion" or "mild dilation of the cerebral sulcus and cerebral cistern". ,
The classification is based on the anatomical site where the finding for each finding item such as "lateral ventricle" or "Sylvian fissure" exists.

Each item of each category has an arbitrary code for generating an information symbol later, for example, 0001 for a large category “X-ray image”, 0001 for a middle category “brain”, and 0001 for a small category “subarachnoid space”. "Is 0002, subdivision" local dilation "is 1240," mild dilation of cerebral sulci and cistern "
1180, subclass “Sylvian fissure” 1651,
It is attached in advance. In addition, in the case of a thing existing on the left and right, such as “Sylvian fissure”, a sign and / or a sign corresponding to “right” or “left”, for example, “R” or “ L "and / or 00
It is preferable that 01 or 0002 can be added.

The image photographing apparatus includes an X-ray image photographing apparatus, a CT image photographing apparatus, an ultrasonic image photographing apparatus and an MR image photographing apparatus.
The present invention is not limited to the I image photographing apparatus, and may be used if a new image photographing apparatus is adopted. Further, it is also possible to use an image reading device (scanner) for the image photographing device. By using the image reading device, not only an actual diagnosis example but also an image introduced in a technical book or the like can be made into a database as described later.

FIG. 2 is an example of an input screen for sub-classification and sub-classification. In the actual input, instead of inputting the above-mentioned symbols preliminarily attached to the items of the sub-classification (left column) and the sub-classification (right column), numbers simplified for input shown in FIG. It is desirable to enter. In addition, it is desirable to input “R” (right) or “L” (left) in front of the sub-classification number for those existing on the left and right, such as “Sylvian fissure”.

The image is displayed on the image display means after being photographed by the image photographing apparatus, and is used for interpretation by a doctor. According to the finding group item table (FIG. 1), for example, the doctor classifies the findings obtained by the interpretation of the image in the diagnostic information creating means according to the finding group item table (FIG. 1). Enter or select “subarachnoid space”. The sub-classification and the sub-classification are input based on the input screen (FIG. 2). For example, the sub-classification is “ENLARGED 9”.
LOCALLY OF. . . (Local extension of ...) ", L11 in the sub-classification, that is," LEFT, SYL
VIAN FISSURE (left Sylvian fissure) ”and sub-class 3“ SLIGHTLY ENLARGED ”
CEREBRAL SULCI AND CISTE
RN (medium dilation of the cerebral sulcus and cerebral cistern) "and the disease name" cerebral infarction
Enter

The diagnostic information input by the doctor is sent to the information symbol assigning means after being confirmed by the doctor by the confirming means. In the information symbol assigning means, a code pre-attached to each of the above items and an arbitrary unique number pre-attached to the disease name, for example, 7650 in the case of "cerebral infarction", is combined with one information symbol, For example, 0001 (major category) -0001 (medium category) -0002 (small category)-
{(1240 (subclassification) +0002,1651 (subclassification))-(1180 (subclassification))} = 7650 (disease name)
Is formed. The information symbol thus formed is added to both the image and the diagnostic information relating to the image including the disease name. Accordingly, one common information symbol is assigned to one image and its diagnostic information. The information symbols are not limited to those formed by the above-mentioned combination, but may be arbitrarily formed.

Each finding input by the doctor is included in the information symbol as a finding symbol. For example, the notation for extension of the left Sylvian fissure is 0001-0001-00.
02- (1240 + 0002,1651), the finding symbol for mild dilation of the cerebral sulci and cistern is 0001-
0001-0002- (1180). That is,
A finding symbol is obtained by dividing one information symbol for each sub / sub-classification.

The image is stored in the image information storage means together with the above information symbols, for example, as digitized digital image information having 256 × 256 pixels and a density difference from 0 to (2 ¹² -1). . The diagnostic information and the disease name of the image are stored in the diagnostic information storage means as the information symbol. The database is constructed by such storage and accumulation.

If it is not possible to input a disease name when interpreting an image, after diagnosing the disease name, the image and its diagnostic information are searched from the database and displayed on the screen, and then the disease name is input. it can.

The learner performs image diagnosis learning using a computer using the present database and the program according to the method of the present invention. The concept over time of the method of the present invention is shown in the flowchart of FIG. The learner can select a learning case from the disease name, clinical symptoms, site, patient age or patient gender, or any combination thereof, and search and display a learning image on the computer screen. A process in which a learner selects and learns a CT image learning case from a definite disease name, for example, is shown in FIGS. 4 to 12.

When the learner clicks on "from a definite disease name" on the learning menu screen, the definite disease names classified in the form of superordinate concept, subordinate concept, and further subordinate concepts are sequentially displayed in "item selection". 3 shown in FIG. 4 and other figures.
The four-digit numbers such as 000, 3600, and 3610 are finding symbols in the finding group item table in the present database, and are not displayed on the screen. The “selected case list” indicates a situation where a case corresponding to an item displayed in “item selection” or an item selected therefrom is stored in the present case database.

When the learner selects, for example, "ID: 001 (age 52, female)" from the "selected case list", the corresponding learning image and other image information in a necessary range, and finally the diagnostic information are displayed. The information is retrieved from the image information storage means and the diagnostic information storage means of the database and displayed in the "case window" of FIG. Nine slice images of CT are displayed, and these are combined into one. When you click the “Examination method” button in the case window,
The inspection method data of the image is displayed in the “inspection method window” of FIG.

When a learning image is selected and displayed, questions and answer options are automatically created and displayed from the finding group item table of the present database and the diagnostic information storage means.
An incorrect answer portion of the answer options is appropriately selected from related items in the finding group item table, and a correct answer portion is selected from diagnostic information corresponding to the image. The answer option is created for each selection of the interpretation subject in the "interpretation-subject window" of FIG. When, for example, "ventricular" is designated in the "interpretation-subject window" of FIG. 7, answer options are displayed in the "interpretation-finding window" of FIG. Of these options, if you select the answer "the ventricle is locally enlarged", the display will say that "the specific part of the ventricle is expanded". In the "interpretation-part window" of FIG. 9, the answer is given for the enlarged specific part. Here, for example, if the site is specified as “left” and “lateral ventricle triangular part”, the answer “the left ventricular triangular part is locally enlarged” is completed and displayed.

When one answer is completed in this way,
Returning to the "interpretation-finding window" in FIG. 8, the next finding is selected, and the part details are designated to complete the next answer. It is desirable to similarly complete the answer for all the interpretation subjects displayed in the "interpretation-subject window" of FIG.

When all the answers are completed, the answer is shown in FIG.
Are displayed together in the “interpretation-result display window”. Answers are evaluated and scored against diagnostic information already read into the computer. The score is the percentage of correct answers, that is, (the number of correct answers) / (the number of correct answers) × 1
It is calculated by the formula of 00.

The evaluation for all answers and individual answers and the score for all answers are shown in FIG.
Answer / Evaluation Window ". Here, the result of learning all the interpretation subjects displayed in the "interpretation-subject window" of FIG. 7 is shown. "Evaluation marks for findings"
"CR" is a correct answer (correc
t), and “FP” of the “evaluation mark of the specific part” indicates a wrong answer for the specific part, that is, excessive reading (f).
al positive). FP
In the case of, "interpretation-answer / evaluation window" in FIG.
When the "True Positive" button is clicked, the image corresponding to the wrong answer and the diagnostic information thereof are searched and displayed from the image information storage means and the diagnostic information storage means of the present database.

The word "FN" described at the bottom of the figure is a missing answer, that is, an overlooked answer (false).
negative). In the case of FN, it is instructed to return to the interpretation subject related to the overlooked answer, for example, "subarachnoid space" and to perform another interpretation, and "R
When the "OI hint" button is clicked, the missing answer is instructed. When the "ROI hint" button is clicked, the correct answer, the image and the diagnostic information corresponding to the missing answer are stored in the image information storage means and the diagnostic information storage of the present database. Searched by means and displayed.

When an image corresponding to an incorrect answer or a missing correct answer is displayed on the screen, or has already been displayed, the operation of each of the above-mentioned buttons causes, for example, an abnormal finding in the image. A specific part (ROI, that is, a region of interest) to be seen is designated and displayed by a frame or an arrow on the screen.

Further, in the "finding display" window shown in FIG. 12, the correct answers to all the questions and the comments on the learned images prepared in advance by the specialist are displayed.

In any of the images, if necessary,
The display conditions of the CT image can be appropriately adjusted, and a specific part on the image can be designated by a frame or an arrow.

In this embodiment, the case where the CT image is selected from the confirmed disease name “malignant neoplasm, intracranial tumor” and the age and sex of the patient has been described. However, other disease names and sites can be selected. It can also be directly selected from clinical symptoms and sites and other finding items or any combination thereof. Further, not only a CT image but also an X-ray image, an ultrasonic image, an MRI image, and any other image recorded in the present database can be similarly selected and learned.

[0048]

As described above, according to the present invention, it is possible for a specialist to make a database of, as it were, living knowledge obtained at the site of clinical image diagnosis and to make questions, answers, and evaluations on a computer screen. The learner can learn standard image interpretation techniques and judgment skills for the image while directly touching the image at his own pace. In addition, according to the present invention, a learner can select and search and display a learning image based on not only the disease name but also clinical symptoms, specific site, patient gender / age, etc. become able to. Furthermore, since the selection of the answer to the question can be performed by clicking, a quick response is possible, and an effect is provided that the learner is provided with a more efficient learning environment. Furthermore, images corresponding to incorrect answers or missing answers and their diagnostic information can be easily searched and displayed, so that the learner's understanding and knowledge can be further increased and deepened.

On the other hand, since the instructing doctor can convert the knowledge acquired at the clinical image diagnosis site into a database and automate the creation of questions and the evaluation of answers, the time and labor spent on instruction can be significantly reduced, and the image diagnosis can be greatly reduced. The education itself can be standardized, and the effect of easily maintaining content consistency with past education contents is produced.

[Brief description of the drawings]

Fig. 1 An example of classification in the construction of the database

FIG. 2 is an example of an input screen for sub-classification and sub-classification in the construction of the database

FIG. 3 is a flowchart showing the concept of the present invention.

FIG. 4 is a “learning menu” in the embodiment of the present invention.
Display of

FIG. 5 is a display of a “case window” in the embodiment of the present invention.

FIG. 6 is a display of an “inspection method window” in the embodiment of the present invention.

FIG. 7 shows the display of “interpretation-subject window” in the embodiment of the present invention.

FIG. 8 is a view showing a “interpretation-finding window” in the embodiment of the present invention.

FIG. 9 shows the display of “interpretation-part window” in the embodiment of the present invention.

FIG. 10 is a view showing an “interpretation-result display window” in the embodiment of the present invention.

FIG. 11 shows the display of “interpretation-answer / evaluation window” in the embodiment of the present invention.

FIG. 12 is a view showing “finding display” in the embodiment of the present invention.

Continuation of the front page (56) References JP-A-5-257418 (JP, A) JP-A-5-165394 (JP, A) Ito, Yamakita, Shimoi, Sasaki, coins, "In the image diagnosis learning information database Modeling of Teaching Materials ”Information Processing Society of Japan Research Report (93-DBS-96-2), Vol. 96, 1993 (Heisei 5-10-29) Ito, Hoshihara, Futami, Fuji, Coin, Saegusa "Databases as Components of Learning Information Systems-A Proposal for an Image Diagnosis Learning System-" IPSJ Research Report (92-D) BS-89-6), Vol. 92, No. 61, 1992 (Heisei 4-7-22), p. 51-55 (58) Field surveyed (Int. Cl. ⁶ , DB name) G06F 19/00 G09B 7/00 JICST file (JOIS)

Claims (11)

(57) [Claims] 1. An image photographed by an image photographing apparatus and displayed on an image display means, and diagnostic information of the image including a disease name created by a diagnostic information creating means according to a standardized finding group item table recorded in advance. And a database construction method comprising: a step of providing a common information symbol; and a step of recording the diagnostic information in the diagnostic information storage means while recording the image in the image information storage means. The diagnostic information and / or the diagnostic information including the relevant image and / or the disease name of the image having the same finding symbol in the information symbol of the image as the image information storage means and / or the diagnostic information storage, respectively. Computer-aided image diagnosis learning support method using a database that can be searched and displayed by means When a learner selects a desired learning image, the learning image is retrieved and displayed from the image information storage means, and a question related to the learning image is entered in the finding group item table and the diagnosis. A computer-aided image diagnosis learning support method which is created from information and displayed together with the learning image, and when a learner answers a question, the correctness of the answer is determined and displayed based on the diagnosis information. 2. The learning support method according to claim 1, wherein the question includes a plurality of answer options. 3. The learning support method according to claim 1, wherein the question comprises a plurality of questions, and the correct answer rate of the answer of the learner can be calculated and displayed. 4. The learning support method according to claim 1, wherein an image corresponding to an incorrect answer can be searched and displayed. 5. The learning support method according to claim 1, wherein a correct answer and an image corresponding to the missing answer can be searched and displayed. 6. The learning support method according to claim 1, wherein a search for a learning image can be performed by a disease name, a clinical symptom, a site, a patient age and / or a patient gender. 7. The learning support method according to claim 1, wherein a specific part can be designated and displayed on the image. 8. The learning support method according to claim 1, wherein a display condition of an image can be changed. 9. The learning support method according to claim 2, wherein an answer can be selected by clicking on a display screen. 10. The question is related to a disease name, a normal finding, an abnormal finding, a clinical symptom, a site and / or a site specification.
The learning support method according to claim 1. 11. The learning support method according to claim 1, wherein guidance items for a learner can be displayed.